Professional Documents
Culture Documents
Overview:
The Chemical and Manufacturing subcommittee of the Industrial Wastewater Committee chose a goal in
2019 of building a living depository of committee recommended papers related to Chemical and
Manufacturing (C&M) from 2012 to present. The goal was to select the best papers related to C&M, as per
committee members, that could be used by new members and outside parties, to get a broad understanding
of the best and current techniques for treating waste water for chemical and manufacturing applications.
First, we gathered the abstracts of all WEF accepted papers and presentations from 2012 to 2018. We
assigned two (2) years to each active team member and asked them to choose 2 - 5 potential papers to
review in detail. From this exercise, 27 potential papers were reviewed. We assigned three (3) papers to
each team member to give an up or down vote for inclusion in the compendium. From this, we chose 11
papers to be included.
This compendium is available for review in AccessWater.org. The papers range in subject matter from
pretreatment processes (MBR, flocculation, etc.) to heavy metal removal, and general membrane treatment
and desalination.
7. Biological Treatment of Coke Plant Wastewater with Activated Sludge MBR Technology
Proceeding paper from WEFTEC 2015
https://www.accesswater.org/publications/-277753/biological-treatment-of-coke-plant-
wastewater-with-activated-sludge-mbr-technology
9. Unconventional Bench and Pilot Approaches for Chemical Adhesive Wastewater Treatment
Proceeding paper from WEFTEC 2014
https://www.accesswater.org/publications/-282370/unconventional-bench-and-pilot-approaches-
for-chemical-adhesive-wastewater-treatment
10. One Automotive Manufacturer: Three Membrane Applications for Wastewater Pretreatment
and Reuse
Proceeding paper from WEFTEC 2014
https://www.accesswater.org/publications/-282879/one-automotive-manufacturer--three-
membrane-applications-for-wastewater-pretreatment-and-reuse
11. A Combined Biological and Advanced Oxidation Process for the Treatment of Wastewaters
from the Microelectronics Industry
Proceeding paper from WEFTEC 2014
https://www.accesswater.org/publications/-282876/a-combined-biological-and-advanced-
oxidation-process-for-the-treatment-of-wastewaters-from-the-microelectronics-industry
The team plans to review 2019 and 2020 papers for candidates and going forward will review the
previous year for additional papers.
WEFTEC 2018
ABSTRACT
INTRODUCTION
Perrigo is a leading global healthcare supplier. Perrigo has a facility in Yerucham, Israel, that
is dedicated to the formulation of creams, ointments, gels and pharmaceutical foams (topical
formulation) as well as formulation of tablets (enteral formulation). Production activities run
from Sunday to Thursday, 5 days per week, 8.5 hours per day, and are stopped during the
weekend, from Friday to Saturday. On Fridays, the production lines are cleaned.
Being a global supplier, the production operates in compliance with the current good
manufacturing practice (CGMP) guidelines, a regulatory standard enforced by the FDA to
ensure pharmaceutical quality. This standard includes, amongst others, maintaining a clean
manufacturing area, which implies that complying with the CGMP regulation can result in an
increased wastewater output with detergents content.
The wastewater from Perrigo plant used to be treated together with wastewater from a
neighboring industry. The only treatment of the wastewater before being discharged to the
public sewer was a dissolved air flotation (DAF) unit. In 2014 the two production plants were
separated and a new dedicated wastewater treatment plant was required for the Perrigo
formulations plant. A decision was made to install a Moving Bed Biological Reactor
(MBBR) system to treat the wastewater. The MBBR technology utilizes free-floating plastic
biomass carriers with a high specific surface area of attached biomass (biofilm). The MBBR
technology was selected because it can robustly cope with fluctuations in flow and organic
load and because the MBBR technology has a short recovery time after toxic events. Other
advantages of the MBBR technology are the simplicity of operation and maintenance, the
possibility of future expansion of the system by addition of biomass carriers, and small
footprint when compared to conventional biological processes.
This paper summarizes the operations of Perrigo Yerucham wastewater treatment plant and
describes the initial process design parameters and the plant configuration, the start-up phase
and the results from one year of operation. The wastewater from the plant is characterized by
high fluctuations in COD concentration and varying flow. Nevertheless, the effluent quality is
not affected by these conditions. The experience in Perrigo showed that the MBBR is an
adequate and robust technology, which can cope well with the challenges of pharmaceutical
wastewater treatment.
The main wastewater parameters that were considered as design basis are summarized in
Table 1. The average design wastewater flow is 75 m3/d, the maximum design COD
concentration is 3,600 mg/l after an equalization tank and the maximum design TSS
concentration is 100 mg/l. The effluent of the treatment plant is being discharged into the
public sewage system with the required COD effluent concentration of 800 mg/l and TSS
concentration of 400 mg/l.
Figure 1 shows the plant configuration to achieve the required effluent quality.
The wastewater flows after the fine screen to an equalization tank of 27 hours. In addition to
the equalization tank, two emergency tanks were installed. pH of the raw wastewater can
have a wide range, depending on the produced pharmaceutical and cleaning agents.
Therefore, the pH is controlled before the equalization tank. In case of extreme pH values, a
part of the wastewater can be diverted into the emergency tanks to avoid extreme pH
downstream.
The temperature of the raw wastewater varies between 25 and 50°C. In order to enter the
MBBR reactor, the wastewater has to be maintained below 35°C. Therefore, the wastewater
flows from the equalization tank to a cooling system that cools down the wastewater if
necessary. Since there is a lack of nutrients in the wastewater, Urea and Phosphoric acid are
being dosed before entering the MBBR for biological growth.
The biological reactor includes two stages of MBBR for COD removal. The two reactors
work in series, each reactor has a 50% carrier fill ratio and the total design HRT is 25.6
hours. An antifoam dosing system is installed to avoid foaming in the reactors. Foaming in
the biological reactor can be caused by surfactants in the wastewater. The surfactants in the
wastewater originate on one hand from the cleaning agents used to clean the production line
between batches. On the other hand it can originate from products that are manufactured in
the plant such as shampoo. After the biological reactor a DAF unit is installed for secondary
solid liquid separation. Sludge treatment includes a sludge tank and a centrifuge.
Given that the treatment plant is near the production plant, an odor treatment system is
installed to avoid odor emissions. The odor treatment system collects air from the
equalization and emergency tanks, the biological reactors, the DAF unit, the sludge tank and
centrifuge.
Startup
Operation of the Perrigo Yerucham wastewater treatment plant started by the end of
December 2014. The site view of the plant is shown in Figure 2.
After only three weeks of operation, the system was acclimated and the start-up phase was
completed. Table 2 presents inlet and effluent COD concentrations measured during the last
two weeks of the start-up. It can be seen that the measured effluent values are within the
required effluent quality of 800 mg/l COD.
Biofilm formation
After the start-up was finalized, fully developed biofilm was observed on the carriers. In
Figure 3 the carriers from the first and second MBBR reactor are shown. It can be seen that
the biofilm on the carriers is thicker in the first reactor than in the second one. The higher the
loading rate on the carriers, the more compact the biofilm produced (Ødegaard, 1999). Since
most of the organic load is removed in the first MBBR reactor, the second stage is low loaded
and only a thin biofilm is developed.
Operation results
Figure 4 presents the COD inlet and outlet concentration in the course of one year of
operation (August 2016 – July 2017). Values are shown in two different locations in the
plant. The inlet samples were collected after the equalization tank, before entering the MBBR
reactor, and the outlet samples after the secondary DAF. Also shown in the graph are the
design inlet quality and the required effluent quality. The required COD effluent quality was
achieved during the whole year, except for a few days in September 2016. The incident in
September was due to an exceptional peak of inlet COD concentration of 8,570 mg/l, which
was 2.38 times the inlet design basis of 3,600 mg/l (Figure 4, circled in red). This high COD
value was the result of a one-time cleaning action in the production line. Notably, the results
show that the system recovered fast from this extreme event. There were other occasions in
which inlet COD concentration exceeded the design basis and in all of those cases the
required effluent specifications were still met (Figure 4, circled in green).
In Figure 5 the TSS concentration during one year of operation is shown. TSS values are
shown at inlet to the MBBR, at outlet of the MBBR and at outlet of the DAF.
The TSS inlet concentration varies between 5 and 30 mg/l, which is relatively low compared
to the design inlet value of 100 mg/l.
The biomass to be separated after the MBBR, includes the inlet TSS and the surplus biomass
as a result of sludge yield. It can be seen that TSS concentration after the MBBR reactor
shows an increasing trend around May 2017. An increase in TSS concentration after the
MBBR reactor would normally be attributed to either higher TSS concentration in the
influent or higher COD load, leading to higher biomass production during the biological
process. No increase in inlet TSS nor increase in COD load during this period was observed.
A possible explanation could be that the additional TSS concentration is coming from another
stream that enters the reactor. One possible stream is the supernatant stream from the
centrifuge that is being re-diverted to the MBBR reactor. In cases where the centrifuge does
not work properly the supernatants may contain high amount of suspended solids. Another
source for the higher TSS could be the washing water from the DAF unit, which also can
contain high concentration of suspended solids and is being diverted to the MBBR reactor.
Further sampling and measurements are necessary in order to establish the source of the
increasing TSS concentrations at outlet of the MBBR reactor.
Although the TSS after the MBBR increased, the required TSS effluent quality is achieved
during the whole year.
To provide a better picture of the weekly variation of COD, the daily inlet concentration
during one typical month of operation (July 2017) is presented in Figure 6. The figure
additionally shows the effluent COD concentration from the MBBR reactor and the
secondary DAF as well as the average daily flow.
Figure 6: COD inlet/ outlet concentration and daily flow, July 2017
The daily inlet flow to the treatment plant during this period varies between 37 m3/d and
84 m3/d. During the week, flow decreases on Fridays and the lowest value is measured on
Saturdays. This correlates with the working regime of the Perrigo manufacturing plant. Since
the production is stopped from Fridays to Saturdays, less flow enters the treatment plant
during weekends. Yet, flow does not stop completely and it consists of effluent from cleaning
processes of the production line and of RO effluent from production.
It can be seen that there are high fluctuations in the COD inlet concentration. The COD inlet
concentration varies between 485 mg/l and 4,852 mg/l. The values measured on Sundays are
the lowest values during the week and reflect the diluted wastewater concentration after the
weekend. Despite the high variation in inlet COD concentration, the daily measured effluent
COD values from the MBBR reactor do not show any fluctuations and are consistently within
the required effluent concentration.
The wastewater trend during the week is also visible in conductivity values. Figure 7 shows
the conductivity at inlet to the reactor and outlet of the DAF during the same month (July
2017). Conductivity varies between 635 and 1,740 µS/cm. It is visible that also the lowest
conductivity values are on Sundays. This indicates the dilution of the wastewater during the
weekend with cleaning water and RO effluent from the production which contains lower
conductivity values than the wastewater. The conductivity values at outlet of the DAF are in
the range of the inlet values. Fluctuations in the outlet values are attributed to chemicals
dosing during the process such as pH correction and coagulant dosing.
It has been shown that the MBBR system recovers rapidly after organic peak loads that
exceed the design values.
In addition, the MBBR system acclimated quickly after initiation of operation and start-up
was finalized after three weeks.
Altogether, these results show that the installed system performed robustly and that the
MBBR technology is well suited to cope with the challenges of pharmaceutical wastewater
treatment.
REFERENCES
Ødegaard H. (1999). The Moving Bed Biofilm reactor. Water environmental engineering and
reuse of water 575314: 205-305.
ABSTRACT
BACKGROUND
Let us look at why we need coagulant and flocculant additives in the first place. We can
calculate the settling rate of a small sphere in a viscous fluid by applying Stoke’s Law,
represented in figure 1.
squared (r2, m2), the density difference between the sphere (ρ, kg/m3 ) and the fluid (ρ o , kg/m3),
and is inversely proportional to the viscosity of the fluid (η, kg/m/s). With all other factors being
constant, Stoke’s Law states that the velocity of settling is proportional to the radius of the
sphere squared. If the radius doubles, the settling velocity increases by a factor of four. Holding
the size of the sphere constant, the settling velocity increases when the density difference
between the particle and the fluid (ρ – ρ o ) increases. The settling velocity also increases when
the viscosity (η) of the fluid decreases. Since the viscosity of water is greater at lower
temperatures, settling velocity decreases as the water becomes cooler. Note that if the density of
the particle is less than the density of the fluid (ρ < ρ o ), the velocity term (ν) becomes negative,
indicating a change of direction. The negative velocity term indicates particles rising through the
fluid rather than settling.
The overall goal of coagulation and flocculation is to agglomerate many small particles into
fewer larger ones so that settling (or flotation) occurs more quickly. One important factor to
realize is that coagulants and flocculants do not have a significant effect on the intrinsic density
of the primary particles or the flocs that result. That means that a particle that is naturally
heavier than water will sink and is not able to be floated by the action of coagulants or
flocculants alone. Conversely, a particle that is intrinsically lighter than water cannot be made to
sink by the action of coagulants and flocculants alone. Weighting agents such as bentonite clay
or colloidal silica can sometimes be added to get lighter than water particles to sink. To get
heavier than water particles to float, gas bubbles can be introduced which attach to the flocs and
make them lighter.
POLYMER CHEMISTRY
Coagulants
Coagulants are used in liquid/solids separation applications to neutralize the ionic charges that
surround solid particles dispersed in water. Most naturally occurring solid particles in water
have a negatively charged surface due to the release of cations such as H+, Li+, Na+, K+, Ca+2,
etc. from the surface of the particle into the surrounding water. When microscopic particles of
like charge approach one another, they repel and cannot aggregate to form larger particles. This
leads to very stable systems of particles in water that will not settle. The higher the anionic
charge on the surfaces of the particles, the more stable the colloidal system becomes. Cationic
coagulants adsorb onto the negatively charged particle surfaces thereby neutralizing the negative
charges that are causing repulsion. Optimum coagulation occurs when the particle surface
charge is reduced to near zero. An overdose of cationic coagulant will impart a positively
charged surface on the particle and re-stabilization will occur. For this reason, the dosage of
coagulant is critical. Note that there are many more complex theories of coagulation that are
beyond the scope of this paper.
The first documented use of coagulation for water purification was in ancient Egypt. Figure 2
pictures an ancient Egyptian cave carving showing a device using coagulation and settling of fine
particulate contaminants.
The first coagulants used in industry were aluminum sulfate and ferric chloride. What these two
compounds have in common is a trivalent metal ion (Al+3 and Fe+3) that is strongly attracted to
any negatively charged surface. It was found that trivalent cations were much stronger
coagulants than divalent cations like Ca+2 and Mg+2. Monovalent cations like Na+ and K+ have
very weak coagulating abilities. Since the Middle Ages, the most widely used coagulants have
been aluminum sulfate, aluminum chloride, ferric chloride and ferric sulfate. A 20th Century
development has been the polymerization of aluminum and iron to form inorganic polymers that
carry a positive charge. The aluminum compounds go by the names polyaluminum chloride and
aluminum chlorohydrate when they are more fully neutralized. See table 1.
Table 1: Commonly used inorganic coagulants. Water soluble metal salts with +3 valences
There are also many divalent cations used in liquid/solids separation in spite of the fact that they
are weaker coagulants. See table 2. One benefit of these is that they don’t lower the system pH
as much as trivalent cations.
Table 2: Commonly used inorganic coagulants. Water soluble metal salts with +2 valences
The final category of inorganic coagulants is the Alkaline Earth Oxides & Hydroxides. These
are typically used to raise the system pH to the point where heavy metals precipitate from
solution (pH 9.5-11.0). See table 3.
Organic coagulants are typically based on nitrogen chemistry since nitrogen, which normally
accepts only three bonds, can use its extra pair of electrons to accept a fourth bond in a process
known as quaternization. This induces a positive charge to the nitrogen atom which is then
available to neutralize the negative charges on a particle surface in water. Commercially
available products of this type are water soluble polymers containing the quaternized nitrogen
moiety. One of the most widely used products is a condensation polymer of Epichlorohydrin and
Dimethylamine. The two reactants come together in a reactor to form the polymer commonly
known as Epi/DMA Polyamine. Figure 3 shows the two reactants and the Epi/DMA product.
The molecular weight using this process can be 10,000 to 100,000 daltons. To produce higher
molecular weights (up to 500,000) a third agent is added to the mix to branch the polymer
structure.
Another very useful monomer is Diallyl Dimethyl Ammonium Chloride, or DADMAC (also
known as DMDAAC). Figure 4 illustrates the monomer polymerizing with itself to form
Poly[DADMAC].
Using addition polymerization, Poly[DADMAC] can be made with an average molecular weight
between 50,000 and 1,500,000.
For both Poly[DADMAC] and Epi/DMA polymers, higher molecular weight usually translates
into a larger and stronger floc with faster settling and a more compact final sludge volume. The
stronger faster settling floc may result in a more turbid supernatant than a slower settling floc.
Selection of the proper molecular weight for a given application requires consideration of clarity
requirements, chemistry, and liquid/solids separation process equipment capability.
In general, inorganic coagulants differ from organic coagulants in the ways shown in Table 4.
Blue-shaded cells in table 4 indicate an advantageous aspect.
When the use of pH adjustment chemicals and sludge disposal is added, organic coagulants can
have a rather large economic advantage over inorganic coagulants. The corrosion potential from
lower pH and lower alkalinity effluent water may also be a detriment caused by use of inorganic
coagulants.
Because there are advantages and disadvantages associated with each type of coagulant, there
has been a trend in the industry to use blends of inorganic and organic coagulants to optimize the
treatment program. The optimum blend can produce sparkling clean effluent water over the
existing pH range with the proper floc size and settling rate while minimizing final sludge
volume.
Flocculants
Flocculants are high to very high molecular weight polyelectrolytes that can be anionically or
cationically charged, or nonionic. With some exceptions, commercially available flocculants are
based on polyacrylamide and its copolymers and derivatives. The high reactivity of acrylamide
monomer leads to products with average molecular weights between 1,000,000 and 20,000,000
daltons. High molecular weight means long chains of molecules that have the capability of
physically bridging from one microscopic or colloidal particle to another.
In practice, wastewater is often treated with a coagulant to neutralize negative surface charges to
near zero. This process lowers the electrical repulsive barrier preventing microscopic particles in
water from aggregating. The system is then treated with a flocculant which acts to bridge the
neutralized primary aggregates. A single high molecular weight flocculant molecule is large
enough to bridge several primary aggregates simultaneously. The loops and tails that extend
from these microflocs can further interact with other microflocs thereby creating macroscopic
floc structures. Relative to coagulation alone, flocculation results in larger structures that settle
faster with less pinfloc and carryover of solids in the supernatant water.
Flocculants are also used in sludge dewatering applications to assist in the release of water and
concurrent production of higher solids concentration. This results in a lower volume, drier
dewatered sludge. Added to sludge, the flocculant bridges from particle to particle and creates
large flocs with free water between the flocs. This creates avenues of egress for water drainage
from sludge. Biological sludges, for example, contain 95% or more water even after being
concentrated as much as possible by gravity settling. Looked at from the other angle, this is only
5% solids. Dewatering can produce a sludge cake of 25% that is 1/5 of the original volume.
Industrial sludge generators use a wide variety of dewatering equipment to achieve this goal.
Most popular are belt presses, centrifuges, plate and frame presses, vacuum filters, screw presses
and drying beds, all of which require the addition of flocculants for maximum performance.
Each type of sludge and each type of dewatering equipment will require a chemical program that
is optimized for that application. For example, biological sludges almost always require a highly
cationically charged, high molecular weight flocculant, but the optimum molecular weight and
degree of polymer branching will depend on whether the application uses a belt press or a
centrifuge. Inorganic sludges such as those found in coal preparation plants and mining
applications, normally require an anionic or nonionic, high molecular weight flocculant. Paper
mills, which have a mixture of biological and inorganic sludges can use flocculants with lower
cationic charges. In all cases, the choice of product depends on the type of sludge, the dewatering
equipment in use and the results desired. Selection of flocculants for dewatering requires skill,
knowledge of chemistry and mechanics, testing and experience.
Figure 5 shows the acrylamide monomer, which is the basic building block of most industrial
grade flocculants. Acrylamide carries no net electrical charge (cationicity or anionicity) and can
be classified as a nonionic monomer.
Acrylamide can polymerize with itself when a free radical initiator is added to a solution of
acrylamide monomer. The asterisk (*) in figure 5 indicates a free radical which initiates, then
propagates the reaction. The chain can be as long as 300,000 monomer units long.
There are also cationic monomers that polymerize quite well with acrylamide. The most popular
worldwide is an ester based cationic monomer. When this monomer is charged to a reactor with
acrylamide and reacted, a cationic copolymer is formed. As with the anionic copolymers, the
ratio of cationic monomer to acrylamide controls the final cationicity of the polymer.
Figure 7: Upper left: Acrylamide monomer; Upper right: Ester-based cationic monomer;
Bottom: Polyacrylamide ester-based cationic copolymer
There are many other cationic monomers that may be used in reaction with acrylamide to form
cationic copolymers or terpolymers. Table 5 shows some of these.
Table 5: Polymers and copolymers used in wastewater treatment (WERF Project 91-ISP-5,
1993)
One less commonly produced amide-based cationic monomer reacted with acrylamide produces
amide/acrylamide copolymers as shown in figure 8.
Figure 8: Upper left: Acrylamide monomer; Upper right: Amide-based cationic monomer;
Bottom: Polyacrylamide amide-based cationic copolymer
The advantage of acrylamide and amide based monomer technology is its stability and
performance at high pH. The ester based monomers are known to hydrolyze and lose their charge
starting at about pH 8. The higher the pH, the faster they lose their cationic charge. At pH 10,
ester based polymers can lose their cationicity within minutes. The amide based copolymers
retain their cationic charge for many days even at pH 12. This provides an advantage when
dewatering lime-stabilized sludges and paper mill sludges which are subject to black
liquor/caustic spills.
Molecular Weight
Because polymerization reactions are random, reported molecular weight (MW) for polymeric
compounds represent averages with a Gaussian distribution as shown in figure 9.
In this example, the polymer sample in question would be said to have an average MW of about
15 million daltons. Some molecules are over 20 million daltons, but an equal number are under
10 million.
There are large differences in average molecular weight trends between anionic and cationic
polyacrylamide copolymers. Because sodium acrylate is a small molecule like acrylamide, it
polymerizes very efficiently and generally very high molecular weights are formed. The cationic
monomers, on the other hand, are relatively large and bulky. They don’t polymerize quite as
well with acrylamide monomer and so generally lower molecular weight trends occur.
Figure 10: Typical molecular weight distributions for cationic and anionic flocculants
Every polymer manufacturer uses viscosity to estimate their MWs and the specifications for each
grade’s MW are also based on viscosity. Because each manufacturer uses a different viscosity
test, it is difficult to compare MWs from one company to another. The only way to compare two
polymers from different companies is to run the same viscosity test on both. Here are some
examples of viscosity tests being used in the marketplace:
(1) Standard Viscosity: 0.1% active Polymer in 1.0 N NaCl (pH 7) at 60 rpm Shear in UL
Adaptor.
(2) Solution Viscosity: 0.5% active polymer in deionized water (pH 7) at 12 rpm shear with
spindle 1, 2 or 3.
(3) Intrinsic Viscosity (IV): Measures the time to pass polymer solution through a small
orifice at constant temperature.
(4) There are many more tests being used and it depends on the manufacturer. Also, each
distributing company may have their own viscosity tests so their customer cannot tell
who the original manufacturer was.
The chart below lists the generally accepted terminology with regard to MWs in liquid/solids
separation applications (actual ranges may vary from source to source or company to company):
In actuality, molecular weight is not the property that is meaningful in liquid/solids separation
applications. Molecular weight is used because it can be assigned a number that can be easily
compared product to product. The more meaningful indices are the hydrodynamic volume and
the average end to end distance of the polymer, which are functions of the molecular weight and
the degree of branching of the polymer.
Coagulation and flocculation are the two steps employed for bringing finely divided particulate
or emulsified contaminants first into very small aggregates, then into larger groupings. The term
“floc” is used in this context to represent these gathered groupings. Floc sizes are increased,
therefore enhancing solids separation from liquid as described in Stoke’s Law (Figure 1).
Soluble material includes organic and inorganic materials that are truly solubilized. The scope of
this paper does not extend to removal of truly soluble contaminants.
Figure 12 illustrates some of the important interparticle forces at work. As particles generally
have a net negative charge, particles exhibiting like charges will naturally repel one another.
This is what is called electrostatic repulsion.
van der Waals forces are the relatively weak interparticle forces of attraction. Other interparticle
forces exist, but are beyond the scope of this paper.
Figure 13 illustrates the forces between two particles approaching each other. Let the left sphere
represent a stationary particle and the right sphere represent a particle approaching the stationary
particle. The blue curve represents strength of repulsive forces such as electrostatic repulsion.
The red curve represents forces of attraction between the two particles, and the dotted curve
represents the sum of forces. One can see that there is a net peak repulsion force at a certain
distance as the right particle approaches the left particle. This net repulsive force is responsible
for keeping finely divided particles in suspension, ie., the non-settleable solids shown in Figure
11.
As we look at the single particle shown in Figure 14, we observe several layers of ions
surrounding the particle. The negatively charged particle will have a layer of positive
counterions at the surface, labeled the fixed layer, or Stern layer. These would be cationic. Just
outside the cationic counterion layer would be a more diffuse layer of counter-counter ions,
largely negatively charged. The outer perimeter of this layer is called the slip plane or shear
plane, and represents the difference between stationary ions on the surface of the particle and
movable ions just outside the fixed layer. Zeta potential meters measure charge density in a
solution by shearing the movable layer of ions and measuring electrical potential. Outside of the
shear plane is the bulk liquid. These are mathematically derived and modeled layers that
describe layer thicknesses. The thickness of the layers is inversely related to the electrolytic
strength of the liquid solution. In other words, higher electrolytic strength compresses the layers.
Figure 14: Ionic charge on single particle, counterions in water near surface (Fixed Layer),
counter-counter ions some distance from surface (Shear Plane, Zeta Potential) and bulk
water ions
Referring back to Figure 13, addition of a cationic coagulant neutralizes enough of the negative
charges on particles so that the hump representing net interparticle forces is lowered, allowing
the particles to more easily approach and then pair with each other. This is the basis of
coagulation and yields small, sometimes unseeable by the naked eye, agglomerations, referred to
as pinfloc.
Figures 15 and 16 give pictorial representations of the two-step process of i) coagulation – initial
agglomeration of finely divided particles, and ii) flocculation – grouping of initial agglomerates
into larger flocs, respectively.
Figure 15: Coagulation; a) Charged particles repel each other; b) Charge neutralized by
cationic coagulant; c) Particles form small agglomerates
Figure 16: Flocculation; a) Large flocculant contacts agglomerates, b) wraps and entraps
agglomerates and other particulate matter to form large “floc” particle.
Figure 17 illustrates the four steps of physical/chemical treatment in a clarifier. The coagulation
step entails feed of a chemical coagulant, rapid mixing to promote distribution of coagulant and
interparticle collisions. This is followed by feed of a flocculant to promote buildup of floc size,
and slow gentle mixing. Gentle mixing is necessary to preserve formed floc. The third step is to
provide a quiescent zone where formed floc can freely settle. Finally, proper sludge removal is
essential to prevent buildup of sludge and possible scouring of built-up solids inventory into the
clarified effluent.
Note that chemical additives generally have indiscernable effects on weight of floc groupings;
they simply enlarge floc. If floc particles are more dense than surrounding liquid, then floc will
tend to settle. If floc particles are less dense than surround liquid, or if gas bubbles or floatable
oils are associated with floc groupings, making net density of floc less than that of surrounding
liquid, then floc will tend to rise.
We’ll start out with a generic wastewater treatment system. Many industrial wastewater
treatment systems will have some or all of the unit operations that are represented in Figure 18.
General statements are given concerning each unit operation depicted. Because of the
uniqueness of some treatment situations, there may be a few exceptions to each general
statement given.
1. Pretreatment systems can include bar screens, grit chambers, free oil float separators
and other unit operations. They are used to remove relatively large or relatively dense
solids that are readily separable by settling or coarse screening, or to remove freely
separating oils by gravity flotation. Coagulants and flocculants are not recommended in
pretreatment systems. Physical separation of grit or free oils is meant to take place in
these systems, leaving the more difficult separations for downstream equipment.
2. Equalization tanks are meant for reducing variability of flow and quality going to
downstream equipment in order to promote smoother and more efficient operation.
Chemical additives are not fed to equalization tanks. (Again, there are a few exceptions.)
3. Primary Treatment may include any of the physical/chemical separation unit operations
such as sedimentation clarifiers, induced or dissolved gas flotation. In each case,
coagulants or flocculants may be used to attain desired contaminant removal.
4. Secondary Treatment encompasses any of the biological treatment unit operations
including suspended and attached growth systems along with bacterial solids separation
and recycling. If a wastewater treatment system is loaded according to design and free of
upsets, floc-forming bacteria do the work of building floc size for settling in a
sedimentation clarifier and no chemical additives are required to aid settling. As is the
case in many operating industrial wastewater systems, sometimes there are upsets, or
there may be some wastewater components that are toxic to the working bacteria and thus
hinder flocculation and settling. In these cases, coagulants or flocculants may be used to
capture solids and build floc size for settling, thus preventing excessive carryover of
solids from secondary clarifiers.
5. Tertiary Treatment includes anything beyond secondary treatment including filtration,
oxidation, membrane treatment, tertiary clarification. It can include physical/chemical
treatment such as for the precipitation of trace phosphorous by metal salts addition, and
optionally flocculant addition as an aid to settling.
Also included below are warnings on where use of chemical additives is inadvisable, and where
they could, in some cases, exacerbate problems.
Figure 19: Common areas for use of coagulants or flocculants in HPI and CPI, shown with
purple arrows
Figure 20: Common areas for use of coagulants or flocculants in PPI, shown with purple
arrows
SUMMARY
Chemistry and application of coagulants and flocculants for wastewater contaminant removal has
been described. Coagulants neutralize charge on fine particulate matter to enhance initial
agglomeration of solids. Flocculants bridge agglomerates in order to make larger groupings or
“flocs,” capable of gravity separation in clarification systems. This paper provides a ready
reference for those contemplating or using coagulant and flocculant chemistry in solids/liquid
separation. Proper use of chemistry in conjunction with operating equipment is addressed, as
well as some misapplications to avoid in industrial wastewater systems.
REFERENCES
Betz Laboratories, Inc., (1991) Handbook of Industrial Water Conditioning, 8th edition; Ch. 35,
36; Trevose, Pennsylvania.
Dentel, S. K.; Abu-Orf, M. M.; Griskowitz, N. J. (1993) Guidance Manual for Polymer Selection
in Wastewater Treatment Plants. Water Environment Research Foundation, Project 91-
ISP-5, Alexandria, VA; p.1-14.
Drew Chemical Corp., (Div. of Ashland Inc., now Solenis LLC), (1985), Drew Principles of
Industrial Water Treatment; Boonton, NJ.
Rubin, A. J. (1978) Chemistry of Wastewater Technology. Ann Arbor Science, Ann Arbor, MI.
ABSTRACT
Jurong Water Reclamation Plant receives a high proportion of industrial used water. Currently, the
industrial used water is treated via conventional activated sludge process and the effluent is
discharged to the sea. A demonstration plant with 1-MGD capacity was constructed at Jurong Water
Reclamation Plant to produce high quality effluent through a combined process of up-flow anaerobic
sludge blanket reactor (UASB) and ceramic membrane bioreactor (CMBR). Since commissioning in
January 2014, the demonstration plant has been in operation and is optimised continually in terms of
energy consumption. Although influent COD fluctuated widely, the final effluent COD was kept
consistently lower than 50 mg/L. It is concluded that the demonstration plant can treat industrial used
water and produce high quality effluent which can be reused as industrial water after blending with
MBR permeate from the domestic stream.
1. INTRODUCTION
Singapore is a small island with only a limited land area of 700 km2. The fast growing population and
thriving economy exert pressure on land use, so that water catchments have to compete for land with
activities such as housing, commerce, industry, transport, recreation and schools. The PUB, the
national water agency of Singapore, takes care of all aspects of water management in Singapore; from
the collection of used water and surface water, to produce and supply potable water, and treating and
recycling of used water. PUB has ensured a diversified and sustainable supply of water with “Four
National Taps Strategy” where it sources water from local catchments, imported water, reclaimed water
and desalinated water. PUB started recycling used water for non-potable and indirect potable uses
(known as the NEWater) in December 2002 and seawater desalination in September 2005. PUB
further aims to maximize the collection of rain water and capture storm water from the entire
Singapore to cover 90% water catchment area by 2060.
Although Singapore is a small city state, it is a busy manufacturing hub with the goods producing
industries contributing 27% to the country’s GDP in 2012. These manufacturing facilities produce
industrial wastewater that flows to PUB’s Jurong Water Reclamation Plant (JWRP) for further
treatment to produce industrial water. There are future plans to recycle used industrial water in
Singapore to supply to the industry. Based on the 2022 projection of used water master plan, the
future Tuas Water Reclamation Plant will treat a total of 800,000 m3/day of used water of which
600,000 m3/day is domestic used water and 200,000 m3/day industrial used water. PUB and Meiden
Singapore Pte Ltd (Meiden) had conducted a pilot study for combined process of up-flow anaerobic
sludge blanket reactor (UASB) and ceramic membrane bioreactor (CMBR) to treat the industrial used
water at JWRP. After the success of the pilot study, both organisations forged further collaboration
and jointly conducted the demonstration trial with 1-million gallon per day (MGD) (or 4,550 m3/day)
capacity UASB-CMBR process at JWRP. Details of the demonstration plant and the results of
12-month plant operation are presented in this paper.
2. DEMONSTRATION PLANT
2.1. Background
JWRP receives a high proportion of industrial used water (40% industrial, 60% domestic) compared
to other WRPs in Singapore which receives only domestic used water. The industrial used water is
currently treated via conventional activated sludge process to remove contaminants and the treated
effluent is discharged to the sea. In 2010, in order to produce industrial water at JWRP, a deliberate
effort was made to segregate the low strength domestic used water from the high strength industrial
used water to maximise the recovery of the domestic used water for water reuse. Currently, a
15-MGD membrane bioreactor (MBR) plant was then installed in JWRP to treat the relatively
low-strength domestic used water to produce industrial water for distribution to Jurong Island. In
order to free up domestic used water for reclamation into the high grade water (NEWater) in the long
term, studies were conducted to explore the feasibility of treating the high strength industrial used
water for reuse.
PUB and Meiden had conducted a pilot study to combine two unique technologies i.e. UASB-CMBR
process to tackle the high strength industrial used water at JWRP. The results of the pilot trials are
encouraging as the UASB reactor is robust to handle the high strength used water. Ceramic
membranes can handle the corrosive nature of the industrial used water stream and are more robust
than polymeric membranes. As a follow-up from the success of the earlier pilot study, PUB and
Meiden then embarked on a demonstration trial of a 1-MGD capacity UASB-CMBR process at
JWRP.
Meiden’s ceramic flat sheet membranes technology for MBR process forms a major component of
the 1-MGD demonstration used water treatment plant. The demonstration plant was commissioned in
January 2014. The incoming industrial used water has a varied Chemical Oxygen Demand (COD) in
the range of 1,000 – 2,000 mg/L with occasional peaks of 4,000 to 5,000 mg/L. The UASB reactor
was able to achieve an averaged 30-40% removal of the COD. There are plans to utilize the biogas
generated from UASB reactors for power generation after chemical scrubbing. Downstream
biological process (anoxic-aerobic) followed by CMBR was able to produce consistent final effluent
with a COD level of below 50 mg/L, ammonia levels of below 5 mg/L and turbidity levels of below
0.2 NTU. The conductivity of the final water remains unchanged throughout the treatment processes
(in the range of 2,000 – 3,000 µs/cm). This superior water quality could be mixed with the permeate
of the 15-MGD MBR plant treating domestic used water before supplying the blended water to
industries for their processes and cooling needs.
The concept for treatment of industrial used water at the demonstration plant is a two-step process
with two core technologies. The first is flat-sheet ceramic membrane technology at MBR system and
the second is UASB technology. Features of these two technologies are shown as below:
The flow diagram of the demonstration plant is shown in Fig. 1. Industrial used water from the
distribution chamber is first supplied to EQ tank. Bulk materials are removed through the 2 mm
screens. The used water is then pumped to the UASB tanks followed by CMBR. CMBR consists of an
aeration tank and membrane separation tanks. Sludge is circulated between the aeration tank and two
membrane separation tanks. The effluent from membrane tanks will be reused after blending with
MBR permeate of domestic stream.
Specification of demonstration plant is shown in Table 1. Design capacity of the plant is 4,550 m3/day.
Capacities of the EQ tank and UASB reactors are 760 m3 and 1,200 m3, respectively. Flat-sheet
ceramic membranes with pore size of 0.1μm are installed to the membrane separation tanks. The total
membrane area is 9,600 m2.
Distribution Permeate
Chamber
UASB Membrane
reactor 1 Separation
EQ Aeration tank 1
tank Reuse
tank Membrane
Screens UASB
reactor 2 Separation
Tank 2
Sludge circulation
Excess sludge
2.3.1. UASB
The UASB process bears the kernel of the pre-treatment in this system. Comprising of two reactors in
parallel, it targets mainly the removal of COD by 50% or more. However, if the microorganisms are
very active and over-consume degradable substances, the overall treatment may become unbalanced.
This may result in insufficient organic matter required in the post-stage for activated sludge in the
MBR. Therefore, a line that bypasses the sewage on the feed side of the UASB is provided to supply
biodegradable organics in the form of Biochemical Oxygen Demand (BOD 5 ) to the MBR tank. In
addition, the online measuring instruments such as COD and pH analysers provided constant
monitoring. Furthermore, the recovery of energy in the form of electricity and reduction of
green-house gases (CH 4 ) emission can be achieved by gas-engine generators with the recovered
biogas separated by the Gas Solid separation (GSS), as there is an option to feed the biogas to the
gas-engine generators in JWRP.
2.3.2 CMBR
CMBR consists of an aeration tank and membrane separation tanks. The oxygen transfer efficiency in
the aeration tank is kept high using a high-efficiency air diffuser system. In addition, an agitator is
separately provided in the aeration tank in order to adjust the agitation capacity. A quasi-anoxic tank
can be produced by partially stopping air diffusion hence maintaining the alpha factor. The blower is
stopped and only the agitator is operated in order to reduce power consumption. The system has the
flexibility to cope with changes in water quality and the dynamic control of air in the aeration tank
demonstrates optimal use of energy.
The membrane separation tanks house the flat-sheet ceramic membrane units. Each unit has 400
modules of flat-sheet ceramic membrane with pore size of 0.1 μm (Table 2). The basic operation of
the system is designed to keep low level of fouling on the membrane surface with air scrubbing.
Periodic backwashing is performed to minimizing fouling. The demonstration plant aims to further
increase the flux and save energy. Photo illustrations of the membrane unit and the membrane
separation tanks are shown in Fig. 2 and Fig. 3, respectively.
3. Performance
The demonstration plant has been in operation for more than a year. The main objective of this project
is to reduce the COD from varying influent to below the control limit of 50 mg/l by biological
processes. Figure 4 shows the variation of COD in EQ tank, effluent of UASB and MBR product. The
influent COD fluctuated between 260 and 4,440 mg/L. Although the influent COD fluctuated widely,
the effluent COD was consistently lower than 50 mg/L.
The UASB-CMBR product water will be blended with MBR permeates from the domestic stream.
Based on the mass balance, we have determined a specific water quality from UASB-CMBR
demonstration plant. Table 3 shows a typical water quality, which was obtained as average values
during operation in March 2015. COD, BOD 5 and turbidity of product water were 47 mg/L, 7.6 mg/l
and 0.19 NTU, respectively. These levels met the target values for the product water from the
UASB-CMBR process.
5000
4000
COD (mg/L)
3000
2000
1000
0
0 20 40 60 80 100 120 140 160 180
Duration (days)
EQ tank UASB eff. MBR product
The plant was mainly tested at two different membrane fluxes, 21 LMH and 25 LMH. Variation in
transmembrane pressure (TMP) of the two membrane separation tanks are shown in Fig. 5. It was
observed that the TMP was range bound from 8 to 10 kPa for 75 days at 21 LMH which indicated
stable membrane performance at this flux. After increasing the flux to be 25 LMH, the TMP was
ranged between 8 and 15 kPa for first 20 days (85 days to 105 days in Fig. 5). The TMP, however,
fluctuated widely after 20 days of operation at 25 LMH (After 105 days in Fig. 5). High levels of oil
and grease (O&G) contaminant were observed during this period which might cause severe fouling
on membrane. After 150 days operation, the membrane flux was set to be 15 LMH to stabilize the
membrane filtration and fluctuation of the TMP suppressed.
30
21 LMH 25LMH 15 LMH
25
20
TMP [kPa]
15
10
5
0
0 20 40 60 80 100 120 140 160 180
Duration (days)
Energy consumption is an important parameter for the success of this process. The target of 0.85
kWh/m3 was set prior plant commissioning. Energy consumption as a function of plant production is
shown in Fig. 6. Over the trial run period the energy consumption has been brought down to below
0.8 kWh/m3 at above 3,500 m3/day of plant production. The project team intends to achieve an
ambitious target of 0.7 kWh/m3 in coming months by varying air flow to reactors and membrane flux.
2.0
1.8
Power consumption per Production
1.6 21
24LMH(Power consumption)
1.4
25LMH(Power consumption)
[kWh/m3]
1.2
25LMH(Power consumption-Biogas
1.0 generation)
0.8
0.6
0.4
2000 2500 3000 3500 4000 4500
Plant Production in JWRP [m3/day]
4. EVALUATION
The achieved target values of the 12-month demonstration testing are shown in Table 4. Water quality
of MBR product, membrane filtration flux, power consumption met targets. However the COD
removal and biogas generation in UASB did not reach the internal target value because of lower COD
value in the influent water (compared to the designed influent COD value of 2,000 mg/L). The COD
removal is likely to be improved if the influent of UASB has higher COD value.
5. CONCLUSIONS
It can be concluded that the demonstration plant with a combination of UASB and CMBR technology
taken from industrial used water of JWRP could produce high quality water to be re-used as industrial
water after blending with MBR permeate of domestic stream. In the future, the stability of the
long-term operation, influence of high flux operations to the TMP, and possibility of opex cost
reduction will be studied. CMBR permeate will also be used as a feed to a small reverse osmosis (RO)
unit to study the impact on RO membranes and observe the RO permeate quality with focus on the
Total Organic Carbon level in the RO feed and permeate.
ACKNOWLEDGEMENTS
The project team would like to acknowledge the Economic Development Board, Singapore, for their
support for this project. The project team is also thanking the staff of JWRP for their help during the
project execution.
REFERNECES
van der Roest,H.F., van Bentem, A.G.N. and Lawrence ,D.P.(2002). MBR-technology in municipal
wastewater treatment:challenging the traditional treatment technologies,Water Sci.
Technol.,46,273-280.
Tao,G., Kekle,K., Zhao,W., Ting,C.L., Viswanath,B. and Seah,H.(2005). Membrane bioreactor for
water reclamation ,Water Sci. Technol.,51,431-440.
Tao,G., Kekre,K., Oo,M.H., Viswanath,B., Lew,C.H., Meng,L.M. and Seah,H.(2009). Large scale
membrane bioreactor plant design(retrofit) and optimization, IWA membrane technology conference
& exhibition 2009.
Noguchi,H.,Niwa,T.,Lie,H.,Kekre,K.,Tao,G. (2010). Development of Ceramic Membrane
Bioreactor(MBR) using Flat-sheet type Ceramic Filter,Proceedings in water convention at SIWW
2010
ABSTRACT
General Motors (GM) owns and operates a large manufacturing complex in Silao, Guanajuato
Mexico. GM is embarking on a comprehensive water recycling program to maximize water
reuse for manufacturing processes, and to augment the current source of water for the Site, which
is local fresh groundwater wells. The Site currently recycles approximately 35% of the
biologically-treated and tertiary-filtered wastewater effluent for use in cooling towers, irrigation,
and to toilet rooms. The conceptual treatment train for a new water reuse treatment system
includes a membrane bioreactor (MBR) and three-stage reverse osmosis (RO) in Phase I, to
recycle approximately 80% of the site’s wastewater. A future phase of the project would include
on-site treatment of the brine from the RO process, recovering even more water through
application of zero-liquid discharge technology (evaporator/crystallizer combination with
mechanical vapor recompression), achieving over 95% recovery.
INTRODUCTION
General Motors (GM) manufactures engines, transmissions, and full-size trucks (e.g., the
Chevrolet Silverado) at a large manufacturing complex in Silao, Guanajuato, Mexico. The Site
employs approximately 6,000 personnel. Additionally, there are approximately 4,000 contracted
employees on site in any given day. The GM Silao site has experienced dramatic growth since
1995 and is currently under construction to add nearly 3 million square feet of new production
facilities. As a responsible corporate citizen, GM is dedicated to protecting human health,
natural resources and the global environment. Along with other environmental goals, GM has
committed to a 15% reduction in water usage per vehicle by 2020, with 2010 as the baseline year
(GM, 2016).
GM employs water reuse and water reduction techniques at many facilities, including rinse water
cascading, automatic shutoffs, and controls to prevent water waste. Table 1 summarizes the
water reuse strategies used at existing GM plants throughout the world.
Quantity
Plant Location Facility Type Recycle Summary (m3/year)
San Luis Potosi, Complex, Vehicle Zero Liquid Discharge Unknown
Mexico Assembly,
Transmission, Engine
Ramos Arizpe, Complex, Vehicle Reuse Treated Wastewater for 72,000
Mexico Assembly, Cooling Towers, Paint
Transmission, Engine Humidification and Sludge,
toilets and irrigation
Sao Caetano do Sol, Complex, Vehicle Reuse Treated Wastewater for 121,000
Brazil Assembly, Toilets and Cooling Towers
Components, Head
Quarters
Joinville, Brazil Engine Plant Unknown 5,670
Detroit Hamtramck, Assembly Plant Stormwater Reuse to cooling 189,000
USA towers and high purity water
system
Wentzville Assembly Plant Stormwater Reuse (2017 180,000
Assembly, USA implementation), Foundation
drainage used for paint sludge
and wastewater usage
For water, the Site relies on six groundwater wells ranging from 300 meters to 500 meters deep..
The Site must frequently rotate which wells are in service to obtain the water needed for
manufacturing operations. In addition to very high taxes for groundwater extraction, the aquifer
levels are declining. The Silao-Romita aquifer has been identified as stressed and the region is
expected to have continued growth in population, agriculture and industry (World Bank, 2004).
Considering its environmental goals, the status of the region and aquifer, and to reduce
production disruption risk, GM is embarking on a comprehensive water recycling program to
maximize water reuse for manufacturing processes. The Site currently recycles a portion of the
biologically-treated wastewater effluent for use in irrigation (100-150 m3/day), cooling towers
(350-500 m3/day), and to toilet rooms (170-240 m3/day); however, the project aimed to increase
overall recycle by a considerable amount through the addition of an advanced treatment and
recycle plant.
An array of concepts for treatment and recycle were developed and vetted relative to key project
criteria (e.g. percent of wastewater recycled, robustness, operability, CapEx and OpEx, etc).
The conceptual treatment train for a new water reuse treatment system includes a membrane
bioreactor (MBR) and two-stage reverse osmosis (RO) in Phase I; the performance/operability of
these processes benefit from strategic upstream and waste segregation components to the project.
A future phase of the project would include on-site treatment of the brine from the RO process,
recovering even more water through application of zero-liquid discharge technology
(evaporator/crystallizer combination with mechanical vapor recompression).
The current groundwater withdrawals average slightly less than 2,000 m3/day, and it is not
anticipated that any additional withdrawals will be sustainable in the future due to other demands
in the region. Currently, the site has groundwater rights for 1,000,000 m3/year (equivalent to
approximately 2,700 m3/day); however, the aquifer level has been dropping steadily and a 2014
hydrogeological study indicates low recharge rates such that groundwater yields from the
existing wells are not anticipated to meet future demands. As a result of these water challenges,
actual present-day yield for the wells is approximately 2,000 m3/day, which meets current water
needs, but is insufficient to accommodate plant growth.
A critical step in the project was developing a clear understanding of future water demands, both
relative to water quality and volumes, following plant modifications and expansions. This was
important as it directed the treatment and recycle strategy for the project. The team first
developed a detailed site water balance for the current site operation via data from existing in-
line flow meters and supplementing with data from temporary strap-on flow meters. This
‘current case’ balance was expanded to include water quality characteristics using existing (water
and wastewater) analytical data from the site and supplementing with new water and wastewater
sampling and analyses where needed.
Building from the ‘current case’ balance, the team developed a ‘projected future’ balance; future
water demand projections were developed through a combination of interviews with existing and
future site process water users, CUC personnel, evaluation of design documents for plant
expansions, chemical supplier requirements (e.g. paint shop), as well as industry and site
experience. To ensure a recycle process is designed and sized correctly, it is important to
consider both typical/average conditions as well as peak/extreme conditions. To this end, the
‘projected future’ balance included average and peak values for flows and key water quality
characteristics.
As the site expands, and new automotive manufacturing processes are added that increase site
water need, the site will need to recycle – as much wastewater as possible to supply the projected
water needs of the manufacturing processes (up to 3,200 m3/day). As cycles of concentration
increase due to the internal recycling of water at the Site, there will need to be a removal
mechanism for TDS and specifically such parameters as silica and total hardness. Refer to
Figure 1 for a summary of the existing site water balance and the anticipated future water
balance.
Wastewater streams are discharged to wastewater treatment facilities at the Central Utilities
Complex (CUC), and segregated based on origin (Engine, Transmission, Stamping, Body Shop,
Painting, Assembly, etc.). Oily waste is treated through screens, oil/water separators, emulsion-
breaking batch reactors, and is then combined with other Industrial Wastewater streams for
treatment at the Industrial WWTP. The Industrial WWTP employs equalization, pH adjustment
with lime, coagulation and gravity separation with lamella plate settlers, and is then sent to
biological treatment. Sanitary wastewater is also treated in the biological WWTP, and comprises
about 17% of the average inflows. At the Biological WWTP, wastewater is treated in an anoxic
selector tank followed by extended aeration. The biological effluent is treated through sand
filters and a portion is discharged to the local municipality treatment works under permit.
However, a significant portion is currently reused on-site for toilet flushing, lawn irrigation, and
the CUC cooling towers.
Investigations have led the team to identifying options for improving the source water that will
be fed to a new water reuse plant. First, the oily wastewater – while only representing 2 percent
of the flow to the Industrial WWTP – contributes almost half of the chemical oxygen demand
(COD) loading to the Biological WWTP facility. In addition, the batch treatment of the oily
wastewater made the daily COD loading unpredictable. Segregating this stream will improve the
consistency of the influent to the Industrial WWTP, improve the biological treatment unit’s
operations, and mitigate potential fouling on the proposed RO membranes. The segregated oily
wastewater will be pretreated and separately biologically treated for discharge to the
municipality.
Second, after a condition assessment, items at the existing wastewater treatment facilities have
been identified for rehabilitation or replacement. Fixing or replacing equipment such as mixers,
lamella plates, and expanding sludge management should improve the operability of the
existing facilities and make it easier to produce a consistent, high-quality effluent for reuse.
As expansion occurs at the site, the flows to the WWTP are projected to increase, as described
above. The primary source of water use and industrial wastewater discharge to the CUC is from
the manufacturing painting operations in the Truck Assembly plant. Flow metering, sampling,
and other process investigations at the paint shop indicated potential for recycling highly treated
wastewater back to the paint processes.
A key to recycling the increased wastewater for reuse is the biological treatment system. There
are space constraints that would make expansion of the existing extended aeration biological
treatment units very difficult. To obtain more capacity out of the existing biological treatment
basins (without increasing the volume), a Membrane Biological Reactor (MBR) has been
proposed as a retrofit to the existing facilities. MBR trains produce exceptional effluent quality,
and are an excellent pretreatment step upstream of RO membranes. The MBR configuration is
not dependent on the settleability of the mixed liquor; therefore, the biological reactors can be
operated at a much higher mixed liquor suspended solids (MLSS) concentration, and can also be
maintained at a higher solids retention time (SRT) than a conventional activated sludge or
extended aeration system. Refer to Table 1 for a generic comparison of MBR technology versus
conventional activated sludge (CAS) systems.
Aside from enhancing biological treatment and reducing the COD to a lower, more consistent
level, the MBR retrofit will significantly reduce the total suspended solids (TSS) of the
biological treatment effluent, which is important for the feed to the proposed Water Reuse RO
system. Turbidity readings from the effluent side of an MBR system, utilizing ultrafiltration
(UF) membranes, are generally below 0.1 NTU.
Arcadis and a local subconsultant, Intema, made several visits to the Site to assess the condition
of the existing water treatment assets, and to collect wastewater samples in order to characterize
the influent wastewater streams to the wastewater treatment plant (WWTP). In places where the
Site lacked operating flow data, portable flow meters were installed to ascertain the typical flow
rates through those lines during normal operation.
These collected data were synthesized into a Site-wide flow balance, which was also augmented
with a material balance for several key parameters of interest to the new treatment facilities
design. The influent design criteria for biological treatment is shown in Table 1 below, not
including the oily wastewater that will be segregated.
Parameter Value
Influent Temperature 10-30°C
COD, mg/L 2,000
TKN, mg/L 140
Phosphate, as P, mg/L 40
Total Suspended Solids, mg/L 250
Total Dissolved Solids, mg/L 1,500
Oils and Grease, mg/L 10
Alkalinity, mg/L as CaCO 3 170
Hardness, mg/L as Ca/Mg 60
pH 6.0 – 9.0
The Site currently treats an average of 1,200 cubic meters per day (m3/day) of influent flow to
the biological WWTP; however, based on the planning-level flow balance for the recycled water
scenario, including the potential increased water usage of the new Paint Shop
Pretreatment/Electro-Deposition systems (PT/ED), this influent flow to the WWTP is expected
to increase. Future flow scenarios show that the anticipated total flows to the WWTP may be as
high as 2,000 – 2,500 m3/day. Therefore, the basis of design capacity for the new MBR
treatment system has been set at 2,500 m3/day. The MBR vendor will be required to provide
enough redundancy to treat 2,500 m3/day at the design concentrations shown above in Table 1,
with one full train out of service.
The effluent requirements for the MBR system (which is the feed to the RO system) are
summarized in Table 2 below. Any flow not forwarded to the RO will be sent to the site’s
existing tank for reuse to cooling towers, toilets, and irrigation, or the municipality.
Parameter Value
BOD 5 (mg/L) <5
COD (mg/L) < 50
TN (mg/L) <8
Oil and Grease (mg/L) < 0.1
pH (S.U.) 6.0 – 8.0
Total Suspended Solids, mg/L < 0.1
Turbidity (NTU) < 0.1
A reverse osmosis system is proposed for the recycled water treatment scheme, to remove
dissolved salts and residual organic compounds from the water before it is reused. Since the
main customer of the new water reuse plant is the new PT/ED, which requires very low chloride
and silica, a dual train full-flow RO system will be sized for planning purposes. Piping will
allow a portion of the flow to bypass the RO to provide operational flexibility, but the bypass
valve will be normally closed for initial operations. It will be able to treat up to 2,000 m3/day
feed water with one skid out of service.
The target overall recovery for the two-stage system is 80%, producing an RO permeate flow
rate of 1,600 m3/day. To achieve the high overall recovery, the skids will be equipped with scale
inhibitor chemical feed and the option to feed an acid to lower the pH of the feed water as well.
Lowering the pH can reduce scaling potential of many common precipitants (a notable exception
is silica). The preliminary RO modelling projections show that the limiting species to the design
of the overall recovery / scaling potential is calcium fluoride (CaF 2 ). The system is subject to
high concentrations of fluoride in the feed due to a fluoride component from the manufacturing
process.
Based on samples retrieved from the effluent of the biological WWTP and analytical results, the
design feed concentrations to the RO system are as shown in Table 3.
Calcium mg/l 40
Carbonate mg/l, as CO 3 <5
Chloride mg/l 250
Fluoride mg/l 20
Iron, dissolved mg/l 1.0
Magnesium mg/l 25
Manganese, dissolved mg/l < 0.2
Nitrate mg/l, as N 200
Nitrite mg/l, as N <0.1
Nitrate + Nitrite mg/l, as N 200
Organic Carbon, dissolved mg/l 10
Organic Carbon, total mg/l 10
Orthophosphate mg/l, as P 2
Potassium mg/l 65
Silica, dissolved mg/l 45
Sodium mg/l 300
Sulfate mg/l 300
pH s.u. 6-8
Total Dissolved Solids mg/l 1,500
Total Suspended Solids mg/l < 0.1
Temperature °C 10-40
The target RO permeate water quality is shown for reference in Table 4 below.
Besides developing recommendations for the main water recycling treatment train, the condition
assessment of the existing facilities also identified several modifications to upstream processes
that would be beneficial to the overall Site-wide water recycling program.
The site currently has segregated oily waste collection sewers and industrial process waste
sewers. The oily wastewater is treated through a separate pretreatment train before being
combined with the industrial wastewater. However, it was noted during site visits and sampling
activities, that the oily wastewater flow rate is low, but it contributes almost half of the total
chemical oxygen demand (COD) loading of the combined influent to the biological WWTP.
Therefore, since it is high-strength and variable (exerting potentially detrimental spikes in
organic loading on the WWTP), it was identified as an ideal candidate to isolate and divert from
the main biological WWTP in the future. Hence, piping and valves will be installed to re-direct
the oily wastewater to a smaller biological WWTP, originally used for sanitary treatment. From
there, the effluent is sent to the municipal sewer. This change will improve the treatability of the
remaining wastewater sent to the MBR and avoid process upsets that could impact the water
recycling RO facilities.
The sanitary wastewater collection system currently employs an in-channel grinder system to
macerate solids before they enter the wet well to the main lift station that transfers sanitary
wastewater to the biological WWTP. The grinder generates small solids that pass through
screening devices and do not settle easily in the secondary treatment process; therefore, it is
proposed that the existing grinder will be demolished as part of this project. The in-channel
grinder will be replaced with a vertical mechanical bar screen to keep large solids out of the
influent wet well, thereby protecting the influent lift station pumps. The existing sanitary lift
station pumps will be replaced with solids handling wastewater pumps, which have been shown
to operate very well with minimal maintenance in similar situations.
Screening of the sanitary wastewater will be accomplished with new rotary drum fine screens
(<2 mm perforated screens) to maintain the MBR system’s membrane warranty.
Storage of the recycled water is required to ensure that above-average hydraulic demands (or
supplies) are buffered so that enough water is always available and that during periods of excess
recycled water supply, the water does not have to be discharged (i.e. wasted). Two locations
have been identified for additional storage: (1) MBR filtrate tank, upstream of the RO process;
and (2) recycled water storage tanks upstream of the pumped distribution to the water users.
Both of the tanks are proposed to be 200 m3, or approximately 50,000 gallons nominal storage.
These each provide approximately 2.4 hours of storage at design flows through the water recycle
plant. Constant-pressure factory-built pump skids are used to transfer and distribute water from
each of the tanks. From the recycled water storage tank, a portion of the water will be pumped to
the new PT/ED (largest primary water user, and excess recycled water is available for use by
Non-Potable Water (NPW) customers, such as the CUC cooling towers, irrigation and toilet
make-up water. This will be accomplished by tying in new lines to existing water reuse and
softened water supplies and storage tank(s). Check valves or air gaps will be used to ensure that
water from the NPW system does not flow back into the recycled water storage or distribution
piping.
The two major equipment packages (MBR and RO) underwent an extended, thorough pre-
procurement process to ensure that GM obtains the best overall value for the installed system.
To facilitate efficiency during the detailed design phase of the project, the team limited the
design to the selected vendor from the procurement phase. Doing so also ensures that the
ancillary systems are not over-designed (to accommodate the specifics of multiple vendor
systems). The on-site construction is scheduled to begin in the 4th Quarter, 2017. Equipment
delivery is scheduled to occur within the 1st Quarter, 2018 and the system is scheduled to be
mechanically-complete by mid-2018.
A workshop was conducted with GM in November 2016 to discuss the results of the alternatives
evaluation, cost models for various treatment trains, and the potential recycled water users on-
site. Projections for site water usage indicate that the demands for recycled water will grow over
time, as production capacity increases and reliance on local fresh water sources decreases.
Therefore, one of the key outputs from the meeting was determination of the initial size of the
water recycling plant (1,600 m3/day finished water), and to ensure that the design includes
provisions for expandability, redundancy, and employs modular construction to the maximum
extent possible. This means that the design requires vendor equipment to be skid-mounted and
factory-built (pre-piped and pre-wired) to the extent practical. The modularity of the design and
fabrication will also facilitate an accelerated construction schedule.
The MBR system’s membrane trains will include expandability within the membrane tanks, as
each train is designed with module racks that include 25% spare tank space. To increase flow in
the future – without adding more membrane trains – the Site would install the additional
membrane modules to fill the spare spaces. Larger expansions in the future may require
additional membrane trains, along with additional biological treatment rearrangements.
The GM manufacturing complex in Silao, Guanajuato, Mexico has an urgent need to augment
existing fresh groundwater supplies with reuse of combined on-site industrial wastewater and
sanitary wastewater. The selected option to accomplish this economically and sustainably is to
The first phase, which is to be implemented in 2018, will allow the Site to recycle approximately
80% of the wastewater generated on-site, while discharging under permit the remainder,
including brine concentrate from the RO to the local municipality. Recovery of most of the
remaining 20% is planned for a future phase that will employ a thermal evaporation and recovery
to further concentrate the brine and recover a fresh water stream to blend back in with the RO
permeate. Eventually, GM anticipates recycling 95% of the water used on-site, sustainably
reusing water in an arid region of Mexico where many other public and private interests also
need groundwater.
REFERENCES
Foster, S.; Garduno, H.; Kemper, K. (2004) The Cotas – Progress with Stakeholder Participation
in Groundwater Management in Guanajuato-Mexico. World Bank LCR. Mexico.
General Motors (2016) Moving Forward – GM 2016 Sustainability Report. Detroit, MI.
ABSTRACT
KEYWORDS
Industrial wastewater, selenium, selenite, selenate, iron co-precipitation, ferrous sulfate, lime,
zero valent iron
INTRODUCTION
Selenium in wastewater can exist in four oxidation states (+6, +4, 0, and -2). As elemental
selenium, selenium is present in the particulate, non-soluble form and can be removed through
filtration. In the selenite form, or the +4 oxidation state, selenium is soluble and can be removed
through chemical addition treatment, such as iron co-precipitation. In the selenate form, or +6
oxidation state, selenium is soluble and very difficult to remove through simple chemical
addition treatment, but instead requires a highly reduced environment to convert selenate to
selenite, which can then be reduced to elemental selenium or removed through iron co-
precipitation or other more simple chemical addition techniques. Selenium in the (-2) oxidation
state exists as hydrogen selenide. This form of selenium was not expected or found to be present
in the facility wastewater.
The facility historically discharged the selenium bearing process wastewater streams, which
contained total selenium concentrations ranging from 0.5 mg/L to 200 mg/L, to the local
publically owned treatment works (POTW) in accordance with their pretreatment discharge
permit. In 2016, the POTW established a total selenium pretreatment limit of 0.5 mg/L. This
permit change prompted the facility to take internal action to identify sources of selenium and
develop a pretreatment approach to their selenium containing wastewater streams prior to
discharge.
Due to the unique nature of process wastewater at the facility and the known elevated
concentrations found in specific wastewater streams, removal of selenium to levels below the
new permit limits was expected to be difficult. Further, selenium is not a common parameter
requiring removal to low levels from industrial wastewaters that are being discharged to POTWs;
therefore, limited treatment research and results for selenium removal on this specific type of
wastewater were available.
Commonly used treatment technologies for selenium removal include chemical reduction and
precipitation such as zero valent iron (ZVI), media adsorption, ion exchange, reverse osmosis,
and biological reduction.
The purpose of this project was to identify the most cost effective approach to segregation and
treatment of the facility process wastewater stream such that the facility could meet the new
selenium limit without constraining production at the facility.
METHODOLOGY
The project began by examining the raw materials used in each wastewater producing process at
the facility. For processes that were expected to contain selenium in the wastewater discharges,
the speciation of the dissolved selenium was predicted. Additionally, expectations for the
influence on treatment approaches due to the combination of selenium species, concentration of
acids, and concentration of hydrogen peroxide in the wastewater were anticipated.
The facility process evaluation determined that three process wastewater streams at the facility
contained appreciable concentrations of selenium. Due to client confidentiality, these waste
streams are referred to as Waste Stream 1, Waste Stream 2, and Waste Stream 3. Waste Stream
1 was identified to have a low concentration discharge (Waste Stream 1A) and a high
concentration discharge (Waste Stream 1B).
Following the desktop evaluation, AECOM worked with the facility to sample and analyze the
identified wastewater streams to determine the type and concentration of parameters in each. As
part of this process, both current and expected future flow rates and selenium concentrations
were defined. Additionally, samples were sent to an outside analytical laboratory for selenium
speciation testing.
Selenium was found to be present in the facility wastewater streams as either selenite only,
selenate only, or a combination of both selenite and selenate. A summary of characterized
selenium speciation and design basis concentrations for the identified selenium bearing
wastewater streams is presented in Table 1 below. It should be noted that the client
recommended a notable increase to the selenium concentrations to be anticipated in each waste
stream, as compared to the results of the wastewater characterization.
Table 1. Expected Selenium Speciation and Concentration in Selenium Bearing Waste Streams
Wastewater Stream Selenium Speciation Selenium Concentration (mg/L)
Waste Stream 1A Selenite only 1.0
Waste Stream 1B Selenite only 100
Waste Stream 2 Mostly selenite with some selenate 100
Waste Stream 3 80% selenate and 20% selenite 200
Based on the wastewater characterization results, bench-scale laboratory testing was performed
to evaluate a number of waste segregation and treatment options for technical and cost
feasibility.
Treatability testing was performed at the facility over a six (6) month period, which included
multiple rounds of onsite bench-scale testing.
As part of bench-scale testing, wastewater samples collected from the facility were spiked with
concentrated solutions of selenite and selenate, based on wastewater characterization results and
the projected long-term increase of selenium use in the manufacturing process. Samples were
then blended at varying rates to simulate potential full-scale operational scenarios. Following
sample spiking and blending, treatment options were evaluated based on the type and
concentrations of parameters present in the wastewater streams.
The client’s initial preference regarding treatment was to combine all selenium bearing
wastewaters and examine a single treatment process. The combined wastewater stream was
anticipated to be a treatment challenge given that it would put selenate in the presence of nitric
acid and selenite in the presence of hydrogen peroxide. With selenate in the presence of nitric
acid and the anticipated treatment approach including chemical reduction of selenate, the nitrates
present in the wastewater were identified to be a major treatment issue, as the nitrates would be
preferentially reduced prior to selenate reduction. Additionally, selenite in the presence of
hydrogen peroxide would risk conversion of selenite to the more difficult to remove form of
selenium, selenate.
Therefore, based on the results of wastewater characterization, the project team anticipated
treatment challenges associated with wastewater composition and treatment approach. After the
results from the initial round of combined wastewater treatability testing were obtained, it was
determined that wastewater segregation based on selenium speciation was the most efficient
approach to selenium removal, as successful results were not able to be achieved through the
treatment of a combined stream. This wastewater segregation included the separation of Waste
Streams 1A and 1B (selenite bearing) from Waste Streams 2 and 3 (selenate bearing).
For selenite bearing wastewater streams (Waste Streams 1A and 1B), iron co-precipitation
treatment was used for selenium removal with ferric chloride addition. Raw Waste Stream 1A
wastewater was found to have an initial pH of approximately 2.0 and was spiked with a 5 mg/L
selenium concentration as sodium selenite. Wastewater pH was then increased to a target pH
setpoint of 6.0. Once the pH setpoint was reached, varying ferric chloride dosages were
evaluated ranging from 0 mg/L to 800 mg/L.
The highly concentrated Waste Stream 1B was found to contain elevated concentrations of iron
and nitrates. Due to the elevated iron concentrations this waste stream was evaluated as a
potential iron source to supplement the need for ferric chloride addition through calculated
blending with Waste Stream 1A.
During the initial testing with combined wastewater streams, it was found that ferrous sulfate and
lime chemistry was able to achieve the best effluent results. Therefore, this treatment approach
was further evaluated on the selenate bearing waste streams.
For selenate bearing wastewater streams, it was determined that a multi-step treatment approach
was required. In order to eliminate the oxidizing characteristics of the wastewater and to prevent
further oxidation of selenite to selenate, the bench tests were designed such that the hydrogen
peroxide present in Waste Streams 2 and 3 was removed through the addition of either catalase,
an enzyme used to quench residual hydrogen peroxide, or sodium bisulfite.
Selenate bearing waste streams were blended based on full-scale operating conditions and the
initial pH was found to be <2.0. For catalase to work properly, the wastewater pH was required
to be increased to at least 3.0. Following pH adjustment, catalase was added at a rate of
approximately 0.005 milligrams (mg) catalase per mg of hydrogen peroxide to be removed. A
reaction time of 30 minutes was then provided, with hydrogen peroxide readings collected
periodically using test strips.
Once hydrogen peroxide was confirmed to be removed, samples were spiked with selenium in
the form of sodium selenate. Selenium removal using ferrous sulfate and lime precipitation
chemistry was then initiated. Ferrous sulfate was added to the wastewater samples at varying
concentrations ranging from 1,000 mg/L to 6,000 mg/L as iron, with the most appropriate
concentration found to be 6,000 mg/L. Following ferrous sulfate addition, lime was added to
adjust the wastewater pH to approximately 8.5. Once the pH setpoint was reached, samples were
allowed to react with the added chemicals for 1 hour. Polymer was then added to the wastewater
samples, flocculation mixing was performed, and solids were allowed to settle for 30 minutes
before supernatant samples were collected and settled sludge volumes were recorded.
Each chemical treatment scheme was evaluated through a number of iterations looking at
different dosages and pH set points to determine optimized conditions.
In parallel with the bench-scale testing program, samples of selenate bearing wastewaters (Waste
Stream 2 and 3) were sent to vendors for testing with ZVI chemistry and selenium specific
resins.
RESULTS
Initial project data gathering and waste characterization yielded a large data set with multiple
sample points for each waste stream and a wide range of parameters evaluated. Although
selenium was the focus for treatment, knowledge of other parameters and their concentrations
present in the waste streams was critical to determining the direction for treatment approach.
An analysis was performed for a wide range of parameters on each identified waste stream.
Relevant parameters and their analytical results (range of results) are summarized in Table 2.
As can be seen from a comparison between Table 1 and Table 2 selenium data, the actual
measured concentrations are notably less than the client recommended future full-scale
concentrations.
For the selenite bearing wastewater streams (Waste Streams 1A and 1B), it was found that iron
co-precipitation using ferric chloride at a concentration of approximately 100 mg/L (as Fe) with
a pH between 5.0 and 6.5 was capable of removing selenium to <0.05 mg/L in the presence of
elevated nitrate concentrations. Additionally, through calculated blending of Waste Stream 1B
with Waste Stream 1A, adequate concentrations of iron were determined to be present for
consistent removal of selenium to <0.05 mg/L after pH adjustment and polymer addition.
Therefore, using this approach, ferric chloride addition was not required.
Results from testing of selenite bearing waste streams are summarized in Figures 1 and 2. Figure
1 displays results from testing with Waste Stream 1A only, while Figure 2 displays results from
testing after Waste Streams 1A and 1B were proportionally blended.
Figure 1. Results from Iron Co-Precipitation Testing on Selenite Bearing Wastewater – Waste
Stream 1A Only
Selenate bearing wastewater streams proved to be more difficult to treat, but were successfully
treated to provide an effluent discharge of 0.5 mg/L by removing the hydrogen peroxide with
catalase, neutralizing the excess sulfuric acid to bring the pH to 6, and applying a ferrous sulfate
dosage of 6,000 mg/L (as Fe) followed by a subsequent lime dosage to adjust the pH to
approximately 8.5. A dosage rate of 180 mL of 2.5% lime per liter of wastewater was found to
be required for pH adjustment to 8.5. Results from testing of selenate bearing waste streams are
summarized in Figures 3.
Figure 3. Results from Ferrous Sulfate and Lime Testing on Selenate Bearing Wastewater –
Proportionally Blended Waste Streams 2 and 3
Sludge production for the two segregated treatment approaches was evaluated after 30 minutes
of settling time during treatability testing and it was found that approximately 140 mL of sludge
per liter of wastewater treated was produced for treatment of Waste Streams 1A and 1B. A
majority of the sludge was attributed to the elevated iron concentrations in Waste Stream 1B.
Additionally, for treatment of selenate bearing wastewaters (Waste Streams 2 and 3), it was
found that approximately 410 mL of sludge per liter of wastewater treated was produced. This
elevated sludge production, as compared to the selenite treatment process, was due to the very
high concentration of ferrous sulfate and lime required to reduce and remove selenium in the
highly oxidized selenate form.
Results from the vendor testing using ZVI chemistry and selenium specific resins were found to
be unsuccessful for removal of selenate bearing waste streams to target levels (<0.5 mg/L
selenium).
DISCUSSION
Based on treatability testing results, it was found that successful treatment of selenium bearing
wastewater streams could be performed through use of waste stream segregation and treatment
using either iron co-precipitation with ferric chloride or using a highly reduced iron chemistry
with ferrous sulfate and lime chemistry. Due to the nature of selenium speciation and oxidation
states found in the wastewater streams evaluated, selenite was found to be the most readily
available for removal and was successfully treated using ferric chloride addition and pH
adjustment to approximately 6.0. On the other hand, selenium in the selenate form was much
more difficult to remove, as it required a reduction step using a highly reduced environment
created by ferrous sulfate and lime addition. Due to the elevated chemical dose requirements
needed to create an adequately reduced environment for selenate removal, cost feasibility was
identified to be a challenge.
A wide range of treatment approaches were evaluated during the bench-scale treatability testing.
However, many of these technologies were found to be either unsuccessful or did not provide an
adequate confidence level for consistent removal efficiencies. The following approaches were
evaluated:
With regards to chemical addition treatment for selenium removal, multiple ZVI chemistries
were evaluated and determined to have varying removal efficiencies. However, in all cases ZVI
was found to produce an effluent that exceeded the target selenium concentration of 0.5 mg/L,
and results were found to be variable. Ferrous sulfide chemistry was found to have success on
the selenite bearing wastewater streams (Waste Streams 1A and 1B); however, the removal
efficiency was consistently found to be less than that of the ferric chloride addition approach,
likely due to the influence of nitrate present in the waste stream.
Further, a wide range of polishing treatment resins and media were evaluated. These evaluations
included both isotherm-style testing and subsequent flow through column testing on the most
promising resins. Effluent results from polishing treatment were inconclusive, as mixed results
were obtained. Given that effluent selenium levels could be reduced below future permit limits
using chemical addition treatment only, it was found that polishing resins or media were likely
not required. In general, polishing resins and media were found to be most successful in
removing selenium in the form of selenite.
CONCLUSIONS
Given the treatment goals identified for the project, that is, removal of selenium to levels less
than or equal to the new facility permit limit of 0.5 mg/L, it was found that project success could
be achieved through use of wastewater stream segregation based on selenium speciation and
treatment through ferric chloride addition or ferrous sulfate and lime chemistry. Additionally,
the purpose of the project was to identify the most cost effective approach that did not constrain
production at the facility.
Although the laboratory testing results indicated technical feasibility for treatment of both
selenite bearing and selenate bearing wastewater streams after stream segregation, ferrous sulfate
and lime chemistry for the treatment of Waste Streams 2 and 3 was found to be not cost
effective. This cost feasibility determination was based on both the high chemical dosage
requirements and sludge management and disposal, as it was found that settled sludge would be
approximately 40-45% of the treated volume after a 30-minute settling period.
Therefore, in conclusion, it was found that treatment of selenium bearing wastewater streams at
the facility could successfully be performed through the following steps:
ABSTRACT
Alternative approaches were developed and costed to treat and recover RO reject generated from
high purity water production within an electronics facility located in the southwestern US, where
reject disposal to a surface water body or wastewater treatment plant would exceed allowable
total dissolved solids (TDS) limits. Technologies utilized in the development of alternatives
included those that employed a combination of precipitative softening using lime, caustic soda
and soda ash implemented using pellet and conventional high-rate solids contact reactors and
tubular microfiltration; granular media and membrane filtration; brackish and seawater RO using
conventional and special configurations (e.g., HERO™, OPUS™, and disc tube); followed by
mechanical-driven evaporation and crystallization. This arrangement was compared with an
approach that used only evaporation and crystallization. The alternatives analysis revealed that
the all-distillation alternative, although higher in capital cost, was estimated to have lower life
cycle cost along with several non-cost benefits.
BACKGROUND
This paper describes a water management study conducted in the arid southwestern U.S. to
evaluate treatment options for a high salinity RO reject that is produced as a waste stream from
high purity water production. Increasing RO reject flows, coupled with a fixed volume within
existing solar evaporation ponds for concentrate disposal, required the development of new
approaches to reducing overall reject flows. Furthermore, water supply limitations required the
treatment and reuse of the clean water recovered from the reject.
Technologies considered in the development of reject treatment and reuse schemes for this study
included precipitative softening using lime, caustic soda and soda ash followed by conventional
and novel solids clarification processes; tubular and hollow fiber micro- and ultra-filtration,
nanofiltration (NF), electrodialysis reversal (EDR) and various configurations of RO (brackish
and seawater spiral wound, high efficiency RO [HERO™], and disc tube); thermally and
electrically-driven evaporation; and passive and enhanced solar evaporation ponds.
The cumulative total RO reject produced from the manufacturing campus’ high purity water
production facility is estimated to be approximately 2,000 gpm. This reject is essentially
concentrated city (potable) water. City water is sourced from groundwater from local wells and
surface water supplied from state or regional storage and conveyance systems. For the purpose of
this study, historical groundwater data was assumed as the water supply since it has higher
concentrations of TDS and inorganics that would result in the production of higher-salinity RO
reject. To estimate the quality of the RO reject, IMSDesign™ software was used based on the
RO system design summarized in Table 1. The historical groundwater and projected RO reject
quality are presented in Table 2.
Lime and soda ash – Lime is added to increase pH and soda ash to provide additional carbonate
to precipitate the desired amount of calcium, as well as strontium and barium and fluoride.
Significant silica removal is also accomplished as silica co-precipitates with magnesium
hydroxide. The sludge produced is easily dewatered.
Caustic soda and soda ash – This chemistry is similar to lime softening except that sodium
hydroxide is used in place of lime to raise pH. Softening performance is similar to that using
lime, however less solids are generated but chemical costs are higher due to the higher cost of
caustic soda when compared to lime. The resulting solids are not as easily dewatered as
with lime.
Once formed, the precipitated solids must be separated from the softened water, either by gravity
or by a physical solids-liquids separation process such as a membrane.
The equipment types that were considered for precipitation and solids separation processes are
presented below in Table 3.
Limited bench testing was combined with process modeling of the various precipitative softening
approaches to determine which would be most cost effective based on chemical doses and
reaction times needed to produce the target softened water quality. Bench testing was particularly
valuable as an adjunct to modeling in that it identified the need for either increased reaction
times and/or increased chemical doses to achieve the desired removals of calcium and
magnesium. Combining the results of the modeling, bench testing, site specific requirements
(e.g., footprint), together with cost, lime and soda ash chemistry executed using a high rate
SCRC was considered to be the most cost-effective and reliable softening and clarification
process to achieve the desired water quality. Table 4 presents the predicted quality of the
precipitative softened reject and Table 5 presents the predicted operating conditions of
the SCRC.
and silica, and would require polishing using additional processes (such as RO). For these
reasons, EDR was not considered any further. Table 6 provides a summary of the three
desalination systems examined and the respective advantages and disadvantages of each.
In comparing HERO™ (or OPUS™) to conventional RO, the primary value of the two
proprietary technologies lies in their ability to achieve maximum recovery without being limited
by mineral precipitation, particularly that of silica. However, the scaling potential of the softened
reject (as shown in Table 3) is sufficiently low that a similar recovery can be achieved both
conventional RO and HERO™/OPUS™ because recovery is limited by feed (osmotic) pressure
rather than saturation of sparingly soluble salts. In both cases, a nominal recovery of 89 percent
can be achieved based on a maximum feed pressure of 1,065 psig, reducing the brine flow from
2,000 to 215 gpm. This is illustrated in Table 7. With the softened RO reject, maximum recovery
is in large part dictated by the salinity of the feed. Based on achieving similar recoveries,
conventional RO was considered to be the most appropriate desalination technology because it
has the lowest capital and operating cost. Still, it should be noted that HERO™ and OPUS™
offer additional process and operational benefits over conventional RO when treating a surface
water-based RO reject because of improved TOC removal and reduced organics fouling at high
feedwater pH. Additionally, prevention of mineral precipitation with HERO™ and OPUS™ is
not dependent on the advertised performance of a scale inhibitor. If the RO reject contained
higher levels of silica, HERO™ or OPUS™ would have been the preferred membrane
desalination process.
A technical and cost evaluation of these technologies indicated that the thermal evaporative
processes is the best choice for RO concentrate volume reduction, primarily due to the
significantly higher recovery that can be achieved. Even at 1,600 psig feed pressure, the disc-
tube RO technology can only reduce the conventional RO concentrate volume by 33 percent,
whereas mechanical evaporation can achieve 67 percent recovery, allowing the conventional RO
concentrate flow be reduced to 70 gpm. Recovery of the former is again limited by osmotic
pressure considerations inherent with the RO process whereas recovery with the mechanical
evaporator is limited by sodium sulfate solubility. The TDS of the brine discharged from the
mechanical evaporator is estimated to be 270,000 mg/L consisting mainly of sodium sulfate and
sodium chloride.
Evaporation Ponds
In addition to evaluating softening, membrane desalination and distillation technologies for
volume reduction and water reuse, an evaluation of the existing solar evaporation ponds was
conducted to determine their ability to evaporate the brine from the mechanical evaporator. The
existing pond system is comprised of 5 ponds with a maximum depth of 5 ft and a maximum,
combined, top surface area of 27 acres. An evaporation pond model was developed by
CH2M HILL to determine the evaporative and hydraulic capacity of these ponds. Even though
the ponds are all slightly different in size, for modeling purposes, it was assumed that the
27 acres was divided equally over all five ponds. Other inputs to the model besides pond area
and depth include side slope ratio, monthly pan evaporation rates for the location region,
monthly rainfall/precipitation data, as well as quantity and salinity of the brine flow.
Based on these inputs, the model was used to calculate real-time values for water level, salt
accumulation, and ultimately, the duration of pond operation until the time that either overflow
occurs or the ponds must be taken out of service to remove the accumulated salt. Pond water
level is directly related to the water (brine) volume in the ponds. The brine TDS concentration in
The costs presented show that the base case has a very high O&M cost, mainly because of the
chemicals and solids disposal requirements associated with precipitative softening. These costs
were significant enough to investigate an all-distillation approach that included multiple
mechanical evaporators and a crystallizer. Table 11 presents the capital and O&M costs for this
approach and shows that while the alternative has higher capital cost and annual energy
requirements, there is a net savings in O&M costs. A comparison of the life cycle costs shows
that the alternative approach would result in cost savings of $30 million over 20 years (at 10
percent discount rate). In addition, this alternative would produce higher quality product water,
be much simpler to operate due to the reduced number of unit processes and would eliminate
truck traffic associated with delivery of lime and soda ash as well as removal of lime-based
sludge. Because the operating cost of all-distillation approach is nearly all from electrical energy
use (minor amounts of chemicals are used), it can also be considered a more sustainable solution
if the electricity is provided from renewable sources (solar or wind or both). Based on these
advantages and lower life cycle cost, the all-distillation approach was recommended for RO
rejection volume reduction and water recovery (reuse).
present value cost, despite a higher capital cost. The distillation-only alternative uses significant
power (~10 MW), but the power costs are less than the cost associated with chemical use and
solids disposal of the base alternative that uses precipitative softening. In addition, the
distillation-only process is less complex, and therefore easier to operate, especially given the
absensce of lime and soda ash slurry and feed systems and the need for proper chemical dosing
and precipitation pH control.
The findings from this case study illustrate that disposal of brackish waste streams, exemplified
here by a RO reject produced from purification of a higher salinity municipal water supply, is
both challenging and expensive. While the use of membrane-based desalination may appear to be
less expensive based on its considerably lower energy use, the need for chemically intensive
pretreatment can result in a higher total water cost than thermal based desalination methods,
depending on what degree of volume reduction is necessary. As importantly, the use of
evaporation ponds for final (zero) liquid discharge must carefully consider a number of factors to
ensure that adequate pond capacity is provided, particularly when disposing of hypersaline
brines, where evaporation rates are low salt accumulation is significant.
The results presented herein underpin the need for an effective industrial water management
approach that moves away from the ‘use it once and discharge’ approach to one that balances
high purity water production with sustainable management of saline waste steams with the
objective of maximizing recovery and reuse and minimizing impact to the environment.
ABSTRACT
Effluent from a coke oven by products plant contains elevated concentrations of organic and
nitrogenous compounds and mostly in the waste ammonia liquor. Due to more restrictive effluent
limitations guidelines from USEPA, it was evident that advanced wastewater treatment
technology would be required before plant production could resume.
This paper will focus on the subsequent design, seeding, start-up and operation following a
successful pilot scale demonstration on a coke plant’s wastewater using an immersed MBR.
Results demonstrated 80%-99% removal of targeted compounds following partial and full scale
start-up of the coke ovens in early-2014. Process design and performance trending data will be
presented for the full scale MBR treatment system, the first ever for a coke producer in North
America.
INTRODUCTION
ArcelorMittal Monessen operates a Coke Production Plant on the Monongahela River located in
Monessen, Pennsylvania. Typical coke production is approximately 1,000 tons per day derived
by heating coal in the absence of air in refractory overs. By-products recovered include crude
coal tar, elemental sulfur, ammonium sulfate, and crude light oil. Effluent from a coke oven by-
products plant contains elevated concentrations of organic and nitrogenous compounds from
sources including coke oven gas desulphurization, crude light oil recovery, ammonia still
operation, coke oven gas condensates, final gas coolers and barometric condensers. The greatest
volume of effluent is waste ammonia liquor (WAL), used to scrub coke oven gas to condense
tars and moisture and is recycled at a high rate. WAL has high concentrations of BOD 5 , COD,
ammonia, cyanide, sulfide, benzene, phenols and polyaromatic hydrogen compounds (PAHs), all
of which goes to the facility’s by-product wastewater treatment system.
With the advent of more restrictive effluent categorical standards from the USEPA, it became
evident that more technologically advanced wastewater treatment techniques to the facility’s
conventional activated sludge treatment system were required. Tetra Tech and GE Water &
Process Technologies (GE) were commissioned by ArcelorMittal to conduct a 4-month pilot
*determine the effectiveness of the MBR to treat coke plant wastewater in achieving effluent
quality by the Effluent Limitation Guidelines for the Iron and Steel Making Industry CFR Part
420.
*demonstrate the capability of the MBR to successfully perform and achieve 75% to 99%
removals of targeted compounds, during both steady state and spiked loading conditions.
*collect necessary information (i.e., sludge retention time, sludge wasting rate, nutrient dosing,
etc.) to complete the design for a full scale MBR treatment system.
To evaluate the process technical feasibility, a pilot-scale system was erected and operated at the
ArcelorMittal Warren (OH) coke plant; this was required since the Monessen plant was non-
operational in “idle-hot” status since 2009, while undergoing repair preceding a capital
rehabilitation and replacement program until mid-2014.
The Tetra Tech/GE ZeeWeed® pilot-scale demonstration unit operated from September, 2011,
through January 2012. Since activated-sludge nitrification processing of coke plant wastewaters
is well known technology, the primary operational and maintenance objectives of the pilot plant
were: 1) demonstrate the operational feasibility of the membrane modules to resist fouling and/or
chemical degradation; 2) determine the attainable membrane wastewater flux rates; 3) determine
the membrane chemical maintenance cleaning and recovery cleaning requirements to maintain
membrane flux and permeability.
The pilot testing results indicated promising performance by the MBR process. Table 1
summarizes the relative removals for the critical parameters.
SYSTEM DESCRIPTION
As a result of the pilot test program’s success and subsequent correspondence with local and state
regulators, ArcelorMittal authorized Tetra Tech in February 2013 to perform comprehensive
design/supply services for a full-scale wastewater treatment system. It was important to the Owner
(ArcelorMittal) that the Engineering Services Contractor (Tetra Tech) accept total system
responsibility for the project’s duration, for not only managing its subcontractor (GE) and
equipment suppliers but for ensuring regulatory compliance during the start-up period and
performance testing. Figure 1 depicts a simplified flow schematic of the 150,000 gallon-per-day
(100-GPM nominal) wastewater treatment facility for satisfying the overall objectives.
The primary system component is the ZeeWeed® ultrafiltration (UF) MBR system by GE. The
ZeeWeed® based water treatment system is a low energy immersed membrane process that
consists of outside-in, hollow-fiber modules immersed directly in the feed-water.
Figure 2. ZeeWeed® Membrane Fiber (left), ZeeWeed ® 500 Membrane Module (right)
The MBR’s aerobic suspended-growth environment is maintained in the bioreactor via diffused
aeration and continuous mixing of the mixed liquor at concentrations of 10,000 mg/L. The net
effect of the higher biological solids concentration is that the reactor size and hydraulic retention
time are reduced while higher sludge ages (of 90 days or more) are maintained. The membranes
are configured as small diameter tubules that are bundled in modules and housed in a cassette for
easy handling. The membrane modules are immersed in the membrane tank, whereby wastewater
passes concentrically into the membranes under a small vacuum induced by pumping versus
under high pressure. This results in the solids being retained in the reactor, while the treated
wastewater (permeate) is discharged. Every 10 to 12 minutes, the process pump is reversed for
approximately 30 seconds to “backpulse” permeate back into the bioreactor; this helps to clear
the membranes and thus prevent fouling while maintaining outflow.
Airflow is introduced at the bottom of the membrane module to create turbulence that scrubs and
cleans the outside of the membrane fibers. This reduced the solids accumulation on the
membrane surface and also provides mixing within the process tank to maintain solids in
suspension. The ZeeWeed® 500 immersed membrane system can remove particles that are
greater than 0.04 microns in size. Furthermore, if dissolved components can be first converted to
insoluble components, the membrane will subsequently remove them as well.
Figure 3 (on the following page) illustrates the flow schematic for the entire wastewater
treatment system.
The entire system, except for the Equalization Storage Tanks and Bioreactor Tank are housed in
a 4,250-square-foot metal-sided pre-engineered building.
The wastewater treatment plant (WWTP) consists of preliminary treatment for oil/tar removal
and the biological treatment system for treatment of excess flushing liquor (waste ammonia
liquor), wastewater from coke oven gas (COG) desulfurization, by-product wastewater (from
light oil recovery), and some site stormwater. The ammonia still effluent is discharged to the
Equalization Tanks. From here, the wastewater is cooled via spiral heat exchangers to <85 F
and an oil coalescer for free oil removal to mitigate membrane fouling. Nutrient augmentation
for maintaining effective biodegradation in the MBR is supplied via phosphoric acid addition.
Sludge dewatering for maintaining effective MLSS concentrations in the bioreactor is provided
with a belt filter press. These filter cake solids are beneficially recycled to the coke ovens with
the coal. Effluent following MBR treatment is collected, sampled, and pumped to the
Monongahela River.
The bioreactor entails both anoxic zone and aerobic zone. Denitrification takes place in the
anoxic zone where nitrates and nitrites are reduced to nitrogen gas with the introduction of
carbon-rich (phenolic-laden) influent from the equalization tanks.
For every milligram of nitrate (NO 3 ) reduction, 3.57 milligrams of alkalinity (as CaCO 3 ) is
produced.
Nitrification occurs in the aerobic zone with process air addition and here the ammonia forms
nitrates and nitrites via autotrophic nitrifying bacteria.
For every milligram of ammonia that is oxidized, 7.14 mg of alkalinity (as CaCO 3 ) is consumed.
Therefore, the addition of caustic soda (KOH or NaOH) to maintain effective pH (6.85 – 7.15
range) in the bioreactor is essential for nitrification and good microbiological activity.
Following the system design in late August 2013, and pre-purchase of the long lead time delivery
equipment, Tetra Tech provided technical assistance during the WWTP’s construction, beginning
October 2013. System start-up commissioning began in early February 2014. The wastewater was
then seeded with activated sludge from the United States Steel (USS) Clairton Works beginning
March 2014 for three weeks to acclimate the MBR’s bioreactor tank. Since start-up, the system has
operated for nearly a year and has been in conformance with all regulatory criteria. Coke oven
wastewater was introduced in early April 2014. The bioreactor’s seed sludge became acclimated
and the MBR system began discharging effluent one week later.
Following hydrostatic (i.e., clean water) testing of the entire wastewater treatment system for
leaks and field programming verification of the automated PLC, the system was ready for
commissioning on February 3, 2014. The bioreactor was then seeded with thickened waste
activated sludge starting March 3, 2104, and for 20 days, as shown in Figure 4.
Adequate nutrients were provided at a Carbon: Nitrogen: Phosphorous ratio of C:N:P of 100:5:1
to support biological activity in the reactor prior to the addition of coke oven wastewater.
Sucrose (C), urea (N), and phosphoric acid (P) were added to the bioreactor in addition to caustic
soda for alkalinity control.
During a ten-day (one-week) period, until April 7, 2014, 100,000 gallons of remaining
Equalization tank wastewater from 2009, as shown in Figure 5, was introduced to the bioreactor
on a batch basis (at a rate of 5 – 10 gpm) for confirmation that the waste was biodegradable.
Bioreactor temperature, dissolved oxygen and pH were also monitored, as well as COD
concentration and phosphoric acid usage rates. Following aeration of the bioreactor, an OUR
expressing mg/L oxygen consumed per minute (mg/L O 2 /min) was obtained on the biosolids.
These results, which showed an increase in the OUR with a decrease in dissolved oxygen (DO),
demonstrated that the microorganisms were acclimating and that the waste was being
biodegraded. There were both active stalked and free swimming ciliates observed. Rotifers and
nitrifiers were also observed and quite active.
Table 2 presents the results of analyses for key constituents in the wastewater during the system
start-up and up to February 2015.
Start-up of the batteries (coke ovens) and coal charging followed for five days starting April 11,
2014, while generating new stripped liquor wastewater as feed to the Equalization Tanks and to
the new WWTP. Treated effluent from the WWTP began on April 16, 2014, and flows increased
to 70 gpm for the next 50 days, while the coke ovens reached full production capacity. Figure 6
depicts the aeration pattern of the bioreactor at the time of WWTP effluent discharge along with
the anoxic zone with spray water addition for foam suppression.
Key bioreactor operational parameters such as dissolved oxygen, hydraulic retention time, solids
retention time, mixed liquor suspended solids, nutrient addition, pH and sludge wasting are
discussed in detail in the following sections.
Table III summarizes the key operating conditions for the bioreactor including pH, dissolved
oxygen (DO) mixed liquor suspended solids (MLSS), solids retention time (SRT), average
hydraulic retention time (HRT) and food-to-microorganism ratio (F/M). The bioreactor has
operated at ambient temperature conditions, which averaged 84 degrees F and ranged from 60
degrees F to 95 degrees F. The DO was consistently higher than the optimal 1.5 to 2.0 mg/L due
to the desirability to maintain process blower airflow at optimal efficiencies. HRTs of 100 hours
enabled effective COD/BOD 5 removal objectives. An F:M ratio of 0.11 along with an SRT of
over 200 days yielded effective organic removals and sludge solids rates.
160000
140000
120000
100000
Flow (gpd)
80000
60000
40000
20000
18,000
16,000
14,000
12,000
MLSS and MLVSS (mg/L)
10,000
MLSS
8,000 MLVSS
6,000
4,000
2,000
SRT was controlled by wasting excess activated sludge from the membrane tank. During steady
state operation the SRT was maintained at 275 days. In order to maintain mixed liquor suspended
solids in the bioreactor of between 8000 to 10000 mg/L, it was determined that wasting 2,000
gallons of sludge per day from the 550,000 gallon bioreactor volume be attained.
0.14
0.12
0.1
F:M (kg COD/ KG VSS / D)
0.08
0.06
0.04
0.02
condensed in the coke wastewater and attributable to a majority of the organics requiring
treatment in the MBR. The system was designed to handle a range of phenols from 100-600
mg/L. Cyanides are formed during coke-making operation and most peak concentrations
appeared during the new coke plant re-start with new COG piping and ensuing corrosion.
Complex iron cyanides (Prussian Blue, Prussian Green) formed as fixed cyanides end up in the
liquor systems and made removal in the ammonia stills and WWTP difficult. The fixed cyanide
concentrations eventually subsided following metal passivation and barrier films forming on the
previously corroded metal surfaces. This is evidenced in late-2014 at which time normal cyanide
levels (10-30 mg/L as CNT) in the wastewater influent were attained. Figure 10 depicts the
relationships for influent wastewater ammonia, thiocynate, cyanide, and phenols.
1800
60
1600
50
1400
Phenol, NH3, SCN (mg/L)
1200
40
CNT (mg/L)
1000
30
800
600
20
400
10
200
0 0
RESULTS OF IMPLEMENTATION
20
NH3-N, SCN & CNT (mg/L)
15
10
Phenols
Phenols have shown over 99% removal rate in the aerobic zone of the bioreactor to levels below
0.01 mg/L. Thus; the MBR has demonstrated its ability to remove phenols at concentrations well
below the regulatory limit of 0.062 mg/L as shown in Figure 12.
0.045
0.04
0.035
0.03
Phenol (mg/L)
0.025
0.02
0.015
0.01
0.005
A more comprehensive listing of the MBR permeate characteristics for all constituents were
analyzed is depicted in Table 5.
Flux
Flux is a measure of the rate at which the permeate passes through the membrane per unit of
outside surface area of membrane per day. It is reported in units of gallons per square foot per
day (gfd). The net flux is a calculation that takes into account the frequency and duration of
backpulsing, accounting for the lost production time as well as the actual volume of permeate
consumed during a backpulse.
The targeted flux demonstrated throughout normal operations was 8-10 gfd. The method
incorporated to maintain the design flux included operations in a backpulse mode. During startup
(April 11 – June 2014) a conservative flux of 6 gfd was used to allow the system to acclimate
while coke plant production to full capacity increased. In June, the flux was increased to 8 gfd as
influent wastewater volumes increased. It can be noted that the increased membrane flux had
little impact on the transmembrane pressure.
Transmembrane Pressure
Transmembrane pressure (TMP) is the amount of pressure (psi) required to pull permeate
through the membrane pores. The ZeeWeed® UF system is designed to maintain a constant flux,
therefore, as the membrane becomes fouled, the transmembrane pressure increases. The
ZeeWeed® UF system is equipped with several tools for membrane fouling control. Air scouring
is used to remove solids away from the membrane surface while a backwash cycle removes fine
particles loosely inserted in the membrane pores. Also, chemical maintenance and recovery
cleanings are used to remove organic and inorganic membrane fouling and reduce operational
TMP. Although the ZeeWeed® UF system has a maximum TMP control setpoint of 8 psi, the
average operational TMP observed during normal operations was much lower at under 1 psi.
Since start-up, the ZeeWeed® UF system is able to successfully operate under 1 psi and well
within the TMP range.
An “N-1” test was conducted on one membrane train to demonstrate that one (of the two)
membrane trains could produce 150 gpm (maximum wastewater flowrate) for 24 hours. A TMP
of <3 psi was observed during the “N-1” test, but was restored to < 1 psi following a return to
normal flow conditions. Overall, the test was successful and produced the required flow and met
effluent discharge criteria.
For the entire eight-month duration since start-up, the operational TMP was at less than 10% of
the maximum allowable TMP. A minor step up in TMP was observed from September 22 to
September 29, where the TMP reached 40% of the allowable range but with only one of the two
trains operating. Through regular operations with both membrane trains operating, the TMP
observed is <10psi.
Backwash Properties
The backwash is used as a method of cleaning in the ZeeWeed® 500 process. Typically, every 12
minutes, flow is reversed through the membrane (backpulsed) for 30 seconds, pushing clean
water from the inside of the membrane lumen to the outside. The water used for the backpulse is
permeate that has been collected in the backpulse tank.
Backpulse flux refers to the rate at which the backpulse water (permeate from the backpulse
tank) passes through the membrane per unit of surface area of membrane. The backpulse flux is
typically set to equal the permeate flux. During normal operations, the backpulse flux is
maintained at 10 gfd.
Recovery cleans are performed less frequently, using a higher concentration of cleaning
chemicals. The objective of a recovery clean is to recover the ZeeWeed® UF membrane
permeability to a base line condition observed after the initial break-in period for the membrane.
Two different cleaning agents are used for maintenance and recovery cleans. Sodium
Hypochlorite at a concentration of 250 mg/L was used to remove membrane fouling caused by
organic compounds that have not been biodegraded in the activated sludge system as well as the
bio film that develops due to microorganism growth on the membrane surface. Citric Acid was
used to remove inorganic scaling that forms on the membrane surface. Maintenance cleans are
scheduled once every two to four weeks, or more frequent whenever the filtration output declines
by 10-15% or the differential pressure increases by 15%. Due to low TMP’s, a recovery
maintenance clean was deemed not necessary to maintain ZeeWeed® UF membrane TMP. It is
planned for a recovery cleaning to be performed after one year of operation with a 1 g/L solution
of bleach and a 2 g/L solution of citric acid.
Once the bioreactor tank microorganisms were acclimated, the effluent quality except for
cyanide demonstrated that the MBR consistently exceeded the permeate requirements. By the
end of start-up, an HRT of 100 hours versus the design HRT of 86 hours consistently met the
removal targets. During steady state operation, the SRT was maintained at 225 days.
Good to excellent removal of all primary target compounds was attained. 80 to 95 percent
removal of BOD 5 , COD, TSS and soluble FOG is achievable. 95 to 99% removal of ammonia,
thiocyanates, phenols, benzo(a)pyrene and naphthalene is achievable. Cyanide removals of 75%
are attainable.
The bioreactor’s performance appeared to be most vulnerable to low pH, low alkalinity
wastewater conditions (i.e., below 6.85 pH). However, removal efficiencies appeared to improve
quickly in response to pH adjustment via the addition of sodium hydroxide. Optimum bioreactor
performance appeared to be in the pH range of 7.0 to 7.5.
The UF membrane system treatment capacity is successfully maintained through the use of
inorganic and organic fouling removal methods. Maintenance chemical cleanings utilizing
sodium hypochlorite and citric acid were implemented to further improve the membrane
condition. To date, no irreversible loss of permeability has occurred.
Flux rates of 8 gal/ft2/day (gfd) were maintained throughout operations. This accounts for losses
due to back pulsing of permeate via reverse pressure cleaning of the membranes to prevent
organic fouling of the membranes while maintaining outflow. Permeability rates of 80 gfd/psi
can be maintained during “steady state” conditions. Transmembrane pressures (TMP) of 0.1 to
0.2 psi before backwash were achieved, and considered to be well within industry standards for a
UF. Membrane maintenance cleaning of the UF membranes was induced every 2 to 4 weeks with
a 250 mg/L solution of sodium hypochlorite to remove any non-biodegradable organics. Citric
acid was administered with a 1,000 mg/L solution to remove inorganic scaling that forms on the
membrane surface.
Also notable is that the MBR demonstrated an ability to withstand relatively short periods of
insufficient oxygen supply such as could be encountered during a severe shock load or
emergency shutdown from a power outage. The response to this short-term anoxic condition was
primarily a reduction in biological activity, but one which appeared to recover upon return to
aerobic conditions.
Periodic microscopy of microbiology showed a viable population present since start-up. The
operation supports the following process design criteria: a dissolved oxygen supply of > 4.0
mg/L, an HRT of four days and an OLR of 40 lb. COD/day/kcf reactor volume.
Overall, the effluent quality is well within categorical standards for Coke Making Subcategory
40 CFR 420 and is able to consistently comply with all discharge requirements. No irreversible
loss of membrane permeability has occurred throughout normal operations.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the cooperation and dedication of ArcelorMittal Monessen’s
Paul Champagne and Brian Bartels, and ArcelorMittal Warren’s Joe Magni for their efforts
throughout the project’s implementation. We also acknowledge the technical sales support,
system commissioning and start-up assistance provided by GE’s Water & Process Technologies
Bob Brazier and Glenn Conners.
* Email: BillMatheson@Duraflow.biz
ABSTRACT
Steel manufacturing plants are large water consumers that use a tremendous amount of water in
their production process. As a result of the increasing demand in resource sustainability, high
quality process water and stringent discharge compliance, the U.S. steel industry has faced tough
challenges in smart water management to survive today’s global economic environment. The
challenge is compounded by the complexity of the contamination of wastewater characterized by
their high concentrations in dissolved solids, suspended particles, toxic metals, difficult-to-treat
chemicals and complex organic compounds. Recent studies have revealed the significance of
proper waste stream segregation, coupled with proper membrane filtration, to achieve maximum
water quality and reuse at minimal capital and operating cost. This presentation shall identify the
various wastewater sources of concern and assess the associated technical challenges that need to
be recognized and overcome by the industry. The discussion will present exploration of the
conceptual approaches and treatment alternatives via a pilot test program along with a full scale
system case study. The case study will depict the design, engineering, and operations details of a
Tubular Membrane/Reverse Osmosis (RO) wastewater recycle plant installed in a large mid-west
US high speed steel plating mill as the replacement of a conventional system to meet the NPDES
(National Pollutant Discharge Elimination System) permit requirements and the company’s long-
term water resource management objectives.
INTRODUCTION
Steel manufacturing plants are large water consumers that use a tremendous amount of water in
their production process. As a result of the increasing demand in resource sustainability, high
quality process water and stringent discharge compliance, the U.S. steel industry has faced tough
challenges in smart water management to survive today’s global economic environment. The
challenge is compounded by the complexity of the contamination of wastewater characterized by
their high concentrations in dissolved solids, suspended particles, toxic metals, difficult-to-treat
chemicals and complex organic compounds. This presentation shall identify the various
wastewater sources of concern and assess the associated technical challenges that need to be
recognized and overcome by the industry. The discussion will present exploration of the
conceptual approaches and treatment alternatives via an extensive pilot test program along with a
full scale system case study.
This mid-west steel plating mill was facing increasingly tighter discharge limits on its industrial
wastewater NPDES permit. At the same time, their existing wastewater treatment system was
nearing the end of its expected life. Maintenance costs were rising while reliability was
declining. It was clear that a new system was in order and objectives were established. Several
technologies were reviewed with regard to meeting the priorities and it was decided that the
technology that best suited their wastewater stream and their objectives was tubular membrane
microfiltration (MF) followed by reverse osmosis (RO). The other technical options considered
were conventional lamella clarifier, polishing filtration and RO, ion exchange and evaporative
technologies.
This plating mill has three high speed plating lines in operation, plating strip steel. The
predominant metals plated are nickel, zinc, and occasionally cobalt. The nickel baths are mostly
Watts nickel and the zinc is acid zinc. They also do acid pickling and alkaline cleaning. The
spent acid pickle baths are sent to the wastewater treatment system. The spent aqueous alkaline
cleaners are mainly hauled away.
Plant Survey
Prior to the commissioning of a pilot trial to confirm the efficacy of the chosen tubular
microfiltration and reverse osmosis technology, a plant inspection was conducted to determine
what, if any, streams might need to be segregated for pretreatment or treated outside the water
recycling system. As is common in most plating operations aqueous cleaners are part of the
process and their potential negative impact was discussed. Plans were already in place to
segregate the cleaners and to haul the spent concentrates. If allowed to come to the treatment and
recycling system, spent cleaners would foul the membranes with organic compounds and
severely impede the operation.
MF Process
The chemically pre-treated wastewater is processed through the membrane microfiltration
modules designed for separation of the precipitates from water. The wastewater is pumped at a
velocity of 3.6 - 4.5 m/sec (12 – 15 ft/sec) through up to 18 membrane modules (Figure 1)
connected in series (Figure 2). The turbulent flow, parallel to the membrane surface, produces a
high-shear scrubbing action which minimizes deposition of solids on the membrane surface.
During operation, filtrate permeates through the membrane, while the suspended solids retained
in the re-circulation loop are periodically purged for further de-watering. An automatic back-
RO Process
Reverse osmosis is a desalination process that has been employed in the metal finishing industry
to purify raw water (e.g. city water) before use as rinse water, recover plating chemicals from
rinse water, and polish wastewater treatment effluents (usually for reuse as rinse water). The
effectiveness of the RO process is highly dependent on the ability of the pretreatment process to
remove scale forming elements and other contaminants that foul RO membranes.
PILOT METHODOLOGY
- Microfiltration skid with concentration tank and a single 1-tube module. The membrane
was 72 inches (1.8 m) long and 1 inch (2.5 cm) in diameter with 0.14 sq m (1.5 square ft)
of membrane surface rated at 0.1 micron nominal porosity.
- Activated carbon was added to the concentration tank of the microfilter as a precaution to
remove organics such as oil and grease but also to adsorb organics that can foul the
downstream RO system;
- Reverse Osmosis system consisted of a high pressure pump feeding one 10.2 cm by 101.6
cm (4 inch by 40 inch) spiral wound TFC RO membrane with 7.2 sq m (78 square ft) of
surface area.
In stage one, site preparation was executed while the microfiltration skids were being built. Site
preparation included vacating the space required for the microfiltration skids and installation of
piping and electrical runs around existing operations. The existing filter press was moved to a
new location with a new superstructure and the clarifier was re-plumbed to become the
equalization and holding tank. Temporary diesel pumps were installed to pump the wastewater
from the collection pits to the reactor tanks by-passing the clarifier in order to confirm
performance of the microfilters and allow cleanout of the clarifier and switch-back to clarifier
operation if necessary.
In stage two, the microfilters with module specifications and system configurations presented in
Table 2 and Table 3, respectively, would be installed and commissioned. There would be a
period of time that they would run proving performance and compliance with the NPDES
discharge requirements. At this point, the system could still be switched back to the previous
conventional clarifier and leaf filter mode of operation at any time. After proof of performance,
the clarifier became the Equalization and Holding tank and the diesel pumps would be retired.
Once confidence was established that the microfilters were working, decommissioning of the DE
leaf filters would begin to make room for the carbon columns and the reverse osmosis systems.
In stage three, the ROs and carbon columns would be installed and started.
By September, Stage 1 was complete and the microfilters were installed. By late October, the
microfilter was started up and operated according to the operating condition specified in Table 4.
The microfilters continued to process all of the wastewater and discharged compliantly to the
NPDES permitted discharge. The contaminants of concern for the raw wastewater, MF filtrate
and RO permeate are summarized in Table 5. Between late October and December the process
was fine-tuned, chemically optimized, and further management of the cleaners segregation plan
were accomplished. By the end of 2011, stage three was begun. The ROs and carbon columns
were installed and commissioned in February of 2012. As shown in Figure 5, each RO skid was
designed as a 5 feeding 2 arrays, producing 200 GPM of permeate and recycles 30 GPM from
the first array concentrate back to the raw water feed. The 30 GPM added to the 250 GPM feed
from the microfilters adds up to a total first array feed of 280 GPM.
By the end of February the RO systems were fully operational and the entire system was
recycling up to 80% of the wastewater treated.
SUMMARY
The initial company water management objectives were met, as well as the secondary objectives
and more. The reuse of the RO recycled water (75+% of all wastewater influent) throughout the
facility proved to be very effective. By providing a more pure source of process water, chemical
demands were decreased at the plating operations, boiler house, and DI regeneration process
areas resulting in significant annual savings?). Metals reduction was stabilized at a higher
efficiency achieving consistent effluent quality meeting the more stringent compliance standards.
Sludge disposal costs have decreased?) and sludge quality is such that it is highly desired by the
offsite smelter. Other added benefits included improved environmental and safety liabilities,
electrical and control enhancements, waste minimization at plating lines, improved operator
knowledge, better use of existing plant space, as well as upgraded water management practices to
sustain long-term business growth. The RO brine is of such quality that the facility is now
pursuing discharging to the city sewer and surrendering their river NPDES permit, improving the
local surface water qualities and thus benefiting the local communities. The implemented water
quality management practices resulting from the full scale installation has prepared the facility
for upcoming regulatory compliance demands, streamlined wastewater treatment, and promoted
green-technology water savings.
ACKNOWLEDGEMENTS
We would like to thank the host facility for giving us the opportunity to perform this pilot study
and full scale installation, especially the management, operators, and process personnel. We
would also like to thank J&T Technologies of Pottstown, Pennsylvania for execution of such a
large scale installation wastewater project, Crane Environmental of Tampa Florida for their
expertise in reverse osmosis, and also Water and Wastewater Laboratories for their dedicated
analytical lab services.
REFERENCES
Morton, Craig A. (Jun 2010) Summary Report: Micro-filtration/Reverse Osmosis Pilot Study,
Midwest Steel Plating Mill, Host facility proprietary information.
ABSTRACT
A major chemical adhesive products business operates a manufacturing plant in central Pennsylvania.
The production process involves the formulation of adhesives that use organic polymers, anionic and
nonionic surfactants, solvents, alcohols, oils, acids, and bases in the process. The site operates a 25,000
gallon per day wastewater treatment system consisting of a chemical precipitation process designed for
removal of particulates in the wastewater and a plate and frame press for solids dewatering. The plant
was recently notified of new, more stringent, effluent Biochemical Oxygen Demand (BOD) and Total
Suspended Solids (TSS) requirements for discharge to the local Publicly Owned Treatment Works
(POTW). The original treatment system was designed for particulate removal and did not have the
ability to meet the new effluent limits. A combination of unconventional physical/chemical treatment
and biological treatment methods were applied at bench scale.
INTRODUCTION
A major chemical adhesive products business operates a manufacturing plant in central Pennsylvania
that produces adhesives used in label and packaging materials. The production process involves the
formulation of adhesives that use organic polymers, anionic and nonionic surfactants, solvents, alcohols,
oils, acids, and bases in the process. Wastewater generated during the manufacturing process is
directed to an on-site wastewater treatment system (WWTS) for pretreatment prior to discharge to the
local Publicly Owned Treatment Works (POTW) facility. The POTW has imposed new “Local Limits” that
include more stringent limitations for specific parameters including Biochemical Oxygen Demand (BOD)
and Total Suspended Solids (TSS). The existing WWTS has a design capacity of 25,000 gallons per day
and consists of a chemical precipitation process designed for removal of particulates in the wastewater
and a plate and frame press for solids dewatering. The existing WWTS does not have the ability to treat
organics in the wastewater or efficiently dewater chemically precipitated solids.
SUMMARY OF PROBLEM
The plant management team planned to increase production at the facility and make improvements to
the original wastewater treatment system. The site and corporate wastewater project team was given
the challenge to handle the proportional increases in WWTP flow/loading with BOD and TSS
concentrations up to 1,600 mg/l and 440 mg/l, respectively, and meet the new BOD and TSS
Page 1 of 28
pretreatment limits (BOD 300 mg/l and TSS 250 mg/l) with minimal capital expenditures. The site was
also tasked with a company-wide initiative to minimize landfill disposal, which required improving
overall solids dewatering. The adhesive nature of the particulates within the wastewater causes a
significant problem within the collection/conveyance system and the onsite wastewater treatment
plant. An assessment of the original wastewater treatment infrastructure showed significant issues with
the existing equipment (i.e. tanks, piping, and pumps at the end of useful life), and expansion
constraints due to limited available space. It was determined that the existing chemical precipitation
system required replacement and relocation to allow for expansion. Similarly, the existing plate and
frame press required significant attention to operate due to the composition of the solids, therefore, the
facility wanted to explore other alternatives for dewatering. In addition, the majority of the BOD
discharged from the original pretreatment system was in soluble form and therefore required an
alternative treatment process to achieve the new effluent limitations.
One of the major challenges with the plant’s wastewater is the adhesive nature of the suspended solids
within the wastewater, which required several bench and pilot treatment processes to be evaluated.
The main focus of this paper is the innovative and unconventional bench and pilot studies that were
conducted which ultimately helped the facility choose an effective treatment process for management
of wastewater and solids generated from the facility.
APPROACH
The purpose of the project was to identify the “most appropriate” wastewater treatment process to
meet more stringent BOD and TSS pretreatment limits with an eye on improving solids dewatering
processes at the plant. Many treatment processes for soluble BOD removal (after chemical precipitation
for solids removal) were evaluated at the desktop level and screened for viability for the unique
characteristics of the wastewater. Subsequently, six processes were evaluated at the bench level.
These included:
• Activated Carbon
• Chemical Precipitation
• Advanced Oxidation (Fentons and Ferrate Treatment);
• Biological Treatment;
• Air Stripping; and
• Solids Dewatering
The goal of the BOD removal evaluation was to cost effectively meet effluent BOD limits while
minimizing adhesive suspended solids impact on the treatment process. Several solids dewatering pilot
tests were also conducted on the slurry from the existing chemical precipitation system using screw
press and rotary drum filter press technologies and were compared against the existing plate and frame
process. An innovative approach for significant BOD and TSS removal included the use of a centrifuge as
the primary wastewater treatment technology (i.e. in place of the chemical precipitation system and
solids dewatering system) to meet pretreatment limits. All treatment technologies tested at the bench
and pilot scale were compared against each other with respect to: technical effectiveness, compliance,
Page 2 of 28
and other parameters, where applicable, including footprint, capital cost, operational cost, chemical and
labor requirements, robustness, safety, and turnup/turndown capability.
WASTEWATER CHARACTERIZATION
In response to the proposed new discharge limitations and in order to assess the impact of increased
production, the Facility implemented a program to evaluate alternatives to meet new discharge criteria
established for their industrial wastewater pretreatment system. As a first step, the Facility collected
analytical data to identify primary sources of Biochemical Oxygen Demand (BOD), Chemical Oxygen
Demand (COD), Total Suspended Solids (TSS), Nitrogen, Phosphorous, and other compounds in their
process wastewater, and collected analytical data to evaluate the reduction of these parameters
through the existing wastewater treatment system. This wastewater characterization included four key
objectives, including 1) primary source identification to evaluate opportunities for source reduction
and/or substitution; 2) treatment system performance data to evaluation existing BOD, nutrient, and
other constituent removal; 3) assessment of BOD / COD composition to identify the soluble portion of
BOD and COD in relation to total BOD/COD to refine potential treatment techniques; and 4) collect data
on other parameters that may impact the operation of the proposed treatment system.
Analytical results collected from the key process waste streams indicated a high variability of the
process wastewater composition. This data also indicated that soluble BOD is a significant portion of the
total BOD, and that the average ratio of COD concentration to BOD concentration was approximately 60
to 1. Process wastewater was further categorized by process generation into two main categories: a)
typical rinse wastewater; and b) High Pressure Cleaning (HPC) wastewater. Figure 1 provides a
comparison of average COD, TSS, and TDS in the Rinse and HPC wash water.
120,000
Concentration (mg/L)
100,000
80,000
Rinse
60,000
40,000 HPC
20,000
-
COD Sol COD TSS TDS
Figure 1: Average COD, TSS, and TDS in Process Rinse and HPC Wastewater
Review of the potential BOD contributions for source elimination and/or reduction indicated that
surfactants may be a major contributor to soluble BOD; however the stable chemical bonds within the
surfactants may be difficult to treat with conventional treatment options. Material substitution and/or
elimination was considered and removed due to the quality requirements in the manufacturing process.
Analytical data was also compiled from samples collected from the main plant sump (MPS), equalization
(EQ) tank outlet (prior to any chemical addition), and from the clarifier outlet, which represents the
Page 3 of 28
treated effluent quality that is discharged to the POTW. Average wastewater concentrations of most
constituents in the MPS were similar to those from the process Rinse Water, whereas concentrations in
the subsequent unit processes (equalization tank, and clarifier effluent) indicate significant removal of
BOD, TSS, and COD. Figure 2 provides an indication of BOD and COD concentrations from the process
water through the treatment system.
Concentration (mg/L)
1,000,000
100,000
10,000
1,000
100 COD Total
10 BOD Total
1
Rinse HPC Inf EQ Clarifier
Sump Tanks
Process or Treatment Unit
Soluble BOD is approximately 70 to 80 percent of total BOD in the EQ Tanks and Clarifier, as indicated in
Figure 3.
Concentration (mg/L)
2,000
1,500
1,000
BOD Total
500
BOD Sol
-
Rinse HPC Inf EQ Clarifier
Sump Tanks
Process or Treatment Unit
Figure 3: Average Soluble BOD and Total BOD Concentrations by Process/Treatment Unit
Results from the wastewater characterization study indicated that a significant reduction in BOD and TSS
concentrations is achieved through the existing chemical/precipitation treatment system. However, it
does not appear that the treatment system consistently removes BOD and TSS to the proposed
discharge limits for the POTW of 300 mg/l BOD and 250 mg/l TSS, respectively. Since a significant
portion (approximately 70 to 80 percent) of the BOD is in a soluble form, the existing clarifier has limited
capabilities for additional BOD removal, and an additional treatment step will be required to meet the
proposed BOD discharge criteria. Extremely high COD to BOD concentrations in the EQ Tank
(approximately 60 to 1) and high COD to BOD concentrations in the Clarifier effluent (approximately 9 to
1) will have to be taken into consideration in the design of a biological treatment system.
Page 4 of 28
Further assessment of the WWTS indicated that infrastructure showed significant issues with the
existing equipment (i.e. tanks, piping, and pumps at the end of useful life). The existing WWTS consists
of equalization, chemical addition, and clarification. Solids from the clarification process are dewatered
with a plate and frame press and hauled off-site for disposal. The treatment system is designed to treat
up to 25,000 gallons per day of high organic-strength wastewater. Space constraints within the existing
treatment building do not allow much room for new treatment equipment or systems. A schematic of
the WWTS is provided in Figure 4 below.
Page 5 of 28
An initial screening of alternatives was completed to identify options to address the unique composition
of the Facility’s wastewater, including the high variability of the influent wastewater loadings; the stable
nature of the chemical bonds within the surfactants contributing to the waste stream; the high ratio of
soluble BOD to overall BOD, and the high COD to BOD ratio. Table 2 provides a summary of the initial
screening of potential treatment options and considerations for their use at the Facility.
Page 6 of 28
Based on this initial screening, bench-scale screening studies of end-of-pipe treatment options including
Activated Carbon, Liquid/Solids Separation Techniques [chemical precipitation], and Advanced
Oxidation with Fenton’s Reagent were completed on the effluent from the existing treatment system.
These tests were followed by bench-scale screening studies of a more ‘conventional’ approach with
biological treatment, and by an unconventional approach using a centrifuge to simultaneously remove
BOD and TSS and dewater solids. Results and findings are provided below.
The purpose of the bench-scale studies was to conduct a preliminary screening of potential end-of-pipe
treatment options to evaluate the technical and economic feasibility for BOD reduction in the
wastewater effluent to meet the new POTW Pretreatment limits. Results from the bench tests were
used to evaluate overall feasibility and preliminary ‘full-scale’ implementation costs, as well as to
identify operating constraints that may inhibit implementation at the Facility. The bench studies
focused on TSS, BOD, and COD removal, since TSS and BOD are the primary constituents of concern for
the discharge to the POTW.
The goal of the activated carbon bench test was to assess the technical and economic feasibility of using
powdered activated carbon to reduce BOD to meet POTW requirements. A sample of the facility
wastewater was tested in an offsite laboratory where it was mixed with powdered activated carbon.
COD and BOD were measured before and after the test.
During the test, approximately 100 grams (±10g) of powdered activated carbon (PAC) was added to an
empty 1000 mL beaker. This was repeated an additional two (2) times for a total of 300 grams of carbon
in three (3) 1000 mL beakers. 1,000 mL of pretreated effluent from the Facility was used for each test.
Figure 5 below is a photograph of unstirred Facility effluent and carbon; from the appearance, it appears
as though the emulsified materials have adhered to the carbon. Each beaker, containing approximately
100 g of carbon and 1,000 mL pretreated Facility wastewater, was placed under a gang stirrer operating
at 100 rpm for 24 hours. Each beaker was covered with a lid to ensure no backsplash or loss of
constituents.
Page 7 of 28
After operating for 24 hours, the gang stirrer was shut off and the carbon was allowed to settle for
approximately 15 minutes. After 15 minutes, a majority of the PAC appeared to have settled. The carbon
treated effluent was then decanted and filtered through a cheese cloth to collect any residual unsettled
carbon. The decanted effluent was then filtered through a Granulated Activated Carbon (GAC) filter
twice to remove the smaller carbon pieces. This additional carbon filtering was not expected to have
had a significant effect on the effluent water quality because the contact time in the GAC filter was very
brief (on the order of 1-2 minutes) versus the previous 24 hour exposure to the PACT. The filtered,
carbon treated effluent was then placed in sample bottles for analysis of BOD, soluble BOD and TSS and
preserved by packing in ice.
Pretreated
Facility Carbon
Effluent Treatment Reduction
Parameter (mg/l) (mg/l) Q (%)
TSS 24 10 U 58.3
Q= Qualifier
U - Indicates the analyte was analyzed for but not detected.
Reduction % = 1 – (Carbon Treatment result / Effluent result)
Page 8 of 28
Results from the bench-scale study of PAC indicated that both soluble and total BOD were reduced by
>80%, and that soluble and total COD were reduced by >93%. Based on the results of the wastewater
characterization study, this removal efficiency would be adequate to meet the POTW limits of 300 mg/l
at an influent BOD loading of 800 mg/l1. However, the unstirred effluent and PAC appeared to have
emulsified material adhered to the carbon. Due to the ‘sticky’ nature of the Facility’s wastewater, it
appears likely that a granular activated carbon column would foul and/or clog during dynamic, full-scale
operation. Additionally, although a Freundlich Isotherm study could not be done to evaluate removal
efficiencies at different dosages, rough calculations, based on the test data, indicate an excessive carbon
usage (significantly less than <1% carbon efficiency [g BOD removed per g carbon]).
Based on carbon vendor input, it is common for high-strength wastewater with BOD and COD
concentrations similar to those observed in the bench-scale study to have a carbon usage of 1%. At full-
scale operation, up to approximately 110 pounds per day of BOD would need to be removed from the
Facility’s wastewater (at an average flow of 20,000 gallons per day, an influent concentration of 800
mg/l, and a target effluent concentration of 150 mg/l). This would require more than 10,900 pounds per
day of carbon. Granular activated carbon therefore does not appear to be a viable treatment option due
to the high carbon usage and appearance of emulsified material on the PAC.
Chemical Precipitation
As an alternative to carbon treatment, bench studies were also completed on the Facility’s wastewater
to evaluate other liquid/solids separation techniques, including bench tests with chemical-precipitation
processes. Coagulants are multifunctional products which a) convert some soluble compounds to
insoluble compounds; b) destabilize and precipitate colloidal suspended solids; and c) agglomerate
discrete suspended solids and induce settling. Flocculants are high molecular weight polymers, usually
derived from acrylamide, which react with small particles, causing them to coalesce into larger particles
that settle rapidly.
The primary raw materials used in production at the facility includes rosin esters, surfactants, and
emulsion polymers. All of these compounds contribute to BOD/COD. The purpose of evaluating
liquids/solids separation was to evaluate options for alternate chemical treatment that would improve
reduction of BOD in the wastewater. A series of inorganic coagulants, based on iron and aluminum salts
were evaluated. A series of low molecular weight, high cationic charge density organic compounds
derived from polyamine chemistry were also tested. These organic polymers are known to be useful in
removing colloidal solids generated by latexes and rosin esters. pH adjustment was also tested, as
changes in pH are known to destabilize colloidal solids, aiding in their removal.
1
Maximum BOD concentration observed in the Facility’s effluent during the wastewater characterization study =
801 mg/l.
Page 9 of 28
Bench test were completed with untreated wastewater from the influent equalization tank at the
Facility’s IWWTP, as well as on the pretreated effluent from the Facility.
For the untreated equalization tank wastewater bench study, a combination of a polyaluminum chloride
coagulant, in conjunction with a high molecular weight cationic (positively charged) polyacrylamide
flocculant produced the best results for treating the Raw Wastewater. Polyaluminum chloride was
dosed at 1200 ppm, and polyacrylamide was added at 20 ppm. This combination generated large,
rapidly settling solids, and non-turbid (crystal clear) water.
The addition of a flocculant (cationic polyacrylamide at 20 parts per million, ppm or mg/L) and
polyaluminum chloride (at 1,200 mg/L) provided a significant reduction in soluble COD and soluble BOD.
Results of the Chemical-precipitation bench study on the untreated equalization tank water are
provided in Table 4 below.
Untreated
Facility
IWWTP Coagulant
Influent Treatment Reduction
Parameter (mg/l) (mg/l) Q (%)
Q= Qualifier
U - Indicates the analyte was analyzed for but not detected.
Reduction % = 1 – (Carbon Treatment result / Effluent result)
Although the results from the bench-scale screening study show significant TSS, BOD, and COD removal
for the equalization tank wastewater, the results indicated only minor improvement over the existing
chemical precipitation processes currently employed at the Facility. No further studies on
liquid/separation alternatives for the equalization tank wastewater were recommended.
A bench scale study on Liquid/Solids separation (Chemical-Precipitation) was also completed on the
pretreated effluent from the IWWTP. A lower dose or both polyaluminum chloride coagulant and
polyacrylamide flocculant provided sufficient clarity and solids generation for the pretreated Facility
effluent. More specifically, a dose of 267 ppm of PAC and 13 ppm of polyacrylamide was required to
obtain non-turbid (crystal clear) water for the bench test. Results from the chemical-precipitation bench
Page 10 of 28
study on the IWWTP effluent showed no reduction in soluble or total BOD, and no further studies on
liquid/separation alternatives for the IWWTP effluent were recommended.
Advanced Oxidation Processes (AOP) refer to specialized chemical techniques whereby a strongly
oxidizing species, called the hydroxyl radical, is generated. The hydroxyl radical is the most powerful
oxidizing agent readily available for conventional purposes. Hydroxyl radicals can be produced via a
number of techniques, including injecting gaseous ozone into the wastewater, combining ozone and
hydrogen peroxide, or through conventional “wet” chemical reactions.
For purposes of the first AO bench study, a “wet” chemical technique, called Fenton’s Reagent, was
initially evaluated to assess the overall feasibility of advanced oxidation and, if results were favorable,
further studies could be completed with alternative AOP techniques.
The bench scale study was completed on samples of the Facility’s treated effluent, as well as the
untreated wastewater from the influent equalization tank. Test procedures included addition of three
reagents in sequential order, including:
The Fenton’s Process releases energy, which increases solution temperature. The reagent strength was
selected to minimize the exothermic (heat generating) reaction. Use of reagents at these concentrations
was considered safe for plant operations. The progress of the test was monitored by measuring
Oxidation – Reduction Potential (ORP). In most reactions, ORP was expected to increase from baseline
conditions, as the reaction progressed. A sharp increase in ORP would indicate that all of the organic
matter had been oxidized. Once the organic demand for oxidizing agent has been eliminated, then the
residual reagents show up as excess oxidant, and the ORP measurement increases proportionately.
Results of the bench scale study indicated that there was no increase in ORP during the test on
untreated wastewater. This was attributed to the high suspended solids concentrations in the untreated
wastewater. In simplified terms, it was not economically feasible to add enough reagents to degrade all
the suspended organic matter and generate excess reagent. When the estimated annual cost for
reagents exceeded $ 1MM, the experiment was terminated.
Page 11 of 28
However, results from the bench-scale screening studies for advanced oxidation using Fenton’s reagent
on the pretreated wastewater indicated that soluble and total COD was reduced by >50%, although BOD
results could not be fully evaluated due to laboratory issues. Analytical results for effluent BOD treated
with Fenton’s reagent were reported by the laboratory as non-detect (at a reported detection limit of
12.9 mg/l). Since the COD reduction was only approximately 50%, it is suspected that low pH conditions
in the treated wastewater or the presence of excess oxidation reagent may have killed off the biomass
and resulted in an erroneous BOD indication of zero (non-detect). The actual BOD % removal was
assumed to be closer to the observed COD % removal. To confirm actual BOD concentrations in the
treated effluent, additional analytical tests were performed after pH neutralization of the effluent.
For the second bench study with AO using Fenton’s Reagent, additional testing and analytical
procedures were implemented to address the potential pH and residual oxidant issues in order to better
evaluate BOD reduction, and to several tests were conducted at varying chemical dosages to evaluate
dosage requirements and costs for the full-scale system. These tests were completed with the Facility’s
pretreated wastewater, again using Fenton’s reagent (addition of sulfuric acid to lower the pH; addition
of a source of ferrous iron; and addition of hydrogen peroxide). To address potential pH issues with the
BOD analysis, the pH was neutralized as a final step with the addition of sodium hydroxide (NaOH)
approximately 6 minutes after the hydrogen peroxide was added.
Additional steps were implemented to address the issues observed from the initial AO bench studies
regarding the BOD analytical results. After treatment with the Fenton’s process (addition of sulfuric
acid, ferrous iron, and hydrogen peroxide), a 2% sodium hydroxide solution was added to all samples to
raise the pH back to the neutral range before the samples were packaged for analysis. During the tests,
hydrogen peroxide and ferrous iron were added to the wastewater in either a 10:1 ratio or a 5:1 ratio at
concentrations varying form 51 ppm to 154 ppm for ferrous iron and 248 ppm to 1405 ppm for
hydrogen peroxide. Three tests were performed at the 10:1 peroxide to iron ratio and three at the 5:1
ratio with ferrous iron concentrations at approximately 50, 100, and 150 mg/l for each of the three
tests. Table 5 below outlines the characteristics of each bench study sample run.
Hydrogen 2% Sodium
Bench Sulfuric Acid Ferrous Peroxide Hydroxide
Test Dosage Sulfate Dosage Dosage Peroxide to Dosage (mL)
Number (ppm) (ppm) (ppm) Iron Ratio
Page 12 of 28
Although the pH was neutralized with sodium hydroxide as a final step to address potential BOD analysis
issues, preliminary laboratory results and notes regarding BOD data indicated that additional steps had
to be implemented to address residual oxidant (hydrogen peroxide) in the treated samples. Initial BOD
results from the lab were reported as “N/A – The results increased with increasing dilutions…This
indicates a potentially toxic substance is being diluted out. The concentration may be biased low.” To
address these initial findings, the samples were re-analyzed by the laboratory at higher dilutions.
Additionally, the samples were pretreated using Standard Method 5210B Section B (vigorous shaking for
4.5 hours) to eliminate residual peroxide, which potentially could have contributed to the BOD toxicity
test issues. However, after implementing these procedures, the lab reported that BOD results still
increased with increasing dilutions, indicating that the samples were toxic to the biomass (i.e., the
biomass grew as the sample was further diluted.) BOD results were reported by the laboratory as non-
detect for the reanalyzed results. Based on the reported laboratory results and without actual treated
BOD concentrations to calculate removal efficiencies at different dosages, use of AO with Fenton’s
reagent was not considered to be a viable alternative for BOD reduction in the Facility’s wastewater (i.e.,
the use of Fenton’s Reagent most likely resulted in false BOD readings.)
Additionally, laboratory results indicated that COD and soluble COD concentrations increased during the
AO bench studies, for all tests. These results were not consistent with the 50% COD reduction observed
during the initial bench studies. Even after the treated samples were re-analyzed after being pretreated
using Standard Method 5210B Section B, results from the second round of analyses showed no
substantial changes in COD concentrations from the first analyses; in all cases, COD concentrations
increased with AO treatment, potentially due to residual hydrogen peroxide and ferrous iron
concentrations in the samples. Based on these results, use of AO with Fenton’s reagent was not
considered to be effective at COD reduction in the Facility’s wastewater (i.e., the use of Fenton’s
Reagent most likely resulted in false COD readings.)
As with COD concentrations, laboratory results indicated that TSS concentrations increased during the
AO bench studies, for all tests. The untreated TSS concentration was 23 ppm, and treated TSS ranged
from 68 to 445 ppm. This is likely due to an iron hydroxide precipitate, which forms during pH
neutralization, resulting in elevated TSS levels. In a full-scale Fenton’s treatment system, a settling step
would likely be required to remove ferrichydroxide sludge and address TSS concentrations prior to
POTW discharge.
Although AO treatment with Fenton’s Reagent does not appear to be a viable alternative for BOD
reduction in the Facility’s wastewater, other Advanced Oxidation technologies such Ferrate Treatment
were considered to be potentially viable treatment alternatives. Bench tests were recommended for
Ferrate Treatment to evaluate dosage requirements, removal efficiencies, and potential full-scale
treatment cost.
Based on inconclusive results from initial Advanced Oxidation treatability tests using Fenton’s Reagent,
treatability studies with other AO technologies such as Ferrate Treatment were then recommended to
Page 13 of 28
address the potential toxicity issues and, if shown to not be viable, ‘rule out’ AO as a potential treatment
alternative for the Facility’s wastewater. Ferrate treatability tests were subsequently completed on the
Facility’s wastewater to evaluate the feasibility of Ferrate treatment to reduce BOD to concentrations
below 150 mg/l, and, if Ferrate treatment is considered to be a viable option, to evaluate dosage
requirements, removal efficiencies, and potential full-scale treatment costs. As with the previous bench
studies for other AO technologies, this Ferrate Treatability study focused on TSS, BOD, and COD
removal, since TSS and BOD are the primary constituents of concern to the POTW.
Ferrate (Iron VI) Treatment is an advanced oxidation process in which Sodium Ferrate is synthesized to
act as an oxidant, disinfectant, and coagulant to treat wastewater. Ferrate is synthesized from a ferric
iron source such as ferric chloride or ferric sulfate, bleach, and caustic. The resulting compound, sodium
ferrate containing Fe(6+), is a strong oxidant capable of oxidizing organic and inorganic contaminants in
wastewater. Ferrate has reportedly been used successfully in various municipal and industrial
wastewater applications, including oxidation and subsequent reduction of inorganic and organic
constituents, metals removal, etc., but is still considered to be innovative technology in the wastewater
industry.
For the Bench Study, treatability tests were conducted on a sample of treated effluent from the Facility.
Ferrate was tested for its effectiveness in decreasing biological oxygen demand (BOD) in the sample
water. Before each treatment trial, the ferrate was synthesized using a proprietary process by the
Vendor. Using the concentration of the synthesized ferrate and the doses desired (in ppm) for each trial,
the amount of ferrate needed was weighed out, and added during treatment. Trials were carried out
using 300, 500, and 1000 mL samples. Mixing began at 65 rpm at which time the Ferrate(VI) doses were
added. The presence of Ferrate (VI) in the sample was tested using potassium iodide starch paper. Once
the ferrate was determined to be gone, the pH was adjusted using carbon dioxide (CO2) and ferric
sulfate (13.2% Fe2(SO4)3) to a pH of 6.0, 7.0, 7.5, 8.0, or 9.5. The pH of the sample during treatment
changes the life span of ferrate. Therefore, the optimal pH would facilitate the longest life of ferrate and
permit more time for oxidation. The mixing speed was then reduced to 20 rpm to allow coagulation, and
finally sedimentation of flocs. Samples were then filtered using a 2 μm filter, which is used to simulate
samples passing thru a typical media filter in the treatment process, and analyzed. Selective samples
were measured for COD and residual iron (FeR) using a laboratory instrument. Based on the findings
from these tests, selected samples were sent to an offsite NELAC certified private laboratory in Orlando,
FL, to be tested for BOD5.
Tests were completed with ferrate doses ranging from 35 to 500 ppm to identify the optimal dosage by
monitoring COD and residual iron concentrations in the lab. COD analysis was completed with the
laboratory instrument(s) before and after treatment with Ferrate to provide a quick evaluation of bulk
organic reduction, and residual iron tests were also completed in the laboratory to provide an indication
of how well the reaction worked, based on the unreacted iron left in solution.
Once the ‘optimal’ Ferrate dosage was identified in the treatment laboratory, based on COD reduction,
two samples of the ferrate treated water were submitted to an outside laboratory for BOD analysis.
Page 14 of 28
Results from the outside laboratory were inconclusive, as BOD concentrations analyzed by the outside
laboratory indicated an increase rather than decrease in BOD concentrations in the ferrate-treated
wastewater samples.
Based on the Vendor’s laboratory findings, including COD and iron concentrations monitored at their
laboratory, the Vendor reported that they were able to achieve a 39% removal of COD at a 50 ppm dose
of Ferrate. The 39% COD reduction was the highest COD reduction observed in their laboratory, and
the Vendor identified the optimal dosage of Ferrate to be 50 to 100 ppm. However, when the ‘optimal’
test was reproduced in order to obtain a representative ‘split sample’ for analysis by an independent
outside laboratory, the COD reduction analyzed by the outside laboratory was negligible at the ‘optimal’
dosage of 100 ppm of Ferrate. The Vendor subsequently provided input that the reproduced treatment
tests completed for analytical verification were actually done using a larger ‘scaled up’ treatment
volume, and that the intensity of mixing was a factor in the COD reduction at higher volumes – mixing at
higher volumes did not achieve the same results previously observed with ferrate treatment at lower
volumes.
Results on BOD from the two independent laboratories were substantially different. Reported COD
reductions varied widely, and the maximum reduction of 39% reported by the Vendor could not be
substantiated or reproduced in samples treated with the same Ferrate dosage and analyzed by an
independent laboratory. In some cases, BOD concentrations in the treated wastewater actually
increased even when COD concentrations appeared to decrease. These inconsistent and variable results
indicated that AO with Ferrate addition does not appear to be a viable option for the Facility’s
wastewater. Based on results of the Ferrate Treatability Study and previous AO Bench Studies with
Fenton’s reagent, additional studies of other Advanced Oxidation technologies were not recommended.
Biological Treatment Bench studies were recommended as an alternative to evaluate BOD reduction in
the Facility’s wastewater, potential toxicity issues, design parameters, operating issues, and costs for a
full-scale treatment system.
Toxicity Studies
As a precursor to Biological Bench Testing, the relative toxicity of the industrial process wastewater was
evaluated using Microtox testing in order to identify any potential issues that may have to be addressed
during the Biological Bench test. Microtox testing provides a quick snapshot of the relative toxicity of a
wastewater within a short period of time. The Microtox Analyzer measures the toxicity a substance by
monitoring the response of bioluminescent bacteria Vibrio fischeri. The bacteria are exposed to a range
of concentrations of the sample and the measured output of the decreased light emitted by the bacteria
in response to the toxic agent is quantified. The test shows a correlation with standard toxicity and wet
testing methods using larger organisms including fathead minnow, bluegill, zebrafish, and rainbow trout.
Additionally, Microtox testing can offer important information about the applicability of biological
treatment to a particular wastewater. If toxicity is present, the biological treatment processes may be
modified as necessary to include the addition of Powdered Activated Carbon (PAC), or operating
parameters may be modified (e.g. to allow for additional time to develop a properly acclimated sludge.)
In tandem with the Microtox testing, analytical data was collected on the wastewater, in order to
Page 15 of 28
evaluate whether there was a correlation between toxicity results and specific parameters including
COD, BOD, and/or TSS.
The Microtox tests monitor the response of bioluminescent bacteria to a wastewater sample; a decrease
in light emitted by the bacteria indicates the presence of toxicity in the sample. Results from the
Microtox tests are used to rank the relative toxicity of a wastewater in 5 categories ranging from Non-
Toxic (little or no reduction in luminescence of bacteria when exposed to the wastewater, or EC50
results near 100) to Very Toxic (>75% reduction in luminescence, or EC50 <25%), according to Table 6
below. Results over 100 percent are indicative of an increase in biological growth as they represent an
increase in bioluminescence.
51-75% Toxic
Daily composite samples of the industrial wastewater were collected from the clarifier effluent, prior to
discharge to the POTW, five days per week for a period of four weeks. In total, 20 samples were
analyzed. Toxicity results on the industrial wastewater ranged from Non Toxic to Very Toxic with 5
minute-EC50 results ranging from 18 percent to 93 percent and 15 minute-EC50 results ranging from 14
percent to 116 percent. The majority of the samples, 15 out of 20 collected, exhibited slightly toxic to
moderately toxic conditions.
Figure 6 below is a plot of the COD concentrations relative to the 5 min EC50 Microtox test results. This
figure shows an inverse relationship between COD and the 5 min EC50. In general, the data showed that
higher COD values resulted in lower EC50 values indicating higher toxicity. To test the relationship
between COD concentrations and the 5 min EC50 results, an R-squared (R2) value was calculated. This
value measures the amount of variation that can be explained by a particular variable. In general, a
higher calculated R2 value indicates a stronger correlation between the independent variables (e.g.,
COD concentration) and the response variable (e.g., 5 min EC50), and vice-versa. The calculated R
squared value was approximately 67%, which shows a meaningful correlation, especially for the small
data set analyzed. BOD and TSS concentrations did not appear to correlate directly with the toxicity
results, although high BOD values often accompany high COD and toxicity values.
Page 16 of 28
Results of the Toxicity Tests and review of historic COD, BOD, and TSS concentrations and Facility
operations indicated that the industrial wastewater is expected to be variable and will have to be taken
into consideration in the design and operation of a biological treatment system. In order to further
evaluate toxicity issues and the feasibility of biological treatment for BOD reduction in Avery’s
wastewater, a Biological Bench study was conducted as a next step.
Biological Treatment
Biological Treatability testing was completed on the Facility’s pretreated wastewater in two phases,
including Phase 1 Feasibility Testing to evaluate whether the wastewater is amenable to biological
treatment; and Phase 2 Bench Test completed to develop design criteria for a full-scale biological
wastewater treatment plant.
The purpose of the Phase 1 Biological Wastewater Feasibility Study (Feasibility Study) was to perform an
aerobic laboratory biological bench scale study to confirm the biodegradability of Avery’s wastewater
and evaluate the feasibility of using this technology to consistently meet local discharge limits. The
feasibility study was completed on a 40-gallon composited sample which was made up from five
individual daily samples (8 gallons each) of pretreated wastewater from the Facility. This compositing
step was done to allow for a consistent feed to the reactor every day and to neutralize any potential
toxicity. The composited wastewater was well-mixed, sampled, and placed into 5-gallon buckets and
refrigerated for the test.
Page 17 of 28
Analytical results for the composited pretreated wastewater sample are provided in Table 7 below.
Based on results from the Wastewater Characterization Study (Table 1), the wastewater fed to the
reactor during the Feasibility Study was similar to the “typical” pretreated wastewater effluent from the
Facility’s WWTP and therefore considered to be representative of conditions at the Facility.
To seed the reactor, a 150mL sample of mixed liquor from a local POTW was obtained and was added to
the reactor and aerated along with 7.35L distilled water. This allowed for an acclimated mixed liquor to
be grown in the reactor. The Feasibility Study took place over a 14 day period. Feeding of the reactor
was conducted once per day. During the study, the Hydraulic Retention Time (HRT) was set at 2 days,
and Solids Retention Time (SRT) was set at 10 days. In order to meet a (HRT) of 2 days, the daily
operating sequence consisted of settling the reactor contents, withdrawing a supernatant volume of
7.5L, and feeding the reactor 7.5L of composited wastewater based on the schedule in Table 8:
Daily mixed liquor samples were collected from the reactor for TSS and volatile suspended solids (VSS)
analysis. Additionally, raw wastewater and reactor effluent samples were collected periodically over the
duration of the Feasibility Study for BOD, COD, phosphorus, orthophosphate, Total Kjeldahl nitrogen
(TKN), ammonia, oil and grease, TSS, and VSS analysis.
Page 18 of 28
In addition to analytical data, samples were collected of the mixed liquor for viewing under a
microscope and subsequent photomicrographs. Evidence of significant biological activity was observed
with the microscope for each of the samples evaluated.
For the duration of the study, mixed liquor suspended solids (MLSS) and mixed liquor volatile suspended
solids (MLVSS) were monitored on a daily basis. The beginning concentration of MLSS and MLVSS on
“Day 0” after the reactor was seeded with mixed liquor from the POTW, was 40 mg/L and 27.6 mg/L,
respectively. The daily concentrations of MLSS and MLVSS are presented in Figure 7 below:
800
700
Suspended Solids (mg/L)
600
500
400
MLSS
300
MLVSS
200
100
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Time (Days)
The graph above shows a steady increase in both MLSS and MLVSS over the 14-day study period. This
steady increase represents biological growth in the reactor.
Throughout the duration of the Feasibility Study, samples were collected of the raw wastewater that
was fed to the reactor and the effluent after settling the contents of the reactor. These analytical
samples were collected to evaluate the success of biological treatment on the Facility’s wastewater. The
results for COD (total and soluble), BOD (total and soluble), and oil & grease are presented in Table 9
below:
Page 19 of 28
Additionally, the mixed liquor in the reactor was observed and daily pH, dissolved oxygen (D.O.), and
temperature readings were taken and documented. Data on pH (7.69 to 9.35 standard units), DO (9.11
to 10.09 mg/l), and Temperature (53 to 67.3 °F) of the reactor were within typical operating ranges and
indicate that the reactor remained relatively stable over the course of the Treatability Study.
During the Phase 1 Study, a rudimentary air stripping test was completed to evaluate BOD and COD
reductions from aeration. The raw Facility wastewater sample was aerated in a 1 liter cylinder for
approximately two hours, and BOD and COD concentrations were evaluated before and after aeration,
respectively. The results of the air stripping test were inconclusive for BOD reduction (BOD
concentrations were not reduced after air stripping), and indicated that a marginal COD reduction of
approximately 10% was achieved. Based on the results of the study, aeration does not appear to be a
viable means for BOD reduction for the Facility’s wastewater.
The results of the Phase 1 Feasibility Study indicated that the mixed liquor biomass continued to grow
throughout the study and that inhibitory and/or toxic compounds in the Facility’s wastewater did not
inhibit biological growth in the reactor. BOD removal efficiency (using Day 0 and Day 14 data) was
78.8%; similarly COD and O & G removal efficiencies were 67.9% and 97.2%, respectively. Based on
these results, the Facility’s wastewater appeared to be amenable to activated sludge biological
wastewater treatment, and additional biological studies were recommended to further evaluate
fluctuations of toxicity in the Facility’s wastewater and collect key design parameters for a full-scale
system.
Prior to the start of Phase 2 and immediately following the Phase 1 Treatability Study, a “Maintain
Phase” was conducted to continue to increase the concentration of MLSS within the test reactor. During
this Maintain Phase, the reactor was fed with 7.5L – 10.0L of the composited raw Facility wastewater
from Phase 1 every other day to achieve a hydraulic retention time (HRT) of 3-4 days. Additionally,
samples of the mixed liquor were collected for total and volatile suspended solids analysis. The reactor
was monitored and a steady growth in biomass (evident from the increased MLSS concentrations) was
Page 20 of 28
observed. No adverse toxicity effects from the Facility’s wastewater were observed during the
maintenance phase. The Phase 2 Study began after a period of approximately 36 days of the
Maintenance Phase.
The purpose of the Phase 2 Biological Bench Test was to perform an aerobic laboratory biological bench
scale evaluation to obtain data for use in developing design criteria for a full-scale biological
pretreatment plant. The Phase 2 Treatability Study was divided into four (4) sub-phases, as follows:
The Phase 2 Bench Study was completed with samples of the Facility’s wastewater collected daily and
shipped to the test laboratory over a period of approximately one month. The daily volume received
varied as the Phase 2 study progressed from approximately 4 gallons per day during the Acclimation
sub-phase to approximately 26 gallons per day during the HRT 3 Testing sub-phase. The wastewater
was stored in a refrigerated cooler upon receipt in the laboratory.
Samples of the raw wastewater were collected on each day that samples were received from the Facility
(Monday – Friday). The results from the raw wastewater characterization sampling has been
summarized and presented in Table 10 below.
The average analytical results in the ‘raw’ Facility wastewater provided for the Phase 2 Bench Study
were fairly consistent with the Wastewater Characterization study results (e.g., BOD concentrations
were similar), although some constituents (average Total COD, TSS, ammonia and Orthophosphate)
were slightly lower and other constituents (Oil and Grease) were slightly higher. Since the Facility’s
wastewater concentrations are known to fluctuate depending on industrial operations at the facility, the
‘raw’ wastewater samples appeared to be representative of typical clarifier effluent concentrations.
On each day, during the duration of the Bench Study, tests were completed to meet the required HRT
and SRT for the specific sub-phase of the study. The fill-and-draw method was used for feeding and
Page 21 of 28
wasting from the reactor. In order to meet the required HRT and SRT for each sub-phase, the daily
operating sequence consisted of settling the reactor contents, withdrawing a known supernatant
volume, and feeding the reactor with a volume of raw Facility wastewater equal to that removed after
settling. Table 11 provides the following operating conditions that were used during the study:
HRT SRT
Sub-Phase
(hrs) (days)
Acclimation 48
HRT 1 24
10
HRT 2 12
HRT 3 6
Mixed liquor samples were collected on an every other day basis from the reactor for TSS and VSS
analysis. Additionally, untreated wastewater and reactor effluent samples were collected periodically
over the duration of the Bench Study for BOD, COD, orthophosphate, ammonia, oil and grease, TSS, and
VSS analysis.
Over the duration of the Bench Study, the reactor was monitored for pH, dissolved oxygen (D.O.), and
temperature on a daily basis. Additionally, MLSS and MLVSS were sampled for and monitored on an
every other day basis. Settling tests were conducted throughout the duration of the study and used to
determine the Sludge Volume Index (SVI) of the mixed liquor, which is a measurement of the biological
solids ability to settle effectively. Additionally, the Specific Oxygen Uptake Range was evaluated for
each study phase.
Based on these results, the pH within the test reactor remained within a reasonable range between 6.0
and 8.0, since bacteria grow best in a pH range of 6 to 9. It was noted that the pH within the reactor
began decreasing prior to and during the acclimation sub-phase and approached 6.0. In order to
improve the buffering capacity of the reactor and keep the operating pH within a range of 6.0 to 8.0,
sodium hydroxide (NaOH) was added from Day 0 to Day 8. Following Day 8, the buffering capacity
within the reactor was sufficient to keep the pH within the ideal range. Activated sludge systems tend
to have a decreasing pH due to the production of CO2 from the oxidation of organic matter. The test
reactor was proof of this decreasing pH, as the mixed liquor solids concentration increased, the pH
within the reactor decreased.
The D.O. within the test reactor remained within a range of 7.0 to 9.0 over the entire study period,
which was more than adequate to ensure sufficient oxygen for microbial biosynthesis. Similar to pH, it is
important to monitor and control D.O. within an activated sludge system. In general, a D.O.
concentration of 1.0 mg/L to 3.0 mg/L is best for maintaining efficient and healthy microorganisms. If
the D.O. concentration drops below 1.0 mg/L within the system, the treatment efficiency of the
activated sludge system will suffer.
Page 22 of 28
It was noted that the reactor was operated with a D.O. concentration much greater than what is
sufficiently needed due to mixing requirements within the reactor. It is important that the aerated
reactor be well-mixed to keep the microorganisms in suspension and increase the contact of the
microbes with the food source. The system mixing was achieved purely by the air source and no
mechanical mixers were used. If the air was turned down any lower, solids would begin to settle in the
reactor.
The temperature within the test reactor remained close to 70-75°F (21-24°C) during the study, with a
minimum reactor temperature of 68°F (20°C) and a maximum reactor temperature of 78°F (25.5°C)
during the duration of the Bench Study. Mesophilic bacteria, the most prominent for these test
conditions, have an optimum temperature range of 10°C to 45°C, and, therefore, the test reactor was
operated within a temperature range that provided for optimal biological activity. Historic data on
temperature in the Facility’s clarifier indicate that the temperature ranges from 15°C to 25°C, which is
similar to the temperature maintained within the test reactor.
Figure 8 below provides an indication of the MLSS and MLVSS over the duration of the Phase 2
treatability study.
Table 12 below provides a summary of the Specific Oxygen Uptake Range in the reactor over the
duration of the Phase 2 evaluation. These results indicate that the specific oxygen uptake rate (SOUR) of
the mixed liquor ranged from 2.95 to 11.55 mgO2/gVSS/hr. These values are indicative of biological
activity, endogenous respiration, and fast settling sludge.
Page 23 of 28
Table 12: Phase 2 Treatability Study System Specific Oxygen Uptake Rate
Results of the biological treatability study indicated that the Facility’s pretreated wastewater can be
successfully treated using biological wastewater treatment technology to reduce BOD and TSS
concentrations to the meet discharge criteria to the POTW. The Phase 2 evaluation further supported
the Phase 1 conclusions that the Facility’s wastewater is biodegradable.
Adequate total BOD removals were achieved during all sub-phases of the Phase 2 study. BOD and COD
removal efficiencies for each sub-phase are shown in Table 13 below. All removal efficiency data is
based on average sample concentrations.
Table 13: Biological Treatability Study Removal Efficiencies and Mixed Liquor Concentrations
As indicated in Table 14, MLSS and MLVSS concentrations increased overall throughout the duration of
the Phase 2 study and ranged from 1,330 mg/L to 4,750 mg/L for MLSS and from 826 mg/L to 3,150
mg/L for MLVSS. (Optimal MLSS concentrations are 2,500 to 3,500 mg/L and optimal MLVSS is 2,000 to
3,000 mg/L, respectively.)
Page 24 of 28
Successful BOD removal efficiencies were achieved at an SRT of 10 days and at HRTs of 48 hours during
the Acclimation Sub-Phase, 24 hours during the HRT 1 Sub-Phase, 12 hours during the HRT 2 Sub-Phase,
and 6 hours during the HRT 3 Sub-Phase. A SRT of 10 days was the only SRT used during the Phase 2
study.
Based on the results of the Study, the optimal HRT appears to be approximately 6 to 12 hours. This HRT
provided adequate BOD removal to meet the future permit requirements without ‘starving’ the
microorganisms.
Biological treatment appears to be viable alternative for treating the Facility’s wastewater to meet the
required POTW discharge criteria.
Water Treatment with Solids Dewatering
As an innovative approach for significant BOD and TSS removal, a centrifuge pilot test was completed for
the Facility to evaluate the feasibility of replacing the existing primary treatment system (clarifier) and
solids dewatering equipment (filter press) with a single unit process. This pilot study was completed
using a solid bowl scroll discharge decanter centrifuge that provided continuous separation of
particulate solids from the influent wastewater by decanting. Separation of solids from the influent
wastewater is accomplished with the application of centrifugal force.
As with the other bench scale and pilot studies, the centrifuge pilot test was conducted on the Facility’s
wastewater, but untreated wastewater from the equalization was used rather than pretreated effluent.
During the pilot study, a total of 33 Chemical Pretreatment Scenarios were evaluated using chemicals
currently used at the Facility in the existing chemical-precipitation process, including Alum and an
organic coagulant, lime solution for pH control, and C-339 Polymer. The primary objectives of the
centrifuge pilot study were to 1) evaluate optimal chemical pre-treatment and polymer dosage; 2)
evaluate the overall performance of the centrifuge under different throughput conditions; and 3)
evaluate the relationship of torque on the centrifuge to the cake dryness. Although these primary
objectives were oriented at identifying the best operating conditions and centrifuge selection for the
Facility’s wastewater to achieve maximum cake dryness and highest capture efficiency, overall results
would be used to confirm whether the centrifuge is a viable option to replace the primary treatment
system to produce an effluent acceptable for discharge to the POTW, while simultaneously dewatering
the solids to produce a dry cake for separate disposal (e.g., landfill).
Page 25 of 28
There are several advantages and disadvantages associated with using the centrifuge to simultaneously
treat the Facility’s wastewater and dewater the resulting sludge.
Centrifuge Advantages include:
• Reduced Footprint - by combining two treatment steps into one unit
• Continuous Operation – influent wastewater is fed continuously, and solids and centrate
discharged continuously. The current filter press operates in batch mode.
• Automatic Dewatered Solids Removal and Conveyance – solids are moved along a screw
conveyor installed along the inside of the bowl of the centrifuge, are transported up an incline
and out of the machine. Existing filter press must be opened and (54) filter plates scraped
individually for solids removal.
• Less Likely to Clog Due to Sticky Nature of Wastewater – versus the filter press plates (although
the long-term exposure to the sticky solids could potentially lead to increased maintenance if
the solids build up over time)
• Higher Throughput – sedimentation in existing clarifier relies on gravity and is therefore slower
than the forced liquid/solids separation achieved by the centrifuge
• Containment of Odor – versus open clarifier
• Lower Operator Attention (when operation is stable)
Centrifuge Disadvantages include:
• Reduced capacity for capturing smaller particulates. Drier solids are typically produced because
finer particulates (typically less than 45 microns) are not screened.
• Reduced centrate quality - centrifuges are primarily used to dewater solids, and may not offer
the same centrate water quality as a chemical precipitation system.
• Centrifuge may not adjust to fluctuations in the influent wastewater quality as well as the
clarifier, since the clarifier has more equalization; the centrifuge will be impacted by large
wastewater fluctuations quicker than the clarifier
• Higher maintenance requirements and operating costs due to mechanical requirements
o Plant personnel can typically provide maintenance on filter press but may not be able to
provide maintenance on centrifuge
o Sticky solids in the Facility’s wastewater could build up in the centrifuge over time and
create long-term maintenance issue)
Overall test results appeared to be very promising in terms of cake dryness and solids capture rate.
Under optimal pretreatment and polymer dosage (0.7% v/v Alum/Coagulant and 5 – 10% w/w lime to
maintain the pH at 6.5 minimum), 65 to 75% total solids cake dryness was achieved with a throughput
range of 40 to 97 gallons per minute, and the capture rate of suspended solids was estimated to be
>96%. Chemical pretreatment during the pilot was monitored with Oxygen Reduction Potential (ORP)
to prevent over or under-dosage of coagulant. Centrate quality was monitored for solids content (mg/l)
and compared to the measured ORP values during the study. Figure 9 below provides a summary of the
variation of centrate quality (mg/l TSS) with respect to coagulant dosage (ORP values).
Page 26 of 28
The results of the pilot indicated that TSS in the centrate discharge ranged from 35 mg/l to 5,144 mg/l.
Since the POTW requirements include TSS concentrations < 250 mg/l, not all of these tests provided a
centrate that could be discharged directly to the POTW. Even for several of the tests that were
identified as achieving optimal pretreatment and polymer dosage (as evidenced with ORP values > 185
MV), many of the tests showed elevated Centrate TSS concentrations of >1,000 mg/l. These results
indicate that further evaluation of ORP and capture efficiency are required to confirm that the
centrifuge is capable of continuously treating the centrate to meet the discharge criteria for the POTW.
Additionally, fluctuations in the untreated influent wastewater quality were noted to impact the
performance of the centrifuge with respect to solids capture and cake dryness. Information from the
Facility’s wastewater characterization study indicates that influent wastewater concentrations vary
substantially based on industrial activities at the Facility. This would have to be taken into consideration
for full-scale operation of a centrifuge at the Facility. Data collected during the centrifuge pilot study
was not comprehensive enough to make a direct comparison with the Facility’s existing treatment
system with respect to POTW discharge criteria (including TSS and BOD discharge concentrations) and
chemical usage.
Overall, the test results for the centrifuge were promising and the use of the centrifuge to replace the
Facility’s existing primary treatment clarifier and filter press would offer several advantages for primary
treatment and solids dewatering. However, additional information and/or calculations are required to
fully evaluate the performance of the centrifuge based on water quality requirements for discharge to
the POTW, address the fluctuations in the Facility’s untreated industrial wastewater, and evaluate
capital and operation and maintenance costs (including chemical addition requirements) against the
existing system. Additional treatment steps may also be required prior to discharge of the centrate to
the POTW.
Page 27 of 28
CONCLUSIONS
In order to identify the “most appropriate” wastewater treatment process to meet more stringent BOD
and TSS pretreatment limits, address an increase in production, and make improvements to the original
industrial wastewater treatment system, a series of innovative and unconventional bench and pilot
studies were completed on the industrial wastewater from a major chemical adhesive products
business. Potential treatment alternatives were initially screened at a ‘desk-top’ level and the more
promising alternatives for the Facility’s wastewater were then tested at a bench scale level. Results of
these tests indicate that the most viable alternative in terms of overall technical effectiveness and
compliance appears to be biological aerobic treatment of the pretreated wastewater.
REFERENCES
1. Activated Carbon and Advanced Bench Test Summary, prepared for URS by Advanced
GeoServices Corporation (AGC), December 11, 2012.
2. Advanced Oxidation Bench Testing Results and Conclusions, URS Corporation, May 28, 2013
3. Andritz Separation Demo Report No. D-15-13, September 5, 2013
4. Ferrate Treatment Technologies Treatability Report, May 10, 2013
5. Lab Notes for Advanced Oxidation Bench Study Prepared by Water Remediation Technologies,
Contracted by AGC.
6. Phase I Biological Wastewater Treatability Study, URS Corporation, August 2013.
7. Phase 2 Biological Wastewater Treatability Study, URS Corporation, November 2013.
8. Summary of Bench Testing Results, Recommendations, and Conclusions, URS Corporation,
December 14, 2012.
9. Summary of Ferrate Treatability Testing Results, URS Corporation, July 19, 2013.
10. Wastewater Characterization and Source Identification Evaluation, URS Corporation, October 8,
2012.
Page 28 of 28
Lucy Pugh1*, Joan Gauthier2, Eric Van Orman1, Cullum Pakosh3, Duane Dehner4
1
AECOM, Grand Rapids, Michigan
2
Chrysler, Auburn Hills, Michigan
3
Chrysler, Toronto, Ontario
4
Chrysler, Trenton, Michigan
*Email: lucy.pugh@aecom.com
ABSTRACT
Automotive manufacturing operations at two Chrysler facilities, the Etobicoke Casting Plant and
Trenton Engine Complex, have been recently expanded to increase production and modernize
manufacturing processes. The modified manufacturing operations and improved metal working
chemical solutions resulted in significant changes to process wastewater quantities and
characteristics, and necessitated changes to each facility’s wastewater pretreatment system to
provide adequate capacity, optimize pretreatment effectiveness, and minimize life-cycle costs.
This paper presents details of technology evaluation, bench and pilot testing, and application of
ultrafiltration technologies for pretreatment and potential reuse of three wastewater sources
generated from die casting, fluorescent penetrant inspection, and engine manufacturing processes
at the Chrysler facilities. The experimental design and data for bench and pilot testing of
ultrafiltration technology are documented. Details regarding negotiation of alternate wastewater
discharge requirements, full-scale design criteria and approach, and full-scale operating
experience are also provided.
INTRODUCTION
Chrysler has embarked on an initiative to reduce water consumption, increase water recycling,
and improve wastewater pretreatment system performance and compliance as part of its World
Class Manufacturing program, a continuous-improvement system designed to benefit all aspects
of Chrysler’s manufacturing operations. Improvements to Chrysler’s wastewater pretreatment
infrastructure are also being driven by the need for increased capacity as a result of expansions to
manufacturing activities, and modifications to manufacturing practices that result in changes to
wastewater characteristics.
As Chrysler upgrades and expands its automotive manufacturing facilities, increased wastewater
pretreatment capacity and/or modified wastewater pretreatment technology is often required to
address wastewater volumes and characteristics that differ from historical operations. Chemicals
used in metalworking processes are more robust compared to previous formulations, resulting in
longer useful life and reduced wastewater volumes. However, these metalworking formulations
present unique wastewater challenges when the chemicals are spent and discharged to the
wastewater pretreatment system. Chrysler needs to implement best-value, reliable solutions to
these wastewater pretreatment challenges, while considering the company’s objectives to reduce
water consumption and recycle water and wastes to the extent possible.
This paper describes three Chrysler wastewater sources that required new or upgraded
wastewater pretreatment systems. Increases and changes to die casting production capacity at
the Etobicoke Casting Plant resulted in the need for additional wastewater pretreatment capacity
for die casting wastewater. In addition, a new fluorescent penetrant inspection process was
installed at the plant, which required a separate pretreatment system to address the fluorescent
color contributed by the new wastewater source.
The Trenton Engine Complex experienced significant production changes since its wastewater
pretreatment system was last upgraded, including construction of a new engine manufacturing
plant, and demolition and replacement of engine manufacturing processes in the existing engine
plant. Although engine production was significantly increased, due to improved manufacturing
processes and machining coolants, these changes resulted in substantially reduced wastewater
flows with increased concentrations of oils and other constituents. The existing wastewater
pretreatment facility was oversized and inefficient for the projected wastewater flows and
characteristics.
Application of membrane technologies for pretreatment of wastewater generated from
automotive manufacturing activities continues to gain popularity. Membrane separation
technologies can be effective for concentrated wastewater streams that result from improved
manufacturing techniques, and offer potential opportunities for wastewater reuse as well as
reduced operating costs. Various forms of ultrafiltration, utilizing spiral-wound, hollow fiber,
and tubular membranes, were evaluated and applied for the Etobicoke Casting and Trenton
Engine manufacturing facilities.
The objectives of the work described in this paper were to conduct water balances, identify
appropriate ultrafiltration and ancillary pretreatment technologies for each wastewater source,
verify the effectiveness and full-scale design criteria for each application through bench and pilot
testing, and implement the new technologies at the Etobicoke Casting and Trenton Engine
facilities.
BACKGROUND
Die lubricant: a diluted, water soluble die lubricant is sprayed on the die before aluminum is
injected. Some of the die lubricant stays on the die and prevents the next shot of aluminum from
sticking to the tool steel. A small amount of die lubricant may fall to the floor where it is
collected by perimeter trenches and conveyed to the wastewater pretreatment system.
Hydraulic oil: fire-resistant hydraulic fluid, consisting of a mixture of water and glycol, is used
in machines that are subjected to elevated temperatures. The fluid may enter the wastewater
stream if a leak occurs within the hydraulic system.
Die slick and plunger slick: these lubricants may enter the wastewater stream through system
leaks or inadvertent spills.
Soaps and detergents: these materials are used for various cleaning purposes around the plant.
Prior to the plant production upgrades, die casting wastewater was collected and pretreated via
oil/water separation for removal of free oils, and low pressure membrane filtration for removal of
emulsified oils.
A new fluorescent penetrant inspection (FPI) process was installed at the Etobicoke Casting
Plant as part of the overall production modifications. The FPI process involves spraying cast
parts with fluorescent dye, rinsing the parts through a two-stage process, and then inspecting the
parts in an ultraviolet booth to detect the presence of imperfections. Rinsewater is discharged
from the process as a waste stream. A temporary nanofiltration system was installed for
pretreatment of the wastewater generated by the fluorescent penetrant inspection process, and
was prone to frequent membrane fouling. Pretreatment utilizing ultrafiltration was proposed to
minimize fouling of the nanofiltration membrane, but required further evaluation.
Pretreated wastewater is discharged to the sanitary sewer and must comply with the requirements
of the City of Toronto Sewer Use Bylaw (Chapter 681).
Trenton Engine Complex
Chrysler’s Trenton Engine Complex, located in Trenton, Michigan, United States, includes the
north and south engine plants. The original north plant contains 195,000 square-meters of
manufacturing space, and includes machining and assembly operations for production of 3.2 liter
V-6 Pentastar™ engines and Tigershark™ engines. The 76,000 square-meter south plant was
constructed in 2007 and produces 3.6 liter V-6 Pentastar™ engines that are installed in a wide
variety of Chrysler vehicles.
Process wastewater is produced at the engine production complex from the following sources:
Machining: spent coolants and rinsewaters containing coolant dragout are discharged from parts
machining operations.
Washers: spent washing solutions and rinsewaters containing cleaning chemical dragout are
discharged from parts washers used to clean parts prior to assembly.
Housekeeping: spent floor scrubber solutions are discharged at a central collection facility and
transferred to the wastewater pretreatment system.
Utility water blowdown: blowdown wastewater is discharged from cooling towers and boilers to
maintain desired water quality within each process.
Stormwater: site runoff from precipitation and snowmelt is collected in some areas of the
complex and discharged to the wastewater pretreatment system.
The existing wastewater pretreatment system at the Trenton Engine Complex was designed to
process approximately 1,000 cubic meters of wastewater per day, and includes batch chemical
treatment reactors for separation of emulsified oils, followed by an aerobic biological process
(sequencing batch reactors) for removal of 5-day biochemical oxygen demand (BOD5) and
ammonia-nitrogen. The pretreatment system was previously described by Pugh et al. (2010).
Pretreated wastewater is discharged to the sanitary sewer and must comply with the City of
Trenton sewer use ordinance and industrial user permit requirements. Due to reduced
wastewater flows and loads, it was proposed that significant operating costs could be avoided if
the biological process could be shut down, and low pressure membrane filtration substituted for
the chemical treatment process. However, the local municipal sewer authority would require
demonstration that the mass discharge loadings would not exceed those allowed by the facility’s
industrial discharge permit.
METHODOLOGY
The approach for evaluation and application of new and improved wastewater pretreatment
utilizing ultrafiltration technologies was similar for both the Etobicoke Casting Plant and the
Trenton Engine Complex.
Water Balances
As a first step, a baseline water balance was constructed for each facility to identify and quantify
all major processes that use water and generate wastewater. Each baseline water balance was
then adjusted to reflect the expected increase in water use and wastewater generation based on
installation of new manufacturing equipment and machines, as well as ancillary facilities.
Current wastewater quantities and quality, and expected wastewater quantities and quality at
future build-out conditions were reviewed with plant engineering, operations and environmental
staff to establish the wastewater pretreatment basis of design for each facility. Decisions were
also made regarding pretreatment equipment redundancy, and consideration was given to
equipment and design capacity to be installed immediately, as well as equipment to be installed
in the future for the ultimate production build-out condition.
At the Trenton Engine Complex, there was an opportunity to re-route relatively clean wastewater
streams around the wastewater pretreatment process and reduce the hydraulic design
requirements of the pretreatment system. The relatively clean streams included some site
stormwater, as well as cooling tower and boiler blowdown wastewaters. However, the
elimination of relatively clean wastewater streams from the pretreatment system, as well as the
reduced wastewater volumes and associated higher wastewater concentrations that result from
operation of new machining equipment, lead to significantly increased wastewater
concentrations in the process wastewater. Close coordination and negotiation of discharge
requirements with the local municipal sewer authority were required to ensure that discharge
standards could be reliably achieved with reduced process wastewater volumes and increased
concentrations.
The ultrafiltration pilot system was installed in the lower level of the existing wastewater
pretreatment system process building, and included the following equipment:
Transfer pump
Bag filter
Oil/water separator
Process tank
Ultrafiltration skid, including eight tubular membranes, each 3 meters in length with 2.54
centimeter diameter, manufactured by Koch Membrane Systems
Permeate storage tank
Clean-in-place/batchdown tank
A photograph of the pilot equipment installed at the Trenton Engine Complex is provided in
Figure 1. The process tank was equipped with a float level switch to automatically control the
transfer pump and maintain a setpoint wastewater level in the tank. During operation of the pilot
system, the pump transferred wastewater from the facility’s influent wetwell through a bag filter
for removal of large solids that could damage the membrane system, and through an oil/water
separator for removal of free oil. Effluent from the oil/water separator was discharged to the
process tank, which provided equalization and served as the source of influent wastewater to the
ultrafiltration membrane skid.
The pilot ultrafiltration system was operated in a modified batch mode, with process wastewater
continuously recirculated from the process tank at an average rate equal to the permeate removal
rate. Permeate that passed through the membranes was discharged to the permeate storage tank.
The remaining wastewater, containing solids and other materials rejected by the membranes, was
returned to the process tank. Thus, the concentration of wastewater constituents increased in the
process tank throughout each batch cycle. At the completion of each cycle, the transfer of fresh
influent wastewater to the process tank was discontinued, and the wastewater in the process tank
was recycled through the membranes until the permeate flow decreased and the transmembrane
pressure increased to unacceptable levels. The membranes were then cleaned by recirculating a
mixture of water and cleaning chemicals from the clean-in-place tank, and the system was
returned to service for the next cycle.
Ultrafiltration Skid
Oil/Water Separator
Bag Filter
Permeate Tank
CIP/Batchdown Tank
Process Tank
Transfer Pump
Figure 1. Pilot equipment was installed in the lower level of the process wastewater
pretreatment building at the Trenton Engine Complex.
A photograph of the ultrafiltration skid is provided in Figure 2. The pilot system was operated to
evaluate a number of variables, including membrane type, wastewater characteristics,
temperature, and membrane cleaning effectiveness.
Figure 2. The lower portion of the tubular ultrafiltration skid with permeate outlets is
shown.
During each trial, samples of influent wastewater and permeate were collected and transported to
TestAmerica Laboratories in Brighton, Michigan, USA for analysis. Sample preservation and
analytical methods conformed to U.S. EPA methods.
Design of the Trenton Engine Complex wastewater pretreatment facilities is currently underway.
RESULTS
The existing spiral-wound ultrafiltration system used for die casting wastewater provided
acceptable service, and it was therefore decided to install a similar system in parallel with the
existing system to avoid multiple membrane filtration styles and simplify operations. The new
ultrafiltration system was provided with an increased level of instrumentation and automation,
and control signals for both the existing and new system were transmitted to the factory
information system to provide remote monitoring of the pretreatment equipment at various
locations within the manufacturing facility.
Figure 3. Untreated fluorescent penetrant inspection wastewater is shown on the left, and
permeate produced from hollow fiber ultrafiltration bench-scale treatment is shown on the
right.
Based on the observed effectiveness of hollow fiber ultrafiltration for the FPI wastewater,
Chrysler elected to proceed with full-scale design of a pretreatment system. The effectiveness of
ultrafiltration in eliminating the fouling issues experienced with the existing spiral-wound
nanofiltration membrane was not confirmed, however. Alternate technologies were identified to
replace the nanofiltration membrane system for removal of residual color, if necessary, including
chemical oxidation and granular activated carbon adsorption. The basis of design for the FPI
wastewater pretreatment system, including pertinent wastewater characteristics and pretreatment
requirements for discharge, are summarized in Table 2.
Table 2. Fluorescent Penetrant Inspection Process Wastewater Characteristics and
Pretreatment Requirements
Sanitary Sewer Discharge
Parameter Units Minimum Maximum Requirements*
Daily volume m3/day 8 27 -
pH std. units 6 8 6.0 – 11.5
Wastewater temperature °C 5 35 <60
Animal and vegetable oil & grease mg/L 75 500 150
Mineral and synthetic oil & grease mg/L 0 20 15
Chemical oxygen demand (COD) mg/L 5,000 15,000 -
Total suspended solids (TSS) mg/L 300 650 350 (surcharge)
*where surcharge is noted, discharge requirements may be exceeded but a surcharge will be incurred
In addition to the sanitary sewer discharge requirements summarized in Table 2, the Sewer
Bylaw also prohibits the discharge of dyes or coloring materials which may or could pass
through the treatment system and discolor the effluent.
Wastewater Collection and Transfer – Die casting wastewater streams are collected in trenches
and sumps throughout the manufacturing facility and flow by gravity to a sump in the lower level
of the wastewater pretreatment area. Pumps are used to transfer the wastewater to a waste
collection tank equipped with a belt skimmer for removal of free oil. Pumps transfer wastewater
from the collection tank to the wastewater equalization system.
Wastewater Equalization – Two existing equalization tanks, each with 125 m3 capacity, were
supplemented with a third, 90 m3 equalization tank to provide additional capacity for projected
die casting wastewater flows. One of the existing equalization tanks needed to be relocated to
provide space for expansion of the wastewater pretreatment process building, and installation of
new ultrafiltration equipment.
Oil/water Separation – The existing coalescing oil/water separator, designed for a treatment flow
rate of 285 liters/minute, was maintained in service for the upgraded die casting pretreatment
system. Wastewater is pumped from the equalization system to the oil/water separator for
removal of free oil.
Pre-filtration – The existing bag filter, equipped with 75 micron bags, was maintained in service
for the upgraded system. Wastewater is pumped from the discharge tank of the oil/water
separator through the bag filter for removal of larger solids that may damage the ultrafiltration
membrane system.
Ultrafiltration – The existing ultrafiltration system was supplemented with a new process tank,
ultrafiltration skid, and clean-in-place system. The existing 19 m3 working tank was replaced
with a larger, 70 m3 working tank to provide additional capacity for two ultrafiltration systems
and extend ultrafiltration batch cycles. The new ultrafiltration skid was equipped with a 100 HP
feed pump and eight horizontal spiral-wound ultrafiltration cartridges. The feed pump transfers
wastewater through a new bag filter system for pre-filtration, and then to the ultrafiltration
membranes. Permeate is produced at an average rate of 110 liters/minute, and is discharged to
the municipal sewer system. The concentrate is returned to the process tank until it reaches a
concentration where it can no longer be effectively treated, which typically occurs after
processing approximately 110 m3 of permeate. The remaining volume in the process tank
continues to be processed to concentrate it further and reduce the volume which requires off-site
disposal, and typically represents approximately 5 percent of the original wastewater volume.
Figure 4. Process flow diagram for Etobicoke Casting Plant die casting wastewater
pretreatment system; new piping and equipment are shown in bold
Figure 6. Die casting spiral-wound ultrafiltration skid and piping between clean-in-place
tank
Figure 7. Controls and instruments for the die casting ultrafiltration system
The new FPI wastewater pretreatment system was located within the manufacturing facility in
close proximity to the FPI process. The new system includes an 11 m3 process tank, feed pump,
bag filter, hollow fiber membrane skid, and clean-in-place system. A process flow diagram for
the new FPI wastewater pretreatment system is provided in Figure 8. Photographs of the
installed equipment are provided in Figures 9 and 10.
Figure 8. Process flow diagram for Etobicoke Casting Plant fluorescent penetrant
inspection wastewater pretreatment system
Figure 9. FPI hollow fiber ultrafiltration membrane skid and control panel
Figure 11. Simplified water balance for the Trenton Engine Complex
Wastewater design flows, characteristics, and requirements for discharge to the local sanitary
sewer, are shown in Table 3.
Table 3. Engine Manufacturing Process Wastewater Characteristics and Pretreatment
Requirements
Sanitary Sewer Discharge
Parameter Units Minimum Maximum Requirements
Daily volume m3/day 40 60 -
pH std. units 5.5 9.5 5.5 – 9.5
Wastewater temperature °C 50 100
Oil & grease, HEM mg/L 50 25,000 70
Chemical oxygen demand (COD) mg/L 700 500,000 -
5-day Biochemical oxygen demand (BOD5) mg/L 500 50,000 682
Total solids mg/L 4,000 50,000 -
Total phosphorus mg/L 1.0 50 4.64
Ammonia-nitrogen mg/L 25 200 14.56*
Total suspended solids (TSS) mg/L 5 10,000 -
Boron mg/L 10 50 40.84
Chromium mg/L <0.05 0.2 3.45
Copper mg/L 0.5 4.0 0.703
Lead mg/L 0.1 3.0 0.802
Zinc mg/L 0.5 50 2.82
*negotiated alternate mass limit of 18 kg/day
measurement of performance for comparison after each batch trial, to determine whether the
membranes were impacted over time and repeated operation.
Each modified batch trial was conducted over a period of several days. The pilot system
operated automatically and was typically unattended with the exception of the concentration and
cleaning cycle operated at the completion of each batch. A comparison of performance for the
two membranes during one six-day trial of unattended operation is shown in Figure 12. This
graph shows the changes in flux rate for each membrane as a function of temperature and
concentration factor. The concentration factor is calculated from the total volume of wastewater
processed through the membranes at a given point in time, divided by the original wastewater
volume.
30.0
FLux (liters/cm2‐day*100); Temperature (C)
25.0
20.0
15.0
10.0
5.0
0.0
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00
Concentration Factor (X)
Figure 12. Graph illustrates flux rate for FEG 276 and FEG 251 membranes as a function
of temperature and concentration factor during a continuous unattended run over six days;
the flux rate for both membranes decreased as the wastewater concentration increased, but
the FEG 276 membrane exhibited better overall performance.
During the concentration, or batchdown process, the introduction of fresh wastewater into the
process tank was discontinued, and the wastewater was allowed to continue to concentrate
through the membranes to minimize the overall volume of concentrated wastewater for disposal.
A comparison of flux rate for the FEG 276 and FEG 251 membranes during a batchdown cycle
is shown in Figure 13, along with temperature, as a function of concentration factor. This
batchdown cycle was conducted over a period of two days, to provide better information
regarding flux rate as a function of wastewater concentration.
0.60 25
0.50
20
Flux (liters/cm2‐day)
Temperature (C)
0.40
15
0.30
10
0.20
5
0.10
0.00 0
0.00 5.00 10.00 15.00 20.00 25.00 30.00
Concentration Factor (X)
Figure 13. Graph illustrates flux rate for FEG 276 and FEG 251 membranes as a function
of temperature and concentration factor during a batchdown cycle; the flux rate for both
membranes increased as temperature increased. The flux rate for the FEG 276 membrane
exhibited better overall performance.
Based on the higher flux rates observed for the FEG 276 membranes over several modified batch
trials, it was concluded that FEG 276 membranes provided superior performance, and the four
FEG 251 membranes on the pilot system were replaced, providing a total of eight FEG 276
membranes for future evaluations. The clean water flux rate for the FEG 276 membranes did not
change significantly over the three months of pilot system operations.
Representative performance of the FEG 276 membranes for treatment of typical Trenton Engine
Complex wastewater, including analytical data for influent and permeate samples, is illustrated
in Table 4.
pH (std. units) 9.32 9.29 8.05 8.13 8.93 8.92 8.37 8.67 7.74 7.29 7.24 5.5 – 9.5
O&G, SGT- 220 <9.5 1,500 <9.4 270 <9.7 300 <9.5 52 <9.5 <9.4 -
HEM
Total solids 3,900 1,600 33,000 1,000 6,600 2,500 37,000 1,900 4,900 2,100 2,900 -
TSS 520 <4.0 3,600 8.0 250 5.0 1,400 4.0 50 <4.0 <4.0 -
COD 30,000 2,600 63,000 1,200 34,000 3,600 230,000 1,500 5,800 1,300 1,700 -
BOD5 2,600 1,300 >19,500 480 3,700 1,300 15,000 530 890 410 440 682
Total P 10 2.1 18 2.1 6.5 0.51 <0.10 0.69 5.0 3.1 3.6 4.64
Arsenic <0.10 0.010 <0.010 <0.010 <0.10 <0.010 <0.10 <0.010 <0.010 <0.010 <0.010 0.345
Barium <2.0 <0.20 <2.0 <0.20 <2.0 <0.20 <2.0 <0.20 <0.20 <0.20 <0.20 20.78
Boron 46 46 20 31 58 66 26 29 10 7 9 40.84
Cadmium <0.020 <0.0020 >0.020 <0.0020 <0.020 <0.0020 0.026 <0.0020 <0.0020 <0.0020 <0.0020 0.115
Chromium <0.050 <0.0050 0.064 <0.0050 <0.050 <0.0050 0.17 <0.0050 0.0051 <0.0050 <0.0050 3.45
Copper 1.2 0.068 3.7 <0.025 2.7 0.064 16 <0.025 0.5 <0.025 <0.025 0.703
Lead 0.20 0.022 0.52 <0.0030 0.27 0.015 2.1 0.0083 0.08 0.004 0.009 0.802
Mercury (ug/L) <0.20 <0.20 <2.0 <0.20 <2.0 <0.20 <2.0 <0.20 <0.20 <0.20 <0.20 <0.20
Nickel <0.40 <0.040 <0.40 <0.040 <0.40 <0.040 <0.40 <0.040 <0.040 <0.040 <0.040 1.552
Selenium <0.050 <0.0050 <0.050 0.0056 <0.050 0.0050 <0.050 <0.0050 <0.0050 <0.0050 <0.0050 0.075
Silver <0.050 <0.0050 <0.050 <0.0050 <0.050 <0.0050 <0.050 <0.0050 <0.0050 <0.0050 <0.0050 3.751
Zinc 6.6 0.83 17 0.53 6.0 0.67 47 0.44 0.9 0.1 0.05 2.82
*second permeate sample represents higher degree of concentration
**negotiated alternate mass limit of 18 kg/day
Wastewater characteristics at the Trenton Engine Complex can vary significantly when coolant
systems are emptied and replaced with fresh solution. The spent coolant solution is discharged
to an equalization tank at the wastewater pretreatment system, and the solution is slowly blended
with the incoming process wastewater until the tank is emptied. This scenario was evaluated by
operating several pilot system trials utilizing a concentrated coolant solution as the wastewater
feed source. As expected, the ability to further concentrate the used coolant solution was a
function of the initial concentration of oil & grease and other constituents in the feed wastewater.
It was possible to concentrate the used coolant solution by a factor of approximately 4.5, and a
50/50 mixture of used coolant and typical wastewater could be concentrated by a factor of 9.
However, permeate from these trials appeared cloudy after the concentration reached a factor of
1 or 2. Analysis of permeate samples revealed that HEM oil & grease (vegetable/animal-based)
was present in significant concentrations, although mineral-based oil & grease remained low.
Recovery of membrane flux rate through cleaning after the coolant trials was also shown to be
more difficult, and additional cleaning chemicals were needed to recover performance. It was
demonstrated that coolant dumps will need to be blended with influent wastewater to achieve
optimum performance.
Figure 14. Process flow diagram for Trenton Engine Complex process wastewater
pretreatment
DISCUSSION
The unique wastewater characteristics and circumstances for each automotive wastewater
pretreatment application lead to the use of three different types of ultrafiltration membranes. A
comparison of membranes and associated design characteristics is provided in Table 5.
The hollow fiber ultrafiltration system for FPI wastewater was installed and commissioned as a
pretreatment stage for removal of suspended and colloidal solids ahead of the nanofiltration
system for removal of color. While the ultrafiltration system was effective in removing
particulate matter from the FPI solution, the spiral-wound nanofiltration system continued to
perform poorly, exhibiting signs of fouling or membrane damage. The nanofiltration membrane
was removed from service for inspection, and it was observed that the membrane backing
material showed signs of compression and damage. The damage may have been caused by high
feed pressures, or by incompatibility between the backing material and the chemicals used in the
FPI dye solution. Due to the ongoing issues associated with operation of the nanofiltration
system, Chrysler elected to utilize an alternate technology to remove the residual color from the
FPI wastewater after pretreatment by ultrafiltration. A simple chemical oxidation system was
installed for batch treatment of the FPI wastewater. Ultrafiltration permeate is discharged into a
batch treatment reactor. When the batch reactor is filled, sodium hypochlorite is added and
mixed with permeate in the reactor. This method provides cost-effective and reliable removal of
residual color.
Several opportunities for reuse of pretreated wastewater at the Etobicoke Casting Plant were
evaluated. Pretreated die casting wastewater was considered for reuse in the facility’s cooling
towers, and for make-up water in the die casting machines. However, since the pretreated die
casting wastewater contains relatively high concentrations of organic material that passes
through the ultrafiltration membranes, additional pretreatment or chemical additives would be
required to avoid biological growth and minimize impacts to parts quality. The cost for
additional pretreatment and/or chemical additives, in addition to water storage and transfer
facilities that would be required for effective reuse, did not justify the benefits of wastewater
reuse at this time. Use of pretreated FPI wastewater for makeup water in the FPI process is
prohibited due to quality control requirements. Wastewater reuse will continue to be periodically
evaluated by Chrysler as part of its World Class Manufacturing initiative.
The on-site pilot study provided valuable information regarding the expected performance and
full-scale design requirements for the new ultrafiltration system. Through operation of the pilot
system over an extended period of time, it was possible to observe membrane performance under
varying wastewater characteristics, and to ensure that the membrane cleaning cycle was able to
routinely restore the membranes to near original performance. The pilot study also provided data
for selection of the optimum ultrafiltration membrane for the Trenton Engine Complex
wastewater, as well as evaluation of performance under non-routine conditions when
concentrated coolant solutions must be treated.
CONCLUSIONS
The use of advanced automotive manufacturing equipment and metalworking solutions can
result in increased production capability and quality. However, these advanced manufacturing
operations can also have significant impacts on process wastewater flow rates and
characteristics, and existing wastewater facilities may not have adequate pretreatment capacity,
or provide cost-effective pretreatment. Careful consideration of the impact of modified
manufacturing activities on existing wastewater pretreatment systems is warranted to ensure
continued discharge permit compliance, optimized pretreatment, and minimum life-cycle costs.
Although the application of low pressure membrane filtration is well demonstrated in the
automotive industry, wastewater characteristics are unique for each manufacturing facility.
Bench and/or pilot testing can provide important information regarding expected pretreatment
system performance and full-scale design considerations. Due to strict quality requirements in
the automotive manufacturing process, reuse of wastewater streams may be limited to those
wastewaters which contain relatively low concentrations of contaminants, or reuse for utilities or
other processes that do not directly impact the quality of automotive parts.
Finally, close communication and coordination with the municipal sewer discharge authority can
provide significant benefits with respect to wastewater discharge requirements. In cases where
wastewater flow rates have decreased as a result of modified manufacturing activities, it may be
possible to negotiate mass-based limits or increased concentration-based limits to allow
optimization of pretreatment processes and minimize operation and maintenance costs.
REFERENCES
Sunil Mehta1*, Nabin Chowdhury1, Denise Horner1, Antonio Lau2, Barbara Schilling3
1
Degremont’s North American Research and Development Center, Ashland, VA 23005
2
Degremont Technologies - Infilco, Richmond, VA 23229
3
Ozonia North America LLC, Leonia, NJ 07605
Abstract:
Micro-electronics manufacturing is one of the fast-growing industries in America, Europe, Asia
and Australia, and the semiconductor market grew by 61% in North America in 2011 compared
to the previous year. Exponential growth of thin-film-transistor liquid crystal display (TFT-LCD)
manufacturing generates a large amount of high-strength wastewater containing
monoethylamine, dimethyl sulphoxide, tetra-methyl ammonium hydroxide (TMAH), and many
other organic and inorganic compounds. This study was focused to develop a sustainable and
cost-effective processto treat the wastewaters from the micro-electronics industries.
Several bench-scale and pilot-scale studies were conducted to investigate complete degradation
and/or removal of TMAH employing biological (activated sludge) and chemical (advanced
oxidation) processes. Combination of aerobic oxidation followed by advanced oxidation and
biological denitrification was the most successful option for TMAH degradation and nutrient
removal from TFT-LCD wastewaters. The aerobic-AOP-anoxic system removed >98% TMAH,
Total Organic Carbon (TOC), ChemicalOxygen Demand (COD) and >96% Total Nitrogen (TN)
from wastewater containing TMAH of 5,000±100 mg/L, TOC of 2800±50 mg/L and TN of
750±30 mg/L. This patent pending biological-chemical process will benefit micro-electronics
industries and the ailing publicly owned treatment works (POTWs) attributed to the increased
nutrient loadings coupled with stringent regulations.
Background:
Micro-electronics industries (e.g., semiconductor, thin-film-transistor liquid crystal display
(TFT-LCD), chip manufacturing) generates a large amount of high-strength wastewater
containing monoethylamine (C2H5ONH2, MEA), dimethyl sulphoxide ((CH3)2SO, DMSO), tetra-
methyl ammonium hydroxide ( (CH3)4NOH, TMAH), and many other organic and inorganic
compounds. Many of these chemicals (e.g. TMAH) are poisonous, corrosive, slowly-
biodegradable and eutrophic to the aquatic environment. Furthermore, stringent discharge
regulations coupled with ailing POTWs are enforcing industries to remove organic and toxic
chemicals prior to discharge into industrial effluents and/or municipal sewers.
Physicho-chemical and/or biological oxidation of TMAH are options to treat this corrosive
alkaline chemical rich wastewater. Hirano et al. (2001), and Kim and Choi (2002) reported
catalytic oxidation and photocatalytic degradation of TMAH wastewater. Again Shibata et al.
(2006) studied ion-exchange resins for TMAH recovery from the wastewater. Even though these
treatment methods avoid production of ammonia (NH3) and nitrogen oxides (NOX), no pilot-
scale and/or full-scale application was found in the literature.
Few studies (Chen et. al., 2003; Chang et al., 2008; Hu et. al., 2012) reported biological
treatment of TMAH-containing wastewater using activated sludge where methylotrophs and
Paracoccus spp. degrade TMAH (Anthony, 1982). The author proposed metabolic pathways for
complete TMAH oxidation to formaldehyde and ammonium (Figure 1). TMAH is hydrolyzed to
tri-methyl ammonium (TMA), di-methyl ammonium (DMA), mono-methyl ammonium (MA),
and finally MA is hydrolyzed to ammonium and formaldehyde. There are two routes in each of
the oxidation reactions, one in the presence of oxygen and the other one in absence of oxygen
(Anthony, 1982).
Chen et al. (2003) reported 92-98% COD removal from TFT-LCD wastewater using a 2-stage
anoxic-aerobic activated sludge process. However, the authors experienced ≤70% Nitrogen
removal in the same experimental set-up. Lin et al. (2011) attempted anaerobic oxidation of
ammonium (ANAMMOX) to remove organic nitrogen using anaerobic hydrolysis. Although
biological processes such as anoxic-aerobic, anoxic-aerobic-membrane bioreactor, aerobic-
anoxic-aerobic process (Whang et al., 2012), methanogenic up-flow anaerobic sludge blanket
(Chang et al., 2008), sequencing batch reactor, and membrane bioreactor (Wu et al., 2008) were
able to degrade TMAH, inhibition of the biological processes e.g. nitrification, biodegradation,
etc. were documented by several studies (Lin et. al., 2011; Hu et. al., 2012; Whang et al., 2012).
Even though few studies reported TMAH degradation and nitrogen removal, complete removal
of TMAH and organic nitrogen is still warranted. This study was focused to develop a
sustainable and cost-effective wastewater treatment process for the semi-conductor industries..
Objectives:
Treatability of TMAH loaded wastewater was conducted to remove >95% TOC, COD and TN
from organic rich liquid wastes from the semiconductor industry.. The detailed objectives of this
study were as follows:
• Develop sustainable processes for >98% TMAH and TOC removal from
semiconductor wastewaters.
• Achieve >95% Total Nitrogen removal employing advanced oxidation coupled with
biological treatment for wastewaters generated from the semiconductor industry.
• Develop a Conceptual Process Design for a full-sale treatment plant to treat micro-
electronics industrial wastewaters .
Methodology:
Industrial waste streams, such as the Liquid Organic Wastes (LOW) stream and the TMAH
Waste stream, were obtained from a semiconductor manufacturing facility in the USA. Initially,
batch tests were conducted to evaluate the treatability of TMAH-laden waste streams employing
advanced oxidation processes (AOP) and/or biological treatment. AOP alone was found
ineffective to remove carbon and nitrogen associated with TMAH and other complex organic
matter present in the waste streams. However, biological treatment of AOP treated stream was
promising to meet the targeted effluent discharge criteria. Hence, a combination of biological
and advanced oxidation processes (anoxic-aerobic-AOP-anoxic and aerobic-AOP-anoxic) were
employed to achieve the complete destruction of organic and toxic matter present in the waste
streams. The LOW stream was treated in the anoxic-aerobic-aerobic/AOP-anoxic system (Figure
2A) and the TMAH stream in the aerobic-AOP-anoxic system (Figure 2B). The individual
treatment processes have the following functions:
• Pre anoxic biological treatment for degradation of organic present in the LOW stream
• Aerobic biological treatment for TMAH breakdown into ammonia
• AOP for the oxidation of ammonia to nitrate
• Post anoxic biological treatment to denitrify the nitrate produced in the AOP
0.95 L 0.95 L
15 L 0.8 L
27 L 4L
Two continuous flow biological treatment pilot units were set-up at Degremont’s DENARD
R&D facilities in Ashland, VA. The biologically treated waste streams were sent to Ozonia’s
R&D Laboratory for AOP treatment on a weekly basis. The wastewater was then returned to
DENARD for further treatment employing a biological system (post- anoxic, denitrification).
Analytical Methods:
Water quality parameters (alkalinity, sCOD, tCOD, sBOD5, MLSS, MLVSS, NH4-N, NO3-N,
NO2-N, sTKN, TKN, PO4-P and TP) were analyzed following Standard Methods for
Water/Wastewater Analysis (APHA, 1998). Reactor pH, temperature, DO and ORP were
monitored using a multimeter (HQ40d, HACH). TMAH analysis was outsourced to the
commercial analytical laboratory, BALAZS NanoAnalysis (Fremont, CA).
Standard Methods for COD analysis using the HACH kit was not applicable for organic
amines (e.g. TMAH) as shown in Figure 4(B). Good agreement between the
“Theoretical” and “Analytical” TOC and TKN in TMAH (Figure 3) was observed,
therefore, TKN and TOC values were employed as indicator of amines. Correlations
between “TMAH and TOC”, and “TMAH and TKN” were developed to estimate TMAH
degradation in the reactors based on TKN and TOC concentrations.
600
B
500 A R² = 0.999
R² = 0.990
TMAH (mg/L)
400
300
200
TOC
100
TKN
0
0 50 100 150 200 250 300
TKN, TOC (mg/L)
Figure 3: (A) “TMAH-TOC” and “TMAH-TKN” Correlations; (B) Theoretical and Actual
Concentrations of TOC and TKN in TMAH
Wastewater Characteristics:
Three waste streams (Stream-1, Stream-2, and Stream-3) were received from a semiconductor
manufacturing facility. Detailed wastewater characteristics are shown in Table 1. A blend of
Stream-1 and Stream-2 (at 4:5 ratio) was termed as the Liquid Organic Waste (LOW) stream.
TMAH-laden (>3%) Stream-#3 was treated separately primarily due to the concentrated and
toxic nature of TMAH.
Table 2 shows additional characteristics of the waste streams: organic content parameters (COD,
BOD5, TOC) and nitrogen containing species. TMAH was analyzed by a certified laboratory,
primarily due to the complexity of the analytical procedure.
The biological systems shown in Figures 2A and 2B were fed with the LOW stream and TMAH
stream at 1:60 and 1:6 dilution, respectively. Table 3 shows influent characteristics of the TMAH
loaded liquid organic waste (LOW) stream and the TMAH only stream. The influent was
blended with essential micro/macro nutrients (metals, alkalinity, and phosphorus) for proper
microbial metabolism. Influent pH was first adjusted to 7±0.1 using concentrated HCl.
COD Performance
4,400 120%
4,000
COD Removed, %
3,200
2,800 80%
2,400
60%
2,000 Inf COD
1,600 Eff COD 40%
1,200 % COD Rem
800 20%
400
0 0%
9/4/2012 9/16/2012 9/28/2012 10/10/2012 10/22/2012 11/3/2012
TOC Performance
2500 100%
Inf TOC
TOC Concentration, mg/L
Eff TOC
2000 80%
% TOC Rem
TOC Removed, %
1500 60%
1000 40%
500 20%
0 0%
9/4/2012 9/14/2012 9/24/2012 10/4/201210/14/201210/24/2012
TN Performance
350 100%
Inf TN
300 Eff TN
TN Concentration, mg/L
% TN Rem 80%
TN Removed, %
250
200 60%
150 40%
100
20%
50
0 0%
9/4/2012 9/16/2012 9/28/2012 10/10/2012 10/22/2012 11/3/2012
Figures 4A, 4B, 4C: Anoxic-Aerobic Reactor Performance (A) COD, (B) TOC, and (C) TN
The aerobic-AOP-anoxic system efficiently removed >98% TOC and >95% TN. Furthermore,
the effluent TMAH of <0.5 mg/L shown in Table 6 as reported by the certified laboratory
(BALAZS) confirmed complete degradation of TMAH in the aerobic bioreactor. Figures 5A and
5B show aerobic-AOP-anoxic system performance. Even though the influent TMAH
concentration was 4,000 mg/L, no toxicity and/or inhibition to the biological treatment system
was observed over the experimental test duration of >120 days.
TOC Performance
4,000 100%
80%
TOC Concentration, mg/L 3,000
TOC Removed, %
60%
2,000 Inf TOC
Eff TOC 40%
% TOC Removal
1,000
20%
0 0%
11/12/201211/17/201211/22/201211/27/2012 12/2/2012 12/7/2012
TN Performance
1000 100
800 80
TN, NH4-N Concentration, mg/L
TN Removed, %
600 Inf TN 60
Final Eff TN
NH4-N in Bio-Effluent
400 40
Bio Eff TKNs
% TN Removal
200 20
synthetic
0 0
11/12/2012 11/20/2012 11/28/2012 12/6/2012 12/14/2012 12/22/2012
Figures 5A, 5B: Aerobic-AOP-anoxic system performance (A) TOC and (B) TN
Conclusions:
A combination of aerobic biodegradation of TMAH followed by AOP and biological
denitrification increased the overall removal efficiency of TN, TOC and TMAH from TFT-LCD
manufacturing wastewaters. The major findings of this study are as follows:
• >98% removal of TMAH, TOC, COD and >96% removal of TN were achieved in the
activated sludge processes coupled with advanced oxidation.
• Aerobic biological treatment alone degraded the TMAH intoammonia (NH4-N).
• Advanced oxidation of the biologically treated TMAH-laden wastewater completely
converted the ammonia(NH4-N) into nitrates (NO3-N).
• Post-anoxic biological denitrification of the AOP treated effluent achieved complete
conversion of nitrates (NO3-N) into nitrogen gas and increased the overall TN removal
efficiency of the treatment system.
References:
1. APHA, Standard Methods for the Examination of Water and Wastewater, 1998
2. American Water Intelligence: Market Leading Analysis of the International Water
Industry, 2012, 3(6)
3. Anthony, C. The Biochemistry of Methylotrophs, 1982
4. Chang, K.F.; Yang, S.Y., You, H.S.; Pan, J.R. IEE Trans. Semiconductor Manu., 2008,
486-491
5. Chen, T.K.; Ni, C.N.; Chen, J.N. J. Environ. Sci. Health, 2003, 38, 2157-2167
6. Global Water Intelligence: Market Leading Analysis of the International Water Industry,
2009, 10(10)
7. Hirano, K.; Okamura, J.; Taira, T.; Sano, K.; Toyda, A.; Ikeda, M. IEE Trans.
Semiconductor Manu., 2001, 14, 202-206
8. Hu, T.H.; Whang, L.M.; Liu, P.W.; Hung, Y.C.; Chen, H.W.; Lin, L.B.; Chen,C.F.; Hsu,
S.F.; Shen, W., Fu, R.; Hsu, R. Bioresource Technol., 2012, 113, 303-310
9. Kim, S.; Choi, C. Environ. Sci. Technol., 2002, 36, 2019-2025
10. Lin, H.L.; Chen, B.K.; Hsia, H.P.; Yang, G.H.; Yang, Y.F.; Chao, Y.C.; Cheng, S.S.
Sustain. Environ. Res., 2011, 21, 155-160
11. Shen, W.; Huang, C.K.; Fu, R.; Chang, B. Bioresource Technol., 2012, 122, 70-77
12. Shibata, J.; Murayama, N.; Matsumoto, S. Resources Processing, 2006, 53, 199-203
13. Whang, L.M.; Wu, Y.J.; Lee, Y.C.; Chen, H.W.; Fukushima, T.; Chang, M.Y.; Cheng,
S.S.; Hsu, S.F.; Chang, C.H.; Wu, Y.J.; Whang, S.J.; Huang, Y.F.; Yang, Y.F.; Lei, C.N.
Water Sci. Technol., 2008, 1085-1093