ENVIRONMENTAL PROTECTION AGENCY
            OFFICE OF ENFORCEMENT
                EPA 330/1-75-001
  Waste Treatment and Disposal Methods
                  for the
          Pharmaceutical Industry
NATIONAL FIELD INVESTIGATIONS CENTER-DENVER
             DENVER.COLORADO
               FEBRUARY 1975

-------
   ENVIRONMENTAL PROTECTION AGENCY
        OFFICE OF ENFORCEMENT
WASTE TREATMENT AND DISPOSAL METHODS

               FOR THE

       PHARMACEUTICAL INDUSTRY
                 by

         E. J. Struzeski, Jr.
National Field Investigations Center
          Denver, Colorado
             March 1975

-------
                              CONTENTS
  I.    INTRODUCTION  .....................     1

 II.    SUMMARY ........................     5
III.    INDUSTRY CATEGORIZATION   ...............    13
       STANDARD INDUSTRIAL  CLASSIFICATION  ..........    13
       KLINE GUIDE  .....................    I4
       GULF SOUTH RESEARCH  INSTITUTE   ............    14
       NATIONAL FIELD INVESTIGATIONS  CENTER-DENVER   .....    16

 IV.    PROCESS  DESCRIPTION   .................    20
       UNIT PROCESSES ....................    20
         Fermentation ....................    20
         Synthesized Organic Chemicals  ...........    20
         Biologicals  ....................    20
         Drug Formulation ..................    20

       TYPICAL  PLANT OPERATIONS ...............    20
         Fermentation ....................    21
           Vitamins  .....................    21
           Citric Acid  .........  * ..........    24
           Antibiotics  ...................    25
           Enzymes  .....................    31
         Synthesized Organic Chemicals ............    31
           Vitamins  .....................    31
           Antibiotics  ...................    32
           Sulfa Drugs  ...................    33
           Steroids  .....................    34
           Prostglandins  ..................    40

         Fermentation/Synthesized Organic  Chemicals .....    41
           Filtration ....................    41
           Funda Filter ...................    41

         Drug Formulation ..................    44
  V.    WASTE CHARACTERISTICS  ................    45

       FERMENTATION  .....................    46
         The Upjohn  Company, Kalamazoo,  Mich .........    46
         Bristol Laboratories,  Syracuse, N.Y .........    47
         A Pharmaceutical Plant in India  ..........    48
         Antibiotic  Wastes, Great Britain  ..........    48

       SYNTHESIZED ORGANIC  CHEMICALS .............    50
         Squibb and  Sons, Humacao, Puerto  Rico ........    50
         Berkeley Chemicals, Summit,  N.J ...........    50
         Parke-Davis, Holland,  Mich .............    53
         M/s Indian  Drugs,  Hyderabad, India .........    54
         Upjohn Laboratories, Kalamazoo, Mich ........    58
                                   m

-------
                      CONTENTS (cont.)


         BIOLOGICALS	    58
          Eli Lilly, Greenfield, Ind	    58

         DRUG FORMULATION	    60
           Dorsey Laboratories, Lincoln, Nebr	    60

  VI.    WASTE PARAMETERS OF SIGNIFICANCE	-.  .  .  .    63

         BOD:  A QUESTION OF RELIABILITY	    63
           Abbott Laboratories, N.  Chicago,  111	    63
           Dorsey Laboratories, Lincoln, Neb	    65
           Experiments in Great Britain 	    66

         TOXICITY   	    66
           Antibiotics	    66
           Phenol-Mercury Compound  	    69
           Hormone	    69

         NITROGEN REDUCTION 	    74

 VII.    WASTE RECOVERY AND CONTROL 	    75
         SOLVENT RECOVERY 	    75

         BYPRODUCT RECOVERY-ANTIBIOTICS PRODUCTION  	    76

         BYPRODUCT RECOVERY-TECHNOLOGICAL TRANSFER	    78
           United States  	    78
           England	    80

         WASTEWATER REDUCTION AND RECOVERY METHODS  	    81
           Bristol Laboratories, Syracuse, N.Y	    81
           Upjohn Laboratories, Kalamazoo, Mich	    83
VIII.    WASTE TREATMENT AND DISPOSAL 	    85

         ACTIVATED SLUDGE 	    85
           Fermentation 	    85
             Design Criteria for Pharmaceutical  Wastes  ....    85

           Synthesized Organic Chemicals  	    87
             Biological  Treatment of Pharmaceutical
                 Chemical  Waste 	    87
             Nitroaniline Isomers 	    88
             Squibb and Sons,  Humacao,  Puerto Rico	    90
             M/s Indian Drugs, Hyderabad, India  	    91
             Hoffman-LaRoche,  Belvidere, N.J	    93

           Fermentation/Synthesized Organic  Chemicals  	    98
             Abbott Laboratories, N.  Chicago, 111	    98
             Wyeth Laboratories, W.  Chester, Pa	108

           Fermentation and  Biologicals  	   110
             Lederle Laboratories,  Pearl River,  N.Y	110
                               IV

-------
                         CONTENTS  (cont.)
         Drug Formulation 	  .....   113
           Dorsey Laboratories,  Lincoln,  Neb	   113
       TRICKLING FILTRATION 	   117
         Fermentation 	   117
           The Upjohn Co.,  Kalamazoo,  Mich	   117
           Great Britain  Plants	   118
         Fermentation/Synthesized  Organic Chemicals  	   119
           The Upjohn Co.,  Kalamazoo,  Mich	   119
           American Cyanamid,  Willow  Is., W.  Va	   120

         Fermentation and Biologicals  	   122
           Lederle Laboratories, Pearl  River,  N.Y	   122
         Biologicals	   124
           Eli Lilly, Greenfield,  Ind	   124
         Drug Formulation	   127
           Merck, Sharpe, and  Dohme,
             West Point,  Pa	   127

       OTHER TREATMENT METHODS	   130
         Anaerobic Filters.  .  . -	  .  .  .   130
         Spray Irrigation	   131
           The Upjohn Co.,  Kalamazoo,  Mich	   131

         Oxidation Ponds	   132
           Studies in India	   132

         Sludge Stabilization  	   133
           Lederle Laboratories, Pearl  River,  N.Y	   133
         Deep Well Injection	   135
           Parke-Davis, Holland, Mich	   135
           Abbott Laboratories,  Barceloneta,  Puerto  Rico.  .  .   136
           The Upjohn Co.,  Kalamazoo,  Mich	   136
IX.     DEVELOPMENT OF EFFLUENT LIMITATIONS	   141

       EXEMPLARY PLANT	   141
       AVAILABLE TREATMENT  AND DISPOSAL PROCESSES  	   142
       MODEL SYSTEMS	   144
       EFFLUENT LIMITATIONS 	   149

       BIBLIOGRAPHY 	   153

       APPENDIX:  CASE HISTORIES OF THE
         PHARMACEUTICAL INDUSTRY	   163

-------
                              CONVERSIONS
multiply
Metric Unit*
'Celcius (C
centimeters (cm)
cubic meters (m3)


grams (g)
hectares (ha)
kilograms (kg)
liters (1)
meters (m)
metric tons
micrometer (^m)
million gallons/day (mgd)
square meters (m2)
       by
  9/5 (C)  + 32
(absolute:  9/5 C)
     0.394
     35.32
     1.307
     0.035
     264.2
     0.035
     2.471
     2.205
     0.264
     1.057
     3.281
     1.094
     1.102
     0.039
     3,770
     10.76
     1.196    ,,
  2.471  X 10"
                                    to obtain
                                    English Unit
                                    "Fahrenheit (F)
                                    inches (in)
                                    cubic feet (ft3)
                                    cubic yards   (yd3)
                                    bushels (bu)
                                    gallons (qal)
                                    ounces (oz)
                                    acres
                                    pounds (Ib)
                                    gallons (gal)
                                    quarts (qt)
                                    feet (ft)
                                    yards (yd)
                                    short tons (tons)
                                    inches (in)
                                    cubic meters/day  (m3/day)
                                    square feet  (ft2)
                                    square yards   (yd2)
                                    acres
*Metric prefixes:
mega (M)	106
kilo (k)	103
hecto (h)	102
deca- (da)	101
                deci  (d)	1Q12
                centi  (c)	10~3
                milli  (m)	10~6
                micro  (u)	10"

-------
                                  ABBREVIATIONS
ASTM       American Society for
              Testing and Materials
avg        average
BOD        biochemical oxygen demand
             (BOD5 unless otherwise
             indicated)
COD        chemical oxygen demand
cpm        cycles per minute
DO         dissolved oxygen
FDA        Federal Drug Administration
F/M        Food to micro-organism ratio
FWQA       Federal Water Quality
             Administration
gpd        gallons per day
gpm        gallons per minute
gsfpd      gallons per square
             foot per day
hp         horsepower
hr         hour
IU         International Units
kwh        kilowatthour
y          micron
ym; yg     micrometer; microgram
mgd        million gallons
             per day
ML(V)SS     mixed liquor (volatile)
             suspended solids
MSG         monosodium glutamate
NOD         nitrogenous oxygen demand
OH"         (hydroxide group)
O&M         operations & maintenance
PE          population equivalents
pH          (hydrogen-ion activity)
ppb         parts per billion
ppm         parts per million
psi(a)      pounds per square inch
              (absolute)
psig        pounds per square inch
               gage
Pt-Co
units       Platinum-Cobal units
SRT         solids retention time
TDS         total dissolved solids
TLm         median toxic limit
TOC         total organic carbon
TS          total solids
TSS         total suspended solids
TVS         total volatile solids
                                       vn

-------
                           I.  INTRODUCTION

     In early 1974, NFIC-D was asked by EPA Regions II,  III,  IV  and  V  to
prepare National Pollutant Discharge Elimination System  (NPDES)  permits
for ten major pharmaceutical  manufacturing plants in Pennsylvania,
Virginia, Indiana and Puerto Rico.   Effluent guidelines,  the  basis for
establishing effluent limitations specified in the permits,  had  not
been developed for the pharmaceutical  industry.  It was  thus  necessary
to develop a rationale for establishing effluent limitations  that would
fulfill two requirements.  They must be neither too lenient  nor  too
stringent relative to prevailing waste treatment and control  practices
in the industry, and waste treatment improvements necessary  to meet
assigned limitations must be practicable.

     The ten specified plants were visited to obtain process  descrip-
tions, information on waste treatment facilities and wasteload data.
Also, available literature on these plants was reviewed.   Because of the
diversity of manufacturing and waste treatment processes  at  the  ten
plants, as well as the different levels of treatment efficiency  achieved,
a review of waste treatment and control practices at other pharmaceutical
plants became necessary to develop an acceptable rationale.   The study
was expanded to include visits to five additional plants  and  an  extensive
literature review of 120 articles covering all aspects of waste  treat-
ment and control in the industry.  In all, data were collected on 39
pharmaceutical plants.

     Case histories (see Appendix)  were prepared for nine of  the ten
plants for which permits were drafted (listed on the following page).
These case histories summarize available literature on production pro-
cesses, waste treatment and control and waste!oads at each plant, give
detailed information on current waste treatment facilities,  summarize
recent wasteload data, and present effluent limitations  established  for
each plant.

-------
          Commercial Solvents Corporation, Terre Haute, Ind.  '(3)
          Eli Lilly and Co., Clinton Laboratories, Clinton, Ihd.  (3)
          Eli Lilly and Co., Mayaguez, Puerto Rico (2)
          Eli Lilly and Co., Tippecanoe Laboratories, Lafayette,  Ind.(3)
          McNeil Laboratories, Fort Washington, Pa.  (5)
          Merck and Company, Inc., Cherokee Plant, Danville,  Pa.  (3)
          Merck and Company, Inc., Stonewall  Plant,  Elkton, Va.  (3)
          Pfizer, Inc., Terre Haute, Ind. (3)
          Wyeth Laboratories, Marietta-, Pa. (4,5)
          Wyeth Laboratories, Paoli, Pa. (5)
     To expand the data base for developing the effluent limitations,

varying combinations of literature review and plant  visits were 'used  for

these additional 29 plants:

          Abbott Laboratories, North Chicago, 111-. (3,5)*
          Abbott Laboratories, Sidney, Australia (3)
          Abbott Laboratories, Barceloneta, Puerto Rico (3)
          American Cyanamid Co., Calco Chemical Division-,
               Willow Island, W. Va. (3)
          Berkeley Chemicals Division, Summit, N.J.  (2)
          Boots Pure Drug Co., Ltd., Nottingham, Great Britain (2)
          Bristol Laboratories, Syracuse, N.Y. (3)
          Dorsey Laboratories, Lincoln, Neb.  (5)
          Eli Lilly Greenfield Laboratories,  Greenfield, Ind. (4)
          Hoffman-LaRoche, Belvidere, N.J.(2)*
          Hoffman-LaRoche, Nutley, N.J.(3)
          Lederle Laboratories, Pearl River,  N.Y. (3;4)*
          Merck and Company, Inc., Albany, Ga. (2)
          Merck and Company, Inc., West Point, Pa. (5J
          Merck and Company, Inc., Barceloneta, Puerto Rico (2)
          M/s Indian Drugs and Pharmaceuticals, Hyderabad, India  (2)
          Parke-Davis and Co., Holland, Mich. (2,3)
          Parke-Davis and Co., Hounslow, Great Britain (2)
          Pfizer, Inc., Groton, Conn. (3)
          Pfizer, Inc., Brooklyn, N.Y. (3)
          Pfizer, Inc., Folkestone, Great Britain (5)
          Pfizer, Inc., Sandwichj Kent, Great Britain (3)
          Salisbury Laboratories, Charles City, Iowa (2)
          Schering-Plough, Inc., Kenilworth,  N.J. (5)
          Schering-Plough, Inc., Union, N.J.  (3)*
          Squibb, E.R.  and Sons, Inc., Humacao* Puerto Rico (2)
          Upjohn Company, Kalamazoo, Mich. (3)
          Warner-Chilcott, Morris Plains, N.J. (5)*
          Wyeth Laboratories, West Chester, Pa. (3)

This study thus evolved into a broad and detailed review of waste treatment

and control  practices in the pharmaceutical industry.
t  Parenthetic numbers indicate NFIC-D industrial  categories* described in
   Section III
*  Plants visited

-------
     Although difficult to define, the pharmaceutical  industry is  gen-
erally involved in the production of medicinal  chemicals,  biologicals,
and botanicals having therapeutic value for humans and animals.   Pro-
ducts of the industry are used for industrial,  agricultural,  and domes-
tic purposes and are found in such familiar forms as cleaning agents,
pesticides, fertilizers, cosmetics and baked goods.*  The  industry is
growing at the rate of 9 percent annually.   In  1973, an estimated  $1.4
billion of medicinals ($665 million), biologicals ($540 million) and
botanicals ($185 million) were sold as Pharmaceuticals or  food and feed
supplements.

     The pharmaceutical industry employs a  vast array of complex pro-
cesses, many of which are proprietary.  This report is not intended to
be an exhaustive presentation of such processes; appropriate  texts of
Pharmaceuticals and chemical engineering will provide greater detail on
processes.  However, pertinent processes are discussed in  sufficient
detail to establish the unit processes involved at plants  of  interest  as
they determine wasteloads and characteristics.

     Wastes from pharmaceutical operations  are  extremely strong and
concentrated, difficult to handle, and require  some of the most complex
and expensive treatment and control systems of  any industry.   This
report presents a detailed description of waste characteristics and
current waste treatment and control practices in the industry.

     The information derived from this study of the pharmaceutical
industry is presented in ten main sections  of this report, including:
Summary, Industrial Categorization, Process Description, Waste Char-
acterization, Waste Parameters of Significance, Waste Recovery and
Control, Waste Treatment and Disposal, Development of Effluent Limita-
tions, and Case Histories (Appendix).
    For an extensive list of pharmaceutical  products see:
    Pharmaceuticals (104), Drugs in Current  Use and New Drugs (85),
    Chemistry in the Economy (116), The Merck Index (71).

-------
                             11.   SUMMARY
     This report provides the rationale for establishing effluent
limitations which were used by NFIC-D in preparation of NPDES permits
for ten major pharmaceutical manufacturing industries in the United
States and Puerto Rico.  Establishing the effluent limitations entailed
plant visits, data collection and evaluation, and an extensive litera-
ture review for 39 plants.  Thus, this report serves as a state-of-the-
art review of current waste treatment and control technology existing
throughout the pharmaceutical industry.

     Information derived from the study is condensed in the following
Section summaries.
INDUSTRY CATEGORIZATION

     A wide variety of processes and raw materials are used by the
pharmaceutical industry to produce a broad spectrum of final products.
The characteristics of resulting wastewaters are equally varied.  To
facilitate evaluation of waste treatment practices, the industry was
divided into categories having similar processes, waste disposal prob-
lems and waste treatment practices.

     Five major categories were defined.  Fermentation Plants employ
fermentation processes to produce medicinal chemicals.  In contrast,
Synthesized Organic Chemicals Plants produce medicinal chemicals (fine
chemicals) by organic synthesis processes.  Most plants are actually
combinations of these two, yielding the third category, Fermentation/
Synthesized Organic Chemicals Plants.   Biologicals Production Plants
produce vaccines and antitoxins.  In the final category, the Drug
Mixing, Formulation and Preparation Plants produce pharmaceutical
preparations in final form, such as tablets and capsules.
PROCESS DESCRIPTION

     Process Description includes many typical manufacturing schemes,
unit processes, and typical plant operations.  Unit processes are
enumerated for fermentation, synthesized organic chemicals, biologicals
production and for drug formulation.  Under fermentation are described
fermentation media, the chemistry and some of the manufacturing steps
for vitamins, particularly vitamins B2, B,2, and C; citric acid; the
antibiotics, particularly penicillin, Terramycin, and streptomycin;
and also enzymes production.  Penicillin preparation is explained for
Bristol Laboratories and Wyeth Laboratories.  The Terramycin process is
detailed both in theory and actual production at two Pfizer, Inc. plants
in Great Britain.

-------
     Synthesis of organic chemicals focuses on chemistry and typical  or
special operations necessary for vitamins B^ and C, the synthetic
antibiotic Chloromycetin, the broad group or sulfa drugs, and the
important classifications of steroids and prostglandins.  Chloromycetin
processing is illustrated for the Parke-Davis plants at Hounslow, Great
Britain and Holland, Mich.  Special attention is devoted to the early
Upjohn Co. process for cortisone, the first important steroid.  Corti-
sone was a forerunner in steroid chemistry and many ensuing discoveries.
Steroids include the anti-inflammatory agents, sex hormones, tissue
building agents and contraceptives.  Prostglandins represent one of
newest of pharmaceutical chemicals available.  They simulate functions
of the prostrate gland in the body but their reactions are still yet
largely unknown.  Filtration techniques as employed in pharmaceutical
works are related to intended uses, and absolute and depth-type filters
are compared.

     Description is also given of typical drug formulation, mixing and
packaging operations leading to final products as purchased in the local
drug store.
WASTE CHARACTERISTICS

     This Section of the report reviews major types of waste resulting
from fermentation, synthesized organic chemicals and biologicals pro-
duction.  Specific waste sources and general character of wastes are
cited.  Data are presented for penicillin, streptomycin and steroids
manufacturing with emphasis on the composition of fermentation broths
and their handling and disposal.  BOD content of spent fermentation
beers can exceed 35,000 mg/1.  The disposition and degree of solvent and
mycelium recovery will greatly influence waste strength and loads.
Characterization of waste streams from synthesized organic chemicals
production is given for Squibb, Inc., Humacao, Puerto Rico (synthetic
penicillin and antifungals), Berkeley Chemicals Division, Summit, N. J.
(various chemicals), Parke-Davis and Co., Holland, Mich, (largely
Chloromycetin), and M/s Indian Drugs, Hyderabad, India.  Extensive
technical data on products, type of waste, and toxicity rating culmin-
ating from more than ten years of study at M/s Indian Drugs are given.

     Synthesized chemical production wastes are strong, difficult to
treat, and frequently inhibitory to biological treatment systems.
Chemical wastes from the factory can change significantly from day to
day.  Biological production wastes are strong and contain animal re-
mains, culture media, pathogenic organisms, toxic elements, and possibly
pesticidal traces.  Information is made available on waste characteris-
tics associated with the antitoxins, antisera and vaccines production
facility of Eli Lilly at Greenfield, Indiana.


WASTE PARAMETERS OF SIGNIFICANCE

     Special  attention is given to studies dealing with BOD, toxicity,
and nitrogen  reduction.  Major parameters of importance in the pharmaceutical

-------
 industry include BOD5 and long-term BOD, COD, TOC, solids, pH, a wide
 array of trace metals and other troublesome ions potentially present,
 unoxidized nitrogen, phosphorous, color, odor, temperature and toxicity
 of pharmaceutical effluents.

     Difficulties have been cited in using "Standard Methods" or equiva-
 lent procedures in analyzing some pharmaceutical wastes.  Concern has
 been expressed on validity of the BOD test due to toxicity.  Nitrifi-
 cation has also been reported during the course of the 5-day BOD test,
 causing further difficulty.  Opinion seems fairly evenly divided on
 appropriateness of BOD.  COD or TOC would serve to augment BOD results
 in the event of question.  Toxicity impact upon biological treatment
 systems was described using various antibiotics, a phenol-mercury com-
 pound having bactericidal properties, the female sex hormone DES, and
 formaldehyde and methyl alcohol wastes.  The alcohol and formaldehyde
 wastes showed no toxicity.  DES could be degraded by activated sludge
 but only over unusually long treatment periods.  The antibiotics demon-
 strated toxicity, but the biological system could be acclimated.  In
 some cases, activated sludge could successfully treat the phenol-mercury
 waste, whereas in other cases the biological system could tolerate but
 not adequately degrade this waste.

     Ammonia and organic nitrogen levels -in typically treated pharma-
 ceutical waste may range upward to a few hundred or as much as a few
 thousand mg/1 when expressed as Kjeldahl nitrogen.  Nitrogen loads may
 be appreciably high as to even exceed BOD effluent loads.  'Discharge
 loads have been reported in the range of 454 kg (1,000 lb)/day to 3,360
 kg (8,000 lb)/day.  Methods used in removing and reducing the various
 forms of unoxidized and oxidized nitrogen are reviewed.


 WASTE RECOVERY AND CONTROL

     Industry situations are described which have significantly reduced
 waste problems by recovery of valuable solvents, recovery of selected
 waste streams, especially spent fermentation broth as animal feeds,
 and specific wastewater reduction and recovery programs instituted by
 certain companies.  The collection, evaporation and sale of dried fer-
 mentation solids has not only been practiced in the pharmaceutical field
 but has been implemented by the distilling and brewing industries as far
 back as the early 1950's.  A number of cases are documented on fermen-
 tation solids recovery.

     A Hiram Walker distillery without feed recovery had a wasteload of
 50 to 55 population equivalents of BOD per bushel  of grains processed.
 With an "entire plant" approach to recovery and without resorting to
 customary conventional treatment, the distillery showed 5.6 PE/bu grains
 in 1951-52.   By virtue of further improvements, the Company reduced its
wasteload to about 1.25 PE/bu grains, equivalent to about a 96 to 97
 percent waste reduction.

-------
     In pharmaceutical manufacturing, the value of nutrients and undeter-
mined growth factors contained in fermentation broth solids is thought
quite high when used as animal feeds.  Examples of fermentation solids
recovery include Abbott Laboratories at North Chicago, 111., the Upjohn
Company, Kalamazoo, Mich., and Abbott Laboratories, Barceloneta, Puerto
Rico.  Detailed information is presented on these practices, including
cost data for Abbott and Upjohn.  Both Pfizer, Inc., Terre Haute, Ind.
and Lederle Laboratories at Pearl River, N. Y. previously evaporated
spent fermentation beers throughout the 1950's and 1960's, but they have
subsequently discontinued such practice.  Abbott, North Chicago in 1972-
73 recovered beers with a BOD5 load potential of 20,000 Ib/day or greater,
and together with activated siudge these practices were responsible for
the overall plant BOD removals achieved of 98.7 percent.  Upjohn re-
ported that BOD reductions directly due to triple-effect evaporation
ranged from 96 to 98 percent for four different types of antibiotic
spent beers.  Recovery of spent beers saves on expensive alternative
biological treatment.

     Intense recovery of solvents represents an extremely important
aspect for pharmaceutical plants.  A commercial ketone solvent was
reported as having a BOD of approximately 2 million mg/1, or some 9,000
times stronger than untreated domestic sewage.  A gallon of this solvent
was calculated as equivalent in BOD to the sewage coming from a city of
77,000 people.  Amyl acetate, another common solvent, is reported to
have a BOD strength of about 1 million mg/1, and acetone shows a BOD of
about 400,000 mg/1.  A very small variation in quantity of solvents
lost to the plant sewer can have a strong impact on treatment facilities
and the receiving stream.

     A preventive program at the Upjohn Company, Kalamazoo, Mich, in the
mid-1950's served to reduce raw chemical wastes from a peak flow of
2,840 mj (750,000 gal)/day to 284 rrT(75,000 gal)/day.  BOD loads corres-
pondingly dropped from 13,600 kg (30,000 lb)/day to about 1,360 kg (3,000
lb)/day.  These reductions were in part due to installation of a spent
solvents incineration system.

     Bristol Laboratories, Syracuse, N.Y., in the 1950's prior to waste
control implementation, showed a BOD. of 0.74 PE/gal fermentation wastes.
When contaminated or spoiled fermentation batches were controlled and
with improved monitoring and recovery of solvents, Bristol Laboratories
lowered their wasteload to about 0.29 PE/gal, or a 61 percent BOD reduction.


WASTE TREATMENT AND DISPOSAL

     The pharmaceutical industry employs a wide variety of waste treat-
ment and disposal  methods, and therefore this Section of the report
contains extensive information on the unique and different approaches
taken by various companies.   This Section has been divided into Activated
Sludge, Trickling Filtration, and Other Treatment Methods.  The latter
includes data on anaerobic filters, spray irrigation, oxidation ponds,
sludge stabilization, and deep well injection.

-------
     The literature gives some perspective on the merits  of handling
strong and sometimes toxic pharmaceutical  wastes by activated sludge vs
trickling filters.  Trickling filter installations appear somewhat more
prevalent in Great Britain compared to a greater predominance of activated
sludge in the U. S.  Two of the better performing activated sludge
installations are at Abbott Laboratories,  North Chicago,  111. and Hoffman-
LaRoche, Belvidere, N. J.  Extensive technical  information was compiled
for these two diversified pharmaceutical manufacturing plants.  The
available data represents the equivalent of case history  documentation
for both Abbott and Hoffman-LaRoche.

     Some wastes may be resistant to biodegradation by activated sludge,
such as certain nitroaniline forms used in the production of sulfanila-
mides and the phenol-mercury waste (cited earlier in the  report).
Examples of activated sludge waste treatment are illustrated for E. R.
Squibb, Humacao, Puerto Rico, M/s Indian Drugs and Pharmaceuticals,
Hyderabad, India, Wyeth Laboratories at West Chester, Pa., Lederle Labs,
Pearl River, N.Y., and Dorsey Labs at Lincoln, Nebr.

     Wyeth Labs and Lederle Labs, the latter recently incorporating the
Unox pure oxygen aeration process, represent examples of  extensive
pretreatment followed by secondary treatment by the municipal works.
Overall BOD reductions for the Wyeth and Lederle wastes are about 97 to
98 percent.

     Hoffman-LaRoche has a unique sodium sulfate recovery complex which
receives select waste streams primarily from sulfa drug manufacturing.
The recovery process consists of a fluidized bed followed by evaporation
and drying of the salt.  The most interesting features at Abbott's
multi-million dollar treatment facility at North Chicago, 111. are
thought to be pasteurization of plant effluent in place of chlorination
and a spent fermentation beer recovery system integral with an extensive
incinerator ducting system.  Exhaust air, from the drying of the spent
beers, is taken into a specially-designed duct system and carried to the
main plant boilers and incinerated.  Other odorous air streams collected
into the duct system include exhaust from the fermenters, and vents from
the enclosed activated sludge tanks, degassing chambers and sludge
holding tanks.

     The activated sludge process, with adequate waste equalization and
proper sludge disposal, has been shown to operate well at drug formu-
lation and other pharmaceutical plants.  Treatment of pharmaceutical
wastes may become very difficult due to extreme variation in manufacturing
levels, product mix and consequently in waste volume and  strength.
Because of batch processing, a 5-day work week and an 8 to 10 hr work
day at many plants, considerable waste holding and treatment system
recycling may be required especially over weekends.  Waste equalization
of about 2 to 3 days may be essential at certain establishments.  Poor
removal efficiencies may not be the fault of the treatment system but
often can be attributed to lack of control by management  over waste
quantities.  Since manufacturing expansion can occur frequently, and
also due to other factors, treatment plant design criteria are too often
exceeded.   High diligence and control must be continuously maintained,
owing to the special nature of pharmaceutical wastes.

-------
10
     Trickling filter installations and their performance in treating
pharmaceutical waste are described for the Upjohn Company, Kalamazoo,
Mich., American Cyanamid at Willow Island, W. Va., Lederle Laboratories
at Pearl River, N. Y., Eli Lilly and Co.  at Greenfield,  Ind., and the
Merck, Sharpe and Dohme plant at West Point, Pa.   Furthermore, the
Appendix contains trickling filter data for the Merck and Co., Inc.,
Elkton, Va. pharmaceutical plant.

     The Upjohn Company, Kalamazoo, Mich, in the 1950's  successfully
handled spent fermentation beers by two-stage trickling  filtration.
Pharmaceutical wastes amounting to 2,270 m /day (600,000 gpd) and con-
taining-about 1,400 kg (3,100 Ib) BOD/day demonstrated BOD reductions in
the range of 95 to 98 percent.  Influent and effluent BOD concentrations
averaged 600 mg/1 and 20 mg/1, respectively.

     Wastes from the Merck, Sharpe and Dohme manufacturing plant, West
Point, Pa., in the 1950's were subjected to two-stage high-rate trickling
filtration groviding high treatment performance.   The raw wastes approxi-
mated 378 m  (100,000 gpd) with an average of 330 kg (725 Ib) BOD/day.
Overall BOD and TSS reductions were reported by the Company as greater
than 98 and 91 percent, respectively.  Experiences in Great Britain  have
shown that single- and multiple-stage filters can accept heavy loadings
of mixed fermentation and other spent effluents and provide about 96
percent waste reduction.  Trickling filters were also noted as being
less susceptible,to shock loads compared to other forms  of treatment.

     Anaerobic filters represent a potentially promising method of
treating pharmaceutical wastes.  Large-scale spray irrigation of pharma-
ceutical wastes has been employed in at least two known  instances.  The
Upjohn Co., Kalamazoo, Mich, in the 1950's used spray irrigation as  the
principal means of disposing of heavy organic wastes, largely antibiotic
spent broths.  Including a series of waste connections to the city,  the
Commercial Solvents Corporation plant at Terre Haute, Ind. achieves  an
overall 91 percent BOD reduction.  This is noteworthy since the remain-
ing 9 percent wastes receive no treatment.

     Other available treatment across the industry includes deep well
disposal used by Parke-Davis, Holland, Mich., Abbott Laboratories at
Barceloneta, Puerto Rico and the Upjohn Company, Kalamazoo, Mich.
Incineration of concentrated organic wastes is -found at  a number of
plants including Abbott Laboratories, Barceloneta, Puerto Rico;  Abbott
Laboratories, North Chicago, 111.;  Upjohn Co., Kalamazoo;  Eli Lilly,
Clinton, Ind.;  Eli Lilly, Lafayette, Ind.;  Eli  Lilly,  Mayaguez, Puerto
Rico;  E. R. Squibb, Humacao, Puerto Rico;  Lederle Laboratories, Pearl
River, N. Y.;  and Merck and Co., Inc., Danville, Pa.
DEVELOPMENT OF EFFLUENT LIMITATIONS

     Once the prescribed information was gathered and properly analyzed
by NFIC-Denver, distinct categories of pharmaceutical plants were estab-
lished, and plants achieving high performance levels of treatment were

-------
                                                                      11
identified.  From these plants together with other data, model  systems
were selected for each industrial category to define wasteload  reduction
currently attainable by well designed and operated waste treatment and
control systems.  Effluent limitations were then developed for  important
waste parameters in each category, based on performance levels  of the
model systems.

     The more sophisticated waste treatment and control facilities at
pharmaceutical plants almost always consisted of activated sludge and/or
trickling filtration processes deployed as multistage systems.   These
systems appeared nearly equally applicable to all types of pharma-
ceutical establishments.  With the exception of Abbott Laboratories at
Barceloneta, Puerto Rico, Eli Lilly's Clinton Laboratories, and Commercial
Solvents Corporation, Terre Haute, Ind., the known plants all employ
some variation of the activated sludge process.

     This Section, in addition to the model systems, describes  alterna-
tive treatment and disposal techniques available today throughout the
industry.  Effluent limitations were developed in terms of percentage
reductions in raw wasteloads or final effluent concentrations for param-
eters of significance.  Recommended average daily limits for categories
1, 2 and 3 were set at 95.0 percent BOD removal, 82.0 percent COD removal,
82.5 percent TSS reduction and 75.0 percent ammonia nitrogen reduction.
For categories 4 and 5, BOD and COD reductions were established as a
minimum of 92.5 and 80.0 percent, respectively.  Limits for metals and
trace ions in pharmaceutical effluents have been suggested.  Also, pH
and fecal coliform limits and toxicity criteria are given for necessary
inclusion in NPDES permits.


APPENDIX:  CASE HISTORIES

     The Appendix contains case histories prepared for nine pharma-
ceutical plants.  As a result of many meetings, a considerable  data base
incorporating much technical information was compiled for each  of the
companies.  The data spectrum was analyzed, NPDES permits were  completed,
and conclusions were drawn therefrom for the overall Pharmaceuticals
report.  Plants described in the Appendix include:


Case           Plant and            Category      Available Treatment
History        Location


  A       Eli Lilly and Co.             3         Chemical destruction,
          Clinton Laboratories                    principally Carver-
          Clinton., Ind.                            Greenfield evaporators
                                                  and John Zink units

  B       Eli Lilly and Co.             3         Multistage activated
          Tippecanoe Laboratories                 sludge, John  Zink unit
          Lafayette, Ind.

-------
12
Case
History
     Plant and
     Location
Category
Available Treatment
  G


  H
          Pfizer, Inc.
          Vigo Plant
          Terre Haute, -Ind,
Commercial Solvents
Corporation
Terre Haute, Ind.

Merck and Co., Inc.
Stonewall Plant
Elkton, Va.

Merck and Co., Inc.
Cherokee Plant
Danville, Pa.

Wyeth Laboratories
Marietta, Pa.

Wyeth Laboratories
Paoli, Pa.

McNeil Laboratories
Fort Washington,  Pa.
   4,5


    5


    5
              Multistage activated
              sludge, aerobic stabili-
              zation plus standard and
              high-rate trickling
              filtration

              Spray irrigation,
              connection to the  city
Multistage activated
sludge and trickling
filtration

Activated sludge plus
roughing trickling
filter

Activated sludge
Activated sludge


Activated sludge

-------
                     III.  INDUSTRY CATEGORIZATION
     The pharmaceutical industry uses a wide variety of processes to
produce an even broader variety of final products.   The characteristics
of wastewaters resulting from the various processes differ substantially,
necessitating different approaches to waste treatment.  Limitations on
the wasteloads discharged by a pharmaceutical manufacturing facility
must take into account these variations in waste characteristics and
treatment.  To achieve a uniform and equitable application of effluent
limitations to substantially different pharmaceutical plants, it is
desirable to divide the industry into categories of plants with similar
operational characteristics.  Several approaches to categorizing the
industry were found in the literature review.
SIC
     A common means of grouping industrial plants is the Standard Indus-
trial Classification (SIC)(28)*.  For the "drug industry," three broad
SIC groups have been established as follows:

Biological Products (SIC-2831) -- Establishments primarily engaged in
    the production of bacterial and virus vaccine, toxoids serums,
    plasmas, and blood derivatives for human or veterinary use.

Medicinal Chemicals and Botanical Products (SIC-2833)  Establishments
    primarily engaged in manufacturing bulk inorganic and organic
    medicinal chemicals and their derivatives and processing of
    bulk botanical drugs and herbs.

Pharmaceutical Preparations (SIC-2834)  Establishments primarily
    engaged in fabricating or processing drugs into pharmaceutical
    preparations, mostly in finished form for human and veterinary
    use.  Products include ampules, tablets, capsules, vials, ointments,
    medicinal powders,  solutions and suspensions.
   Bibliography entries are referred to throughout the text parenthetically;
   see p 153 for listing.

-------
14
     These SIC groupings are primarily based on final products.  There
are substantial differences in processes and waste characteristics among
these three groups, as discussed in later sections of this report.  From
the process and waste characteristic viewpoint, however, there are also
major variations within the Medicinal Chemicals and Botanicals group.  A
categorization different from the SIC groups is thus needed.
KLINE GUIDE

     The Kline Guide to the Chemical Industry-1974 (46) includes as
"bulk Pharmaceuticals" the medicinal chemicals, biologicals, and botan-
icals having therapeutic value for humans or animals in the forms of
pills, capsules, syrups, injectables, and ointments.  Bulk pharmaceuti-
cals are manufactured by a wide variety of processes, including chemical
synthesis, fermentation, extraction, and other complex methods.  The
list on the following page presents Pharmaceuticals according to the
Guide.

     The Kline Guide groups are based entirely on products and do not
differentiate by process or waste characteristics.  Thus the Guide does 
not present a suitable categorization scheme.
GSR I

     A report on the State-of-the-Art on Pollution Control in the
Pharmaceutical Industry (3) was prepared in 1973-74 by Gulf South
Research Institute (GSRI) for the EPA and the Pharmaceutical Manufac-
turers Association (PMA).  Based on that report and findings of GSRI,
pharmaceutical plants were thus categorized:


          Pharmaceutical

          Chemical and Pharmaceutical/Chemical

          All Others - combining two or more of the following groups:

                       pharmaceutical compounding and formulating,

                       chemical, fermentation, biological, and natural
                       product extraction operations

These categories are not suitably definitive for the application of
effluent limitations.

-------
                                                                      15
                            PHARMACEUTICALS

                           according to the

               KLINE GUIDE TO THE CHEMICAL INDUSTRY (46)



Medicinals                          Antihistamines

Antibiotics                         Dermatological agents and local
     Penicillins                      anesthetics
     Tetracyclines                       Salicylic acid

Vitamins                            Expectorants and mucolytic agents
     B-complex
     E                              Renal-acting and edema-reducing  agents
     C
     A
                                    Biologicals
Anti-infective agents
     Antiprotozoan agents           Serums
     Anthelmintics                  Vaccines
     Sulfonamides                   Toxoids
     Urinary antiseptics            Antigens

Central depressants and stimulants  Botanicals
     Analgesics and antipyretics
     Barbiturates                   Morphine
                                    Reserpine
Gastro-intestinal agents and        Quinine
  therapeutic nutrients             Curare
     Choline chloride               Some alkaloids
     Amino acids and salts-b        Some codeine, caffeine
                                    Senna, aloe, podophullum, terpin
Hormones and substitutes

Autonomic drugs
     Sympathomimetic

-------
16
NFIC-D

     Based on plant visits, discussion with industry representatives,
and the literature review summarized in this report, NFIC-D further
divided the SIC groups.  Although the pharmaceutical industry can un-
doubtedly be classified in other ways, the NFIC-D categorization was
useful in developing and applying effluent limitations.

     Most bulk pharmaceutical manufacturing plants employ two major
types of processes:  fermentation and/or synthesis of organic chemicals.
The characteristics of wastewaters from these two processes may be
substantially different.   Most plants are actually a combination of
these two types, yielding the third category.  The fourth and fifth
categories are self-explanatory.  The NFIC-D pharmaceutical industry
classification follows:
 r(l)  Fermentation Plants -- Plants primarily employing fermentation
          processes to produce medicinals.


 (2)  Synthesized Organic Chemicals Plants  Plants primarily
          engaged in the synthesis of organic medicinal 'Chemicals
          (fine chemicals).


 (3)  Fermentation/Synthesized Organic Chemicals Plants -- Plants
          employing both fermentation and synthesized organic chemi-
          cals processes.  Most moderate- to large-sized plants are
          in this category.


 (4)  Biologicals Production Plants -- Plants primarily engaged in the
          production of vaccines and antitoxins.


 (5)  Drug Mixing, Formulation and Preparation Plants  Plants primarily
          engaged in the production of pharmaceutical preparations in
          final form, such as tablets, capsules, and solutions.
*  Parenthetic numbers reflecting industrial categories identify
   specific plants listed in Section I

-------
                                                                      17


     A wide variety of products is made by pharmaceutical companies,
and many of the products overlap into other fields.  Such products
include industrial chemicals, paints, plastics, fertilizers, pesticides,
cosmetics, animal feed supplements, and ingredients used in candies,
baked goods, cleaning agents, metal etching compounds, car radiator
compounds and blueprints, just to mention a few.  Product lists and
technical description of many pharmaceutical ingredients are contained
in these references:  Pharmaceuticals (104), Drugs in Current Use and
New Drugs (85), Chemistry in the Economy (116), The Merck Index (71).

-------
                       IV.   PROCESS DESCRIPTION
     The unit processes used to produce Pharmaceuticals are best cate-
gorized as fermentation, synthesized organic chemicals, biologicals,  and
drug formulation and preparation.   However, this categorization is
somewhat artificial since more than one series of processes is used in
virtually all plants; for instance, fermentation is often combined  with
the production of synthesized organic chemicals.  Unit processes are
described below, followed by examples of typical plant operations which
are considered both by category and products.
                            UNIT PROCESSES
Fermentation

     The unit operations of fermentation are:  seed production, fermen-
tation (growth), chemical adjustment of broth, evaporation, filtration,
and drying.
Seed Production

     The seed culture is grown usually on a small scale and for a short
     time.  Scale of seed production equipment approximates 5 to 10
     percent of the size of the equipment used in the fermentation step.
     The seed can be used immediately or stored for use as required.
Fermentation

     The seed is combined with an appropriate medium and allowed to
     incubate in agitated, constant-temperature fermentation vessels.
     The type of medium varies with the material to be produced, but
     typical mediums may contain one or more of glucose, soy meal, corn
     steep liquor, lactose, lime, and mineral salts.
Chemical Adjustment (Optional)

     This step is required only if the final fermented material cannot
     be properly segregated from the medium at existing chemical con-
     ditions.  Adjustment may occur in pH, conversion to a salt, addi-
     tion of a solvent, or any number of these or other refinement
     techniques.
Evaporation (Optional)

     In some cases, the fermentation product may be evaporated to
     concentrate the material.

-------
20


Filtration
     A variety of filters are used to extract the desir.ed product,
     including vacuum filters, filter presses and pressure filters.
Drying
     Drying is used when the product is to be marketed as powders,
     tablets or capsules.
Synthesized Organic Chemicals

     Because of the wide variety of chemicals produced,  the unit pro-
cesses involved in their synthesis vary.   Generally,  however,  synthesis
includes chemical reaction in vessels, solvent extraction (solid-liquid
and/or liquid-liquid), crystallization, filtration,  and  drying.   Any
number of the steps can be used, combined in a wide  variety of per-
mutations.  In some cases, the exact process is classified as  company
confidential.
Biologicals

     To produce biological  Pharmaceuticals, specially prepared viruses
and cultures of bacteria are injected into biological organisms such as
eggs, horses, or other appropriate animals.  The serum of interest is
generally produced in a specific organ -- pancreas,  blood, liver,  etc.
That organ is processed^ and the raw serum is extracted and purified by
evaporation, crystallization, filtration, and/or other purification
steps.  Increasingly, animals are used only for testing and organisms
are being replaced by artificial mediums, such as beef broth.   (Biological
unit processes are not discussed in this report; however, for  more
detail see the Appendix, "Case History 6, Wyeth Laboratories,  Marietta,
Pa.," a biologicals production and drug formulation  plant.)
Drug Formulation

     Drug formulation processes use standard steps of mixing (liquid or
solid), pelletizing, encapsulating, and packaging.  This processing
is generally free of excessive wastes, which are therefore considered
here only briefly.
                       TYPICAL PLANT OPERATIONS
     Typical  plant operations are discussed both by the dominant unit
process employed and by the products produced.

-------
                                                                      21


FERMENTATION

     In 1968, Lines discussed his experiences on fermentation wastes
generated in Great Britain (10).   He focused on the manufacture of
antibiotics, such as penicillin and the tetracyclines,  and vitamins,
such as riboflavin (B2) and cyanocobalamin (B12).

     Appropriate organisms are cultured in relatively complex nutrient
solutions, stirred mechanically,  and generally vigorously aerated.  The
desired crude product is removed from solution or from the cells of the
organisms by solvent extraction,  ion exchange, or absorption.  A large
proportion of the starting raw materials and cell material created
during fermentation is disposed of as waste, or is partially recovered.

     The particular strains of organisms are usually obtained when
individual companies increase the productivity of previous organism
strains by mutation and species selection.  Otherwise,  these strains may
be acquired under license from other producers.  Formulas for suitable
fermentation mediums are likewise developed by the company or are
purchased.  Consequently, detailed information on fermentation processes
and pharmaceutical plant manufacturing waste effluents is difficult to
obtain.

     Some data on starting fermentation solutions for antibiotics have
been provided by Lines (10) as follows:
     Benzyl penicillin  (g/1)                Bacitracin-        (g/1)







a/
Lactose
Glucose
Corn Steep Liquor
KH2PO.
CaCty
Vegetable oil

Antibiotic from Baci
35
10
35
4
10
2.5

11 us subtil is
Starch
Peanut meal
Yeast
Calcium acetate
K9HPO.
MgSO.Vo
NaCr  *

10
45
3
0.5
1
0.2
0.01

Other materials typically used in fermentation mediums are sugars,
distillers' solubles, fish or whale solubles, soy bean meal, fish meal,
molasses and trace minerals.  Sugars, alkalis and organic acids may be
added to the fermentation process to maintain levels of essential
nutrients and desirable pH.  Product yields, at least in the mid-19601s,
were seldom more than 10 mg/1.


Vitamins

     Four of the more important vitamins are manufactured by fermentation
processes, including vitamin B2 (riboflavin); vitamin B12 - from selected

-------
22


molds in fermentation vats; vitamin C (ascorbic acid) derived from
sorbose - produced by a bacteria feeding on sorbitol, the latter prepared
from corn sugar; and vitamin D - manufactured by ultraviolet irradiation
of ergosterol (ergosterol represents a substance obtained from yeasts).
Vitamin' B0 (Riboflavin)
 " ' " "      &
     Riboflavin was discovered about 1933 and its chemical  formula is
C17H206IV  Riboflavin generally functions in the form of specific
enzymes bound to certain proteins; a wide variety of human and animal
body reactions depend on these protein systems.   Riboflavin deficiency
causes formation of oral, dermal and corneal lesions.  Although riboflavin
is widely distributed in plant and animal tissues, greater amounts are
present in certain foods such as liver,  milk, cheese, lean meats and
leafy vegetables.  Riboflavin is one of the many ingredients added to
"enriched" flour, bread and various cereal products.  It is added as a
usual supplement to animal feeds to the extent of 2 to 8 g/ton feed;
most goes into poultry and swine rations.

     Manufactured both by chemical synthesis and microbiological fer-
mentation, one-sixth of riboflavin in the U.S. as of the late 1960's was
produced by fermentation.  Most of the latter riboflavin was used in
crude concentrates as animal feed supplements.

     In fermentation, three groups of micro-organisms have been found to
efficiently synthesize riboflavin:  1) bacteria of the butanol-acetone
group, chiefly represented by Clostridium acetobutylicum;  2) selected
Candida yeasts;  and 3) two yeastlike fungi -- Eremothecium ashbyii and
Ashbya gossypii.

     Using C. acetobutylicum, various carbohydrate-containing mashes
such as cereals, corn, rice, whey and semisynthetic starch materials
have been used as appropriate media.  Iron-sequestering agents are often
added to these particular broths.  The mash is generally incubated at 37
to 40C (99 to 104F) for 2 to 3 days.  During fermentation, mixed ethyl
alcohol, acetone and butanol vapors are commonly collected, condensed
and fractionally distilled.  The riboflavin may be recovered by absorp-
tion and elution, extraction with butanol followed by precipitation from
the extract by petroleum ether, or addition of a reducing agent causing
precipitation of the riboflavin.

     When E.  ashbyii is used in fermentation, the medium may include
proteinaceous materials, carbohydrates and vegetable oil lipids.  Fermen-
tation is usually conducted for 50 to 90 hr, then the final beer is
heated to free the riboflavin from the mycelium.  This procedure has
provided riboflavin yields of 500 yg/ml broth, although some reports
indicate returns threefold greater (94).

     At Dawes Laboratories in Newaygo, Mich. (92), the medium for riboflavin
was dextrose, corn steep water and animal stick liquor.  The organism

-------
                                                                      23
used was Ashbya gossypii, a yeastlike fungus.   Described as  a  whole-
broth process, the entire spent fermentation mash was dried, eventually
yielding a final feed supplement product rich in vitamin B,,.

Vitamin B^ (Cyanooobalamin)

     Hester and Ward (92) in their 1954 report describe vitamin B,2,
obtained almost exclusively from microbial sources either as a primary
fermentation product or concurrently with certain antibiotics.  The
chemical formula for vitamin B,2 is C63H9ni4I!li4PCo-  B.l2 is used in
the treatment of pernicious anemia ana nutritional deficiencies in
humans, but the market for it is also extremely important in animal feed
preparations.  The amount of B,,, necessary in a ton of feed is only a
few milligrams.  Furthermore feed-grade vitamin does not require the
complicated purification processing as for human use.  Merck in 1948
found that B,2 could be produced by certain micro-organisms, notably
those used in commercial production of streptomycin.  Streptomyces
olivaceus is one of the most promising of these organisms.  Lederle
Laboratories produces vitamin B,p simultaneously with Aureomycin (chloro-
tetracycline), using Streptomyces aureofaciens in fermentation.

     Dawes Laboratories in Newaygo, Mich. (92), started producing
vitamin B,p and vitamin B~ in 1950.  The organism used for  vitamin B-^
production was S. olivaceus, with a typical medium composition of
distillers' solubles, dextrose, calcium carbonate and CoClp-6 HpO.  The
seed fermenters were 1.5 m  (400 gal) vessels and the production fermen-
ters were 20 m  (5,200 gal) tanks.  Foaming of the aerated  medium
presented certain problems which were minimized by the addition of oils,
such as corn oil, soybean oil, and lard oil, before cooking, and/or the
addition of sterile oil during fermentation.  The fermentation tempera-
ture was maintained at about 28C (82F).  The cultures were grown in
the seed tanks for about 2 days, and subsequently the production fer-
menters were operated from 3 to 5 days.

     In the early 1950's, the fermentation broths were usually harvested
when containing about 1.5 to 2.5 ug B,2 /ml of broth.  Yields, however,
have no doubt improved greatly since tne 1950's.  After fermentation,
the active ingredient, vitamin B,2, is stabilized by reducing the pH to
about 5.0 with sulfuric acid and adding sodium sulfite.

     Vitamin B,- for animal feeds is a whole-broth concentration pro-
cess.  The Dawes Lab process involved drying the spent fermentation mash
in an evaporator or vacuum pan.  The resulting syrup was then fed into a
double-drum dryer yielding a solid product with only about  5 percent
moisture.  The dried material in bulk was then passed through a hammer
mill, a mixer, and eventually bagged or drummed as vitamin  B,9 feed
supplement.  The finished product assayed at 10 to 30 mg vitamin B
The product also contained about 35 percent protein and appreciable
quantities of niacin, pantothenic acid, pyridoxin, riboflavin and
thiamin.

-------
24
     Dawes Labs and others in the 1950's were seeking a micro-organism
     _   _     -  ,    ..    .      IN      I    I     I ^  __ l_ ___!_ J _  _^	
      e of producir
      ncorporation
                                           t \f ^^wrxiii** M 1111^1 x* ^  ^ * i  i
capable of producing both vitamin B,2 and a desirable antibiotic for
dual incorporation into their animal  feed supplements (92).
Vitamin C (Ascorbic Acid)

     Vitamin C can be produced at a reasonable price by modification of
the natural sugar, glucose, which provides a cheap and abundant starting
material.  The commercial synthesis is an interesting example of a
combination of chemical manipulation together with the specific abilities
of micro-organisms.

     Glucose is reduced to the chemical sorbitol.   Fermentation with
Acetobacter suboxydans produces 1-sorbose.  Further chemical  reactions
involving acid and alkalai additions produce ascorbic acid, which is
recovered for use (116).
Citric Acid

     In 1923, full-scale production of citric acid was initiated in the
IL S. utilizing fermentation via the fungi, Aspergillus niger.  Deep-
tank fermentation with A. niger was introduced in 1952.  As of about
1961, U.S. production of citric acid was 27 to 36 kkg (60 to 80 million
lb) annually, practically all derived from fermentation.  Citric acid
has extremely widespread use.  It is also the most extensively employed
organic acid in the food industry (95).

     In the early surface culture method of fermenting citric acid by A.
niger from sugar solutions, beet sugar molasses derived from straight-
house sugar factories constituted the best source of carbohydrate.
After 6 to 12 days, the fermentation was terminated and the mycelium was
washed and' pressed to remove residual adhering citric acid.
      V

     The deep-tank or submerged fermentation process for citric acid was
conducted at low pH's and ammonia could be added as a nutrient during
fermentation.  After 5 to 14 days, the mycelium was filtered off and the
desirable citric acid was recovered.

     Citric acid is recovered by either precipitation as the calcium
salt, or by crystallization upon concentration of the filtrate.  With
the first method, the filtrate is heated to about 60C (140F) and
calcium hydroxide is added.  The precipitated calcium citrate is filtered
out and washed thoroughly.   The calcium citrate is acidulated with
sulfuric acid giving filterable calcium sulfate.  In the second method,
dilute citric acid is purified by decolorization and demineralization,
then subjected to evaporation.  The mixture is centrifuged and the
crystals are washed and dried.

     When conditions have permitted, citric acid has been produced from
lemons, generally the culls and low-quality fruit.  Juices from shredding

-------
                                                                      25
and pressing the fruit pulp are combined with washwaters and subsequently
fermented by yeasts.  The fermented juice with 3 to 4 percent citric
acid is filtered and limed, giving calcium citrate.  After it is filtered
and washed, this product may be marketed as crude calcium citrate, or
the citric acid may be recovered and purified by the steps described
above  (95).
Antibiotics

Penicillin

     The Penicillium mold was nurtured by Fleming in London in 1928-29.
A very small quantity of penicillin was extracted which was found extremely
effective against Staphylococcus infections.  However, mass production
facilities were not available then.  Florey and Chain of Oxford University
directed new efforts toward working with Fleming's mold.  Later they
solicited U.S. pharmaceutical experience from Pfizer, Merck and Squibb
toward full-scale production of penicillin vitally needed during World
War II.  Eventually PeniciIlium chrysogenum was developed which, under
proper culture conditions, gave 200 times more penicillin than Fleming's
Penicillium notatum.  Full-scale production was achieved with fermentation
procedures using corn steep liquors as the basic medium, and ample
supplies of penicillin were finally available in the early 1940's (112).

     In 1958, Ross (8) provided detailed criteria on antibiotic fermen-
tation operations.  In some of the earliest penicillin processes, yields
as low as 30 to 60 lU/ml of broth were obtained.  With submerged fermentation
techniques using large fermenters, yields were eventually increased to
5,000 lU/ml and higher.  Significant advances also occurred in developing
mutant organism strains which served to decrease fermentation time and
increase antibiotic yields.

     The respective sizes of full-scale fermenters utilized in the late
1950's were generally 38 to 76 m  (10,000 to 20,000 gal), but some units
were 189 m  (50,000 gal) or greater.  A height to diameter ratio of 2 to
3:1 is generally employed.  The operating volume is usually about
three-quarters of the total volume available to allow for foaming.  Seed
fermenters were used in the size range of 1,100 to 1,900 1 (300 to
500 gal).  Fermentation periods were about 90 hr.  The type of nutrient
broth is usually corn steep liquor, a byproduct of the cornstarch
industry and previously a discard material.  Peanut oil or equivalent is
added to the fermenters to break the foam or at least keep it under
control.  Mechanical agitation is necessary and power rating on the
agitators is generally from 0.8 to 1.0 hp/100 gal of fermenter capacity.

     The principal difficulties in operating the fermentation process
are maintaining a delicate balance between the desired growth micro-
organisms and other competing forms of growth such as wild yeasts and
bacteria.  The latter forms may be present in the' nutrient materials,
water,  air, or on the surfaces of the fermenter, or pipelines.   The
undesirable organisms can affect the fermentation process significantly.

-------
26


In the case of penicillin, many organisms produce penicillinase, an
enzyme which destroys the penicillin.  Contaminating organisms consume
the nutrients in the broth, retarding antibiotic generation.   Unwanted
byproducts may not necessarily be harmful, but they could be  extremely
difficult to remove in the final recovery processes.  Very stringent
sterilization procedures are used in every fermentation step.

     The penicillin fermentation broth leaving the fermenters is filtered
to remove mycelium, and the active compound is recovered from the filtrate
by solvent extraction (8).

     In the early 1950's, Mann (32) and Gallagher et al. (65) reported
on Bristol Laboratories at Syracuse, N.Y., which began penicillin fer-
mentation soon after it was formed in 1943.

     The mold penicillin notatum was grown on the surface of  quiescent
nutrient solution in flasks, bottles or trays.  A standard practice was
to use 1.9 1 (2 qt) bottles and incubate them on their sides  to obtain
the' greatest quantity of culture.  In some plants, as many as 30,000
bottles were inoculated each day.  The surface culture method was re-
placed by the submerged, deep-vat fermentation tank process in the
1940's. J"hese fermentation vessels increased from 4.5 m  (1,200 gal) up
to 114 m  (30,000 gal) or larger in the late 1940's and early 1950's.

     Seed for the fermentation tanks is obtained daily from the laboratory
where a master culture of the mold is kept in test tubes.  From the
master culture, larger quantities of the mold are grown in flasks and
small fermenters, which in turn serve as seed source for the  full-scale
fermentation tanks.  The fermentation vessels may be up to three stories
tall, but in some cases they are horizontally arrayed.

     To the fermenters are added corn steep liquor, lactose,  lime and
mineral salts mixed with water to form a medium which is sterilized with
heat or steam before inoculation.  The penicillin mold is added to the
tanks and the medium is incubated near room temperature under intense
agitation and aeration for about three days.  During the 73 hr incubation
the mold has excreted the desired chemical, which is subsequently
refined and concentrated into the finished penicillin.

     The fermentation broth is removed from the tank and the  growths are
inhibited with formaldehyde.  The mold or mycelium is removed by passing
the entire ferment liquor through vacuum filtration.  The mycelium may
be burned, buried, used as a fertilizer, or dried and mixed with animal
feed.  The filtered broth from fermentation is acidified and  extracted
with amyl acetate or other solvents.  The solvent is separated from the
medium by centrifugation, and the penicillin values are transferred to
an aqueous buffered solution.  The spent broth passes through a solvent
recovery column, is neutralized with caustic, and then is sent to the
plant sewer.

-------
                                                                      27
     Finishing process operations include:  filtration,  extraction with
further solvents, discard of the spent buffer solution,  washing the
filter solids, crystallization and drying of the penicillin crystals.
All operations are conducted under sterile conditions.

     Various tests are made on the product before FDA certification and
release, including assay and pharmacological tests in rabbits and mice,
the pyrogen test, and animal safety, sterility and toxicity tests.

     At Wyeth Laboratories in West Chester, Pa., fermentation broths are
made of corn steep liquor, lactose and mineral salts.  After appropriate
fermentation, the fungi mycelium is separated from the spent broth by
vacuum filtration.  The penicillin is extracted from the broth with a
solvent such as amyl acetate in an acid solution.  After retrieving the
penicillin, the solvent is recovered by stripping for further reuse (4,
36, 50).
     The Terramycin plant of Charles Pfizer, Inc. in Sandwich, Kent,
England was reported to be Europe's largest plant for the production of
antibiotics in the mid-1950's (41, 80).  The antibiotic Terramycin
(trademark for oxytetracycline), is generated via fermentation by
Streptomyces rimosus.  It is used in the treatment of more than 100
diseases including typhus, pneumonia, peritonitis and dysentery, and is
also used as an animal feed supplement.  It was reportedly discovered in
1949 by the laboratories of Charles Pfizer in the U.S. after successive
screening of micro-organisms from thousands of soil samples collected
world-wide.  The Kent facilities were completed in Oct. 1954.  Refining
of the Terramycin is carried out at Pfizer's nearby pharmaceutical plant
in Folkestone, England.
             
     In the sterile area of the Kent plant, the master culture spores of
Streptomyces rimosus are stored freeze-dried until needed.  The first
stage of Streptomyces multiplication occurs in small containers under
strict aseptic conditions.  Large quantities of the mold are needed to
set up the fermentation medium.  While the mold is multiplying, the
medium for the large, full-scale fermenters is prepared.  Initially, the
medium is prepared in 7.6 m  (2,000 gal) tanks to which are added
desirable nutrient salts.  This nutrient medium is then batched with
water and heated.  From the media preparation room, this mass is pumped
to the full-scale fermenters and sterilized with live steam.  The mold
is subsequently added to the large fermenters.

     The main fermentation tanks are under pressure to preclude entry of
airborne contamination.  Large quantities of sterilized air are blown
through the fermentation mixture while the contents of the tank are
continuously agitated.  The ongoing reaction generates considerable
heat, and the fermenters are cooled by continuously circulating cool
water to maintain a predetermined constant temperature in the tanks.
Antifoam agents are added to the process.  After a prescribed number of

-------
28


days, the fermentation process is considered complete, the useable food
in the media having been consumed by the organisms.   During the first
day, the original molds may have increased more than 100-fold,  con-
currently expelling the product from which the antibiotic is later
extracted.

     The broth in the fermenters now contains the highly diluted and
crude Terramycin, the spent molds, and the spent medium.   The total
broth passes from the fermenters to the clarification area .where chemicals
are added to promote coagulation of the mycelium (molds)  and to aid
subsequent filtration.  The mycelium is removed by straining the fermen-
tation broth through a Dorr-Oliver stainless steel rotary vacuum filter.
The filtrate which contains the active Terramycin, is then subjected to
various chemical additions and processing to produce the crude inter-
mediate salt, Terramycin hydrochloride, inside pressure platen filter
presses.  After successive washing of this salt, the Terramycin hydro-
chloride is partially air-dried and milled before further purification.

     Refining of the Terramycin hydrochloride is carried out in glass-
lined reactor vessels.  The intermediate salt is mixed with a solvent
and decolorizing agents, and the solution is pumped through a platen
frame filter.  The insoluble matter is removed and a clear solution  of
salt is passed to a vacuum tank with steam ejectors for concentration of
the dissolved solids.  This concentrated solution passes into a glass-
lined crystallizing vessel where first-stage Terramycin hydrochloride is
precipitated.  Centrifugation separates the hydrochloride from the
mother liquors.  The resulting hydrochloride is then redissolved,
filtered clear, and recrystallized as the pure salt.  After washing, the
pure Terramycin hydrochloride is dried in vacuum ovens.  The refined
Terramycin at Kent is bright yellow. A batch-fractionating unit is
available for solvent recovery.

     Some of the final refining processes for Terramycin hydrochloride
are carried out at the nearby Pfizer Folkestone plant.  The products
received there are incorporated into the desired forms for subsequent
sale such as tablets, oral and nasal suspensions, intravenous and
intramuscular solutions.  Besides Terramycin, the Kent plant manufactures
other antibiotics (41, 80).

     Reeves (64) described the Pfizer program for 1947-50, from the
discovery of Terramycin to the subsequent development'of process opera-
tions that led to full-scale manufacturing.  The first experimental
process in 1950 extracted Terramycin from filtered broth using n-
butanol, but it was difficult to properly distribute Terramycin between
butanol and the fermentation broth.  To enhance this distribution,
ammonium salts were used.  Ensuing development improved and enlarged
these extraction and refining processes for Terramycin.

     Capacity of the Pfizer fermenters in the 1950's varied from 8 m
(2,000 gal) for the pilot plant tanks in Brooklyn, N.Y. to 95 m  (25,000
gal) for some of the production fermenters at the Groton, Conn, plant.

-------
                                                                      29
The air supply for the fermenters passes through special filters for
sterilization before being introduced through distributors at the bottom
of the fermentation tanks.

     When the fermenter has reached maximum activity after several days
of incubation, the fermentation beer is removed and passed through
continuous rotary precoat vacuum filters.  The screened mycelium is sent
through a drum dryer and the residues are recovered for use as animal
feed supplements. The filtered beers are delivered to a treatment tank
to which quarternary ammonium salts are also added.  The solution is
then passed through a second rotary vacuum filter.  The filtrate con-
taining the spent fermentation beers passes through evaporators and a
spray drier for recovery of large amounts of material useable as animal
feed supplements.  The filter cake containing the Terramycin "Q" salt is
recovered, dried by warm air in cabinet-type driers, and transferred to
the refining operations.

     Dried Terramycin Q salt is treated within glass-lined stirred
reactors together with methanolic hydrochloric acid.  The treated
solution is routed through a series of filtering steps for removal of
filter-aid and other insoluble products, and a crude grade of Terramycin
hydrochloride is obtained.  The final process steps designed to yield
purified Terramycin hydrochloride involve solvent recovery, vacuum
distillation, recrystallization, washing of the crystals, centrifuging,
and a number of unknown steps.  The bulk purified drug receives final
drying in vacuum trays or equivalent equipment.  The 8 m  (2,000 gal)
fermenters, seeded with from 57 gm (2 oz) to 19 1 (5 gal) of pure
culture, were reported to yield about 4 kg (9 Ib) of antibiotic in the
early 1950's.  Antibiotic production efficiency has undoubtedly in-
creased tremendously up through the present.  A high degree of solvent
recovery should necessarily be integrated into the Terramycin refining
processes (64).


Streptomycin

     In 1958 Bartels (83) described the transition from laboratory to
full-scale manufacturing of streptomycin, which resulted in development
of the "Whole Broth Process."  Streptomycin is produced by the Strepto-
myces species under favorable conditions of fermentation over a 4- or 5-day
incubation.

     Some of the first broth commercially produced contained less than
50 ppm of streptomycin.  The early process involved acidification of the
broth followed by filtering on pressure precoat filters.  The antibiotic
was absorbed from the clear, neutral  filtrate by a carboxylic-type ion
exchange resin.  The resin was then eluted with dilute acid for recovery
of the streptomycin, and this compound subsequently passed through a
series of refining steps to obtain the pharmaceutical grade product.
Although the yield of crude streptomycin in the broth was increased to
more than 2,500 ppm, the concurrent filtration rates became extremely
low, causing high filtering costs and considerable mechanical  losses.

-------
30
This led to the eventual development of the streptomycin whole broth ion
exchange process.  In this procedure, the mycelium is not filtered out
before entering the ion exchange columns.  Comparison of the whole broth
process vs the original process is demonstrated as follows:
Parameter           Whole Broth Process      Mycelium Removal Process


                                        unit/g resin
Activity of
  whole broth            320,000                       320,000

Loss during
  pressure filtration                                45,000

Activity of feed
  to resin column        320,000                       275,000

Column loading (avg)     279,000                       233,000
Percent column recovery,
  column eluate/feed        84.7                          84.7
Percent overall recovery,
  column eluate/whole broth 84.7                          72.9
     Ross (8) described fermentation of streptomycin in the late 1950's.
The respective fermenter designs and operating conditions were the same
as those previously described for penicillin.  The type of nutrient
broth used was a mixture of soya meal and glucose in water, to which
some metallic salts were added.

     After fermentation, the streptomycin broth was acidified and re-
quired considerable filter aid for filtration.  Streptomycin was re-
covered from the filtrate by adsorption onto charcoal or onto a cation
exchange resin.  Acidified alcohol was used as the eluate for charcoal
and dilute acid for the resin.  The eluate is further purified by passing
a solution of streptomycin hydrochloride in 80 percent methanol through
alumina or activated carbon (8).

     Mudri and Phadke (7) described streptomycin production in 1968.  A
pure culture of Streptomyces griseus is subjected to aerobic fermentation
within'a medium of glucose and a nitrogen source (such as corn steep
liquor).  Following fermentation, the broth is filtered to remove the
insoluble solids.  The filtrate is absorbed on a resin or charcoal,

-------
                                                                      31
and the active streptomycin is eluted by dilute acids  from the  charcoal
or resin.  The eluate is neutralized and concentrated  under reduced
pressure.  The crude streptomycin product solution is  then precipitated
out by the addition of acetone and subsequently purified.
Enzymes

     Enzymes, fundamental catalysts triggering specific chemical  reactions
without becoming a part of them, constitute one of the most unusual
classes of substances produced by fermentation.  Some of the more
familiar of the enzymes are pepsin, ptyalin, and rennin which are
digestive tract enzymes.

     Enzymes are manufactured as commercial products for use in the
brewing, meat packing, baking, cleaning, cheese making and the leather
tanning industries.  Examples are "Sure-Curd," a bacterial enzyme akin
to rennin from animal stomachs, which is derived by fermentation for use
in the cheese making industry; the bacterial enzyme, invertase, used in
making candy with liquid centers; and amylase, an enzyme derived from
molds and bacteria used for the conversion of starch into sugars.  A
number of bacterial enzymes are used in medicine to dissolve blood clots
and destruct unwanted chemicals within the human body, particularly in
sensitive individuals (112).
SYNTHESIZED ORGANIC CHEMICALS
Vitamins

Vitamin B^ (Riboflavin)

     Chemical synthesis of riboflavin (94) generally involves modi-
fications and refinements of early Kuhn and Karrer reactions.  Kuhn's
procedure involved reaction of 6-nitro-3, 4-xylidine with d-ribose
through a series of complex steps yielding riboflavin.  The Karrer
procedure increased the d-ribose-based yield and gave N-d-ribityl-3,
4-xylidine, an important intermediate for making riboflavin.   Overall
reactions in producing the intermediate include condensation, reduction,
epimerism, and acetylation.

     The intermediate can be condensed with violuric acid to give ribo-
flavin, directly transformed through an azo-dye intermediate, then con-
densed or reduced to phenylenediamine, then chemically treated.  All
three procedures satisfactorily yield riboflavin.

Vitamin C (Ascorbic Acid)
     In manufacturing vitamin C, the natural  sugar glucose (CKH190,.) is
reduced to sorbitol (CgH^Og), then is converted to sorbose (CfiRfJ)fi)
by the synthesizing capability of micro-organisms such as AcetQbactOr

-------
32
suboxydons.  Further chemical modifications give the intermediate
CfiH]Q07, which when subjected to alkali and/or acid treatment is even-
tual Yy converted to ascorbic acid.
Antibiotics

Chloromycetin (Chloramphenicol)

     Chloromycetin (trademark for Chloramphenicol), discovered in 1947,
was one of the first of the so-called broad-spectrum medicinal antibiotics.
The drug is efficacious against both gram-positive and gram-negative
organisms.  The compound is reported effective in combatting typhoid,
paratyphoid, H. influenza, bacterial meningitides, and staphylococcal
infections (96).

     As of May 1952, the Parke-Davis and Co. plant in Hounslow, Great
Britain was reported to be the only producer of an antibiotic by synthesis,
Eleven reactions, subdivided into four main stages, constituted the
basic operations (45).

     The first three steps of the first stage, the initial condensation
processes, are performed in glass-lined reaction vessels.  Wide variation
in solids content, viscosity and other characteristics create difficult
agitation problems for this part of the operations.  Three-point sus-
pension stainless steel centrifuges are used for filtrations in this
stage.

     The second stage of the process uses stainless steel units and
glass-lined refluxing units.  Glass-lined steel vessels are used in acid
hydrolysis.  In the third stage of processing, stainless steel refluxing
units and stainless steel reaction vessels are used.

     Before reaching the final product, there are several drying opera-
tions.  While a battery of drying ovens fulfills major drying needs,
vacuum shelf driers are also available.

     Since large quantities of solvents are necessary in the processes,
extensive solvent recovery is on hand.  Stills and fractionating columns
are enclosed in the Hounslow plant.  Due to the uncommon amounts of
solvents and other flammable materials used inside the factory, fire
design and precautions have received a great deal of attention.

     The building has a comprehensive fume extraction system.  All
vessels are served by high velocity extraction ducts whereas the vacuum
oven, shelf drier trays, centrifuge balance tanks and associated areas
are served by fume hoods.  The air extraction system has its own ex-
traction fans and air washing equipment (45).

     Chloromycetin was previously prepared from deep-tank submerged
fermentation using various strains of the actinomycete, Streptomyces
venezuelae.  The fermented broth was extracted with amyl acetate, and
the extract was concentrated, washed, and again concentrated.  Crude

-------
                                                                      33
crystals were filtered, dried and redissolved  in  hot water.   The  hot
solution was decolorized, clarified and cooled to induce  recrystallization.
The purified crystals of chloramphenicol  were  filtered, washed, dried,
pulverized, sieved and then packaged.   This biochemical process scheme
has now been largely replaced by organic synthesis (96).

     The most commonly-used synthesis  process  starts with p-nitroaceto-
phenone.  The consecutive steps involve bromination in a  variety  of
inert solvents, addition of a chlorinated solvent followed by alcoholic
mineral acid hydrolysis, acetylation with acetic  anhydride and alkali,
the addition of formaldehyde in the presence of sodium bicarbonate,
chemical reduction, acetylation, nitration with fuming nitric acid, acid
hydrolysis, recovery of the d-base by precipitation with  alkali,  and
finally acetylation of the dried base to give a good yield chloram-
phenicol (96).

     Adinoff in 1953 (59) and Melcher in 1962 (13) provided data  on
process description and methods of waste treatment in the manufacture of
synthetic Chloromycetin at the Parke-Davis plant in Holland, Mich.

     The synthesis of Chloromycetin introduced at Holland, Mich,  in the
early to mid-1950's, was divided into eight major steps involving some
40 independent manufacturing procedures.  The plant is at the mouth of
the Black River and the head of Lake Macatawa.  Treated waste discharges
enter almost directly into the Lake which is used for swimming, boating,
fishing, and service water for various other companies on the lake.


Sulfa Drugs

     The sulfa drugs, or Sulfonamides, represent a very large group of
compounds; several thousand derivatives have been synthesized.  Sulfa
drugs were the first drugs to control  systemic bacterial  disease, and
today still have important application.  Early studies in sulfa drugs
led to the discovery of Prontosil, the first drug to cure bacterial
septicemias (blood poisoning).  Prontosil was found to quickly dis-
sipate, and sulfanilamide was formed in the body.  Activity of the drug
was consequently attributed to sulfanilamide.   Thereupon, extensive
testing of thousands of sulfanilamide derivatives was conducted to find
active compounds with a broader spectrum of action (98).

     Sulfa drugs may be strictly defined as sulfonamides, derived from
sulfanilamide.  The chemical formula for sulfanilamide is H?NCfiH.SO?NH?.
Sulfa drugs are antibacterial agents,  but their action is bactenostatic
(growth of bacteria inhibited without destruction) rather than bactericidal
(destruction of bacteria).  Sulfa  drugs are still considered prominent,
even though they have been largely displaced by antibiotics.  These
drugs continue to provide a principal  treatment of urinary-tract  disease
for which they are claimed to be less  expensive and possibly superior by
some physicians.  Other specific applications  include the treatment of
fungus-related nocardiosis, rheumatic  fever and ulcerative colitis.
Although antibiotics have largely replaced sulfa  drugs for treating

-------
 34


 human  disease,  this has been less true in animal therapy because of the
 relativelyvlow  cost of the sulfas.  Sulfaquinoxaline and Sulfamethazine
 are  important drugs used to combat coccidiosis in turkeys and chickens.
 Other  prominent and typical sulfa drugs are sulfathiazole, sulfadiazine
 and  sulfaguanidine.

      In the manufacture of sulfa drugs, acetylsulfanilyl chloride  is
 usually reacted'with the appropriate amine.  Extra amine or a base is
 generally available to neutralize HC1 made free within the reaction.
 The  resulting acetyl product is then usually hydrolyzed with alkali to
 give the desired sulfanilamide derivative.  Many variations are possible
 in these reactions, some of which may involve acetylation, diazotization,
 or amination.  The sulfa drugs reached a maximum production of 4,536 kkg
 (10  million Ib) annually in 1943 which dropped to less than half this
 amount in 1944  when antibiotics were commercially introduced.  Since
 that time, the  level of the sulfas has remained fairly constant.   U. S.
 production of sulfa drugs and antibiotics is shown below (98).
Year            1942        1943        1946        1952        1956        1966
   Drug	
                       (left col. = kkg)              (right col. = Ib x 1,000)

Total  Sulfa
  Drugs       2,465  5.435 4.539  10.006  2,315 5.104  2.625 5.786  1,731  3,817  2,472  5,450

Sulfathiazole   723  1.594           915 2.016   328   724  .  0  . 0    0   -v.  0
Total Anti-
biotics
Penicillins

0
0

0
0

0
0

0
0

17
16

38
35

675
304

1,487
671

892
286

1,967
631

4,378 9.652
949 2.092
 Steroids

      Steroids manufacturing as early as 1959 on a dollar  volume  basis
 was  in second place among ethical Pharmaceuticals, holding  24  percent  of
 the  $1.9 billion Pharmaceuticals market, and outranked only by antibiotics
 (73).   In 1958-1959, the anti-inflammatory steroids, mainly cortisone
 and  related compounds, comprised 75 to 80 percent of the  steroid market.
 The  remainder of sales come from sex hormones used in treatment  of
 endocrine disorders (i.e., glandular and other associated body organs).
 In the 1960's, the anabolic agents (those promoting tissue  building
 functions and particularly important to the growing population of geriatric
 patients) were expected to lead the market gain for steroids.  A very
 large market for steroid hormones capable of acting as contraceptive
 agents was also visualized at that time (73).

      The major classes of steroids include the sterols, the bile acids,
 the  adrenal cortical hormones, the sex hormones, various  contraceptive drugs,

-------
                                                                      35
insect molting hormones, cardiac-active lactones,  sapogenins,  certain
alkaloids, and certain antibiotics.  The basic steroid compound has
three benzene groups together with a pentene group,  constituting 17
available carbon positions as shown below:
     The first important steroid, introduced into use in 1949, was cor-
tisone.  The beneficial action of cortisone as an anti-inflammatory
agent was related to its ability in minimizing the inflammatory response
of tissues to infective or toxic attack.  Besides combating rheumatoid
disease, cortisone preparations appeared to have great value in treating
skin disorders, blood diseases, eye infections, and endocrine disorders.

     Cortisone is chemically comprised of the four rings of the cyclopen-
tanophenanthrene nucleus, depicted below.  It has methyl groups at the
C-10 and C-13 positions, a two-carbon side chain, the hydroxyl group at
C-17, an oxygen atom at C-ll, and a ketone at C-3 in conjunction with a
double bond at the 4, 5 position.
                               Cortisone

     The only production process available for cortisone in 1949 in-
volved desoxycholic acid, a constituent of ox bile, as the base material
for generating the steroid.  Some 37 chemical steps were necessary to
produce cortisone, with about 10 steps necessary to shift the oxygen
from position 12 to 11.  Clinical studies have shown there must be an
oxygen in position 11  to have an active compound.

-------
36


     Complexity of the synthesis, relatively short supplies of ox bile,
and low yields all limited the supply of cortisone and kept prices very
high.  Improvements in cortisone and other steroid production were
sought, and a breakthrough was reported by Upjohn in 1952 (73).


Progesterone as cm Intermediate to Cortisone

     Uy.iohn Company.  In 1952, Upjohn perfected introduction of an 11-
hydroxy group into progesterone (female sex hormone) by utilizing Rhizopus
arrhizus, a common mold.  This was a microbiological oxygenation of
steroids that meant a variety of starting materials could now be used to
derive progesterone.  This precursor material not only to cortisone but
also many other active steroids is termed 11-hydroxyprogesterone.

     At Upjohn, soybeans were a major starting material in producing
soybean sterols, including stigmasterol.  Stigmasterol in turn is syn-
thesized to progesterone, which proceeds by microbiological oxygenation
to the chain of various desired steroids.  Soybean oil byproducts are ,
prepared from soybeans, and from these byproducts is extracted a sterol
mixture containing 12 to 25 percent stigmasterol, plus large amounts of
mixed sitosterols.  Unfortunately, since the physical properties of
stigmasterol and sitosterols are nearly identical, the next problem was
how to best separate these two types of sterols (73).  the chemical
configuration for stigmasterol is shown on the following page with the
overall schematic for steroid production from soybeans.

     The older commercial process for separation of stigmasterol and
sitosterols was quite time-consuming, expensive, and has low yields.
Uphohn reported development of highly specialized procedures whereby
stigmasterol could be selectively isolated in high yield and high purity
from mixtures containing sitosterols and extraneous materials.  These
operations comprised multistage countercurrent crystallization using
diverse selected solvents.

     The Upjohn countercurrent crystallization process starts with
soybean-derived feed solids containing about 20 percent stigmasterol.
Through six successive crystallizations of this material, the final
product is built up to 97 percent stigmasterol.  Essentially, steady
state conditions in the Upjohn process make it possible to achieve in
excess of 85 percent of the total theoretically available stigmasterol
in the feed solids (73).  The solvent used in the Upjohn process is an
azeotropic mixture of 63 percent ethylene dichloride and 37 percent n-
heptane.

     Pressure filtration is employed throughout all processing because
of the handling of volatile solvents, the presence of flammable mixtures
of air and solvent vapors, and the need to increase solvent recovery to
the maximum extent practicable.   A very high degree of solvent recovery
is necessarily expected from the overall processing.  The molten sterol

-------
             Stigmasterol
                                                           37
                 REFINED GIL
          RESIDUE
       SOYBEAN STEROLSJ
                      CRYSTALLIZATION
                          PROCESS
        STIGMASTEROL
       ...4.-.
        PROGESTERONE

    II-HYDROXYPROGESTERONE

             I
        CORTISONE
        HYDROCORTISONE
        PREONISONE
        PREDNISOLONE
        METHYLPREDNISOLONE
Steroid production from soybeans

-------
38
residue discard from the process containing about 2.8 percent stigmas-
terol (a black tarry liquid) is drained into large load-lugger truncated
containers.  In about 7 days the melt solidifies into a semi crystal line
mass.  The containers are discharged to outside storage or disposal by
simply inverting the load-luggers.  An alternative to casting the
residues in pigs is to process this material for recovery of B-sitosterol,
although again a discard is formed (73).  Problems of solid waste
disposal and associated Teachability appear to merit critical attention.

     Boots Pure Drug Company.  A close association was maintained in the
1950's between Boots Pure Drug Co. of Great Britain and the Upjohn
Company in the U.S.  From this association and continued research in
Great Britain, Boots constructed a new plant at Beeston near Nottingham
in 1955 for the preparation of cortisone, hydrocortisone, and a-1-
hydrocortisone (40).  The Beeston facilities included fine chemicals
manufacturing plus fermentation, with full packaging capability.

     The base production material was diosgenin, obtained from the wild
vegetable Dioscorea, otherwise known as Testudinoria sylvatica, or
elephant's foot, believed primarily obtained from South Africa.  Essen-
tially, diosgenin is converted to progesterone, an intermediate product,
by a five-stage chemical process.  Fermentation and microbiological
oxygenation convert progesterone into 11-a-hydroxyprogesterone.  This
precursor is converted to the three end products of hydrocortisone,
cortisone, and a-1-hydrocortisone via multistage chemical synthesis;
however a-1-hydrocortisone requires an extra fermentation step for
incorporation of an additional double bond into this compound.  The
latter chemical is reported to have greater -steroidal activity and less
undesirable side-effects (40).

     The production of the three cortisone compounds requires from 10 to
30 steps, depending on the particular starting material and the specific
synthesis.  In contrast to the Upjohn process which employs stigmasterol
from soybeans as the starting material, Boots Pure Drugs uses the plant
extract, diosgenin.  Both processes yield, the intermediate product,
progesterone, a female hormone.  Formation of the important precursor
11-hydroxyprogesterone is conducted by fermentation using the mold
Rhizopus arrhizus, which is apparently identical in the British and
American processes.  Boots Pure Drugs process operations are outlined in
the figure on the following page.
                                                                       3
     Conversion of progesterone to the hydroxy form takes place in 15 m
(4,000 gal) stainless steel fermentation tanks and requires several
days.  The hydroxyprogesterone is precipitated in the fermenters in a
finely dispersed form.  The fermentation beers are removed and passed
through a rotary vacuum filter for separation of the mycelium.

     Extraction of 11-hydroxyprogesterone is necessary both from the
aqueous fermentation broth and from the mycelium.  The broth is ex-
tracted with a suitable solvent in a Podbielniak centrifugal extractor.

-------
                                                                   39
              H,C
                                                             COCH,
                     AcO
             WOSGENIN
                  CO-CH,
                              PSEUDO-DIOSGENIN DIACETATI  PREGNADIENOLONE ACETATE
    ^
ACO'
                                             COCH,                   CO.CB,



                                   ,   Ci}             C\\
                          MM      I   I    J	I M|OPr), __       J     	I
                         1   *-  _x*sw 1^*^^. sS"^     *owiNAUEnr ^S^**. .X'^V^ j/*
                              f    1^   If         MACTIONI J^  ^> ^   ><

                          noxkx^SX           ^^^
       rUCNENOLONE ACETATE
                                   PREGNENOLONE
                                                 HO
KXROeiOLCGICAL
OXTCiNATlON

V.TH RK1ZQHJS ORGANISMS
          HO.
                                                          HYOROCORTISONE
                                                       (17-HYDftOXTCORTICOSTERONE)
                     CO-CHj
                                                            CORTISONE
                                                           (l7ii-HYDROXr.ll
                                                         DEMYOROCORT[COSTWONE>
        I l^-HYDROXYPROCESTERONt
                                                           A-I-HYDROCORTISONE
                                                       (l-DEHYDAOHYDROCOJUISONt)
                  Processing of Corticosteroids

       Boots Pure Drug Co., Ltd., Nottingham,  Great Britain

   The process,  starting with diosgenin, proceeds through the im-
   portant intermediates of progesterone and  11-hydroxyprogesterone
   to yield three final steroid products:  cortisone, hydrocortisone,
   and A-1-hydrocortisone, (40).

-------
40


Hydroxyprogesterone is removed from the mycelium by agitation with
solvent in a fixed vessel.  Tfie desired chemical is carried with the
filtrate resulting from subsequent plate and frame filters.  This fil-
trate is combined with the appropriate.aqueous layer fronrthe 'Podbielniak
extractor and concentrated in a film evaporator equipped with distilla-
tion.  The slurry -from the evaporator is crystallized into-a -white
powder, mainly 11-hydroxyprogesterone.  Many items of standard chemical
equipment are used with heavy reliance on''Pfaudler glass-enamelled
reaction vessels and Pyrex glass conveyance pipelines (40.).
Prostglandins

     In the past couple of years, the Prostglandins have teen introduced
to the medical field.  Chemically, the prostglandin products are a
family of 20-carbon unsaturated fatty acids'.  They are characterized by
a five-membered carbon ring and two long side chains.  One side chain is
seven carbon atoms long and ends with a.-carboxyl group.  The other-chain
is eight carbon atoms long with a hydroxyl group in the C-l5 position.
The synthesized prostglandins are' intended to simulate compounds 'origin-
ally thought to be produced by the prostate gland in the body. The
prostglandins are involved at the cellular level in regulating many
bodily functions including gastric acid secretion, inflammation and
vascular permeability, contraction and relaxation of smooth muscles,
body temperature, food intake, and blood platelet aggregation (115).

     The ability of the pharmaceutical companies to synthesize prost-
glandins has apparently outstripped the ability to understand what these
compounds are, what they do, and their biological effects and implications.
Clinical-trials, animal testing and toxicological studies are-being
conducted in earnest.  Only two prostglandins, have s'o far reached the
market for human use.  The Upjohn Company distributes two'naturally-
occurring prostglandins, called E~ and F2  in Great Britain for
inducing labor and for terminating pregnancy, respectively.  ,In the
United States, p2a is available but only'on a restricted basis.-

     Problems with prostglandins include their rather rapid metabolism
in the body and the lack of tissue specificity.  Whereas-certain -tissues
may be properly impacted, at the same time the drug may adversely
affect other tissues or organs.  One of the more important syntheses for
prostglandins is characterized by developing a y-lactone fused to the 5-
membered carbon ring at the C-8 and C-9 positions.  The two side chains
are added at a later stage in the synthesis.  Natural sources of prost-
glandin have been found, such as the sea whip (a species of Caribbean
coral) in the Cayman Islands, but commercial manufacture of prostglandins
by the major pharmaceutical companies has been based upon total synthesis

-------
                                                                      41
FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS

Filtration

     Filtration techniques are extensively used in process operations
but also may have application in pharmaceutical waste treatment.   The
absolute filter removes ultrafine particulate matter from final  pharma-
ceutical products, such as liquids and intravenous solutions.   Almost
total elimination of particulate matter is feasible today.  Apparently,
contamination of the product due to settling of airborne particles in
manufacturing is not a major problem compared with liquidborne con-
tamination, as such may enter with deionized water or other sources.


Absolute vs Nominal Filters

     Of the various filter types, only the wire mesh screen filter and
membrane filter are considered absolute.  These absolute filters have a
pore-size rating (a definite largest particle that can penetrate the
filter) which remains the same throughout the life of the filter.  Both
of these filters have a continous matrix, do not migrate or slough off,
and rely solely on their pore size for particle retention.

     When the solution to be filtered is nonaquaeous or contains a
solvent incompatible with the filter, solvent-resistant membrane filters
may be used.  The most common membrane filter is the 1.5y pore size.
For fine particle removal, the 1.2u pore-size membrane filter is fre-
quently used.  Therefore, all particles larger than 1.2y in their
maximum dimension will be removed, providing nearly complete protection
from all contaminants.

     In contrast to the absolute filters, depth-type filters have no de-
finite pore-size rating.  Because of their random orientations of sub-
structures, they are rated on a nominal basis.  They effectively remove
gross matter, slimes, and sometimes even bacteria, but they do migrate.
Generally, the ideal system is a combination of a large depth-type
filter followed by an absolute filter (14).


Funda Filter

     The Funda Filter is used for filtering, washing and drying such
products as streptomycin and penicillin crystals, sorbite, sala'zoDyrine,
phenacetin, steroids, and intermediate chemicals for cortisone manufac-
turing.  Also, it is used with activated carbon for filtering penicillin
liquors either in the aqueous or solvent stages (26).

     The Funda Filter is a vertical unit with a series of horizontal
filter leaves enclosed by the filter vessel.  Filtration takes place
only on the upper surfaces of the horizontal leaves.  Two forms of
filter units are available.   One, the precoat filter (or slurry dis-
charge filter), is designed to discharge the separated filter cake as

-------
42


slurry at the end of the run.  The other, a dry-residue filter, discharges
the collected solids from the filter as a dry or semidry cake.  Both
forms of filters may or may not be precoated and incorporate filter
aids.  The filtering media is a fine metal screen or a cotton, or synthetic
cloth which retains particles 5u or smaller.  The dry residue filter
becomes increasingly important because operating requirements to,day are
specifying a dry cake discharge ready for disposal.

     The filter vessel under pressure can be used as a reactor if Acces-
sary.  It is particularly suitable for handling of volatile, toxic, ex-
plosive, or noxious materials that must be confined in a pressure
vessel.  It can be fully automated'and instrumented.  .The collected cake
can be washed with either water or solvent.  The Funda Filter also can
treat the residual heel left after filtration by completely disgorging
the final tank liquids as clear filtrate.  If warranted, the filter cake
can be dried to any degree within the filter vessel;  Centrifugal force
and air or inert gas sparging can quickly and easily release the cake
from the filter leaves and the vessel.

     In straight filtration, if the filtrate is turbid, it may be recycled
to the feed tank and again through the filter until it becomes clear.
After filtration, washing the cake may involve relatively large or small
amounts of wash liquid.  A very expensive solvent used for washing would
be limited in volume.  Successive washings may be conducted with different
wash liquids, each to dissolve specific materials in the cake - a
process known as selective washing.  Considering the variety of process
operations in pharmaceutical manufacturing, the chemical values may be
largely contained in the filter cake, or otherwise in the filtrates and
within particular washes.  But, the chemical values are generally in the
liquid state.  The filter discs need not be cleaned after each cycle,
although cleaning is simple and requires little time.

     Wastes associated with filtration include the discarded filter
cake, wash liquids, precoat and filter aid materials, and washwater from
filter cleaning (26).
Specific Applications of the Funda Filter

     In the manufacture of para-aminosalicyclic acid, bleaching and deo-
dorizing with activated carbon is necessary and this process step is
undertaken through the Funda Filter.  A subsequent step in para-amino-
salicyclic acid production involves the washing of these crystals inside
the filter-and then dissolving these solids for the next crystallization
step.  A third application is the filtration of the sodium salts from
the para-aminosalicyclic acid which is followed by washing of the cake
with alcohol and hot air drying inside the filter.

     In the manufacture of carotene (a precursor of vitamin A), the Funda
filter is used for decolorizing with activated carbon in more or less
the same manner for para-aminosalicyclic acid above.  The filter is used
in some cases for the selective separation of alpha and beta-carotene
using different solvents.

-------
                                                                        43
     A pharmaceutical plant  in  England utilizes the Funda Filter for
filtration of antibiotics  in the  solvent state and a second filter  1s
used for recovery of the antibiotic crystals from the crystallizer.   In
another antibiotics production  line shown below (Antibiotic A), the
Funda Filter is an essential  part of overall operations.  The completed
fermentation broth is adjusted  with sulfuric acid and mycelium is re-
covered through a vacuum filter.   To the antibiotic filtrate are added
additional sulfuric acid,  surfactants, and solvent.  This mixture passes
through a DeLaval Model ABE-216 countercurrent multistage extractor for
the separation of spent broth,  stripped solvent and an aqueous buffer
solution carrying the active ingredient.  Spent solvent is recycled to
the filtered fermentation  broth.   Very high antibiotic recoveries are
possible through the DeLaval  unit.

     The Funda Filter receives  and filters the aqueous buffer solution
containing the antibiotic  values  through activated carbon.  A solvent  is
added to what is believed  to be the filtrate, which subsequently passes
through a centrifugal separator for separation of the spent buffer  from
the main process line.  Anhydrous ammonia is added to the active liquid
phase which is filtered before  entering the crystal!izer.  The last
stage employs a Funda Filter for  final separation of the antibiotic
crystals which are then ready for packaging (26).  A number of waste
lines are present which, however, are not shown on the figure.
             ABE 216
                        FCRMENTER
DE LAVAL MODEL
ABE 216 COUNTER
CURRENT EXTRACTOR
                           AQUEOUS BUFFER
                             SOLUTION
                                  CENTRIFUGAL
                                 SEPARATOR FOR
                                  LIQUID LIQUID
                                  SEPARATION

                                   ANHYDROUS
                                    Na2S04

                                 FILTER
                                                          CRYSTALLIZER
                                                          FUNDA FILTER
                                                  ANTIBIOTIC
                                                   CRYSTALS


                                                  COMMERCIAL
                                                   PRODUCTS
                     Antibiotic A production (26)

-------
44
DRUG FORMULATION

     The Folkestone plant of Charles Pfizer and Company in Great Britain
receives Terramycin hydrochloride and various other bulk Pharmaceuticals
and prepares and formulates these drugs into the desired form for
customer sales.  The Terramycin salt, upon reaching Folkestone,  is  first
blended with additives to stabilize the product and also pH adjustment
is made (41, 80).

     All vials and bottles for antibiotics are washed and sterilized in
a separate department.  Two washing machines are available, one  a single
jet unit capable of handling 2,500 bottles/hr, and the other a multijet
unit handling 8,000 vials/hr.  Sterilized containers are kept within a
sterile area which is under pressure to prevent entry of other than
carefully controlled air supply.  Air to the sterile area is dried  to
low relative humidity, cooled, and joins with recirculated air,  the
latter having passed through a cyclone and filters.  This combined  air
passes through electrostatic precipitators, filters and glass wool
before entering the sterile area.  Washed rubber stoppers and clothing
enter the sterile area via a steam autoclave.  All filling of bottles
and vials is conducted in the sterile atmosphere.

     In the tablet manufacturing department, a binding agent and a
disIntegrant are mixed with the Terramycin, and the powder mixture  is
made into slugs.  These are passed through a rotary granulator,  next
through a sifter, and then fed into a rotary tablet-forming machine.
The tablets subsequently receive finishing coatings of shellac varnish,
gelatin, white sugar and a wax polish.  Other operations at the  Folkestone
plant include filling capsules, and the preparation of sterile and  non-
sterile liquid solutions, non-sterile powders, and various ointments
(80).

-------
                       V.   WASTE CHARACTERISTICS


     Major types of wastes from the pharmaceutical  industry result  from
the process of fermentation or the production of  synthesized organic
chemicals and biologicals.

     Process waste streams from fermentation, which produces antibiotics,
vitamins, steroids and associated products,  may include:

     1. Liquid and solid wastes from fermentation operations
     2. Liquid wastes from extraction and purification processes
     3. Liquid and solid wastes from recovery processes
     4. Floor and equipment washdowns
     5. Sanitary and miscellaneous waste streams.

The strength of fermentation wastes can vary appreciably  with different
batches and between different manufacturers.  Howe (6) estimates that
the production of 0.45 kg (1 Ib) of antibiotics generates from 11.4 to
13.2 nT (3,000 to 3,500 gal) of wastewater.   Based on 4,540 kg (10,000
Ib) of antibiotics produced daily during 1958, the average wastewater
volume from antibiotics was 114,000 to 132,000 m /day (30 to 35 mgd).
The concentration of pollutants in the washwaters depends on the extent
of washing and the particular cleanup procedures employed.

     Wastewaters from the manufacture of synthesized organic chemicals
consist of complex mixtures of organic and inorganic materials having
varied characteristics.  The organic wastes  include solvents, salts,
acids and some plant and animal derivatives.  These wastes usually
contain high COD and TDS.   The pH is usually very low or  high and fre-
quently the wastes are colored.  The chemical derivatives extracted from
natural plants or animal organs are unusually strong in TS, TSS, BOD and,
pH, and they are generally toxic to fish, aquatic life and animals.
With careful production scheduling and waste stream control, it is
possible to segregate the waste sources and  separately treat the strongest
or most complex chemical waste streams.

     Biological Pharmaceuticals production,  including antitoxins,
antisera and other associated compounds used in the treatment and pre-
vention of specific diseases, results in the generation of large quanti-
ties of wastewaters which are difficult to handle.  Such  wastes contain
animal droppings, animal carcasses and organs, blood, body fluids,  fats,
egg fluid, egg shells, biological culture media,  feathers, solvents,
antiseptic chemicals and herbicidal compounds.  They are  generally
characterized by high BOD, COD, TS, toxicity, colloidal solids, color
and odor.  Sometimes the antitoxin and antisera wastes contain highly
dangerous pathogens together with toxic components such as benzene,
phenols, cresols, mercury compounds and a variety of other bactericidal
materials.  The presence of these compounds  can cause great difficulty
in determining waste characteristics by standard  analytical methods (6).

     The following data on waste characteristics of the pharmaceutical
industry were reported for fermentation, synthesized organic chemicals,
and biological plants.

-------
46


FERMENTATION

The Upjohn Company, Kalamazoo, Michigan

     The Upjohn Company (43) in the late 1950's and early 1960's analyzed
spent beer wastes, defined as the end product after the extraction of
either antibiotics or steroids.  After solvent extraction, significant
amounts of solvent and the antibiotic or steroid may still remain in the
colored, odorous spent beers.  Composition of the spent beers is as
follows:
          Component                     Content

          Total solids                   1 to 5%
          Ammonia-N                    100 to 250 mg/1
          BOD                        5,000 to 20,000 mg/1
          pH                             3 to 7


Makeup of the total spent solids is typically:


          Component                     Content (%)


          Protein                       15 to 40
          Fat                            1 to 2
          Fiber                          1 to 6
          Ash                            5 to 35 (which includes
                                                  Ca, P and K)
          Total Carbohydrates            5 to 27
          Steroid(s)                    Present
          Antibiotic(s)                 Present

The vitamin content of the beer solids is given as:


          Component                     Content (yg/g)

          Thiamine                       4 to 12
          Riboflavin (B2)               10 to 150
          Calcium                       35 to 2,000
          Pyridoxine HC1                125 to 75
          B1?                           Present
          FAfic acid                     1 to 5


     Howe (6) reports that spent fermentation broth solids can (potentially)
amount to 0.2 kg (0.4 Ib) solids per gallon of broth.  This is equivalent
to about 47,000 mg/1 or 4.7 percent waste total solids. 'These solids,
mostly dissolved, are characterized as high in nitrogen, phosphates, vita-
mins and may contain trace amounts of antibiotics.   Howe emphasizes re-
covery of these solids, particularly for animal feeds.

-------
                                                                         47
47


Bristol Laboratories, Syracuse. New York

     Bristol Laboratories (32, 65) analyzed penicillin production wastes
which were a mixture of solids, spent broth, chemicals, waste acids and
caustics, unrecovered solvents, and washing and cooling waters.  The
solid penicillin growth, or mycelium, is reported to have a high nitro-
gen and BOD content.  The spent penicillin broth contains sugars,
proteins and organic acids and is quite high in BOD and suspended
solids.  Depending on the amount of food remaining at the completion of
fermentation, and the extent of in-plant control and recovery of solvents
and mycelia, the spent broth will have a BOD ranging from 4,000 to
13,000 mg/1.  In addition to day-to-day wastes, the disposal of contamin-
ated fermentation batches presents a special problem since individual
batches may run as high as 30,000 mg/1 in BOD and 20,000 mg/1 in TSS.

     About mid-1948, before waste control procedures were fully im-
plemented at Bristol Laboratories, the total wastes from the fermen-
tation facility had a BOD population equivalent from 0.1 to 0.3 PE/1
(0.37 to 1.0 PE/gal) of discharge, averaging 0.2 PE/1 (0.74 PE/gal)
waste.  By controlling contaminated or spoiled fermentation batches and
with improved control and recovery of solvents, these BOD loads were
reduced 61 percent to 0.04 to 0.11 PE/1 (0.16 to 0.43 PE/gal) of dis-
charge, averaging 0.08 PE/1 (0.29 PE/gal).  After passing through
equalization and partial neutralization (pH adjustment up to 4.5) the
wastes were directed to municipal treatment.

     In their early experiences, personnel at the Syracuse, N.Y. muni-
cipal treatment works recorded 35,000 mg/1 BOD and TSS in the pharma-
ceutical plant penicillin sewer.  These extreme values occurred before
Bristol Laboratories initiated mycelium recovery and solvent stripping.
BOD and TSS concentrations were later reduced to 5,000 and 3,000 mg/1,
respectively.  The strength of penicillin wastes can vary appreciably
with different batches and from one manufacturer to another.

     Impact on the municipal treatment plant appeared to be caused more
by the quantity rather than the strength and nature of,the wastes re-
ceived from the Bristol Laboratories.  The main effects were:

       1. Solvents imposed additional loads on aeration units with
          greater costs for air supply

       2. Rates were higher for recirculation of return sludge and
          other flows to reduce shock loads

       3. Considerably larger volumes of sludge were produced as a
          result of the penicillin wastes.  Much higher solids
          loads were subsequently imposed upon the digesters.

     It was concluded that penicillin wastes are amenable to aerobic and
anaerobic treatment, provided no substances are present which may poison
or retard biological activity.   However, over the years a greater than
normal  number of upsets, some serious, have apparently occurred at the
Syracuse, N.Y., biological  treatment facilities.

-------
48


A Pharmaceutical Plant in India

     Waste characterization data was obtained from a pharmaceutical
plant producing penicillin and streptomycin (7).  In penicillin production,
molds of the PeniciIlium notatum-chrysogenum group are cultured in a
medium of corn steep liquor, peanut meal, lactose and mineral salts.
The wastewaters include mycelium and the spent filtrate and wash waters
from succeeding steps.  In streptomycin production, pure cultures of
Streptomyces griseus are cultured in a medium consisting of a sugar
source (glucose) and a nitrogen source (corn steep liquor or equivalent).
The filtrate is absorbed on charcoal or a resin and the active ingredient
is eluted from the resin with acid.  The eluate is neutralized and
concentrated.  Acetone is used to precipitate the,crude product which is
purified in subsequent process steps.  Streptomycin wastewaters also
included mycelium and the spent filtrates and washwaters.  Certain
amounts of mycelium were sold as manure or stock feed.

     Representative raw wastewaters were collected from the penicillin
and streptomycin production areas.  The average results of eight 24 hr
composite samples and results based upon eleven grab samples are given
in Table V-l.  The streptomycin wastes were stronger than those origin-
ating from penicillin production.  The nitrogen concentration was low
but the phosphate, sulfate and chloride values were high.  More than 40
percent of the total solids were volatile solids, indicating high
biodegradability potential.

     Reference was made to information previously reported by Muss in
1951 which showed combined wastes from penicillin manufacturing as
having an average BOD of 2,100 mg/1, a variation in BOD from 750 to
5,000 mg/1, and total solids about 5,000 mg/1.  Heukelekian reported an
average of 4,475 mg/1 BOD for penicillin wastes with a variation from
2,400 to 8,150 mg/1 BOD.  Spent streptomycin broth was shown to have an
average BOD of 2,450 mg/1  with a range in BOD values from 825 to 5,900
mg/1.

Antibiotic Wastes, Great Britain

     Lines (10) describes fermentation wastewaters as being contaminated
with solvents, disinfectants and various solids including filter aids.
If the fermentation broths and mycelia can be evaporated and dried into
salable end products, then the waste sources can be reduced to evapora-
tor condensates and washwaters.  The highly organic nature of fermen-
tation waste creates a favorable situation for biological treatment, but
treatment plant biota can be adversely affected by the antibiotic
residues.  The BOD of fermentation wastewaters is usually in the range
of 5,000 to 30,000 mg/1.  Analysis of a principal antibiotic waste in
Great Britain proceeding recovery and treatment is depicted below:

-------
                                                   Table  V-l
                                  Penicillin  and  Streptomycin  Raw Wastewaters
                                       Fermentation  Plant  in  India  (7)
Parameter
Color
Odor
PH
24-Hour Composites Grab Samples
Penicillin Wastes
Colorless
Fru i ty
6.3

Streptomycin Wastes Penicillin Wastes
Pale yellow
Septic
6.2 3.9 to 7.8
(mg/1 )
Streptomycin Wastes
2.9 to 8.7

BOD (37C; 99F)              1,490
Total  free and albuminoid
     ammonia (N)               17.7
Organic nitrogen (N)            17.3
Nitrates (N)                    0.3
Phosphates (POJ               72.0
Sul fates (SO.r                51.0
Chlorides   *                  91.0
TS                            1,910
TVS                             880
TSS                             420

Settleable solids,  ml/1          8.0
1,800

 31.0
 29.1
  0.7
 65.0
 52.0
204.0
3,590
1,450
1,750

 81.0
650 to 5,500

3.1 to 29.4
0.8 to 35.3
0.1 to 0.5
 18 to 700
 26 to 192
 16 to 200
480 to 26,200
200 to 12,180
 70 to 1,080

2.0 to 56
500 to 2,800

3.0 to 57.1
7.8 to 40.9
0.0 to 0.8
  9 to 700
 25 to 765
 66 to 1,200
960 to 4,950
480 to 3,070
 80 to 1,800

2.3 to 214
                                                                                                                  10

-------
50
          Component                     Content
                                   (mg/1 except pH)

          pH                            9.3
          TS                            23,690
          TSS                           18
          BOD                           7,120
          Total nitrogen (N)            1,260
          N(UN)                        41
SYNTHESIZED ORGANIC CHEMICALS

Squibb and Sons, Inc., Humacao, Puerto Rico

     Squibb conducted treatability studies (about 1969-70) prior to
construction of its integrated pharmaceutical plants in Humacao, Puerto
Rico (23).  Synthetic penicillin was to be produced at one plant, and
two antifungals, Amphotericin and Mycostatin, at the other.   Major unit
operations consisted of hydrolysis, solvent washing, solvent recovery,
crystallization, distillation and filtration.

     Initial studies showed that wastewaters should be segregated into:

                 1. A strong process wastewater stream
                 2. A weak process waste stream together with methanol
                    solvent recovery still bottoms, plus sanitary
                    sewage
                 3. Service waters, which included boiler and cooling
                    tower blowdowns and spent demineralizer regenerants.

     The Squibb report emphasized that none of the dilute process
wastewaters, including the methanol solvent recovery still bottoms,
showed any sign of acute toxicity although no data were presented in
this direction.

     The proposed waste treatment and raw wastewater characteristics are
given in Table V-2.

Berkeley Chemicals Division, Summit, New Jersey

     Berkeley Chemicals Division (58, 102) produced various  synthesized
organic chemicals which were not specifically named.  Raw materials
and/or final products included sodium formate, formaldehyde, isopropyl
alcohol, citric acid, dehydroxynapthelene sulfonic acid, cupric acetate,
and xylol.  Process wastes amounting to 94 to 190 m /day (25,000 to
50,000 gpd) and containing 8,000 to 20,000 mg/1 BOD were trucked from the

-------
                                              Table  V-2

                 Proposed Pharmaceutical  Wastewater  Treatment and Characteristics
                         Squibb and Sons, Inc.,  Humacao,  Puerto Rico (23)
Stream
(1) Strong
Process
avg
max
(2) Dilute
Process
avg
max
(3) Service
Water
avg
max
Composite
Treatment Flow
(m3/day)
Incineration
44.7
65.9
Biological
127.9
141.6
Neutralization
and Settling
133.6
306.2

(gpd)

11,800
17,400

33,800
37,400

35,300
80,900
BOD^-7 COD-7
(kg/day) (Ib/day) (mg/1^-7) (kg/day) (Ib/day) (mg/1^7)

21,500 47,300 480,000 30,700 67,600 687,000
33,700 74,200 48,000 105,800

82 180 640 113 250 890
86 190 127 280


21,500 47,500 30,800 67,850
a/  COD/BOD ratio of 1.4 assumed
b/  Calculated
                                                                                                                tn

-------
52
plant and bled into the Summit, N.J.  sewerage system.   These wastes in
1962-63 were rated around 2,360 kg (5,200 Ib) BOD/day mainly derived
from spent mother liquors and equipment and floor washings.   Cooling
waters and condenser waters, relatively uncontaminated, were discharged
untreated to the Passaic River.

     Process wastes from five major sectors in the plant were described
as:

     Waste A - Sodium formate, formaldehyde, hexaldehyde; 1,510 to
               1,890 I/day (400 to 500 gpd) mother liquors.
     Waste B - Solutions from urea and decomposition products;
               227 kg (500 Ib) urea discharged/day.
     Waste C - Isopropyl alcohol, citric acid.
     Waste D - Waste from purification of dehydroxynapthelene
               sulfonic acid; becomes deep red when  combined with iron.
     Waste E - Cupric acetate and xylol.
     Overall process waste characteristics were given as:
          Component                     Content


          Flow                          94.6 to 189 m3/day
                                        (0.025 to" 0.05 mgd)
          BOD                           8,000 to 20,000 mg/1
          Phenolics                     5 to 500 mg/1
          TSS                           350 to 850 mg/1
          VSS                           70 to 85% of TSS
          pH                            8.2 to 10.1
     Various laboratory and pilot plant studies were conducted on pre-
treatment of the chemical wastes.  An effort was made to segregate the
strongest waste streams to determine if separate treatment were justified.
The following analyses were obtained:


     Source                                         BOD
                                   mg/1            kg/day         Ib/day
Still bottoms No. 1                 840,000          38              85
Still bottoms No. 2                 166,000          36              80
Mother Liquor No. 1                  27,700         351             775
Mother Liquor No. 2                  27,000         340             750
Mother Liquor No. 3                  20,000         254             560

-------
                                                                      53
       The BOD load in the three spent mother liquor streams amounted to
946 kg (2,100 Ib) BOD/day, representing about 40 percent of the total
BOD load from the pharmaceutical plant.

     It was difficult to determine BOD analyses for the process waste-
waters.  Using the BOD dilution techniques, a sliding BOD scale was
evident in the various samples.  BOD values increased with increasing
dilutions, indicating a striking inhibitory effect.  In a series dilu-
tion of the combined process wastewaters (using manometric techniques)
ranging from 1 through 100 percent waste concentration, the BODy of the
pure waste was recorded as 600 mg/1, whereas the 1 percent waste (cor-
rected to the 100 percent level) gave a BOD7 of 13,000 mg/1.  The latter
value was more than 20 times greater than the former.  For BOD results
collected during the study, the reference author indicates it is very
likely that maximum (desired) BOD values were not recorded.

     The wastewaters were extremely strong and inhibitory to biological
treatment.  It was noted over the two years of study that not only did
the overall plant BOD load increase from 1,810 to 2,720vkg (4,000 to
6,000 lb)/day, but also the character of the wastes was constantly
changing.

Parke-Davis and Co., Holland, Michigan

     As a first step in waste characterization and treatment facility
planning at the Parke-Davis Chloromycetin plant in the 1950's, Company
personnel studied each operation, the various reactions, chemicals used,
respective yields, unreacted compounds, wastes resulting from the wash-
ing of solids during filtering operations, solvent recovery possibili-
ties, byproducts, and auxiliary data (50).  All wastes to be disposed of
were listed, and a program was set up for testing the major wastes.
These tests included pH, BOD, TSS and TDS, solvent content, toxicity,
color and chlorine demand of the wastes.

     The most deleterious wastes were defined as:

                 1. Bottoms from the solvent recovery plant

                 2. Wash "A" from one of the filtering operations

                 3. Wash "P" from another filtering operation

                 4. Filter and spent wash stream termed liquor "S."


     Analysis on the composite of these four waste streams, consisting
of 52.9 percent bottoms, 30.1 percent wash "A," 7.5 percent wash "P,"
and 9.5 percent liquor "S," was as follows (59):

-------
54
          Component                     Content


          pH                            3.7
          BOD  (20-day)                  45,000 mg/1
          IDS                           60,000 mg/1
          Color                         Reddish-brown; 12,000 units
          Toxicity                      2 to 5%

     The wastes were said to contain complex, unidentified organic
compounds resulting 'from the processes.  Ions present included sodium,
chloride, ammonium, acetate, bromide and nitrates.  No flow values for
the chemical waste system were reported.  Chemical wastes were treated
within the chemical process sectors for the removal of aluminum salts,
sulfates, phosphates and immiscible solvents (13).  The types of waste
materials remaining in the chemical waste sewer were reported as:
          Component                     Content
                                    (kg/day)  (Ib/day)
Acetic acid
Ammonium acetate
Sodium acetate
Sodium chloride
Sodium and ammonium
bromide
PH
COD
910
910
770
450

230


2,000
2,000
1,700
1,000

500
4 to 5
40 to 60,







,000 mg/1
                                              but sometimes as high
                                              as 100,000 mg/1
          Large quantities of methanol,
          misc. tars and dissolved organics,
          xylene


The various in-plant procedures required for removal of aluminum salts,
sulfates and phosphates were sometimes more extensive than the chemical
process steps that produced these compounds.

     As the chemical wastes pass through retention-equalization tanks,
some precipitates are settled out.  Xylene formed an emulsion on the top
of the retention tank, and was periodically skimmed off.  The separated
waste materials were, trucked away to land disposal.


M/s Indian Drugs and Pharmaceutical Plant. Hyderabad. India

     Preliminary production figures projected in 1962 for the various

-------
                                                                      55
chemicals to be manufactured at the Hyderabad plant are shown in Table
V-3.  The groups of synthetic drugs include the sulfanilamides,  the
antipyretics, (i.e., phenacetin), the vitamins B^,  62 and folic  acid,
and the antitubercular drug, isonicotinic acid hydrazide.  Large quanti-
ties of both inorganic and organic basic chemicals  are necessary for
intermediate chemicals production.  Some slight shifting of product mix
was indicated between 1962 and 1970.  Patil (30) lists 35 chemicals to
be manufactured whereas Mohanrao (54) later cites  up to 70 various
chemicals associated with the Hyderabad production  facilities.
                                            q
     The Russians estimated that about 600 m /day  (0.15 mgd) industrial
wastewaters could be expected from the factory, excluding the sanitary
sewage and the spent cooling and condenser waters.   Waste constituents
consist largely of inorganic and organic salts, almost all of which are
dissolved.  Approximate composition of the untreated synthetic drug
wastes are given below (30):
     Parameter                                    Concentration (mg/1)


     Calcium chloride                                    600-700
     Sodium chloride                                   1,500-2,500
     Ammonium sulfate                                 15,000-20,000
     Calcium sulfate                                     800-:21,000
     Sodium sulfate                                      800-10,000
     Sulfanilic acid and related products                800-1,000
     Various sulfa drugs                                 400-700
     p-amino phenol, p-nitrophenolate,
          p-nitrochlorobenzene, etc.                     150-200
     Amino-nitrozo ami no-benzene antipyrene sulfate      170-200
     Analogous substances                                150-200
     Alcohols (methyl, ethyl, isopropyl, etc.)         2,500-3,000
     Benzene, toluene, etc.                              400-700
     Chlorinated solvents (dichloroethylene,
          chloroform, etc.)                              600-700
     Other substances                                    Unknown
     The M/s Indian Drugs and Pharmaceuticals installation, completed in
the early 1970's, consisted of twelve blocks or sectors (54).   Nine of
these were to be used for the manufacturing of drugs and chemicals, and
the remaining three for storage of raw materials, intermediates and
final products.  The nature and volumes of wastes discharged from each
sector are shown in Table V-4.

     The majority of wastes discharged are acidic.   Sector 8,  which
produces intermediates for the manufacture of synthetic drugs, con-
tributes slightly more than one-half of the concentrated industrial
wastewaters from the factory.  Wastes from Sectors 3 and 8 and the

-------
56
                               Table V-3

      Design Production,  Synthetic  Drugs and  Intermediates  (1962)
            M/s Indian Drugs  and Pharmaceutical  Plant  (30)
                           Hyderabad, India
Drug
Sulfadimidine
Sulfadiazine
Sulfaquanidine
Sulfacetamide sodium
Acetozol amide
Su If anil amide
Urosulphan
Ethazol (Globucid)
H.N.H.
Chloroquin
Luminal
Phenacetin
Ditrazine
Pyrami done
Novalgin
Piperazine sulphate
Vitamin B-l
Vitamin B-2
Vitamin PP (nicotinamide)
Folic acid
Annual
Production
(metric tons) (tons)
181
136
68
45
23
45
23
23
18
4.
9
91,
9
36
9
9
23
2
27

200
150
75
50
25
50
25
25
20
5 5
10
100
10
40
10
10
25
2
30
9 1
                         Total  tons/year      782           863

              (Intermediates)

          Acetanilide Sulphonylchloride     1,360         1,500
          Acetyl  acetone                       91           100
          Aceto acetic ester                   45            50
          Aceto propylic alcohol                45            50
          Phyenyl  ethylnalonic  ester           18            20
          Sodium bisulfite (38% soln.)         544           600
          Dichloroethylene                    272           300
          Propionyl  chloride                   18            20
          Ortho-formic ester                   18            20
          Hydrazine  sulfate                   181           200
          Diethyl amine                         36            40
          Sulfanilamide tech.                 408           450
          Beta picoline, purified              27            30
          Gamma picoline, purified             18            20
          Piperazine  hydrate                   36            40

                         Total  tons/year    3,117         3,490

-------
                          Table V-4

             Character of Industrial  Wastewaters
From Various Sectors of the Hyderabad, India Plant, 1970 (54)
Wastes
Sector
1
2
3
4

5
6
7
8
9

12

13
14
Activity
(
Sulfa
Vitamin B,
Vitamin B~ and folic acid
Antipyretic and anti tubercular
drugs
Products using sodium metal
Sodium bisulfite
Phosgenation and chlorination
Intermediate chemicals
Pilot plant and central
laboratory
Liquids and solids chemical
storage
Acid and alkali storage
Raw materials
Acid
o
m /day)(gpd)
28
7
4
15
5.5
1
8.5
262

6

0
15
0
7,400
1,850
1,060
3,960
1,450
260
2,240
69,200

1,590

0
3,960
0
Alkaline
(m3/day)(gpd)
33
4.2
3
18
8
0
0.3
45

4

0
0.1
1.5
8,720
1,110
790
4,760
2,110
0
80
11,900

1,060

0
26
400
Neutral
(m3/day)(gpd)
0
15
10
23
28
0
0
0

20

0.1
0
0
0
3,960
2,640
6,080
7,400
0
0
0

5,280

26
0
0
Total
Hastes
(m3/day)(gpd)
61
26.2
17
56
41.5
1
8.8
307

30

0.1
15.1
1.5
16,100
6,920
4,490
14,800
11,000
260
2,320
81,100

7,920

26
3,990
400
     Total
352  93,000
117.1   30,900
96.1  25,400   565.2 149,300
                                                                                           Ul

-------
58
overall composite wastes were highly acidic.  The remaining sectors
contributed predominately alkaline wastes.  Specific characteristics of
the wastes from the individual sectors are presented in Table V-5.  The
various sectors (except 3 and 5} had high BOD values ranging from 11,800
to 20,000 mg/1.  Total suspended solids (not shown in Table V-5) were
found to be negligible; however, some settleable solids were evident.
These settleable solids when washed with distilled water are readily
redissolved into solution.  All wastes had high nitrogen content with
Sectors 3 and 5 containing the least amounts.  The .phosphorous content
was negligible.

     Because the wastes were known to'contain large amounts of various
organic and inorganic chemicals, they were expected to be toxic.  Con-
sequently, toxicity studies were conducted on the flows from Sector 8,
the composite flow from all sectors, and the total wastes from all
sectors excluding Sector 8.  The standard bioassay test was employed
with the fish species, Barbus ticto, Puntius puntius and Cyprus carpio
communicus.  The 1970 bioassay results at the Hyderabad plant were as
follows (54):


          Source of Waste                              48-hr TLm

                                                       (% by vol)

Sector 8, after lime neutralization
     to pH 7.0, followed by sand filtration               26.0

Composite flow from all sectors, after
     sodium hydroxide adjustment to pH 7.0                0.29

Composite waste from all sectors,
     excluding "Sector 8, pH 9.2                           0.27


     With or without extensive neutralization, the wastes continued to
be extremely toxic.  Flows from Sector 8 were less toxic, but only
after this waste stream received pH adjustment from 0.6 to 7.0.
Upjohn Laboratories, Kalamazoo, Michigan

     [See Section VIII:  WASTE TREATMENT AND DISPOSAL: OTHER TREATMENT
METHODS.]
BIOLOGICALS

Eli Lilly Greenfield Laboratories, Greenfield, Indiana

     The production of antitoxins, antisera and vaccines generates
wastewaters containing animal manures, carcasses and organs, body fluid,
blood, fats, egg fluids and egg shells, spent grains, biological culture

-------
                                                                  Table V-5


                            Composition of Wastewaters from Individual  Sectors,  Hyderabad,  India  Plant,  1970  (54)

Parameter

Flow, cm /day
pH
TS (percent)
TVS (percent)

Total acidity (CaCO,)
Total alkalinity
(CaCO,)
Chlondls (Cl)
Sul fates (SOJ
Total nitrogen (N)
Phosphorous (P)
BOD
COD
Sector 1
(Sulfa)

61
9 0
11 8
3.2

*

12,700
22,500
25,600
10,000
*
17,500
31 ,600
Sector 2
(Vitamin B, )
1
26 2
10.0
0.7
0.2

*

24,500
4,500
1,700
8,600
*
20,000
33,100
Sector 3
(Other
Vitamins)
17
3.0
5.1
0.5

2,500

*
4,000
14,800
130
*
1,600
3,200
Sector 4
(Antioyretic,
Anti-TB)
56
10.9
9.7
2.0

*

8,900
8,500
14,800
2,400
*
15,000
27,800
Sector 5
(Sodium
Metal Use)
41.5
11.9
3.5
0.1
(mg/D
*

4,100
14,500
2,400
430
*
1,600
3,500
Sector 7
(Phosgen-
ation)
0.3
11 0
5.6
1.1

*

235,000
8,000
4,900
*
*
19,300
43,100
Sector 8
(Intermed-
iates)
307
0.6
8.3
6.7

57,600

*
20,500
37,000
6,200
*
9,400
13,700
Total
All Sectors
but #8)
202
9.3
7.7
1.7

*

10,800
17,000
14,800
5,200
*
15,300
28,500
Total
All
Sectors
509
0.8
8.6
5.0

32,200

*
18,500
28,000
6,100
*
13,000
19,700
*  Negligible
                                                                                                                                                       tn
                                                                                                                                                       vo

-------
DU
media, feathers, solvents, antiseptic agents, herbicidal components,
sanitary loads, and equipment and floor washings (6, 31).  These wastes
are characterized by high BOD, COD, TS, colloidal solids, toxicity,
color and odor.  In many instances antitoxin and antisera wastes from
production and research can carry highly dangerous pathogens and special-
ized toxic chemicals as benzene, phenols, cresols and mercury compounds.
Presence of these materials often causes serious difficulty in analyzing
the wastes by standard analytical procedures.  In the mid-1950's the
mixed wastes from the Greenfield Laboratories amounted to 56,800 I/day
(15,000 gpd) and had the following characteristics (31):
          Component                     Content

          BOD                     1,000 to 1,700 mg/1
          TS                      4,000 to 8,500 mg/1
          TVS                     3,000 to 7,500 mg/1
          SS                        200 to   800 mg/1
          pH                        7.3 to   7.6
          Color                      Brownish-red
     By the late-1950's, the facility had grown appreciably (70).  The
problem was not the large wastewater volume but rather a host of scatter-
ed waste sources requiring special pretreatment measures.  The wastes
were categorized as:

          1. animal wastes from test animals
          2. pathogenic or infectious wastes from laboratory
             sectors conducting research on animal disease
          3. toxic chemical wastes originating from laboratory
             sectors conducting research upon bacteriological,
             botanical and zoological problems
          4. wastes from production of antisera and antitoxins
          5. overall plant sanitary wastes.
                                                  o
     The combined wastewaters were averaging 680m /day (180,000 gpd)
with the maximum flow rate approximating 1,820 m /day (480,000 gpd).
The unit waste loads from various animals including horses, cattle,
calves, pigs, rabbits, mice, monkeys, dogs, cats and pouts varied from
0.8 to 80 I/day (0.2 to 20.0 gpd) per animal and from 0.007 to 2.9 kg
(0.016 to 6.50 Ib) BOD/day/animal.  Raw waste characteristics are given
in Table V-6.
DRUG FORMULATION

Dorsey Laboratories. Lincoln. Nebraska

     [See Section VIII:  WASTE TREATMENT AND CONTROL: ACTIVATED SLUDGE;
Drug Formulation.]

-------
Antisera/antitoxi ns
     production

Additional anti-
     influenza vaccine
     production

Plant science and
     animal research
Monkey storage areas

     Total (avg)
                                                                 Table V-6
                                                         Raw Waste Characteristics
                                              Eli Lilly Greenfield Laboratories, Indiana (70)
Plant Area Flows
Ave Max
(ir3/day) (gpd) (m3/day) (gpd)
BOD
(mg/1) (kg/day) (Ib/day)
Total Solids
(mg/D (kg/day) (Ib/day)
Settleable
Solids
(kg/day) (Ib/day)
                               204
54.000    655    73,000     200-400     45-73   100-160   760-1,520  154-308  340-680    41-77    90-170
                                                                                                                 60-110   4.5-14
                                                                                                                                    10-30
                               250     66,720     383    101,000     40-80      9-23   20-50    100-200     27-50


                               204     54.000     708    187,000    810-1,620 186-367 410-810 1,220-2,450  249-499  550-1,100  64-354  140-780

                                18      4,700     83    22,000    800-1,600  14-27   30-60    770-1,540  136-272  300-600   136-272  300-600


                               680    179,400   1,820    480,000      380       254     560*      860        567     1,250*     122      270*

*  Maximums of 1,080 Ib/day BOD, 2,690 Ib/day  TS, and  1,080 Ib/day settleable solids

-------
                 VI.  WASTE PARAMETERS OF SIGNIFICANCE
     Some important parameters in disposing of wastes generated  from the
pharmaceutical industry are given by Molof and Zaleiko,  1965 (25).  Major
monitoring parameters for waste streams include BOD,  variation in BOD
from hour to hour, toxic chemicals, color, odor, solids  (dissolved,
colloidal and suspended), temperature, and variation  in  hydraulic load-
ing. The ammonia and organic nitrogen content in pharmaceutical  wastes
is of increasing concern.  These loads sometimes equal  or exceed BOD
loads in the wastewater and cause damaging  effects on fish and  wildlife.
Within the treatment system, key parameters include DO,  COD, and turbidity.

     This section focuses on studies which have dealt with BODS  toxicity,
and nitrogen reduction.


BOD: A QUESTION OF RELIABILITY

Abbott Laboratories, North Chicago, Illinois

     Nedved, Bergmann and Comens of Abbott Laboratories  (75) defend the
usefulness of the Biochemical Oxygen Demand (BOD) test if the nature of
the analysis is appropriately interpreted and understood.  The Abbott
Environmental Pollution Abatement and Control Group felt that:

     1.   The standard BOD test, when carefully and conscientiously
          performed, is significantly reproducible

     2.   The test provides extremely valuable information in a  ...
          variety of ways and can be applied to many  ... practical
          problems

     3.   In general, there is no other single analytical parameter
          applied to water analysis that even approaches the information
          value of a standard BOD determination

     4.   The BOD test should not be used independently as a basic pa-
          rameter, but it should be augmented with companion analyses
          and admittedly subjective evaluations.

     The Group enumerates a number of specific uses for data derived from
the BOD test, including determination of toxic and inhibitory effects of
waste materials.  They point out that the "logical and uncomplicated ap-
proach is to accept the BOD test for what it is -- an indirect indication
of biological activity measured under somewhat controlled conditions by
the uptake of oxygen in unstable, 'individualistic' biological systems."

     The Abbott personnel indicate the limitations of the test must be
recognized and accepted.   A major disadvantage of the test is that it
requires 5 days for completion.  The ASTM has described  additional

-------
64
limitations of the test as:  a) BOD,- cannot be considered as a quanti-
itative expression without an approximation of the rate of oxidation and
the ratio of BOD5 to ultimate oxygen demand;  b) the BOD,- values of
different industrial wastes are not (necessarily) additive; and c)  the
efficiency of a biological treatment process may not be accurately
determined on the basis of a BODginfluent and effluent.  Abbott states
it was not the intent of its paper to deny or dilute the limitations of
the BOD test, but rather to promote the advantages of the BOD analysis
within the confines of the known limitations.

     The BOD analytical procedure was studied in detail, including  quan-
tification of the seed used in the BOD test since pharmaceutical samples
generally required seeding.  The seed was obtained from the Abbott  sani-
tary sewer, settled and gauze-filtered before use.  A study was conducted
on a formulated sample of acetone, sodium lauryl sulfate, sodium propionate,
and water.  The theoretical TOC was calculated to be 125,728 mg/1 and the
theoretical TOD (^ ultimate BOD) was 444,022 mg/1, which gives a calculated
TOD to TOC ratio of 3.54.  When conducting BOD on replicate samples at four
concentrations, the results shown in Table VI-1 were obtained.
                              Table VI-1

              BODg Determinations of Acetone Mixture (75)
Sample
Level
Concentration
(%)


BOD
(mg/1 )
BOD
Replicate
Averages
(mg/1)
Replicate
Standard
Deviation
(mg/1)
     0.003
     0.002
     0.0015
     0.001
170,327
169,640
183,820

335,980
329,160
335,460

333,950
352,830
339,350

361,260
365,300
371,380
174,595
333,533
342,043
365,980
+6,529
+3,100
+7,940
+4,182

-------
                                                                      65


     As the concentration of the sample increases,  the  BOD decreases,
first gradually between the 0.001  and 0.002 percent concentrations
(attributable to slight inhibitory effects), and  then  substantially
between the 0.002 and 0.003 percent concentrations  (attributable  to
marked inhibitory effects).  This type of inhibitory behavior was com-
monly observed for many chemical process waste streams.

     From these results, it was speculated that the BOD of an extremely
dilute acetone mixture sample would be about 400,000 mg/1, compared to
the calculated theoretical TOD of 444,022 mg/1.  Conversely,  an acetone
mixture of 0.004 percent or greater concentration,  would show a high
level of toxicity approaching a zero BOD^ value.

     The following conclusions were drawn:

          1.  One of the major reasons for "perpetual  misinterpre-
              tation" of the results of a BOD analysis is  the futile
              and unnecessary search for an absolute definition of
              BOD.  A universally acceptable definition is highly un-
              likely, and this very probability adds to the attractive-
              ness and value of the test.  Since the test  uses a  dynamic
              system, the response is necessarily difficult to interpret.
              The test also suffers from over-study and misinterpretation,
              which creates undue preconceptions.  This test  does not  pro-
              vide an absolute measurement, but rather (according to the
              authors), it should be construed "as a subtle suggestion, a
              reflection, of an oxygen demanding property of a sample".

          2.  BOD values should not be utilized independently, but  TOC,
              TOD, and COD should be obtained to compliment BOD results.

          3.  A subjective evaluation of the composite information  and
              optimistic interpretations of the BOD results are justi-
              fied on the basis that, at the present time, no other pro-
              cedures can adequately supplant the BOD test (75).
Dorsey Laboratories. Lincoln, Nebraska

     Anderson (2,38), in 1968-69, studied the character, treatability
and reduction in pollution loads passing through the Dorsey extended
aeration treatment facility.  Toxic constituents were thought possible
in the wastes, and consequently BOD samples were seeded with specially-
developed mixtures of river water and sewage, taken from the city of
Lincoln and the Dorsey Labs' treatment works.

     Besides conventional 5-day BOD's, samples were prepared for
"continuous monitoring".  Data thus obtained was used to correct for
time lag and nitrification.  Even with seeding, lag periods up to 2
days were encountered in certain samples.  No chemicals were added to
inhibit nitrification.  Corrections for nitrification were necessary on
a number of the BOD samples.  The BOD value was found to increase with
dilution, indicative of a waste containing toxic or inhibiting substances,

-------
66
     The authors emphasize the difficulty of obtaining reliable BOD data
for industrial wastewaters, especially when inhibiting substances are
present and/or the characteristics of the waste stream are highly variable.


Experiments in Great Britain

     Based upon his experience in Great Britain, Lines (10) has observed
that apparent BOD's obtained from using fully acclimated seed may be
more than three times greater than normally determined values.  This was
found even for fermentation wastewaters, which are relatively biodegrad-
able although strong.

     Micrbbial breakdown by organisms may be impeded either by the
nature of the organic materials or by antibiotic residues.  If oxytetra-
cycline, an antibiotic residue, is present in the raw wastes, a possible
counter measure is treatment with ferric chloride, whereby an inactive
complex is formed through the OH" group.
TOXICITY

     In the handling and disposal of pharmaceutical effluents, toxic
chemical compounds are frequently encountered.  Certain toxic chemical
compounds are rated as extremely dangerous physiologically to humans and
animals.  Other compounds may seriously impair biological growths and
metabolism within waste treatment systems.

     Howe, 1961 (62), reported on experimental studies with potent anti-
biotics, a phenol-mercury substance, a selected hormone, and substances
containing formaldehyde and methyl alcohol.
Antibiotics

     Certain antibiotics are bactericidal whereas others are only bac-
teriostatic.  Bactericidal refers to actual killing or splitting of bac-
terial cells.  Bacteriostatic refers to the properties of inhibition,
that is decreasing, or halting, the growth of bacterial forms.  No
claims are made regarding fungal or viral forms.  A very small dosage of
antibiotics may inhibit specific organisms, but a higher concentration
is generally necessary to inhibit biological growths in an oxidation
system because of the diversity of organisms present.

     When the antibiotic concentration in a biological oxidation system
reaches the toxic level, the biological floes disperse or disintegrate.
The microbial population may become acclimated to the antibiotic, par-
ticularly if a carefully controlled amount of the chemical is contin-
ously introduced; but then again, the system may not acclimate at all.
Also, the system may acclimate to one antibiotic, but may fail when a
second antibiotic is introduced.  Table VI-2 presents concentrations of
antibiotics known to be inhibiting to various bacterial forms.

-------
                                                  TABLE VI-2

                                     Minimum Concentrations of Antibiotics   ,
                               Required for Complete Inhibition of Organisms- (62)
Affected Organisms
Ps. Aeruginosa
B. Proteus
Esch. Coli
A. Aerogenus
K. Pneumoniae
Penicillin
(mg/1)
Profuse Growth
10-1,200
38-6,000
375-6,000
5-600
Streptomycin
(mg/1)
50-1,000
12.5-1,000
6.3-1,000
0.8-1,000
0.8-1,000
Aureomycin
(mg/1)
50-1 ,000
12.5-400
3.1-25
6.3-200
6.3-50
Chloramphenicol
(mg/1)
200-1,000
3.1-50
3.1-25
3.1-50
1.6-25
Ilotycin
(mg/1)
250
250
62.5
--
62.5
a/ Original  data from Frank, Wilcox,  and Finland (1950)  and from Wick, Eli  Lilly & Co.  (1961).   The word
   "complete" is not defined by Howe  (62).   From 18 to 25 strains of bacteria utilized  for each numerical  entry.

-------
68
     Lab experiments were conducted on the effects of three antibiotics
on small-scale activated sludge reactors.  Total solids content in the
systems was maintained at about 8,000 mg/1 with acclimated sludges.
Nutrients were added, the pH level was maintained at 7.4 to 7.6, the
temperature was 32C (90F), and air was supplied. Antibiotic A was
introduced to the system at an initial concentration of 210 mg/1, and the
aeration system operated for 24 hr.  Results showed that the biological
system performed satisfactorily, and at the end of 24 hr, there was 93
percent BOD removal and almost 93 percent removal of the antibiotic
(determined by spectrophotometric assay).  Initial and final BOD values
approximated 4,200 mg/1 and 300 mg/1, respectively.  BOD reduction was
slow in the first 4 to 8 hr, after which the removal rates increased
significantly.

     The second antibiotic was a wide spectrum type known as Antibiotic B.
This substance was introduced to the system at an initial concentration of
900 mg/1, and pH 7.8 was maintained.  Over a 12 hr aeration period,
measurements showed little or no decrease in antibiotic concentration, no
removal of BOD, and not only a decrease in microbial population but also
a general dispersion of the biological floes.  A second test was run using
an initial concentration of 200 mg/1 of Antibiotic B and the mixture was
aerated for 24 hr at pH 7.4.  In this case, the antibiotic remained
stable over the first 4 hr or so but then it was reduced about 94 percent
over the next 16 to 20 hr.  The microbial population increased, after an
initial setback during the first 8 hr.  The BOD was reduced 96.7 percent
from 6,200 mg/1 to 200 mg/1.

     Antibiotic C was a relatively powerful agent known to be stable
under any pH condition.  This antibiotic was added to the aeration
system at an initial concentration of 400 mg/1.  After 12 hr the system
showed no reduction of BOD or antibiotic activity.  Furthermore, the
microbial population declined and was converted into dispersed forms.
Fresh activated sludge previously acclimated to Antibiotic C in another
series of experiments was added to the system.  After about 8 hr, the
system started to function normally, and over 24 hr of full aeration,
the antibiotic was reduced 96 percent, and BOD was reduced 94.5 percent
from approximately 5,500 mg/1 to 300 mg/1.

     The above experiments on biological oxidation of antibiotics demon-
strate that biological growths can be significantly, if not totally
disrupted by various antibiotics.  However, remedial measures in accli-
mating the system and achieving recovery are also demonstrated.

     Both Antibiotics A and C are described as "macrolide" antibiotics,
produced by two different species of Streptomyces organisms.  The chemi-
cal features of this class of antibiotics are a large lactone ring, a
ketone function, and an amino-sugar attached by glycosidic linkage.
Antibiotic A contains a phenoxymethyl group.  Biological degradation is
thought to proceed by oxidation and hydrolysis of the antibiotic.  When
excessive amounts of these macrolide antibiotics are present in the
aeration system, the antibiotics and their degradation derivatives-are
both toxic to certain micro-organisms.  Some organisms may be able to

-------
                                                                      69


acclimate to the antibiotic(s), lowering antibiotic concentration, whereby
levels are favorable" for continuous functioning of the biological  treat-
ment system.  Unfortunately, no data was obtained relative to full-scale
systems and their adaptability and/or vulnerability to the dynamic sit-
uation in an ongoing pharmaceutical manufacturing installation.


Phenol-Mercury Compound

     The particular phenol-mercury studied was a highly bactericidal
compound of sodium ethylmercury thiosalicylate (Thimerosal, Lilly) and
sodium o-phenylphenate (Lilly).  This compound is used both for disin-
fection and sterilization.  A wastewater containing this compound re-
ceived treatment.  This waste was found to contain 6.9 mg/1 mercury, 286
mg/1 total phenylphenate, 15,000 mg/1 emulsified oil, 83,400 mg/1 total
solids, and 160,000 mg/1 BOD with a pH of 7.9.

     The experimental treatment consisted of two alternate systems.
Primary Treatment A involved breaking the oily emulsion of the phenol-
mercury waste by sulfuric acid treatment at pH 1.5.  Primary Treatment B
consisted of acid cracking, plus lime-aluminum coagulation together with
bentonite and sludge additions to aid coagulation, followed by settling.
These two effluents were then aerated 8 hr employing acclimated, acti-
vated sludge.  The results of the two treatment systems are shown in
Table VI-3.

     After chemical treatment and settling, the phenol-mercury waste was
successfully treated by activated sludge with about 99.9 percent re-
duction of phenylphenate, and the mercury concentration decreased from
6.9 mg/1 to 0.05 mg/1.  However, data are lacking on the conventional
parameters to sufficiently assess the biological oxidation capabilities
of these systems.  In the biological degradation of the phenol-mercury
compound, breakdown of the sodium ethylmercury thiosalicylate is thought
to result from oxidation and/or reduction cleavage of the compound with
attendant release of mercury from the organic structure.  The mercury
can be oxidized to HgO and/or Hg203<  Sunlight is also believed respon-
sible for dissociation of mercury.  The sodium thiosalicylate and the
ethyl components are subsequently broken down into the intermediates
such as benzoic acid, succinic acid, acetic acid, etc., and eventually
into the simple end products.

     Table VI-4 gives the results of additional experiments in which
varying concentrations of raw waste and Primary Treatment A and B ef-
fluents were fed into the activated sludge process and aerated for 8 hr.
The activated sludge process could tolerate but not successfully treat
the given concentrations of raw emulsified waste.


Hormone

     The next phase of investigations involved the experimental degra-
dation of the female sex hormone, DES, by aerobic and anaerobic treatment.

-------
                                             TABLE VI-3
                        Effluent Quality From Experimental Treatment Systems
                                Receiving Phenol-Mercury Wastes (62)
Emulsified
Parameter Raw Waste
PH
TS
TDS
Total Hg
Total
Phenyl -Phenol
(C6H5C6H4OH)
TL , 48 hr
T% cone.)
7.9
83,400
8,600
6.9
286
0.3
Primary
Treatment A
1.5
26,900
21,000
0.85
14.0
0.5
Primary
Treatment B
6.7
7,740
5,000
0.02
23.7
12.0
Primary Treatment A
+ Activated Sludge
7.4
1,600
1,320
0.05
<0.005
b/
Primary Treatment B
+ Activated Sludge
7.3
1,350
1,210
0.05
<0.005
b/
a/ Data given in mg/1, except pH and TL .
b/ No toxicity recorded.               m

-------
                                             TABLE VI-4

                                    Activated Sludge Treatment of
                     Raw and Chemically Pre-treated Phenol-mercury Effluent (62)
Waste
Concentration
Initial BOD
Final BOD
Toxic to
test fish
Emulsified
Raw Waste (mg/1 )
1% 2% 5%
2,800 5,700 9,100
1,700 3,700 8,100
Yes
Primary
Treatment A (mg/1)
1%
400
210

2% 5%
700 1,620
450 820
No*/
Primary
Treatment B (mg/1 )
1%
250
110
--
2%
590
270
No*/
5%
1,100
550
--
a/ At 40% concentration, both A and B were toxic to fish.   Unfortunately,  methodology not explained
   in arriving at toxicity observations.

-------
72


DES in very low concentrations is highly bactericidal.   The limit for
instantaneous contact with DES is 0.22 g/kg (0.1 g/lb)  of carrier material
and, therefore, extreme care was taken in handling this waste.   A 3.8
liter (1 gal) chamber was used for the activated sludge studies and a
3.8 liter container for the digestion experiments.  DES was mixed into
a carrier and this waste mixture was placed into the aerator in the
ratio of 1 volume waste to 6 volumes activated sludge.   Activated sludge
treatment was conducted for 14 continuous days, the results of which are
shown in Table VI-5.

     Treatment required an unusually long time, although the DES reduc-
tion was 94 percent after 7 days and 99 percent over 14 days.  When the
initial concentration of DES in the feed to the aerator was lowered from
204 mg/1 down to 5 mg/1, reduction of DES over 24 hours of aeration was
measurably improved.  However, even at these low concentrations, some
inhibition of activated sludge growths was evident.  The experimental
data suggested that perhaps the strong DES wastes could be treated in a
long-term aerated lagoon, and the diluted waste could be treated by
activated sludge.  However, neither approach appeared feasible because
hygiene practices and safety precautions associated with handling of DES
from the batch production program were considered too strict and time-
consuming.

     In the anaerobic digestion experiment, the. DES waste was added to
the digestion chamber in the ratio of 1 volume waste to 2 volumes di-
gested sludge.  Temperature was maintained at approximately 32C (90F)
and digestion was conducted for 14 days.  After 7 days of digestion, BOD
and DES removals were respectively 83.9 and 86.0 percent.  After 14
days, BOD reduction was 94.6 percent and DES reduction was 99.0 percent.
These results were comparable to those obtained by activated sludge
treatment of the strong DES waste for 14 days.


Formaldehyde and Methyl Alcohol

     Howe (62) also conducted limited evaluation studies on the treat-
ability of formaldehyde and methyl alcohol wastes by activated sludge.
Activated sludge and the waste material were added to a 3.8 liter (1 gal)
aeration chamber with the MLSS maintained at 6,000 mg/1.  In separate
experiments, the initial methyl alcohol and formaldehyde concentrations
in the aerators were 1.05 and 0.25 percent, respectively, and aeration
was conducted for maximum periods of 6 hr.  Methyl alcohol was analyzed
by the refractometer from the distillate of each sample.  Formaldehyde
was determined by the spectrophotometer.  The methyl alcohol level was
reduced 81 percent after 2 hr, and the formaldehyde was reduced 87
percent after 4 hr.   The organics may have been partly removed via
volatilization during aeration.  No toxicity effects were observed in
these test runs.

     The biochemical decomposition of formaldehyde may proceed two ways.
The formaldehyde may combine with protein molecules in the sludge and be
removed by clarification, or the formaldehyde may be oxidized into

-------
                      TABLE VI-5



Activated Sludge Treatment of Hormone (DES) Wastes (62)
Aeration
Period
0 days
7 days
14 days
204 mg/1 DES
TS
pH (mg/1 )
7.6 50,000
8.4 27,900
8.5 13,200
Waste; 14 Days Aeration
TDS
(mg/1)
45,000
10,800
5,900
DES
(mg/1)
204
10
2
DES
Reduction
(%)
__
94.0
99.0
BOD
(mg/1 )
28,000
3,000
1,000
BOD
Reduction
(%)
__
89.3
96.4

Run Number
1
2
3
4
5 mq/1 DES
Initial DES
pH (mg/1 )
8.2 5
8.4 4
8.5 3
8.6 1
Waste; 24 hr
Final DES
(mg/1)
1.7
1.2
0.8
<0.2
Aeration
DES
Reduction
(%)
66.6
70.0
73.6
>80.0
Initial BOD
(mg/1)
18,100
17,000
15,800
14,600
Final BOD
(mg/1)
3,100
2,800
2,400
2,100
BOD
Reduction
(%)
82.9
83.5
84.8
85.6
                                                                                          oo

-------
74
formic acid and methanol as the intermediates, which in turn break down
into carbon dioxide and water.  The biochemical decomposition of methyl
alcohol is thought to be somewhat similar to that for formaldehyde.
NITROGEN REDUCTION

     NPDES discharge permits for bulk pharmaceutical  manufacturing in-
stallations are focusing on necessary and substantial reduction of
ammonia and organic nitrogen in the associated wastewaters.  Waste
streams from these installations, even after reasonably high levels of
biological treatment, may still contain from a few hundred to a few
thousand mg/1 of Kjeldahl nitrogen.  Ammonia and organic nitrogen loads
may substantially exceed BOD loads in the final discharges and range
from 454 kg (1,000 lb)/day up to 3,630 kg (8,000 lb)/day.

     Impacts due to nitrogen compounds include': the oxygen demand of
nitrogen compounds; ammonia toxicity to fish and wildlife; increased
chlorine demand due to ammonia and related compounds; formation of
chloramines and similar constituents during chlorination.which are
decidedly toxic to fishlife; off-tastes to water supplies; and health
problems to humans and animals.

     Also, with pharmaceutical wastes, even moderate levels of unoxi-
dized nitrogen can give varying and possibly distorted results in run-
ning the BOD test.  Regulatory authorities are now initiating standards
for nitrogenous waste materials in streams and effluents.  Treatment
facilities will increasingly be required to produce effluents low in
both unoxidized nitrogen and total nitrogen.

     Loehr (117) summarizes available management methods in reducing the
nitrogen content of wastewaters as shown below.  Other important refer-
ences include Adams (118) and Ehreth and Barth (120).
            Treatment
Nitrogen Compounds Removed
      Physical and Chemical
        Land Application
        Electrochemical
        Ammonia stripping
        Ion Exchange
        Electrodialysis
        Reverse osmosis
        Breakpoint chlorination
      Biological
        Nitrification
        Algal  utilization
        Microbial  denitrification
        Land application
NH3, NHU+, organic N
NH4+
NH3
N03-, NHu+
N03-, NH4+
N03-, NIV
NHit+, organic N
NH4+ (NH3) conversion to N02- and N03-
All forms
N03-, N02-
All forms

-------
                     VII.  WASTE RECOVERY AND CONTROL
      In general, wasteloads from pharmaceutical manufacturing operations
can be reduced either by recovery of valuable solvents, or by recovery
of certain components in the form of animal feed.  Both types of in-
plant control measures reduce the raw wasteload to treatment facilities.
Examples of each type of control are given below.


SOLVENT RECOVERY

     Solvent pretreatment has a twofold objective:  to recover solvents,
and to favorably impact the waste disposal problem.  This pretreatment
lowers the BOD and reduces toxicity of the spent liquors.  As a result of
the heat of distillation, the character of these wastes can be modified
by the formation of solids and tarry compounds.  Ultimate disposal of
the most deleterious wastes is simplified by concentrating them in as
small a waste volume as possible.  Weaker wastewater from other sources
in the plant can then be treated or disposed of by more conventional
methods (13, 59).

     The impact of solvent wastes on waste treatment facilities is
significant because of the high BOD levels.  As an example, a commer-
cially used ketone solvent has been reported (25) to yield a BOD level
of 2,000,000 mg/1, about 9,000 times stronger than normal untreated
domestic sewage.  A gallon of this solvent is equivalent to raw domestic
wastes from a city of 77,000 people.  A second solvent, amyl acetate,
was reported to have an approximate BOD of 1,000,000 mg/1 (32,65).
Thus, a small variation in solvents lost to waste streams has a great
impact upon a treatment plant and/or receiving watercourse.

     Among the methods used in recovery are single- or multi-effect
distillation units.  In other cases, solvents may be incinerated.
Molecular sieves have also been used in purifying spent organic sol-
vents (6).

     Solvent recovery not only represents savings in cost of raw materials
to the production department but also means considerable savings in
waste disposal costs.  Automated monitoring has been recommended on all
effluent lines especially for developing a material balance analysis.
Monitoring is particularly important on waste streams originating from
the solvent recovery areas.  The monitors can be wired into an alarm
system to immediately detect and warn of malfunctioning processes,
surges, spills, or other accidental discharges (25).

     At Bristol Laboratories, Syracuse, N.Y.  in the early 1950's, sol-
vents were recovered by two methods (65).  Solvents and fermentation broth
were passed downward through a 1.8 m diam. by 9.1 m high (6 by 30 ft)
stripping column.  The broth percolates down  and around a series of plates
and bubbler caps.  Live steam enters the bottom of the column.   Distillates

-------
76
off the top of the columns are condensed for solvent recovery and spent
broth leaves the column at the bottom.  The second system for solvent re-
covery at Bristol Laboratories is a combined stripping and rectifying col-
umn from which it was possible to remove all but. 0.25 percent of the water
from the spent solvents. -
BYPRODUCT RECOVERY - ANTIBIOTICS PRODUCTION

     Feasibility studies have been made for recovering spent fermen-
tation broths and mycelia from antibiotic production (10).  At least
some of these products have been shown to contain unidentified growth
factors for animal feeding.  Recovery is accomplished by filtration,
evaporation, and drying; the processed material is then sold as animal
feed.  While use may be limited in certain cases because of undesirable
trace amounts of antibiotics, this material has been shown to be a rich
source of vitamins, especially EL and B,2-  A typical analysis of dried
penicillum mycelium is given below (10).
  Content          Percent                 Content
Moisture
Crude Protein
Fat
Fiber
Ash
Carbohydrate

8
32
7
7
20
26

Aneurin
Nia'cin
Pantothenic acid"
Riboflavin
Choline1
Pyridoxin
Biotin
6
7
64
37
3,700
13
5
     A typical scheme developed for recovery of antibiotics production
wastes is that based on a feasibility study conducted for a waste
recovery system to be installed on a new antibiotics plant operated by
Abbott Laboratories in Barceloneta, Puerto Rico (22).  According to this
scheme, fermentation beers would be segregated and solvents and anti-
biotics stripped.  It was believed that virtually all the mycelium,
solvents and antibiotics could be recovered.  After a thorough study of
alternate treatment and disposal methods, the company decided to complete-
ly segregate the spent fermentation beers.  Subsequently, this bulk
material would be concentrated by triple-effect evaporators to 30 per-
cent solids, or higher (a method similar to that used at Abbott's North
Chicago, 111. facility).

     Immediate plans called for handling this concentrated material on-
site in an odor free, refractory incinerator with the exhaust air
discharging into the main boiler stack. .Future provisions called for
adding suitable drying equipment when a. local market "could be developed
for the fermentation solids.  The company stated that concentration and

-------
                                                                      77
incineration would provide for complete destruction of the BOD and
solids in the spent beer streams, and that no conventional secondary
treatment operation could equal this degree of treatment.   Although
Abbott's development plans for Barceloneta did not envision complete
treatment for the overall antibiotics complex, zero waste  discharge was
indicated as being economically feasible for the fermentation processing
areas.  More recent information, however, creates some doubt as to
whether this proposed waste recovery was accomplished.

     As early as 1952, Edmondson (60) described the large-scale recovery
and disposal of penicillin and other antibiotic spent beers by triple-
effect evaporators at Upjohn Company's Kalamazoo, Mich, plant.  In 1951,
when Upjohn expanded its fermentation production facilities, BOD loads
entering the existing trickling filter waste treatment plant rose from
1,530 to 3,600 kg (3,500 to 8,000 lb)/day.  Various antibiotics were
generated from deep vat fermentation, and at least four major types of
antibiotic spent beers were handled.  Most of the residual extraction
solvents were removed by azeotropic distillation before evaporation.
Certain broths were quite low in pH and were consequently neutralized
with 50 percent caustic before evaporation.  The spent beers were stored
to equalize loads onto the evaporator.  Excess beers not capable of
being handled in the evaporator system were sent to existing biological
treatment.  Certain of the antibiotic spent beers contained substances
reported as toxic to the biological system, and therefore evaporation
was intended to partially minimize impact on the biological treatment
works.

     The spent beers from the equalizing tank with a solids content
between 2.5 to 3.5 percent were received into the triple-effect evap-
orator at the rate of 9.5 to 11 kl (2,500 to 3,000 gal/hr.)  Barometric
condenser water from the third or last effect of the evaporator could be
recycled as cooling water.  The condensates from the first effect were
sent to the boiler house.  Condensates from the second and third effects,
possibly contaminated from entrainment or carryover, were generally
discharged to an aeration tank and then to the trickling filter plant.
The multiple effect evaporator system for handling spent beers at the
Kalamazoo plant was put on-line Dec. 9, 1952.  The solids concentration
leaving the evaporators, found to be a function of the type of beers
entering the system, varied from 15 to 35 percent.

     The BOD reductions directly associated with the triple-effect
evaporation system were 96 to 98 percent for the four types of anti-
biotic spent beers.  Residual BOD in the condensate streams varied from
90 to 450 mg/1.  This organic BOD was substantially attributed to
solvent carryover in the condensate streams.  For example, Antibiotic A,
which was not stripped prior to evaporation, showed the highest BOD of
450 mg/1  and the least BOD reduction of 96 percent.  Initiation of
solvent recovery and improvement of evaporation procedures would help
reduce the residual BOD.  Total solids in the condensates  were rela-
tively low (75 to 150 mg/1).  There was concern that the carryovers
could contain some trace antibiotic activity, thereby depressing bio-
logical  growths on the subsequent trickling filters.  "Checks" for

-------
78
antibiotic activity, run both on the condensates and the concentrates,
showed that practically all of this activity was carried down in the
concentrates (60).

     A cost analysis was made of initial  capital investment,  -plus the
annual operating costs of the evaporator system vs conventional  biologi-
cal treatment in handling the antibiotic beers.  In this case, no credit
was given for potential profit return on a saleable byproduct.   Biolo-
gical treatment was characterized by a much higher initial  cost  and
lower operation and maintenance cost compared to the evaporator  system.
By decreasing the initial investment with the evaporator system, the
standby cost is reduced, and the operating cost varies somewhat  more
proportionately with the volume of beer production.  This situation
would seem to provide the high degree of production flexibility  gen-
erally required in pharmaceutical manufacturing plants.  With a  lower
initial cost for the evaporator, the capital cost savings could  be
invested in a secondary profit-deriving process, producing an annual
profit.

     Edmondson (60) in 1953 indicated it should be possible to find a
market for this concentrate material, thereby reducing annual cost of
the evaporator system.  The concentrate, found to be a rich source of
protein and vitamins, was thought suitable for animal feed either as a
dry product or a syrup.  Annual cost of the evaporator system was about
$0.01/gal spent beer if operated only one shift per day, but decreased
to $0.0058/gal if operated over three shifts.  Annual cost was otherwise
estimated as $0.06/lb BOD removed.
BYPRODUCT RECOVERY - TECHNOLOGY TRANSFER FROM RELATED INDUSTRY

United States

     Blaine and Van Lanen (42) describe techniques developed by the
distilling and brewing industries in the 1950's and early 1960's for
achieving practical waste abatement.  These techniques, they felt, may
be adaptable also to the newer branches of the fermentation industry,
including Pharmaceuticals.  The "entire plant" approach has been used
successfully by the distilling industry in attaining desired objectives.
First, an inventory of waste sources is made, which when totalled
approximates the findings of waste loads in the combined plant effluents,
Such an inventory provides assessment of the relative costs of various
waste abatement alternatives, including in-plant controls.   It also
demonstrates the cost or profit involved in major production rearrange-
ments, such as changing types and amounts of primary products and
byproducts that can be generated from a given quantity of raw materials.
When BOD, COD and other analyses are diligently applied to  each process
unit, an accounting of waste is also an accounting of product loss.

-------
                                                                      79
     At a Hiram Walker distillery, substantial  waste abatement has
been made entirely within plant, without the use of so-called  conven-
tional treatment.  Wasteloads were reduced from 5.6 PE BOD/bu  of grain
processed in 1951-52 to 2.8 PE BOD/bu grain in  1959-60, and this was
expected to drop to about 1.2 to 2.0 PE BOD/bu  in the early 1960's.
A 56 Ib bushel is the commonly accepted unit of production in  the dis-
tilling industry.

     The basic processes used by Hiram Walker are described as follows.
Grains are received, milled, slurried, then pressure-cooked in eight
49 kl (13,000 gal) cookers.  The mash is cooled in the cookers by vacuum
blowdown, saccharified with malt, and further cooled in a five-effect
vacuum cooler.  After yeast is added, the mash  is sent to a battery of
twenty-four 380 kl (100,000 gal) fermenters for 55 to 90 hr.  The fer-
mentation beer, containing the spent grain slurry and about 7  percent
alcohol, is pumped to the stills.  After distillation, the spent fer-
mentation beers are sent to the feed recovery plant.  This whole still age
contains about 25,000 mg/1 BOD, 8.3 Ib BOD/bu (^ 50 PE BOD/bu), 5 to 7
percent solids  (17 to 19 Ib/bu), and a unit volume of 40 gal/bu grain
(42).

     The whole still age is screened, then the screenings are pressed and
dried in rotary driers.  Seven rotary indirect  steam tube driers provide
a total finished product capacity of 10 to 12 tons/hr of dried light
grains.  Press liquors are returned to the feed recovery process.

     The thin stillage passing through the screens is centrifuged in
twelve basket centrifuges.  Then it is evaporated using two parallel
sets of triple-effect evaporators and finishing pans.  Evaporated syrup,
at about 35 percent solids, may be dried directly on drum driers to
produce distillers solubles.  Otherwise, it is  mixed with the  screened
grains and sent to rotary driers to give a product called distillers
dried grains with solubles.  The following table summarizes the Hiram
Walker in-plant waste survey, completed in 1953.
                                             PE BOD/1,000 bu       (%)
Process Area                                     mashed
Cooking and fermenting, incl. cleanups
Distillation
Feed recovery plant
Powerhouse
Total
371
55
2,510
73
3,009 = 3 PE/bi
12.3
1.8
83.4
2.4
j 100.0

-------
80
     In 1954, of the 2,510 PE produced from the feed recovery plant,  945
PE were eliminated by completely recovering and converting into feeds
the solids from the drum dryer vapor scrubber and from the dust col-
lector on the dried grains dryers.  The remaining 1,565 PE were attri-
butable mainly to condensates from the triple-effect evaporators.   The
company stated these wasteloads can be reduced at least 50 percent if
the existing equipment is operated under tightly controlled flow and
temperature conditions.  Residual PE from the feed recovery plant  is  now
estimated at 750 PE, or less.  The residual with these changes is  esti-
mated as about 1.25 PE BOD/bu grains processed (42).

     Without feed recovery, waste materials represent 50 to 55 PE
BOD/bu.  Recovery of screenings reduces the wasteload to 40 to 45  PE,
whereas drying of total stillage leaves less than 2 PE BOD/bu, or  a 96
tc 97 percent waste reduction accomplished by good in-plant control and
recovery procedures.  Blaine and Van Lanen describe various feed re-
covery operations as having produced net profit and in most cases  gen-
erating enough revenue to have paid for capital equipment costs.  Annual
market value of feed byproducts from the entire U.S. grain distilling
industry as far back as the 1950's. was estimated at $20 million annually,
which undoubtedly has increased greatly through the present (42).
England

     Based on experiences in Great Britain, Jackson (48) compares the
various fermentation industries and examines the status of each.  The
fermentation industries are those in which micro-organisms such as
yeasts, molds and bacteria are either used under closely controlled
conditions to produce desirable end products, or are incorporated into
processes where their presence contributes significantly.  Apart from
the brewing, wine and distilling industries, Jackson defines other
activities as including the production of antibiotics, bakers'  yeast,
organic acids, solvents, cheese, silage, and processes such as  the
retting of flax.  The effluents from the distillation of whiskey contain
about one-third of the organic matter originally used in fermentation,
whereas for antibiotics the proportion is much higher.  Breweries and
wineries are in a more fortunate position, since most of the raw ma-
terials used in these industries eventually end up in the bottle or cask
with less potential  wasteloads in the final effluents.  From many of the
fermentation processes, the micro-organisms are simply released into the
effluents, and one example is cited of yeasts being freely wasted from
the production of alcohol.

     From malt and grain whiskey production, waste is reduced signifi-
cantly by recovering the solids which otherwise would be present in the
effluents.  The husk and coarse insoluble matter from the mashing stage
is sold as "wet grains," at about 80 percent moisture content,  or

-------
                                                                      81


 subsequently dried for lower transportation costs.  Additional solids
 can be recovered from grain-spent washes by filtering through fine
 screens followed by evaporation of the liquors.

     Solids recovery produces greater economic return when the opera-
 tions are relatively large and the price for dried materials approxi-
 mates or exceeds production cost.  The solids in malt spent wash are
 nearly twice those in grain spent washes, which indicates greater po-
 tential revenue return for malt whiskey distilleries.  However, Jackson
 mentions that malt distilleries are considerably smaller than grain
 distilleries.  Evaporation of malt whiskey spent wash in Scotland has
 been determined as economical only if a central evaporation plant ser-
 ving many distilleries is fed a minimum of 1,500 kl (400,000 gal) of
 spent wash per week.  Most grain distilleries are large enough to give
 an economical return on evaporation.

     Since 1966, the economics of spent grains recovery have generally
 become much more favorable.  Byproduct conversion of only these waste
 streams does not, however, represent an adequate solution in many cases.
 Other wastes, including the evaporation condensates, steep water, 'spent
 lees, and miscellaneous washes must also be treated, converted to
 suitable byproducts, or otherwise safely disposed of.  The subject
 byproducts are a good source of protein and contain appreciable amino
 acids, B vitamins, and unidentified growth factors for compounding into
 desirable animal feedstuffs (48).

     In the production of antibiotics and vitamins by fermentation,
 yields of the desired values are quite small and therefore the potential
 wastes are nearly equal to the quantity and strength of the incoming raw
 materials.  Jackson (48) states if the bacterial cells or fungal mycelium
 are not contaminated with filter aids, it is possible to recover and
 use this mycelium for animal feeding either in the wet or dry form.
 Penicillin wastes have been found to contain valuable growth factors.
 Mycelium and likewise evaporated and spray-dried solubles from peni-
 cillin, riboflavin, streptomycin and vitamin B12 fermentations have been
 used for animal feeds or supplements.
WASTEWATER REDUCTION AND RECOVERY METHODS

Bristol Laboratories. Syracuse. New York

     The Bristol Laboratories plant pretreats its fermentation waste
before disposal to the municipal treatment works.  After studying the
plant and the municipality receiving the wastes, the following measures
were recommended to the industry in 1951 (32):

-------
82
          Remove formaldehyde from the industrial wastes to eliminate
          bacterial inhibition within the digesters.

          Install stripping towers for removal of solvents.

          Remove all mycelium.

          Eliminate sources of possible leakage of process materials.

          Conduct a program of sampling and testing solvents in waste-
          water flows.

          Carefully program the dumping of contaminated or spoiled
          fermentation batches.

          Neutralize acid fermentation wastes to protect sewers and
          preclude interference with biological units at municipal
          treatment plant.

          Provide waste stream equalization.

          The manufacturer informs the proper authorities of changes in
          process operations so that careful consideration could be
          given to effects upon the municipal treatment facilities.

     Bristol Laboratories incorporated many of these measures and
reduced the wastewater volume by installing a recirculating cooling
water system.  This reduced the~need for new water in this circuit from
7,570 m /day (2.0 mgd) to 150 m /day (40,000 gpd).  Water consumption
was reduced in the following areas:
          Location                           Volume Reduction
                                             (1/min)(gpm)
Air conditioning condensers
Vacuum pump cooling
Nash pumps
Vacuum ejectors
Bleed steam condensation
Total Reduction
189
76
95
246
757
1,360
50
20
25
65
200
360

-------
                                                                      83
     Water reduction on the air conditioning units was the result of
converting to tower water for condensing, and also routinely cleaning
the (fouled) heat transfer surfaces in the evaporative coolers.   Pumps
and ejectors were converted to tower water for cooling and condensing.
The volume of steam bleed condensation waters was reduced by installing
control valves on the condensing water headers leading to the various
bleeds required in sterilization.
Upjohn Laboratories, Kalamazoo, Michigan

     [See Section VIII:  WASTE TREATMENT AND DISPOSAL:  OTHER TREATMENT
METHODS.].

-------
                  VIII.  WASTE TREATMENT AND DISPOSAL


     During the 1950's, there was substantial increase in the manu-
 facture of fermentation product Pharmaceuticals1 such as vitamins,
 steroids and sex hormones, and antibiotics.  These products created
 complex waste treatment problems for the industrial facilities and the
 municipal treatment works which accept these pharmaceutical wastes.

     Pharmaceutical plants generally combine the fermentation process
 with the production of synthesized organic chemicals and/or biologicals.
 The combinations result in a complex mixture of wastes, presenting
 unusual and difficult problems.  The waste flows often have extremely
 high BOD and COD and contain toxic chemicals detrimental to biological
 treatment.  And, depending on production scheduling and whether proces-
 sing .is batch or continuous, the wastes can vary widely both in strength
 and quantity.

     The treatment of wastewaters from antibiotic production as of 1961
 was thought to almost exclusively encompass biological methods, a
 situation that has not changed radically in the 1970's.  Kempe (37)
 states that both activated sludge and trickling filters have been
 successfully employed on fermentation waste.  Activated sludge is more
 vulnerable to shock loads, but it is less of a problem where fermen-
 tation specialists are available for good treatment plant operation.
 Nevertheless, severe impact on activated sludge plants is due to widely
 fluctuating loads and the likely presence of antibiotics and toxic
 solvents.  These problems appear less troublesome in trickling filter
 installations.

     Aside from activated sludge and trickling filtration^ other treatment
 methods, such as anaerobic filters, spray irrigation, oxidation ponds, and
 deep well injection have been employed by the pharmaceutical industry.

     Kempe concluded that the wide diversity, in itself, of treatment
 processes used for pharmaceutical wastes attests to the fact that all
 the answers to these waste problems are not yet available.


                           ACTIVATED SLUDGE


 Fermentation

 Design Criteria for Pharmaceutical Wastes

     Eckenfelder and Barnhart (27) give design considerations and criteria
 in activated sludge treatment of pharmaceutical  wastes, particularly
 those originating from fermentation activities.   For wastes containing
 spent broths, nitrogen and phosphorous are usually present in sufficient
quantities for biological  oxidation, but this may not always be the
case.

-------
 86
      BOD loads  to  an  activated  sludge  treatment  system are most frequently
 reported in  terms  of  Ibs  daily  BOD/1,000  ft  aeration basin or Ibs daily
 BOD/lb MLVSS.   At  high  concentrations  of  BOD in  the aeration basin, the
 rate of removal  of BOD  will  be  roughly proportional to the concentration
 of viable activated sludge  present  (MLVSS).  However, at relatively low
 concentrations  of  BOD in  the basin,  the rate of  removal of BOD progres-
 sively decreases and  is more or less proportional to the remaining
 concentration of BOD.   Significant  differences are expected between
 batch treatment vs a  continuous,  completely mixed treatment process.   In
 batch treatment, the  BOD  removal  rate  will.be more or less constant and
 proportional to the MLVSS level  until  a limiting BOD concentration is
 reached.   Below this  point,  the rate of BOD removal will decrease and  be
 proportional to the concentration of BOD  remaining at any time.   In a
 continuous system, the  BOD  concentration  of the  aeration basin is equal
 to that leaving the tank, and the rate of BOD removal will be dictated
 by the effluent BOD concentration.  "Rates of BOD removal must be deter-
 mined experimentally  for  each pharmaceutical waste by developing data
 from laboratory bench studies or  pilot-scale plants.

      In selecting  aeration  equipment for  the activated sludge basin, the
 oxygen requirements of  the  micro-organisms and the transfer characteris-
 tics of the  oxygen into the  wastes  must be known.  At high BOD loading
 levels, oxygen  utilization  rates  as  high  as 120  mg 0?/hr/g VSS have been
 observed.  When  BOD removals are  nearly complete, the utilization rate
 decreases  to very  low levels (5 to  10  mg  02/hr/g VSS).  Total oxygen
 requirements are related  to  the BOD  removed and  the MLVSS concentrations
 by the following expression  which is determined  experimentally in the
 lab or by  pilot-plant setup:

      kg (lb) 02/day = a x kg (Ib) BOD  removed/day + b x kg (Ib) MLVSS.

 High-speed turbine systems  for  transferring high rates of oxygen into
 very strong BOD  wastes  have  been  used  successfully by the pharmaceutical
 industry.

      The excess  or net  sludge resulting from activated sludge treatment
 may be  determined  from  the expression:

      kg  (lb) VSS/day =  a  x kg (lb) BOD  removed/day - b x kg (lb) MLVSS

 where  "b"  relates  to the  amount of sludge  consumed through auto-oxidation.

      Following the aeration  process, the  biological  sludges can be
 separated  from the aeration  liquor by either settling or flotation.   At
 low BOD loadings into the aeration process, settling may be successfully
employed,  but at high loadings, the sludges leaving the aeration basin
are oftentimes highly active and will frequently gasify and float
without benefit of oxygen.   In this case,  pharmaceutical  plants have
utilized dissolved air flotation for solid-liquid separation,  particu-
 larly those plants dealing with spent fermentation  beers.

-------
                                                                      87
     The various processes employed for sludge disposal are thickening
by settling or flotation, aerobic oxidation, anaerobic digestion,
centrifugation, air drying, vacuum filtration, evaporation and recovery,
burning, spreading onto land, landfilling or burial.  Aerobic oxidation
of sludges has been used successfully by pharmaceutical companies.
However, 20 to 50 percent of the sludge is usually reported as resistant
to oxidation which becomes treatment process residue.
 Synthesized Organic Chemicals

 Biological Treatment of Pharmaceutical Chemical Waste

     Chemical manufacturing wastes can usually be treated biologically
 to acceptable standards economically, but extensive study is necessary
 to give satisfactory results (47).  If within a complex effluent there
 are  significant amounts of refractory materials, it is best to separate-
 ly treat these streams by other means than to overload the activated
 sludge system.  Wastes containing phenolics, aldehydes, organic acids
 and  similar constituents can generally be treated economically with up
 to 99 percent BOD reductions.

     A high level of organic removal from chemical wastes using activa-
 ted  sludge treatment is usually achieved only in the lower loading range
 of less than 0.5 kg (Ib) of BOD applied per kg (Ib) of activated sludge
 dry  solids.  At loadings of 1.0 kg (Ib) BOD per kg (Ib) solids, re-
 ductions will often decrease to less than 50 percent.  With more diffi-
 cult-to-treat chemical wastes, the choice must be made between single
 stage vs two-stage aeration.  Two-stage aeration is considered by most
 to represent two or more separate aeration tanks in series, but strictly
 speaking, a system is two-stage only when it involves intermediate
 settling.

     A chemical plant where the pollution loads comprise 70 percent
 citric acid process wastes and 30 percent other pharmaceutical wastes
 had  a BOD of 1,800 mg/1.  This wastewater was treated to yield a final
 discharge of less than 50 mg/1 BOD in^a single-stage activated sludge
 system with a loading of 4.3 kg BOD/m  (270 lb/1,000 ft^j/day.

     A study was made of a chemical plant manufacturing a wide array of
 pharmaceutical products, dyestuffs and intermediate products.  Some of
 the  contaminants were readily degraded but the remainder proved a
 difficult problem for the activated sludge process.  With single-stage
 treatment it was possible only to reduce the influent BOD of 1,000 mg/1
 to 200 mg/1, even with 48 hr aeration.  It was subsequently shown by
 utilizing a two-stage system with 6 hr aeration in the first stage,
 intermediate settling, and 12 hr of aeration in the second stage, that a
 final effluent of less than 50 mg/1 BOD could be expected.

     The oxygen demand of chemical wastes in biological  treatment will
generally vary between 1.1  and 2.2 kg (Ib) oxygen/kg (Ib) BOD removed.
Oxygen demand for pharmaceutical  wastes is thought to be in the range of
1.1  to 1.3 kg (Ib)  oxygen/kg (Ib)  BOD.   Excess biological  sludges

-------
produced from chemical wastes will usually approximate 0.2 to 0.5 kg
(Ib) sludge/kg (Ib) BOD removed.  If the sludges are thickened followed
by chemical conditioning, the mixture can be passed through a rotary
vacuum filter producing a cake having up to  25 to 30 percent solids.
Since land disposal of these and other sludges is becoming increasingly
difficult, incineration is being viewed in a much more favorable manner.

     Burgess reports that the multiple-hearth furnace can provide com-
plete incineration of the organic matter in the sludges, the end products
comprising sterile ash and clean flue gases.  Scrubbers, precipitators
or other pollution control devices may be necessary on the exhaust air
streams.  Collected ash may possibly be reused as a filter aid conserving
chemical costs.  Although sludge incineration requires fairly large
capital and operational costs, in some cases supplemental fuel or heat
may not even be necessary.  It is essential in every case to obtain
maximum dewatering of sludges prior to incineration and to select a
furnace with high heat economy.  The multiple-hearth furnace is said to
be well-proven and reliable for the incineration of sludges and also
claimed to have the highest thermal efficiency of the various types of
incinerators commercially available (47).
Nitroaniline Isomers:  Biodegradation and Treatment Capability

     Extremely long-term biodegradation tests were conducted on the
isomers of nitroaniline, which were being used as raw materials in the
manufacturing of bulk Pharmaceuticals and special chemical products at
an organics plant in the midwest (5, 119).  The subject isomers consisted
of the ortho-, meta-, and para- forms of nitroaniline used in the
organic synthesis of various sulfanilamides.  These sulfanilamides
included formulated organic dyestuffs and a special coccidiostat used in
treating poultry diseases.  Of the three isomers, the company reported
greatest use of the para form; the ortho was used only to a limited
extent; and the meta form was eventually discontinued in production.
The nitroanilines were studied because difficulties developed during
treatability tests being conducted on the company's wastewaters.
Although these industrial wastes were being sent to municipal facilities,
the company investigated the feasibility of pretreatment and separate
treatment schemes.

     The first bench-scale and pilot-scale biological treatment units
showed poor organic carbon removals.  A persistent yellow-orange color
indicated that the problems were due to the nitroanilines which were
known to have toxic properties.  Subsequently, long-term biodegrada-
bility studies were run on these compounds.  The findings indicated that
process modifications would not offer a satisfactory solution and that
almost complete recovery of these materials was necessary (5, 119).

     Certain organic industrial wastes are exceedingly difficult to
treat biologically because they are either toxic, virtually
nonbiodegradable, or only slowly oxidized so that they are not removed
during the normal course of biological treatment.  The company conducted

-------
                                                                      89
biodegradability tests for 180 days to specifically determine the
treatability potential of the wastes involved.

     Although previous work on the biodegradation of nitroaniline was
quite limited, some related investigations had been carried out on
nitrobenzene, aniline, nitrophenols and nitro-benzoic acid.  These
results showed a wide degree of susceptibility to breakdown.  The
literature also indicated the possibility of cross-adaption or cometa-
bolism, i.e. when a culture of micro-organisms has adapted to one
chemical or isomer, it shows a definite tendency to degrade a related
chemical or isomer.  For purposes of the long-term biodegradation
studies, the isomers of nitroaniline were mixed with sewage, an electro-
lytic respirometer measured oxygen uptake rates, and nitrification was
suppressed by means of the chemical inhibitor, 2-chloro-6-(trichloro-
methyl) pyridine.  The biochemical oxygen demand of the sewage substrate
and the added nitroaniline could be separately accounted for in the
oxygen utilization-time rate curves that were developed.

     The first tests essentially indicated over a period of 60 days that
both ortho- and meta-nitroaniline were virtually nondegradable.  The
para-nitroaniline, although showing no susceptibility to breakdown for
the initial 20 days, exhibited significant oxygen uptake thereafter.  It
was indicated that all three isomers at the concentrations used in the
tests were not toxic to the micro-organisms in the sewage seed.  Cross-
seeding was attempted with micro-organisms previously adapted to p-
nitroaniline applied to new samples of meta- and ortho-nitroaniline.
The results of these tests were only partially successful.  Essentially,
about 25 percent of the meta-isomer was degraded over a period of 50
days.  There was virtually no biodegradation of the ortho-isomer.

     Testing was continued, adding relatively high levels of p-nitroaniline
to previously acclimated seed to determine if these high concentrations
could be sufficiently oxidized and if toxic effects would occur.  Initially,
50 mg/1 p-nitroaniline was added to each of four samples of sewage seed.
After twelve days practically all of the original p-nitroaniline had
successfully degraded, and at this point further additions of the
chemical were made in amounts of 100 mg/1, 200 mg/1 and 400 mg/1 to each
of three samples.  Oxygen utilization curves were developed over 74
days.  In all three cases, the oxygen curves rose immediately after the
extra chemical  additions (i.e., the 12th day).  The BOD curves were
considered normal although protracted for the two situations in which
100 mg/1 and 200 mg/1 doses had been added.  The curve reflecting the
addition of 400 mg/1 p-nitroaniline demonstrated a different shape.
Young and Affleck (5) concluded that there appeared to be no toxic
effects up through about 200 mg/1 p-nitroaniline.  However, at the 400
mg/1 level, the oxidation rate was definitely slowed, but "complete"
toxicity was not reached, since the p-nitroaniline was eventually
degraded (45 to 55 days).

     A further  test was conducted to determine temperature effects upon
biodegradation  of para-nitroaniline.   Comparison of results obtained at

-------
90


35C vs 20C (95F vs 68F) showed rather unexpectedly that there was no
degradation of the nitroaniline isomer at 35C.  Apparently all previous
testing had been carried out around 20C.  It was also noted that no
other series of micro-organisms was capable of adapting to the p-nitroaniline
at 35C throughout the entire test period of 35 days.  The sewage
portion of the substrate responded in anticipated fashion and showed
about twice the degradation rate in going from 20C to 35C.

     It may be concluded from the above investigations that both ortho-
and meta- nitroaniline were not satisfactorily degraded after tremendously
long periods of holding.  Reasonably large amounts of p-nitroaniline can
be biologically treated but careful controls must be exercised and
reaction periods are still in the order of many days.  It was observed
that the yellow color of the p-nitroaniline solution gradually diminished
when biodegradation was progressing smoothly.
E. P. Squibb and Sons* Humacao^ Puerto Rico

     The Squibb Plant planned to segregate their wastes from synthetic
penicillin and antifungals production into three main streams:  the
strong process stream; the weak process stream together with sanitary
sewage; and miscellaneous blowdowns plus spent demineralizer regenerants.
Company plans stipulated that 99 percent or greater of the organic
wastes generated from manufacturing would receive a high degree of
treatment (23).

     Strong process wastes amounting-to about 30,800 kg (68,000 Ib)
COD/day in an average flow of 44.1 m /day (11,800 gpd) were to be sent
to storage tanks and continuously fed to a vertical liquid incinerator
supplemented by fuel oil.  Preliminary test runs had shown that these
concentrated wastes could be combusted with a minimal amount of visible
plume.  However, the wastes originating from the methanol solvent re-
covery unit, largely still bottoms, were diverted from going to the
incinerator because fuel costs for burning this particular waste were
thought to be excessively high.  They were transferred to the weak
process stream (economic considerations overshadowed technical expediency)

     The dilute process waste stream amounted to 113 kg (250 Ib) COD/day
in an average flow of 129 m /day (34,000 gpd).  This COD apparently does
not reflect the addition of the still bottoms mentioned above.  These
still bottoms demonstrated no acute toxicity but adequate data was
lacking in the Squibb report.  The stream was to receive biological
treatment consisting of waste equalization, activated sludge (probably
extended aeration), chlorination, settling and filtration.  The effluent
would then be sent to a storage tank for cooling tower makeup water.
Nutrient addition and pH adjustment, if necessary, would be integrated
into biological treatment.  Excess sludges will be hauled to land
disposal.  The biological treatment system was designed for a minimum of
85 percent BOD removal.

-------
                                                                       91


     The miscellaneous "service waters," amounting to about 132 m /day
 (35,000 gpd), will enter an equalization tank with 2 days'  detention,
 then be neutralized, settled, and pumped to an emergency fire pond as
 makeup against evaporative losses.  Overflow from the emergency pond
 will be directed to Pridco Ditch serving as the receiving watercourse.
 Unfortunately, no performance data is available on the Humacao instal-
 lation, presumed completed, and therefore no judgment can be made
 concerning the full-scale system.


 M/s Indian Drugs and Pharmaceuticals, Hyderabad, India

     The Hyderabad, India plant manufactures a variety of synthetic
 drugs and chemical intermediates.  The  Indians and Russians in the late
 1950's had concurred on a conventional  type of treatment facility
 consisting of equalization, pH adjustment, preaeration (extended aeration),
 sedimentation, biological filtration with recirculation, secondary
 sedimentation and lagooning.  The primary and secondary settlers and  the
 trickling filter units were planned to  be installed in duplicate providing
 parallel treatment.  These plans called for pH adjustment and neutralization
 of the industrial wastes at the plant site.  Domestic sewage would be
 mixed with the process wastes.  Known toxic materials, such as cyanide or
 arsenic, would be segregated within the plant (30).

     A wastewater was simulated in the  laboratory based upon the com-
 position of expected factory effluents.  The laboratory-scale treatment
 system was the type described above, and was designed to handle 15 to 19
 liters (4 to 5 gal) of synthetic drug waste daily.  The trickling filter
 was studied first.  Both the simulated  drug waste and 50 mg/1 and 100
 mg/1 of simulated phenolic wastes were  applied to the filters.  P-
 aminophenol and p-nitrophenol and sewage were used for preparation of
 the phenolic waste feed.  The full-strength simulated drug waste contained
 about 2,500 mg/1 COD, which was too much for the laboratory filters,  so
 the waste was cut to about 300 mg/1 COD.  Four runs were made for each
 of the waste streams.  The trickling filtration runs provided 27 to 37
 percent COD removals on the drug waste.  COD removals on the 50 mg/1
 phenolic waste ranged from 70 to 87 percent.  However, when the phenolic
 level was increased to 100 mg/1, the COD removal dropped, ranging from
 51 to 70 percent.  Patil (30) commented that the concentration of
 phenols in the actual drug waste is in  the order of 150 to 200 mg/1,
 which means very little removal of phenols by trickling filtration.
 Further, all aromatic compounds which are present in high concentrations
 in the drug waste will also leave the biological system without being
 broken down to simpler or less noxious compounds.  It was concluded that
 trickling filtration was not a suitable method to adequately handle the
 highly complex wastes.

     In 1970, Mohanrao (54) described the characteristics and volume of
wastes coming from the various manufacturing sectors of the completed
Hyderabad, India synthetic drug factory (Tables V-4, 5).  -Because Sector
8 waste (constituting about 50 percent of the total  568 m /day (0.15

-------
92


mgd) process flows) was fairly dissimilar from the remaining process
wastes, strong consideration was given to segregation and separate
treatment of the Sector 8 flows.  Sulfonilic acid was thought to be the
only organic in Sector 8 wastes, and later recovery of acid was deemed
possible.  Treatability studies were conducted on Sector 8 waste and
subsequently on the combined wastes from all other sectors.  Presumably
because the factory had not yet opened, simulated rather than real
wastewater feed was used in all the treatability investigations.

     Neutralization of Sector 8 waste was attempted using a 10 percent
lime slurry, but sludge settling and separation were poor and there was
only minimal removal of BOD, COD and sulfonilic acid.  Neutralization
and chemical coagulation of wastes from the remaining sectors likewise
did not produce significant removals.

     Total process waste exclusive of Sector 8 flows was diluted to give
1 and 2 percent waste solutions, which were treated with unacclimated
activated sludge for 23 hr.  Maximum reductions in BOD and COD were
respectively 40 and 32 percent, decreasing on subsequent days and indi-
cating progressive sludge inactivity.  This same waste diluted to 0.5 to
8.0 percent (percent volume/volume) was then subjected to acclimated
activated sludge, with aeration periods of 23 hr followed by 1 hr of
settling.  The waste feed concentration could not exceed 7 percent in order
to guarantee an effluent containing less than 50 mg/1 BOD, and a COD of
250 to 280 mg/1.

     With an acclimated sludge and a constant influent BOD of about
1,000 mg/1, the relation between aeration period and MLSS levels vs BOD
removal efficiencies at various F/M ratios was developed.  Rates of
sludge synthesis and settling characteristics were also determined.
These results show at a BOD/MLVSS ratio of 0.29 or less, the efficiency
of BOD removal approximates 90 percent with an aeration period of 8 hr
giving a final effluent BOD of about 100 mg/1.  The amount of organic
matter in the excess sludges was 87 to 93 percent, and sludge settleability
was very good yielding final effluent TSS levels of 30 to 40 mg/1 after
20 min.  Because of phosphorous deficiency, all of the above wastes were
supplemented with dipotassium hydrogen phosphate.

     From the laboratory studies, the drug wastewaters are judged amenable
to activated sludge treatment but required an absolute minimum of 14
times dilution for efficient organic removals.  Information supplied by
the factory shows that 2,000 m, (0.5 mgd) of domestic sewage containing
100 mg/1 BOD, and some 4,000 m  (1.0 mgd) of spent condenser wastewater
with a BOD about 50 mg/1 would be available for diluting the 300 to 350
m /day (0.07-0.09 mgd) of process wastes exclusive of Sector 8 flows.
Consequently domestic sewage and spent condenser waters were proposed to
be used for the dilution of process wastes within biological treatment.
This would provide dilution ratios from 17 to 20:1.  Process wastes
exclusive of Sector 8 flows plus sanitary sewage and condenser waters
will be treated via a two-stage high-rate trickling filter recently
installed by the factory.  This will be followed by an oxidation pond,

-------
                                                                      93
and the effluent will then pass to the plant sewer.   Sector 8  wastes
will receive only lime neutralization and sand filtration before  finally
discharging to the same sewers (54).

Hoffman-LaRoche, Inc. Belvidere,  New Jersey

     Information on the Hoffman-LaRoche bulk pharmaceutical manufac-
turing installation at Belvidere, New Jersey was obtained during  an
NFIC-D visit made on June 29, 1972, and from continuing correspondence
regarding the NPDES permit between the Company and EPA, Region II,  New
York, N.Y. (119).  The Belvidere, New Jersey facility initiated opera-
tions in 1969 and by June 1972, the plant had approximately 350 employees.
Considerable production expansion was experienced during 1972  and 1973,
and the Company reported 738 employees in September 1973.

     Process Description.  In 1972, the Belvidere installation was
divided into three process divisions:  1) the Dry Powders Plant;  2) the
Sulfa Drugs Plant: and 3) the Vitamin C Complex.  Dozens of products  are
manufactured in the Vitamin Powders plant.  A number of chemical  inter-
mediates were also included in the above production.  Vitamin  C manu-
facturing did not fully come online until late 1972-early 1973.  The
Company uses at least 80 different raw materials including lead,  mercury,
palladium and nickel.

     According to the 1973 NPDES permit application, vitamin powders  are
produced by spray drying, spray chilling, and a spray tower process
followed by blending operations,  yielding 2,040 kg (4,500 Ib)  daily.
The vitamin products require gelatin, vegetable gum, starches, edible
waxes and other materials or coatings, covering, and fillers.   Since
Belvidere directly manufactures both vitamin C and riboflavin, it is
assumed that these and other vitamins are converted into vitamin  powders
and mixes at this same location.

     The plant manufactures about 541 kg/day (1,200 Ib/day) of hard-
shell gelatin capsules.  Sulfa drugs consisting of sulfamethoxazole and
sulfadimethoxine amount to an average production of 2,490 kg (5,500
Ib/day).  The sulfa drugs plant is said to present special problems in
waste treatment and likely the source of much of the mixed waste  salts
undergoing special recovery and disposal at Belvidere.  Within sulfa
manufacturing, at least two inorganic waste streams, spent carbon and
filter aid, are separately collected and taken to off-site disposal.

     The Riboflavin Manufacturing Process, involving organic synthesis
operations, was expected to begin during 1974.  The Vitamin C  or  Ascorbic
Acid Manufacturing Process consists of oxidizing glucose to sorbitol,
then to sorbose by a fermentation step, through diacetone, and eventually
to vitamin C, mostly by organic synthesis procedures.  Production of
vitamin C in September 1973 was given as 27.2 metric tons (30  tons) per
day.  Aromatics Manufacturing appears to be producing minor amounts of
undesignated complex aromatic organics.  Sodium sulfate is a principal
product originating from the unique Sodium Sulfate Recovery complex,
involving a fluidized bed process and drying of the salt.  Various  waste
streams including selected streams from Sulfa Drug manufacturing  are

-------
94


directed to this salt recovery complex.  Incoming feed must be carefully
controlled to minimize chlorides into the recovery process.  The byproduct
manufacturing rate was 72.6 metric tons (80 tons)/day sodium sulfate in
1972.  Very large amounts of water are evaporated in these operations.
Sodium sulfate is said to be marketed for use in pulp and paper manufac-
turing and detergent chemical manufacturing.

     Approval and Startup of Belvidere Plant.  Consent was given by the
State of New Jersey on July 23, 1968, followed by concurrence of the
Delaware Rier Basin Commission (DRBC) in Sept. 1968 for industrial waste
treatment facilities at Hoffman-LaRoche, Belvidere, N.J.  Besides sulfa
drugs and vitamin C manufacturing, the establishment was said to include
an animal testing farm for basic research, although the latter function
probably was not constructed.  The proposed treatment facilities were to
provide screening, waste equalization and pH adjustment, chemical floc-
culation and settling, activated sludge aeration with final settling,
followed bw holding lagoons and chlori nation.  The design flow was cited
as 3,780 m /day (1.0 mgd).  Excess biological sludges were to be lagooned
and dried in open beds, but only as a temporary measure.  A fish aquarium
tank was to be installed in the plant laboratory and would receive
treated effluent continuously as a monitor on toxicity.

     The initial flows of sanitary sewage and industrial waste were not
expected to exceed an average of 150 m /day (40,000 gpd) with a maximum
BOD for the raw wastes of 185 mg/1.  This loading was equal to only 28
kg (62 Ib) BOD/day.  Under full operation, the raw wastes at the design
flow of 3,780 m /day (1 mgd) were expected to have a BOD of 1,900 mg/1
and a TSS of 1,000 mg/1.  These values are equivalent to daily loadings
of 7,190 kg (1-5,860 Ib) BOD/day and 3,790 kg  (8,350 Ib) TSS/day.  The
treatment plant design approved by the Delaware River Basin Commission
specified effluent loadings of 50 mg/1 BOD =  190 kg (417 lb)/day BOD and
20 mg/1 TSS = 76 kg (168 lb)/day TSS.  The design conditions provided
for BOD removals of 97.4 percent and TSS reductions of 98.0 percent.

             1972, the Company reported a waste flow averaging about
1 ,020 m /day (0.27 mgd) and containing a raw wasteload of about 5,440 kg
(12,000 Ib) BOD/day.  Startup of the vitamin C complex was anticipated
during the latter half of 1972 at which time the raw wasteload was
expected to reach to 13,600 kg (30,000 Ib) BOD/day.  These Company
figures are noted as being much higher than design estimates previously
given to the State and the DRBC in 1968.  By Sept. 1973, the Company on
is revised permit application increased its flow estimate to 6,100
m /day (1.6 mgd) together with a maximum flow of 7,950 m /day (2.1 mgd).
This compares to the design flow estimate of 1.0 mgd given in 1968.

     The treatment works were a multifunction system.  As of June 1972,
total construction expenditures were estimated at about $2.6 million.
The Company in 1970 had estimated that full system costs for abatement
of water, air and solids waste pollution would eventually approximate
$4.0 million.  Waste streams include industrial  and sanitary flows,
water treatment plant sludges, regenerants, and various blowdowns all
directed to treatment.

-------
                                                                      95
     Raw wastes enteced the treatment plant through a barminuter and bar
screens into a 284 m  (75,000 gal) preclarifier which was integral  with
a 2,780 m  (1 million gal) equalization basin.  Very little solids
accumulation was noted in these first two holding basins.  The equal-
ization basin is equipped with two turbine type surface aerators, oper-
ating at 25 hp each.  Full load on each aerator is 100 hp.  No chemicals
are added to this side of the treatment plant.  The TSS in the equalizing
basin were only about 200 mg/1, but waste loads during this period  were
only about one-fourth of full normal loads.   The equalized wastewaters
passed through a parshall flume, received pH adjustment, and entered a
17 m 55 ft) diameter x 3 m (10 ft) deep flocculator-clarifier with a
673 m  (178,000 gal) capacity.  This chamber was not in full use during
June 1972.  The wastewaters then proceeded to an aeration basin 6 m
(20 ft) deep by 36.6 m (120 ft) long by 12 m (40 ft) wide.  The aeration
basin had common wall construction with a second basin of similar size
being used as3an aerobic sludge digestion tank.  A 7.6 m (25 ft) diameter
tank of 135 m  (35,700 gal) capacity was available as a sludge thickener
but was not in use.  Aeration basin effluents were routed to a secondary
clarifier which was identical in size to the flocculator-clarifier
previously described.  Secondary clarifier effluents were passed through
two 3,780 m  (1 million gal) shallow oxidation ponds (no artificial
aeration), then chlorinated before final discharge to the Delaware
River.  Excess biological sludges were being taken to experimental
sludge drying beds with the dried sludge cake believed to be ultimately
carted away to landfill.

     Testing was conducted for BOD, COD, TOC, pH, conductivity, dis-
solved oxygen, color, chlorides and temperature.  The live continuous
flow-through bioassay aquaria used fathead minnows as the test species.
Three aquaria were used, one serving as the control (also holding extra
fish), the second containing only Delaware River water with fish, and
the third stocked with test fish containing treated effluents diluted
200-fold with Delaware River water.

     By Sept.  1973, the Company declared it was no longer possible  to
consistenty attain the effluent design loads of 1968.  The permit appli-
cation of Sept. 14, 1973 lists treated average discharge loads for  the
single outfall 001 as follows:
               Component           	  Load	
                                   (kg/day)        (Ib/day)
BOD
TSS
COD
TDS
NH-, N
POJ, P
TOC
Sul fates
Chlorides
Aluminum
260
435
1,324
17,900
5
10
422
454
998
0.3
574
960
2,920
39,500
11
23
930
1,000
2,200
0.6

-------
96
               Component           	  Load
 I
Iron
Lead
Nickel
Sodium
Zinc
Oil /Grease
Phenol ics
Fecal Col i forms
(kg/day)
0.5
0
0
1,840
2.4
0
0
--
(Ib/day)
1.2
0
0
4,060
5.3
0
0
5/100 ml
     Current raw wasteloads were not available.  Nevertheless, minimum
reductions can be calculated using raw waste values of 7.2 kkg (15,860
lb)/day BOD and 3.8 kkg (8,350 lb)/day TSS given as design conditions in
1968, and a BOD raw waste loading of about 13.6 kkg (30,000 lb)/day
reported for the end of 1972.  The BOD removals, based respectively on
7.2 kkg (15,860 lb)/day and 13.6 kkg (30,000 lb)/day raw loads, are 96.4
percent to 98.1 percent.  The TSS removal, based on 3.8 kkg  (8,350
lb)/day, is 88.5 percent.

     The above calculations are thought to represent minimum waste
reductions, and therefore typical removals through the treatment systems
are presumed to be about 97.5 percent BOD and at least 90 percent TSS.
It is noted from the Sept. 1973 permit application data that ammonia and
organic nitrogen loads in the final effluents are exceptionally low.

     The COD parameter is always considered important in properly
characterizing pharmaceutical wastewater but also EPA Region II recently
asked Hoffman-LaRoche to conduct a study of the sources of TOC in the
discharge.  This study was not only intended to identify TOC components
but also to set the basis for removal of such compounds from the dis-
charge and enable effluent limitations under the NPDES permit.  The
superiority of COD vs TOC or TOD is undergoing evaluation.  The addition
of TOC to continuously collected BOD and COD information by the company
would seem very desirable.  Review of the literature indicates that
certain reduced inorganic ions can be oxidized under the COD test giving
abnormally high results.  The dichromate oxidation of sulfites, sulfides,
nitrogen and ferrous iron was suggested.  Ongoing study by Hoffman-
LaRoche tentatively shows that refractory-type compounds are resistant
to decomposition in the COD and BOD tests, whereas they may be measured
by the TOC analysis.  The opposite may also be true.

     Assuming in the study that ultimate BOD is about 90 percent of
theoretical oxygen demand and BODg is about 48 percent BOD|||T, the
ratios below were developed.

-------
                                                                      97
               BOD,   Oy   32
               	i =-.=   X  (0.9 X 0.48) = 1.15
               TOC    C    12

               COD    09   32
               	  =_-=_  = 2.66
               TOC    C    12
     In turn, these ratios were compared to the ratios derived on the
wastes entering and leaving the Belvidere treatment system during 1972.
BOD/TOC
Influents
Effluents
0
0
.64
.03
- 0
- 0
.88
.05
2.
2.
COD/TOC
32 -
00 -
2
2
.83
.83
COD/BOD
3.13
33.3
- 4.35
- 142.

3
Evaluation of this data spectrum shows that:

     1.   The BOD/TOC ratios from the treatment system were less than
          calculated values and dropped considerably with increasing waste
          stabilization.

     2.   The COD/TOC ratios from the treatment system approximated the
          calculated values and did not vary significantly between the
          influent and the effluent.

     3.   The COD/BOD ratios for treated effluents were very high.

     The Company comments that a nonbiodegradable residue will  probably
remain in the effluents even after biological  treatment.   Even  a well-
oxidized secondary sewage treatment plant effluent is shown to  contain
10 mg/1 BOD, perhaps 10 to 20 mg/1 TOC, and 60 mg/1  COD.   The Company
thought that somewhat more material might be present in the effluents
than in the influents that would be' slightly less amenable to dichromate
oxidation (COD/TOC ratios).  However, the data collected  does not
warrant this conclusion.  The presence of nonbiodegradable materials
and/or inorganics were said to contribute to the relatively low BOD/TOC
ratios.  More data was necessary, and Hoffman-LaRoche indicated that
other analyses such as gas chromatography, thin-layer chromatography,
and spectrophotometric techniques will probably be necessary to de-
termine specific nonbiodegradable and inorganic fractions (119).

     Known air pollution control measures at the Belvidere facilities
include cyclones with water sprays serving the Vitamin Powders  plant

-------
98
together with baghouse units and a wet scrubber on the vitamin spray
dryer exhaust.  Belvidere has centralized solvent recovery, but it also
collects large quantities of nonrecoverable or nonuseable solvents and
conveys them to offsite incineration or landfill.


Fermentation/Synthesized Organic Chemicals

Abbott Laboratories3 North Chicago, Illinois

     The Abbott installation represents a large Pharmaceuticals manufactur-
ing plant with approximately 6,000 employees operating on a 3-shift, 24
hr/day, 7 day/week continuous schedule.  They have extensive fermentation
and antibiotic manufacturing facilities.  Two of the most important
antibiotics at Abbott in past years have been erythromycin and penicillin.
Of the hundreds of fine chemicals and other medicinals made by Abbott,
majar types include the sedatives, diuretics, the antihypertensives,
anticoagulants, anticonvulsants, laxatives and antidandruff preparations.
Abbott also produces animal feed supplements, intravenous solutions and
associated equipment, irrigation solutions, vitamin preparatons, cough
medicines and a host of other health products.  At least one "biological
insecticide" was manufactured there in 1972.  They conduct highly varied
processing of synthesized organic materials including drug preparation,
formulation and packaging operations.  Furthermore the Company maintains
a small manufacturing, packaging and research center located at Abbotts
Park near North Chicago.  This center is engaged in the production of
radiopharmaceuticals.

     Early Production and Waste Treatment.  The plant, less than one-
half mile from Lake Michigan, draws its water intake from and also
returns treated effluents to the Lake.  Over the years, because of the
magnitude and complexity of the waste streams, it has been desirable and
also necessary for the Company to operate its own waste treatment works.
Starting with modest penicillin production during the early 1940's,
fermentation activities grew steadily from the 1940's through the 1960's,
with rapid expansion occurring through the 1960's.  Although heavily
engaged in antibiotics production via fermentation, there is consider-
able activity in the manufacture of fine chemicals through multiple-step
chemical reactions.  In the middle-1960's, fermentation wastes were said
to contribute about 80 percent of the organic matter in the combined
Abbott process waste streams.

     Waste treatment in the early 1940's consisted of neutralization,
lagooning and chlorination.  An activated sludge plant was constructed
in 1953 which consisted of twin primary tanks, twin aeration tanks,
settling tanks, chlorination, sludge filtration and land disposal.
Design capacity of this plant operating at 90 percent BOD removal was
2,720 kg (6,000 Ib) BOD/day raw waste loading within^an average flow of
1,890 m /day (0.5 mgd).  The primary tanks were 76 m  (20,000 gal), the
aeration tanks were 1,140 m  (300,000 gal), and the secondary settlers
both had a 150 m  (40,000 gal) capacity (22,56,103).  Sanitary eff-
luents, spent process streams, and storm flows were segregated.  San-
itary flows were sent to the nearby municipal treatment plant; cooling
waters were released to the lake without treatment; and process wastes
were received into the biological treatment system.

-------
                                                                      99


     Turbine Mixer Aeration.   In 1957,  a pilot plant  was  built  to  thoroughly
investigate the merits of turbine mixer aeration  which  promised up to 15
times higher allowable BOD loadings per unit volume of  aeration tank
than previous processes.  The process wastes from Abbott's  manufacturing
operations were averaging 2,400 mg/1 as predominately soluble BOD. The
pilot plant using sparged turbine mixers demonstrated power consumption
as low as 0.34 kwh/lb of BOD removed compared to  0.6  kwh  for conventional
municipal treatment plants, and 0.56 kwh per Ib BOD removed for the
Abbott plant previously using hydraulic ejectors.  Whereas  conventional
municipal treatment works employing activated sludge  were being designed
at BOD loadings around 0.5 kg/nT (30 Ib) BOD/1,000 ft0  of aeration
volume, the Abbott pilot plant showed the ability to  handle loads  up  to
7 kg/m (440 Ib) BOD/1,000 ft  concurrently giving 80-90% reductions.
This led to replacement of existing aeration equipment  with sparged
turbine mixers by Abbott in 1958.  Also, the primary  settlers were
modified into sludge flotation units.  These units produced 6 percent
sludge solids from a 1 percent feed and served successfully to concentrate
activated sludge (22, 56, 103).

     For full-scale sparged turbine mixer aeration, a tgtal of 300 hp
mixer capacity was incorporated into the single 1,250 m  (330,000 gal)
aeration basin.  The turbine mixer was intended to operate at a 90
percent efficiency.  The aeration tank was subdivided into three compart-
ments, each 11 m (36 ft) long by 10 m (32 ft) wide by 3.8 m (12.5 ft)
deep.  Each chamber had two turbine mixers.  The pre-existing system of
two aeration tanks was reduced to one tank, and the second was converted
to a waste equalization basin.  The treatment plant provided for neutrali-
zation of raw wastes, flow equalization in the 1,250 m  (330,000 gal)
tank, followed by a three-compartment activated sludge basin, secondary
settling and chlorination.  Flows progressed through  the three activated
sludge compartments in series.  A flotation thickener served to concen-
trate underflows from the secondary clarifier for return to the aeration
basin or for ultimate disposal.

     When all three aeration compartments were thoroughly agitated, good
separation of solids in the secondary settler could not be attained
because the mixed liquors were highly gasified.  Consequently, the two
mixers in the last aeration compartment were turned off and the air was
reduced.  This change considerably  improved sludge settling.  The last
compartment was used more for degasification than for aeration of the
mixed liquors.  Sustained full-scale operation of the activated sludge
basin under the conditions outlined above permitted BOD loadings in the
range of 7.2 to 8.8 kg/nT (450 to 550 lb/1,000 ft3) tank capacity/day.
BOD reductions of 75 to385 percent were obtained.  Air requirements
averaged3about 13,750 m /kg (220 ft  air/1b) BOD applied, and about
20,000 m /kg (320 fr/lb) BOD removed.  Power needs in  removing BOD were
quite low, averaging about 0.41 kwh/lb BOD (6, 55).

     Special Laboratory Programs.  To provide ample laboratory support
of the treatment plant, a series of special lab procedures  and  programs
were developed by Abbott.  These included a shaken flask test,  a sludge
activity test, bench-scale activated sludge units designed  to provide
answers on long-term treatment responses, and routine analysis  of the
raw wastes for nutrients and inhibitory components (35).

-------
100


     The shaken flask test was designed to give quick indications on the
amenability of new wastes to the activated sludge system.   It was neces-
sary to know whether occasional poor treatment performance was due to
decreased sludge activity, to a substrate having low oxidation potential,
or to some inhibitory component being present.  The shake  test uses a
500 ml flask receiving 100 ml of substrate together with 2,500 to 5,000
mg/1 of activated sludge obtained from the treatment system.   The flask
is shaken at 325 cpm for a prescribed period, samples are  withdrawn and
analyzed, and percent waste removals are calculated.  The  test showed
that a certain cyclic amine being disposed of intermittently was inhibi-
tory to activated sludge when present in amounts greater than 100 mg/1.
This data prompted the Company to improve recovery of the  cyclic amine
by adding a stripping column and closely regulating the rate of release
of this compound.  The flask test was also utilized to determine the
effect of temperature upon BOD reduction and to assess the minimum
temperature level at which steam injection and/or activated sludge
reseeding was advisable.  The tests were run for periods of 1, 7 and 24
hr.  Drops in activated sludge treatment efficiencies were observed at
temperatures of 18 to 24C (65 to 75F) or below, and corrective measures
were implemented.

     The sludge activity test told whether a low BOD reduction may be
due to the raw waste having a low oxidation rate, or the sludge itself
having poor activity.  For relatively low oxidation of the waste, considera-
tion can be given to nutrient addition, adaption of the sludge to the
waste, or isolation and/or removal of the particular process waste
stream.  Bench-scale activated sludge units were used in initial attempts
to discover if weekend shutdowns at the main manufacturing plant would
have a significant impact upon waste treatment (35).

     The I960's.  In 1964, a large degassing and settling tank was added
to improve the solids separation in the secondary clarifier.  The secondary
settler consisted of a standard peripheral feed circular tank.  The
degasification chamber was added to the outer circumference of the
settler at the feed end.  This operation was designed with drilled pipe
air sparging for the purposes of promoting flocculation and enhancing
the settling characteristics of the activated sludge clumps.

     Also in 1964, a change was made from chlorination to pasteurization
of the treated process effluents.  In this unique pasteurization
process, the turbine exhausted into a newly added barometric condenser
with the former chlorine contact chamber becoming the hot well.  Steam
at 5 to 10 psig was readily available from plant operations, and was
reused in driving the condensing steam turbine.  The exhaust steam, in
turn, was condensed in the barometric condenser with the treated wastewater
effluent serving as the coolant.  In this process, the effluent wastewaters
were heated up to 65 to 71C (150 to 160F) and held for 30 min in the
hot well, thereby completing the pasteurization process (22).

     In 1965, more aeration of the sparged turbine type was added to
existing facilities.  Expansion of fermentation activities (primarily
erythromycin) was steadily occurring through the 1960's with the largest
additions in 1966-67.  This caused a major re-evaluation of overall

-------
                                                                     101
waste handling.  Fermentation wastes were said to contribute about 80
percent of the organic matter in the combined process waste streams.
The plant decided to recover spent fermentation broths and sell  them  as
animal feed supplements for these reasons:

     1.   Only spent beer drying could meet the stringent effluent
          limitations expected in the future.

     2.   All materials dried would receive the equivalent of 100
          percent treatment.

     3.   Efficiency of the existing treatment system would be
          improved because of reduced BOD load to the aeration tanks
          and lower hydraulic load placed on final clarification

     4.   There would be considerably less "excess" activated sludge
          generated for final disposal

     5.   A marketable product would be created

     6.   The odor control system designed for the drying operations
          could be, and was subsequently expanded to include the exist-
          ing activated sludge plant, thereby solving another problem.

     Processes selected for the drying of spent beers consisted of:

     1.   Isolation and separate collection of the spent beer streams

     2.   Concentration of the spent beer streams from 4 to 5 percent
          solids up to 30 to 35 percent solids via triple-effect forced
          circulation evaporators

     3.   Subsequent drying of the concentrates by steam-heated drum
          dryers

     4.   Bulk solids storage, packaging and marketing

     5.   Necessary odor control

     One of the more critical aspects of the beer drying operations was
odor control  which constituted a major undertaking.  Incineration was
considered the most feasible and effective method of destroying odors,
although carbon absorption, wet scrubbing, chemical oxidation, odor
masking chemicals, and tall stack dispersion were evaluated.  The possible
incineration schemes contemplated were straight incineration at the
site, catalytic incineration at the site, and ducting of the odorous  air
streams over and into the main plant steam boilers.  These streams would
in effect be used as combustion air supply.  The latter system proved to
be the most economical approach.  The duct was designed in such fashion
to allow for future additional connection of the exhaust air from all
the fermenters and full connection from the waste treatment plant. The
combined odorous air stream was to serve as combustion air supply to

-------
102


three 60 kg/cm2 (850 psig)  steam boilers.   This  air stream would  be
oxidized at boiler temperatures of 980C (1,800F)  to 1,090C  (2,000F).
The ducting system ran 365  m (1,200 ft)  in length with a  diameter varying
from 122 to 137 cm (48 to 54 in).

     It was reported shortly after the incinerator  ducting system had
been installed in late-1967 that not only were the  fermenter exhausts
hooked up but also the activated sludge tanks were  enclosed  and their
exhausts were likewise connected to the main ducting'system.  Odors  from
the degassing chambers and sludge holding tanks were handled similarly
(22).

     Transition* 1969-71.  Additional waste abatement measures from 1969
through 1971 included:

     1.   Two more aeration tanks
     2.   A second combination degassing and secondary settling tank
     3.   A 67 percent increased waste equalization capability
     4.   A centrifuge system for sludge removal
     5.   An electrostatic precipitator and necessary equipment on the
          main boilers to allow  for burning of waste sludge
     6.   Surface condensers to  replace most barometric condensers
     7.   Major process and storm sewer renovations, extension and
          modifications

Peripheral waste recovery and abatement programs were:  solvent recovery/
recycle/incineration; ammonia recovery; accidental  spill control; and
solid waste disposal-recycle operations.

     Separate solvent recovery systems were installed, and  in 1973,
about 114 m /day (30,000 gpd) of aniline, benzene and amyl  acetate were
being recovered for reuse.  Some mixed solvents are collected for sale
and  some are incinerated in a smoke-free solvent burner.

     Ammonia is recovered in dilute form from at least one  of the chemical
manufacturing operations.   It is then concentrated and sold in bulk as
fertilizer.  The spill control program at Abbott incorporated unusual
design  in storage units, installation of extensive diking and curbing,
and  sewer separation.  Contamination throughout the separate sanitary,
process, and cooling water  sewer networks is monitored and  recorded to
quickly pinpoint leaking equipment, faulty sewers, and process irregularities
(103, 109).

     New Wastewatev Treatment System.  Presently, Abbott has three
separate wastewater drainage systems.  About 1,885 m /day (0.5 mgd) of
sanitary sewage and cafeteria wastes are discharged to the  North Shore
Sanitary District North Chicago  Sewerage Works.  Spent cooling waters
are  collected, chlorinated,,and  discharged more or less without treatment.
This flow averages 55,600 m /day (14.7 mgd).  The Company reports that
only cooling waters enter this sewer.  However, as of 1972,  the sewer
was  also collecting drips and condensates from the first-effect spent
fermentation beer evaporator, boiler blowdowns, some floor  drainage,
packing gland coolants, liquid effluents from wet scrubbing of air
streams, barometric condensates  from the spent broth evaporation system,
and  undefined tank and miscellaneous washings (103, 119).

-------
                                                                     103
     As of 1972, the Company was able to  meet  all  imposed effluent
criteria by the State and local  authorities.   The  treatment works as of
1973 had a reported hydraulic capacity of about  3,780 m  /day  (1.0 mgd)
and an organic loading capability of 13,600 kg (30,000 Ib) BOD daily
with greater than 90 percent BOD removals (16, 103,  119).  This  system
is described as follows:
     1.   Raw Waste Handling - Fermentation  and  chemical  wastewaters are
          passed through a fine bar screen,  the  flows  and pH  are  re-
          corded, and the wastes enter a  wet well.   The  pH of the raw
          wastes can vary from 3 to 13 in a  few  minutes.   The wastes are
          neutralized and pumped from the wet well  to  the equalization
          basins.

     2.   Equalization - Two equalization basins are available, each
          1,890 m^ (500,000 gal), providing  an effective capacity of
          3,600 m  (960,000 gal) and giving  1.0  to  1.5 days detention at
          average flow rates.  One of these  basins  is  divided into three
          compartments and can be used for the segregation and partial
          storage of potentially toxic or inhibitory chemical  compounds.
          Each basin is equipped with three  25 hp mixers to provide
          necessary agitation during equalization and  holding.

     3.   Aeration - The activated sludge aeration  tanks consist  of six
          378 m  (100,000 gal) compartments.  Four  of  the compartments
          are equipped with two 50 hp sparge air turbines, and the two
          remaining compartments each have a single 50 hp unit.   A total
          of 500 hp is available in the aeration basins.   Waste retention
          time is about 24 hr, excluding  the sludge return volumes.
          Operating temperatures are closely controlled  at 38C +_ 2C,
          the F/M ratio is maintained at  about 0.25, and the  MLVSS is
          considered optimum at 8,000 to  12,000  mg/1.  All the equaliz-
          ation basins and the activated  sludge  chambers are  enclosed by
          a  flat slab concrete covering.   As described previously, a
          ducting system continuously conveys the odorous exhausts from
          the equalization and aeration chambers to the  central boiler
          house for incineration.

     4.   Degassing - The mixed liquors leaving  the aeration  tanks are
          heavily entrained with fine air bubbles.   Since this greatly
          inhibits subsequent solids separation, degassing is necessary.
          The waste stream enters a common wall  degassing/launder
          chamber which surrounds the settler and has  a  capacity  of
          about 625 m  (165,000 gal).  Diffused  air at the bottom of the
          degassing chamber accelerates the  release of entrained  gases
          and breaks down the foamy layer.  Mean residence time in the
          chamber is about 8 hr at high hourly flow rates.

     5.   Final Settling - Two 18 m (60 ft)  diameter secondary clarifiers
          can provide a surface overflow  rate of 1.3 m /day/m  (180
          gsfd).  Sludge return rates of  up  to 15,100  1/min (4,000 gpm)

-------
104
          are possible to the aeration basins, although sludge return is
          usually maintained at 500 percent of the effluent rate.
          Excess sludges are sent to a pair of sludge centrifuges
          handling about 570 1/min (150 gpm).

     6.   Pasteurization - Pasteurization at 66C (150F) and 20 min
          wastewater detention offers three distinct advantages over
          chlorination:  a) a more positive coliform control can be
          maintained; b) a potential safety hazard that chlorine pre-
          sents can be eliminated; c) lower operating cost.  The chlor-
          ine equipment was retained on an emergency standby basis.

     7.   Spent Beer Recovery - Spent beers are concentrated via mul-
          tiple-effect evaporators to a 30 percent solids concentra-
          tion.  The resulting thick syrup is shipped in tank cars and
          sold as an additive in chicken feeds.  Excesses are incin-
          erated in plant boilers.  Although Abbott had previously
          installed driers, operating problems were experienced and
          drying was discontinued.

     Abbott Labs reported achieving an average overall BOD removal of 94
percent throughout 1971, 95 percent in 1972, and 96 percent removal
through the early part of 1973.  The spent beer recovery program was
judged largely responsible in achieving BOD removals greater than  95
percent.

     In 1972, the process wastes averaged 2,270 m3/day (600,000 gpd) and
had a population equivalent of about 100,000 persons ^6,800 to 9,100 kg
H5.000 to 20,000 Ib) BOD/day.  Waste characteristics were given  as
follows (103):
Parameter                Flow	     _JJOD_    _[S!L     _IDS_      PH

                     (md/day)(mgd)    (mg/1)    (mg/1)     (mg/1)


Chemical wastes         990  0.262     2,520      510      5,690     5.4
Fermentation
Wastes
Combined
1,180
2,180
0.312
0.575
3,620
3,120
1,660
1,140
3,590
4,620
6.7
6.1
Abbott calculates that their recycling and recovery operations in 1972
were responsible for reclaiming waste materials having a population
equivalent of about 300,000 persons ^23,000 to 27,000 kg (-\-50.000 to
60,000 Ib) BOD/day:  Quality data on the combined treated process plus
cooling water flows after final chlorination for July through Dec.  1972
are shown below:

-------
                                                                     105
Parameter
BOD
TSS
IDS
Phosphorous, P
Phenol ics
Coli forms, counts/100 ml
Mercury
pH 3
Flow m /day (mgd)
Actual Concentration
(mg/1)
16
20
400
0.6
0.02
11
0.0003
7.5 units
56,800 (15.0)
1972 State Effluent
Standards
(mq/1 )
20
25
750
1.0
0.3
400
0.0005
5-10 units

     Vlaste treatment performance data compiled from Monthly Summary
Sheets submitted by the Company to the State are given in Tables VIII-1
and VIII-2 for Jan. 1972 through Dec. 1973.  Steadily improving waste
removal efficiencies are apparent.  The overall BOD and TOC removals for
1972 averaged 94.6 percent and 86.0 percent, respectively.  In 1973, the
average BOD and TOC reductions were 96.7 percent and 83.0 percent,
respectively.  In some months in 1973, BOD reductions approached and
even exceeded 98.0 percent.  If the spent fermentation beers going to
the evaporators are considered an equivalent part of the process waste
loads, average BOD reductions for 1972 and 1973 would be 97.8 percent
and 98.7 percent respectively for the two years.  These recovered beers
are thought to have a BOD load potential of 9,070 kg (20,000 Ib) BOD/day
or greater.  Percentage removals of total suspended solids were below
desired levels, although average removals were still in the 71 to 74
percent range.  Significant carryover and entrainment of gases within
the mixed liquors leaving the aeration basins is likely yet occurring.
If spent fermentation beers were accounted for as part of the TSS raw
waste loads, reduction of equivalent raw loads may have easily exceeded
80 percent.

     The technical feasibility of incinerating excess biological sludges
in the coal-fired boilers had been previously demonstrated but volumes to
be disposed of by this means were originally too large.  A continuous
feed disk nozzle centrifuge consistently produced a 5 percent concentrated
sludge feed at 328 m  (100,000 gal)/week to the boilers.  Based upon a
rate of 30,000 m  (7.8 million gal) of sludge being incinerated yearly,
Abbott calculated incremental costs as $219,300 or $28.10 per 1,000 gal
of sludge incinerated.  This boiler system operating at 980C (1,800F)
or above was found to be a most efficient method for eliminating both
process odors as well as the sludges (16,103,119).

     The summary data show that from 1972 to 1973 hydraulic and organic
loadings increased from 8 to 16 percent, and a similar trend may be
likely in 1974.  This indicates that production may be exceeding the
available capacity of the treatment works.  Ammonia and phosphate loads
in the combined effluents are quite low.  Mercury may be approaching
marginal  conditions and arsenic concentrations during certain months are

-------
                                                                 Table VIII-1
                                       Influent and Effluent  Loadings Through  Biological  Treatment  Works
                                              Abbott Laboratories,  North Chicago,  111.,  1972-1973

Month
1972
Jan
Feb.
Mar.
Apr
May
June
July
Aug
Sept
Oct.
Nov.
Dec
Avg
1973
Jan.
Feb
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct
Nov.
Dec.
Avg.
Flow

Influent
BOD

a /
Effluent Red-'
TOC
Influent


Effluent
Red^
(m3/day)(mgd) (kg/day)(lb/day) (kg/day )(lb/day)^(%) (kg/day)(lb/day) (kg/day )(lb/day) (%)

2,020 0.534
2,180 0.577
2,080 0.549
.
2,240 0 593
2,340 0 619
2,100 0.555
2,310 0 610
2,340 0 619
2,160 0.570
1,930 0.511
2,020 0 534
2,160 0.570

2,250 0.594
2,520 0.666
2,270 0.599
2,400 0.635
2,490 0.658
2,430 0.643
2,420 0.639
2,420 0.640
2,540 0.672
2,720 0.719
2,820 0.746
2,670 0.705
2,500 0.660

5,770 12,730
5,790 12,760
6,530 14,400
-
8,300 18,310
7,150 15,760
7,120 15,700
7,340 16,190
8,650 19,070
6,590 14,530
5,950 13,120
4,900 10,810
6,740 14,850

6,670 14,710
8,020 17,680
6,970 15,370
7,770 17,120
7,790 17,165
8,490 18,710
6,860 15,120
8,200 18,070
8,420 18,570
6,640 14,640
5,910 13,035
6,010 13,260
7,310 16,120
a/ Reduction
b/ Monthly BOD concentrations leaving

283
249
340
-
525
365
411
489
466
412
267
216
366

272
487
236
156
139
204
184
182
295
204
155
411
244
treatment

625
549
751
-
1,158
806
908
1,079
1,028
909
588
477
807

600
1,073
520
344
307
451
405
401
651
450
342
906
538

95 1
95 7
94.8
-
93.7
94.9
94 2
93.3
94.6
93.7
95.5
95.6
94.6

95.9
93 9
96.6
98.0
98.2
97.6
97.3
97.8
96.5
96.9
97.4
93.2
96.7
plant during 1973

_ _
3,220 7,110
3,210 7,075
-
4,140 9,120
4,170 9,200
3,810 8,410
4,680 10,320
4,960 10,930
4,470 9,860
3,500 7,710
3,140 6,925
3,930 8,665

3,790 8,365
5,020 11,060 1
4,610 10,170
4,550 10,025
4,820 10,620
4,830 10,655
4,290 9,450
4,540 10,005
4,570 10,080
4,200 9,250
3,730 8,235
3,810 8,395
4,400 9,690
were in the range

_
390
493
-
552
548
498
503
810
693
514
522
552

632
,050
610
627
740
696
779
681
865
721
675
880
750
of 56

_
852
1,086
.
1,217
1,209
1,098
1,110
1,787
1,527
1,134
1,150
1,217

1,393
2,312
1,345
1,383
1,631
1,535
1,717
1,501
1,907
1,590
1,488
1,941
1,645
to 193

_
88.0
84.7
.
86.7
86.9
86.9
89.2
83.7
84.5
85.3
83.4
86.0

83.3
79.1
86.8
86.2
84.6
85.6
81.8
85.0
81.1
82.8
81.9
76.9
83.0
mg/1;
f
TSS
Influent
(kg/day }(lb/day)

2,390
2,330
2,300
-
2,230
2,350
3,100
3,300
2,690
2,140
2,220
2,330
2,490

2,680
2,740
2,470
1,740
2,080
2,510
2,640
2,450
2,540
2,840
2,220
2,390
2,420

5,270
5,140
5,080
-
4,915
5,190
6,830
7,270
5,925
4,715
4,890
5,140
5,490

5,905
6,052
5,435
3,830
4,580
5,530
5,380
5,400
5,600
6,250
4,895
5,265
5,345
corresponding TSS
Effluent
Red^
(kg/day)(lb/day) (%)

676 1
380
713 1
-
1,100 2
708
563
654
645
969
783
527
702 1

652 1
1 ,030 2
426
555
625
545
796
370
552
509
511
1 ,020 2
633 1
levels were

,490
838
,571
-
,435
,560
,241
,441
,421
,136
,727
,163
,547

,438
,273
940
,224
,378
,202
,754
817
,217
,122
,127
,241
,395
153

71.7
83.7
69.1
-
50.5
69.9
81 8
80 2
76.0
54.7
64.7
77.4
71.4

75.6
62.4
82.7
68.0
69 9
78.3
67.4
84.9
78.3
82.0
77.0
57.4
73.8
to
409 mg/1.  Monthly concentrations for total  plant effluents  during  1973 were 9  to 20 mg/1  for  BOD averaging  13 mg/1;  corresponding  TSS
values were 10 to 53 mg/1  averaging 27 mg/1.

-------
                     Table VIII-2
            Nutrient and Trace Contaminants
Treated Process Combined with Cooling Water Discharges
  Abbott Laboratories, North Chicago, 111., 1972-1973

Month
1972
Jan.
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Avg.
1973
Jan.
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Avg.
Flow
Phosphates
(m /day)(mgd)

_
50,500
52,300
-
67,200
71,100
67,900
82,000
52,200
52,000
44,700
41 ,400
58,100

46,300
50,200
57,200
64,900
77,100
94,700
94,000
97,900
72,600
64,700
51,100
41,700
67,800

_
13.34
13.82
-
17.76
18.78
17.95
21.66
13.79
13.74
11.80
10.93
15.36

12.24
13.26
15.12
17.14
20.37
25.02
24.83
25.87
19.19
17.11
13.50
11.01
17.90
as P
(mg/1) (kg/day) (Ib/day)

_
0.9
0.9
-
0.9
0.8
0.6
0.6
0.6
0.6
0.9
0.9
0.8
f
1.1
1.1
0.8
0.6
0.6
0.5
0.4
0.3
0.2
0.3
0.5
0.6
0.6

_
45
47
-
60
57
41
51
31
31
40
37
44

51
55
46
39
46
47
38
29
15
19
25
25
36

_
100
104
-
133
126
90
112
69
69
89
82
97

112
122
101
86
102
104
83
65
32
43
56
55
80
Ammonia as N
(mg/1) (kg/day) (Ib/day)

_
2.8
2.6
-
4.1
4.9
3.0
3.2
2.8
1.8
2.6
1.5
2.9

3.0
1.8
1.8
3.7
5.4
1.9
1.4
1.8
2.1
1.4
2.3
0.9
2.3

_
145
136
-
276
348
204
262
146
93
116
62
179

139
91
103
240
416
180
131
176
152
91
117
38
156

_
320
300
-
608
768
450
578
322
206
256
137
395

306
200
227
529
918
397
290
389
336
200
259
83
345
Arsenic
(mg/1)

0.7
0.6
0.5
-
0.3
0.4
0.2
0.1
0.2
0.1
0.2
0.3
0.4

0.2
0.2
0.4
0.3
0.3
0.3
0.3
0.3
0.2
0.1
0.3
0.2
0.3
Phenol ics
(mg/1)

_
0.05
0.02
-
0.03
0.06
o.oi
0.03
0.02
0.01
0.03
0.04
0.04

0.02
0.02
0.01
0.01
0.04
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Cyanides
(mg/1)

_
0.004
0.002
-
0.009
0.008
0.004
0.001
0.003
0.005
0.005
0.003
0.004

0.004
0.003
0.005
0.006
0.004
0.005
0.005
0.007
0.007
0.005
0.008
0.006
0.005
Selenium
(ppb)

_
< 0.5
< 0.5
-
0.9
1.3
1.4
0.9
3.2
< 0.5
4.1
4.4
x- 2.0

4.9
4.6
2.7
2.2
4.6
3.6
2.3
3.8
2.3
2.7
4.1
2.7
3.4
Mercury
(ppb)

_
3.0
< 0.5
-
0.3
-
-
-
< 0.2
< 0.3
< 0.2
-
v 0.7

0.1
-
0.2
0.1
< 0.1
0.1
< 0.1
< 0.1
< 0.1
0.2
0.1
< 0.1
x. 0.1

-------
108


higher than desirable.  Arsenic should be maintained between 0.1  and 0.2
mg/1, or lower.

     The Abbott waste treatment works represents a multimillion dollar
installation.  The spent beer recovery system was indicated as having an
additional fixed cost of $2.0 million or more.  Annual  operation costs
for the waste treatment plant are said by the Company to approximate
$1.2 million which prorates to $1.20 to 1.30/nT ($4.50 to $5.00 for each
1,000 gal) of process wastes treated.  Of this amount,  about $0.50 is
for the depreciation of existing treatment facilities.

     In view of increasingly tight effluent requirements by the State of
Illinois for discharges into Lake Michigan, the Company as of 1971-72
made plans to connect their treated waste effluents to the advanced
waste treatment plant to be built by the North Shore Sanitary District
at Gurnee, 111.  Abbott intends to continue full operation of their
waste treatment facilities before discharge of the treated effluents to
the District.  The Company will also continue release of the spent
cooling waters into Lake Michigan.  The Gurnee plant as of 1973 was
designed to produce an effluent of 10 mg/1 BOD and 12 mg/1 TSS for
eventual discharge into the Des Plains River,.  State effluent criteria
for releases into Lake Michigan by 1975 will be 4 mg/1  BOD and 5 mg/1
TSS.  It is noted that the combined  Abbott plant discharges in 1973
averaged 13 mg/1 BOD and 27 mg/1 TSS (16, 103, 119).  The Gurnee plant
was originally scheduled to be completed in 1973 but apparently certain
delays have occurred.
Wyeth Laboratoriesj West Chesterj Pennsylvania

     In the 1950's this plant was a major manufacturer of penicillin.
Factory wastes of spent fermentation broths and sanitary sewage were
combined with the town's wastes and sent to a newly constructed trick-
ling filter plant under auspices of the town.  Problems soon developed,
and Wyeth Labs initiated pilot studies to determine if their wastes
could be treated by the "complete-mixing" activated sludge process
before discharge to the municipal sewer.

     The fermentation broths were made up of corn steep liquor, lactose
and mineral salts.  After appropriate fermentation, the fungi mycelium
was separated from the spent broth by vacuum filtration.  The penicillin
was extracted from the broth with a solvent such as amyl acetate in an
acid solution.  After retrieving the penicillin, the solvent was re-
covered by stripping for further reuse.  The remaining spent fermen-
tation broth had an extremely high BOD concentration and low pH.  Sep-
arate disposal of the mycelium filter cake served to produce a fermen-
tation wastewater quite low in TSS, if not in soluble organic matter.

     The Company studies focused on treatment and disposal of spent
fermentation broths and culminated in the construction of full-scale
pretreatment facilities during 1957-58.  The full-scale Wyeth treatment
installation was designed to handle-spent fermentation broths and
domestic sewage, amounting to 850 rrT/day (225,000 gpd) at pH 2.0 and

-------
                                                                     109


with COD waste strength of 2,560 mg/1 equivalent to 2,176 kg (4,800
Ib) COD/day.  No primary settling was employed.  Prior to entering the
aeration tanks, the wastes were to be neutralized with caustic to a pH
of about 6.5.  Two aeration basins were constructed in parallel each of
1,160  m3.,(41,000 ft ) capacity.  The tanks were based on removal of 0.96
kg COD/nT  (60 lbs/1,000 ft )/day.  It was estimated that 3 kg  (7 Ib) of
sludge would be generated for each 45 kg  (100  Ib) COD removed.  The
wastes then entered two final clarifiers  ooeratinq in paralleMesigned
to give a  hydraulic overflow rate of 6.5  m /day/m  (160 gpd/ft  ) and a
waste  detention time of 7 hr.  No sludge  digestion facilities were
built, since excess sludges would be sent to the municipal treatment
plant. The treated effluent was expected to contain less than  50 mg/1
soluble BOD and 180 to 260 mg/1 volatile  suspended solids (36,  50).

      Initially, only spent fermentation broths were sent to  the treat-
ment  plant.  Data collected Aug. 1958 to  Feb.  1959 showed an average
waste  flow into treatment of only 113 nT/day (30,000 gpd), containing
about  1,130 kg  (2,500 Ib) COD/day.  The average COD in the raw waste was
9,750  mg/1  but COD strengths varied widely from 2,400 mg/1 to  over
30,000 mg/1.  Data subsequently collected,during Feb. and Mar.  1959
showed an  average raw waste flow of  132 m /day (35,000 gpd)  commensurate
with  an average COD strength of 17,300 mg/1, thereby giving  a  daily COD
loading about 2,290 kg  (5,050  lb)/day.  The latter loading appeared to
exceed the design level of the treatment  plant.  The treatment of spent
broths by  themselves resulted  in an extremely  strong waste with very
long  treatment  retention time  (about 23 days). The effluent COD's
stabilized around 880 mg/1, indicative mostly  of non-biologically
oxidizable materials remaining.

      During the first year of  pretreatment, very high effluent TSS
values were traced to the waste neutralization practices and excessive
buildup of inert organic solids in the aeration basin.   Process changes
were  instituted to minimize these problems.  Nitrogen content,  TSS and
pH levels  were  found somewhat  interdependent.  Ample nitrogen  in the
spent  fermentation broths provided the opportunity of delivering an
effluent at almost any stage of nitrification. When approaching m'-,
trification, periods of low flows tend to throw the plant into den-
itrification and bulking sludge.  Incoming raw wastes were adjusted to
pH 6.5 by  soda ash before entering the aeration basin.   Biological
oxidation  resulted in the release of ammonium  ions which kept  the pH
level  above 6.5.  Conversion of ammonium  to nitrates and nitrites with
excess aeration caused a loss  in the ammonium  carbonate  buffering capacity
of the system leading to lower pH's, greater potential for filamentous
growths, and eventually a poor-settling sludge.  Careful control of the
pH level chemically was therefore necessary.   If operations  were con-
ducted almost entirely within the ammonia range, slight  overloads or
reductions  in oxygen transfer could also  result in undue Sphaerotilis
growths and sludge bulking (36, 50).

-------
no
     An  increase in fine chemicals production was experienced at the
Wyeth  plant around 1959-60.  This not only greatly increased COD loads
but also changed the character of the combined wastes from acid to
alkaline.  The wastewaters furthermore became heavily saturated with
inorganic salts.  Preliminary estimates in 1960 indicated that total COD
loads  would approximate 2,720 kg (6,000 lb)/day, and hydraulic loads
could  be about 378 to 568 m /day (100,000 to 150,000 gpd).  Pilot plant
studies  were resumed by the Company to determine impact upon the pre-
treatment works.

     With the introduction of the fine chemical wastes in the early
1960's,  aerator loads were about 1.2 kg COD/day/nT (73 Ib COD/day/1,000
ft  ) aerator volume.  Predictions were made for 4.5 kg (10 Ib) sludge to
be produced for each 45 kg (100 Ib) COD removed, an effluent COD of 600
mg/1 from the pretreatment facilities and a 90 percent BOD removal
efficiency.  An equalizing tank was added to achieve full blending of
acid'and alkaline wastes.  After collecting only one month of data,
peaks  as high as 4,500 kg (10,000 Ib COD)/day were recorded.  An average
daily  hydraulic load of 500 m /day (138,800 gpd) was also measured  (50).

     It  was reported in 1963(4) that studies were conducted on a waste
mixture  that contained 60 percent combined plant sewage plus 20 percent
each of  the3waste broth and fine chemical waste, the latter two amount-
ing to 76 m /day (20,000 gpd) each.  Blending of the alkaline and acid
wastes followed by pH adjustment to 7.0 with sulfuric acid gave an
incidental 30 to 35 percent COD reduction.  After neutralization and
equalization, the waste mixture was subjected to complete-mixing ac-
tivated  sludge.  COD removal in a series of eight pilot runs varied from
81.5 to  95.0 percent, averaging 90 percent.  With proper modifications,
the pretreatment facilities could successfully cope with the new fine
chemical wastes.  The system consistently showed 90 percent BOD removal
and up to 90 percent COD removal.  Industrial pretreatment followed by
municipal treatment were thought capable of providing overall BOD re-
movals of 97 to 98 percent for the Wyeth Laboratories' wastes.


Fermentation and Biologicals

Lederle Laboratoriest Pearl River,  New York

     Lederle Laboratories is a major manufacturer of antibiotics.  Main
processing consists of fermentation,  although synthesized organic chemicals
such as vitamins are also produced.  Biologicals produced include vaccines
and antitoxins (12, 24, 44,  119).

     In the plant,  spent fermentation beers are generally recovered in
wet form.  Some may be dried and sold as animal  feed supplements, but
most is taken to landfill  or spread onto nearby farmlands.   Bad fermentation
batches usually go  to the waste treatment plant.

-------
                                                                     Ill
     Molof (12) reviewed 15 years of waste treatment acitivity at Lederle.
The treated industrial effluents are transported by gravity to the
Orangetown, N.Y. municipal trickling filter treatment plant wherein
municipal and industrial waste streams are mixed roughly in a ratio of
1:1.  The final municipal effluent is released to the Hackensack River
Basin, an important watershed for domestic water supply.

     In the early 1960's Lederle had a two-stage trickling filter plant.
In 1961 the Company completed the conversion of existing tanks into an
activated sludge plant serving in tandem with the dual-stage trickling
filter system.  The activated sludge system consisted of an aeration
tank divided into four compartments with a total capacity of 830 m
(220,000 galKwhich was followed by a secondary clarifier and a storage
tank of 757 m  (200,000 gal).  Clarifier sludges could be returned to
the activated sludge chamber.

     Although wastes were settled before entering the trickling filters,
there were no provisions for settling prior to activated sludge treat-
ment.  The two biological processes were linked by a flow division box.
Three separate flows entered the division box, one having relatively low
BOD, a second with medium BOD, and the third with high BOD.  The divi-
sion box thus controlled the BOD loading to each biological process.
The activated sludge process was designed for a hydraulic load of 1,890
m /day (0.5 mgd), or about half of the total wastewater flows being
treated.  Another link between the two processes resulted from the
storage during the daytime hours of the activated sludge effluents in a
757 m  (200,000 gal) tank.  At night, this liquid was sent either to the
trickling filter plant or to the municipal sewers.  This storage was
necessary due to limits on maximum flows that could be accepted by the
municipality during the daytime.

     Special  pretreatment provisions were made for wastes from the
solvent recovery sector, the mash filtrates, and discarded fermenter
mashes.  Empty tanks were used for the separate aeration of collected
and stored wastes.  The most concentrated waste streams generally re-
sulted from solvent recovery.  A selected solvent may have a BOD of
about 2 million mg/1 or 1.6 kg BOD/1 (13 Ib BOD/gal).  This solvent is
some 6,250 times stronger than sanitary sewage in BOD strength.  There-
fore, variations at the solvent recovery unit can create wide variance
in the raw waste loadings.  These pretreatment operations at Lederle
were said to significantly absorb surges in both organic and hydraulic
loadings.

     Operational data were somewhat sparse at the time of the 1962 status
report because the activated sludge process had been in operation only
eight months.   The activated sludge system was receiving, with no odor
problem, about 1.9 kg/m (120 lb/1,000 ft0} BOD/day.  The trickling
filter system was receiving about 1.4 kg/m  (90 lb/1,000 ft ) BOD/day
with definite odor characteristics.  The activated sludge process also
tended to produce a clearer effluent.  Disadvantages of the activated
sludge process were that it required more technical  control than the
filters and its lesser ability to handle shock loads (12).

-------
112


     After two years of study and in large part due to severe odor
problems, the"Company selected the Unox Pure Oxygen Aeration Process,
and this unique activated sludge system was finally completed full-scale
in March 1972.  This new system, at a reported cost of $2.5 to $3.0
million replaced all previous systems.  The Unox system was designed to
treat an^average flow of 5,680 m /day (1.5 mgd) with hourly peaks up to
11,400 m /day (3.0 mgd).  The average organic load of the treatment
facility was designed at 18,100 kg (40,000 Ib) BOD/day with acceptable
peak loads up to 29,500 kg (65,000 Ib) BOD/day.  The treatment works
expects to produce an effluent with less than 250 mg/1 BOD and 250 mg/1
TSS.

     The main treatment unit is the closed reactor, a multichambered
aeration tank fitted with a gas-tight precast concrete lid.  A typical
reactor has three separate chambers or stages, separated by baffles,
with tank depths from.,3 m (10 ft) up to 9 m (30 ft).  The Lederle
reactor has a 3,100 m   (820,000 gal) capacity with a dual train, a total
of six bays.

     Oxygen is supplied to the aeration tank on demand.  When oxygen
pressure decreases in the reactor, a flow controller is activated per-
mitting more oxygen to enter. 'Oxygen is produced on site by a molecular
sieve generator.

     Mechanical surface aerators bring the pure oxygen into solution.
Liquids and oxygen are introduced into the first chamber, the oxygen
filling the space between the liquid level and the tank lid.  An agi-
tator then whips up the wastewater to contact the oxygen.  The reactor
has a normal liquid depth of 3 to 3.6 m (10 to 12 ft) with a 1.2 m (4
ft) blanket of pure oxygen above.  Oxygen and wastewater pass concur-
rently through each successive contacting stage, and the effluents from
the last stage leave the Unox chamber, for settling in conventional
clarifiers.

     The Unox unit provides an oxygen dissolution capacity of 12,200 kg
(27,000 Ib) oxygen/day at a 90 percent utilization efficiency.  The
reactor effluents are calculated to contain a minimum of 5 mg/1 dis-
solved oxygen.  Activated sludge is recycled to the head end of the pure
oxygen reactor.  A return sludge to incoming sewage flow ratio of about
1:1 is maintained.  The last pure oxygen contact stage is provided with
means of venting exhaust gases.

     Other treatment units at Lederle include a flash mixer and a clari-
flocculator, 18 m (60 ft) in diameter, preceding the Unox process.
Following the activated sludge process, there are three 12 m (40 ft)
diameter clarifiers arranged in parallel, plus chlorination of final
effluents.  Since the Unox system utilizes pure oxygen, a combustible
gas analyzer and a TOC analyzer were installed on the effluents from the
clariflocculator to warn of dangerous elements such as solvents and
explosive gases in the system.  Given certain conditions, the oxygen
delivery is shut off and/or the  wastewaters can be diverted around the
reactor.  In the event the exhaust gases from the reactor may create

-------
                                                                     113
 odors,  ozonation provisions have been made available.  A series of
 cooling towers are utilized for significant reuse of cooling water,
 principally  in fermentation.  The COD:BOD ratio of incoming raw feed
 into  treatment approximates 1.4 and that of the outflows averages 2.0.
 The latter ratio is  surprisingly low.

      Before  entering treatment, the raw process pharmaceutical wastes
 contain from 9,070 to  11,300 kg (20,000 to 25,000 Ib) BOD/day and 4,540
 kg  (10,000 Ib) TSS/day or more.  In June 1972, the Company verbally
 reported  an  average  BOD entering the  industrial waste treatment system
 of 9,800  kg  (21,600  lb)/day with outflows approximating 975 kg (2,150
 lb)/day (about 90 percent BOD removal prior to municipal "secondary"
 treatment).   Primary and secondary sludges from the treatment plant are
 vacuum  filtered.  The  dewatered biological sludges may be mixed with
 straw and sawdust and  composted; otherwise, these materials are in-
 cinerated or taken to  landfill for ultimate disposal (12, 24, 44, 119).
 Drug  Formulation

 Dorsey Laboratories3 Lincoln, Nebraska

      Anderson, et al.  (2, 38) conducted detailed studies in 1968-69 of
 the characteristics  and treatability of wastewaters from the Dorsey
 Laboratories' Pharmaceuticals formulation plant at Lincoln, Nebr.
 Attempts were made to  correct serious deficiencies in the existing
 extended aeration treatment facilities.  Correction procedures were,
 however, deemed largely impractical and future plans called for con-
 nection to the city  of Lincoln municipal sewerage following industrial
 pretreatment.  The Laboratories employ 225 people during a single-
 shift, 40-hr, 5-day  work week.  The plant conducts batch processing and
 the product mix changes frequently.  Products are primarily of the non-
 prescriptive type either liquid or tablet form.  These include seda-
 tives, digestive aids  and various medications for arthritis, coughs,
 colds, hay fever, sinus and bacterial infections.

      Wastewaters to  the extended aeration treatment facilities included
 sanitary and cafeteria wastes and spent waters from production and
 cleanup operations.  Spent cooling waters and boiler blowdowns were
 excluded from the treatment system.  The aeration basin and settling
 chamber were 79.5 m  (21,000 gal) and 7.3 mj (1,935 gal) respectively.
 Design retention time  in the aeration basin was 25 hr, and the design
 organic load (inflow) was 30.8 kg (68 Ib) BOD/day.  There were no
 provisions for wasting excess sludge or sludge recycle.   Furthermore,
 there was no equalization of the raw waste flows occurring over the 8 to
 10 hr process cycle each day, or of the high hourly variation in waste
 strength.

     Since 1965, the extended aeration plant had experienced serious
operational  problems such as bulking sludge, excessive loss of biologi-
cal solids from the system,  and overload conditions.   Resulting efflu-
ents were far from satisfactory.

-------
114


     The studies by Anderson et al. were divided into three phases:

          1.   Initial evaluation of the existing treatment system
               and nature of wastewaters;

          2.   Treatability of the wastes in a laboratory-scale treat-
               ment system;

          3.   Incorporation of findings into a detailed appraisal
               of the existing treatment plant.

     During the first phase, daily composites were taken of the influent
and effluent (Table VIII-3).  TSS and VSS in the effluents were much
higher than the levels entering the plant.  Also, BOD reductions were
only in the 30 to 70 percent range.

     During the second phase, the laboratory extended aeration treatment
system was operated continuously for 114 days.  The laboratory tests
indicated that the wastewaters coming from the Dorsey Laboratories
contained relatively high concentrations of mono- and di-saccharides,
determined to be quite conducive to filamentous growths of Sphaerotilis
natans.  Sphaerotilis is a leading cause of sludge bulking, and it was
reported as one of the major contributing factors to the operational
problems encountered in the existing treatment plant.  Nitrogen deter-
minations made on the raw wastewaters suggested a possible nitrogen
deficiency.  Difficulties were experienced in performance of the lab-
oratory system and the system was reported to have failed at least
twice.

     In the third phase, operational modifications were made to the
treatment plant and there was additional daily composite sampling and
evaluation from May 26 to Sept. 30, 1969 (Table VIII-4).  Again, the
influent flows varied greatly over the recording period.  On the 18th
day of the study, high strength sugar wastes started to enter the system
from the manufacturing of cough syrup, and Sphaerotilis forms promptly
increased.  To preclude system failure, the Company consented to keeping
the strong sugar waste out of the circuit from the 42nd to the 72nd day
of the survey.  Nevertheless, operational problems were still evident.
About the 73rd day, discharge of the sugar wastes resumed, and in spite
of a daily application of 5.4 to 6.8 kg (12 to 15 Ib) ammonium phosphate
to the basin, S. natans growths were increasing.  Measurements were not
made from the 87th until about the 121st day of operation.

     Results of grab samples collected on Sept. 21 showed that MLSS
concentrations had dropped from previous levels of 2,000 and 3,500 mg/1
to much less than 1,000 mg/1, and dispersed growths were highly pre-
valent.  Discussions with Company personnel also revealed during this
latter period that sulfa drugs had been produced at the factory.  Over
the last few days of study culminating on Sept. 30, 1969, the system
demonstrated partial recovery with a MLSS level of 2,400 mg/1 reached on
Sept. 30.  Besides wide hourly variations in waste strengths and flows,

-------
                                                                     115
                             Table  VIII-3

               Waste Treatment Performance Data,  Phase 1
                      Dorsey Labs,  Lincoln,  Nebr.
Parameters

Avg. flow
Max. flow
Detention time, days
BOD influent
BOD effluent
BOD % removal
COD influent
TVS influent
TVS effluent
TSS influent
TSS effluent
VSS influent
VSS effluent
VSS % removal

Jan. 3
(m3/day)(gpd)
23.5 6,200
110 29,000
3.4

820
570
1,520
832
900
218
688
147
646

Sample Dates, 1968
Feb. 13
(m3/day)(gpd)
20 5,400
107 28,300
3.9
(mg/1)
1,140
350
30 69
2,770
1,510
785
249
630
199
532
0 0
kg/day (Ib/day)

Mar. 19
(m3/day)(gpd)
58 15,400
173 45,600
1.4

840
250
70
1,960
486
365
81
250
72
241
0

BOD influent
19
(42.4)     23     (51.4)    49     (108.0)

-------
                                                 Table VIII-4

                                   Waste Treatment Performance Data,  Phase 3
                                          Dorsey Labs, Lincoln, Nebr.
Composite Sample Dates,
Parameters
Cumulative Days of Study
Avg. Flow: 2
m /day
(gpd)
Max. Flow: 3
m /day
(gpd)

BOD effluent
COD Influent
COD effluent
COD, % removal
TVS Influent
VSS influent
VSS effluent
VSS, % removal
6/12
17

15
(4,000)

38
(10,000)

3
375
50
87
279
80
97
0
6/25
30

20
(5,200)

40
(10,500)

9
375
30
92
264
140
155
0
7/15^
50

35
(9,300)

83
(22,000)

-
540
90
83
178
48
42
13
7/31-X
66

68
(18,000)

23
(61,000)
(mg/1)
30
410
110
73
156
48
98
0
8/6^
72

72
(18,900)

360
1969
8/14
80

23
(6,000)

102
(95,000) (27,000)

-
680
-
-
364
131
261
0

50
3,210
210
93
2,240
203
105


8/21
87

98
(25,900)

235
(62,000)

-
660
-
-
544
77
36
48 53

9/30
127

18
(4,800)

8
(21,500)

475
3,580
1,340
63
1,770
129
400
0
a/ High-strength sugar wastes being bypassed around treatment plant on or about these dates.   Results for these
   dates presumably reflect treatment performance only for'the raw wastes actually "reaching and passing
   through treatment system.

-------
                                                                     1-17
hydraulic loads to the treatment plant on certain days were in excess of
450 percent of the design criteria.  Organic loads frequently appeared
to be at least 50 percent higher than design specifications.  The study
concluded:

     1.   The existing aeration system was incapable of providing a high
          level of treatment for this waste.

     2.   BOD determinations were complicated not only by variability in
          amounts and character of the wastes but also by the presence
          of inhibiting substances.

     3.   Abundant growths of S. natans leading to bulking sludge, was
          one of the primary reasons for continued failure of the system.

     4.   Many operational problems were prevailing within the system.

     5.   Even though problems were identified, the investigators could
          see no practical solutions possible, other than costly pre-
          treatment and piping modifications.
                         TRICKLING FILTRATION


Fermentation

The Upjohn Companyj Kalamazoo3 Michigan

     Tompkins (17) reviewed performance of the trickling filter treat-
ment plant at the Upjohn Pharmaceuticals plant from 1948 through 1956.
The facilities were primarily receiving antibiotics fermentation wastes,
but also sanitary wastes.  A balance was attempted between the amount of
spent fermentation beers that could be accepted into the treatment works
vs that which were necessarily hauled to field disposal.  Edmondson (60)
indicated the possibility that substances in the antibiotic spent beers
would depress and/or be toxic to biological growths on the trickling
filters.

     The fermentation processes increased rapidly in the early 1950's.
The trickling filter plant was stressed under heavy loads at that time,
and, in spite of significant expansion of the biological treatment
facilities in 1953, hauling of spent fermentation beers had to be
continued.

     The treatment plant discharged to a small trout stream.  Plant
effluent was limited by the State to 45 kg (100 Ib) BOD/day, which was
significantly exceeded through most of 1951-1952.   The effluent was also
subject to chlorination provisions from May 15 to Sept. 15, yielding a
0.5 mg/1 minimum residual (17).

-------
118

                                                         Q
     The original plant had a maximum capacity of 2,650 m /day (700,000
gpd) with two 27 m (90 ft) diameter filters, operated both in parallel
and series, with a pair of primary settlers and a pair of final  set-
tlers.  The spent beers, before entering the main treatment plant,  were
held in two pre-aeration tanks, each 95 m  (25,000 gal).

     An antibiotic introduced in the spring of. 1952 created a spent beer
of 20,000 to 30,000 mg/1 BOD, whereas previous beers were about 5,000
mg/1.  This caused a serious treatment plant overload.  Hauling spent
beers to land disposal was greatly increased, and these wastes were
heavily chlorinated.  At times, the effluent of the treatment plant was
almost as high as the influent in TSS (17).

     Treatment plant expansion in spring 1953 added an additional
primary settler, converted the pair of final settlers to intermediate
clarifiers, added two final sedimentation basins, a new 37 m (120 ft)
diameter filter, and increased pumping capacity.  Improvements were also
made on the fermentation beer pre-aeration tanks, and a grit chamber was
added.  Even though the spent beers had a pH range of 2 to 11,  when
mixed with large volumes of washwaters the pH extremes were minimized
and pH adjustment was unnecessary.

     In spite of doubling the design capacity of the anaerobic sludge
digester, problems with solids existed from 1948 onward.  There were
breaks in the digester waste gas line, and a poor supernatant quality
from the digester was attributed to greasy sludges entering the di-
gester, thereby decreasing the solids settling ability.  Grease or lard
oil used as a defoamer in the antibiotics production decreased trickling
filter performance and caused shock loads.  The oils remained emulsi-
fied, causing solids buildup and ponding on the filters and signifi-
cantly hampered BOD reductions.  The use of Daphnia demonstrated the
presence of toxic wastes.

     Since 1953, BOD reductions have been 95 to 98 percent.  At an
average flow rate of 2,300 m /day (600,000 gpd) and 1,200 kg (3,080 Ib)
BOD/day to the treatment plant, final effluents contained an average of
41 kg (90 Ib) BOD/day.  Average influent and effluent BOD strengths were
respectively 600 mg/1 and 20 mg/1.  During the warm months, the unchlor-
inated final effluents amounting to 380 m /day (100,000 gpd) were used
for irrigation of lawns around the plant.  Besides other benefits,
chlorine consumption was reduced significantly.  Nevertheless, it was
reported that hauling of excess spent antibiotic beers was still
continuing through the mid-1950's (17).

Great Britain Plants

     In the mid-1960's, biological trickling filters used on mixed
wastes (from fermentation, and other industries) showed that the filters
may be loaded at about 0.6 kg/m  (1  Ib/yd ) BOD/day to achieve about 96
percent removal.  High-rate filters in series, with and without   -
circulation, have been successfully used with loads up to 1 kg/m


-------
                                                                     119


 (2 lb/yd3 BOD)/day on the first stage.  Resulting sludges are digested
with the production of fuel gas, and the digested sludges are dried,
 incinerated and/or disposed of by land spreading or landfill.

     Biological filtration was apparently preferred in Great Britain to
 the activated sludge process.  It was inferred that trickling filtration
may be somewhat less susceptible to shock waste loads, or become coloni-
 zed by micro-organisms more resistant to the antibiotic residues con-
 tained in the wastes.
Fermentation/Synthesized Organic Chemicals

The Upjohn Company* KalamazoOi Michigan

      In view of rapidly increasing production in the 1950's, the Upjohn
Pharmaceuticals plant evaluated many forms of waste treatment for
technical and economic feasibility.  Cushman and Hayes (57) reported in
1956  on pilot plant trickling filtration studies'using various combin-
ations of fermentation, synthesized organic chemical and sanitary
wastewaters from Upjohn.  Paradise in 1955 (68) noted that due to the
high  volumes and high BOD of the fine chemical wastes that the cost for
biological facilities would exceed $2 million and were therefore not
feasible.  Deep well disposal installed in 1952 also temporarily minimized
the need for additional treatment of the chemical wastes.  Nevertheless,
Paradise maintained these wastes would be amenable to biological oxida-
tion, and from the data obtained either two-stage trickling filtration
or two-stage activated sludge could give 95 percent reduction of the BOD
loads.

     The types of wastewaters received into the trickling filter pilot
plant included:

      1.   Synthesized organic chemical waste in the BOD range of 2,000
          to 10,000 mg/1, chiefly consisting of spent mineral acids and
          their salts, alcohols, chlorinated and unchlorinated hydro-
          carbons, organic acids, aldehydes, ketones, and residual
          fractions of heterocyclic steroids and their intermediates.

     2.   Sanitary wastes plus miscellaneous industrial wastes from the
          fine chemicals and antibiotic sectors, including varying
          amounts of spillage products, tank washings, breakage, filter
          cake, organic solvents, oils, greases and some fermentation
          byproducts.  This stream had a BOD strength of 200 to 1,400
          mg/1.

     3.   Spent antibiotic fermentation broth with a BOD range of 3,000
          to 15,000 mg/1 (57).

     Each waste blend was neutralized into the pH range of 6.5 to 7.0.
Trickling filter units were used in a two-stage setup with primary
settling, intermediate settling between the two filters,  and final

-------
120
settling following the second filter.  Pre-aeration of the feed tank to
the pilot plant was employed for four of the seven waste blends.   In-
fluents and effluents from the pilot plant were tested for antibacterial
activity using B. subtil is (neomycin test organism) and M. aureus ATTC
9114 (erythromycin test organism).

     The test results showed that the spent fermentation beers generally
contained ample nutrients for proper biological growths, and no sup-
plement was necessary.  However, the beer content had to be maintained
below 30 percent by volume in the total waste mixtures.  With the vari-
ous mixtures of chemical, sanitary and fermentation wastes, the incoming
BOD levels ranged between 550 and 4,080 mg/1, and BOD reductions were
93.9 to 99.5 percent.  Effluent BOD values were generally below 30 mg/1
except for the chemical-sanitary waste blend, which was much higher.
Influent nitrogen levels were in the range of 18 to 168 mg/1.  Cor-
responding nitrogen reductions for the various blends of wastes through
the two-stage trickling filter system were between 55.7 and 95.4 per-
cent, but generally in excess of 80 percent.  The pilot plant was operated
18 months.

     The study report concluded that all wastes investigated were treat-
ed by biological filtration to a high degree (57).
American Cyanamid* Willow Island* W. Virginia

     Fine chemicals and antibiotics waste treatment at the American
Cyanamid plant in the early 1950's was described in a series of three
papers by Vogler, Brown and Griffin (18).  Vogler outlined the treatment
system utilized for chemical wastewaters in 1950-51; Brown described the
expected characteristics of the new antibiotic aureomycin wastes; and
Griffin gave design details on the additional treatment facilities
necessary for the fermentation liquors.

     Synthesized organic chemicals production included the manufacturing
of melamine (used for melamine-urea and melamine-formaldehyde resins);
various pigments such as iron blue, chrome yellow, chrome green, and
molybdate orange; various brighteners; folic acid (which can be used as
a vitamin); and a host of pharmaceutical intermediates.

     The early treatment system was an earthen holding basin with a
surface area of 3.6 hectares (7.2 acres) and a capacity of 27.5 mg.
This system received sanitary, cafeteria, power plant and manufacturing
effluents, and the storm waters collected over the plant property.  At
an average wastewater flow of 12,700 m /day (3.4 mgd), the lagoon was
providing about 8 days' waste detention.  Final discharge was made to
Cow Creek, a small tributary which flowed about 0.8 km (0.5 mi) to the
Ohio River.

     Raw manufacturing wastewaters from the plant largely consisted of
spent mother liquors, filtrates, column slops, washing and cooling
waters.  Principal pollutants were organic byproducts, acids, alkalies

-------
                                                                      121


 and salts.   Total  waste  flow  reported as 12,700 m3/day  (3.4 mgd) in 1950
 was subsequently reduced to 9,400 m /day (2.5 mgd) by in-plant water
 conservation measures.   Raw wastes had high color and ranged in BOD from
 46 to 150 mg/1,  averaging 80  mg/1; pH varied from 2.5 to 8.0, averaging
 5.9; and  TS  and  TSS  approximated 1,500 mg/1 and 200 mg/1, respectively.
 The lagoon effluents were pale  green with average values of 33 mg/1 BOD,
 a pH of 7.6, and TS  and  TSS of  1,300 mg/1 and 100 mg/1, respectively.
 The lagoon influent  loading was about 1,000 kg  (2,300 Ib) BOD daily
 equivalent to 12 kg  (27  lb)/acre-foot of lagoon capacity/day.  Average
 BOD and TSS  reductions were about 50 percent.

      In 1950, the decision was  made to manufacture the  antibiotic
 aureomycin at the Willow Island plant.  Aureomycin (chlorotetracycline)
 was one of the first broad spectrum antibiotics available in the pharma-
 ceutical  industry.

      Aureomycin  wastes were divided according to source as follows:

      1.   Strong fermentation beers and filtrates of high BOD, ranging
          from 4,000 to  7,000 mg/1.  Aureomycin was fermented in a
          medium of  sugars, corn steep liquors  and nutrient salts.
          After  separation of mycelium, the fermentation broth is passed
          through layers of Magnesol which absorb the desired chemical
          values.  Absorption of B,p, a growth-producing vitamin, from
          the fermentation broth was cited.  The relationship of vitamin
          B|2 to aureomycin was not precisely defined.  Spent filtrates
          ana washes were reported discharged to the sewer, a source  of
          high BOD wastes.

      2.   Washings of equipment and plant floors ranging in BOD from  600
          to 1,500 mg/1

      3.   Inorganic  solids including filter cakes and precoat from the
          vacuum filters.  Precoat (Filter-Cel) and diatamaceous earth,
          serving as filter aids, both enter the sewer.

      4.   Chemical wastes, including spills of acids and alkalies, and
          spent  butanol  and brines, resulting from refining of the
          drugs.  Most butanol was reported as recovered and reused,  but
          a  significant  quantity was lost as wastage and/or contained
          within  the residual "heels" remaining after distillation.
          Butanol has an  extremely high BOD content, i.e.^0.68 kg H.5 Ib)
          BOD  for each kg  (Ib) of butanol lost.   The heels may contain
          up  to  500,000 mg/1  BOD and the total sewered  load from this
          source alone may amount to 1,400 kg (3,000 Ib) BOD daily.
          Associated  spent brines may total  54 metric tons (60 tons)/day,
          expressed  as the dry salt.

     ull fermentation production was expected to generate an average of
600 m /day (160,000  gpd) wastewaters  containing  3,600 kg (8,000 Ib) BOD
daily.  At this flow, BOD raw waste strength was calculated as 6,000
mg/1.  Because fermenters occasionally run  wild  and  may be suddenly dumped

-------
122

                                            0
the treatment plant was designed for 2,400 m /day (450 gpm) with a
hydraulic maximum of 3,300 m /day (600 gpm).  Mycelium and diatamaceous
earth removed by filtering of the spent fermentation beers were estimated
as amounting to 1,100 m  (600 ft /day) sludges at 8 to 10 percent solids,
requiring satisfactory disposal.

     Based on previous Company experience, trickling filtration was
selected as the prime means of treating the Willow Island fermentation
wastes.  Solvents mixed with salts from the refining end of the anti-
biotics process were judged adverse to biological treatment, and it was
therefore planned to sewer these streams into the existing chemical
waste treatment lagoon.  These streams were estimated to contribute 390
kg (900 Ib) BOD/day to the 390 kg/day already in the lagoon.  The
objective of 100 mg/1 BOD in the combined effluent flows of 19,500
m /day (5.16 mgd) gave a total plant allowable BOD load of 2,000 kg
(4,300 lb)/day, thereby leaving 1,100 kg (2,500 lb)/day of BOD as a
maximum coming from the fermentation treatment works.  The TSS effluent
objective was 80 mg/1.  The trickling filtration plant had to be designed
for an average BOD removal of 68.8 percent, which was not overly demanding.

     The sequence of treatment steps to properly handle the antibiotic
waste would comprise pH control by adding lime to the entering wastes;
pre-aeration and 12-hr wastewater holding within an equalization basin;
parallel 5 m (16 ft) diameter primary clarifiers; parallel 24 m (80 ft)
diameter high-rate biofilters; and parallel 5 m (16 ft) diameter secondary
clarifiers.

     Secondary sludges would be returned ahead of the primary clari-
fiers, and sludges removed from the primary settlers were to be pumped
and dewatered on a 2 x 2 m (6 x 6 ft) vacuum filter.  Lime and ferric
chloride were to be added to the sludges to aid in dewatering.  Fil-
trates from the vacuum filter would be returned to the pH control box.
The trickling filters were designed for an average recirculation ratio
of 3 volumes of filter effluent to 1 volume of raw wastes passing onto
the filters.  Dewatered sludges were to be disposed of to landfill.

     The final Willow Island plant effluents representing the combined
treated antibiotic and chemical waste flows would receive full chlorin-
ation before ultimate discharge to Cow Creek and the Ohio*River (18).


Fermentation and Biologicals

Lederle Laboratories, Pearl Rivev3 New York

     Since the end of World War II, the Pearl River plant has produced
antibiotics (penicillin and aureomycin), vitamin B1? and biologicals.
Brown classified the waste streams from Lederle (18; 101) as follows:

-------
                                                                     123
     1.   Strong fermentation beers with BOD's from 4,000 to 8,000 mg/1

     2.   Inorganic solids, including precoat and filter aid used in
          filtering out mycelium

     3.   Washings of floor and equipment contributing a large part of
          the hydraulic load and containing from 600 to 2,500 mg/1 BOD

     4.   Chemical wastes containing spent or lost acids, alkalis, salts
          and solvents.  A pound of solvent lost, such as acetone or
          butanol, can exert 0.5 to 0.7 kg (1 to 1.5 Ib) of BOD.  Salt
          loads can amount to many tons per day found in the waste
          streams.

     5.   Barometric condenser waters resulting from the evaporation and
          drying processes.  These waters usually are of extremely large
          volume and therefore are not generally discharged through the
          treatment plant.  The BOD of these waters can range from 60 to
          120 mg/1; these loads can be high.

     Solid wastes from the Laboratories including infected manures,
glassware, filter cakes, cafeteria and miscellaneous trash,  were sent
to a double-hearth incinerator.  Sewage sludge, previously dried to
about 70 percent moisture, was also received into this incinerator.
Through the late 1940's, spent penicillin fermentation broth was being
evaporated to give a slurry with about 10 to 15 percent solids.  This
was then spray dried to produce an animal feed supplement.  However, the
aureomycin spent broth when spray dried was found to carmelize, making
the procedure impractical.  Excluding weak sewage, the main process
waste stream from Lederle was about 1,600 m /day (0.425 mgd) with 2,500
mg/1 to 3,000 mg/1 BOD.  The raw waste equivalent was estimated in the
order of 3,600 to 4,500 kg (8,000 to 10,000 Ib) BOD/day.

     Following pilot plant studies for a high-rate trickling filter
treatment plant, the first 31 m (100 ft) filter was completed in July
1949, the second in July 1950 and the third in 1952.  These filters
operated in parallel following three 12 m (40 ft) diameter parallel
primary clarifiers.  A tank providing plain aeration and 4 to 4.5 hr
waste detention was installed ahead of the primary clarifiers in the
fall of 1950.  The remainder of the treatment plant, apparently com-
pleted during 1951, consisted of a storage tank for partial equalization
of raw wastes from the manufacturing sectors, followed by pH adjustment,
and grease removal by flotation.  These units preceded the 4.5 hr aer-
ation tank, the primary clarifiers and the three large trickling filters.
A 12 m (40 ft) diameter (single) secondary clarifier was added in the
early 1950's, together with chlorination facilities for disinfection of
final effluents.  The treated industrial effluents from Lederle were
subsequently delivered to the local municipality for further biological
treatment.

     Waste  studies on the treatment plant showed BOD removal through the
tank providing plain aeration was about 30 percent, with the influent

-------
124
and effluent averaging 2,500 and 1,750 mg/1, respectively, in December
1950.  Raw waste TSS was in the range of 1,000 to 1,200 mg/1, but 60
percent of the solids were nonvolatile.  This high ash content was due
to diatamaceous earth and filter aid lost from mycelium separation.   The
primary and secondary sludges, after dewatering via vacuum filtration,
at times reached 9 to 12 percent solids with an ash content as high as
70 percent, again due to abnormally high loss of fermentation filtered
materials.  Although the diatamaceous earth entrapped with the sludge is
a good soil conditioner, it is relatively damaging to pumps and the
cloths on the vacuum filters.  Analytical results were collected from
the Lederle treatment works in 1949-50, but this was before installation
of the third trickling filter and the final clarifier.  The following
averages were found:  raw wastes - 2,170 mg/1 BOD and 1,330 mg/1 TSS;
and the filter effluents - 812 mg/1 BOD and 225 mg/1 TSS (18, 101).

     By the late 1950's, the Lederle waste treatment plant had been
expanded into a two-stage trickling filtration system.  The trio of 31 m
(100 ft) diameter filters in parallel had become the primary biological
stage, whereas a single 41 m (135 ft) diameter filter, served as the
second stage.  This modified trickling filter facility was designed to
operate with a recirculation ratio of 3:1.

     Mauriello in 1958 (67) cited prevailing odor problems associated
with the treatment plant.  The anaerobic condition of the filters during
excessive loading represented one odor source of considerable concern.
Corrective measures included better flow equalization, increased recy-
cling of partially treated flows, more aeration capability and possible
addition of cooling towers to the system.  Greater control over concen-
trated wastes coming .down to the treatment plant was emphasized, es-
pecially concerning solvent recovery wastes, and spent fermentation
beers (which were no longer being dried for animal feed supplements).
Also suggested was separate treatment of intermittent small batch amounts
of deleterious waste.
Biologicals

Eli Lilly Greenfield Laboratories> Greenfield, Indiana

     In 1960, Howe explained that this type of pharmaceutical facility,
which produces antitoxins, antisera, vaccines and other substances for
the prevention and treatment of specific diseases, generally yields
large quantities of potent wastes (6).  At Greenfield, Howe indicated
that the veterinary and plant science research and production wastes
were pretreated before passing to the main treatment works.  Pathogenic
wastes were separately pasteurized, and the toxic wastes were pretreated
with lime and hypochlorite.  The antitoxin and vaccine wastes were taken
directly to the main treatment facility consisting of primary settling,
biological filtration, final settling, sludge digestion, and chlorin-
ation.  Animal wastes were handled by 8 m  (2,000 gal) septic tanks and
absorption fields.  Various waste solvents, screenings, animal carcasses
and miscellaneous solids were collected and taken to incineration facil-
ities for final disposal.

-------
                                                                     125
     The combined wastes went to an Imhoff tank; the overflows passed
to a trickling filter then a final settling basin.  Sludges from the
Imhoff tank went to drying beds.  Secondary-settled sludges and drainage
from the sludge drying beds were pumped back to the head of the treat-
ment system.  Sludges off the drying beds were either disposed of on
land or incinerated with trash, agg solids, manure and miscellaneous
solids.  The Company encouraged the use of spoiled eggs by nearby hog
farmers to decrease the wasteload on the treatment plant (31).

     Animals inoculated with pathogenic micro-organisms were isolated,
and after autopsy the bodies were incinerated.  Pathogenic wastewaters
from the isolation holding areas and the autopsy building were dis-
charged into a collection and holding chamber for future steam sterili-
zation at 93C (200F) and 30 minutes minimum detention.  Effluents from
the collection chamber were pumped to the main treatment plant.

     Toxic wastewaters from herbicide, insecticide and other research
were collected and given separate batch treatment before release to the
main treatment plant.  Dichromate solutions were reacted with ferrous
sulfate and lime to reduce hexavalent chromium to the trivalent form.
Phenolic wastes were reacted with lime and hypochlorite.  Waste DDT and
other insecticides together with solvent carriers, if available in
appreciable quantities, were collected in cans and then burned.

     Wastewaters associated with the production of smallpox, typhoid and
other vaccines were released to the central treatment works.  If waste
eggs could not be given away to nearby farmers, the eggs were crushed
and screened with the egg fluid passing to the treatment works.  The
screenings and egg shells were burned or buried with other wastes.
Sanitary waste from the 200 employees at Greenfield was also discharged
to the central treatment plant.

     The main treatment works in the late-1950's reflected some ex-
pansion from the early 50's, including provisions for prechlorination, a
primary clarifier, and conversion of the previous standard-rate trick-
ling filter to a high-rate filter.  The^primary settler was 53 m
(14,000 gal) and the final settler 27 m  (7,000 gal), both quite small.
Final chlorination remained optional (31, 70).  Waste loads entering and
leaving the treatment facility during the late 1950's are shown in Table
VIII-5.
     Bloodgood (53) in 1966 described the next stage in progression of
the Greenfield Laboratories.  Employment increased to 500 persons and,
besides main production capabilities, four peripheral areas of research
were added:  veterinary science, animal nutrition, plant science, and
human medicine.  New construction was initiated to conduct toxicity
studies and tissue culture work in the area of human medicine.  Blood-
good describes the new Greenfield installation as comprising five sep-
arate waste treatment plants, i.e. Plants 238, 237, 236, 226 and 219.
Plant 219 is the largest of these facilities and is essentially the same
as described above consisting of a primary clarifier, a high-rate tritk-
ling filter, secondary clarifier and sludge digester.

-------
                                                                                                                                                       CTl
                                                                     Table \llll-S



                                                  Eli Lilly Greenfield Laboratories' Wasteloads (70)
Influent
Plant Area
Anti sera/antitoxins
production
Additional anti-
influenza vaccine
production
Plant science and
animal research
Monkey storage areas
Total (avg)
Flows

BOD TS
(m3/day) (gpd) (m3/dS) (gpd) (m^ > ^d^ (1b/da*> ^ > (^/da^ <1b/da*>
204 54,000
250 66,720
204 54,000
18 4,700
680 179,400 1
655 13
383 101
708 187
83 22
,820 480
,000 200-400 45-73 100-160 760-1,520 154-308 340-680
,000 40-80 9-23 20-50 100-200 27-50 60-110
,000 810-1,620 186-367 410-810 1,220-2,450 249-499 550-1,100
,000 800-1,600 14-27 30-60 770-1,540 136-272 300-600
,000 380 254 560-X 860 567 1,250^
Settleable
Solids
(kg/day) (Ib/day)
41-77 90-170
4 5-14 10-30
64-354 140-780
136-272 300-600
122 27(X

Effluent
Parameter
BOD mg/1
TSS mg/1
TS mg/1
Coll form plate count
pH
Present (late-1950

18-40
20-60
400-475
400 to 1,000/100
7.3 to 8.0
's)

ml 400
Future

10-20
20-40
400-450
to 1,000/100 ml
7.3 to 8.0



a/   Maximums of 1,080 Ib/day BOD, 2,690 Ib/day TS, and 1,080 Ib/day settleable solids

-------
                                                                     127
     Plant 219 affords both primary and secondary treatment principally
serving the older portions of the facilities.   Animal  wastes constitute
a major part of flows received into this treatment works.   Flows average
303 m  (80,000 gpd) with influent and effluent BOD's of 105 mg/1 and 35
mg/1, respectively.  Steam-sterilized pathogenic wastewaters from
animal isolation sectors, which were cited above as entering the plant
sewer, actually enter the Plant 219 system.  Treated flows from giant
Systems 219 and 226 (described below) are combined, giving 946 m
(250,000 gpd) to 1,140 m  (300,000 gpd) of effluents with an average BOD
of 13 mg/1.  Final comeosite discharges in the mid-1960's were estimated
at about 1,140-1,320 nT/day (300,000 to 350,000 gpd).  3Bloodgood (53)
stated that they were prepared to handle up to 3,780 m  (1 mgd) of total
wastewater flow in the future at this location.

     The other treatment systems primarily handled animal wastes in-
cluding excreta, spilled feeds, washings and other losses, arising from
horses, cattle, calves, pigs, rabbits, mice, monkeys, dogs, and cats.
The wasteloads vary from 0.8 to 78 I/day (0.2 to 20.0 gpd) per animal
and from 0.007 to 2.9 kg (0.016 to 6.50 Ib) BOD/day/animal, ranging from
the smallest to the largest animals.  Wastewaters from "normal" animal
holdings are discharged into 7,600 1 (2,000 gal) septic-holding tanks;
excess overflows are discharged to absorption trenches within outlying
agricultural fields.  The septic tanks are cleaned annually or bian-
nually and the resulting sludge is disposed of on nearby lands.  The
septic tank effluents originating from the "sick" animal isolation barn
are diverted into the wet well of the treatment plant for prechlor-
ination and subsequent bio-oxidation.  The various peripheral treatment
systems are tabulated in Table VIII-6.
Drug Formulation

     j Sharpe3 and Dolvne3 West Point, Pennsylvania
     In 1951, the West Point facility consisted of the main warehouse,
the pilot plant and fermentation area, the boiler house, the research
laboratories, the synthetic chemical process building and the blood
plasma processing laboratories.  Liontas (34) reports that a high-rate
trickling filter installation was completed in late 1950 - early 1951 at
the West Point Pharmaceuticals works.  Wastewaters were made up of
organic and inorganic acids, fermentation wastes, salts, acidic sol-
vents, and various aliphatic and aromatic organic chemicals.  Sanitary
wastes were also received into the biological treatment works.  From
available data, fermentation wastes are thought to comprise only a small
portion of the entire plant wasteload.

     The plant had three sewer systems:  one for storm and cooling
waters, one for sanitary wastes, and one for industrial wastes.  Storm
and cooling waters are discharged directly to the receiving stream.
Some of the industrial wastewaters flow directly to .the trickling
filtration treatment works, whereas others first receive special hand-
ling and pretreatment.  Pretreatment of selected industrial wastes was
intended to protect the biological system.   The wastes are essentially

-------
128
                            Table VIII-6
             Eli Lilly Animal Waste Treatment Facilities
                       Greenfield, Indiana (53)
  Plant       Type and Source
                 of Waste
                                   Treatment
   236
 Swine  Barns
   226
   237
  238
Toxicology Building,
Tissue Culture complex,
Agriculture and
Industrial Products
buildings, Veterinary
Isolation Barn, En-
gineering and Main-
tenance shops.

Feed mill and wash-
downs from chicken
barns
Sheep and cattle
barn
 Flow =  19 m3  (5,000  gpd) est.
 Influent and  effluent
 BOD  = 510 and 75 mg/1,  respectively
 Screening,  holding tank
 (56.8 m3; 15,000 gal),
 aerated tank  (90.8 m3;  24,000 gal),
 settling tank (26.5  m3; 7,000 gal),
 final effluent to seepage
 Flow =  660  m3 (175,000  gpd)
 Segregation of bedding, screening,
 dual  aeration and settling tanks
 (980 m3; 260,000 gal),  sludges to
 underground (340 m3; 90,000 gal)
 sludge  digestion tank,  overflows
 to (1,360 m3;  360,000 gal) earthen
 lagoon, cooking of selected
 pathogenic  wastes.
 Two  underground holding
 tanks in series, first  aerated
 (49.2 m3; 13,000 gal);  second for
 settling (34 m3; 9,000  gal); to a
 stone filter  bed then to ground

 Underground tank subdivided into
wet well (6.2 m3; 1,650 gal) and
 sludge digestion (69.3 m3;  18,300 gal);
from wet well  to lagoon (378 m3;
 100,000 gal),  final  percolation into
ground.

-------
                                                                     129
equalized, and the acidity is neutralized with concurrent precipitation
of sludge containing heavy metals.  The sludge slurry is pumped to
sludge lagoons for solids concentration.  The strong lagoon supernatants
and underdrain filtrates, containing about 10,000 mg/1 BOD, are sent
back to the treatment works.  This represents a major load imposed upon
the treatment plant.

     The waste treatment works consisted of:  a two-stage, high-rate
trickling filter preceded by a comminutor, primary wet well and primary
settling; followed by secondary settling, intermittent sand filtration
and final chlorination.  Thre was no intermediate settling between the
two filter stages, but a secondary wet well was available for waste
application to the second-stage trickling filter.  A portion of the
first stage filter effluent is returned to the primary wet well as
recycle to the first-stage filter, the remainder passing to the se-
condary wet well.  Part of the effluent from the secondary settler is
returned to the secondary wet well for recycle to the second-stage
filter.  Nonrecycled effluent from the second-stage filter passes to an
alternating dosing chamber which intermittently doses the sand filter
beds.  Sludges from the secondary settler are drawn into the primary wet
well and theoretically removed from the primary clarification unit.  The
primary sludges are conveyed to the industrial pretreatment equalization/
neutralization tanks.  The (precipitated) sludges are then sent to the
sludge lagoons.  Lagoon solids are periodically removed and disposed of
(34).

     Design of the biological treatment plant was based on the following
loading criteria:  flow - 380 nT/day (113,000 gpd); BOD - 1,450 mg/1
-V620 kg H.360 Ib) BOD/day; TSS - 220 mg/1; filter recycle ratio of 3:1
for the primary stage and 2:1 for the secondary stage.  The total cost
of the treatment plant was approximately $600,000, including the sewers
and engineering services.  When the facility began operating in November
1950, it received only sanitary sewage.  Sludge lagoon supernatants were
introduced to the treatment works in September 1951.

     During 1952, an average of 330 m /day (98,000 gpd) wastewater
containing 330 kg (725 Ib) BOD/day was sent to the biological treatment
works (85 percent of hydraulic design load, but only 53 percent of
organic design load).  Overall average BOD reduction was greater than 98
percent through treatment, and TSS removal was 91 percent, giving an
average of only 2.5 kg (5.6 Ib) TSS/day in the final effluents.  Annual
O&M cost in 1952 was $40,500.  It also cost $1.10/1,000 gal waste
treated which equates to $4.70/lb BOD removed.  These costs were high be-
cause full Pharmaceuticals production had not yet been reached at West Point.

     Elimination, control and/or pretreatment of industrial wastes were
closely engineered into the process operations.  For example, highly
viscous organic chemicals were collected into drums and burned, and
blood wastes were incinerated.  Once, the pilot plant wanted to discard
a fermenter loaded with Terramycin, but the antibiotic would have likely
disrupted the biofilters.  Since it was found that high alkalinity
inactivates the antibiotic, sodium hydroxide was added to the fermenter
before discharge to the sewer.  Also, the pilot plant can shift complex
or difficult wastes to land irrigation or land disposal.

-------
130


     In colder months, such as January 1952, overall  BOD removal  dropped
about 5 to 7 percent below the average.  When the plant was starting up
recycling was too high, causing significant icing in  the system,  so the
recycle rates during extremely cold weather were reduced.  The filters
could withstand shock loads with as much as a 200 percent variation in
BOD loads from one day to the next.  Although the treatment facilities
were operating well under the design levels, occasionally the filters
were exposed to loadings in excess of twice the design criteria.   Large
quantities of antiseptics and antibiotics twice reached the treatment
system causing appreciable sloughing of the filters,  but in a few days
conditions appeared to return to normal.

     Nitrogen data collected in July 1953 showed a consistent drop in
nitrates across the first-stage filter, but then a significant increase
occurring across the second-stage trickling filter and the sand filter.
Effluents contained 3.3 to 8.2 mg/1 nitrate (N) compared to 1.0 to 2.6
mg/1 nitrate in the incoming raw wastewater.  The Company described the
biological treatment works as giving excellent performance (34).
                        OTHER TREATMENT METHODS

     A number of special treatment methods are practiced at pharma-
ceutical plants.  These methods, discussed below, include: anaerobic
filtersj spray irrigation, oxidation ponds3 sludge stabilization^ and
deep well, injection.

Anaerobic Filters

     Laboratory-scale anaerobic filters were studied by Dennis and
Jennett (1) in treatment of pharmaceutical wastes.  These strong wastes
were ordinarily sent to the city of Springfield, Mo. municipal treatment
works.  The anaerobic filters were intended to establish pretreatment
capacity for the industry prior to municipal discharge.  Anaerobic
filters were also selected because of the relatively low excess solids
generated by such treatment.  The waste was found deficient in nitrogen
and phosphorous for purposes of biological treatment, so nutrients
were added.

     COD was used as the major parameter during the study.  The waste
stream contained about 16,000 mg/1 COD, less than 50 mg/1 ammonia and
organic nitrogen, less than 1 mg/1 total phosphorous, 30 mg/1 TSS, and
trace amounts of heavy metals.  The wastewater contained about 1 percent
methanol, which accounted for the very high COD concentration.  The feed
to the filters consisted of the pharmaceutical wastes diluted downwards
to between 1,000 to 8,000 mg/1 COD.

     The soluble wastes could be anaerobically treated without the need
for solids recycling.  Four plastic columns were used 15.2 cm (6 in) in
diameter by 1 m (3 ft) high, each having a 14.3 1 (3.8 gal) capacity.
The filter media was quartzite stone, 2.5 to 3.8 cm (1.0 to 1.5 in) in
diameter.   The filters were started up on a mixture of sewage sludge,
glucose and nutrients before introduction of the pharmaceutical waste.

-------
                                                                     131
Complete conversion to the pharmaceutical wastes was made after about 25
days.  Full acclimation was considered achieved when constant gas
production and a high COD removal were recorded, which occurred about
the 40th day of testing.  The feed was brought into the bottom of the
filters and the system(s) was maintained at 35 to 37C (95 to 99F).

     Detention ranged from.,12 to 48 hr and organic loads were varied
from 0.22 to 3.52 kg COD/nT (13.8 to 220 lb/1,000 ft3) filter/day.
Effluent TSS for all filters was generally below 50 mg/1.  COD removals
were normally 90 percent or more.  However, for loads over 1.76 kg
COD/nT (110 lb/1,000 fr)/day the effluent COD's were greater than 200
mg/1 and effluent quality was considered deficient.  Loading changes
caused initial drops (less than 10 percent reduction) in COD treatment,
but the filters were capable of readapting nearly to previous conditions
within 7 to 21 days.  These results demonstrate the ability of the
filters to operate successfully under shock loading conditions.

     Biological solids taken from the inside of the filter showed very
high settleability.  The supernantants containd low TSS, and these
solids in turn contained less than 3 percent volatile content.  Settling
of the filter effluent seemed necessary.  Although the influent con-
tained small amounts of toluene, this compound was not removed by the
anaerobic filters.  Furthermore, the toluene imparted a distinct odor
which pervaded the treatment system and effluents.

     The study concludes that besides low amounts of solids generated,
for organic influent loadings in the range studied and at waste strengths
greater than 1,000 mg/1 COD, the anaerobic filters were capable of COD
reductions from 93.7 to 97.8 percent.  The filters could be operated  for
periods up to six months without needing filter solids unloading.  The
filters also appeared to recover quite rapidly from shock organic load-
ings (1).
Spray Irrigation

The Upjohn Company* Kalamazoot Michigan

     During the 1950's, severe waste abatement problems led this Company
to investigate different treatment methods, one of which was spray
irrigation.  Excess spent fermentation beers had to be hauled from the
plant to external land disposal sites.  The merits of spray irrigation
were reported by Colovos and Tinklenberg of Upjohn in 1962 (43).

     Initial spraying tests, conducted on land behind the manufacturing
plant, produced mixed results.  In the early tests using straight fer-
mentation beer liquors, odors developed from the beers, from the wind-
carried sprays, and from anaerobic ponding.  Pretreatment with chlorine
was selected as the best means to reduce these odors.  Amounts of chlor-
ine found necessary were about 500 mg/1 for the steroid beers and 700
mg/1  for the antibiotic beers, compared to a total chlorine demand in
the range of 1,000 to 2,000 mg/1.  These figures indicate high chlorine
consumption.

-------
132
 More extensive spray irrigation tests were carried out on unused
farmlands about 22.5 km (14 miles) from the plant.  The ground cover
consisted of the normal weed and grass growths normally found on farm-
lands in the midwest, not tilled for many years.  The upper 20 to 25 cm (8
to 10 in) of soil was a Fox loam (grayish-brown to brown gritty loam),
underlain by 25 to 30 cm (10 to 12 in) of yellowish to light brown loam,
with 0.3 to 0.9 m (1 to 3 ft) of reddish-brown gravelly sandy clay
beneath.  After tall grass and weeds had been mowed, a dense ground
cover was obtained.  Orchard irrigation equipment (aluminum pipe and
rotating sprinklers) was used.  Chlorinated beers were stored in a 114
m  (30,000 gal) underground tank.  A maximum of 22 rotating sprayers
were employed, each rated at 83 1/min (22 gpm) and with a nozzle pressure
of 4 kg/cm  (60 psig).  The total assembly was capable of handling about
110 m   (29,000 gal)/hr of spent beers.

     Results of spray irrigation studies were mostly qualitative rather
than quantitative.  Chlorinated beers were sprayed onto a particular
area until the equivalent of about 3.8 cm (1.5 in.) of liquid were
applied.  The sprays were then transferred to another sector and re-
turned  to the original area about one month later.  Observations indi-
cated no anaerobic conditions associated with pooling or puddling, which
was not true with the unchlorinated beers.  An, adequate ground cover was
considered important in promoting percolation and minimizing runoff.
When parched grass was sprayed, the grass recovered rapidly and turned
dark green.  Areas sprayed in the fall were judged in very good con-
dition  the following spring, with very dense and dark green ground
cover.  Wintertime spraying did not appear to induce any undesirable
effects, and no special problems were encountered.  Further testing with
barley  ground cover at the spray application rate of 0.64 cm (0.25 in)
per day, or 66 cm (26 in) applied over 5 months gave mixed results,
probably indicating an excessive application rate.

     Cost of the system exclusive of pumps, storage tank and land was
$1,070.  Consumption of chlorine was 0.48 to 0.72 kg/m  (4 to 6 Ibs/
1,000 gal) equivalent to $0.05 to 0.08/m  ($0.20 to $0.30/1,000 gal) of
spent beer sprayed.  Further studies were to include drilling of test
wells to discern the presence or absence of ground water contamination,
adaptability of other types of ground cover, maximum rates of appli-
cation, and long-term effects of high chloride content of the wastes on
the soils.  It was concluded that spray irrigation of spent fermentation
beers,  based on the evidence gathered up to that time, was proving very
satisfactory (43).
Oxidation Ponds

Studies in India

     Laboratory and field-scale studies were conducted in India on
treating synthetic drug waste mixtures by pure cultures of algae (39).
Reported results were somewhat difficult to interpret and therefore only
summary findings are given.

-------
                                                                     133
     In a series of experiments, synthetic drug wastes were diluted with
water and settled sewage in the range of 1:100 to 1:2 (drug waste:  water
+ sewage).  The objective was to obtain the least dilution whereby  the
organic and toxic levels would be maximum but still  capable of sup-
porting a healthy algae population over a reasonable period of time.
Harvesting of the algae was not mentioned.

     The various waste mixtures were inoculated with pure cultures  of
Chlorella pyrenoidosa and Scenedesmus quadricauda.  In some cases,
acidic pH levels were adjusted before algae inoculation, whereas in
other cases it was not.  Dissolved oxygen, pH, BOD and color of the
waste mixtures were measured at various intervals.  Maximum waste hold-
ing and treatment was 10 days.

     Synthetic drug wastes diluted 20 times with water and settled
sewage gave initial BOD's in the range of 120 to 352 mg/1 (at 37C; 99F)
for the various mixtures.  S. quadicauda was generally more efficient in
waste reduction than C. pyrenoidosa.  Interference in some of the DO
tests was caused by a deep pink color in the raw wastes due to the
presence of p-aminophenol.  Fish (species unidentified) were acclimated
to the drug waste mixtures and during acclimation could survive at  BOD
concentrations at about 60 to 65 mg/1.  After full acclimation, they
were capable of survival in the 1:20 synthetic drug diluted waste mixtures.

     In general it was concluded that treatment of these synthetic  drug
wastes by algae oxidation ponds would be very difficult.  At a 1:20
dilution (1 part synthetic drug wastes: 20 parts water + settled sew-
age), Chlorella and Scenedesmus were kept viable for 5 to 8 days, and
BOD's were reduced from 310 to 340 mg/1 to 30 to 40 mg/1, or 85 to  93
percent BOD reduction.  However, reductions varied widely within the
individual experiments.

     In view of the large dilutions and land areas required, this treat-
ment method presented some formidable problems, solutions to which  were
considered expensive (39).


Sludge Stabilization

Lederle Laboratories* Pearl River,  New York

     In 1953-54, Lederle Laboratories initiated pilot plant evaluations
on composting and reusing excess organic sludges from overall process
operations (61).  Organic sludges accumulated from the fermentation of
antibiotics, extraction of active ingredients from animal organs,
housing of animals, and sanitary sewage from 4,000 persons, with most of
the sludge coming from the secondary treatment works.  In the mid-1950's
5 to 9 metric tons (6 to 10 tons) of wet sludge were being processed
daily.   Given available land, it was hoped that composting would solve
sludge disposal problems.

-------
134
     Early results showed that one of the best blends for good composting
and suitable physical handling was a mixture of 65 percent treatment
plant sludge, 25 percent animal manure, and 10 percent sawdust.  Before
being added to the compost, the treatment plant sludge was dewatered,
using lime, on vacuum filters.  Phosphate rock was added to the compost
mixture, giving a moisture content from 50 to 60 percent.  Both moisture
and temperature were important in composting.  The materials were
repeatedly shreddded, .blended and turned over.

     In 1955, the Company converted 4 ha (10 acres) of open landfill
to full-scale composting.  Windrows, made up of shredded, blended and
properly-mixed materials, were about 7.6 m (25 ft) wide, as much as 4.6 m
(15 ft) high, and ran up to 91 m (300 ft) long.  They were allowed to
compost for 5 months or longer.  The aged compost was eventually trucked
into an enclosed building and reshredded, giving a final moisture con-
tent of about 40 percent.  The windrow method had a capital investment
of only about $10,000 and an annual operations expenditure of about
$15,000.

     In final form the compost was a granular, free-flowing, dark brown
material with a slight humus scent.  The finished compost was reportedly
used on Lederle grounds, by adjoining communities, and sold through a
local outlet for lawn treatment and soil conditioning.

     By the mid-1960's, organic sludge at Lederle had increased to about
18 metric tons (20 tons) daily, consisting of 10.9 metric tons (12 tons)
of wet biological sludge from the 3,785 m /day (1 mgd) industrial waste
treatment works and 7.3 metric tons (8 tons) as antibiotic cake residues,
animal cage wastes and manure (15).  Windrowing worked well for some
time, but eventually the local commercial product was phased out.
Procedures were changed, whereby the compost was held in windrows for  up
to 2 years and the materials were turned only every 6 months.

     Eventually, whether due to production shifts creating greater amounts
of acid sludges or, more probably, due to the natural encroachment of
homes around the Lederle property, odors arising from the composting
became a pubjic problem for the Company.  In 1968, the Company turned
the piles in the fall and winter and reverted to smaller-sized windrows
with more frequent turning.  These changes were only partially effective.

     After 15 years, composting was to continue, but only if the odor
problems could be solved.  Economically alternative means of disposal
were being sought.  Compared to other methods of solid waste disposal,
composting has cited advantages of low capital and operations expense,
minimal air pollution problems and the conservation of natural materials.
Disadvantages include the greater land requirement, potential odors, and
the absence of ready markets for the final product.  As a fertilizer,
unsupplemented compost is rated quite low in N:P:K content, i.e. 2:1:1.
Thus for buyer education, emphasis should be placed on compost as providing
trace elements, organic matter, chelating agents, and perhaps growth
factors that are not found in commercial fertilizers (15).

-------
                                                                     135
Deep Well Injection

Parke-Davis Company, Holland, Michigan

     At Parke-Davis, concentrated chemical manufacturing wastes are
injected by a deep well system into a limestone strata 427 to 518 m
(1,400 to 1,700 ft) below ground surface (see Waste Characteristics
Section for nature and strength of this^waste).  Maximum pumping rate of
injected chemical wastes is about 163 nT/day (43,000 gpd) (13, 59).

     Before the Company committed itself to the deep well system in the
early 1950's, pilot plant studies served to eliminate other waste
handling/treatment alternatives including waste neutralization, activated
carbon, evaporation of the liquors, and trickling filtration.  Since the
chemical wastes contained high dissolved salts and had considerable
buffering capacity, neutralization was judged too expensive and would
not cope with other objectionable properties of the process wastes.
Activated carbon could reduce the color level of the wastes from 11,000
Pt-Co units to 2,000 units, but this was still much too high.  Activated
carbon was also prohibitively expensive in view of the high O&M costs
and remaining problems of toxicity and BOD.  Evaporation necessitated
difficult residue disposal and high fixed and operating costs.  However,
in the event of a breakdown in the well disposal works, the Company de-
cided to have a converted evaporator system available as a standby.
Trickling filter studies indicated at least a 50 percent reduction in
BOD, but removal of color, toxicity and salinity were unsatisfactory.
It could not compare favorably with deep well disposal chiefly because
of high initial investment, the filter may be incapable of handling
wastes from new products, and in the event of plant shutdown the filter
growths could experience complete dieoff.

     In the early 1960's, the concentrated chemical wastes were pumped
from the process reactors to a 303 m  (80,000 gal) tank for^settling.
This effluent was cascaded" through a second and third 303 m  tank, then
passed through sand gravity filters and injected underground by a pair
of deep disposal wells.  Compatibility tests of the wastes with the
limestone and brine in the underground formation were conducted before
the wells were put into use.  The tests have been rerun each time there
is a major process change inside the manufacturing plant.

     The formation is 10 to 15 percent voids.  The Company believes it
has benefitted from the large quantity of acetic acid pumped, which has
apparently opened up the strata even more.  The quality of the chemical
waste sewer system is checked with COD rather than BOD tests.  Cor-
relations have seemingly been obtained between the COD and the BODC and
BOD2Q results.                                                    b

     The first well is 436 m 1,432 ft) deep and began in 1951.  Initial
pumping pressure was 21 kg/cm  (300 psi) which has dropped to 18 kg/cm
(250 psi), and well pressure at rest is 14 kg/cm  (200 psi).  Production
increases necessitated a second well in 1956, which is 503 m (1,649 ft)

-------
136

                                                   2
deep.  Initially, the puwping pressure was 21 kg/cm  (300 psi) but then
rapidly fellpto 11 kg/cm  (150 psi), and the well pressure at rest is
only 4 kg/cm  (50 psi).  These pressure heads appear abnormally low.
The pumps used for waste injection are triplex, positive displacement,
porcelain-lined units.

     Advantages of the deep well system are reported to be:  low cost of
installation, maintenance and operation; and complete elimination of
hard-to-dispose-of wastes.  Reported disadvantages are:  what is occur-
ring cannot be seen; a high element of risk if the area has not pre-
viously been surveyed by the drilling of several other wells; and the
possibility of contaminating the underground strata.  The injection well
system is reported practical when the waste volume is not excessive, the
underground strata has the proper characteristics, chances of under-
ground contamination are slight, the wastes are compatible with and can
be adequately assimilated into the strata, and no other method appears
possible or feasible to handle the waste problem' (13, 59).

     A recent appraisal of the Parke-Davis deep well disposal works
indicates this system may need to be upgraded with respect to the EPA
Administrator's Decision, Statement No. 5, made in 1973 on deep well
waste injection.  This system, unless supported by additional data, may
not be in accord with the present philosophy of alternately treating and
disposing of the subject waste by surface means.  Past injection history
shows that fracturing of the underground strata has indeed occurred,
with possibile vertical fracturing of the overlying shale layers or the
confining aquaclude.
Abbott Laboratories* Baroeloneta^ Puerto Rico

     Plans in 1968-69 for the new Abbott antibiotics facility in Bar-
celoneta, Puerto Rico more or less stipulated there would be no external
waste discharges from the fermentation sectors (22).  Spent fermentation-
beers, mycelium and solvents would be fully recovered.  Miscellaneous
process wastes like floor washings, ion exchange rinses, and tank
cleanings originating from the antibiotics or other processing areas
were to be merged with boiler blowdown, tower blowdown and the plant
sanitary sewage, and then discharged into deep injection wells.  It was
reported the only discharges leaving the plant would be clean waters
originating as roof drainage, surface runoff occurring during periods of
rainfall, and some miscellaneous clean or cooling waters without organic
content.  The injection well was to be drilled to a sufficient depth
until the salt content in the underground aquifer reached a minimum of
5,000 mg/1.  The Company was prepared to go to a well depth of 762 m
(2,500 ft) costing $100,000.  There is strong indication these waste
treatment and disposal plans were changed substantially before the
startup of full-scale process facilities.
The Upjohn Companyf Kalamazoof Michigan

     In the early 1950's, Upjohn's production was greatly outdistancing
the capability of its waste treatment facility to handle the wasteloads

-------
                                                                     137
(19, 68).  The synthesized organic chemicals manufactured included
cortisone, hydrocortisone, folic acid, adrenal  cortex and other cortical
steroid products.  The chemical wastewaters with up to 9,070 kg (20,000 Ib)
BOD/day and considerable organic solvents were  discharged to a 36 ha (90
acre) swamp on Company property, a method long  obsolete, and producing
serious odors.  Besides, additional manufacturing of newer steroids was
anticipated.  The table below shows a preliminary inventory of wastewater
discharges from the synthesized organic chemicals manufacturing.


Parameter                          Content

Flow      1,510 to 2,650 m3/day.(400,000 to 700,000 gpd)
BOD       2,000 to 10,000 mg/1-' = 2,720 to 13,600 kg (6,000 to 30,000 lb)/day
TDS       500 to 7,000 mg/1
pH        2.0 to 8.0, generally acidic
Color     Highly visual
Odor      That of organic solvents

a/ Waste streams of up to 500,000 mg/1 BOD not  uncommon.

     The liquid waste streams were laden with spent mineral acids and
their respective salts, short- and long-chain alcohols, both chlorinated
and unchlorinated hydrocarbons, various organic acids, aldehydes, ke-
tones, and the residual fractions of the steroids and their interme-
diates.  The survey showed high variability in  wastewater composition
from day to day, and new processes and modifications were causing steady
increase in the volume and strength of wastewaters.  Two programs were
undertaken to cope with the pervading waste disposal problems:  a
preventive waste abatement program; and a search for new waste treatment
methods, this phase culminating in a 57,0 m /day (150,000 gpd) deep well
disposal system.

     Major procedures established under the preventive waste abatement
program included:

     1.   Separating clean spent waters from contaminated wastewaters
          and discharging the former to the storm sewer
     2.   Recovering solvents wherever practical
     3.   Collecting and hauling selected high  organic wastes to land
          disposal
     4.   Collecting and incinerating non-reusable combustible solvents
          and residues
     5.   Exhausting steam jets to the atmosphere
     6.   Recycling seal  waters on the vacuum pump systems
     7.   Providing bypass bleeder lines on drinking fountains
     8.   Employing good housekeeping practices such as shutting off
          drinking  fountains  during night hours and weekends,  turning
          off hoses when not  in use,  and repairing leaking valves.

-------
 38                                                                  138
     The preventive program was surprisingly effective.  From September
 1953 to April 1954, the raw chemical wastewater flow was reduced from a
 peak of 2,840 mj/day  (750,000 gpd) to 284 nT/day (75,000 gpd).  The
 corresponding drop in daily BOD loads was from 13,600 kg (30,000 Ibs)
 /day at peak to 1,360 kg  (3,000 lbs)/day in March-April 1954.  Waste-
 water volume reduction was in large part attributed to the many steam
 vacuum jets used  in manufacturing.  These jets were made to discharge to
 the atmosphere, dispensing with water contact condensers.  Reduction in
 BOD wasteloads was primarily attributed to the newly installed spent
 solvents incineration system.

     During the preventive program, various treatment was evaluated, in-
 cluding biological degradation, evaporation and concentration, under-
 ground disposal into  the  brine strata about 396 m (1,300 ft) below
 grade, spray irrigation,  chemical treatment, carbon absorption, con-
 trolled use of the swamp, and hauling all wastes from the site.  From
 these possibilities,  and  due to the waste volume reduction experienced
 during the waste  preventive program, deep well injection was selected as
 most feasible.

     To start the well program, one test well was drilled and found to
 have an injection capacity of at least 378 m /day (100,000 gpd).  Fol-
 lowing this, two  full-scale waste disposal wells were drilled into the
 brine strata at respective depths of 467 and.,450 m (1,532 and 1,476 ft).
 This system was designed  to dispose of 568 m /day (150,000 gpd) of
 synthesized organic chemicals process wastes (19, 68, 72).

     Final segregation and collection of spent flows resulted in the
 following waste handling  and abatement system:
                  2
     1.   3,785 m /day (1.0 mgd) of uncontaminated cooling waters from
          various coolers and condensers were released to a storm sewer
          and then to the swamp  a 36.5 ha (90 acre) existing pond.

     2.   Strong  process  and sanitary wastes with an average flow of 284
          m /day  (75,000  gpd) and a maximum flow of 568 m /day (150,000
          gpd) and containing about 1,810 kg (4,000 Ib) BOD/ day, were
          forwarded to the deep disposal wells.
               2
     3.   150 m   (40,000  gal)/month of solid waste, including chemical
          sludge, was hauled away by tank truck and spread onto nearby
          lands.
              3
     4.   llm /day (3,000 gpd) of nonreusable waste solvents directed
          to a specially designed solvent-incinerator, the waste heat
          from which was reused for incineration of animal  refuse.

     In disposing of the strong process wastes, pretreatment was neces-
sary before deep well  injection.   Liquid wastes were combined in a
vitrified tile sewer resistant to acid, alkali  and solvents.   Pretreat-
ment consisted of pH adjustment with high calcium lime followed by
flocculation,3settling in a 227 mj (60,000 gal) clarifier,  a  first  suKge
tank of 219 m  (58,000 gal),  filtration,  a second surge tank  of 219 m3
(58,000 gal),  and the  pump station.   A stainless  steel  vertical  leaf

-------
                                                                     139
Niagara pressure filter was used for filtration.   Diatamaceous earth was
used to precoat the filter leaves.  Suspended solids were reduced in the
prefilter flows from 100 mg/1 to less than 10 mg/1 in the filtrate.
Filter cake was sluiced and combined with chemical sludge from the
clarifier to a dewatering pit, the material from which was eventually
hauled away by tank truck for field disposal.  The final  pumping system
used two plunger pumps (with provisions for a third pump  if necessary)
to inject the filtered wastes into tha underground formations under
operating pressures of 35 to 63 kg/cm  (500 to 900 psig).  This disposal
plant was placed on line in October 1954 and thought to be operating
through 1975 (19, 68).

     With the development of the deep well disposal system, the natural
36.5 ha (90 acre) swamp or pond which previously received the strong
synthesized organic chemical process waste was converted  into a large
recnarge basin.  The storm sewer carrying 3,785 m /day (1.0 mgd) of un-
contaminated cooling water was directed into this pond.  About the same
time, the Company also developed a 0.47 ha (1.15 acre) artificial
recharge pond.  Storm and spent cooling waters from the plant were
directed into the 0.47 ha (1.15 acre) basin.  In 1955, the artificial
pond and the natural pond were recharging 9 and 16 percent, respectively,
of the installation's total wellwater drawn from groundwater resources
(72).

-------
               IX.  DEVELOPMENT OF EFFLUENT LIMITATIONS


     As discussed in previous Sections, data on waste characteristics,
production processes, and waste treatment and control practices for 39
plants in the pharmaceutical industry were compiled from literature
searches, industry contacts, and plant visits.  The industry was divided
into five categories based on differences in production processes, final
products, and associated waste characteristics.  Waste treatment data
was reviewed and evaluated and plants that were achieving exemplary
levels of wasteload reductions were identified.

     From these plants, model systems were selected for each industry
category to define the levels of wasteload reduction that are currently
attainable by properly designed and operated waste treatment and control
systems.  Effluent limitations were then developed for waste parameters
of significance for each category, based on performance levels of the
model systems.  This Section of the report describes the model systems
for each category and the effluent limitations developed.
EXEMPLARY PLANTS

     A number of bulk pharmaceutical manufacturing plants are currently
and consistently attaining exceptionally high levels of organic wasteload
reduction, as especially measured by removals of BOD and COD.  Most of
these plants are in the Fermentation/Organic Synthesized Chemicals category.
Fermentation Plants

     There are no known plants that employ only fermentation processes
for the production of Pharmaceuticals.


Organic Synthesized Chemical Plants

     High-performance plants identified in this category are the Hoffman-
LaRoche facility at Belvidere, N.J. and the E. R. Squibb plant at Humacao,
Puerto Rico.


Fermentation/Organic Synthesized Chemical Plants

     The majority of pharmaceutical manufacturing plants are in this
"combined" manufacturing category.  The plants have both extensive
fermentation facilities and significant organic chemical synthesis
capability.  Fermentation operations usually account for about 30 to 60
percent of the plant production.

     Outstanding plants for this category are:  the Pfizer facility at
Terre Haute, Ind.; Abbott Laboratories, Inc. plants at North Chicago, 111.

-------
142
and Barceloneta, Puerto Rico; and Eli  Lilly's Clinton Laboratories  at
Clinton, Ind.  In addition, Lederle Laboratories,  Pearl  River,  N. Y.,
and Wyeth Laboratories, West Chester,  Pa.  achieve  high levels of waste-
load reductions in their pretreatment  facilities prior to discharge to
municipal treatment systems.
Biologicals Production Plants

     Only a few plants of this type exist and data on waste treatment
practices are somewhat limited.  No exemplary plants were identified.
Drug Mixing, Formulation and Preparation Plants

     No exemplary plants were identified for this category.   Waste
treatment performance at McNeil Laboratories, Fort Washington, Pa., and
Wyeth Laboratories, Paoli, Pa. approached exemplary levels.
AVAILABLE TREATMENT AND DISPOSAL PROCESSES

     Of the wide variety of techniques and approaches to waste control,
treatment and disposal available in the industry, some of the better
adapted and designed systems have performed extremely well.   Other
systems for a number of reasons are yielding relatively poor results.
The industry has employed all of the following aggregated subsystems at
one time or another:

          Separate filtration of mycelium, and drying and recovery of
          fermentation broths and mycelium for use in animal feed
          supplements.

          Solvent recovery at centralized facilities or at individual
          sectors.  Reuse and/or incineration of collected solvents.

          Collection of biological, synthetic and pathogenic wastes when
          present, within plant for incineration or disposal by separate
          means.  Included are steam cooking and sterilization of pathogenic
          wastes, separate dry cleaning of animal cages, and overall
          animal wastes.

          Special recovery and subsequent sale of sodium sulfate salts.

          Scavenging and recovery of high-level ammonia waste streams,
          sold in bulk as a fertilizer base and for other needs.

          Elimination of barometric condensers.

          Extensive holding and equalization of wastewaters  prior to
          treatment systems.

-------
                                                                143
Incineration of mycelium and excess  biological  sludges.   In-
cineration system may also receive pathogenic wastes,  unrecover-
able solvents, fermentation broths or syrups, semi sol id  and solid
wastes, etc.  System can furthermore be integrated with  the
burning of odorous air streams.   Concurrently,  waste systems
may be covered or enclosed for odor control  and for maintaining
optimum temperatures for biological  treatment.

The activated sludge process including multiple-stage, ex-
tended aeration, the Unox aeration system, and  other variations.

The trickling filtration process, including high-rate and
bio-oxidation roughing towers, and multiple-stage systems.

Acid cracking at low pH's.

Extensive neutralization and pH adjustment.

Spray irrigation of fermentation beers and other pharmaceutical
wastes.  Other forms of land disposal are employed.  Deep
burial is unacceptable without the highest possible precautions.

Treatment of selected waste streams by activated carbon, ion
exchange, electro-membranes, sand filtration, chemical coagula-
tion, etc.

For proper handling and disposal of excess biological sludges,
a number of possibilities are available, including sludge
flotation, thickening and vacuum filtration, sludge centrifugation,
degasification, aerobic and/or anaerobic digestion, lagooning,
drying, evaporation, converting into a useable  product, incinera-
tion, land spreading, crop irrigation, composting or landfilling.

Anaerobic or submerged filters.

Cooling towers are necessary at pharmaceutical  plants for
reuse of cooling and jacketting waters.

Chlorination, pasteurization and/or other equivalent means of
disinfection.  Extensive disinfection is generally utilized
inside vaccine-antitoxin production facilities.

Municipal waste treatment.

Multiple-effect evaporation-steam and/or oil, multiple hearth and
rotary kiln incineration, and special thermal oxidation systems.

Extensive air cleaning by electrostatic precipitators, venturi
and water scrubbers and other equivalent systems.

-------
144


MODEL SYSTEMS

     The more advanced waste treatment and control  systems at pharmaceutical
plants almost invariably include activated sludge or trickling filtration
processes, usually as multistage systems, although other suitable treatment
processes are available.  From all indications, these systems are equally
applicable to all types of pharmaceutical plants.  With the exception of
Eli Lilly's Clinton Laboratories, Ind., Abbott Laboratories at Barceloneta,
Puerto Rico, and Commercial Solvents Corporation, Terre Haute, Ind., the
known plants all employ some variation of the activated sludge process.

     Burgess (47), in 1967 British literature, cites at least two cases
where the activated sludge process gives very high organic reductions on
pharmaceutical type effluents.  Wastes containing phenolics, aldehydes,
organic acids or similar compounds generally can be treated economically
by biological means with up to 99 percent BOD reductions.  A chemical
plant, with 70 percent citric acid wastes and 30 percent other pharma-
ceutical product effluents, employed the activated sludge process in re-
ducing the raw waste BOD from 1,800 mg/1 to 50 mg/1 BOD; that is equiva-
lent to a 97.2 percent BOD removal.  It was also indicated for a chemical
plant manufacturing a wide array of pharmaceutical products, dyestuffs
and intermediates that two-stage activated sludge could reduce the raw
waste BOD from 1,000 mg/1 to about 50 mg/1 in the final effluents,
obtaining a BOD reduction of about 95 percent.

     The Upjohn Company, Kalamazoo, Mich., demonstrated successful
treatment of spent fermentation beers (category 1) through the 1950's
by two-stage trickling filtration.  Pharmaceutical raw wastes amounting
to 2,270 m3 (600,000 gal)/day and containing 1,397 kg BOD (3,080 lb)/day
were reduced by trickling filters to about 41 kg BOD (90 lb)/day in the
final effluent, equivalent to BOD reductions of 95 to 98 percent.
Influent and effluent BOD concentrations were respectively 600 mg/1 and
20 mg/1.

     Wastes from Merck and Co., West Point, Pa. (categories 5- and 1) in
the 1950's were subjected to high-rate trickling filters giving high
treatment performance.  Overall BOD reductions were reported as greater
than 98 percent and TSS removals about 91 percent.

     Throughout 1972, the Stonewall, Va. plant of Merck and Co. (cate-
gory 3), utilizing three-stage biological treatment and with relatively
high raw wasteloads, showed average annual BOD and COD removals of 95
and 80 percent, respectively.

     Spray irrigation is relied on as the principal means of waste
treatment and disposal at the Commercial Solvents Corporation (CSC)
plant, Terre Haute, Ind. (category 3).  Including a series of waste
connections to the city, CSC accomplishes 91 percent BOD reduction.
This is noteworthy because 9 percent of the raw wastes receive no treat-
ment whatsoever.

     Another interesting case is the E. R. Squibb plant in Puerto Rico
(category 2).  According to Company reports, more than 99 percent of the

-------
                                                                     145
total organic raw wasteloads from this plant were to be incinerated,
amounting to 30,840 kg COD (68,000 lb)/day.   A dilute process stream
would be subject to biological treatment, producing a final  effluent  of
no more than 23 kg COD (50 lb)/day.

     The model systems presented below detail the unit processes used at
each high-performance plant and the associated wasteload reductions.


Organic Synthesized Chemical Plants

     Treatment steps at the Hoffman-LaRoche facility at Belvidere, N. J.
include the following:

     '  Screening
     "  Pre-clarifier
     '  Equalization basin equipped with turbine aerators,
        providing about a 1-day waste detention time
     '  pH adjustment and/or neutralization
     '  Flocculator-clarifier
     '  Activated sludge aeration
     '  Secondary settler
     '  Two shallow oxidation ponds in series
     '  Sludge thickening tank
     '  Aerobic sludge digestion tank and sludge drying beds
     '  Final chlorination
     '  Sodium sulfate recovery system,  a very unique  feature that
        includes a fluidized bed process and an anhydrous sodium
        sulfate plant

     Limited  treatment results have been received for  the Hoffman-
LaRoche system.  Based on treatment plant design and Company data of
Sept. 1973, minimum BOD reductions are calculated to be in range of 96.4
to 98.1 percent (average of 97.5 percent for BOD), and TSS removals in
the  range of  90 percent.  Design criteria called for BOD and TSS re-
ductions of 97.4 and 98.0 percent, respectively.  COD  removals of at
least 90 percent are apparently being attained.  Other data show that
effluent phosphorous and ammonia loads are exceptionally low.
Fermentation/Organic Synthesized Chemical Plants

     The waste treatment system at the Pfizer, Inc. plant in Terre
Haute,  Ind., is essentially a five-stage biological system with a re-
tention time for process wastes of 45 to 65 days.  Both the activated
sludge  and trickling filter processes are used.  An extensive case
history on this plant is presented in the Appendix.  System components
are as  follows:

     '  Primary settling
        Two extended aeration (activated sludge) basins, generally
        operated in series, providing up to 12 days' waste retention
        Secondary settling

-------
146
     '   Two clarigesters in parallel
     '   Two standard-rate trickling filters in parallel
     '   High-rate bio-oxidation tower
     '   Final clarifier
     '   Two 2 ha (3.75 acre) aerated stabilization ponds in series
     '   17 ha (35 acre) stabilization pond
     '   Chlorination
     '   Aerobic sludge digester
     '   20 ha (40 acre) sludge stabilization pond
     '   Land and crop application of stabilization pond  sludges
     '   Holding pond giving 1-day detention of spent cooling waters

     In 1972, Pfizer reported average BOD and TSS waste  removals through
the treatment works of 98.0 and 97.5 per cent, respectively.  Analysis
of data for 17 consecutive months through May 1974 showed average waste
reductions over this period of 99.1 percent for BOD and  97.8 percent
for- TSS.  The treated process stream averaged 10 to 15 mg/1 BOD and 20
to 30 mg/1 TSS.  Including the untreated cooling water flows, minimum
waste removals were about 98 percent for BOD and 95 percent for TSS.
The Pfizer system was capable of giving 50 percent phosphorous removals.
Unoxidized nitrogen levels in the raw process wastes were high, but the
biological treatment works provided average removals in  Kjeldahl, ammonia,
and organic nitrogen wasteloads of 75, 67 and 81 percent, respectively.
Summer nitrogen removals were'significantly higher than  winter removals.

     Waste treatment system components at Abbott Laboratories, North
Chicagox, 111., include the following:

     '   Waste screening and neutralization
     '   Two equalization basins providing 1.0 to 1.5 days' waste
        detention, equipped with auxiliary aeration
     '   Six 380 kl (100,000 gal) activated sludge aeration compartments
        equipped with sparged air turbines
     '   Degasification chambers for the mixed liquors
     '   Two final settlers in parallel
     '   Pasteurization of final process effluents
     '   Chlorination of combined process plus spent cooling water flows
     '   Evaporation and drying of spent fermentation broths
        Complete enclosure of waste treatment works except for
        secondary settlers
     '   Centrifuging of excess biological sludges
        Ducting of odorous air streams from waste treatment works
        and from fermentation sectors into the main plant boilers
        Incineration of excess biological sludges together with odorous
        air streams in the main plant boilers
        Recovery of selected waste stream(s) high in ammonia for bulk
        fertilizer sales
        Spill control program, including diversionary dikes and curbing
        around tank farms and other critical plant areas
        Provisions to connect highly treated process wastes to the
        Gurnee, 111. municipal  advanced waste treatment  plant in the
        near future

-------
                                                                     147
Data for 1972 at Abbott Laboratories showed average BOD and TOC'removals
from the process waste streams of 94.6 and 86.0 percent, respectively.
In 1973, average BOD and TOC reductions were 96.7 and 83.0 percent,  re-
spectively.  If one considers the spent fermentation beers going to the
evaporators (with a BOD load of 9,000 kg or 20,000 Ib/day) as being an
equivalent part of process raw wasteloads, average BOD reductions for
1972 and 1973 would then equate to 97.8 and 98.7 percent, respectively.
Removals of TSS were in the 71 to 74 percent range, but if the spent
fermentation beers had been accounted for as part of the TSS raw waste-
loads, equivalent TSS reductions would have exceeded 80 percent.  Total
phosphorous and ammonia nitrogen loads in the final Abbott effluents
were low during 1972-73, averaging 36 kg (80 lb)/day and 156 kg (345
lb)/day, respectively.

     The waste treatment and control system at Eli Lilly's Clinton
Laboratories, Ind., is unique in that it almost entirely relies upon
chemical destruction processes rather than biological processes.  The
system consists of the following:

     '  Concentration of waste streams within the manufacturing sectors
        down to minimal volumes
     '  Over-sized strippers for solvent recovery
     '  Stripper system to precondition the wastes entering multiple-effect
        evaporation
     '  Carver-Greenfield multistage, oil-dehydration, steam evaporator
        system receiving strong fermentation wastes
     '  Two John Zink thermal oxidation incinerator systems receiving
        highly concentrated chemical wastes
     '  Bartlett-Snow rotary kiln incinerator receiving plant trash and
        fermentation mycelium
        Small biological treatment plant for handling sanitary sewage
        Extensive cooling water towers and cooling water recirculation
        systems
     '  Scrubbing of air effluents from all incinerators and from the
        waste heat boiler on the Carver-Greenfield system

Clinton Laboratories is currently achieving 90 percent or better removal
of both BOD and COD from its raw wasteloads.  Minor treatment modifications
in the future, however, are expected to increase overall reductions up
to 95 percent for BOD and 93 percent for COD.  The Clinton plant continues
to demonstrate COD reductions equivalent to BOD reductions.  The COD to
BOD ratio for the final effluents is surprisingly about 2.2, almost the
same as for the incoming raw wastes.  An extensive case history of this
plant is presented in the Appendix.

     Other treatment and control systems in the pharmaceutical industry
have also produced high treatment performance results for this category
of wastes.  Two of these systems provide extensive pretreatment of
wastes prior to discharge to municipal treatment systems.  Wyeth Labora-
tories at West Chester, Pa.  consistently showed 90 percent BOD removal,
and up to 90 percent COD removal in  the 1960's with an activated sludge
plant.   Lederle Laboratories at Pearl River, N.  Y. with the Unox aeration
system and the activated sludge process is currently demonstrating slightly

-------
148


over 90 percent BOD reduction.  The additional secondary treatment of
these highly treated effluents provided by the municipal systems is
thought to give overall BOD removals of 97 to 98 percent for Wyeth
pharmaceutical wastes, and about 97 percent BOD reductions on Lederle's
wastes.
Biologicals Production Plants

     No exemplary plants were identified for this category; therefore,
no model system could be specifically defined.  A case history giving
details of the treatment system for the Wyeth Laboratories at Marietta,
Pa., a serums, vaccine and antitoxins production plant, is presented in
the Appendix.  Unfortunately, the activated sludge treatment plant at
Wyeth, Marietta was experiencing possible overload and toxicity diffi-
culties.  Reductions in BOD averaged 92.3 percent, but BOD results were.
likely affected by toxicity; COD reductions were only about 50 percent.
This was not an exemplary treatment situation.  The data base for this
category is relatively limited because very few of these plants exist.
Good waste treatment performance is cited elsewhere in the report for
Eli Lilly Greenfield Laboratories, Ind. (category 4).  Waste treatment
removal efficiencies have also been high at Lederle Laboratories, which
in part is also a category 4 plant.  Technical judgement based on the
literature and field experience indicates that waste treatment results
comparable to those achieved by category 5 plants (described below)
should be achievable by this category.
Drug Mixing, Formulation and Preparation Plants

     Two plants were identified in this category that have waste treat-
ment systems that approach but do not attain exemplary levels of waste-
load reduction.  Both showed reasonably good waste removals with rather
small, compact activated sludge plants.

     The treatment system at the Wyeth Laboratories, Paoli, Pa. includes
the following components:

     '  Screening
        Two equalization basins equipped with auxiliary air spargers,
        providing up to 2 days' waste detention
        Three activated sludge chambers operating in parallel
     '  Secondary settling
        Aerobic digestion of excess biological sludge with residues
        taken to sanitary landfill
     '  Chlorination of final effluent

     During the period Jan. 1973 to Jan. 1974, the treatment system BOD
and COD removals averaged 94.5 and 85.0 percent, respectively.  A case
history of this plant is in the Appendix.

     At McNeil Laboratories, Fort Washington, Pa., process and sanitary
waste streams are segregated and receive separate handling before passing

-------
                                                                     149
through an activated sludge system.   Treatment system components  include
the following:

     Process Wastes

     '   Primary settling tank
     '   Two equalization basins equipped with auxiliary air and
        providing waste retention up to 2.5 days

     Sanitary Sewage

     '   Comminutor

     Combined Wastes

     '   Activated sludge chambers, 24 hr aeration
     '   Two secondary clarifiers in parallel
     '   Chlorination of final effluent
     '   Provision for sludge recycling around the aeration chambers
     '   Thickening and^aerobic digestion of excess sludges

In spite of hydraulic and organic overload conditions, the McNeil system
over the period Jan. 1973 through Jan. 1974 achieved average waste
removals of 92.4 percent BOD and 87.7 percent COD.  Without overloading,
the system could likely have produced about 95 percent BOD removal and
90 percent COD removal.
EFFLUENT LIMITATIONS

     The end products of the compilation, review, and evaluation of the
data presented in this report are effluent limitations.  They are used
as the basis for specifying numerical limits on the allowable wasteloads
that may be discharged from individual pharmaceutical plants.  In most
industries, effluent limitations have been developed that specify an
allowable wasteload per unit of production.  It has been inferred through-
out this report that waste limits per unit of production could not be
established due to:  the wide diversity and large number of products
made; the trading of intermediates between plants; lack of agreement in
the industry concerning what to measure and how to measure products; and
extreme reluctance of the pharmaceutical companies to disclose appropriate
production data.  The effluent limitations developed herein are thus
specified in terms of percentage reductions in raw wasteloads or final
effluent concentrations appropriate for each parameter of significance.


Limits for BOD. COD and TSS

     Wasteload reductions achieved by high performance plants varied
only slightly between systems handling wastes from organic synthesized
chemical  plants (category 2) and systems treating wastes from plants

-------
150


with a large component of fermentation processes (category 3).   Demon-
strated BOD removals ranged from 96 percent for fermentation operations
to 94 percent for organic synthesized chemicals.  Comparable variation
in COD reductions were 83 vs 81 percent.  No difference in TSS  removals
was observed. ' Effluent limitations for fermentation (category  1),
organic synthesized chemicals (category 2), and fermentation/organic
synthesized chemical (category 3) plants were set at 95 percent BOD
removal, 82 percent COD removal and 82.5 percent TSS removal.  For
category 3 plants, a refinement in effluent limitations can be  made by
determining the percentage production of Pharmaceuticals by fermentation
vs organic synthesis and adjusting BOD and COD limits between the two
sets of values noted above.

     For biologicals production plants (category 4) and drug mixing,
formulation and preparation plants (category 5), effluent BOD and COD
limitations were established as 92.5 and 80 percent removals, respec-
tively.  Proper data were not available on which to base broad  TSS
limits.

     The detailed information given in this report indicates that in
applying suitable treatment processes the BOD, COD'and TSS limitations
can be reached with a wide degree of flexibility, certainty, and a
considerable margin for safety.  Maximum daily limitations are  recom-
mended as 2 times the average daily limits for BOD and COD, and 2.5
times for TSS.  These limits are intended in all cases to represent
baseline or minimum requirements.


Ammonia Nitrogen Limits

     Pharmaceutical plants that have fermentation or organic synthesis
operations generate large loads of unoxidized nitrogen (ammonia and
organic nitrogen).  Ammonia nitrogen limits are generally necessary for
plants in categories 1, 2 and 3 because effluents from these plants even
after biological treatment may still contain thousands of pounds per day
of unoxidized nitrogen.  Ammonia and organic nitrogen loads for certain
plants have been shown to exceed BOD loads in the final effluents.

     Oxidation of ammonia to nitrate nitrogen can theoretically consume
about 4.6 parts of oxygen for each part of ammonia present in wastewaters
in comparison to about a 1.4:1 ratio for BOD.  Organic nitrogen correspond-
ingly has a high dissolved oxygen demand for oxidation into nitrates.
Ammonia is decidedly toxic to fish and aquatic life under varying stream
conditions.  Also, even moderate levels of unoxidized nitrogen  can give
distorted results in running the BOD analysis on plant effluents.  And
finally, chlorination of effluents containing ammonia and other organic
nitrogenous compounds can form chloramines and associated compounds that
can, in turn, be highly toxic to fishlife, impart off-tastes to water
supplies, and create other undesirable side effects.

-------
                                                                      151
     Based on intensive discussion with industry members over the past
few months and practical considerations, effluent limits for ammonia
nitrogen were established as 25 percent of existing plant discharges
(i.e. 75 percent reduction), or 100 mg/1 times the internal plant process
flows, whichever gives the lesser amount.  The level of 100 mg/1 can be
reached by in-plant controls, scavenging and process substitution in
most factories.  There would therefore be no need for extensive treat-
ment.  Ammonia levels of 20 mg/1 times prevailing process flows are
technically achievable and at least one industry member has indicated
tentative concurrence with the 20 mg/1 limit.  The lower limit may
ultimately be specified for the industry in the future.
Limits on Metals and Trace Ions

     Because of their presence in most pharmaceutical effluents, metals
and trace ions have been indicated as being troublesome.  Limits for
metals have indeed been set on a number of actual NPDES pharmaceutical
permits.  Individual plant conditions will often dictate particular
allowable limits; consequently only suggested effluent limits are given
here.  Previous pharmaceutical permits have employed the following
ranges:  iron - 1.0 to 1.5 mg/1; manganese - 0.5 to 1.0 mg/1; phenolics -
0.25 to 0.5 mg/1; total chromium - 0.25 to 0.5 mg/1; aluminum - 1.0 to
2.0 mg/1; sulfides -  about 0.5 mg/1; zinc - 1.0 to 1.5 mg/1; lead - 0.1
to 0.25 mg/1; copper -  0.5 to 1.0 mg/1; and mercury (total plant) - 0.05
kg (0.1 lb)/day.

     Other effluent limits may be equally appropriate depending on
specific conditions and the type of facility involved.  Additional
constituents deserving attention may include:  cyanide, copper, tin,
cadmium, nickel, arsenic, chlorinated hydrocarbons, pesticides, and
Ra-226, gross alpha and beta (the latter only if radionuclides are being
handled by the plant).


Other Limitations

     Because of the potential toxicity of pharmaceutical plant effluents,
a monitoring requirement for toxicity in effluents has been developed.
Every six months, the permittee shall supply TL -96 hr bioassay test
results for the effluent from each outfall.   Fe
-------
                              BIBLIOGRAPHY

 1.   Pharmaceutical Waste  Treatment  with an Anaerobic Filter, N.  D.  Dennis
     and J.  C.  Jennett,  Paper Presented  at 29th Annual  Purdue Industrial
     Waste Conference, May 7-9,  1974.

 2.   Study of Pharmaceutical Manufacturing Wastewater Characteristics and
     Aerated Treatment System,  D.  R. Andersen, et.al.,  Proceedings of 25th
     Purdue Industrial Waste Conference, Lafayette,  Ind.,  Part  I, p  26-35,
     1970.

 3.   Pharmaceutical State-of-the Art Report  (Draft), Gulf  South Res.  Inst.
     for Pharmaceutical  Manufacturers' Assn.  and  USEPA,  54 p plus Appen-
     dicies, Draft Report  effective  Feb. 1974.

 4.   Pilot Plant Studies on Biological Treatment  of Pharmaceutical Wastes,
     R.  W. Herion, H. 0. Loughhead and A.  L.  Veliutao,  Industrial Water &
     Wastes, 8, 28-32, Mar.-Apr. 1963.

 5.   Long-Term  Biodegradability Tests of Organic  Industrial Wastes,  J.  C.
     Young and  S. B. Affleck, Paper  Presented at  29th Annual Purdue  Indus-
     trial Waste Conference, Lafayette,  Ind., 26  p,  May 7-9, 1974.

 6.   Handling Wastes from  the Billion Dollar  Pharmaceuticals Industry,
     R.  H. L. Howe, Wastes Engineering,  3J_,  728-731, 752-753, Dec.  1960.

 7.   Characteristics of Antibiotic Wastes, S. S.  Mudri  and N. S. Phadke,
     Environmental Health, (India),  ]Q_,  40-42,  1968.

 8.   The Production of Antibiotics,  G. Ross,  British Chemical Engineering,
     3_,  8-13, Jan. 1958,

 9.   Pollution  Control  in  the Pharmaceutical  Industry,  B.  S. Lane,  Indus-
     trial Pollution Control Handbook, Ch.  17,  H. F. Lund, ed., McGraw-Hill,
     N.  Y.f 1971.

10.   Liquid Wastes from the Fermentation Industries, G. Lines,  Water Pollu-
     tion Control, (Brit.) p 655,  1968.

11.   Spray Irrigation of Organic Chemical  Wastes, R. A. Woodley, Proceedings
     of 23rd Purdue Industrial  Waste Conference,  Lafayette, Ind., Part  I,
     p 251-261, May 7-9, 1968.

12.   Pharmaceutical Waste  Treatment  by the Trickling Filter, Activated
     Sludge, and Compost Processes,  A. H.  Molof,  Biotechnology  and  Bio-
     engineering, W^, 197-208, 1962.

13.   Pharmaceutical Waste  Disposal by Soil Injection,  R. R. Melcher, Bio-
     technology and Bioengineering,  |V_,  147-151,  1962.

-------
154
14.  Detection and Reduction of Particulate Matter in Pharmaceuticals,
     B. Trasen, Chemical and Engineering Progress, 64_, No. 2, 64-68, 1968.

15.  Recovery of Useable Waste Products, D. G, Reihard, Sanitary Engineer-
     ing Conference, Tenth Proceedings, Univ. of Illinois, Dept. of Civil
     Engineering, Urbana, 111., 6 p, Feb. 6-7, 1968.

16.  Coal-Fired Boilers Burn Waste Sludges and Odors in an Integrated
     Pollution Control System, D. E. Quane and M. R. Stumpf, Paper Pre-
     sented at 46th Annual WPCF Conference, Cleveland, Ohio, 19, Sept.
     30-Oct. 5, 1973.

17.  Two-Stage Filter Operation at the Upjohn Haste Treatment Plant,
     L. B. Tompkins, Sewage & Industrial Wastes, 39_, 1161-1165, Oct. 1957.

18.  Chemical and Antibiotic Wastes Treatment at Willow Island, West
     Virginia:  1. Present Chemical Wastes Treatment, J. F. Vogler,
     11. Aureomycin Wastes, J. M. Brown, 111. Design of Additional Treat-
     ment Facilities, G. E. Griffin, Sewage and Industrial Wastes, 24_,  4,
     p 485-495, Apr. 1952.

19.  What to do with Wastes When Volumes Overtake Capacity, S. M. Paradise,
     Industry and Power, 69_, 35-39, Dec. 1955.

20.  High-Rate Activated Sludge Treatment of Fine Chemical Wastes, F. E.
     Dryden, P. A. Barrett, J. C. Kissinger and W. W. Eckenfelder, Jr.,
     Sewage'& Industrial Wastes, 2, 183-194, Feb. 1956.

21.  Complete Biological Treatment of Antibiotic Production Wastes, R.  H. L.
     Howe, Biotechnology and Bioengineering, r^, 161-170, 1962.

22.  Engineering Processes for Waste Control, W. G. Barker and D. Schwarz,
     Civil Engineering Progress, 6_5_, 1, p 58-61, Jan. 1969.

23.  Squibb Solves Its Pharmaceutical Wastevtater Problems in Puerto Rico,
     J. R. Lawson, M. L. Woldman and P. P. Eggerman, Chemical Engineering
     Progress Symposium Series No. 107, 1971, article from Water-1970,
     p 401-404, 1970.

24.  Unox Gets First Industrial Test, Environmental Science & Technology,
     6, 10, p 878-879, Oct. 1972.

25.  Parameter of Disposal of Waste from Pharmaceutical Industry, A. H.
     Molof and N. Zaleiko, Ann. New York Academy Sciences, 130. 851-857,
     1965.

26.  The Funda Filter in the Pharmaceutical Industry, D. M. Landis, Chem-
     ical Engineering Progress, 64_, 2, p 74-78, Feb. 1968.

27.  Biological Treatment of Pharmaceutical Wastes, W. W. Eckenfelder,  Jr.,
     and E. L. Barnhart, Biotechnology and Bioengineering, IV, 171-180,
     1962.

-------
                                                                     155
28.   Standard Industrial Classification Manual - 19?2,  Executive Office
     of the President, Office of Management and Budget, U.  S.  GPO, Wash-
     ington, D. C., 1972.

29.   Design of Pharmaceutical Wastes Plant was Based on Lab Studies,
     J. F.  H. Walker, and R.  W.  Herion, Wastes Engineering, 3_3_, 610-614,
     639, Nov. 1962.

30.   Treatment and Disposal of Synthetic Drug Wastes, D. M. Patil, T. K.
     Srinivasan, G. K. Seth and Y. S. Murty, Environmental  Health (India),
     4_, 96-105, 1962.

31.   Antitoxin and Vaccine Wastes Treated at Eli Lilly Plant,  R. H. L.
     Howe and D. G. Coates, Wastes Engineering, 26. 235, May 1955.

32.   Effects of Penicillin Wastes in Ley Creek Sewage Treatment Plant,
     Syracuse, N. Y., U. T. Mann, Sewage & Industrial Wastes,  23^, 1457-
     1460,  Nov. 1951.

33.   Make Pure Water from Strong Wastes, R. H. L. Howe and R.  W. DeMoss,
     Industry and Power, 72^,  14, 1957.

34.   High-Plate Filters Treat Mixed Wastes at Sharp and Dohme,  J. A.
     Liontas, Sewage and Industrial Wastes Journal, 26_, 3, p 310-316,
     Mar. 1954.

35.   Laboratory Testing of Pharmaceutical Wastes for Biological Control,
     R. Otto, W. Barker, D. Schwarz and B. Tjarksen, Biotechnology and
     Bioengineering, VJ_, 139-145, 1962.

36.   Complete Mixing Activated Sludge Treatment of Antibiotic Wastes,
     R. E.  McKinney, Biotechnology and Bioengineering, J^, 181-195, 1962.

37.   Introductory Remarks on Symposium on Fermentation Waste Disposal,
     L. Kempe, Biotechnology and Bioengineering, IV_, 127-128,  1962.

38.   Pharmaceutical Wastewater : Characteristics and Treatment, D. R.
     Andersen, et.al., Water & Sewage Works, IW/2, 4 and 6, 118, 3, Mar.
     1971.

39.   Use of Algae for Treatment of Synthetic Drugs Wastes, K.  R. Khan,
     Environmental Health (India), , 193-201, 1962.

40.   Cortisone from Diosgenin, The Industrial Chemist, 32_, 167-174, Apr.
     1956.

41.   British Terramycin Production, The Industrial Chemist, 31, 561-563,
     Nov. 1955.

42.   Application of Waste-To-Product Patios in the Fermentation Industries,
     R. K.  Blaine and J. M. Van Lanen, Biotechnology & Bioengineering,  IV,
     129-138, 1962.

-------
156
43.  Land Disposal of Pharmaceutical Manufacturing Wastes,  G.  C.  Colovos
     and N. Tinklenberg, Biotechnology & Bioengineering, JW, 153-160,  1962.

44.  Oxygenation Controls Odors3 BOD from Fermentation,  Chemecology,  p 7-8,
     MCA, Washington, D. C., Nov. 1972.

45.  Chloramphenicol Synthesis, The Industrial  Chemist,  28, 209-214,  May
     1952.

46.  The Kline Guide to the Chemical Industry,  2nd Edition, P. Noble,  ed.,
     Charles H. Kline & Co., Inc., Fairfield, N. J. 1974.

47.  Treatment of Chemical Plant Effluents,  J.  V. Burgess,  Effluent Water
     Treat. Journ., 7_, 533-535, Oct. 1967.

48.  Fermentation Waste Disposal in Great Britain, C.  J. Jackson, Proceed-
     ings of the 21st Purdue 'Industrial Waste Conference, Lafayette,  Ind.,
     p 19-32, May 3-5, 1966.

49.  Incinerators, Domes Overcome Odors, Chemecology,  MCA,  p 3, Washington,
     D. C., June 1973.

50.  Two Treatment Installations for Pharmaceutical Wastes, R. W. Herion
     and H. 0. Loughhead, Proceedings of 18th Purdue Industrial Waste
     Conference, p 218-232, Apr. 30-May 2, 1963.

51.  Anaerobic Lagooning - A New Approach to Treatment of Industrial
     Wastes, D. 0. Howe, A. P.  Miller and J. E. Etzel, Proceedings of 18th
     Purdue Industrial Waste Conference, Lafayette, Ind., p 233-242,  Apr.
     30-May 2, 1963.

52.  Drug Firm Cuts Sludge Handling Costs, E. Braez, Water and Wastes Engi-
     neering, 9_, 1, p A-22, A-23, Jan. 1972.

53.  Treatment of Animal Wastes at the Greenfield Laboratories of Eli Lilly
     & Company, T. W. Bloodgood, Proceedings of the 21st Purdue Industrial
     Waste Conference, Lafayette, Ind., p 56-61, May 3-5, 1966.

54.  Waste Treatment at a Synthetic Drug Factory in India,  G.  J.  Mohanroa,
     P. V. Subrahmanyam, S. B.  Deshmukh and S.  Saroja, Journal WPCF,  42_,
     8, Part I, p 1530-1543, Aug. 1970.

55.  Pharmaceutical Waste Disposal Studies,  II.  Turbine Mixer Aeration in
     an Activated Sludge Plant, W. G. Barker, et.al.,  Proceedings of 15th
     Purdue Industrial Waste Conference, Lafayette, Ind., p 58-67, May 3-5,
     1960.

56.  Pharmaceutical Waste Disposal Studies I.  Turbine Mixer Aeration in
     an Activated Sludge Pilot Plant, W. G.  Barker, et.al., Proceedings of
     13th Purdue Industrial Waste Conference, p 1-11,  May 5-7, 1958.

-------
                                                                     157
57.  Pilot Plant Studies of Pharmaceutical Wastes at the Upjohn Company,
     J. R. Cushman and J. R. Hayes, Proceedings of the llth Purdue Indus-
     trial Waste Conference, Lafayette, Ind., p 62-72, May 15-17, 1956.

58.  Use of Research Techniques for Determination of Design Parameters,
     G. J. Genetelli, H. Heukelekian and J. V.  Hunter, Paper of the
     Journal Series, New Jersey Agricultural  Experiment Station, Rutgers-
     the State University of New Jersey, Department of Environmental
     Sciences, New Brunswick, N. J. 1967.

59.  Parke-Daois Engineers Convert Holland Plant, D. K. Foraker and J.
     Adinoff, Industry and Power, p 67-69 & p 80-84, July 1953.

60.  Disposal of Antibiotic Spent Beers by Triple Effect Evaporation,
     K. H. Edmondson, Proceedings of 8th Purdue Industrial Waste Conference,
     Lafayette, Ind., p 46-58, May 4-6, 1953.

61.  Composting Waste Sludge from Pharmaceutical Manufacturing, A. J.
     Gabaccia, Sewage and Industrial Wastes Journ. 3_U 10, p 1175-1180,
     Oct. 1959.

62.  Biological Degradation of Wastes Containing Certain Toxic Chemical
     Compounds, R. H. L. Howe, Proceedings of 16th Purdue Industrial  Waste
     Conference, Lafayette, Ind., p 262-276,  May 1961.

63.  Miracle Drug Wastes and Plain Sewage Treated by Modified Activated
     Sludge and Biofiltration Units, R. H. L. Howe and S. M. Paradise,
     Wastes Engineering, 27., 210-213, May 1956.

64.  Terramycin : From Dirt to Drug, R. V. Reeves, Chemical Engineering,
     59_, 145-147, Jan. 1952.

65.  Pharmaceutical Waste Disposal, A. Gallagher, et.al., Sewage and In-
     dustrial Wastes Journ., 26_, II, p 1355-1362, Nov. 1954.

66.  Treatment of Pharmaceutical Wastes, W. R.  Home and U. S. Rinaca,
     Proceedings of 15th Purdue Industrial Waste Conference, Lafayette,
     Ind., p 235-239, May 3-5, 1960.

67.  Biological and Pharmaceutical Wastes, C. G. Mauri ello, Sewage and  In-
     dustrial Wastes Journ., 30, II, p 1397-1398, Nov. 1958.

68.  Disposal of Fine Chemical Wastes, S.  J.  Paradise, Proceedings of the
     10th Purdue Industrial Waste Conference, Lafayette, Ind., p 49-60,
     May 9-11, 1955.

69.  City-Industry Cooperation in Waste Disposal, T. B. Henry, Journal  WPCF,
     37^ 8, p 1171-1175, Aug. 1965.

-------
158
70.  Waste Treatment for Veterinary and Plant Science Research and Pro-
     duction at Eli Lilly Greenfield Laboratories, R. H. L. Howe and R. A.
     Nicoles, Proceedings of 14th Purdue Industrial Waste Conference,
     Lafayette, Ind., p. 647-655, May 5-7, 1959.

71.  The Merck Index, P. G. Stecher, ed., Merck Co., Inc., Rahway, N. J.,
     1968.

72.  Recharge Operations at Kalamazoo, W. H. Sisson, Journ. American Water
     Works Assn., 47^, 9, p 914-922, Sept. 1955.

73.  Separation of Sterols by Countercurrent Crystallization, A. Poulos,
     J. W. Greiner and G. A. Fevig, Industrial and Engineering Chemistry,
     53_, 12, p 949-962, Dec. 1961.

74.  Science, Industry and the State, G. T. Smith, ed., Symposium Publica-
     tions Division, Pergamon Press, Oxford, London and N. Y., 100 p, 1965.

75.  Pharmaceutical Laboratory BOD Studies, a Matter of Philosophy, T. K.
     Nedyed, D. E. Bergmann and A. A. Comens, Paper Presented at 2nd Sym-
     posium of Hazardous Chemicals and Disposal, Indianapolis, Ind., 1971.

,76.  BOD of Synthetic Organic Chemicals; C. B. Lamb and G. F. Jenkins,
     Proceedings of 7th Annual Purdue Industrial Waste Conference, Lafa-
     yette, Ind., p 326-339, May 7-9, 1952.

77.  The Biochemical Treatability Index (BPI) Concept, C. H. Thompson,
     D. W. Ryckman and J. C. Buzzell, Proceedings of the 24th Purdue Indus-
     trial Waste Conference, Lafayette, Ind., p 413-435, May 6-8, 1969.

78.  A Biodegradability Test for Organic Compounds, R. L. Bunch and C. W.
     Chambers, Journ. WPCF, 39_, 2, p 181-186, Feb. 1967.

79.  Chemical Structures Resistant to Aerobic Biochemical Stabilization,
     F. J. Ludzack and M. B. Ettinger, Journ. WPCF., 32, II, p 1173-1200,
     Nov. 1960.

80.  Antibiotic Production Plant, The Engineer (Great Britain), 201, 779-781,
     1956.

81.  Behavior of 3-Amino-l9 2, 4 - Triazole in Surface Water and Sewage
     Treatment, F. J. Ludzack and J. W. Mandia, Proceedings of the 16th
     Purdue Industrial Waste Conference, Lafayette, Ind., p 540-554, May
     2-4, 1961.

82.  Metabolism of Organic Sulfonates by Activated Sludge, J. M. Symons and
     L. A. Del Valle-Rivera, Proceedings of the 16th Purdue Industrial
     Waste Conference, Lafayette, Ind., p 555-571, May 2-4, 1961.

83.  A Novel Ion-Exchange Method for the Isolation of Streptomycin, C. R.
     Bartels, et.al., Chemical Engineering Progress, 54, 8, p 49-51, Aug.
     1958.                                           ~

-------
                                                                     159
84.  The Theory and Practice of Industrial Pharmacy,  L.  Lachman,  H.  A.
     Lieberman and J.  L. Kanig, eds., with 35 contributors, ,811  p,  Lea
     and Febiger, Philadelphia, Pa., 1970.

85.  Drugs in Current Use and Hew Drugs,  1969,  ed., W.  Model 1,  Springer
     Publishing Company, Inc., N. Y., 1969.

86.  Experimental Pharmaceutical Technology,  E. L.  Parrott and  W.  Saski,
     Burgess Publishing Company, Minneapolis, Minn.,  1970.

87.  Pharmaceutical Technology,^ Fundamental Pharmaceutics, E.  L.  Parrott,
     Burgess Publishing Co., Minneapolis, Minn., 1970.

88.  The Long White Line, (Story of Abbott Laboratories), H.  Kogan,  Random
     House, Inc., N. Y., 1963.

89.  Solid Waste Management in the Drug Industry,  G.  L.  Huffman,  Proceed-
     ings of the 26th Purdue Industrial Waste Conference, Lafayette, Ind.,
     p 444-449, May 4-6, 1971.

90.  The Removal of Mercury from Industrial Wastewaters by Metal Reduction,
     M.  D. Rickard and G. Brookman, Proceedings of the 26th Purdue Indus-
     trial Waste Conference, Lafayette, Ind., p 713-720, May 4-6,  1971.

91.  Anaerobic Lagoon Treatment of Low Sulfate Chemical Wastes,  R.  A.
     Woodley and T. F. Brown, Proceedings of the 26th Purdue Industrial
     Waste Conference, Lafayette, Ind., p 844-856,  May 4-6,  1971.

92.  vitamin BI% Feed Supplement, A. S. Hester and G. E. Ward,  Industrial
     and Engineering Chemistry, 46_, 2, p 238-243, Teb.  1954.

93.  Trickling Filter Studies on Fine Chemical Plant Wastes,  F.  E.  Reimers,
     U.  S. Rinaca and L. E. Poesse, Sewage & Industrial  Wastes  Journ.,  26,
     1,  p 51-58, Jan.  1954.

94.  Riboflavine, Kirk-Othmer, Encyclopedia of Chemical  Technology,  2nd
     Ed., Interscience Publishers, John Wiley & Sons, Inc.,  N.  Y.,  London,
     17, 445-457, 1969.

95.  Citric Acid, Kirk-Othmer, Encyclopedia of Chemical  Technology,  2nd
     Ed., Interscience Publishers, John Wiley & Sons, Inc.,  N.  Y.,  London,
     5_,  524-541, 1964.

96.  Chloramphenicol,  Kirk-Othmer, Encyclopedia of Chemical  Technology,
     2nd Ed., Interscience Publishers, John Wiley & Sons, Inc., N.  Y.,
     London, 4_, 928-937, 1964.

97.  Fermentation - An TEC Unit Process Review, A.  E. Humphrey and F. H.
     Deindoerfer, Industrial & Engineering Chemistry, 53, 11,  p 934-945,
     Nov. 1961.                                       ~

-------
 160
 98.  Sulfonamides (SuZfa Drugs),  Kirk-Othmer,  Encyclopedia  of Chemical
      Technology, 2nd Ed., Interscience Publishers,  John  Wiley & Sons,  Inc.,
      N. Y., London,  19, 261-279,  1969.

 99.  Steroids,  Kirk-Othmer, Encyclopedia of Chemical  Technology, 2nd  Ed.,
      Interscience Publishers, John Wiley & Sons,  Inc., N. Y., London,  18,
      830-896, 1969.                                                    ~~

100.  Spent Antibiotic Broth at Eli Lilly & Co.,  S.  M.  Paradise and  R.  H.  L.
      Howe, Industrial Wastes, 3_,  101-104, July-Aug. 1958.

101.  Treatment of Pharmaceutical  Wastes, J. M.  Brown,  Sewage & Industrial
      Wastes,  Journ., p 1017-1024, Aug. 1951.

102.  A Rational Approach to Design for a Complex Chemical Waste, E. J.
      Genetelli, H. Heukelekian and J.  V. Hunter,  Proceedings of the 5th
      Texas Water Pollution Association Industrial Water  & Waste Conference,
      p 372-396, 1965.

103.  Unconventional  High Performance Activated Sludge  Treatment of  Pharma-
      ceutical Wastewater, W. G. Barker, H. R.  Stumpf and D.  Schwarz,  Pro-
      ceedings of 28th Purdue Industrial Waste  Conference, Lafayette,  Ind.,
      May 1-3, 1973.

104.  Pharmaceuticals, Chemistry in the Economy,  Ch. 7, American Chemical
      Society  Publication, p 171-195, Washington.  D. C.,  Oct. 1973.

105.  Outline  of Details for Microbiological Assay of Antibiotics :  Second
      Revision,  B. Arret, D. P. Johnson and A.  Kirshbaum, Journ. of  Pharma-
      ceutical Sciences, 60, II, p 1689-1694,  1971.

106.  Collaborative Study Sf Aerobic Media for  Sterility  Testing by  Membrane
      Filtration, F.  W. Bowman, M. White and M.  P. Calhoun,  Journ. of  Pharma-
      ceutical Sciences, 60, 7, p  1087-1088, 1971.

107.  Analysis of Eadiopharmaceuticals, M. Jaffe and L. Ford, Journ. of the
      AOAC, 54,  4, p  879-883, 1971.

108.  Waste Control Highlights Plant Design, Environmental Science and Tech-
      nology,  4, II,  p 898-900, Nov. 1970.

109.  Environmental Coexistence, Newspaper Article Reprinted by Eli  Lilly &
      Co., 1973.

110.  Innovations in  Industrial Waste Control at Clinton  Labs, Paper Pre-
      sented by Robert Ellis, Eli  Lilly & Co.,  18 p, Indianapolis, Ind.,
      May 4, 1970.

111.  Waste Water Treatment at Vigo Plant, Pfizer, Inc.,  W.  L. Kindrick and
      M. E. Johnson,  Paper Presented to Effluent Standards and Water Quality
      Advisory Committee, Chemicals Workshop,  Purdue University, Lafayette,
      Ind., 19 p, May 1, 1973.

-------
                                                                     161
112.   The Story of Fermentation,  Pfizer,  Inc., 235 East 42nd St., N.  Y.

113.   Waste Water Treatment - Hgo Plant;  A Paper Presented to the Water
      Quality Committee, Wabash Valley Assn. by M. E.  Johnson, Pfizer,
      Inc., 9 p, Nov. 11, 1971.

114.   Microbiological Process Report,  I960 Fermentation Process Review,
      A. E. Humphrey and F. H. Deindoerfer, Journ. Applied Microbiology,
      ]Q_, 359-385, 1962.

115.   Pros taglandins : Chemical Foundation is Laid, R. L.  Rawls, Chemical
      and Engineering News, 52_, 25, p 18-20, June 24,  1974.

116.   Pharmaceuticals, J. N. T. Gilbert and L. K. Sharp, Butterworths,
      London, 206 p, 1971.

117.   Nitrogen Control, Agricultural Waste Management, Ch. II, R. C.  Loehr,
      Academic Press, N. Y., London, 1974.

118.   Removing Nitrogen from Waste Water,  C. E. Adams, Environmental  Sci-
      ence and Technology, _7_, 8,  p 696-701, Aug. 1973.

119.   Personal Communication, Files, Trip Reports, EPA, National Field
      Investigations Center-Denver, 1974.

120.   Control of Nitrogen Wastewater Effluents, D. J.  Ehreth and E. Barth,
      USEPA Technology Transfer Design Seminar, Office of Research and
      Development, Washington, D. C., revised Mar. 1974.

-------
       APPENDIX
    CASE HISTORIES
        OF THE
PHARMACEUTICAL INDUSTRY

-------
                PHARMACEUTICAL INDUSTRY
                    CASE HISTORY  A
        ELI LILLY AND CO.,  CLINTON LABORATORIES
                   CLINTON, INDIANA

(FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS  PRODUCTION)

-------
                                                                     A-l
    ELI LILLY AND CO., INC., CLINTON LABORATORIES,  CLINTON,  INDIANA
          (FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS PLANT)
BACKGROUND

     Eli Lilly operates a large bulk pharmaceutical  plant at Clinton.
The highly automated plant is one of the newest in the industry,  having
been completed about 1971.  The facility was described in 1970 (108,110)
as a $60 million complex with $6 to 8 million attributed to "environ-
mental control measures."  The plant operates continuously with current
employment slightly more than 500.  Employment is expected to reach
1,000 by the end of 1976.

     Permit applications submitted by the Company in 1971 and 1974
indicated that the plant is engaged in the manufacture of bulk uncom-
pounded antibiotics and medicinal organic chemicals  and their deriva-
tives.  The plant primarily manufactures cephalosporin antibiotics, a
new type of human broad-spectrum antibiotics directed to the treatment
of respiratory, urinary tract, and skin and soft tissue infections and
diseases which in many cases are reported not to be  effectively treated
by other antibotics.  The cephalosporins are believed to be a highly
modified and restructured form of penicillin.  Major products made at
Clinton are as follows:

          Monensin Sodium, the desired ingredient in the registered trade-
          mark product, "Coban."  Coban is prepared  in bagged form as  a
          medicated premix for use in broiler (chicken) feed.  Coban is
          reported to materially aid in the prevention of coccidiosis.

          Cephalexin Monohydrate or "Keflex," a semi-synthetic cephalos-
          porin antibiotic intended for oral administration.

          Cefazolin Sodium or "Kefzol," the cephalosporin antibiotic
          prepared in the parenteral or injectable form.

     Manufacture of these products would indicate that the Clinton Labora-
tories are heavily involved in both fermentation and the synthesis of
organic chemicals.

     There are close corporate and manufacturing ties between the
Clinton Laboratories and the Eli Lilly Tippecanoe Laboratories in
Lafayette, Ind., and headquarters and manufacturing  operations in Indian-
apolis, Ind.  Cephalosporin antibiotics and their derivatives are also
manufactured in Lafayette, and trading of intermediate products between
Clinton and Lafayette is common.  For example, a product manufactured  in
intermediate stages at Clinton may be shipped to Lafayette for further
processing, returned to Clinton, and perhaps then sent to Indianapolis
for final packaging.  In waste materials handling, this same condition
exists.  Highly concentrated waste materials may be  traded between
Lafayette and Clinton for final disposal by incineration at whichever
facility has maximum capacity.  Most final drug packaging and prepara-
tion is believed centered in Indianapolis.

-------
   A-2
Eli Lilly plans to double production at Clinton Laboratories and has
asked for increased allocations on waste load discharges to the Vlabash
River.  Full production for current Phase I is expected about the third
quarter of 1975.  Production levels in mid-1974 were about 60 to 70 per-
cent of full Phase I capacity.  With a doubling of full Phase I capacity,
Phase II represents significant expansion both in terms of processing and
treatment.  Phase II operations are expected to come on line by the 4th
Quarter 1975 or 1st Quarter 1976 and will extend through the first part
of 1977.  The NPDES permit has been written for Clinton Laboratories in
staged sequence to cover both Phase I and II production schedules through
June 1977.
WASTE TREATMENT AND CONTROL

     Plant water supply obtained entirely from Company wells is used
for sanitary needs, boiler feed makeup, process cooling water, makeup
for cooling water recycling systems in the purification and fermentation
process sectors, process needs, and sprinklers.

     Treatment Plant design data were based on the following estimated
characteristics of waste streams from the manufacturing areas.


                              Table A-l.
                     Design Waste Loads (108,110)
            Wastes                     Load                      BOD
                              (kkg/day)  (17000 lb/day) (kkg/day) (1,000 Ib/day)
Mycelia
General Plant Trash
Concentrated Chemical Wastes
Primary Wastes
Secondary Wastes
Dilute Chemical Wastes
Watery Process Waste
Sanitary Sewage
Untreated Clear Water Stream
TOTALS
>32
5

17
48
54
309
337
23,150
%23,950
>70
10

38
106
120
681
743
51 ,041
^52,800
>115


42
10
2
11
0.1
0.1
-x-180
>253


93
23
4
25
0.2
0.2
^398
     The maximum BOD discharged to the Wabash River was estimated in
1970 as 1,100 kg (2,500 lb)/day.  The potential raw waste load was
claimed by Eli Lilly as amounting to 181,000 kg (400,000 Ib) of BOD
daily equivalent to the BOD in raw sewage discharged by a city of 2.5
million persons.  The Clinton waste reduction system was declared to
have a maximum BOD reduction of 99.4 percent.

-------
                                                                     A-3


     Design waste handling objectives included:  1)   recovery of solvents
and process chemicals even if recovery proved uneconomical:  2)   no burial
of organic wastes either on or off the 280 hectare  (700 acre) Clinton in-
dustrial site; and 3)  no deep well waste disposal.   These objectives re-
sulted in the development of a waste handling, treatment and disposal
scheme at Clinton that is rather unique in the industry (108,110).  In
contrast to most pharmaceutical plants that primarily use biological  oxi-
dation procedures for treatment of "conventionally"  dilute wastes, Clinton
uses chemical destruction techniques.  Concentrated  semi-solid, semi-liquid
and liquid waste streams are mostly converted into  innocuous gaseous  em-
issions by stripping, distillation, evaporation and  incineration.   In most
plants, the concentrated semi-solid and semi-liquid  wastes are  transferred
off-site for ultimate disposal or separated from the treatment  plant.

Mycelia and Plant Trash

     Mycelia recovered from the fermentation broths  is burned with plant
trash.  About 136 kkg (150 tons) of wet mycelia equal to 32 kkg (35 tons)
of dry product are recovered daily.  The extremely  high load of 115 kkg
(253,000 Ib) of BOD/day associated with this waste  [Table A-l]  is  un-
doubtedly due to the presence of other organics in  addition to  mycelia.
Future plans call for evaporating and converting the mycelia into  a dried,
high protein animal feed supplement (108,110).

     Incineration of mycelia and trash is accomplished in a Bartlett-Snow
rotary kiln incinerator equipped with an afterburner and wet scrubber.
Ashes amounting to 270 kg (600 lb)/day are buried on site (108,110).   Air
scrubber effluent is discharged to plant sewers.  At full Phase I  opera-
tional levels, this effluent would have a flow rate  of 760 1/min (200 gpm)
and would contain estimated daily COD and BOD loads  of 54 kg (120  Ib)
and 18 kg(40 Ib), respectively.  These incinerated  wastes are not  normally
considered part of the raw waste loads received by  treatment works at
pharmaceutical installations.
Chemical Wastes

     Two types of concentrated chemical wastes are identified in Table A-l
"Primary" wastes are defined as those wastes capable of supporting com-
bustion while "Secondary" wastes require supplementary fuel  for burning.
These wastes originate primarily in the chemicals manufacturing and
purification operations.  Some are semi-solid, and even solid.   These
concentrated wastes have a flow rate of about 61 1/min (16 gpm)(108,110).

     Dilute chemical wastes with a flow rate of about 38 1/min  (10 gpm)
originate as bottoms off solvent stripping columns and other unrecover-
able wastes.  Lilly has reported that some solvent streams containing
about 3 to 5 percent solvents are uneconomical to recover and are mixed
with the chemical waste streams.

-------
A-4
     Both the concentrated and dilute chemical wastes are sent to two
John Zink thermal oxidizers equipped with adjustable venturi  scrubbers
for removal of particulates prior to stack discharge (108,110).   Scrubber
effluents totaling about 760 1/min (200 gpm) and containing a COD load
of"about 140 kg (300 lb)/day equivalent to about 90 kg (200 lb)/day of
BOD at full Phase I levels are discharged to plant sewers.

     The two John Zinc incinerators receive relatively small  amounts of
wastes but these are extremely concentrated.  At full Phase I operations,
the quantity of wastes entering the thermal oxidation system is only
76-95 1/min (20-25 gpm), but BOD and COD strengths of these materials
assay a few hundred thousand mg/1.  It is fairly evident that practi-
cally all of the material sent to the John Zink units cannot be con-
sidered as normal raw waste loads even under highly unusual conditions
in the pharmaceutical industry.  In addition to Clinton wastes, highly
concentrated wastes from other Eli Lilly plants are incinerated in the
John Zink units.
Watery Process Wastes

     Watery process wastes, defined as those containing no components
more volatile than water, originate primarily from the fermentation op-
erations.  The Company would not verify the possible presence of fermen-
tation broths in these waste streams.  These wastes, with a flow rate of
about 247 1/min (65 gpm), are decanted and then discharged to a Carver-
Greenfield evaporation system discussed below.

     A stripper operation, associated with the evaporator system, serves
the fermentation and product sectors primarily for the reclaiming of sol-
vents from fermentation broths.  The stripper system discharges recovered
solvents, watery wastes that go to the Carver-Greenfield unit, and a
distillate that goes to the plant sewer.  The volume of the distillate
has not been given but it is expected to contain COD and BOD loads of 320
and 140 kg (700 and 300 lb)/day, respectively, under full Phase I con-
ditions.  Because of the extremely large volumes of solvents entering the
stripper system, BOD and COD loads in the influent are enormous.  This
process stream (except for residuals) cannot normally be considered as a
raw waste at pharmaceutical plants.

     The Carver-Greenfield unit uses a multi-step oil dehydration pro-
cess and is equipped with a centrifuge, waste heat boiler and a venturi
scrubber.  Under full Phase I operation, the raw waste load to the unit
is predicted to be about 14,900 to 17,400 kg (32,800 to 38,300 Ib)
COD/day or 7,700 kg (17,000 Ib) BOD/day.  This unit relies upon an oil
carrier stream that is continuously blended into the evaporator and
waste heat boiler.  The oil carrier stream contains loads of about 5,170
to 6,030 kg (11,400 to 13,300 Ib) COD/day equal to about 2,270 to 2,720 kg
(5,000 to 6,000 Ib) BOD/day.  Dehydrated solids from the evaporator are
burned in the waste heat boiler to produce steam.  Three waste streams

-------
                                                                     A-5
leave the Carver-Greenfield unit.  The primary effluent is a distillate
with exceptionally high organic waste concentrations.  Under full Phase I
conditions this 150 to 230 1/min (40 to 60 gpm) waste flow is expected
to contain 1,500 to 1,720 kg (3,300 to 3,800 Ib) COD/day or about 770 kg
(1,700 Ib) BOD/day.  From January to July 1974, COD levels in this
effluent were in the range of 1,500 to 2,450 mg/1.  This waste stream
contains the largest waste load from the Clinton facilities continuing
to be discharged to the Wabash River.  The waste heat boiler produces
about 270 to 320 kg (600 to 700 Ib) of ashes daily.  Bottom ash is
hauled to landfill while flyash is collected in a venturi scrubber and
washed to the sewer.  This scrubber effluent will contain about 64 kg
(140 Ib) COD/day or 18 kg (40 Ib) BOD/day in a flow of 190 1/min (50 gpm)
at Phase I operational levels.

Solvents
     Barometric condensers are virtually absent at the Clinton Labora-
tories.  Published information states that approximately 80 percent of
the solvents from chemical manufacturing and purification were being
recovered by evaporation, distillation and other processes and the
Company was attempting to increase this recovery up to 90 percent.  Some
of the major solvents included benzene, acetone and ethanol.   Unfor-
tunately, in the critical area of solvent recovery, much higher effi-
ciencies should be attained by Clinton (108, 109, 110).
Sanitary Wastes

     Sanitary wastes are received into a small 379 m /day (100,000 gpd)
"package" sewage treatment plant that represents the only biological
treatment unit present in the Clinton operations.  The sewage flow
approximates 57 to 114 1/min (15 to 30 gpm) and design BOD input and
output loads were rated at 68 kg (150 lb)/day and 7 kg (15 lb)/day, res-
pectively, for this unit at full Phase 1 plant capacity.   Aerobically-
digested excess sludges from this sub-system are sent to the John link
thermal oxidizers.
Cooling and Storm Water

     Various cooling water streams and storm water from all  quadrants of
the plant ultimately are discharged without treatment through the main
plant outfall (Outfall 001).   The Company reports that the daily BOD
load contributed by these wastes ranges from 34 kg (75 Ib) during dry
weather to 68 kg (150 Ib) during wet weather with an average of 45 kg
(100 Ib).

     In summary, residual waste discharges from the Clinton  Labora-
tories' chemical wastes destruction facilities consist of the scrubbing
effluents  from the two John Zink incinerators, the Bartlett-Snow incin-
erator and the Carver-Greenfield waste heat boiler; the concentrated

-------
A-6


distillates leaving the Carver-Greenfield evaporator;  effluent  from the
stripper operation preceeding the Carver-Greenfield .evaporator; effluent
from the sanitary sewage treatment plant; water treatment  plant eff-
luents; blowdowns from the extensive cooling  water circuits  (i.e.  having
20,000 to 40,000 gpm recycle) and other blowdowns; some  floor drain and
storm sewer discharges; an array of spent cooling  waters;  and waters
from undefined or undesignated sources.  Total3effluent  flows were
anticipated in the range of 12,500 to 22,000  nr/day  (3.3 to  5.8 mgd)
containing around 1,130 kg (2,500 lb)/day of  BOD.

     The sum of the above waste loads predicted under  "full" Phase I
production capacity approximates 1,090 kg (2,400 Ib) BOD/day and  2,180 kg
(4.800 Ib) COD/day assuming treatment capabilities remain  more  or less
the same as currently existing.  It is noted  these waste loads  are very
close to the conditions approved by the State on September 21,  1971 for
the Clinton treatment works as discussed in a later  section.

     Another interesting point with regard to the  above  loads  is  that
air pollution control scrub streams contribute about 120 kg  (270  Ib)
BOD/day or about 10 to 12 percent of the total  effluent  loads  from the
Clinton plant.

     A cost summary for the Clinton, Indiana  waste treatment facilities
prepared by Eli Lilly in 1970 showed, at that time,  projected  installa-
tion expenditures would be around $4.9 million and total costs  including
depreciation would approximate $1.58 million  (108,109,110).

WASTE LOADS

NPDES Permit Application

     The initial permit application submitted in July  1971 gave the
following average waste loads based on "100 percent  operation":

                              Table A-2.
                  Summary of Permit Application Data
                               July 1971


           Parameter                           Load

Flow
BOD
COD
TDS
TSS
Total P
NH3-N
kg/day
23, 70^
1,120
4,380
7,370 ,
73^-
5
23
Ib/day
6. 26^
2,470
9,650
16,250 .
160^-
10
50
a/ Flow in m /day.
b/ Flow in mgd.
c/ These values were later declared to be in error.

-------
                                                                     A-7
     The Company submitted updated flow [Table  A-3]  and  waste  load
[Table A-4] data on February 23, 1973.
                               Table A-3.
                        February 1973 Flow Data
Waste Stream
Liquid effluents from the thermal oxidizers
Distillates etc. from oil dehydration CG evaporator
Venturi scrubber on CG evaporator
Scrubbing effluents from the trash incinerator
Treated sanitary effluents
Spent cooling waters and miscellaneous blowdowns
Estimated Total
Flow
(m3/day)
7, 090^
350b/
980
110
3.900
12,700

(gpm)
1 ,300^
65h/
50^
180
20
715
2,330/
a/ Reported later as 1,090 m /day (200 gpm).
b/ Reported in 1974.
/ 2,330 gpm = 3.36 mgd.
                              Table A-4.
                  Summary of February 1973 Load Data
                                                 Loads at Full  Phase I
Parameter           February 1973 Loads           Production Capacity
                (mg/1)  (kg/day)  (Ib/day)    (mg/1)  (kg/day)   (Ib/day)
BOD
COD
TSS
Zinc
Fecal
77
177
25
0.8
Col i forms -
980
2,280
310
9
-
2,170
5,030
680
20
-
81
191
40
3.0
1,500/100
1,040
2,590
540
40
ml (max.)
2,300
5,720
1,200
90
-

-------
A-8
Company Sampling Results

     Waste loads observed in the plant effluent (Outfall  001)  during
1972, 1973 and January through April of 1974 are summarized in Table A-5.
These data, taken from the Company monthly reports to the State of
Indiana, indicate that BOD and COD concentrations and loads have de-
creased steadily from-1972 to the present.  Total suspended solids
concentrations have remained relatively stable although loads  have
increased.  Chlorides increased significantly from 1972 to 1973 but the
chloride concentrations and loads have since continued at more or less
the same level.  Bromides showed the largest percentage increase of all
parameters.  Zinc levels have stayed about the same since 1973.  Con-
sidering the steady increase in Clinton production from 1972 to 1974, it
is somewhat surprising that there have not been greater increases in the
major parameter waste loads.

     The 1972 data were erratic, perhaps indicating early efforts to
shake down plant facilities.  During the January through April 1974 per-
iod, daily BOD, COD and TSS loads averaged about 770, 1,700, and 570 kg
(1,700, 3,700 and 1,250 Ib), respectively.  These loadings were about
70 percent of the design values predicted for full Phase I production.
The total suspended solids loads approximated about 70 percent of the
BOD loads.

     One very striking observation from all the Clinton results is the
very low COD to BOD ratio of around 2.0 to 2.5 consistently recorded for
Clinton wastewaters.  This is undoubtedly due to the unique waste treat-
ment facilities at this location.

     The Company for some time has been conducting sampling at selected
locations inside the Clinton Laboratories' complex.  COD values for the
period January 1974 through July 1974 are summarized in Table  A-6 for
seven internal sampling points.  Stations 5 and 6 flows are included in
Station 7 data.  Adding the COD loads from stations 1, 2, 3, 4 and 7
gives a total COD loading of 1,490 kg (3,284 lb)/day as an average over
the period January through July 1974.  It is not known if the  stripper
effluent associated with the Carver-Greenfield system is included in
this total.  The Outfall 001 loads for the period of January through
April 1974 in Table A-5 averaged 1,680 kg (3,703 Ib) COD/day.   Con-
sidering that different time periods are involved for the two  sets of
data, the COD loads seem to agree fairly well.  These loads, repre-
sentative of Clinton operations over the first part of 1974, were about
65 percent of those predicted under full Phase 1 production capacity.
Also from the data in Table A-6, it is evident that the Carver-Greenfield
distillate was lower in COD than predicted.  Conversely, significant and
unaccounted for waste sources were contributing large waste loads at
sampling locations 1 and 4, which were reported as representing pri-
marily non-contact cooling water streams.  Extraneous V/aste sources are
likely present even through no barometric condensers are utilized in the
Clinton plant.  Non-contamination is generally evidenced by COD concen-
trations in the range of 10 mg/1 to no more than 20 mg/1.  Sampling lo-
cations 1 and 4 over the period January through July 1974 demonstrated
COD levels of 33 mg/1  and 126 mg/1, respectively.  Such sources should be
identified and controlled and/or eliminated.

-------
                       TABLE A-5
 Summary of Monthly Effluent Waste Loads (Outfall  001)
Eli  Lilly and Co., Clinton Laboratories, Clinton,  Ind.
                Jan.  1972 to Apr.  1974
Parameter
FLOW
mgd
3
m /sec
BOO
mg/1
Ib/day
kg/day
COD
mg/1
Ib/day
kg/day
TSS
mg/1
Ib/day
kg/day
CHLORIDES
mg/1
Ib/day
kg/day
BROMIDES
mg/1
Ib/day
kg/day
ZINC
mg/1
Ib/day
kg/day

Average

3.38
0 148


70
2,390
1,084

196
6,770
3,071

27
912
414

72
2,270
1,030

0.7
25
11

0.5
14
6 4

1972


Range

2.90 -
0.127 -


48 -
1,846 -
837 -

113
4,320
1,960

20
501
227

57
2,250
1,021

0.4
15
6.8

0.2
9
4

4.90
0.214


185
7,556
3,427

- 421
- 17,200
- 7,802

- 53
- 2,126
- 964

- 143
- 3,680
- 1,669

- 1.3
- 50
- 23

- 1.2
- 29
- 13


Average

3.45
0.151


69
1,963
890

140
3,980
1,805

35
1,035
70

161
4,660
2,114

2.4
70
32

0.9
27
12
Year
1973


Range

2.46
0.108


43
1,326
60

104
3,073
1,394

21
520
236

119
3,150
1,429

0.5
12
5.4

0.4
13
5 9

- 4.15
- 0.181


- Ill
- 3,018
- 1,369

- 218
- 5,333
- 2,419

- 65
- 2,040
- 925

- 214
- 6,710
- 3,044

- 9.5
- 293
- 133

- 2.3
- 62
- 28


Average

4.34
0.190


48
1,714
77

102
3,703
1,680

34
1,250
567

150
5,410
2,454

10.0
370
168

0.9
31
14

1974

Range

4.23
0.185


40
1,403
636

80
2,835
1,286

24
806
336

104
3,760
1,706

5
185
83 9

0.4
16
7.3

- 4.47
- 0.195


- 53
- 1,915
- 86

- 121
- 4,007
- 1,818

- 55
- 1,990
- 903

- 180
- 6,480
- 2,939

- 16
- 590
- 268

- 1 0
- 36
- 16
                                                                                                          3=-

                                                                                                          vo

-------
                                                  Table A-6.

                            COD  Values  for Sampling Points  in Clinton Laboratories
                                              Jan - July,  1974

No.
1

2

3

4



5
N
6

7



Sampling Location
Description
Non-contact cooling waters from chemical
operations and north utility equipment area
plus surface runoff from NW quadrant
Boiler house blowdowns, south cooling tower
waters, plus runoff from central service
Non-contact cooling waters from fermentation
with surface runoff from SW quadrant
Non-contact cooling waters from Product Recovery,
the east utility equipment area and product
recovery strippers plus surface runoff from SE
quadrant
Effluent from Smith-Loveless sewage treatment
plant
Distillate from Carver-Greenfield evaporator
system
Effluents from the two John Zink incinerators,
the Bartlett-Snow incinerator, the Carver-
Greenfield evaporator, plus storm water from
the NE quadrant
Estimated Flow
(m /day) (gpm)

10,900 2,000

270 50

410 75



1 ,360 250

80 15

220 40



3,820 700

Range^
(mg/1)

20 - 62

56 - 145

9-39



83 - 165

19 - 55

1,492 - 2,452



183 - 286
COD
Average
(mg/1) (kg/day)

33 360

103 28

15 6



126 172

29 2

2,045 446



242 924


(lb/day)

793^

62

14


h /
380-

5

983



2,036
a/    Monthly concentration.
b/    Significant waste loads contained in  these  flows  not  previously accounted for by Company.

-------
                                                                    A-ll
     The Company conducted trace metal analyses on the Outfall  001  ef-
fluent on random days during 1973-1974 that yielded the following re-
sults:
                               Table A-7
                 Summary of 1973-1974 Trace Metal Data

Parameter           Median (mg/1)                Maximum (mg/1)
Copper
Iron
Lead
Mercury
Aluminum
Chromium
<1.0
<1 .0
<0.05
<0.004
<0.10
0.03
3.10 ,
< 1 . 0 
0.26
<0.004
<0.10
0.08
a/ Concentration of 1.4 mg/1 noted during August 1974 EPA-State
   survey that also showed zinc values of 1.4 to 1.6 mg/1 and some
   136 kg (300 lb)/day of NH3-N in the effluent from Outfall 001.
DEVELOPMENT OF NPDES PERMIT CONDITIONS

State Waste Load Limitations

     In February 1970, the Company requested the State of Indiana for per-
mission to discharge up to 1,430 kg (3,150 lb)/day BOD based upon a raw
load of 180,600 kg (398,200 lb)/day BOD.  Appropriate waste treatment was
described which would provide a 99.2 percent BOD reduction.  Major con-
tributing waste sources were as described in Table A-l.   Normally only the
watery process wastes, sanitary sewage and untreated clear water stream
are expected to contribute waste loads to the treatment facility at most
pharmaceutical plants.  Equivalent raw waste loads for the Clinton facility
are thus estimated to be in the range of 11,300 to 13,600 kg (25,000 to
30,000 Ib) BOD/day.

     On April 21, 1970, the State gave notice of considering the Company's
preliminary plans and specifications and attached certain conditions to
the plans.  Anticipated wastewater characteristics of the final  effluents
were 1,137 kg (2,507 lb)/day BOD, 1,157 kg (2,550 lb)/day chlorides,
243 kg (535 lb)/day bromides and 250 kg (550 lb)/day zinc contained in a
total  flow of 21,200 nT/day (5.6 mgd).  The State also specified that the
Carver-Greenfield effluents should receive additional treatment  because
of the very high organic concentrations in these particular waste streams.

-------
A-12
     On September 21, 1971, the State of Indiana gave approval  to plans
and specifications for the Clinton waste treatment facilities.   The  State
apparently accepted the conditions of 1,137 kg (2,507 lb)/day BOD an
34 kg (76 lb)/day TSS to be contained in a total wastewater flow of
15,200 m /day (3,97 mgd).  The State recommended that the Company design
and install adequate facilities for the treatment of the distillates from
the Carver-Greenfield operations.

     In mid-1972, the State requirements for the single effluent from the
Clinton facilities (Outfall 001) were given as:   BOD - 1,137 kg
(2,507 lb)/day or 53 mg/1 maximum; TSS - 40 mg/1 maximum; chlorides  -
1,157 kg (2,550 lb)/day or 54 mg/1 maximum; bromides - 243 kg (535 lb)/day
or 11.4 mg/1; and zinc - 1 mg/1.

Rationale for Effluent Limitations

     Various factors including a lack of data, the nature of the Clinton
facilities and/or perhaps the reluctance of the Company to provide suf-
ficient interpretive information made it difficult to determine an equiva-
lent raw waste load for the Clinton Laboratories.  Our best estimates of
raw waste load under full Phase I  production capacity are 11,300 to
13,600 kg (25,000 to 30,000 Ib) BOD/day and 22,700 to 27,200 kg (50,000
to 60,000 Ib) COD/day.  Treatment performance achieved by the Clinton fa-
cilities is currently estimated to be roughly 90 percent for both BOD and
COD.

     Based upon past data and from all indications given by the Company
to the State, Clinton Laboratories up through full Phase I production is
more than capable of holding average daily BOD, COD and TSS loads down
to 1,137 kg (2,507 Ib), 2,310 kg (5,100 Ib) and 820 kg (1,800 Ib) re-
spectively.  Accordingly, average daily immediate load limits in the
Clinton permit were established as 1,137 kg (2,507 Ib) BOD, 2,310 kg
(5,100 Ib) COD, and 860 kg (1,900 Ib) TSS.

     Evaluation of existing waste control, treatment and disposal prac-
tices identified several areas where improvements resulting in reduced
waste loads could be made.  The most significant reduction could be
achieved by providing treatment for the waste distillate from the
Carver-Greenfield evaporator unit.  This small 150 to 230 1/min (40  to
60 gpm) waste stream contributes a substantial waste load because of
excessive BOD and COD concentrations calculated at 2350 and 4700 mg/1,
respectively, for full Phase I operations.  These distillates are readily
amenable to additional inexpensive treatment and should be receiving
treatment today.  It is noted that this same recommendation was made by
the State of Indiana on September 21, 1971, when approval was given  to
the Clinton waste treatment facilities.  A minimum requirement of a
62.5 percent BOD reduction and a 50 percent COD reduction was selected
for this waste stream.  Assuming the addition of a biological treatment
unit on the distillate waste stream, it should be possible to reduce

-------
                                                                    A-13


full Phase I BOD loads from 1,137 to 658 kg  (3,507  to  1,450  lb)/day  and
COD loads from 2,310 to 1,540 kg (5,100  to 3,400  lb)/day.  A correspond-
ing reduction of TSS loads from 800 to 660 kg  (1,770 to  1,450 lb)/day
should be achievable by addition of the  distillate  treatment unit.   This
would be an overall  20 percent reduction.

     Although the percentage waste load  reductions  achieved  by treatment
of the Carver-Greenfield distillate do not appear particularly outstanding,
they are considered  sufficent to meet waste  treatment  requirements  for  the
bulk pharmaceutical  plants in the industry.  The  Clinton facility  has been
given certain extra  credits "for achieving exemplary COD  reductions.  Also,
the plant has other  alternative means available for reducting waste  load-
ings such as removing or controlling waste streams  connected to cooling
water discharges.

Recognizing the uncertainties in developing  raw waste  loads  for Clinton
and other exigencies, accordingly the permit limits for  full Phase I
operations on an average daily basis were set  at  (1,550  Ib)  BOD, (3,750 Ib)
COD and (1,525 Ib) TSS.  These figures result  in  average waste reductions
of around 94.5 percent for BOD and 93.0  percent for COD.

     Regarding nitrogen and phosphorous  effluent  loads,  based upon past
data, phosphorous would appear to present no problems.  Ammonia nitrogen
levels in the Clinton effluents are marginal with respect to BPT industry
limits.  Data provided by the Company in mid-1974 indicated  levels of
110 to 200 kg (240 to 450 lb)/day ammonia-N  vs. 150 kg (340  lb)/day es-
tablished in the NPDES permit for Phase  I Clinton conditions.  However,
limited data collected by Eli Lilly in September-October 1974 reports
that ammonia-N is now averaging about 290 kg (635 lb)/day.   Clinton at-
tributes the higher  "recorded" levels of ammonia  to better measurement
techniques and ammonia resulting from breakdown of  raw materials used  in
reactions, through incineration and other waste processing.   Identifica-
tion of contributing nitrogen souces and possible treatment  may be neces-
sary, partially dependent upon a more adequate data base.   If ammonia  re-
duction is necessary, first priority should  be given  to  selected waste
stream scavenging and direct in-plant recovery means.

     Clinton laboratories, over the past 24  months, has  apparently re-
duced zinc levels in the final effluent  to less than  1.0 mg/1.  Tentative
and/or long-term limits hagve been set on trace constituents including
phenolics, zinc, iron, lead, copper and  sulfates  because these elements
are indicated to represent possible or probable problems.  Load limits
for these elements have been based upon  an average  total effluent flow
of 16,428 m /day (4.34 mgd) during 1974  and  respective concentrations
for these parameters of 0.25, 1.0, 1.0,  0.1, 0.5, and  0.5 mg/1.

-------
A-14
     Phase II load limits were set at twice the Phase I  limits.   Maximum
daily limits for all parameters except fecal  coliforms were established
as 150 percent of average daily limits.   Fecal  coliform  limits,  as a
matter of policy, were set at 200 and 400 organisms/100  ml  for the av-
erage and maximum daily conditions, respectively.

     Important dates for the Clinton permit include the  date of  permit
issuance assumed as of around November,  1974; the  attainment of  full
Phase I production expected to occur around mid-1975; and the attainment
of full Phase II production expected in  late 1976  or early  1977.   The
expiration of the permit has been set as June 30,  1977.   The numerical
limitations have been staged to conform  to the  expansion phases  through
1977.

-------
              PHARMACEUTICAL INDUSTRY
                  CASE HISTORY B
 ELI LILLY AND CO.,  INC.,  TIPPECANOE  LABORATORIES
                 LAFAYETTE,  INDIANA

(FERMENTATION/SYNTHESIZED  ORGANIC  CHEMICALS  PLANT)

-------
                                                                     B-l
 ELI LILLY AND CO., INC., TIPPECANOE LABORATORIES,  LAFAYETTE,  INDIANA
          (FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS PLANT)
BACKGROUND

     The Eli Lilly and Company Tippecanoe Laboratories at Lafayette,
Indiana, manufactures bulk pharmaceutical fine chemicals, antibiotics,
herbicides, and miscellaneous products using fermentation and synthetic
chemical processing technology.

     The plant was reported first opened in 1954 as an antibiotics
plant, and production was broadened in 1957 to include chemical  manu-
facturing.  Since 1957 the Company indicates growth has been extensive,
approximating a tenfold increase.  In 1972, antibiotics and antibiotic
derivatives comprised 50 percent or more of total plant activities.
However, by 1974, the ratio of antibiotic production to total plant
activities had somewhat decreased.  Tippecanoe operations are currently
categorized as roughly 40 percent fermentation; 20-30 percent herbicides
production; and 30-40 percent production of various organic synthesized
drugs and medicinals.

     The plant makes in the order of 166 chemical intermediates  and  88
finished bulk products.  The major products include analgesics,  anaes-
thetics, antibiotics and antibiotic intermediates, antihistamines,
barbiturates and barbituric derivatives, alkaloids, mercuric compounds
(i.e. merthiolate), antidiabetics, opium derivatives, tranquilizers,  the
herbicide Treflan or trifluralin, and numerous other chemicals and
intermediates.  In the last few years, the plant has embarked upon large
scale recovery and sale of sulfuric acid.  No production figures were
obtained for sulfuric acid.  A changing product mix undoubtedly has  a
less than favorable impact upon the industrial waste treatment facili-
ties.

     In 1972, essentially three antibiotics were made with erythromycin
predominating.  Today the Cephalosporin-type antibiotics are largely
manufactured with some "Tylan" (an animal antibiotic) and possibly
others.  The production of trifluralin initiated in the past 3-4 years
is now quite significant.  All processing is supposedly carried  forth by
batch methods.  Final packaging is primarily done at Eli Lilly's Indian-
apolis, Indiana plants.

     Tippecanoe is operated 24 hours per day, 7 days a week, and has
about 1,050 employees.  Around May 1974, NFIC-Denver was asked to pro-
vide assistance in developing the NPDES permits for Eli Lilly and Com-
pany plants at both Lafayette and Clinton, Indiana.  The Tippecanoe
Laboratories at Lafayette was visited by NFIC-Denver on May 10,  1974,
and again on June 26, 1974.  The visits provided valuable information on
current plant practices.

-------
 B-2
 PROCESSES

     The Lafayette plant has four main boilers, three on coal and one on
 fuel oil.  Water supply for the plant, obtained from Company-owned
 wells,  is  treated with polyphosphates and chlorinated.  The potable
 water supply  is softened by lime treatment and the boiler water receives
 softening  principally by zeolite resins.  Wastewaters from the treatment
 plant are  discharged to the sanitary sewer system.

     "Treflan," the principal herbicide manufactured, is mainly a pre-
 emergent crab grass and weed killer utilized for cotton and soybean
 fanning.   According to the literature, the main process for synthesizing
 Treflan consists of nitrification in successive steps with p-chloro-
 benzotri fluoride, hydrochloric acid, sulfuric acid and fuming nitric
 acids as raw  materials.  The p-chlorobenzotrifluoride is converted over
 to  the  mononitro intermediate then to the dinitro intermediate, dissolved
 in  chloroform and aminated with soda ash to yield trifluralin.  Trifluralin
 is  mixed with aromatic naphtha and emulsifiers to the desired final
 formulation.  Spent acids are reported being recovered.  Residual acids,
 salt water layers, and chemical washwaters together with unrecoverable
 solvents are  discharged to the plant sewers.  It is noted that Treflan
 has a reported toxicity in the range of 14 to 58 ppb expressed as an
 LC   over  24  hours derived for relatively sensitive fish species.
WASTE TREATMENT AND CONTROL

Waste Treatment in the  1950's

     The early waste  treatment works, completed in April 1954, are
described  in a series of papers by Howe and Paradiso in 1956  (63), Howe
and DeMoss  in 1957 (33), and Paradiso and Howe in 1958 (100).

     The early treatment facilities consisted of neutralization, preaer-
ation (later converted  to modified activated sludge), biofiltration,
settling,  anaerobic digestion, nitrification, chlorination and sludge
drying.  Installation cost was cited in 1954 as $750,000.  Wastewater
problems were principally high dissolved and colloidal solids, BOD,
color, odors, and solvents.  Wastes being received into the system
/included antibiotic wastes consisting mainly of spent fermentation
broths and  associated effluents, and sanitary waste streams including
various equipment and floor washings.  Treatment design called for 3,975
m /day (1.05 mgd) of mixed wastewaters containing a maximum BOD load of
5,670 kg (12,500 lb)/day.

     Spent  cooking waters from the Tippecanoe installation amounted to
about 15,140 m /day (4.0 mgd) and received no treatment.  These were
merged with the treated antibiotic and sanitary wastewaters prior to
final discharge to the Wabash River.  It was anticipated that the sani-
tary wastes and.spent antibiotic broths, 290 m /day (77,000 gpd); floor
washings,. 230 m /day  (60,000 gpd); and wash waters, 720 m3/day (190,000
gpd) would  have the following waste loads:

-------
                                                                     B-3
                 Flow              BOD         Total Solids  	TSS
           (m3/day)  (gpd)   (kg/day)(Ib/day)(kg/day)(lb/day)(kg/day)(lb/day)
Min.
Avg.
Max.
flow
flow
flow
510
1,240
1,990
135
327
525
,000
,000
,000
860
3,240
3,360
1
7
7
,900
,140
,400
1960
9330
7640
4330
20,560
16,850
61
240
236
135
530
520
     Howe (6) describes the installation in the mid-1950's as having
biological treatment and chlorination of antibiotic wastewaters.  The
biological treatment system employed both turbine aeration and diffused
air, and had the following characteristics:
Parameter           Antibiotic Wastes             Treated Effluent
                    Influent
BOD
TS
pH
25,000 mg/1
35,000 mg/1
3.9
75 mg/1
580 mg/1
7.5
There were two aeration basins, 17m x 6m x 5m (56 ft x 20 ft x 16 ft)
having a mixed liquor solids level averaging 3000 mg/1.  The basins used
about 17.2 m  of air/kg (275 scf air/lb) of BOD removed.  The BOD
loadings were 7.2 to 9.6 kg/day/m  (450-600 lb/day/1,000 ft ) of
aeration basin capacity (33, 63, 100).  Equipment and floor washings,
sanitary wastes, cafeteria and laboratory wastes were combined into a single
stream that received pre-chlorination and passed through a comminutor
before discharge into the settler that followed the aeration basins.  The
combined wastes passed to a second treatment system consisting of high-rate
trickling filters each 23 m (76 ft) in diameter by 1.5 m (5 ft) deep
operating in parallel with the effluents going to a secondary clarifier.
Final clarifier overflows received post-chlorination before combining
with the cooling waters and discharging to the River.  Howe and Paradise (63)
in 1956 reported that the final effluent from the treatment system after
combining with the cooling water stream had an average BOD of 12.2 mg/1
(equivalent to approximately 230 kg (500 lb)/day BOD discharged); a pH
of 7.0 to 7.4; a total solids level of 380 to 410 mg/1; and no evidence
of color, odor or toxicity.  The clarifier sludges went to single or
double-stage anaerobic digesters, were dried and used as a soil conditioner.

-------
B-4
Digester supernatants went to a nitrification basin 8.8 m x 1.5 m x 3.8
m deep (29 ft x 5 ft x 12.5 ft deep), equipped with auxiliary aeration.
These nitrified supernatants were first routed to the trickling filters
but later diverted to the "primary" clarifier.

     The 1956 Eli Lilly report (63) cites the availability at that time
of vacuum filtration equipment for handling waste sludges, solvent
recovery operations, and evaporators for the recovery of spent broth.
It is not known whether the spent broth evaporators were ever put into
use.

     When the treatment facilities went into operation, it was soon
found that real operating conditions were much different than what had
been assumed under the original treatment plant design criteria.  Actual
waste loads were tremendously higher than the design criteria as shown:
Parameters       Unit         Design Loads        Actual Loads
                            (kg/day)(Ib/day)    (kg/day)(Ib/day)

BOD         Clarifiers        240     530    2,720-5,440  6,000-12,000
            Trickling filters                  910-1,810  2,000-4,000

TSS         Clarifiers      3,450   7,600    5,440-10,880 12,000-24,000
            Trickling
             filters        2,950   6,500    4,540-9080   10,000-20,000

Sludge      Digester          38^ 10,000^      151          40.000-/

a/  Volume in m /day.
b/  Volume in gpd;  5 percent solids content assumed.

     Typical performance results of the treatment system given by Eli
Lilly in 1958 are shown in Table B-l.   The data indicated the Tippecanoe
treatment facility was giving waste removals of 89.9 to 92.4 percent for
BOD and 90.5 to 92.4 percent for TSS.

     Howe and DeMoss (33) cite serious corrosion problems through the
1950's because of the nature of the wastes being handled.  Odors were
also a pervading problem but various approaches considered successful
were:

     1)   Quick removal of odor-producing solids from the biological
          oxidation processes.
     2)   Adequate oxidation of organic matter with sufficient air.
     3)   Addition of odor-suppressant chemicals.
     4)   Control of odor-producing vapors by water mists.
     5)   Evaporation of spent broths.
     6)   Use of oxidizing chemicals such as chlorine-containing agents.
     7)   Complete digestion of solids.
     8)   Complete combustion of sludge gases.
     9)   Close control of treatment processes.

-------
                        Table B-l
 Characteristics of Wastewaters Treated Around 1957-1958
Eli Lilly & Co., Tippecanoe Laboratories,  Lafayette,  Ind.
Source
Antibiotic Waste
Neutralization Tank (A)
Influent
Tank (A) Effluent
to Antibiotic Waste
Holding Tank (B)
Tank (B) Effluent
to Aeration Tank (C)
to Digesters
Sanitary Wastes
to Aeration Tank (C)
Secondary Clanfier
Final Effluent
leaving chlorine
contact Chamber
Temperature
C F
49
49
32
32
18-71
10-24
24-29
120
120
90
90
65-160
50-75
75-85
PH
(kg/day)
3-5 8,620
7 8,620
6.0 4,310
6.0 4,310
2.3-13 363
7.2 680
7.3 680-910
BOD
(Ib/day)
19,000
19,000
9,500
9,500
800
1,500
1,500-2,000
Total
(kg/day)
9,660
9,660
4,830
4,830
725
1,130
1,130
Solids
(Ib/day)
21,300
21,300
10,650
10,650
1,600
2,500
2,000
TSS
(kg/day)
5,440
5,440
2,720
2,720
499
570
454
(Ib/day)
12,000
12,000
6.000
6,000
1,100
1,250
1,000
TDS
(kg/day)
4,220
4,220
2,110
2,110
227
570
454
(Ib/day)
9,300
9,300
4,650
4,650
500
1,250
1,000
                                                                                                                co
                                                                                                                 i
                                                                                                                en

-------
B-6


Waste Treatment in the 1960's

     Howe in 1962 (21) described various waste streams originating from
antibiotics production as:

     1)  Liquid and solid wastes directly from fermentation processes
         (high strength);
     2)  Liquid wastes from extraction and purification processes
         (high strength);
     3)  Liquid and solid wastes from recovery processes (high strength);
     4)  Washings of equipment and floors (varying strength);
     5)  Spent cooling waters;
     6)  Laboratory wastes, sanitary sewage and miscellaneous (varying
         strengths).

     He characterized these wastes as having an irregular flow pattern,
a continuously-changing waste composition, high dissolved and colloidal
solids, fluctuating pH, high BOD, temperature, toxicity, color and odor.

     Changes in the waste treatment facilities during the mid-19501s and
early 1960's basically consisted of the addition of a third-stage
biological treatment step: activated sludge following the high-rate
trickling filters.  The high-rate trickling filter effluents were returned
to the primary clarifiers, thence through the chlorine contact chamber
and the third-stage biological treatment sub-system before final dis-
charge to the river.  The third stage of biological treatment was report-
ed to consist of a 378 m  (100,000 gal.) clarifier and a shallow 1.8 m
(6 ft) deep aeration basin of 3,400 m  (900,000 gal.) capacity.

     The  clarifier following third-stage activated sludge was under
construction in early 1962.  At that time, chemical wastes, presumably
from organic synthesis, were observed as being present in the waste
treatment circuit.  The chemical wastes were being introduced into the
(aerated) nitrification tank, prior to the primary settlers and the
second-stage biological trickling filters.

     Waste characteristics of the waste treatment facility influent in
the early 1960's are shown in Table B-2.

     High-rate two-stage aeration comprising the first stage of bio-
logical treatment was reported as giving 80 percent BOD reduction of the
strong antibiotic fermentation wastes.  The MLSS concentration in this
sub-system was maintained around 6,000 to 8,000 mg/1.  The high rate
trickling filter sub-system constituting the second stage of biological
treatment and receiving the treated first-stage antibiotic wastes plus
equipment and floor washings and chemical wastes, was reported as receiv-
ing BOD loads of 5,260 to 9,980 kg (11,600 to 22,000 lb)/day and pro-
viding 70 to 80 percent  BOD reduction. The BOD removal within the third
stage system was reported averaging about 50 percent.  Overall treatment
plant BOD removal efficiency was rated by Howe as between 90 and 95

-------
                                                                      Table B-2

                                     Characteristics  of Wastewater Entering Treatment System in Early 1960's (21)
                                               Eli  Lilly & Co., Tippecanoe Laboratories, Lafayette, Ind.
Wastewater
                          Flow
                            BOD
                    Total  Solids
                    (m3/day
              Suspended and Colloidal
                       Solids
          (gpd)   (mg/1)  (Ib/day)  (kg/day)  (mg/1)  (lb/day)(kg/day)  (mg/1)  (Ib/day)  (kg/day)    PH
                          Color

                         (Co-Pt)
Temperature

  C   F
Antibiotic Liquid
   Wastes
  568   150,000   15,000-  18,000-  8160-    20,000-  25,000-  11,300-  10,000-  15,000- 6800-   3-10.5
        average   35,000  40,000   18,100   50,000   63,000   13,000   40,000   40,000  18,100
                                                                         20,000-   43.3-48.9  110-120
                                                                         30,000
Equipment and        378-   100,000-    2,000-   4,000-1810-     3,000-  6,100-  2,770-  2,000-  4,000-1810-   2.0-11.5   1,000-  12.8-29.4  55-85
   Floor Washings   1510    400,000    8,000   16,000   7260     15,000  31,000  14,100   9,000  18,200   8,300             5,000
   and Sanitary
   Wastes
Total Wastes
 946-   250,000-
1510    400,000
22,000- 9980-
56,000  25,400
31,100- 14,100-
94,000  42.600
19,000-  8620-
58,200  26,400
                                                                                                                                                          co
                                                                                                                                                           i

-------
B-8
percent.  Ammonia content in the final effluents was 4.3 mg/1.  Compar-
ing waste loads of 1961 (21) with those of the 1950's at Tippecanoe, the
important conclusion may be drawn that raw waste loads more than doubled
over the intervening 5 to 7 years, while expansion in the waste treat-
ment works was far less significant.

     Howe (21) cites special studies conducted by Eli Lilly towards
improving waste removal efficiencies and other investigations on the
toxicity of certain antibiotics to biological treatment.  Some anti-
biotics even if present at very low concentrations, are toxic to the
microflora in the aeration system and can quickly render the entire
system static.  Even though the microflora may be acclimated to the
presence of one antibiotic at a certain concentration, it may fail
completely when another antibiotic is added to the system.  Howe recom-
mends that certain antibiotic wastes should be diluted to ensure succes-
sful biological treatment.  This is best accomplished in large aeration
tanks that can accommodate high hydraulic and BOD loadings.
Sludge Centrlfuging. Late 1960's and Early 1970's

     A 1972 report (52) describes sludge centrifuging within the waste
treatment installation.  After activated sludge treatment, spent anti-
biotic wastes are settled in a series of clarifiers and thickeners.  The
settled sludges are pumped to a Pennwalt Sharpies P-5400 centrifuge.
Supernatants are returned to the waste treatment plant, while sludge
solids are pumped into a truck and hauled away.  The centrifuge receives
265 to 568 1/min (70 to 150 gpm) of sludge liquors with a solids content
of 3 to 6 percent.  The outflows consist of a centrate having no more
than 0.8 percent solids and a readily-handled sludge concentrate of 20
to 30 percent solids, amenable to incineration.  The centrifuge system
is reported as having reduced sludge truck hauling needs by as much as
80 percent.  The 363 kg (800 lb)/minute of sludge previously generated
were reduced to about 75 kg (166 Ibl/minute.   Sludge volume was reduced
from 0.35 to 0.07 m(12.5 to 2.5 fr)/minute.

     Similarly, the synthetic chemical wastes are neutralized and settled,
with the underflows going to centrifuges.  The partially-treated chemical
waste overflows from the settlers are combined with the effluents from
the antibiotics-sanitary waste treatment systems.  The combined flows
receive final  settling in two clarifiers before discharge to the River.

     The centrifuge sub-system was installed  in 1970.   Antibiotic sludges
were admixed with relatively small quantities of Nalco polyelectrolytes
(0 to 11 1/min (0-3 gpm) of 0.02 percent Nalco No. 610 solution).
Nothing was added to the chemical waste sludges.  The Tippecanoe plant
as of 1974 had two centrifuge units in operation.

-------
                                                                     B-9
Odor Abatement. Late 1960's, Early 1970's

     Around 1967, the Company covered the aeration tanks in its first
stage of biological treatment.  The centrifuge operations were also
enclosed within a separate building.  Odors from these processes were
carried in a 1.5 m (5 ft) ducting system to an incinerator put into
operation around this time.  In 1970, the Company converted the two
high-rate trickling filters previously available into aeration tanks,
and gave them dome-type coverings to contain the noxious odors and
eventually carry them to incineration.  In order to supplement the
single incinerator, the Company attempted to duct the collected noxious
air streams into the fire boxes at its main power plant.  However,
because of the large volumes of air being carried and secondary odors
being created, they found it necessary to install two more incinerators,
burning natural gas on an interruptable schedule.  Low sulphur fuel oil
is used as standby.  The incinerator off-gases are water scrubbed with
effluents discharged to the liquid waste.system.  Eli Lilly, as of 1973,
reported that more than $2 million had been invested in odor abatement
equipment, and fuel bills for the three incinerators were running about
$13,000 monthly (49).  The present air ducting system also vents various
fermentation buildings and some of the more critical chemical process
sectors.  As of early 1974, the Company fully covered one of the two
aerated lagoons within the chemical waste treatment system and the
second lagoon was being enclosed as of around June, 1974.

Existing Haste Treatment and Control Systems

     The Tippecanoe treatment works handle waste waters from the fermen-
tation area separate from the chemical wastes coming from general chemi-
cal manufacturing.  The overall works include more or less three sep-
arate waste collection and treatment systems:  1) the spent fermentation
antibiotics broth; 2) sanitary wastes, floor and equipment washings and
miscellaneous; and 3) mixed chemical waste waters.

     The fermentation wastewaters enter the "100 System."  This treat-
ment system consists of waste holding; settling, 378 m  (100,000.,gallon)
capacity; "pre-aeration" in four activated sludge basins 1,700 m
(450,000 gallon) capacity apparently in parallel; and two sludge thicken-
ing tanks.  Thickened sludges are sent to the centrifuge building.
Daily flows being received into this system in mid-1974 were averaging
946 to 1,140 m /day (250,000 to.,300,000 gpd), or 43 to 72 percent over
the system design flow of 662 m /day (175,000 gpd).  New additions were
planned to the fermentation treatment system by the Company, probably
during the latter half of 1974, but the new criteria appear to be less
desirable than original  design objectives.  Manufacturing change and
expansion seem to be occurring much faster than the capability of the
treatment facilities to handle such changes.   Overflows leaving the
"100" Treatment System are directed to second-stage biological  treatment.

-------
B-10
     Sanitary sewage, equipment and floor washings, washwater from the
mycelia storage bins, backwashes from the water treatment system and
miscellaneous streams enter the "200 System."  In this system, the
appropriate waste streams, after passing through a comminutor, join with
the fermentation overflows from the sludge thickening tanks and these
combined streams enter "primary clarifiers" in parallel followed by two
aeration tanks in parallel.  Sludges from the 200 System primary clari-
fiers are sent to the centrifuge building or carted away to landfill.
The aeration basin effluent combines with effluent from the "300 system"
and enters two "final clarifiers" before eventual discharge.  These
final clarifiers were observed to be significantly overloaded especially
with respect to TSS removal.

     The somewhat independent treatment system for chemical wastewaters
including those from trifluralin manufacturing is described as the "300
System."  The system is a series of basins installed at various intervals
since 1954.  Chemical wastewaters enter a holding chamber, overflow to
a clarifier, then flow to a  box where lime is added for neutralization,
a second clarifier, an aerated chamber followed by another aerated
chamber and then an aerated basin, with the overflows then going into
another clarifier and a pump pit.  From the pump pit, the chemical waste
flows enter two "activated sludge" aeration basins or lagoons operated
in series.  Waste detention time is reported to be 1.0 to 1.4 days.
Unfortunately, the chemical waste treatment system has ample oppor-
tunities for bypassing and some of these bypassed streams can possibly
enter the untreated cooling water drainage into Outfall 001. The chemical
waste treatment system is presently utilizing a full-scale but experi-
mental carbon absorption facility, mainly for the removal of color from
the Treflan wastewaters.  These particular wastes resulting from nitri-
fication have an intense yellowish-orange color which is not adequately
reduced through biological treatment.

     Other important aspects of the waste treatment and control works
include the sludge centrifuge building housing a pair of Sharpies centri-
fuges, a centralized solvent recovery sector, extensive air ducting
systems for collecting and carrying noxious odors from select- processing
and treatment sectors to three different incinerator units on site, and
a John Zink thermal oxidizer used for burning concentrated liquid and
solid wastes.  The John Zink and the other three incinerators are equip-
ped with water scrubbing devices which in turn return some organic loads
to the liquid waste collection system.

      Selected  concentrated wastes  having strengths of a few thousand ng/1
 of BOD and COD as  reported by the  Company are stored within six 116 m
 (30,000 gal.)  tanks  built into the side  of the hill  next to the John Zink
 thermal  oxidation  unit.   The  wastes  are  eventually incinerated in the
 John  Zink  system.  The  scrubber effluents from the John Zink together with
 surface drainage from around  the incinerator unit  and associated waste
 storage tanks  appear to  flow  directly into the Outfall  001  and then to the
 Wabash River.

-------
                                                                    B-ll


     Mycelium separated from fermentation broths is collected,  dried and
mixed with animal feeds or otherwise prepared as a pre-mix feed.   Mycelium
not incorporated into feed supplements is carted away to landfill.   In
the recovery area, mycelium is stored in a "live bottom bin" prior to
going to the dryer.  The storage bin is periodically cleaned with liquids
bled to the sanitary waste ("200") system.

     The combined treated fermentation, sanitary, chemical and  miscel-
laneous wastes leaving the final clarifiers from the 100, 200 and 300
waste treatment systems are merged with untreated spent cooling waters
in Outfall 001 that discharges to the Wabash River.  The cooling waters
in this sewer have been reported by the Company as uncontaminated,
originating from condensers and cooling jackets in the plant.  Outfall
001 also captures barometric condenser discharges, and bottoms  from a
distillation column contributing 760 to 1140 rrT/day (200,000 to 300,000
gpd), in addition to scrubber effluents from the John Zink incinerator
contributing 2,270 1/min (600 gpm).  Accompanying data will show these
cooling waters often have greater wasteloads than are contained in the
(100, 200 and 300) treatment plant effluents.  This constitutes a large
part of the existing problems at Tippecanoe.  Flows from the 100, 200
and 300 waste treatment systems total approximately 3,780 to 5,300
m^/day (1.0 to 1.4 mgd), while Outfall 001 has about 30,300 to 45,400
m /day (8.0 to 12.0 mgd) of final wastewaters.

     Outfall 002 serves the west side of the main factory grounds.  This
outfall is reported to contain mainly cooling waters from the jacketing
of vessels manufacturing herbicides plus barometric condenser waters
from units manufacturing medicinal drugs.  This flow, averaging 3,780 to
7,560 m /day (1-2.0 mgd), receives no treatment except for a small
retention pond close to the River.  The Company recognizes some process
contamination coming from barometric condensers together with acidity in
this discharge.  In October 1972, Outfall 002 was the site of a watery
solvent spill, the materials lost at that time comprising tetra-
hydrofuran and diethylanaline from the herbicide manufacturing sector.

     Outfall 003 is a relatively new outfall receiving hillside drainage
from some 80-120 ha (200-300 acres) of land previously used for waste
disposal on top of the promontory overlooking the Wabash River.  This
drainage is now reaching the River via diverse channels downstream of
the main plant, apparently spreading over about one mile of river frontage.
Approximately one-half of the drainage is being collected by the Company
and routed through the levee in a single flow to the River.  This portion
of the drainage is,Outfall 003.  The remainder of the hillside drainage
joins with Outfalls 001 and 002 or reaches the River via other means.
Total drainage is estimated in the range of 189-757 m /day (50,000 to
200,000 gpd).  The NPDES permit calls for collecting and bringing this
entire drainage back to the main treatment works by 1977 at which time
Outfall 003 and other similar drainage will be discontinued.

-------
B-12


     The Company currently has no chlorination facilities either for the
separate treatment plant effluents, or for the 001, 002 and 003 effluents,

     Lilly has been experiencing difficulty in disposing of concentrated
solid wastes.  In Spring, 1974, the Company requested permission of the
State to dispose of selected waste materials to public sanitary land-
fill, but permission could not be secured because of previous leaching
problems at the landfill site.  The wastes were of four general types:

     1)   Residues, principally "still bottoms," from distillation-
          purification operations.  Materials ranged from glossy solids
          to viscous tars.
     2)   Filter cakes.
     3)   Organic compounds - including materials not meeting quality
          specifications for sale.
     4)   Filter papers - composed of glass, rayon or paper fabric
          materials, and discarded chemical compounds.

     The various waste materials, stored in metal drums, consisted of
filter cake from formylglycine, thiophene press papers, off-spec alpha-
methyl homoveratryl amine and homoveratric acid, and various waste
residues resulting from 2-(3 Phenoxyphenyl).propionitrile, "Papaverine,"
ethyl malonic ester, isoamyl malonic ester, phenoxyacetophenone, cyclo-
pentamine, dichlorobenzene and dioxyline phosphate.

WASTE LOADS

Waste Loads 1958-1971
     The 1958 report by Eli Lilly (100) cited raw waste loads at Tippe-
canoe of around 9,070 kg (20,000 Ib) BOD/day, TSS of 5,900 kg (13,000
lb)/day and TDS of 10,400 kg (23,000 lb)/day.  Waste removals were given
as 83-91 percent for BOD and 86-92 percent for TSS.  A 1962 report by
Lilly (21) described total waste flows approaching 946 to 1,510 m /day
(250,000 to 400,000 gpd) containing daily BOD loads of 9,980 to 25,400
kg (22,000 to 56,000 lb)/day and total solids of 14,100-42,700 kg (31,000
to 94,000) Ib/day.  Information sheets provided by the Company to the
State in March 1969 stated that BOD loads into the treatment system at
that time were running 5,700-19,300 kg (12,600 to 42,500 lb)/day, but
future loads could be expected in the range of 18,100-22,700 kg (40,000
to 50,000 lb)/day.  The final effluents to the River were said to approxi-
mate a total of 22,700 rrT/day (6 mgd) containing 47-50 mg/1 BOD and 25-
50 mg/1  TSS.  These latter figures likely constituted the prime basis
for the State's allowable discharge limits of 1,430 kg (3,150 Ib) BOD/day
and 544 kg (1,200 Ib) TSS/day approved on September 21, 1971.  The State
specified that if these limits were exceeded, additional treatment would
be necessary.  Over the past few years, Tippecanoe Labs has consistently
exceeded these limits.

-------
                                                                    B-13
     In December 1971, the State informed the USEPA that Lilly's new
waste treatment facilities being installed at Tippecanoe would have a
final plant flow of 34,200 m /day (9.03 mgd) and achieve 95 percent BOD
reduction, producing by May 1972 an effluent containing 42 mg/1 BOD or
1,430 kg (3,150 Ib) BOD/day, down from the 5,850 kg (12,900 Ib) then
being discharged.  The EPA gave temporary assent to these load limits
but mentaioned that other undesirable waste constitutents were also
present in the final effluents including cyanides, mercury, phenolics,
chlorinated hydrocarbons and pesticides.

Present Waste Loads

     The Company in mid-1974 provided the State and the EPA with a
summary of monthly waste load inputs to the waste treatment works for
the period of May 1973 through April 1974.  In terms of BOD raw waste
loads, these tentative figures showed 7120 kg (15,700 lb)/day origin-
ating from the antibiotics spent broth sector; 2490 kg (5,500 lb)/day
described as sanitary and floor wash raw loads; and 10,000 kg (22,100
lb)/day attributable to chemical process wastewaters.  These raw loads
averaged 19,700 kg (43,500 Ib) BOD/day but monthly values were reported
as high as 27,600 kg (60,900 lb)/day.

     A recent accounting of raw waste loads was made from detailed
Company sampling and analysis sheets on the individual 100, 200 and 300
waste treatment systems for the separate months of March and May, 1974.
March 1974 was reported as one of the highest months during 1973-1974 in
terms of raw and final waste loads at the Tippecanoe treatment works.
Combined treatment plant flows for Maisch through May, 1974, varied on a
monthly basis between 3600 and 4430 m /day (0.95 and 1.17 mgd), averaging
3970 m /day (1.05 mgd).  Calculations by NFIC-Denver, based upon this
data, show total plant raw waste loads as well as waste removal efficiencies
within each of the three treatment systems in Table B-3.  The antibiotics
fermentation sector was estimated to contribute approximately 40 percent
of BOD, TSS and COD plant waste loads.  The average of the total plant
raw waste loads for the two selected months in 1974 were 24,500 kg
(54,000 Ib) BOD/day,  44,400 kg (97,800 Ib) COD/day, 87,700 kg (193,400 Ib)
TDS/day, and 5760 kg (12,700 Ib) TSS/day.

     The results in Table B-3 demonstrate that current treatment re-
ductions are lower than the values previously reported by the Company.
The treatment systems appear to lack sufficient capacity and flexibility
for coping with current waste loads.  On a number of occasions, the
State has raised questions as to the sludge handling and TSS removal
capabilities of the existing systems.  The waste performance data show,
in particular, the highly inadequate TSS removals being experienced.

-------
                                                                                                              OT
                        Table B-3

Treatment Plant Raw Waste Loads  and  Removal  Efficiencies
Eli Lilly S Co., Tippicanoe Laboratories,  Lafayette,  Ind.
                   March and May, 1974
Percent Removals
Parameter
BOO
COD
TSS
TDS
BOD
COD
TSS
TDS
Flow Range
Total Plant Raw
Waste Loads
(kg/day) (Ib/day)
28,300 62,300
50,300 111,000
6,300 13,900
98,900 218,100
20,400 44,900
38,300 84,500
5,200 11,500
76,500 168,700

Antibiotic "100"
Treatment System
78
59
[173% Incr.]
42
77
60
[280* Incr.]
34
795 - 1,135 m3/day
{0.21 - 0.30 mgd)
x
Sanitary + Misc.
"200" Treatment System
March, 1974
[Incr.]
[Incr.]
[Incr.]
[Incr.]
May, 1974
[Incr.)
[Incr.]
[Incr.]
[Incr.]
1,325 - 2,271 m3/day
(0.35 - 0.60 mgd)
Chemical "300"
Treatment System
72
72
70
33
80
67
59
22
946 - 1,514 m3/day
(0.25 - 0 40 mgd)
Overall
Treatment Systems*
71.8
64.8
51.8
28.2
80.0
68.0
19.9
21.6
3,407 - 4,540 m3/day +
(0.90 - 1 20 mgd +)

-------
                                                                    B-15


     In October 1974, Eli Lilly presented the USEPA with a new set of
raw waste loads for Tippecanoe Laboratories in terms of the maximum
operating month:

                                   BOD                        COD
Source                      	           	
                            (kg/day) (Ib/day)          (kg/day) (Ib/day)

Fermentation
  wastes                    10,400    23,000           15,900    35,000
Sanitary and dilute
  process wastes             2,450     5,400            5,900    13,000
Chemical wastes             14,100    31,100           33,100    73,000
Secondary feed to
  John Zink unit             3,810     8,400            9,530    21,000
Spent cooling and
condenser waters
Hillside drainage
Total
540
950
32,300
1,200
2,100
71,200
770
1,450
66,600
1,700
3,200
146,900
     In November 1974, total raw waste loads assumed for the Lilly,
Lafayette NPDES discharge permit were raised by the EPA from 24,500 to
30,200 kg (54,000 to 66,500 Ib) BOD/day, and from 49,900 to 54,700
(110,000 to 120,500 Ib) COD/day.

     A summary of monthly effluent values were derived from Lilly,
Lafayette's Monthly Report Sheets to the State.  These loads applicable
to Outfalls 001 and 002 for the period of January 1973 through April
1974 are presented in Tables B-4, 5 and 6.  The information shows that
waste loads in the combined treated effluents plus cooling water dis-
charges to the Wabash River have significantly exceeded the State
limits approved on September 21, 1971, for all months from January 1973
to April 1974.  Over this 16-month period, BOD loads averaged 5,000 kg
(11,100 lb)/day, with maximum and minimum months of 8,030 kg (17,700 lb)/day
and 2,900 kg (6,400 lb)/day respectively.  The minimum month loading was
about twice the State-stipulated limits for BOD and three times the
limitation for TSS.

     The combined effluents from the three Tippecanoe treatment systems
(100, 200, and 300), which discharge into Outfall 001 were previously
cited as averaging about 3,970 nr/day (1.05 mgd).  Considerable spent
cooling waters plus miscellaneous streams-cause an increase in the
Outfall 001  flow up to 30,800 to 40,500 rrT/day (8.4 to 10.7 mgd) as
illustrated by Table B-4.  Comparing these cooling waters vs. treatment
plant effluents for May 1974, it was found that the spent cooling water
sewer contributed 17 to 48 percent of the waste discharge loads within
Outfall 001.   For March 1974, the cooling water sewer contributed 33

-------
                                                                                                       oo
                                                                                                        i
                         Table B-4


              Plant Waste Loads (Outfall  001)

Eli Lilly and Co., Tippecanoe  Laboratories,  Lafayette,  Ind.

                 January 1973  - April  1974
Flow
Month
Jan. 1973
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 1974
Feb.
Mar.
Apr.
Avg. 16 mos
Data
(m3/day)
40,700
35,200
31 ,800
36,000
33,900
36,600
37,000
37,200
36,300
37,800
37,200
34,000
36,900
35,100
35,500
37,700
36,200
(mgd)
10.74
9.30
8.41
9 50
8.96
9.67
9.78
9.82
9.60
9.99
9.83
8.98 >
9.75
9.27
9.38
9.95
9.56
COD
(kg/day)
12,100
14,700
13,100
14,700
15,000
11,100
10,800
12,000
10,500
11,200
10,400 .
15,300
16,400
18,000
17,100
10,300
13,300
(Ib/day)
26,700
32,500
28,800
32,300
33,000
24,400
23,700
26,400
23,200
24,800
22,900
33,800
36,100
39,600
37,700
22,700
29,300
BOD
(kg/day)
6,580
5,400
4,670
5,400
6,080
2,990
2,590
3,180
2,810
3,810
3,720
6,580
5,850
7,760
6,460
3,040
4,810
(Ib/day)
14,500
11,900
10,300
11,900
13,400
6,600
5,700
7,000
6,200
8,400
8,200
14,500
12,900
17,100
14,250
6,700
10,600
IDS
(kg/day)
55,500
61,100
48,300
52,900
54,100
57,200
51 ,600
53,000
49,000
50,300
51,100
50,500
56,200
61,100
64,000
58,300
54,700
(Ib/day)
122,400
134,800
106,500
116,600
119,300
126,200
113,700
116,800
108,000
110,800
112,700
111,400
123,900
134,700
141,200
128,600
120,500
TSS
(kg/day)
3,450
2,450
2,090
2,900
3,190
2,770
4,720
5,800
6,260
4,170
1,680
2,220
3,760
2,360
2,770
3,270
3,370
(Ib/day)
7,600
5,400
4,600
6,400
7,030
6,100
10,400
12,800
13,800
9,200
3,700
4,900
8,300
5,200
6,100
7,200
7,420

-------
                        Table  B-5
             Plant Waste Loads (Outfall  002)
Eli Lilly & Co., Tlppecanoe  Laboratories,  Lafayette,  Ind.
                January 1973 - April  1974
Flow
Month
Jan. 1973
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 1974
Feb.
Mar.
Apr.
Avg. 16 mos
Data
(m3/day)
4,160
5,450
5,720
6,060
5,110
6,090
7,760
6,700
6,320
6,060
5,110
3,290
5,450
5,720
5,070
6,210
5,640
(mgd)
1.10
1.44
1.51
1.60
1.35
1.61
2.05
1.77
1.67
1.60
1.35
0.87
1.44
1.51
1.34
1.64
1.49
COD
(kg/day)
118
154
399
490
363
744
513
449
884
590
190
118
680
431
240
313
408
(l'b/day)
260
340
880
1,080
800
1,640
1,130
990
1,950
1,300
420
260
1,500
950
530
690
900
BOD
(kg/day)
41
50
218
290
236
472
295
236
317
240
77
50
313
263
141
171
213
(Ib/day)
90
110
480
640
520
1,040
650
520
700
530
170
110
690
580
310
380
470
IDS
(kg/day)
2,090
2,630
3,760
2,900
2,590
3,490
3,950
4,080
3,040
3,360
2,950
1,860
2,720
3,520
2,180
6,170
3,220
(Ib/day)
4,600
5,800
8,300
6,400
5,700
7,700
8,700
9,000
6,700
7,400
6,500
4,100
6,000
7,760
4,800
13,600
7,100
TSS
(kg/day)
54
50
109
95
82
154
122
141
95
68
50
45
109
109
73
95
91
(Ib/day)
120
no
240
210
180
340
270
310
210
150
no
100
240
240
160
210
200
                                                                                                     co
                                                                                                      i

-------
                                                                                                      07
                                                                                                       I
                                                                                                      _j
                                                                                                      00
                       Table  B-6
     Plant Waste Loads  (Sum of Outfall  001  and  002)
Eli  Lilly & Co., Tippecanoe Laboratories, Lafayette,  Ind.
                January 1973 - April  1974
Flow
Month
Jan. 1973
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 1974
Feb.
Mar.
Apr.
Avg. 16 mos
Data
(m3/day)
44 ,800
40,700
37,600
42,000
39,000
42,700
44 ,800
43,900
42,700
43,900
42,300
37,300
42,400
40,800
40,600
43,900
41 ,800
(mgd)
11.84
10.74
9.92
11.10
10.31
11.28
11.83
11.59
11.27
11.59
11.18
9.85
11.19
10.78
10.72
11.59
11.05
COD
(kg/day)
12,200
14,900
13,500
15,200
15,300
1 1 ,800
1 1 ,200
12,400
1 1 ,400
1 1 ,800
10,600
15,500
17,100
18,400
17,300
10,600
13,700
(Ib/day)
27,000
32,800
29,700
33,400
33,800
26,000
24,800
27 ,400
25,200
26,100
23,300
34,100
37,600
40,600
38,200
23,400
30,200
BOD
(kg/day)
6,620
5,440
4,900
5,720
6,300
3,450
2,900
3,400
3,130
4,040
3,810
6,620
6,170
8,030
6,620
3,220
5,030
(Ib/day)
14,600
12,000
10,800
12,600
13,900
7,600
6,400
7,500
6,900
8,900
8,400
14,600
13,600
17,700
14,600
7,100
11,100
IDS
(kg/day)
57,600
63,800
52,100
55,800
56,700
60,700
55,500
57,100
52,000
53,600
54,100
52,400
59,000
64,600
66,200
64,500
57,900
(Ib/day)
127,000
140,600
114,800
123,000
125,000
133,900
122,400
125,800
114,700
118,200
119,200
115,500
130,000
142,500
146,000
142,200
127,600
TSS
(kg/day)
3,490
2,490
2,180
2,990
3,270
2,900
4,850
5,940
6,350
4,260
1,720
2,270
3,860
2,450
2,860
3,360
3,450
(Ib/day)
7,700
5,500
4,800
6,600
7,200
6,400
10,700
13,100
14,000
9,400
3,800
5,000
8,500
5,400
6,300
7,400
7,600

-------
                                                                    B-19
to 70 percent of the total waste loads being conveyed by Outfall  001  to
the River.  In March 1974, the effluent loads were 6,490 kg (14,300 Ib)
BOD/day and 2,770 kg (6,100 Ib) TSS/day, which were about five times
greater than the State limits of September 21, 1971.   Besides less than
desirable waste removal efficiencies by the treatment facilities, a
large portion of the final waste loads is due to the untreated cooling
water streams.  These "undefined" waste sources should be eliminated  or
controlled at the earliest possible time.

     Monthly values on ammonia nitrogen loadings covering the period
January through June 1974 were provided by the Company (Table B-7). The
data clearly shows that the large majority of unoxidized nitrogen is
associated with the chemical wastewaters within the "300" Sector.
Chemical wastewaters contained ammonia concentrations varying from 840
to 1,130 mg/1, equivalent to an average loading of 1090 kg (2,400 lb)/day.
Outfall 001 receiving combined effluents from the 100, 200 and 300
treatment systems plus large amounts of spent cooling waters had an
average ammonia load of 1,720 kg (3,800 lb)/day.  The total ammonia
loads to the Wabash River originating from all three Tippecanoe outfalls
(001, 002 and 003) averaged about 1,810 kg (4,000 lb)/day.  In compari-
son with this 1974 data, the discharge permit applications of April 1972
and May 1974 showed considerably less ammonia-N:  508 kg (1,120 lb)/day
and 1,102 kg (2,430 lb)/day, respectively, for Outfall 001.  The EPA
survey of September 1973 found 658 kg (1,450 lb)/day ammonia-N, and the
State-EPA survey of August 1974 illustrated 1,430 (3,150 lb)/day for
Outfall 001.

     Outfall 002 was discharging 3,785 to 7,950 m3/day (1.0 to 2.1 mgd)
(Table B-5) of contaminated cooling waters as indicated by monthly BOD
loads ranging from 41 to 517 kg (90 to 1,140 lb)/day and averaging 213
kg (470 lb)/day.  As high as 884 kg (1,950 lb)/day of COD was discharged
with the peak recorded in September 1973.

     Outfall 003 represents underground waste percolation and drainage
intercepting the surface strata and finding its way to the Wabash River.
Total drainage is in the range of 189 to 568 m /day (50,000 to 150,000
gpd), or possibly greater.  Company data of July 1974, and an EPA-State
Survey in August 1974, showed the following characteristics for this
outfall:

-------
                       Table B-7

                Ammonia Nitrogen  Loadings
Eli Lilly & Co., Tippecanoe Laboratories,  Lafayette,  Ind.
                   January - June,  1974
CO

ro
o
Waste Stream
Influent to "100 System"

Influent to "200 System"

Influent to "300 System"

Effluent from "300 System

Combined Effluent
"100, 200, 300 Systems"

Outfall 001

Outfall 200


Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Concentrations
(rag/1)

110-470
37
27-52
970
840-1,130

390-1,180
280
240-350
52
38-64
3
1-5

(kg/day)
159-381
91-544
68
41-109
1,089
798-1,393




1,715
1,270-2,082
25
9034
Loads
db/day)
350-840
200-1,200
150
90-240
2,400
1 ,760-3,070




3,780
2,800-4,590
55
20-75

-------
                                                                     B-21
                  Company  Data             EPA State Survey
 Parameter          July 1974                August 1974
                                        (mg/l)(kg/day)(lb/day)
,BOD5
BOD,n
30
COD
TSS
IDS
NH3-N
66-660
__

260-1,240
8-110
1,660-2,800
188-350
170
1,020

410
40

295
37
222

91
9

66
82
490

200
20

145
 DEVELOPMENT OF  NPDES  PERMIT

     Average daily  load  limitations were prescribed for Lilly, Lafayette
 by  1977  based upon  the average of the better performing treatment
 situations across the pharmaceutical industry.  The Lafayette permit
 closely  approaches  the norm of waste load reductions previously written
 into a series of other permits on similar type pharmaceutical plants.
 Specifically, BOD,  COD and TSS reductions based upon maximum monthly raw
 waste loads were respectively 94.9 percent, 79.4 percent, and approximately
 80.0 percent.   The  1977  average daily loads were limited to 1,540 kg
 (3,400 lb)/day  BOD; 11,340 kg (25,000 lb)/day COD; and 1,130 kg (2,500 lb)/day
 TSS.  Maximum daily allowable limits were approximately double the
 average  daily limits.  Ammonia nitrogen limitations for 1977 were established
 as  500 kg (1,100 lb)/day for the average daily condition and 1,130 kg
 (2,500 lb)/day  as a maximum daily limit.  Tippecanoe could technologi-
 cally, but perhaps  not economically, reduce future unoxidized nitrogen loads to
 considerably lower  levels.

     Temperature was  not perceived a problem with Outfall 002 and 003,
 but constitutes potential concern with Outfall 001.  Average temperatures
 for the  001 effluents are in the range of 32-36 C (89-97F), but
 individual daily temperatures have exceeded 38 C (100F), occasionally
 reaching 41 C  (105F).  The State has decided that the Wabash River
 demonstrates sufficient  dilution to preclude any thermal  impacts.
 Temperature limitations  were not found necessary although careful
 monitoring will be  required.

     The May 1974 discharge permit application of Lilly,  Tippecanoe
 reported relatively high average concentrations of trace  metals and ions
 in Outfall 001:   aluminum - 2.0 mg/1, chromium - 0.58 mg/1,  and pheno-
 lics - 0.83 mg/1.   Lead was reported as less than 1.0 mg/1  and mercury
was indicated of probable concern from other sampling information.

-------
B-22
Other data shows that the same parameters are relevant for Outfall 002.
The average daily limit for mercury was set at 0.045 kg (0.1 lb)/day,
whereas concentration limits for chromium, phenolics, and lead were
established as 0.25 mg/1, 0.25 mg/1 and 0.1 mg/1, respectively.- In the
permit, fecal coliforms were not to exceed 200 and 400 organisms/100 ml
respectively for the average daily and maximum daily conditions.  The
permit further specified that 96-hour TL  fish bioassay data would be
collected every 6 months for each outfall.

-------
              PHARMACEUTICAL INDUSTRY
                  CASE HISTORY C

             PFIZER,  INC.,  VIGO PLANT
               TERRE  HAUTE, INDIANA

(FERMENTATION/SYNTHESIZED ORGANIC  CHEMICALS PLANT)

-------
                                                                      C-l
            PFIZER, INC., VIGO PLANT, TERRE HAUTE, INDIANA
           (FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS PLANT)
BACKGROUND

     Located in Vigo County near Terre Haute, the Vigo Plant is a bulk
pharmceutical manufacturing facility producing antibiotics and synthetic
organic chemicals.  The Pfizer Ag Research Center is adjacent to the
manufacturing plant.  These facilities are spread over a land area of
810 hectares (2,000 acres) of which 65 hectares (160 acres) are used for
waste treatment and up to 320 hectares (800 acres) are cropland poten-
tially available for sludge disposal.  Current employment is about 550
persons.  The pharmaceutical plant operates essentially continuously.

     As best determined, the Vigo Plant manufactures four different anti-
biotics by fermentation processes.  The 1971 discharge permit application
lists Streptomycin and Terramycin as two of these.  The other two are con-
sidered by the Company to be confidential information.  The antibiotics
are manufactured to a nearly finished state with final purification and
refining conducted elsewhere, presumably at Pfizer plants in Groton,
Connecticut and/or Brooklyn, New York.  Penicillin is also handled at the
Vigo Plant but is received as the highly purified compound and packaged only.

     Production rates of crude antibiotics fluctuate rather widely through-
out the year.  During 1972, the monthly production of the four antibiotics
ranged from 0 to 71 percent of the theoretical plant manufacturing capacity
(111).  One antibiotic was not manufactured for ten of the months in 1972
and production reached only 8 percent of capacity in the other two months.
For the other three antibiotics, monthly production ranged from none to
58, 71 and 38 percent, respectively, in comparison to full capacity.

     In early 1971, corn meal and soybeans were reportedly used as raw ma-
terials for fermentation.  The Company reported that during 1972 more than
72 different raw materials were employed in the production of antibiotics
and synthetic organic chemicals.

     Two major synthetic chemicals, fumaric acid, and benzoic acid and its
derivatives, (e.g. sodium benzoate), are manufactured.  Fumaric acid is
used" for food acidulation and is also widely used in the plastics industry.
Sodium benzoate enjoys very broad use as a preservative for food products.
Benzene is converted by catalytic oxidation to maleic acid which in turn
is isomerized to fumaric acid.   Toluene is air oxidized with the concurrent
use of co-napthenate catalyst to generate benzoic acid.  Untreated consti-
tuents remaining from benzoic acid processing may include benzol, diphenyl,
benzophenone and anthraquinone.

     Fumaric acid production was doubled in late 1970.  During 1972, mon-
thly fumaric acid production ranged from 8 to 90 percent of plant capacity
(111).   Manufacturing of benzoic acid and its derivatives ranged from 44
to 100 percent of capacity.

-------
C-2
     The Pfizer Ag Research Center represents peripheral  but important ac-
tivities.  Research and development are carried out on improving food sup-
plements for animals including dogs, cattle, hogs, chickens, and turkeys
with emphasis on poultry.  Up to a few hundred animals, mostly small  spe-
cies, are maintained at any one fine at the Ag Research Center.   Studies
are primarily nutritional although medicinal evaluations  may also be  con-
ducted.  The latter may involve introduction of scours followed by ef-
forts to cure the disease.  No vaccine use is reported.  Another such
evaluation involves the dehorning of cattle by chemical or other means.
Certain animals may be sacrificed.  Animal carcasses are  generally dis-
posed of by incineration.  Radiotracer studies are conducted but only
infrequently.  Animals subjected to radiotracers may be subsequently
interred at a specially-designated burial site.  Wastewaters generated
by the Ag Center, thought to consist of laboratory wastes, floor and
equipment washings, some pan wastes and sanitary wastes,  are piped a
considerable distance to the main waste treatment works.

     Spent cooling waters and treated wastewaters are discharged to
Jordan Creek, a very small stream originating on Pfizer property.  Pfizer
effluents comprise the entire dry weather flow of the creek near its  source
(113).  Jordan Creek is a tributary of Honey Creek, another small stream
that terminates in the lower Wabash River.  Both Jordan and Honey Creeks
are classified for agricultural water use and fish propagation.

     Information used in the evaluation of the Vigo Plant was obtained from
a variety of sources.  Two papers by Company personnel were made available.
These were a paper presented by Johnson before the Water Quality Committee
of the Wabash Valley Association in November 1971 (113),  and a paper  by
Kindrick and Johnson on wastewater treatment at the Vigo Plant presented to
the EPA Effluent Standards and Water Quality Information  Advisory Com-
mittee (ESWQIAC) at Purdue University in May 1973 (111).   Waste source data
was available from the original 1971 discharge permit application and from
an April 1972 survey conducted by EPA's Evansville, Indiana, field station.
Survey results included data on the quality of Jordan Creek at the plant
boundary.  The Vigo Plant was visited by NFIC-Denver on June 18 and 20,
1974.  The Company provided follow-up information on July 10 and again
on July 19, 1974.  During June 1974, information on the plant in the  files
of the Indiana State Board of Health was compiled including monthly waste
treatment performance sheets submitted to the State by the Company.
WASTE TREATMENT AND CONTROL

Water Use

     The 1971 discharge permit application showed plant water supply aver-
aging 22,000 m /day (5.8 mgd), derived entirely from groundwater.  Water
use included 17,600 m /day (4.65 mgd) fo cooling purposes, 4,700 m /day
(1.25 mgd) foe process needs and 1,100 m /day (0.30 mgd) for boiler feed.
About 1,500 m /day (0.4 mgd) of cooling water was reused for process
water and boiler feed supply.  Current plant water use averages 24,600
m /day (6.5 mgd).

-------
                                                                     C-3


Waste Sources

     As shown in Table C-l, more than three-fourths of the cooling water
is used in the fermentation and antibiotic recovery sectors of the plant.
These same sectors, plus the organics sector, discharge most of the pro-
cess wastewaters with the antibiotic recovery sector contributing a major
portion.

                              Table C-l.
          Relative Magnitude of Cooling Water Use and Process
              Wastewater Discharges by Operational Sector
Sector
Organics
Fermentation
Recovery of Antibiotics
Boiler and Compressor Houses
Pharmaceutical Packaging
Farm, Ag R & D
Wastewater Treatment
Total
Cooling
Water Use
(Percent)
7
47
35
6
2
1
2
100
Process Sewer
Wastewaters
(Percent)
16
10
61
6
3
4
0
100
     Wastewater in the "spent cooling water" stream includes spent condenser
and jacket cooling water, boiler blowdown, and barometric condensates.  The
condensates are believed to be about 3,780 m /day (1.0 mgd) or about 20 to
25 percent of the total spent cooling water stream.  About 1,500 m /day
(0.4 mgd) of once-through jacket water are reused in vacuum barometric con-
densers.  The Company indicated that the following seven barometric conden-
sers are present:  a three-stage unit rated at 2,460 1/min (650 gpm) and
operated 10 percent of the time, a single-stage unit rated at a few hundred
gpm and operated 45 to 50 percent of the time, a second similar single-stage
unit operated 80 percent of the time, three small units in the chemical re-
covery and refining sectors, and a small unit serving ion exchange operations
at the water treatment facility.  When the largest unit is operating, baro-
metric condenser flows can easily exceed 3,780 m /day (1.0 mgd).  York de-
misters are believed available on only one of the units.  The Company con-
siders entrainment control to be about normal for all units.

     The Company reports there are no central solvent recovery facilities
at the Vigo Plant.  However, aqueous solvents are extensively employed in
place of conventional hydrocarbon solvents within antibiotics purification
processes.  This should appreciably reduce the potential raw waste loads
going to treatment.   Foamovers from the various fermenters are directed to
the industrial  waste treatment plant.

-------
C-4
     The manufacturing operations produce large volumes of semi-solid
and solid waste materials.  The Company reports that about 11,470 m
(15,000 yd ) of such wastes are generated annually.   This is believed to
be equivalent to about 9 kkg (10 tons) of dried waste solids that must
be disposed of daily.  Heel and still bottom discards total about 3,800
1 (1,000 gal) weekly.  These materials are all generally disposed of by
landfill.

     Mycelium filtered out of the fermentation broths was estimated in
1972 as amounting to around 7,650 to 15,300 m  (10,000 to 20,000 yd)
annually.  This mycelium was believed to be disposed of at the Vigo
County Sanitary Landfill.  For Terramycin broths, the mycelium is fil-
tered out from the broth prior to extraction of the active ingredient.
Conversely, Streptomycin is a whole broth process wherein mycelium
separation is not absolutely necessary.  With Streptomycin, the exhaus-
ted or spent broth together with the finely-divided mycelium all end up
in the process waste treatment works.  At the Vigo Plant, practically
all  spent fermentation broths and a significant portion of the mycelium
are delivered to the waste treatment facility.  The waste load to the
treatment facility was substantially increased in 1972.  The Company's
Final Waste Treatment Plans submitted to the State in 1971 indicated
that ... "A spent broth from the Streptomycin antibiotic manufacturing
process in the past (had) been dried and used for feed supplement.
(However) decreased demand for the by-product makes drying economically
unfeasible.  The Company proposes to enlarge the existing waste treatment
facilities to treat the added load."  The enlarged treatment facilities
were completed in 1972.
Waste Treatment

     History of Treatment Facilities - 1947 to 1973 -- The Vigo Plant site
was originally leased from the U. S. Government in 1947.  Until 1958 when
the facility was purchased by the Company, waste treatment was provided by
the original treatment works consisting of two clarigesters, two trickling
filters, two clarifiers and chlorination facilities.  The sizes of these
units designed to handle an influent BOD load of 680 kg (1,500 lb)/day are
shown in Table C-2.  It quickly became obvious that the existing treatment
was inadequate for the industrial waste loads generated.  From 1949 through
1958, the "excess" wastes were handled by anaerobic lagooning, spray irri-
gation of high-level BOD liquors onto nearby fields and concentration and
drying of waste liquors.

     A substantial increase in both production and waste treatment capacity
has occurred since 1958.  An aeration pond and a final polishing pond
[Table C-2] were completed in 1960.  Mechanical surface aerators were added
to the aeration pond in 1966.  A large sludge stabilization pond was added
in 1967.  No effluent is permitted from this pond.  Also, in 1967, a bio-
oxidation trickling filter tower was added, operating in series with the
original plant trickling filters.

-------
                   Table C-2
       Summary of Waste Treatment Units
Vigo Plant, Pfizer, Inc., Terre  Haute, Indiana
Treatment Unit Aerators
Clarigesters (2)
Trickling Filters (2)
Clanfiers (2)
Aeration Pond(s) six-20 HP
Final Polishing Pond
Sludge Stabilization Pond
Trickling Filter Tower
Holding Pond (Cooling Water) three-5 HP
Extended Aeration Pond five to eight-75 HP
Extended Aeration Pond five to eight-75 HP
Aerobic Sludge Digester three to five-75 HP
Earthen Settling Basins
Primary
Secondary
Chlorine Contact Tank
Volume
760 m3 (200,000 gal.)
280 m3 (75,000 gal)
28,000 m3 (7.5 million gal)

1,130 m3 (40,000 ft3)
18,900 m3 (5 million gal)
17,400 m3 (4.6 million gal)
17,400 m3 (4.6 million gal)
9,460 m3 (2.5 million gal)
260 m3(69,000 gal)
400 m3(105,000 gal)
946 m3(250,000 gal)
Dimensions Detention Time
4.1 hr
13 7 m dia x 1.1 m deep
(45 ft dia x 3.5 ft deep)
12.2 m dia (40 ft dia)
3.0 hectare (7.5 acre)
1.2 m deep (4 ft deep)
14 hectare (35 acre) 30 days
1.2m deep (4 ft deep)
16 hectare (40 acre)
12.2 m dia x 9.6 m high
(40 ft dia x 31.5 ft high)
0.3-1.2 m (1-4 ft) deep 1 day
85 m x 85 m x 3m deep 6.2 days
(280 ft x 280 ft x 10 ft deep)
85 m x 85 m x 3 m deep 6.2 days
(280 ft x 280 ft x 10 ft deep)
59 m x 84 m x 3-4 m deep
(194 ft x 274 ft x 10-13 ft deep)
1.1 hr
1.7 hr
4 hr
Remarks
Original Plant
Original Plant
Original Plant
1960 Addition
Aerators added in 1966
Divided into two equal
ponds in 1974
1960 Addition
1967 Addition
1967 Addition
1968 Addition
1969 Addition
1972 Addition
1969 Addition
1972 Addition
1972 Addition

c->
i
01

-------
C-6
     In 1968 a holding pond for retention of spent cooling waters  prior
to discharge to Jordan Creek was constructed.   This is currently the only
control provided for this waste stream.

     An extended aeration pond was installed in 1969 to provide first stage
biological treatment of process wastes prior to discharge to the original
treatment works.  An aerobic sludge digestion pond equipped with mechani-
cal surface aerators was also installed in 1969 to handle excess sludge.

     The waste treatment works handling process wastes, but exclusive of
spent cooling waters, were described in 1971 as receiving 3,520 m /day
(0.93 mgd) waste influent rated at 2,700 mg/1  BOD equivalent to a raw
waste load of some 9,500 kg (21,000 Ib) of BOD/day (113).  The treated
effluent averaged 26 mg/1 BOD which approximated 90 kg (200 Ib) of BOD/day
in the final discharges.  On the basis of process wates, BOD removal    3
through the 1971 treatment works was rated at 99.0 percent.  The 3,520 m /day
(0.93 mgd) of treated process flows were combined with 13,200 m /day (3.5 mgd)
of spent cooling waters before discharge.  Detention time in the overall
waste treatment works was reported by the Company as amounting to 60 days
which is noteworthy.

     Disposal of excess sludge presented continuous problems at the Vigo
Plant and in early 1971 it was evident the system was significantly over-
loaded with respect to sludge being generated.  The clarifiers were not
performing satisfactorily, anaerobic conditions prevailed in the 3 hectare
(7.5 acre) aeration pond, and there was significant risk that anaerobic con-
ditions and serious odors would develop around the large sludge stabili-
zation lagoon.  The sludge storage/stabilization pond plus the aerated
sludge digestion pond simply had inadequate capacity for the sludge loads
being experienced.  Accordingly, in mid-1971, two sludge conveyance and
disposal "vehicles" were pressed into service together with a heavy-duty
farm tractor.  Excess sludge was drawn directly from the clarigesters at
the treatment works.  The sludge was injected into a narrow furrow 15 to
38 cm (6 to 15 in.) below the ground surface in a procedure quite similar 3
to anhydrous ammonia injection.  From June to November 1971, over 13,200 m
(3.5 million gal.) of sludge was injected into Pfizer-owned farmlands.  Some
160 hectares (400 acres) were used for sludge disposal through 1973.  The
subsurface sludge injection program was reported carefully controlled.  In
order to protect crops, the application rates of organic and inorganic ni-
trogen combined were limited to a maximum of 140 kg (300 Ib)/ acre annually.

     In September 1971, the State gave approval for further expansion of
the waste treatment works at Vigo.  Additional units consisted of a second
extended aeration pond (presumably similar to the pond installed in 1969)
equipped with floating mechanical aerators and two auxiliary earthen set-
tling basins [Table C-2].  These relatively small basins consist of a
primary settling and grit removal cell and a secondary settling cell
through which activated sludge is recycled from the effluent side back to
the influent side of either of the extended aeration ponds.  The two extended

-------
                                                                     C-7


aeration ponds are arranged so that they may be operated either in parallel
or in series.  The expanded treatment plant was completed in 1972 with the
capacity to handle 5,700 to 7,600 m /day (1.5 to 2.0 mgd) of process flows
containing a raw waste load around 13,600 kg (30,000 Ib) of BOD/day (113).

     Current Waste Treatment PracticesThe current (mid-1974) waste treat-
ment system for the Vigo Plant consists essentially of all the treatment
units constructed through 1972 listed in Table C-2 with minor modifications.
The units are connected in a manner to provide five-stage biological treat-
ment of process wastewaters.  The stages in order of downstream progression
are an extended aeration activated sludge plant, a two-stage trickling fil-
ter plant composed of the original (pre-1947) waste treatment works and the
bio-oxidation high-rate trickling filter, an aerated stabilization pond,
and a final polishing pond.  Spent cooling water passes through a separate
holding pond, the only control measure.

     The extended aeration stage consists of the extended aeration pond con-
structed in 1969 and the second pond and two earthen settling basins com-
pleted in 1972 [Table C-2].  Process wastes initially enter the earthen pri-
mary clarifier.  Settled sludge goes either to land disposal or the aerobic
digester.

     Primary clarifier effluents enter the extended aeration activated
sludge basins which may be operated either in a parallel or series arrange-
ment.  During the field inspection of June 20, the basins were operating
in series as is generally the case.  The MLSS in the extended aeration
basins were reported usually between 2,000 and 4,500 mg/1.  Effluent from
the extended aeration basins receives settling in an earthen secondary
clarifier with an overflow rate of about 20.4 m /m  (500 gal./ft )/day.
Secondary sludge is returned to the aeration basins or taken to land disposal

     The effluent from the extended aeration stage passes to the3two clari-
gesters operated in parallel.  The overflow rate is about 19.4 m /m
(475 gal./ft )/day.  Settled sludge at about 1.5 percent solids content is
sent to land disposal, the aerobic digester, or possibly back to the exten-
ded aeration basins.

     Clarigester effluents are applied onto two standard-rate trickling
filters operating in parallel.  Hydraulic loading is around 49 m /m
(1,200 gal./ft )/day and the organic loading is about 10.7 kg/m  (670 Ib
BOD/1,000 ft ).  Effluents are applied next in series to a high-rate, plas-
tic-media, bio-oxidation tower.  Wastewaters leaving the trickling filter
units enter into a final clarifier.  Sludge is unloaded to the clari-
gesters, to the aerobic digester, or to the sludge stabilization pond.

     Final clarifier flows were previously sent to the 3.0 hectare (7.5 acre)
aerated stabilization pond.  However in early 1974 this particular pond was
divided into two equal  aerated ponds operating in series.  In June 1974,
all  six aerators were situated in the first pond and the Company was
planning to install additional aerators in the second pond as soon as
possible.

-------
C-8
     Overflow from the two aerated ponds next passes to the large final
stabilization or polishing pond.  There is no supplementary aeration in
this pond which has a detention time of 30 days or slightly greater.
Effluent from the final stabilization pond is passed through a chlorine
contact box with a rather large capacity of 946 m  (250,000 gal.) pro-
viding about 4 hr detention.  Due to the nature of the wastewater and/or
the long detention in the box, no chlorine residual is observed in the
final effluent.  This effluent (Outfall 001) intercepts Jordan Creek im-
mediately below the spillway overflow of the spent cooling water retention
pond.

     The aerobic sludge digestion unit is a pond with a sludge holding ca-
pacity of 9,460 m  (2.5 million gal.) and is equipped with mechanical float-
ing aerators for relatively heavy aeration.  Sludge at 3-5 percent solids
is removed from this unit for disposal into the sludge stabilization pond.
Sludge removal rates are 300 m /day (80,000 gpd) if the solids content
is 3 percent but less at higher solids content.  Sludge holding and dis-
posal facilities are much more critical over the colder winter periods.

     The sludge stabilization pond receives a relatively low loading of
mostly stabilized sludges and no overflow is reported.  Evaporation is
slightly in excess of rainfall.  Sludge banks are slightly exposed at the
center of the pond and the remainder of the sludge deposits are submerged
under the overlying waters or supernatants.  No odors are reported what-
soever from this operation by the Company and none were observed during
the plant visit.

     As indicated previously, disposal of excess sludge by sub-surface
injection into cropland was begun in 1971.  In addition to this practice,
sludge is now being applied to growing crops in side-dressing fashion.
Up to 320 hectares (800 acres) of Company-owned cropland are available
for this purpose with about half in actual use through 1973.  It is re-
ported that every effort is made to minimize runoff from the fields re-
ceiving sludge dosing.  Pfizer is applying sludge to the soil in the
wintertime except for about 60 days when the weather is most severe.
Crops include winter wheat, corn and soybeans.  The fields are apparently
leased to local farmers who harvest the crops.  About 64 mm (0.25 in.)    3
of sludge is deposited during each application which is equivalent to 38 m
(10,000 gal.)/acre/application.  Since the water table is about 46 to 55 m
(150 to 180 ft) below ground surface, the Company feels there is little or no
hazard to ground water.  Quality of groundwaters in the sludge application
sectors is being monitored, especially for nitrate and phosphate levels
(113).   Unfortunately, little or no data is apparently collected by the
Company on survival of pathogenic microorganisms possibly associated with
sludge injection practices.  Adequate data would be welcome on this subject.
Pfizer reports no prevailing odors associated with these sludge disposal
practices.

-------
                                                                      C-9
     The cooling water impoundment pond provides about one-day detention
for  spent cooling waters and possibly some surface runoff emanating from
the  Vigo Plant.  No treatment is provided for the mixed cooling waters
before release to the pond.  The pond is equipped with floating mechanical
aerators.  This pond was heavily laden with algae during the plant visit in
mid-June 1974.  The Pfizer waste treatment system is somewhat unusual in
that overall process wastewater detention time in the system, discounting
cooling waters, is in the range of 45 to 60 days.  In the power cost area,
the  various aerators in the system add up to a total of 1,100 to 1,200 HP.
Together with the other electrical-consuming equipment in the network,
Pfizer personnel in June 1974 estimated a total waste treatment power need
of perhaps 1,400 to 1,500 HP.  These figures are not completely verified and
may  be on the high side.  Pfizer could not conveniently break out power
costs attributable solely to treatment but did indicate that 1,500 HP would
be equivalent to around $8,000 per month or $80,000 to $90,000 annually for
electrical power costs.
WASTE LOADS

NPDES Permit Application

     The original NPDES permit application submitted in 1971 summarized
average conditions for 1970-71 and the combined cooling water and treated
process waste effluents [Table C-3].

                              Table C-3.
                        Permit Application Data
Parameter
BOD
COD
TOC
Total Solids
TSS
Ammonia-N
Organic-N
Kjeldahl-N
Sod i urn
Phosphates as P
Sulphates
Iron
Oil and Grease
Col i forms
Trace Metals
Phenol ics
Flow
Concentration
(mg/D
14
80
18
932
40
33
24
56
164
19
146
2
0
Reasonably
Little data
0
i/

(kg/day)
280
1,600
360
18,500
770
680
490
1,100
3,270
380
2,900
40
_
high levels
given


Lead
(Ib/day)
610
3,500
790
40,800
1,700
1,450
1,070
2,450
7,200
830
6,400
88
_


_

a/ Flow given as 19,900 m /day (5.25 mgd) for the combined effluent in-
   cluding 4,900 to 5,200 m /day (1.3 to 1.5 mgd) of process wastewaters,

-------
 C-10
1972 EPA Survey Results

     A survey of wastewater effluents and receiving waters in and around
the Vigo Plant property was conducted in April 1972 by EPA's Evansville,
Indiana, field station.  Summary results are given in Table C-4 for the
treated process wastewater effluent and for Jordan Creek downstream of
both the process wastewater and spent cooling water discharges.  Flow
rates for the process and cooling water effluents during the survey were
estimated at 4,200 and 15,500 m /day (1.1 and 4.1 mgd), respectively.
                               Table C-4
                  Summary of 1972 EPA Survey Results
Parameter
Treated Process Effluent
         (mg/1)
Jordan Creek at Northerly
Plant Property Line(mg/l)
BOD
COD
TOC a/
Ammonia-N-' ,
Kjeldahl-N-' ..
Phosphorus, total-
Arsenic
Cadmium
Chromium, total
Copper
Fluoride
Iron
Lead
Manganese
Mercury
Nickel
Zinc
Fecal Col i forms/1 00ml.
Color (units)
30 to 40
440
145
145
185
45
0.01
0.002
0.03
0.03
0.20
2.60
0.02
0.25
0.0017
0.10
0.06
690
60
10 to 12
80
33
20
28
6.8
-
-
-
-
0.22
-
-
-
-
-
_
20
-
a/ Ammonia and Kjeldahl-N content of raw process waste respectively
   analyzed as 125  and 280 mg/1.
b/ Phosphorous content of raw process waste analyzed as 88 mg/1.
     The April 1972 survey showed the raw process wastes at that time as
having 125 mg/1 ammonia-N and 280 mg/1 Kjeldahl-N (i.e. Kjeldahl-N equals
ammonia plus organic nitrogen).   The process stream after passing through
the treatment works contained 145 mg/1 ammonia-N (a 12 percent increase),

-------
                                                                       C-ll


whereas the final Kjeldahl-N content was 185 mg/1 (a 34 percent decrease).
With many pharmaceutical plant effluents, excessive amounts of ammonia
and organic nitrogen are present even after extensive biological treat-
ment, and this is likewise the case for the Vigo Plant.  Nitrogeneous
wastes originate both from the fermentation and the synthetic organic
chemical manufacturing sectors.

     During the EPA survey, analytical problems were encountered with re-
spect to the BOD test because of the nature of the wastes.  Consequently,
estimates only were given for BOD values if no analytical control problems
had occurred, and furthermore, only possible ranges in BOD values were re-
ported.  The EPA survey showed high values for iron, phosphorus, and ammo-
nia and organic nitrogen.

Company Data

     Pfizer collects extensive data on its process wastewater stream both
before and after treatment.  In contrast, only limited data is available
on the spent cooling water stream.

     A 1973 Pfizer status report presented extensive treatment data for
the 1972 operational year and additional data for 1973(111).'  Process
wastewater flows averaged 4,430 m /day (1.17 mgd) during 1972.  Average
1972 treatment results are shown in Table C-5.
                               Table C-5
         Summary of 1972 Process Wastewater Treatment Results
Influent
Parameter (mg/1) (kg/day)
BOD
TSS
1,480
1,020
6,670
4,540
(Ib/day)
14,700
10,000
(mg/1)
27
25
Effluent
(kg/day)
120
111
(Ib/day)
264
244
Percent
Removal
98.0
97.5
     On a monthly average basis, raw process waste loads discharged to the
treatment facility varied from 910 to 9,070 kg (2,000 to 20,000 Ib) of
BOD/day and from 450 to 6,800 kg(l,000 to 15,000 Ib) of TSS/day.  A very
substantial majority of the BOD was removed in the extended aeration basins
and the trickling filter units, leaving relatively little BOD going through
the final polishing ponds.  On the other hand, low TSS levels were not at-
tained until passing through the final 3 and 14 hectare (7.5 and 35-acre)
stabilization ponds.

-------
 C-12
     Data reported for 1972-1973 on the combined treated process and spent
cooling streams are summarized in Table C-6 (111).   The data on iron, phos-
phorus and ammonia nitrogen will be discussed in more detail in a following
section.
                               Table C-6
  Residual Pollution Loads, Combined Process Plus Cooling Water Discharges
                               1972-1973

                                Concentration             	Load	
Parameter                           (mg/1)               (kg/day)(Ib/day)
Flow
Low
Normal
High
BOD
TSS
Iron
Phosphorus as P
Ammonia N


10
22
2
19
40

19,870f/
29,530^
200
438
40
380
770

c'ock/
9 . -JL /
7.80^
440
965
88
830
1,700
             3
a/  Flow in m /day
b/  Flow in mgd


     Pfizer generally analyzes treatment plant inflow and effluent on a
daily basis for BOD and solids; COD results are lacking throughout the
system.  An extended series of results on BOD and TSS concentrations and
loadings for the period January 1973 to May 1974 and covering the process
waste stream both before and after treatment, the cooling water stream,
and Jordan Creek at the Company downstream property line, is summarized
in Tables C-7 and C-8.

     The process waste system is providing better than 99 percent removal
of BOD according to recent monthly performance reports by the Company and
also 1973-74 data [Table C-7].  Biochemical oxygen demand removal over the
period of 17 months through May 1974 averaged 99.1 percent with a residual
of 73 kg (160 lb)/day BOD in the treated effluent.  Corresponding average
TSS removals for the process waste stream have been 97.8 percent yielding a
residual load of 132 kg (290 lb)/day TSS in the treated effluent [Table C-8].
The treated process stream has been generally running 10 to 20 mg/1  BOD, and
20 and 40 mg/1 total suspended solids.  The ratio of COD to BOD in the Pfizer
treated effluent appears to be in the range of 5.5 to 6.0, and COD values for
the treated process stream approximate 60 to 80 mg/1.

-------
                                                                       Table C-7
                                                            Summary of Monthly BOD  Data  for

                                              Waste Treatment Facilities,  Cooling Water  and Jordan  Creek

                                                             January 1973  through May  1974
Month
Jan. 1973
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 1974
Feb.
March
April
May
Average,
1973-74
x + 2&
Treated Flow
(m3/day)(mgd)
5,750
5,600
5,980
5,830
4,580
5,380
4,730
2,910
4,160
4,130
4,160
4,850
5,410
4,850
5,070
5,070
5,490
4,920

1.52
1.48
1.58
1.54
1.21
1.41
1.25
0.77
1.10
1.09
1.10
1.28
1.43
1.28
1.34
1.34
1.45
1.30

Process Raw
Waste
(mg/1) (kg/day )(lb/day)
1,100
1,470
1,250
1,340
2,190
2,050
1,640
900
2,200
2,340
2,100
1,700
1,570
2,190
1,820
1,770
1,950
1,740

6,270
8,300
7,300
7,440 -
8,070
10,160
7,440
2,400
9,530
9,890
9,160
9,390
8,800
8,850
8,710
7,030

8,050

13,820
18,300
16,100
16,400
17,800
22,400
16,400
5,300
21,000
21 .800
20,200
20,700
19,400
19,500
19,200
15,500
-
17,740

Treated Effluent
(mg/1) (kg/day )(lb/day)
10
10
10
15
19
29
19
17
12
20
12
9
12
13
11
16
17
15

54
54
64
82
91
159
95
41
54
77
45
50
68
64
59
82
104
73
106
120
120
140
180
200
350
210
90
120
170
100
110
150
140
130
180
230
160
233
Removal
(Percent)
99.1
99.3
99.1
98.8
98.9
98.4
99.0
97.9
99.5
99.1
99.5
99.5
99.3
99.3
99.4
98.8
-
99.1

Cooling Water^
(mg/l)(kg/day)(lb/day)
5
7
9
5
6
6
5
5
5
5
6
6
6
7
6
6
6
6

83
118
154
86
104
104
86
86
86
86
104
104
104
118
104
104
104
102

190
260
340
190
230
230
190
190
190
190
230
230
230
260
230
230
230
225

Combined . Jordan Creek
Effluents- @ Property Line
(kg/day ){lb/day)
141
172
218
168
191
263
181
127
141
163
150
154
168
181
159
181
209
175
204-227
310
380
480
370
420
580
400
280
310
360
330
340
370
400
350
400
460
385
450-500
(mg/1)
5
7
8
9
-
18
14
12
8
11
10
8
8
9
10
11
9
10

a/  Average spent cooling water flows  assumed,to be 17,000  m /day  (4.
    therefore equal  to an average of 21,960 m /day (5.8 mgd).
b/  Excludes June 1973.
5 mgd).  Combined treated process plus cooling water flows
                                                                                                                                                          o

-------
                                                                      Table  C-8
                                                            Summary of Monthly  TSS  Data  for

                                              Waste  Treatment Facilities,  Cooling Water  and  Jordan  Creek

                                                             January  1973  through May  1974
Month
Jan. 1973
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 1974
Feb.
March
April
May
Average,
1973-74
x + 2a ^
Treated
Flow
(m3/day)(mgd)
5,750
5,600
5,980
5,830
4,580
5,380
4,730
2,910
4,160
4,130
4,160
4,850
5,410
4,850
5,070
5,070
5,490
4,920

1.52
1.48
1.58
1.54
1.21
1-.41
1.25
0.77
1.10
1.09
1.10
1.28
1.43
1.28
1.34
1.34
1.45
1.30

Process Raw
Waste
(mg/l)( kg/day) (Ib/day)
760
1,080
870
900
2,760
1,360
1,450
210
1,920
2,460
1,520
1,150
1,720
1,150
1,230
1,240
1,420
1,365

4,373
6,051
5,203
5,248
12,642
.7,258
6,863
612
7,992
10,152
6,328
5,570
9,312
5,570
6,237
6,291
7,802
6,668

9,640
13,340
11,470
11,570
27,870
16,000
15,130
1,350
17,620
22,380
13,950
12,280
20,530
12,280
13,750
13,870
17,200
14,700

Treated Effluent
(mg/l)( kg/day )(lb/day)
.17
22
22
30
23
43
35
26
24
40
24
27
28
21
23
46
27
28

91
127
118
163
95
250
181
68
100
159
95
91
141
100
123
222
145
132
207
200
280
260
360
210
550
400
150
220
350
210
200
310
220
270
490
320
290
457-
Removal
(Percent)
97.9
97.9
97.7
96.9
99.2
96.6
97.4
88.9
98.8
98.4
98.5
98.4
98.5
98.2
98.0
96.5
98.1
97.8

Cooling
Water^
(mg/l)(kg/day)(lb/day)
15
10
14
12
4
4
6
23
5
9
9
15
8
19
13
33
18
13

254
172
240
204
68
68
104
390
86
154
154
254
136
322
222
563
309
218

560
380
530
450
150
150
230
860
190
340
340
560
300
710
490
1,240
680
480

Combined , Jordan Creek
Effluents- @ Property Line
(kg/day )(lb/day)
345
299
358
367
163
318
281
463
186
313
250
345
277
422
345
785
454
349

760
660
790
810
360
700
620
1,020
410
690
550
760
610
930
760
1,730
1,000
770

(mg/1)
13
14
35
21
--
37
21
32
9
16
16
32
18
21
24
46
21
24

a/  Synthesized from monthly TSS concentrations  previously,provided  by Company.
b/  Average spent cooling water flows assumed of,  17,000  m /day (4.5 mgd).  Combined
    equal to an average of 21,960 m /day (5.8 mgd).
c/  Excludes June 1973.
treated process plus cooling water flows therefore

-------
                                                                    C-15
     The cooling water stream receiving no treatment contributes  a like
organic and solids loading but in a much larger flow.   The cooling water
discharge averages 5 to 7 mg/1 BOD and generally 4 to 20 mg/1  TSS.  Over the
period of January 1973 through May 1974, the average BOD and TSS  loadings
for the cooling water stream were observed to be 102 kg (225 lb)/day and
218 kg (480 lb)/day, respectively.  Waste removals through the process
waste treatment system have been exceptionally high.  But if the  BOD and
TSS loads contained in the cooling water stream areadded overall removals
in the Pfizer system are then slightly reduced to about 98 percent for  BOD
and 95 percent for total suspended solids.

     The plant has a potential iron problem in its wastewater effluents
even though some of the local surface receiving streams are fairly high in
iron content.  Pfizer combined treated process and cooling water  flows  con-
tain around 2.0 mg/1 Fe; in June 1974 the total discharge was running about
1.7 mg/1 Fe.  Company data shows the cooling water flows contain  around
1.2 mg/1 Fe, and the treated process flows around 2.6 mg/1 for a  discharge
total load of 32 to 41 kg (70 to 90 Ib) Fe/day.  Importantly, Company re-
sults illustrate only 0.2 mg/1 Fe in the intake ground waters as  compared
to 6.0 mg/1 in the raw process wastewaters.  According to Pfizer, the in-
cremental iron contributed to final discharges mostly arises from tank  cor-
rosion although some is believed to originate from raw materials  used in
fermentation.  No iron salts are purchased by the Company.  There is no
data available on corresponding manganese levels in effluents.  In June
1974 the effluent from the treatment plant was observed to have a rela-
tively persistent yellow color.  The Vigo Plant by virtue of improved
in-plant practices is believed easily capable of reducing iron levels in
the future down to 1 mg/1, equivalent to a total discharge load of around
23 kg (50 lb)/day.

     In June and July 1974, Pfizer, Inc., provided the EPA with significant
data on nitrogenous loads in the waste treatment system and removal efficien-
cies of ammonia, organic and Kjeldahl nitrogen through the system.  During "
the plant visit of June 1974, the Company indicated that ammonia-N in the
treatment works varies from a low of 15 mg/1 in the summertime to a high of
130 mg/1 or greater during the winter months.  The Company also stated  that
just recently a concentrated ammonia waste stream amounting to 140 to 180 kg
(300 to 400 Ib) ammonia daily had been removed from the Pfizer circuit  and
was now being sold for its by-product nutrient value to a nearby  industrial
plant.  Long-term ammonia data collected by the Company during both summer
and wintertime periods showed the following values:  raw process  wastes -
75 to 235 mg/1; treated process stream - 5 to 75 mg/1; spent cooling water
stream - 0 to 6 mg/1; and Jordan Creek near Pfizer northerly property line
- still relatively high at 5 to 40 mg/1.  In June 1974, Pfizer personnel
indicated they believed that direct ammonia toxicity to fishlife  could  in-
deed become evident in the ammonia range of 2.5 to 25 mg/1.

     Pfizer iniated intensive nitrogen data collection over the period  of
May 29 through June 2, 1974, and this information shows the treatment system

-------
 C-16
 is capable, at least during late spring and early summer, of removing con-
 siderable Kjeldahl, ammonia and organic nitrogen waste loads [Table C-9].
 Again it is noted that wintertime nitrogenous levels are generally higher
 than summertime results in this situation.  Ammonia and organic nitrogen
 in the final discharges will undoubtedly continue to represent vexing
 problems for the Vigo Plant.


                              Table C-9.
            Summary of Nitrogen Data, May 29 - June 2, 1974


                  	Process Wastes           Spent Cooling Water and
 Parameter         Untreated   Treated   Percent      Treated Process Wastes
                   (mg/1)      (mg/1)   Removal              (mg/1)


 Kjeldahl-N          360          87       75                   21
Ammonia-N           165          53       67                   11
Organic-N           194          35       81                   10
Nitrate-N           1.3        10:7       -                     5
DEVELOPMENT OF NPDES PERMIT CONDITIONS

     Large amounts of unoxidized nitrogen in pharmaceutical plant wastes
create a significant number of problems which have been elaborated upon
elsewhere in this report.  Many of these problems are quite serious.  Best
practicable technology for removal of nitrogen loads in the Pfizer process
waste stream has been established as capable of yielding 20 mg/1 ammonia-N,
which equates to 95 kg (210 lb)/day ammonia.  This load has been established
in the draft NPDES permit as the average daily limit to be attained'by
Pfizer in 1976-7.  The 20 mg/1 ammonia limit is less stringent than the
ammonia nitrogen levels which may be expected in the effluents from a muni-
cipal plant having reasonably good "secondary treatment" methods available.
Pfizer reports they are capable in the summertime of achieving 15-20 mg/1
ammonia levels, but wintertime values are much higher.   In-plant modifi-
cations, recovery and/or removal, or the treatment of high-strength ammonia
wastes are determined as feasible at Pfizer.

     Company personnel in June 1974 indicated their treatment works is ca-
pable of removing about 50 percent of the incoming phosphorous raw waste
loads.  No limits have been set on the draft NPDES permit regarding phos-
phorous since significant phosphorous reductions are anticipated commen-
surate with future nitrogen removals.

-------
                                                                    C-17
     Fecal coliform data by the Company covering the period January 1973
through July 1974 show a maximum value of 490 organisms/100 ml.   Because
of the nature of wastes involved, the draft NPDES permit has incorporated
monitoring provisions and a limit of 200 organisms/100 ml for fecal coli-
form bacteria.

     Pfizer has provided EPA with appreciable bio-assay data but for various
reasons this information has not been incorporated herein.   Many of the
pharmaceutical permits, including that for Pfizer's Vigo Plant specify that
bioassay and fish survival  monitoring requirements are necessary because
inherent toxicity may be found associated with pharmaceutical wastes.

-------
                PHARMACEUTICAL INDUSTRY
                    CASE HISTORY D

            COMMERCIAL SOLVENTS CORPORATION
                 TERRE HAUTE,' INDIANA

(FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS PRODUCTION)

-------
                                                                     D-'l
         COMMERCIAL SOLVENTS CORPORATION, TERRE HAUTE, INDIANA
        (FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS PRODUCTION)
BACKGROUND

     The Commercial Solvents Corporation (CSC) plant at Terre Haute,
Indiana, is a bulk pharmaceutical chemical and organic chemical  manu-
facturing installation that produces a wide variety of agricultural,
industrial, animal health and human health (final) products.  Fermenta-
tion and a great diversity of organic chemical synthesis operations are
conducted.  Major products as of 1974 include monosodium glutamate, the
methyl amines and various methyl amine derivatives, alcohols, chemical
intermediates, choline chloride (an important animal feed supplement),
surface active agents, waxes, enzymes, starting materials for bacterio-
stat and pesticide-type compounds, and a spectrum of "Baciferm" and
"Bacitracin" products.  Some fermentation is conducted for Eli Lilly and
Company:  The plant was once a major manufacturer of riboflavin (Vitamin
B^) but, due to economic reasons, these particular operations were
discontinued about 6-7 years ago.  Due to shortages of hydrocarbon
feedstocks, the CSC plant seemed to be operating below rated capacity in
mid-1974 in many manufacturing sectors, especially the methylamine
synthesis and derived products.

     Products associated with the CSC, Terre Haute facility are as
follows:

monomethylamine                    monosodium glutamate
dimethylamine                      "Ralgro(P-1492)" (animal growth factor)
trimethylamine                     "Baciferm Soluble 50" (animal medicinal)
butyl lactate                      "Baciferm" (antibiotic/feed supplement)
dibutyl phthalate                  choline chloride (animal feed supplement)
tributyl phosphate                 "Bacitracin" USP (antibiotic)
butanol                            "Zinc Bacitracin" USP (antibiotic)
ethyl alcohol                      "Bacitracin X-l Concentrate"
isopropyl alcohol                  "Regular Bacitracin"
"Cycloserine"                      "Tris Amino" (wide variety of pharma-
"NMPD" (intermediate)                 ceutical uses)
"AMPD" (intermediate)              "Bioban(P-1487)M (bacteriostat, pesticide,)
"NMP" (intermediate)               Hydroxyethyltrimethyl ammonium bicar-
"NEPD" (intermediate)                 bonate (intermediate, alkaline catalyst)
"Alkaterge-C" (surface             Oxazoline waxes (various forms)
   active agent)                   "Adamad Catalyst 20" (acid catalyst)
"Alkaterge-T" (surface             Various enzymes
   active agent)                   "Tylosin" antibiotics (for Eli  Lilly & Co.)
Various nitroparaffinic compounds

-------
D-2


     The plant is an old installation dating back to around the 1880's.
However, extensive refurbishing and modernization have taken place in
recent years.  The Terre Haute plant operates continuously 24 hours per
day, 7 days a week with around 600 employees.

WASTE SOURCES

     The CSC plant currently has three main waste discharges to the
Wabash River.  These are Outfalls 001, 002 and 003 (numbered progres-
sively in a northerly or upstream direction) located on the east (main
manufacturing plant) side of the River.  Three pipelines under the River
convey additional wastes from the plant to an anaerobic lagoon system
and a spray irrigation system on the west side.  Until recently, the
lagoon system effluent was discharged to the River through Outfall 004
on the west side of the River.  In about January 1974, however, wastes
formerly going to the lagoon system were diverted to the City of Terre
Haute municipal sewerage system eliminating discharge 004.  There is no
surface discharge to the River from the spray irrigation system.

     The philosophy of the Company has been to collect the majority of
strong wastewaters for appropriate treatment and disposal.  Remaining
contaminated process streams have been merged with large amounts of
cooling waters and discharged through Outfalls 001, 002 and 003, pre-
sumably because of the large expense that would be otherwise involved in
segregating and transporting these wastes to available treatment.  There
is little treatment on these three lines and the volume ratio of cooling
water to process wastes is roughly estimated to be in the order of 8 to
1.  It is believed that the plant has reasonably high capability in
shifting waste sources from one collection system to another.  Given
below is a comprehensive tabulation of waste sources currently contri-
buting to Outfalls 001, 002 and 003, the waste stream previously known
as Outfall 004 but now going to the city, and the concentrated waste
stream directed to the spray irrigation fields.

Wastewaters Discharged to Outfall 001

          Bottom ash sluice streams when coal is burned in the power house
          boilers.  The boilers may be fired by gas, oil or coal fuel
          supply.  The degree of coal burning will depend upon the relative
          abundance of each type of fuel.
          Regenerants and backwashes from the hot lime-zeolite water
          softening operations.
          Floor washings from the steam power plant, and from the choline
          chloride feed supplement preparation building; also boiler
          blowdowns from the steam power plant.
          Regenerants from the MSG ion exchange columns and some possible
          loss of bonechar from the MSG char column(s).
          Miscellaneous streams including undesignated tank washings,
          air compressor coolant flows, cooling waters from the micro-
          biological sector, and land surface runoff.

-------
                                                                     D-3


     Wastewaters are retained for a short time period within a 0.4 to
0.6 hectare (1.0 to 1.5-acre) shallow settling pond built into and
alongside a coal storage pile.  The pond overflow is Outfall 001.
Average wastewater flow was reported as 4,160 m /day (1.1 mgd) b^y the
Company in their 1971 discharge permit application, and 2,740 m /day
(0.725 mgd) during a July 1973 EPA sampling survey of CSC outfalls.

Wastewaters Discharged to Outfall 002

          Overflows from the fermentation pots capturing foamovers from
          the CSC fermenters.
          Air vent discharges from fermenters which are directly exhausted
          into the plant sewer leading to Outfall 002.
          Floor washings from the warehouse area, from MSG manufacturing,
          and from other undefined process sectors.
          Barometric condenser waters from MSG manufacturing and the
          fermentation product sectors.
          Spent cooling waters from operable fermenters and spent
          compressor coolant flows.
          Unspecified wastes from Baciferm production.
          Roof drainage and land surface runoff.

 3   Average wastewater flow for Outfall 002 was reported as 9,080
m /day (2.4 mgd) in the Company's 1971 waste discharge permit appli-
cation.  The flows were 10,790 to 11,020 m /day (2.85 to 2.91 mgd)
during the July 1973 EPA field sampling study.

Wastewaters Discharged to Outfall 003

          Spent condensates from some 12 barometric condenser units.
          Floor washings from the following:  chemical pilot plant,
          "Cat 20" production area, and the choline chloride, choline
          bicarbonate, butyl lactate and "Bioban" manufacturing sectors.
          Air scrubbing effluents from Baciferm recovery operations, and
          from the methyl amines and choline chloride production sectors.
          Miscellaneous wastes including production laboratory wastes,
          spent cooling waters and surface runoff accruing from undesignated
          areas, together with drainage and runoff associated with
          tank farm sectors.

     Discharge rate for Outfall 003 was given as 12,500 m3/day (3.30
mgd) in the 1971 discharge permit application compared to measured
values of 12,600 to 13,200 m /day (3.33 to 3.48 mgd) during the July
1973 EPA field survey.

Wastewaters Previously Going to Outfall 004 but Now Directed to the
City of Terre Haute

          Air vent discharges from fermenters directly exhausted to the
          plant sewer.

-------
D-4


          Vessel washings and overflows from fermentation pots within the
          fermentation sectors.
          Washings and discards from fermentation cookers.
          Floor washings from Baciferm recovery, zinc bacitracin and
          enzymes production.
          Process waste streams from the microbiological R & D plant and
          laboratory, the chemical pilot plant, the production laboratory
          and the methyl amines production sector.
          Sanitary and miscellaneous spent waters.

                                                               3
Average wastewater flows for Outfall 004 were reported as 450 m /day
(0.12 mgd) in the 1971 discharge permit application.  This raw waste
stream "now" going to city sewers was rated in July 1974 as around 570
m /day (0.15 mgd) containing some 2,300 kg (5,000 Ib) BOD daily.

Wastewaters Discharged to the Spray Irrigation System

          Spent fermentation broths from MSG manufacturing plus other
          leftover fermentation broths.
          Series of process waste streams from Baciferm recovery, and
          the production areas for bacitracin, "Biobane," and methyl-
          amines derivatives.
          The water layer from butanol recovery.
          Various mother liquors and distillates from overall manu-
          facturing areas.
          Floor washings from nitroparaffin and butanol derivatives
          manufacturing sectors.
          Washings and spent solutions from railroad tank car cleaning.

     In early 1971, Company plans called for diverting solvent recovery
still heels to the CSC waste treatment system.  However, in June 1974
the Company indicated it had no distillation still specifically employed
for waste recovery.  However, a large butanol recovery still, when not
used for butanol, is reported as available for refining waste solvents.
It is noted that these still bottoms apparently have not yet been tied
into the treatment works.

     Continuing efforts are made to curb air pollution problems at the
CSC, Terre Haute installation.  These problems have not been completely
solved to date although appreciable air scrubbing is available.  Pre-
viously, air contaminants were cited as potentially originating from the
Baciferm-Riboflavin rotary drive driers, the MSG processing areas, the
cycloserine process, and the trimethylamines products area.  As men-
tioned above, riboflavin is no longer manufactured at this plant.

-------
                                                                     D-5


WASTE TREATMENT AND CONTROL

History and Background

     Past waste abatement objectives of the Company have been to segre-
gate most of the high-strength waste streams from dilute wastes and to
transport these concentrated materials into available waste treatment
facilities located across the Wabash River from the main plant site.
The strongest wastes are treated and disposed of via spray irrigation
over 142 to 150 hectare (350 to 370 acres) of Company lands.   An 0.8
hectare (2 acre) anaerobic pond followed by a 5.7 hectare (14 acre)
aerobic stabilization pond have been employed for handling other strong
wastes.  The spray irrigation complex was designed for complete contain-
ment of applied wastes with no runoff to receiving streams.  About 81
hectares (200 acres) of land are in current use whereas 57 hectares (150
acres) are inactive.  The Company could conceivably purchase more lands
for spray irrigation if necessary.

     Spray disposal of strong wastewaters has been decidedly advan-
tageous to CSC but difficulties have involved odors, excessive ponding,
lack of success in maintaining suitable cover crop, potential back-
flooding of the irrigation site by the Wabash River, and probable salin-
ity pickup by the irrigated soils.  The anaerobic-aerobic pond system
described above was designed for 95 percent BOD reduction and 98 percent
TSS reduction.  However, due either to the inability to reach these
performance levels, excessive odor conditions, or a combination of both
factors, use of the pond system has been discontinued.  The Company
about 8-9 months ago diverted the waste stream entering the pond system
to the City of Terre Haute municipal waste treatment facilities.  The
lagoons are now used only for storing bad batches and spills.  There
were no direct discharges from the two-lagoon system to the River during
all of 1973 and the first half of 1974, except for some unplanned re-
leases that occurred in October 1973.

     Three separate published reports describe the past history and
waste handling practices at the plant from the early 1960's through
about 1970.  The first of these reports discusses test results obtained
from a pilot-scale anaerobic lagoon receiving chemical and fermentation
process wastes.  The test criteria developed from the pilot-scale opera-
tions eventually led to design of a full-scale (20 acre) lagoon (51).
In 1968, CSC issued a report upon results obtained through the spray
irrigation land disposal system.  Performance data on the land disposal
system covered the period August 1965 until about March 1968 (11).  The
Company concluded that land application of pharmaceutical wastes, in
this case mostly derived from fermentation processes, is a viable means
of treating these low volume, high BOD spent streams.  In the last
published paper of 1971, CSC again focused upon the anaerobic lagoon
method of treating strong pharmaceutical wastewaters.  Results are given
for the period covering July 1965 through March 1971 (91).

-------
D-6


Anaerobic Lagooning. Pilot Plant Studies in the 1960's

     In the early 1960's, waste treatment was apparently being consid-
ered for the first time at Terre Haute.  Total effluents at that time
approximated 32,600 m /day (8.6 mgd), but it was found that the very
strong process wastes could be feasibly segregated from the relatively
clean cooling and associated waters.  The strongly contaminated waste
streams potentially available for treatment amounted to 566 m /day
(150,000 gpd) with a composite waste strength of about 10,000 mg/1 BOD
and in excess of 30,000 mg/1 TSS.  The BOD loading was in the area of
4,080 kg (9,000 lb)/day and the total suspended solids were roughly
around 12,700 kg(28,000 lb)/day.  Process waste streams varied in pH
from 3.5 to 10.5.  Waste loads which could not be easily segregated nor
collected amounted to about 390 kg (650 Ib) BOD/day and these were
allowed to discharge to the River.  Sanitary wastes were delivered to
the city sewer system "whenever possible" (51).

     The Company systematically evaluated a number of waste treatment
methods including activated sludge, trickling filtration, anaerobic
digestion, spray irrigation, lagooning and wet oxidation.  Biological
treatment by activated sludge or trickling filtration would provide high
BOD removals, but pretreatment of the waste streams was thought neces-
sary due to excess amounts of toxic metal ions present.  The very strong
nature of the wastes might also have caused unacceptably high unit
loadings upon these treatment processes.  Anaerobic digestion appeared
attractive except for the somewhat long retention time required.  Spray
irrigation was discounted because of the prevalence of sanitary sewage
and the extreme levels of nitrogen and toxic metal ions.  Waste lagoon-
ing was given full consideration because of simplicity, easy solids
separation, and other advantages.  An aerobic lagoon was determined as
requiring too much land.  "Satisfactory" operation of an anaerobic
lagoon was indicated probable at a unit BOD loading of around 500 kg/hectare
(450 Ib/ acre)/day, which in the case of CSC equated to approximately 8
hectares (20 acres) of land.  Wet oxidation of the waste streams at
elevated temperatures appeared meritorious, but system complexity and
high initial costs represented serious constraints.  The Company decided
to pursue pilot-plant studies on anaerobic lagooning.  Treatment objec-
tives were to reduce total waste loads to the Wabash River down to 1,900
kg (4,200 Ib) BOD/day.  CSC was contemplating future raw waste loads up
to 8,200 kg (18,000 Ib) BOD/day.  Accordingly, treatment system BOD
removal efficiencies of 60-80 percent were sought in the feasibility
studies.

     A pilot lagoon 1.2 m x 1.2 m x 1.2 m (4 ft x 4 ft x 4 ft) was
constructed for experimental treatment of strong wastes.  The BOD of the
feed was adjusted to about 10,000 mg/1 and unspecified pretreatment was
employed to convert toxic heavy metal ions to their less harmful form.
It was found necessary during the study to add sodium or ammonium ni-
trate in the feed in order to suppress formation of hydrogen sulfide
arising from high sulfates contained in the raw wastes.  Up to 500 mg/1

-------
                                                                     D-7
nitrate salts were added.  Anaerobic lagooning studies were carried  out
at ambient temperatures around 22C, but also at 15,  10,  and 5C.

     The pilot anaerobic lagoon was operated for a continuous period of
54 weeks.  For the first 16 weeks, the lagoon was maintained at ambient
temperatures of 20-25C.  The BOD results showed steadily increasing BOD
in the lagoon effluent but, near the end of the 16 weeks, equilibrium
conditions were being approached with an effluent BOD  around 1,100 to
1,200 mg/1.  The pH level held in the range of 6.8 to  7.2.   Howe et. al.
(51) state that a balanced anaerobic environment is necessary.  In the
initial phases of anaerobic stabilization, complex organic compounds are
converted to volatile acids.  Concurrently, organic nitrogen materials
must be converted to ammonia in order to neutralize the excess volatile
acids in solution.  This also prevents occurrence of drastic pH changes
in the anaerobic basin.

     Control temperatures within the anaerobic basin were then dropped
to 15C and maintained at this level from the 16th week through about
the 30th week.  The effluent BOD tended to stabilize around 1,200 mg/1,
and BOD reductions were determined to be at least 80 percent as speci-
fied under the study objectives.  Volatile acids (as acetic) and alka-
linity (as CaCOJ stabilized at around 1,000 mg/1 and  2,200 mg/1, respec-
tively.  Unfortunately, at about the 28th week, process changes in the
CSC plant caused a drastic shift in the composition of the feed to the
pilot lagoon.  The ammonium ion content dropped off appreciably and was
substituted for by up to 25,000 mg/1 sodium ions.  The balanced anaero-
bic environment was significantly disturbed from the 30th week through
the 37th week.  Volatile acids, alkalinity and BOD increased to respec-
tive peaks of 2,300 mg/1, 3,800 mg/1 and 2,500 mg/1.  Acclimation of the
system to the changed feed was noted around the 37th week, and at this
point temperature was dropped to 10C and subsequently to 5C.  At the
lower temperatures of 10C and 5C, effluent BOD increased markedly but
leveled off to around 4,000 mg/1.  This corresponded to a BOD reduction
through the system of about 60 percent based upon an incoming feed
concentration of 10,000 mg/1 BOD.  Alkalinity and volatile acid concen-
trations stabilized in the range of 3,700-4,300 mg/1 for alkalinity, and
2,200 to 2,700 mg/1 for volatile acids.  Fairly large  amounts of settleable
solids were observed during operation of the experimental lagoon.

     A full-scale 8 hectare (20 acre) anaerobic lagoon was reportedly
completed sometime around 1962-1963.  The maximum depth of this lagoon
was reportedly only 1.2 m (4 ft) which is somewhat surprising.  A small
forebay area was designed at the influent end of the lagoon specifically
intended for settling of gross solids.  Retention time of wastes in the
lagoon was calculated as 220 days.  The Company indicated that lagoon
temperatures of 5C or below were not expected over long periods.  Based
upon the pilot plant studies and full-system design, CSC personnel
predicted a somewhat optimistic 80 percent reduction of BOD.  The Company
expected lagoon effluent loads to the River of no more than 820 kg
(1,800 Ib) BOD/day (51).

-------
D-8
Fermentation Waste Disposal by Spray Irrigation Year-Round

     Woodley, chief environmental officer of CSC, provided a 1968 status
report upon the Terre Haute pollution abatement facilities with special
emphasis upon the waste spray irrigation system installed in 1965 (11).
The latter system was designed to replace, at least in part, the anaero-
bic lagoon constructed previously.  Woodley describes the overall pic-
ture at Terre Haute during 1965-1968 as follows:

     1)   Spray irrigation of fermentation wastewaters
     2)   Anaerobic lagoon system for chemical product wastewaters
     3)   Solids resulting from vacuum filters taken to sanitary landfill
     4)   Sanitary wastes discharged to city sewer system
     5)   Small pond for settling of sluiced bottom ash from the boilers
          at the power plant
     6)   Release of the remaining low concentration wastes to the
          Wabash River

     The land application system started in August 1965 was principally
intended for the treatment and disposal of low volume, high BOD fermen-
tation-derived wastes.  The spray system covered an available 150 hec-
tares (372 acres) of land, but as of March 1968 only about 65 hectares
(160 acres) had been put into actual use.  Based upon operating exper-
iences from August 1965 through March 1968, the Company contends that
satisfactory waste treatment performance can be obtained during all
weather conditions including wintertime temperatures in the minus 0C
range; furthermore continuous irrigation can be maintained even during
times of severe flooding in the Wabash River valley.  Over this two and
one-half year period, the Company estimates that 6,500 kkg (14,300,000
Ib) of raw waste BOD were "adequately-treated" and disposed of via the
spray irrigation system.  This bulk waste load^equates to an average of
6,700 kg (14,800 Ib) BOD/day which was a very sizeable increase over the
raw waste loadings of the early 1960's.  The disposal site is reasonably
well protected by levees along the Wabash River and adjoining bottom
areas.

     The spray disposal site, located across the River from the CSC
manufacturing plant, as of 1968 comprised roughly 100 active 0.4 hectare
(one-acre} spray plots, 40 inactive 0.4 hectare (one-acre) spray plots,
two 320 m  (85,000 gal.) earthen waste equalization basins, a 0.8 hec-
tare (2-acre) anaerobic lagoon, an inactive lagoon of about 4 hectares
(10 acres), and a refuse dumping area for mycelium and trash.  The 8
hectare (20-acre) lagoon excavated in 1962-1963 was either subsequently
abandoned, or more probably converted to the 0.8 hectare (2-acre) anaero-
bic pond and inactive lagoon as cited in the 1968 status report.

     Fermentation wastes going to the spray irrigation site were report-
ed as having BOD strengths as high as 64,700 mg/1.  The land was gener-
ally sprayed 1.4 hectares (3 acres) (three consecutive plots) at a time,
with any single plot sprayed continuously for 10 hours before resting.

-------
                                                                     D-9
The 10 hours of spreading at 340 1/min (90 gpm)  amounted to around 200
nr (54,000 gal.) being applied to each spray plot.   After dosing,  the
spray plots were rested two weeks or more until  the next application.
Calculations show a maximum daily application rate  onto the spray  irri-
gation fields of around 380 m /day (100,000 gpd).   However, the long-
term spray application rate probably averages closer to 230 to 260
m /day (60,000 to 70,000 gpd).  About 4.9 kg/cm  (70 psi) pressure head
was maintained both on the main header and lateral  spray lines.  Exces-
sive runoff or ponding of spray plots caused individual plots to be
taken out of service.

     The accumulated record of fermentation waste disposal via the spray
irrigation fields from August 5, 1965 through March 31, 1968 (almost 32
months) is summarized as follows:

     Flow      -    214,300m3     (56,614,000 gal.)
     BOD       -      6,471 kkg    (14,268,000 Ib)
     COD       -     14,800 kkg    (32,623,000 Ib)
     TSS       -      5,400 kkg    (11,976,000 Ib)

     Woodley comments that from the time the irrigation system was
Installed in 1965 through 1968 that the rates of waste application more
or less tripled.  Over this same period, the composite data show that
the ratio of COD to BOD for these raw wastes averaged 2.29 and the
corresponding TSS to BOD ratio was 0.84.  A summary of monthly character-
istics of raw wastes delivered to the land disposal site for the years
of 1966 and 1967 is presented below.  In reviewing these high waste
strengths it is noted that no recovery of spent fermentation broths was
practiced, by the company, as far as is known.
                               Table D-l
             Monthly characteristics of CSC, Terre Haute
              Fermentation Wastes Discharged to the Spray
         Irrigation System, January 1966 to December 1967 (11)
     Parameter                     Range               Average


  Flow (m3/day)               150  to  320               250
  Flow (gpd)               40,000 to 85,300            66,800
  BOD (mg/1)               24,100 to 39,500            30,000
  COD (mg/1)               47,000 to 90,000            67,000
  TSS (mg/1)                5,900 to 55,000            26,000

-------
 D-10
     A groundwater monitoring network was developed by CSC and was
reported as showing no subsurface contamination.  The land disposal system
operates essentially via evaporation and adsorption.  Transpiration is
considered inconsequential since vegetation does not grow on the spray
plots.  It is reported however that revegetation does occur after
suitable resting of the soils.  Recontouring of the land surface was
not recommended by the Company because of the costs involved.  A significant
number of problems were encountered in continued operation of the
spray system, but by one means or another, the Company managed to alleviate
these difficulties.  This system was considered highly satisfactory
towards solving the waste disposal needs of CSC.  The report cautions
that monitoring of groundwaters and the elimination or minimization of
surface runoff or ponding are necessary requisites for a successful
land application set-up (11).

Anaerobic Lagoon for Treatment of CSC Chemical Wastes. 1965-1971

     A further report by Woodley and Brown in 1971 more fully describes
the anaerobic lagoon system in receiving and treating the organic chemical
wastes from the CSC, Terre Haute manufacturing facility (91).  A small
anaerobic lagoon of earthen construction was available with an approximate
surface area of 0.8 hectares (2 acres) and a maximum depth of 1.5 m (5 ft).
Monthly performance data collected of this lagoon over a five and
one-half year period terminating in 1971, are shown in Table D-2:


                               Table D-2
    Treatment of Chemical Plant Wastewaters by Anaerobic Lagooning
          Commercial Solvents Corporation, Terre Haute, Ind.
                     July 1965 to March 1971 (91)
              .-""*        	     L-M-*.                  	.   ., . - - . .
     Parameter                Range                    Average


  Flow (gpd)            44,000 to 230,000               95,000
  BOD (mg/1), influent     520 to 11,900                 3,830
  BOD (mg/1), effluent     130 to 5,220                  1,240
  BOD, % Removal         .    0 to 94                        69
  Volatile acids (ma/1)-'    62 to 1,050                    510
  Alkalinity (mg/1)-'       450 to 2,270                  1,060
  Nitrogen (mg/1)-7  .      70 to 780                      220
  Phosphorous (mg/1)-'     ,  0 to 42                        20
  Hydrogen sulfide (mg/1)-   0 to 6                          4
  Sulfates (mg/1), influent 40 to 250                      105
  Sulfates (mg/1), effluent  0 to 230                       40

a?Refers to lagoon contents."~~

-------
                                                                    D-n
     Close analysis of the 1965-1971 lagoon performance data showed
that decreases in lagoon waste BOD removals were not reliably predicted
by volatile acid and alkalinity levels, and their respective ratios.
Periods of low BOD removals appeared to be more so related to low
wintertime temperatures than to any other factor(s).  Consequently, in
September 1970 the anaerobic lagoon was deepened from 1.2 to 1.5 m (4 to
5 ft).  The authors recommended that the depth of the lagoon should have
been made 3.7 to 4.6 m (12 to 15 ft) in order to take full advantage of
heating and anaerobic degradation.

     The anaerobic lagoon through the late 1960's was receiving organic
chemical processing wastes including specifically methylamines still
bottoms; alkaterge condensate; ami no butanol still bottoms; caustic
washes of fermentation equipment; floor washings; and some sewage.  The
wastes were generally basic in composition tending to keep the pH of the
lagoon contents in the 8.0 to 9.0 range.  The wastes were thought to
offer ample alkalinity for neutralization of the volatile acids formed
in the lagoon.

     Odor conditions around the anaerobic lagoon were reported of low
magnitude attributed mostly to a "relatively" low amount of sulfates in
the wastes entering the lagoon.  It was indicated that if the sulfate
content of the raw wastes could be kept below 100 mg/1, then no signifi-
cant odors would arise.  The anaerobic lagoon was operated on as low a
sulfate feed as practicable.  The Company contends that anaerobic lagoon
treatment over 1965 to 1971 at the Terre Haute location was quite succes-
sful.  Treatment efficiencies seemed to be more so controlled by tempera-
ture and pH rather than by the conventional volatile acids-alkalinity
parameters.  Efficiencies decreased markedly when the liquid wastewater
temperatures fell below 17C (63F).  The anaerobic lagoon was operated
for six years at an average BOD loading around 2,000 kg/hectare (1,785
Ib/acre) of pond surface or 0.16 kg/m  (10 lb/1,000 ft ) of pond volume
(91).

Recent Waste Handling and Abatement Procedures

     As described previously in the Waste Sources section, the plant
currently has three direct discharges to the Wabash River.  The most
concentrated wastes from the plant are disposed of by spray irrigation
on land with no surface return to the River.  Other concentrated wastes
were conveyed to an anaerobic lagoon system for treatment before dis-
charge to the River.  This latter waste stream was connected to the City
of Terre Haute municipal sewerage system in January 1974.

     The lagoon system consisted of a 0.8-hectare (2-acre) and then a 8-
hectare (20 acre) anaerobic lagoon during the 1960's.  The latter evolved
around 1971 into the 0.8 hectare (2-acre) anaerobic lagoon followed in
series by a 5.7 to 6.5 hectare (14 to 16 acre) aerobic lagoon before
final discharge to the River via Outfall 004.   The aerobic lagoon was
eventually abandoned in late 1973.   Company design criteria on the
anaerobic and aerobic lagoons as of September 1971 are presented in Table D-3,

-------
D-12
                               Table D-3

                     Lagoon System Design Criteria
     Criterion                Anaerobic Lagoon         Aerobic Lagoon
Area (hectares)
(acres)
Depth (m)
(ft)
Capacity (m )
(million gal.)
Detention
Time (days)
Impressed BOD
Load (kg/rrf)
(kg/hectare),
(lb/1,000 ftj)
(Ib/acre)
BOD Removal (percent)
TSS Removal (percent)
1.0
2.5
1.5
5
15,520
4.1

27

0.15
--
9.2

75
95
5.5
13.5
1.8
6
98,400
26.0

176

--
104
--
93
80
67
     Company plans showed an average raw waste input to the lagoon
system of 570 nT/day (0.15 mgd) at 4,000 mg/1 BOD and 3,000 mg/1 TSS
which equates to daily BOD and TSS loads of 2,270 kg (5,000 Ib) and
1,700 kg (3,750 Ib), respectively.  Waste strengths leaving the anaero-
bic cell were expected to be 1,000 mg/1 BOD and 150 mg/1 TSS.  Final
effluent (Outfall 004) pumped from the terminal pond to the River was
expected to contain in the order of 200 mg/1 BOD and 50 mg/1 TSS.
Company projections indicated that effluent loadings leaving the lagoon
system would approximate (250 Ib) BOD and (62 Ib) TSS daily.  Overall
system waste removals were predicted as 95 and 98 percent for BOD and
TSS, respectively.

     It is again mentioned that due to poor treatment performance, ex-
cessive odors, or other factors, the two pond system was essentially
taken off line at the end of 1973.  It is assumed these strong wastes
amounting to about 570 m /day (150,000 gpd) with peaks up to 680 m /day
(180,000 gpd), are now permanently connected to the city, and will not
return to the River at any time in the foreseeable future.  Apparently
the city has willingly accepted this waste stream.  Future pre-treatment
requirements that may be imposed upon the Company as a result of this
connection, are unclear as of this time.  Actually the Company has two
different tie-ins to the city, the 004 waste connection and a separate
sanitary waste connection.   Plant sanitary sewage is rated as 76 m /day
(0.02 mgd), one-half going to the city in each of the above two lines.

-------
                                                                    D-13
     The spray irrigation system was evaluated by EPA in June of 1974.
The BOD of raw wastes received into the spray disposal  system is in the
range of 30,000 to 100,000 mg/1.   The average flow rate through this
system in 1974 has been around 340 m /day (90,000 gpd), with peak
flows up to 760 rrr/day (200,000 gpd).  Raw waste strengths have
stayed roughly the same since the 1960's but flows have increased from
260 m /day (70,000 gpd) previously to 340 rT/day (90,000 gpd) now.
According to the Company, the spray system is currently operated 5  days
a week and some 4 to 5 hours daily.  CSC strives to keep the
caustic level as low as possible in the system.  Heavy rainfall through
the first half of 1974 has led to unusual amounts of standing water in
the disposal fields.  Recycle pumps have not been able to keep up
with the excess flows experienced.

     Observation of the spray fields showed virtual absence of cover
crops and apparent destruction of wooded stands.  Four sprays are
employed off a single movable header, each spray discharging over an
effective ground surface of 0.4 hectare (1 acre).  The sprays are
moved to an adjacent location when the ground becomes saturated after
a few days.  Operation and maintenance of this spray system is said
to require up to 8 persons nearly full time.  Possibilities of appreciable
odor, mosquito breeding and underground drainage to the Wabash River, if
any, should receive more detailed evaluation in the future.  Based
upon previously reported waste characterization data fgr the spray irri-
gation system and a current wastewater volume of 340 m /day (90,000 gpd),
raw waste loadings in the spray disposal network are estimated around
860 kg (19,000 Ib) BOD, 19,700 kg-(43,500 Ib) COD, and 7,200 kg (15,900
Ib) TSS daily.

WASTE LOADS

1967-1970 Waste Loads

     In April 1971 the Company reported the following raw waste BOD
loads for 1970:
     System                           	Load	
                                      (kg/day)  (Ib/day)
Spray Irrigation Disposal System
Anaerobic Lagoon
"Primary Solid Removal Facilities"
River Discharges 001, 002, 003 & 004
Total Raw Waste Load
11,300
2,300
3,200
1,900
18,700
25,000
5,000
7,000
4,300
41,300
     The Company also reported average annual final BOD loads discharged
for 1967-1970.

-------
D-14
     Year                 Maximum Monthly            Average Monthly
                        (kg/day)  (Ib/day)         (kg/day)  (Ib/day)
1967
1968
1969
1970
4,720
2,760
2,170
2,390
10,400
6,080
4,790
5,260
2,490
1,790
1,790
1,950
5,500
3,950
3,950
4,300
     From these tabulations, the Company claimed an overall 89.6 percent
BOD reduction for 1970.

1972-1973 EPA Survey Results

     Wastewater sampling surveys of the CSC plant were conducted by the
Evansville, Indiana, field station of the USEPA during May 1972 and
July 1973.  The later survey found higher waste loads at all outfalls.
In 1972 the total COD load discharged by Outfalls 001, 002 and 003 was
less than 1,180 kg  (2,600 lb)/day but the 1973 survey observed discharges
totalling about 5,500 kg (12,100 lb)/day from the three outfalls.  For
Kjeldahl nitrogen,  the May 1972 investigations gave 160 kg (360 lb)/day
for discharges 001, 002 and 003, whereas the 1973 survey showed 240 kg
(540 lb)/day.  During the second study, all three outfalls demonstrated
high fecal coliforms in excess of 1,000/100 ml.  As an average of both
surveys, total phosphorous concentrations for Outfalls 001, 002 and 003
were respectively 0.45 mg/1, 0.45 mg/1 and 0.53 mg/1, which were not
particularly high.

     The EPA report prepared in connection with the May 1972 survey pro-
vided special critique upon BOD, COD, ammonia and Kjeldahl results for
the CSC wastewaters.  The report stated that 5-day BOD was a poor
measurement parameter, especially for Outfall 003.  Where the nitrogen
content is relatively high compared to the carbonaceous demand, 5-day
BOD values may result which are actually larger than COD values.  In
effect, nitrogenous BOD may comprise a significant portion of the BOD
results obtained.   Sample dilution seems to play a critical role in
both the BOD and NOD for incubation periods of five or less days when
working with pharmaceutical waste samples.  The EPA report implied that
10 or 20-day BOD values would more accurately describe the CSC waste-
waters than would 5-day BOD.  Unknown and varying amounts of NOD could
cause calculations  on BOD removal efficiencies to be inaccurate.

     The 1972 EPA Evansville report cites an instance where the 20-day
BOD value was larger in magnitude than could be expected from the sum of
TOC plus unoxidized Kjeldahl nitrogen.  This was attributable to in-
complete digestion  of organic nitrogen compounds in the ammonia analytical
procedures.  The EPA Evansville laboratory obtained only a 40 percent

-------
                                                                    D-15
efficiency in digestion of methylamine and only a 10 percent efficiency
in digesting trimethylamine.  The report recommends, with specific
nitrogen compounds such as those manufactured by CSC, that the Company's
efficiency in organic nitrogen digestion, for purposes of running the
nitrogen analysis, should be explicitly stated.

Recent Company Waste Load Data

     Monthly report sheets on pollution abatement routinely supplied by
the Company to the State give extensive analysis of BOD loads for
Outfalls 001, 002 and 003 and some data on BOD loads to the 004 waste
handling system and to the spray irrigation disposal system.  A summary
of this monthly data is presented in Table D-4 covering the 20 month
period from October 1972 through May 1974.  The tabulation shows that
the monthly raw waste loads varied from 9,300 kg (20,500 lb)/day BOD up
to a maximum of 20,800 kg (45,800 lb)/day, averaging out at 13,900 kg
(30,700 lb)/day.  The same sheets show that the final BOD loads being
discharged to the River aggregated for the 001, 002 and 003 outfalls (on
a monthly basis) ranged from 500 to 2,200 kg (1,100 to 4,900 lb)/day,
averaging some 1,250 kg (2,760 lb)/day.  The Company for this 20 month
period reported a combined discharge to the Wabash River averaging
27,300 m /day (7.22 mgd).  The raw waste BOD load for the last 10 months
averaged 13,900 kg (30,600 lb)/day, almost the same as the 20-month
average.  Overall BOD removal efficiencies varied from 87 to 96 percent,
averaging 91.0 percent, in spite of the fact that 9 percent of the raw
waste loads received no treatment.  This data analysis shows somewhat
higher BOD percentage removals compared to the 1967-1970 situation
described previously, and also, 25-50 percent lower organic loads in the
final effluents compared to the former period.

     Information provided by Commercial Solvents Corporation in August
1974 indicates a BOD raw waste load into the spray disposal system
around 8,800 kg (19,500 lb)/day, and a raw waste load via the old 004
waste line which is now going to the city of around 4,800 kg (10,500 Ib)
BOD/day.  The first figure is believed too low and the second too high
but nevertheless the two add up to 13,600 kg (30,000 Ib) BOD/day which
is extremely close to the raw waste load estimates abstracted from the
Company's monthly report sheets.  The long-term average raw waste load
of the CSC plant has been taken as 14,100 kg (31,000 Ib) BOD/day.  The
monthly report sheets from October 1972 through May 1974 enable character-
ization of the 001, 002 and 003 discharges as given in Table D-5.  There
was essentially no discharge from Outfall 004 for this period of record.

-------
                          Table D-4

Summary of BOD Waste Load Data for Outfalls 001,  002 and 003
       Commercial  Solvents Corp.,  Terre Haute,  Indiana
                   October 1972 -  May 1974
Month
Oct. 72
Nov.
Dec.
Jan. 73
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 74
Feb.
March
April
May
Average
Flow
(m3/day)
27,400
28,810
28,810
28,770
26,570
25,510
25,740
26,800
28,280
27,250
28,010
28,770
25,170
26,500
26,500
25,590
29,530
26,310
27,900
29,530
27,330

(mgd)
7.24
7.61
7.61
7.60
7.02
6.74
6.80
7.08
7.47
7.20
7.40
7.60
6.65
7.00
7.00
6.76
7.48
6.95
7.37
7.80
7.22
Raw Waste Load
(kg/day) (Ib/day)
10,700 23,700
11,900 26,300
10,900 28,900
9,400 20,800
9,300 20,500
10,300 22,600
16,900 37,200
16,500 36,300
17,400 38,300
11,800 26,100
15,900 35,100
15,400 33,900
20,800 45,800
13,600 30,000
17,400 38,300
19,800 43,700
11,800 26,000
14,900 32,800
9,500 20,900
12,500 27,600
13,900 30,700
Final Waste Load
(kg/day) (Ib/day)
1,400 3,080
1,310 2,890
1,310 2,890
1,220 2,700
1,120 2,460
1,130 2,490
1,580 3,490
1,090 2,400
1,220 2,680
1,240 2,720
1,030 2,280
1,510 3,320
1,250 2,750
1,490 3,290
2,220 4,900
1,490 3,290
1,170 2,590
930 2,060
860 1 ,890
500 1,100
.1,250 2,760
BOD Removal
(percent)
87.0
89.0
90.0
87.0
88.0
89.0
90.6
93.4
93.0
89.6
93.5
90.2
94.0
89.0
87.2
92.5
90.0
92.0
91.0
96.0
90.6

-------
                                                                      D-17
                               Table D-5

                 Characteristics of Waste Discharges
                                 from
                       Outfalls 001, 002 and 003
                        October 1972 - May 1974
   Parameter                  Outfall 001      Outfall 002     Outfall 003
Flow 2
Average (m /day)
(mgd)
3
Range (m /day)
(mgd)
BOD
Average (mg/1)
(kg/day)
(Ib/day)
Range (mg/1)
(kg/day)
(Ib/day)

2,760
0.73
1,140-4,620
0.30-1.22

29
84
185
13-101
19-319
42-704

11,960
3.16
11,000-13,200
2.90-3.50

37
400
880
20-116
230-1,300
500-2,860

12,600
3.33
10,300-14,600
2.72-3.85

63
770
1,690
22-96
280-1,280
620-2,820
     Since the majority of the organic waste load discharged directly to
the Wabash River from the CSC Plant is associated with Outfall 003, it is
presumed that future waste abatement must focus on this particular outfall
in preference to Outfalls 001 and 002.  It is noted that the very frequent
flooding of overflow weirs on Outfalls 002 and 003 represents a continuing
problem which makes flow measurements impossible and sampling sometimes
impracticable.  It has been recommended that the Company find a solution
to this prevailing compliance monitoring difficulty.

     Partial data concerning COD loads in discharges 001, 002 and 003
over the period January through April 1974 was provided by CSC to the EPA.
The derived average CODjBOD ratio of 3.80 and an associated TSSrBOD
ratio of 0.70 were determined applicable to the waste stream being
discharged to the city and to the River.  These were compared with
previously-reported organic and solids loads discharged to the spray
disposal system.  Accordingly, total average CSC plant raw waste loads
potentially available for treatment and disposal roughly equate to
14,060 kg (31,000 Ib) BOD/day, 33,600 kg (74,000 Ib) COD/day, and 9,750
kg (21,500 Ib) TSS/day.   It is recognized that raw waste loads can vary

-------
D-18


significantly from month to month, principally depending upon the type
and number of ferrnenter tanks dumped over a specified time period, the
number of spoiled fermentation batches experienced, the degree of recovery,
if any, of spent fermentation beers, the nature and degree of solvent
losses, contamination in reactor and still bottoms released to the CSC
waste disposal systems, and various other factors.  Still bottoms are
yet reported as going to the River in undefined quantities.  Filtered-
out fermentation mycelium, generally running about 20 percent solids
content, are usually taken to landfill.  This mycelium may amount to a
few thousand pounds of BOD daily.  This mycelium has not been added into
the basic raw waste load at CSC.  Its removal is either necessary or
included under basic housekeeping procedures employed at a bulk pharma-
ceutical fermentation plant.  Future attention must be given to waste
equalization, and the elimination or removal of selected waste sources,
spills, and surges within the CSC system.

DISCUSSION OF NPDES PERMIT CONDITIONS

     Reservations have been expressed whether BOD can be properly run on
CSC wastewaters.  Consequently both BOD and COD limitations have been
employed in the permit.  Recovery of fermentation spent broths has been
taken as the "minimum" equivalent acceptable treatment for fermentation
wastes within the pharmaceutical industry, meaning a 95 percent or
better reduction of BOD.  Considering prevailing waste handling and
treatment practices, the Best Practicable Control Technology Currently
Available for the Terre Haute plant has been established as a 97.2
percent BOD reduction based upon current raw waste loads.  At Terre
Haute, it is estimated that fermentation activities roughly comprise 45
percent of the BOD raw waste loadings of 13,600 to 14,100 kg (30,000 to
31,000 Ib) BOD daily.  On the organic chemicals production side, a
required minimum of 93.5 percent BOD reduction has been called for.  The
average daily and maximum daily BOD limitations specify 820 and 1,270 kg
(1,480 and 2,800 lb)/day BOD respectively, to be achieved by 1976, as
the total allowable organic loads from all CSC outfalls.

     A COD reduction of 80 percent has been indicated in the proposed
NPDES permit which equates to 6,350 kg (14,000 lb)/day COD as the average
daily allowable value.  However, both the 1971 CSC permit application
and results of previous EPA sampling surveys show that 5,440 kg (12,000
Ib) COD/day are consistently being achieved at the present time.  The
latter figure therefore establishes the controlling 1976 COD condition
for this permit.  With the types of waste control and treatment facilities
envisioned at Terre Haute in the future, an overall 95 percent TSS
removal based upon raw waste loads is expected.  This yields an average
daily limit of 500 kg (1,100 lb)/day TSS to be reached by 1976-1977.
The limitations on BOD, COD and TSS in the permit will necessitate
additional waste removals of zero to 60 percent over the next three or
so years.

-------
                                                                    D-19
     Excessive ammonia and organic nitrogen are present in the waste-
waters from many bulk pharmaceutical manufacturing establishments.
Oxidation of ammonia to nitrate nitrogen can theoretically consume  about
4.6 parts of oxygen for each part of ammonia present in the wastewaters.
This compares to only about 1.4 parts of oxygen necessary for each  part
of BOD.  Organic nitrogen, correspondingly as for ammonia nitrogen, has
a high DO demand for conversion into nitrates.  Secondly, ammonia is
decidedly toxic to fish and aquatic life under varying stream conditions.
Thirdly, even moderate levels of unoxidized nitrogen can give distorted
results in running the BOD test on industrial plant effluents.  Addition-
ally, ammonia and other organic nitrogenous compounds during chlorination
serve to form chloramines and similar compounds which in turn are highly
toxic to fishlife, impart off-tastes to water supplies, and have other
undesirable side effects.  In a series of recent industrial cases,  not
only has the critical need been shown for ammonia and total nitrogen
reductions in wastewaters, but also available technology has been de-
scribed to bring organic nitrogen-laden wastewaters down into the range
of 10 to 20 mg/1.  Because of the prevalence of quite high levels of
ammonia and organic nitrogen in the bulk Pharmaceuticals manufacturing
industry, it has been recommended by 1976 that 25 mg/1 ammonia nitrogen
or less, shall be attained within the treated CSC process effluents.
This recommendation has been incorporated under effluent limitations for
the Pharmaceuticals industry.  For CSC, Terre Haute, average daily and
maximum daily limitations on ammonia-N have been tentatively set at 73
kg (160 lb)/day and 145 kg (320 lb)/day, respectively,  to be achieved
by 1976-1977.  EPA sampling results in 1972 showed an average total of
161 kg (355 lb)/day ammonia-N, and 39 kg (85 lb)/day organic-N in
Outfalls 001, 002 and 003.  EPA 1973 results demonstrated an average
total of 145 kg (324 lb)/day ammonia-N and 117 kg (259 lb)/day organic-N
within the three outfalls.  Required future ammonia nitrogen reductions
are in the range of 25 to 60 percent compared to present conditions.

     The 1973 EPA survey found average chromium and zinc levels in
Outfalls 001, 002 and 003 ranging up to 0.35 and 0.6 mg/1, respectively.
Sulfides, oil and grease, and phenolics were present in amounts consid-
ered sufficient to warrant effluent limits.  A sulfide limitation of 0.5
mg/1 has been established for the future.  Maximum allowable levels of
phenolics and total chromium have been set at 0.25 mg/1 for the average
daily condition.  The upper level of zinc has been set at 0.75 mg/1 and
oil/grease at 10 mg/1.  Sulfates (90-290 mg/1) and alkalinity (165-1,500
mg/1) are quite high but no industry-wide limitations have been proposed
for these parameters.

     Methanol has been tested for in the past and, in May 1972, a value
as high as 1,150 mg/1 was recorded in the anaerobic treatment pond
system.  However, amounts of methanol and of methylamine, dimethyl amine

-------
D-20
and trimethylamine were repeatedly less than 5 mg/1 both in 1972 and
1973 for Outfalls 001, 002 and 003.  No limitations have been established
in the NPDES permit for the latter parameters.

     Monitoring requirements were established for pesticides, chlorinated
hydrocarbons and toxicity.

-------
              PHARMACEUTICAL INDUSTRY
                  CASE HISTORY  E
        MERCK & CO.,  INC.,  STONEWALL  PLANT
                 ELKTON,  VIRGINIA

(FERMENTATION/SYNTHESIZED ORGANIC  CHEMICALS  PLANT)

-------
                                                                     E-l
         MERCK & CO., INC., STONEWALL PLANT, ELKTON,  VIRGINIA
          (FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS PLANT)
BACKGROUND

     The Stonewall Plant of Merck & Co., Inc., at Elkton, Va., on the
South Fork of the Shenandoah River, was started around 1941.   In the
early 1950's, operations included fermentation, processing and the
manufacture of synthesized medicinals consisting of antibiotics, vitamins
and sulfa drugs.  In the late 1950's manufactured products consisted of
fine chemicals plus fermentation-derived ingredients including Vitamin
B,, Vitamin B2, Vitamin B,2, Lysine, Streptomycin, Sulfaquinozaline,
Nicarbazin, Gfycamide and various feed supplements.  Operations were
broadly categorized as organic synthesis, fermentation, extraction and
solvent recovery (66, 93).

     Presently, the Stonewall installation is a moderate-sized bulk
pharmaceutical manufacturing facility producing both human and animal
medicinals consisting of antibiotics, vitamins, coccidiostats, steroids,
ami no acids, and feed supplements.  Major products reported by the
Company in December 1970 included three compounds essentially generated
by fermentation:  Cyanocobalamin, otherwise known as Vitamin B,^'. the
antibiotic, Streptomycin; and the amino acid, Lysine.  Other products
made primarily, if not entirely by organic synthesis reactions, included
Riboflavin, also known as Vitamin B^; Thiamine leading into Vitamin B,;
the animal coccidiostats Sulfaquinoxaline, "Nicarbazin," and "Amprolium;"
and the chemical intermediate Aminothiazole.  Chemical restructuring of
modified penicillin forms is also thought to be conducted at this
installation.  Recent data provided by the Company indicates that Stone-
wall manufactures from 12 to 24 different "bulk" chemicals and some 40
to 60 intermediate chemicals.  Approximately 30 different solvents are
used for extraction and/or overall processing.  Specific product data
could not be obtained from the Company.  The installation operates 24
hours per day, 7 days a week and current employment is between 500 and
550 persons.

     The South Fork of the Shenandoah River is associated with some of
the more demanding water quality criteria in the State of Virginia since
the waters are classified as Mountainous Zone waters and the town of
Shenandoah withdraws its public water supply downstream of Merck.  The
South Fork is classified for use as domestic water supply, agricultural
water supply, livestock watering, for propagation of fish and aquatic
life, and for boating and aesthetic enjoyment.
PROCESSES

     The Stonewall Plant is a major fermentation facility.   There are
approximately two dozen fermenters on site ranging from small through
"standard" (37,900 to 64,400 1, or 10,000 to 17,000 gal.) up to very
large in size.  The very large fermentation tanks can have  serious
impact on the waste treatment system when dumped.

-------
E-2
     In December 1970, the Company indicated that all three cited fermen-
tation products were extracted by adsorption of the active ingredient(s)
onto ion exchange resins.  Spent fermentation broths, from which the
desired products have been removed, from the Vitamin B,2, Streptomycin
and Lysine processes were all being directly released .into the liquid
waste collection system for ensuing biological  treatment.  The Vitamin
B,p production is a whole broth process with soybean meal employed as
one of the fermentation substrates.  In most fermentation plants, after
chemical recovery of the vitamins, the spent broth.is usually dried and
incorporated into animal feed supplements.  However, this is not the
case at the Stonewall Plant.  Company personnel state that making broth'
up into animal feed supplements is not economically competitive.  The
fermentation broths are relatively liquefied with low suspended solids
concentratons apparently discouraging mycelium filtration and the economical
recovery of the stripped broths.  No removal of mycelium by separate
filtration is believed practiced and, consequentially, all of this
material likely enters into the waste treatment plant.

     Streptomycin and Vitamin B,p production are similar income respects;
the fermentation broths have nearly the same color and fermentation
temperatures are similar.

     It was observed that the contents of the equalization basin within
the main treatment works were almost the same color as the broths inside
the factory, reinforcing the conclusion that spent fermentation broths
contribute heavily to raw waste loads at Stonewall.

     The Company in their December 1970 report commented upon the produc-
tion of steroids and animal feed supplements.  Steroid products are
considered a minor manufacturing line with processing consisting essen-
tially of recrystallization and isolation of desired ingredients from.
previously-purchased semi-refined stocks.  No wastes are reported produced
other than floor washes and used solvents, the latter being primarily
incinerated.  The preparation of food supplements involves operations
such as drum drying, blending, milling, screening and packaging.  Waste
streams include floor and equipment washings.

     Organic chemical products are synthesized from a wide spectrum of
inorganic and organic raw materials.  Waste streams generated from these
manufacturing sectors include mother liquors, cake washes, reactor
residues, column and extractor spent solutions, aqueous distillates,
floor and equipment washings and a wide variety of miscellaneous discard
lines.   The majority of these streams end up in the waste treatment
works.

     Solvents such as methanol, ethanol, acetone et.al., used in various
extraction and stripping operations, are mostly collected and recovered
at a centralized solvent recovery station for reuse in processing.
Other solvent recovery is exercised'at individual process stations.  The

-------
                                                                           E-3
Company could not provide the EPA with the efficiency of these operations.
Tarry still botttoms are separately collected and trucked from the plant
site, presumably to landfill.

     The Company emphasized the high variability of product mix possible
at any single time.  Large amounts of salts are apparently utilized
within processing as evidenced by high inorganic loads in the waste-
waters.

     Plant personnel remarked that new surface condensers are mainly in
use, with only two or three barometric condensers allegedly remaining
within the manufacturing plants.

     The Company reports that approximately one-half of the liquid
process wastes and raw waste loads originate from fermentation; the
other half comes from organic chemical synthesis operations.

     All intake water supply, approximately 37,900 to 41,700 m /day (10
to 11 mgd), is obtained from Company well fields; the ground water is
quite cool at about 12.8 C (55F).  Water for cooling and industrial
processing receives only chlorination.  Water for the boilers and associ-
ated needs is treated by softening, filtration and deaeration.  The
potable supply is chlorinated and filtered.

     Plant power is supplied by four boilers:  three on coal and the
fourth on No. 6 fuel oil.

     Air pollution control devices are minimal up to the present time.
However, water scrubbing is believed utilized on the spent air streams
from Vitamin B, process operations.

     The quality control lab was visited during ^larch 1974.  Colonies of
small animal species, mostly rabbits and white mice, are routinely
maintained. The animals are used for acute toxicity testing and occasion-
ally for oral toxicity testing.  A type of bacterial bioassay is conducted
which measures the rate and effectiveness of bacterial die-away when
Streptomycin is introduced into the test.


WASTE TREATMENT AND CONTROL

Early Waste Treatment

     In the early 1950's the Stonewall Plant contributed a total waste
load of 454 to 2,270 kg (1,000 to 5,000 Ib) BOD/day to the River.
Strong wastes including fermentation process spent liquors, fine chemical
mother liquors, and various washes, concentrates and still residues
having a BOD strength in the range of 3,000 to 100,000 mg/1 were evapor-
ated to a concentrated solution which was then incinerated.  "Weak"
liquid wastes were discharged to a common sewer and the combined effluents

-------
E-4
subjected to equalization, oil skimming and sludge removal within two
consecutive basins prior to being discharged to the River.  Spent cooling
waters and "jet" discharges of "negligible" BOD flowed into the second
skimming basin and then to the River.  Sanitary wastes were settled in a
clarigester, chlorinated and then discharged to the second skimming
basin.  Solid wastes and trash were collected and incinerated.

     In the early 1950's, production was rapidly expanding, and labora-
tory and pilot plant studies were conducted on biological treatment,
particularly trickling filtration, in order to provide increased flexi-
bility in waste disposal, to improve the effluent to the River, to
"decrease operating costs" for waste treatment, and to minimize the
possibility of air pollution from incineration of wastes within the
then-existing system (93).  The stated objectives were at that time "to
devise a simplified scheme of waste treatment which would require consid-
erably less supervision and would eliminate the complex system of wastes
segregation and handling."  Ironically, these objectives are now receiv-
ing serious reconsideration in the light of current regulations and
requirements.

     The initial pilot-scale trickling filter unit installed in October
1951, was studied with respect to treating fermentation wastes.  The
filter was operated for nine weeks at recirculation ratios of 12:1 and
18:1.  Strength of the feed material approximated'3,500 to 4,100 mg/1
BOD.  Average BOD Deductions were 66 to 78 percent.  The toxicity thresh-
old value of the filter effluent, determined with Daphnia magna was as
low as 0.65 percent.  Filter growths were very heavy on occasion and the
potential for ponding and odors was present.  A single trickling filter
was judged insufficient for adequate BOD removal.  Heavy filter growths
and high recycling rates commensurate with high operating costs, were
also judged unfavorable from the Company standpoint.  The decision was
made to dilute the fermentation liquors with the relatively weaker
general plant wastes and treat the mixture on two filters in series at~
comparatively low recirculation rates (93).

     Two pilot-scale filters were set up in series with intermediate and
final clarifiers treating mixed fermentation and general plant wastes.
Recycle ratios of 3:1 and 5:1 were employed around the two filters.
Average BOD Of the raw feed was in the range of 700 to 780 mg/1.
Overall BOD reductions were 84 to 87 percent, and the average effluent
toxicity thresholds were 80 percent to 100 percent.

     Next, the strong process wastes previously incinerated, having an
average BOD of 30,000 mg/1, were brought into the trickling filter pilot
plant.  Evaporation and incineration of these wastes were equated with
unacceptably high operating costs by the Company.  The mixture of spent
fermentation liquors, general plant wastes and the strong process
wastes approximated a BOD of 900 mg/1.  The two-stage filter setup was

-------
                                                                     E-5
operated for three months^at a recyclerratic of 5:1  with average BOD
loadings of 0-98 kg BOD/m^ (1.65 Ib/yd )/day on the  primary filter,  and
0.39 kg BOD/m (0.65 Ib/yd )/day on the secondary filter.  Overall  BOD
reduction averaged 84 percent with an effluent BOD of 135 mg/1.   The
toxicity threshold determined with Daphnia magna was 25 to 50 percent on
the final effluents compared to a threshold value of less than 0.1
percent on the raw mixed feed.

     The reported results on the pilot trickling filter studies  ended
with consideration of a preliminary full  scale design.   The tentative
layout comprised three trickling filters  and two clarifiers.   Two of the
filters and one of the clarifiers would constitute the  primary treatment
stage, whereas the third filter followed  by a clarifier were intended to
comprise the secondary treatment stage.  Design loading on the two
primary filters was given as 0.9 kg BOD/m  (1.5 Ib/yd )/day,  and the
estimated loading on the secondary filter was cited  as  0.8 kg BOD/m
(1.35 Ib/yd )/day.  An overalUBOD reduction of 80 percent was pro-
jected.  Disposal of some 45 m  (12,000 gal.) of sludge daily was ex-
pected to be made onto adjoining cultivated lands.

     Other Merck & Co. investigations in  the early 1950's involved
filtration and incineration studies on trickling filter sludges  and  also
anaerobic digestion and land disposal  of  these sludges.  Equipment costs
were found to be too high for filtration  and incineration.  Hydrogen
sulfide problems and other complications  were associated with anaerobic
digestion.  Summertime disposal of sludges onto tilled  soil& showed
acceptable rates of application up to 1.6 1/m (0.04  gal./ft ).  Winter-
time studies comprised surface dosing of  soils with  the sludges  for  up
to four months.   Some odor and fly problems were evident with the onset
of warm weather but were eliminated when  cultivation was resumed.   The
same surface dosing rate was judged satisfactory for the winter  months
(93).   These early practices at the Stonewall Plant  are not considered
the best for handling and treating pharmaceutical wastes in view of
current-day technology.

waste Treatment in the Late 1950's

     The waste treatment system changed appreciably  by  the late  1950's
(66).   Plant effluents essentially comprised three separate streams.
                       3
     1)   Some 26,500 m /day (7 mgd) of reportedly uncontaminated
          waters, containing spent cooling flows and unknown source con-
          tributions, which were discharged more or  less directly to
          the River.

     2)   Approximately 76 m /day (20,000 gpd) of sanitary sewage were
          collected in an individual system and received primary and
          secondary treatment.

-------
E-6
     3)   Chemical wastes made up of strong organics and spent
          fermentation broths, predominately acidic in nature,
          were directed into the main treatment wo'rks.  These strong
          chemical wastes had a volume of 4,730 to 5,680 m /day (1.25
          to 1.5 mgd) and were characterized by a BOD of 1,500 to
          1,900 mg/1; COD of 3,000 to 3,500 mg/1; TSS of 500 to '
          1,000 mg/1; and pH levels varying from 1.0 to 11.0.

     Liquid chemical wastes entered the main treatment plant through a
bar screen and mechanical grit and trash removal unit and then flowed
into an equalization basin which retained the wastes for 5 to 6 hr.
Floating immiscibles were removed by a mechanical skimmer and periodically
burned.  Effluent from the equalization basin was neutralized to around
pH 7.0, combined with the sanitary effluent which had received primary
treatment in a 6.7 m (22 ft) diameter by 5.5 m (18 ft) deep clarigester,
and sent to secondary treatment.  Settled solids in the equalization
basin were pumped to the sludge stabilization basin described below.
The digested solids from the clarigester were periodically taken to sand
filters with the filtrates returned to the clarigester.

     Secondary treatment consisted of activated sludge aeration followed
by two trickling filters in parallel.  The combined flows cited above
entered the mixed liquor activated sludge rectangular basin providing 4
hr waste retention.  This basin overflowed into a flotation chamber
where separation of solids was made.  Part of the solids was returned to
the aeration basin and the remainder pumped to an aerobic sludge holding
chamber.  The underflows from the flotation unit were released into a
wet well serving two large trickling filters in parallel.  Effluents
from the filters were passed to a splitter box which divided the flow
either for recycle or to a final settler.  A recirculation ratio of
about 7:1 was maintained.  Sludge was removed from the clarifier into a
sludge thickening tank.  Clarified effluents from the final settler were
combined with the untreated cooling water stream for final discharge to
the South Fork of the Shenandoah River (66).

     Settled solids from the chemical waste equalization basin, part of
the solids from the flotation chamber and thickened sludge from the
clarifier following the trickling filters were collected in aerobic
sludge holding chamber where air was added and sludge stabilization
effected prior to the sludge being trucked and disposed of onto Company
lands.  Capacity of the sludge stabilization,basin was reported as 908
m  (240,000 gal.), sufficient to provide about 5 days of sludge holding
and stabilization.  Apparently excess sludges were also being taken from
the flotation sludge collection box directly to land disposal.
Recent Waste Handling and Treatment Practices, 1970 through 1974

     EPA, NFIC-Denver was requested by Region III in 1974 to prepare a

-------
                                                                     E-7
NPDES discharge permit for the Stonewall Plant.  Consequently NFIC-
Denver visited the Elkton, Va., manufacturing facility on March 13,
1974, to obtain an update on processing and waste control and treatment
practices.  This same plant had been previously inspected and evaluated
by NFIC-D on July 7, 1972.

     In May of 1970 a roughing biofilter, 18 m (60 ftKin diameter by
6.6 m (21.5 ft) high and containing 1680 m  (60,000 ft ) of filtration
media, was added to the existing Merck treatment works comprising the first
stage of a three-stage bio-oxidation system ahead of the activated sludge
and the trickling filter treatment phases.  The biofilter was provided
with recycling ratio capability of up to 6:1.  The Company, in their
December 1970 written presentation to the State of Virginia, reported an
expected raw waste load of around 18,100 kg BOD (40,000 lb)/day.  The
fin-.'l effluents, including the treated industrial and sanitary wastes
and the untreated cooling waters, were reported as having the following
characteristics:

               Flow      -    30,300 to 45,400 m3/day (8 to 12 mgd)
               BOD       -    35 to 100 mg/1 and 1540 to 3060 kg
                                (3,400 to 6,750 lb)/day
               COD       -    250 to 375 mg/1
               TSS       -    40 to 100 mg/1
               pH        -    6.5 to 8.0

They indicated that their treatment plant would reduce BOD loads by at
least 90 percent.

     Except for the roughing biofilter, the Stonewall waste treatment
works has not changed significantly in recent years.

     The current waste treatment works is essentially a three-stage
biological system for handling process and sanitary wastewaters.  It is
considered somewhat marginal in view of the steadily decreasing organic
removal  efficiencies that have been demonstrated from 1972 through 1974.
The treatment of process wastes basically includes:

     A "tritor" mechanical grit and trash removal unit installed
     within a concrete flumeway.  Solids and grit are removed for
     burning.  The associated bar screen receives automatic cleaning.
     Bypassing around this unit is highly probable.

     Equalization basin, 18 m (60 ft) in diaaeter by 4.6 m (15 ft)
     deep with a holding capacity of 1,190 m  (315,000 gal.) and
     providing 3 to 6 hr detention for process wastes.  The pH in
     the equalization basin ranges from less than 1.0 to greater
     than 13.0.  The rapid variation is frequently caused by dumping
     of fermenter tanks and the large quantities of caustic or acid

-------
E-8
utilized in the fermentation purification and recovery processes.
This basin is served with a 30 HP agitator for mixing of incoming
raw wastes.

Equalized process wastes pass through a small neutralization
basin of 73m  (19,400 gal.) capacity, providing about 20 minutes
waste retention.  Acid and caustic additions are automatically
made, and the outgoing pH values are usually in the range of
6.5 to 8.5.  The neutralization box is equipped with two 7.5 HP
turbine agitators.

Neutralized wastes are lifted into a primary clarifier 18 m (60 ft)
in diameter by 2.4 m (8 ft) deep giving about 2 hr flow detention.
Floatables are skimmed off the top.  Primary sludges are sent to
the aerated sludge stabilization basin, while settled effluents'
are sent to the first stage of biological treatment.

Large roughing biofilter as described above.  Effluents are sent
to the activated sludge subsystem.

The activated sludge portion of the plant consists of activated
sludge aeration chambers, froth tanks for the collection of
excess foam off the aeration tanks, aerobic sludge digestion
or stabilization basins, and a flotation unit to separate the
activated sludge effluents from return activated sludge.  Detention
time in the primary activated sludge basins approximates 4 hr.
Excess foam ends up in the aerobic digestion basins.  The flotation
unit is reported to provide about 3-4 percent solids concentration
in the separated sludges.  Excess activated sludge is routed to the
sludge stabilization basins.  The activated sludge basins
receive oxygen via four 25 HP and four 50 HP sparged turbine
agitators.

The aerated sludge digesters receive sludges from the primary settlers,
from the secondary settlers (following, the high-rate trickling
filters), and also from the activated 'sludge flotation unit.
After some 2 to 3 days of aerobic digestion, the "digested" sludges
are generally passed through SWEECO screens, a pair of centrifuges,
and the sludge residues are then taken to land disposal.  Triton
centrifuges were installed principally to reduce the volumes of
final sludges handled and the land acreages required for solids
disposal.  Supernatants from the sludge centrifuges are returned
to the main treatment plant.

Activated sludge effluents are pumped over two 46.6 m (153-ft)
diameter high-rate trickling filters approximately 1.2 m (4-ft)
deep arranged in parallel.  The filters are capable of attaining
a recycle ratio^of 7.5 to 1 and operating at a hydraulic loading
up to 131,000 m /hectare/day (14 mgad).

-------
                                                                     E-9
     Trickling filter effluents are passed through a pair of 12 m
     (40 ft) diameter by 3 m (10 ft) deep final settlers arranged in
     parallel.  Clarifier overflows are carried underground to the small
     retention pond located near the River into which all plant waters
     combine before ultimate release to the River.  Stonewall provides no
     chlorination of final plant discharges.  The retention pond is
     skimmed when conditions dictate.  In July 1972, discharges
     from this pond to the river where observed as dark in color and turbid.

     Sanitary sewage at Stonewall approximating 75.7 m /day (20,000 gpd)
     or less is passed through a grinder and then to a two-compartment
     clarigester.  Flows from the clarigester enter the main process waste
     treatment works immediately following the neutralization basin.
     Periodically sanitary sludges are directed to "small" sand filter beds.

     Final sludges from the treatment works leaving the centrifuges
     or the aerobic sludge stabilization basins are trucked and applied
     to 47 ha (120 acres) of Company-owned lands about 0.8 km (0.5 mi)
     from the main factory.  The sludges are applied by spreading or
     irrigating.  The fields receive liming, and the sludges are periodically
     harrowed and plowed under.  Sludge application rates are thought to
     be about 100 cm (40 in.) annually and the Company reports no surface
     runoff from these fields.  There have been some indications of
     localized odors from the sludge disposal fields.

     Spent cooling, air conditioning and surface waters from the plant
grounds bypass treatment but combine with the treated process waters in
a small final retention pond just before ultimate discharge to the South 
Fork of the Shenandoah River.  Dilution is approximately 7 to 1, i.e.
volume of cooling water plus treated process stream divided by volume of
process water stream.  Past data shows 378Q to 4920 m /day (1.0 to 1.3
mgd) process wastes vs. 27,200 to 32,200 m /day (7.2 to 8.5 mgd) of
composite cooling water-surface runoff streams.  Stonewall has no cool-
ing towers to permit any significant reuse of spent cooling waters.
Opportunities for process wastes, solvents and/or other incompletely-
treated wastes bypassing treatment and intercepting the cooling and
storm drains are judged reasonably probable.

     In January 1972, the EPA requested Merck to submit a Spill Pre-
vention, Containment and Counter-Measure Plan.  The Company receives raw
materials primarily by tank car thereby necessitating a sound spill
prevention plan.  It is not known whether such plan has been yet sub-
mitted.

     Merck recently indicated to the EPA that essentially no removal of
nitrogen and phosphorous waste loads may be expected through the existing
Stonewall waste treatment plant.  As far back as May 1972, the State had
notified the Company that sometime in the future, denitrification and
phosphorous removal would likely be necessary.  Ammonolysis, amination,

-------
E-10
alkylation, diazotization, coupling of amino groups and other organic
synthesis processes conducted at Stonewall illustrate the prevailing use
of nitrogenous compounds.


WASTE LOADS

     Appropriate wastewater characterization data includes the NPDES
discharge permit application of 1971-1972; results of an EPA field
sampling survey conducted in February 1972; and BOD and COD analysis on
the waste treatment system supplied by the Company to the EPA in April
1974.  Taking the last source of information first, summaries have been
prepared for sampling results of 1972, 1973 and 1974, and these are
given respectively in Tables E-l, E-2 and E-3.  These tables provide
characterization of the raw wastes entering treatment, the spent cooling
water stream, and the combined treated process plus cooling water discharges.

     It is noted that data for each of the three years of record was
collected at somewhat different times of the year.  In spite of this,
the annual BOD input load to the treatment system averaged 18,400 kg
(40,500 lb)/day in 1972, which somewhat surprisingly decreased to 13,200
kg (29,050 lb)/day BOD in 1973, and further decreased to 10,600 kg
(23,400 lb)/day during 1974.  The COD/BOD ratio for the raw process
wastes approximated 2:1.  The Raw process flow rate correspondingly
dropped off from about 6,620 m /day (1.75 mgd) in 1972 to 4,050 m /day
(1.07 mgd) in 1974.  Since no other data of similar nature was obtained
from the Company, no other is believed available.  Based upon this data
it js presumed that the average raw waste loads have declined to around
13,200 kg (29,000 lb)BOD/day and 26,300 kg (58,000 Ib) COD/day.  The
NPDES discharge permit for Stonewall was accordingly set on these bases.
Conversely, the volume of cooling water has steadily increased from
about 22,700 m /day (6 mgd) in 1972 to approximately 32,200 m /day (8.50
mgd) in 1974.  The data array demonstrates that the Company can maintain
the cooling water stream at or below 4 mg/1 BOD and 11 mg/1 COD.  These
concentrations coupled with a 1974 average cooling water flow of 32,200
m /day (8.5 mgd), equate to daily loads of 159 kg (350 lb)/day BOD and
386 kg (850 lb)/day COD.

     The most notable feature of the performance data is that waste
removal efficiencies significantly dropped from 1972 through 1974.  The
treatment plant in early-1972 was removing some 95 percent of incoming
BOD raw waste loads and about 80 percent of the incoming COD.  These
percentage removals are quite close to the treatment levels now expected
and being written for pharmaceutical plant permits by the EPA.  Unfortunately,
treatment performance in 1974 fell to about 87 percent BOD removal and
only about 53 percent COD reduction.  From the Stonewall record, we
readily observe that the plant in 1972 with considerably higher raw
waste loads was achieving better treatment performance compared to 1973
and especially compared to 1974.

-------
                                                                       Table E-l
                                                     Waste Treatment Performance, Merck & Company
                                                 Stonewall Plant, Elkton, Va.,  Feb  28 - April 9, 1972
Equalization
Basin Effluent
(Raw Process Wastes)

Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
(mg/1)
6
2,270
5,610
6
2,790
5,580
6
2,960
5,740
6
2,480
5,510
6
2,520
5,660
6
3,620
5,310
(kg/day)
,760s-/
15
37
,670s-/
18
37
,150s-/
18
35
,660s/
16
36
,740s/
17
38
,81 0s/
24
36
1
,300
,900
1
,600
,200
1
,200
,200
1
,500
,700
1
,000
,100
1
,600
,200
(Ib/day)
.786^
33,800
83,500
.762^
41 ,000
81 ,900
.624^
40,100
77,700
.760^
36,300
80,800
.780^/
37,400
84,000
54,300
79,700
Cooling Water
(Untreated)
Discharge
(mg/1) (kg/day) (Ib/day)
WEEK 1
20,400s-/ 5.40^
4. 91 200
13 270 600
WEEK 2
12,600s-/ 5.70^/
12 270 600
36 770 1,700
WEEK 3
22,000s/ 5. SO6-/
8 180 400
25 540 1,200
WEEK 4
21,200s/ 5.6(/
5 91 200
15 317 700
WEEK 5
24,200s/ 6.40^
7 180 400
22 540 1,200
WEEK 6
28,400s/ 7.50-/
11 317 700
32 910 2,000
Total Plant
Treated Process
Discharges
(mg/1)
27
27
286
71
812
28
40
295
46
832
28
32
271
50
834
28
23
254
37
832
31
33
249
29
791
35
52
276
60
777
Final Effluent, Waste Reduction
Plus Cooling Water Efficiencies (Percent)
(Outfall 001) Treatment Overall
(kg/day)
,300s/
7
1
2
,400s/
1
8
1
23
,000s-/
7
1
23
,000s/
7
1
23
,000s-/
1
7
24
,200s-/
1
9
2
27
725
,800
,950
,210
,130
,350
,310
,600
910
,580
,410
,400
635
,120
,040
,300
,040
,700
900
,500
,810
,710
,130
,400
{Ib/day) Plant
7.20^
1
17
4
48
7.50^
2
18
2
52
7.40^
2
16
3
51
7.40^
1
15
2
51
8.2(/
2
17
1
54
930^
4
21
4
60
,600 95.9
,200 80.1
,300
.800
,500 95.4
,400 79.6
,900
,100
,000 96.0
,700 80.1
,100
,500
,400 96.7
,700 81.4
,300
,400
,300 94.9
,000 81.2
,980
,100
,000 93.9
,400 75 7
,700
,300
95.3
79.5
94.0
78.0
95.1
78.8
96.2
80.7
93 9
80.0
92.7
73.8

Flow
BOD
COD
TSS
TDS
6
,630s-/
18
36
1
,400
,900
.752^
40,500
81,300
AVERAGES
23,100s/ e.lO6-/
190 420
560 1,230
29
,500s/
1
8
1
24
,040
,030
,450
,000
7.8
-------
                                                                      Table E-2
                                                     Waste Treatment Performance, Merck & Company
                                                Stonewall  Plant,  Elkton, VA., Sept.  10 - Oct.  21,  1973

Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Equalization
Basin Effluent
(Raw Process Wastes)
(mg/1) (kg/day) (Ib/day)
1,470s-7 1 180^7
2,970 13,200 29,200
5,550 24,800 54,600
5,860s-7 1.292-7
2,850 13,900 30,700
5,570 27,200 59,900
4,700s-7 1.243^7
2,760 13,000 28,600
5,480 25,700 56,700
4,890s-7 1.293-7
2,600 12,700 28,000
5,710 27,900 61,600
4,640s-7 1 227^-7
2,760 12,800 28,200
5,790 26,800 59,200
4,250s/ 1.1 2<7
3,160 13,400 29,600
5,810 24,700 54,400
4,660s/ 1.23^-7
13,200 29,050
26,200 57,700
Cooling Water
(Untreated)
Discharges
(mg/1) (kg/day) (Ib/day)
WEEK 1
27,300s-7 7 2(7
2 45 100
6 180 400
WEEK 2
25,700s-7 6.8(7
3 91 200
9 227 500
WEEK 3
26,100s-7 6.9(7
3 91 200
8 227 500
WEEK 4
25,000s-7 6.60^/
4 91 200
1 1 270 600
WEEK 5
25,700s-7 6 8(7
3 91 200
8 227 500
WEEK 6
32,900s/ 8.7(7
3 91 200
9 320 700
AVERAGES
27,250s/ 7.2(7
91 200
240 530
Total Plant Final
Treated Process Plus C
Discharges (Outfa
(mg/1 )
31
31
228
55
790
30
42
283
58
805
30
28
267
30
883
29
30
288
30
886
30
- 34
251
39
834
37
18
233
25
860
31
(kg/day)
,800s-7 8.
998
726
1,750
25,100
,700s-7 8.
1,270
8,660
1,780
24,700
,700s/ 8.
860
8,160
920
27,100
,900s-7 7.
1,410
8,610
910
26,500
,300s-7 8
1,040
7,580
1,180
25,300
,100-7 9.
680
8,610
925
31,900
,800s-7 8.
875
8,160
1,220
26,800
Effluent, Waste Reduction
loo ling Water Efficiencies (Percent)
ill 001) Treatment
(Ib/day) Plant
2,200 92.8
16,000 71.4
3,860
55,400
2,800 91.5
19,100 68.9
3,920
54,400
lO^-7
1,900 94.1
18,000 69.1
2,030
59,700
3,100 89.6
19,000 70.1
2,000
58,400
OO*7
2,300 92.6
16,700 72.6
2,600
55,700
80^7
1,500 95.6
19,000 66.4
2,040
70,300
1,930 92.7
18,000 69.8
2,700
59,000
Overall
92.5
70.9
90.9
68.4
93.4
68.5
89.0
69.5
91.9
72.0
95.0
65.5
92.1
69.1
                                                                                                                                                               m

                                                                                                                                                               ro
a/  Flow in m /day.
b/  Flow in mgd.

-------
                                                                       Table  E-3
                                                     Waste  Treatment  Performance, Merck & Company
                                                 Stonewall  Plant,  Elkton, Va., Jan.  21 - March  3,  1974

Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
BOD
COD
TSS
TDS
Flow
COD
TSS
TDS
Equalization
Basin Effluent
(Raw Process Wastes)
(rng/1) (kg/day) (Ib/day)
S.SIO^ 0.926-/
2,048 7,170 15,800
5,074 17,800 39,200
3, 840^ l.Ol^7
2.518 9,660 21,300
5.761 22,100 48,700
4,040^ 1.1 12-'
2,551 10,700 23,600
5.894 24,800 54,600
4,040^ 1.067-'
2,647 10,700 23,500
5,691 23,000 50,600
4, 300^ K171-'
2,432 10,400 23,000
5,411 23,200 51,200
4,430^-/ 1.1 71-'
3,378 15,000 33,000
6,169 31,400 69,200
4,050^ 1 071^
23,000 50,800
Cooling Water
(Untreated)
Discharges
(mg/1) (kg/day) (Ib/day)
WEEK 1
32, 900^ 8 7
-------
E-14
     From the 1972-1974 data on TSS and IDS, it appears that solids in
the final combined effluents have remained fairly constant over this
particular period.  The slight rise in IDS loads is believed attributable
primarily to greater amounts of cooling waters used in 1974 compared to
1973 and 1972, and not to increases in synthesized organic chemicals
manufacturing.  Because of sludge disposal problems, caustic soda is
used in neutralization to minimize sludge volumes, but in turn causes
increase in IDS effluent loads.  There was no data available on TSS raw
waste loads.  However, due to the "soluble" nature of this type of
waste, TSS input loads are expected to be considerably lower than corresponding
BOD and COD.  Nevertheless the effluent TSS loads demonstrate that
improved suspended solids separation and handling facilities are necessary
at this plant.

     Information contained in the Company NPDES permit application and
results of the EPA sampling survey of February 1972 (both sources provide
only limited data), indicate that the following waste parameters are of
probable concern:  nitrogen, phosphorous, chromium, phenolics, aluminum,
lead, zinc, iron, copper, sulfides, cyanides and fecal coliforms.


DISCUSSION OF NPDES DISCHARGE PERMIT

     Comparing future required procedures at Stonewall with the overall
industry, the NPDES discharge permit has been predicated upon a 96-97
percent BOD removal for fermentation wastes and a minimum of 93 percent
reduction of BOD coming from synthesized organic chemicals manufacturing.
An 80-82 percent COD reduction has been additionally called for.  These
reductions are equivalent to loadings of 612 kg (1,350 lb)/day for BOD
and 4990 kg (11,000 lb)/day for COD on an average daily basis.  The past
data on raw waste loads are important, and any possible substantiation
of higher waste loads are important, and any possible substantiation
of higher waste loads by the Company would require full documentation.
If 1974 raw waste loads had been used as the base for effluent limitations
(which was not done) much lower allowable loads would have been obtained.
Average daily future TSS limits have been set at 590 kg (1,300 lb)/day,
but some storm run-off may need to be removed from the waste collection
system in order to reach this level.  The above load limitations have
been based upon recommended process load treatment plus cooling water
loads to be expected from the plant when exercising a high degree of
waste source control on the cooling water system.

     Tentative limits were established for chromium, phenolics, zinc,
iron, sulfides and fecal coliforms which are respectively given as 0.25
mg/1, 0.25 mg/1, 0.5 mg/1, 1.0 mg/1, 0.5 mg/1, and 200 organisms/100 ml.
Sulfide limits are imposed because of the high levels of sulfates in
pharmaceutical wastes.  These recommended limits have been calculated
upon what other industrial categories have been required to achieve with
their process waste effluents alone.  For Stonewall, the applicable flow

-------
                                                                    E-15

                                                         2
for these concentration limits has been taken as 37,800 m /day (10.0
mgd). This represents process flows plus cooling water volumes.  Additionally
the State of Virginia has imposed a temperature limit of 24C (75F)
upon the final effluents.

     Based upon nitrogen removal technology explained elsewhere in this
report and due to serious impacts of nitrogenous wastes upon water
quality, specified nitrogen removal recommendations have been made in
the NPDES permit.  The past record shows that substantial amounts of
ammonia and organic nitrogen are present in the wastewaters.  Ammonium
salts were previously reported as used in waste neutralization which
significantly increased the nitrogen content of the effluents.  Hopefully
this practice has been discontinued.  The Company permit application
reported the combined treated process plus cooling water streams as
containing 41 mg/1 ammonia N and 54 mg/1 Kjeldahl N, which are respectively
equivalent to 1540 kg/day (3,400 lb)/day and 2000 kg/day (4,400 lb)/day.
The process wastes per se are calculated to contain around 350-400 mg/1
Kjeldahl nitrogen, the large part of which is ammonia N.  Process waste
streams have been determined as amenable to nitrogen recovery, in-plant
control and/or treatment, if necessary.  The proposed NPDES permit
specified that 20-25 mg/1 ammonia N shall be attained in the future in
the treated process effluents. Phosphorous levels will be monitored in
the present permit but will ultimately require more serious attention.

-------
              PHARMACEUTICAL  INDUSTRY
                  CASE HISTORY  F

        MERCK AND CO., INC.,  CHEROKEE  PLANT
              DANVILLE, PENNSYLVANIA

(FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS  PLANT)

-------
                                                                     F-l
      MERCK AND CO., INC., CHEROKEE PLANT, DANVILLE, PENNSYLVANIA
          (FERMENTATION/SYNTHESIZED ORGANIC CHEMICALS PLANT)
BACKGROUND

     The Cherokee Plant is a multi-product bulk Pharmaceuticals manufac-
turing facility of medium size.  However, considerable production expan-
sion will be completed by 1975-1976.  Current employment is 400 to 500
persons and operations are continuous, 24 hours per day, 7 days a week.
The Danville complex, situated on the North Branch of the Susquehanna
River, was once the site of a World War II government defense installa-
tion.  Activities starting around 1944-1945 were directed to the manu-
facturing of hexamines used as raw materials in the production of ex-
plosives.  Explosives were subsequently replaced by mono-sodium gluta-
mate (MSG) and other chemicals.  In 1969, MSG was in turn phased out and
replaced by other products.  Dryden et. al. (20) reported upon early
waste handling and treatment experiences at the Cherokee Plant in the
initial years of the 1950's, and the results of their study are summar-
ized in a later section.

     In the early 1950's the plant was, and still is, a fermentation and
fine chemicals production facility.  At that time, principal products
included various antibiotics obtained by fermentation means, synthetic
vitamins and cortisone.  The first extensive waste treatment facilities
at this location consisted of neutralization, waste equalization, alum
treatment and attendant sludge handling and disposal.  It quickly became
obvious that additional treatment was necessary.  The report by Dryden
et. al. in 1956 described pilot plant investigations on the trickling
filter and activated sludge processes eventually leading to a full-scale
activated sludge treatment works in the late 1950's.

     A number of bulk antibiotics and synthesized organic compounds are
currently manufactured, plus a series of chemical intermediates.  Data
submitted by the Company in 1971 on the NPDES permit application refers
to available processes for the manufacture of various antibiotics (mainly
penicillin and its modified forms); the production of vitamins, largely
niacin (nicotinic acid); steroid chemicals; and the processing of various
industrial organic chemicals for the food and other industries, the
principal compound being sorbitol.  It is believed that the plant manu-
factures up to 30 to 40 different compounds.  Currently plant activity
is about evenly divided between fermentation and associated processing
vs chemical intermediates, organic synthesized chemicals and their
derivatives.  This ratio will substantially change in 1975-1976 to 40':60
or possibly 35:65 as fermentation to organic synthesized chemicals.   The
change in 1975 involves the startup of large-scale "Aldomet" or 1-
methyldopa production.

-------
F-2


     Of significant concern from the water pollution standpoint is the
scheduled $60 million production expansion, primarily for the manu-
facture of "Aldomet".  Additional organic waste loads plus unoxidized
nitrogen loads to the river are receiving critical evaluation.  Startup
of Aldomet is expected around September 1975 and production increases
are planned through 1976 and beyond.

     Information regarding the Cherokee Plant of Merck and Co., Inc., at
Danville, Pa. has been derived from 1)  a single literature reference
published in 1956 2)  the 1971 NPDES permit application; 3)  a plant
visit made by EPA, NFIC-D to Danville on April 1, 1974; and 4)  various
detailed followups between the EPA and the Company in order to prepare
the NPDES discharge permit for this installation.

PROCESSES

     Processing schemes at the Cherokee Plant are not precisely known,
but since sorbitol and niacin are two major organic synthesized products
made by Merck, speculation may be offered on the particular manufactur-
ing methods employed.  From the literature, sorbitol is reported as
generated from the catalytic hydrogenation of dextrose giving a 70
percent sorbitol in water solution.  Sorbitol is not only sold to food
processors but is also extensively employed as a precursor or starting
material for Vitamin C manufacturing.  In the above process, dissolved
dextrose in water aided by liberal amounts of a nickel catalyst are fed
into a continuous reactor under high pressure and heat and reacted with
hydrogen.  The sorbitol slurry is filtered and the clear sorbitol passed
through ion exchange.  Further purification is accomplished through
activated carbon and the solution is concentrated in a continuous evap-
orator.  Processing results in losses of nickel catalyst and activated
carbon plus spent acids and alkalis from resin regeneration.  There is
indication of potential "cross-over" fermentation conducted by Danville
in converting sorbitol to sorbol.  Sorbol represents a subsequent possible
by-product.

     Niacin or nicotinic acid is an anti pellagra vitamin usually pre-
pared from pyridine derivatives by processes such as the oxidation of e-picoline
with air as the oxidizer and V20g as the catalyst.  Potassium perman-
ganate with electrolytic anode oxidation has also been employed.  Nicotine
and 2-methyl-5-ethyl pyridine have likewise in the past been oxidized to
nicotinic acid.

     Aldomet is' known to come in at least two forms:  as 1-methyldopa
contained in tablets and as methyldopa hydrochloride prepared as an
ester HC1 injectable.  Aldomet is considered one of the the more important
anti-hypertensives available on the market recommended for patients with
sustained moderate to severe hypertension, blood pressure and related
ailments.

-------
                                                                     F-3
     Methyldopa is generated both in the d-and 1-isomeric forms.   Con-
version is attempted to the latter form since the activity of the com-
pound is thought to reside almost entirely in the 1-isomer.  The chemical
formula of the basic compound is C1QH,3N04.  Alpha-methyldopa is other-
wise known as a-methyl-3,4-dihydroxphenyl-alanine.  Vanillin is one of
the major base materials at the very beginning of the process.   Large
amounts of ammonia are believed employed in the main process.  The 1-a-, or
a-methyldopa, is essentially synthesized from asymmetric intermediates.
The main substrates used for the racemization and resolution processes
(separation into respective optically active components) are likely dl-ot-
amino-ot-vanillyl propionitrile and dl-a-acetamido-a-vanillyl propionitrile.
The more recent techniques involve selective crystallization procedures
in lieu of the more conventional resolution methods.  Hydrolysis of the
1-isomeric material concurrent with the newer techniques gives 1-a-
methyldopa in high yields.  It is importantly noted that ensuing waste-
waters because of liberal quantities of cyanides integrated in processing,
must be passed through a cyanide destruction procedure at high pH before
external release.  Ammonia nitrogen waste loads are also significantly
high.


WASTE TREATMENT AND CONTROL

Waste Sources

     Plant water supply, averaging 7,520 to 9,520 m /day (20 to 25 mgd),
is obtained from the Susquehanna River and passed through a preliminary
settling pond.  About 10 percent of the settled water is routed through
an "iron-manganese removal system" that uses polyelectrolyte, alum and
permanganate treatment.  Cooling water does not receive iron-manganese
treatment.  Backwash sludges and filter washings from the water treat-
ment plant are discharged to a small lagoon.

     Primary plant boilers are coal fueled.  Three of the four boilers
are equipped with mechanical dust collectors.  The sludge incinerator
uses No. 2 fuel oil and is equipped with water spray condensation of
gases and absorption via a packing column prior to' final gaseous dis-
charge.  Bottom ash slurry from the boilers and scrubber effluent from
the incinerator are discharged to the small lagoon receiving water
treatment plant wastewaters.

     Mycelium generated by fermentation activities is a major potential
source of waste loads.  Mycelium is recovered from some fermentation
broths whereas it is not from others.  Mycelium must necessarily be
removed from penicillin fermentation beers, whereas this is not absolutely
the case with many other antibiotics.  Mycelium, if not separately
filtered'out and collected within the fermentation purification sector,
becomes part of the raw waste load impressed upon the treatment plant.
Mycelium that is reclaimed at 20-25 percent solids is admixed with
partially-dewatered activated sludge for burning in a multiple-hearth

-------
F-4


incinerator.  About 45 kkg (50 tons)/day of wet mycelium is said to be
recovered at the Cherokee installation.  Company personnel  have indicated
little interest in animal feed by-product recovery from fermentation op-
erations.

     According to information recently obtained from the Company, eight
barometric condensers continue to discharge to the spent cooling water
collection system.  These units are a major source of waste loads dis-
charged to this sewer system.

     The plant has a centralized solvent recovery system.  Unrecoverable
solvents are collected at the solvent recovery station or otherwise se-
gregated at their point of use and periodically disposed of via outside
scavenger services.  In spite of overall solvent recovery at this fa-
cility, contribution of organic raw waste loads due to "lost" solvents
is thought to be fairly significant.

     WastewateK streams generated by the Cherokee Plant in mid-1974 in-
cluded 3,780 m /day (1.0 mgd) of process 'Wastes, 380 m /day (0.1 mgd) of
sanitary sewage and 71,900 to 83,300 m /day (19 to 22 mgd) of spent
cooling water, boiler blowdown and miscellaneous wastewater streams.
Sanitary sewage is discharged to the local municipal sewerage system.
Process wastes are treated as described in the following section.

     The volume of spent cooling water is surprisingly high as. is the
ratio of cooling water to process wastes which is in the range of 20 to
26:1.  The spent cooling water, 95 percent of the Cherokee Plant effluent,
is released "untreated" to the Susquehanna River.

Waste Treatment

     Early Waste TreatmentWaste Treatment facilities in the early
1950's consisted of neutralization, equalization, alum treatment, and
attendant sludge handling and disposal.  Additional treatment became
necessary and pilot plant studies were undertaken to evaluate treatment
processes and determine design criteria.

     Initially, a pilot-scale trickling filter setup was constructed to
evaluate the treatability potential of the wastewaters (20).  After a
few weeks of operation, formidable problems became evident including an
extremely prolific growth on the filters which could not be dislodged.
Correspondingly, extreme ponding and drastic reductions in waste treat-
ment efficiencies were experienced.  At this point it was decided to an-
alyze the activated sludge process as an alternate means of treatment,
and this was done in an ensuing experimental program.

     The laboratory study phase consisted of using a series of 5-liter
batch aerators, 15 cm (6 in.) in diameter by 41 cm (16 in.) tall.  Based
upon varying sludge concentrations and BOD strengths of the raw waste
feed into the aerators, the relation of BOD removal vs aeration

-------
                                                                     F-5


time was adequately determined.  From these tests, it was concluded that
the activated sludge process was adaptable for treating the industrial
wastewaters, a high degree of BOD removal could be obtained with reason-
able aeration times, and further tests should be conducted at the pilot
plant stage.

     This activated sludge pilot plant was capable of handling flow
rates up to 57 1/min (15 gpm).  A series of test runs was conducted at
various MLSS levels, feed rates and retention times.  Removals of BOD
varied from 50 to 93 percent.  These runs demonstrated that an operation-
ally stable activated sludge could be maintained using all the wastes
from the Cherokee Plant.  It was cautioned that the pH value of the raw
wastes should be kept between 6 and 8.  The investigators stated that a
wide range of treatment performance could be obtained, and also predicted,
for any given set of conditions as to suspended solids, retention times
and BOD loadings.  In disposing of excess biological sludges, filtra-
tion-incineration and anaerobic digestion were rejected as feasible
operations (although sufficient reasons were not given) whereas aerobic
digestion was thought to offer at least a partial solution to the sludge
disposal undertaking (20).

     Design criteria were necessary, and probability data were developed
for Cherokee Plant wastewaters with respect to waste volume and BOD load
and concentrations.  The 90-percent high probability values of flows and
loads were used to give treatment plant design criteria.  The design
volume of raw wastes was established as 3,200 m /day (0.84 mgd) and the
BOD raw waste load as 6,100 kg (13,500 lb)/day.  The 50 and 90 percent
probability values of raw waste BOD strengths were 1,600 mg/1 and 2,600
mg/1, respectively.  Mixed liquor suspended solids in the aeration
basins were assumed as 2,500 mg/1.  The BOD discharge to the receiving
stream was predicted to be 1,800 kg (4,000 lb)/day yielding a desired
"at that time" BOD removal of 70.5 percent via activated sludge.  At the
"average" expected raw waste loading of 3,600 kg (8,000 lb)BOD/day, a
BOD reduction of around 90 percent was predicted.  It is noted that
design retention in the aeration basin(s) was only 2.2 hr.

     From these studies the investigators "concluded that the activated
sludge process appeared superior to other treatment methods for Cherokee.
The process gave consistently "good" BOD reductions and recovered quickly
when subject to toxic substances.  Also, from the standpoint of ease and
efficiency of operation, activated sludge was deemed the best approach
for satisfactorily treating the fine chemical wastewaters from Danville
(20).

     The recommended activated sludge treatment facilities were con-
structed in the late 1950's.  Various improvements have been made since
then with significant modifications and add-ons to the system completed
around December 1972, including a large roughing filter, secondary settling,

-------
F-6


and vacuum filtration and incineration facilities for handling and
disposing of excess biological sludges.  These changes resulted in
improved waste treatment performance beginning in early 1973.   Present
waste treatment facilities are described below.

     Current Waste TreatmentAs of mid-1974, treatment of process wastes
consisted of equalization, neutralization, biological treatment including
a roughing trickling filter and activated sludge, and secondary settling.
Sludge was disposed of by dewatering and incineration.
                                                                         3
     Process wastes flow into a wet well and are then pumped into a 136 m
(36,000 gal.) equalizing tank providing about 9 hr detention.   Waste-
waters in this holding basin are kept agitated by means of a 60 HP aera-
tion support system.  Three cooling towers with a total rating of 5,450
m /day (1,000 gpm) are also available adjacent to the equalizing tank
for chilling the process wastewaters from 35 to 46C (95 to 115F) down
into the range of 29 to 35C (85 to 95F), when found necessary, especially
in the summertime.  The Company reports the latter temperature range as
being optimum for subsequent biological treatment.  Process wastes
entering the equalizing tank hae a pH range of 2 to 12.  Process flows
are currently averaging 3,780 m /day (700 gpm or 1 mgd).

     Equalized process flows enter a neutralization tank for dosing with
sulfuric acid or caustic soda.  This tank has a 85 m  (22,500 gal.)
capacity and provides about 36 min detention.

     Neutralized wastewaters are pumped to a roughing trickling filter
tower 6.6 m (21.5 ft) high by 15.2 m (50 ftkin diameter.  Impressed
flow rate on the filter is 7,630 or 15,260 m /day (1,400 or 2,800 gpm),
depending upon whether one or two pumps are used.  The recirculation
ratio around the filter is 1:1 or 3:1.

     Trickling filter effluent flows into two aeration basins in parallel
with a total capacity of 2,560 m  (675,000 gal) giving 16 hr waste
detention at average flow.  Aerators on both the activated sludge aeration
basins and the equalizing tank include twelve 30 HP sparged turbine
units and two 50 HP units for a total of 460 HP.  Specifications on the
system call for 100 percent recycle of sludge from the secondary clari-
fication tank back to the activated sludge basins.  The Company indicates
that 5,000 to 6,000 mg/1 MLSS are maintained in the aeration basins.

     The activated sludge liquors receive final clarification in a
single 15.2 m (50 ft) diameter tank having a capacity of 555 nr (146,500
gal.).  Average detention time in this unit is 1.5 hr.  Sludge underflow
concentration is reported to be roughly 1 percent solids.

     Excess biological  sludges from the secondary clarifier are sent to
a sludge thickener tank 16 m (50 ft) in diameter.  Thickener supernatants
are returned to the main treatment works.  The sludge is concentrated to

-------
                                                                     F-7
roughly 3 percent solids and then conditioned, dewatered and incinerated.
Conditioning consists of the addition of FeCU and lime.    Conditioned
sludge is passed through a pair of vacuum filters producing a sludge of
about 16 percent solids content.  Filtrates from the vacuum filters plus
filter belt wash waters are returned back to main treatment.  Dewatered
and filtered biological sludge is mixed with mycelium at about 20 percent
solids content recovered from within the fermentation operations and in-
troduced into a 7-tier multiple hearth incinerator.  Temperatures within
the incinerator vary between 650 and 930C (1200 and 1700F) depending
upon the particular hearth level.  The Company in 1972 indicated that
sludge loadings of around 41 kkg (45 ton)/day of activated sludge at 84
percent moisture content plus 46 kkg (51 ton)/day of mycelia sludge at
80 percent moisture content were being experienced.  Design specifica-
tions for the incinerator are set at 80 kkg (88 ton)/day, which means
that all sludges cannot be sufficiently handled by the incinerator and
this unit is apparently underdesigned.  Off-site disposal of sludges is
likely necessary.  The completely oxidized sludges coming out of the
incinerator in the form of a fine ash are collected and scavenged to
landfill.  The above incineration system was not completed until Febru-
ary-March of 1973.

     Water treatment plant sludges, various filter washings, scrubber
effluent from the sludge incinerator with a flow rate of about 1,500
1/min (400 gpm), and ash slurry from the main boilers are discharged to
a small lagoon located adjacent to the final discharge line of the
plant.  Overflow from the lagoon combines with spent cooling waters and
treated process wastes before discharge to the Susquehanna River through
Outfall 001.  In April 1974 it was observed that this lagoon was almost
completely clogged with solids, and little solids settling was occurring.
Later the Company reported that a second settling pond was under construc-
tion.  The State of Pennsylvania permit which represents the last author-
ization for waste discharge by the Cherokee Plant explicitely states
that ..." settled solids shall at no time be permitted to accumulate in
the sedimentation basin(s) to a depth greater than one-third that of the
basin(s) as constructed, and the settled materials removed shall be
handled and disposed of in a (satisfactory) manner ..."

     Treated process wastes leaving the final clarifier, overflows from
the small lagoon, and the relatively large volumes of spent cooling
waters, all combine at about the same point just prior to discharge to
the Susquehanna River via Outfall 001.  Although analytical data exist
for the combined discharge and for the treated process waste stream,
virtually no separate sampling information could be secured from the
Company on the small lagoon overflows and the spent cooling water stream.
Significant waste loads are believed present in the latter two flows,
however, this cannot be verified by available data.  Waste concentrations
in the process flows are about 15-20 times higher than reported for the
combined effluents.

-------
 F-8
     There  is no disinfection or equivalent on the final waste discharge.
 Fecal coliforms are reported present but in low quantities.  The Company
 indicates chlorination is not necessary because sanitary sewage in the
 last 2-3 years has been entirely diverted out of the Cherokee treatment
 works into  the local municipal sewerage system.
WASTE LOADS

Current Production Levels

     Company records indicate that the present waste treatment facility,
including improvements completed in December 1972, has the following de-
sign criteria [Table F-l].
                              Table F-l.
                                                               a/
       Summary of Current Waste Treatment Plant Design Criteria-
   Parameter
(kg/day)
(Ib/day)
    Influent Loads
    Flow
    BOD
    TSS
    NH3-N

    Effluent Loads
    BOD
    TSS
 3,030^
 8,620
 2,130
 1,130
 1,380
   610
   Removal Efficiencies

   BOD                              84 percent
   TSS                              71 percent

   Cooling Water Stream

   Intake Water BOD                590
   Incremental BOD Added in Plant  450
         Total BOD Load          1,040

   Combined Effluent Loads
b_/ Flow in m /day
c/ Flow in mgd
19,000
 4,700
 2,500
 3,050
 1,350
                         1,300
                         1.000
                         2,300
Flow
BOD
TSS
NH,-N
TDS
75, 700^
2,430
610
1,130
27,670
20. 0-/
5,350
1,350
2,500
61,000

-------
                                                                       Table F-2

                                                              Summary of Monthly BOD Data
                                                     Merck and Co., Cherokee Plant,  Danville, Pa.
                                                             December 1972 - February 1974
Process Wastes to Treatment
Month
Dec.
Jan
Feb.
Mar.
Apr.
May
June
Average
(First Half 73)
July
Aug.
Sept.
Oct.
Nov.
Dec.
Average
(Second Half 73)
Jan.
Feb.
Average, 74
Overall
Average
Flow
(m3/day)
3,260
3,520
4,050
4,010
3,630
3,970
6,400
4,280
3,480
2,910
3,560
3,520
3,220
3,260
3,330
3,900
4,320
4,130
3,780

(mgd)
0 86
0.93
1.07
1.06
0.96
1.05
1.69
1 13
0 92
0.77
0.94
0.93
0.85
0.86
0.88
1.03
1.14
1.09
1.00
BOD
(kg/day)
5,585
7,051
7,502
6,658
7,284
8,182
11,450
8,030
5,467
6,242
7,112
6,691
5,815
6,424
6,290
7,150
7,667
7,410
7,080
Load
(Ib/day)
12,312
15,545
16.539
14.678
16.058
18.039
25.244
17,700
12,052
13,760
15.678
14,750
12,820
14,163
13.870
15,763
16,903
16,330
15,600
Raw Water
Intake BOD
(kg/day)
146
202
166
216
288
232
63
195
48
223
190
124
157
148
150
102
131
120
160
(Ib/day)
1972
321
1973
445
365
475
634
511
139
430
106
492
418
273
347
326
330
1974
224
288
260
360
Combined
Flow
(m3/day)
68,500
71,500
75,300
83,300
82,900
87,100
98,400
83,300
83,300
93,100
89,700
87,800
87,800
85,900
87,800
74,200
75,700
75,000
82,900
(mgd)
18.1
18.9
19.9
22.0
21.9
23.0
26.0
22.0
22.0
24.6
23.7
23.2
23.2
22.7
23.2
19.6
20.0
19.8
21.9
Effluents
BOD
(kg/day)
1,930
2,802
2,363
1,991
2,424
1,572
1,238
2,060
937
1,119
899
1,545
1,184
1,251
1,160
1,188
1,568
1,380
1,600
BOD Removal
Load
(Ib/day)
4,255
6,177
5,210
4,390
5,343
3,466
2,729
4,550
2,066
2,466
1.982
3,406
2,611
2,759
2,550
2,619
3,457
3,040
3,530
Through Treatment
(percent)
67.9
63.1
70.8
73.3
70 5
83.6
89.7
75.2-
83.7
85.7
90.1
78.8
82.3
82.8
83.9
84.3
81.3
82.8
79.7
* Overall^
(percent)
65.9
61.3
69.4
71.0
67.9
81.3
89.2
73.4
83.0
82.7
87.7
77.3
80.2
81.0
82.0
83.5
79.9
81.7
77.9
_/  Effluent data for the treatment works was not available;  therefore, actual BOD removals could not be computed.   The percent BOD removal shown
    was based on a treated effluent load computed by subtracting the raw water intake load from the combined effluent load.   This assumes no
    incremental BOD increase in the cooling water stream which is known to be incorrect.   The computed values thus  represent "minimum" removals
    and actual removals would be higher.
b/  Overall BOD removal efficiencies were based on a comparison of the combined effluent  loads with the process waste loads  to treatment plus
    the raw water intake loads.

-------
                  Table F-3
Summary of Monthly TSS and Total  Solids Data
Merck and Co., Cherokee Plant,  Danville, Pa.
        December 1972 - February  1974
                                                                                                           i
                                                                                                           o
Month

Dec.

Jan.
Feb.
Mar.
April
May
June
Average
(First
July
Aug.
Sept.
Oct.
Nov.
Dec.
Average
(Second

Jan.
Feb.
Average
Overall
Total Suspended Solids (TSS)
Raw Water Intake
(kg/day)

4,450

2,360
2,490
10,910
5,220
1,510
2,110
4,100
Half 73)
640
1,020
590
620
703
3,840
1,200
Half 73)

2,540
1,740
, 74 2,100
Average 2,700
(Ib/day)

9,820

5,200
5,480
24,050
11,500
3,320
4,660
9,000
1,410
2,240
1,290
1,360
1,550
8,470
2,700

5,590
3,830
4,700
6,000
Combined
(kg/day)

6,580

6,730
3,390
5,830
4,560
5,920
5,810
5,400
7,080
1,310
1,710
1,580
2,280
7,390
3,540

3,420
1,820
2,600
4,400
Effluents
(Ib/day)

14,500

14,830
7,470
12,850
10,050
13,050
12,800
11,800
15,610
2,880
3,760
3,480
5,030
16,300
7,800

7,530
4,010
5,770
9,600
Net Increase
(kg/day)

2,130

4,350
900
-5,080
- 660
4,400
3,670
1,270
6,440
270
1,130
950
1,590
3,540
2,300

880
80
500
1,600
(Ib/day)
1972
4,700
1973
9,600
1,990
-11,200
- 1,460
9,700
8,100
2,800
14,200
600
2,500
2,100
3,500
7,800
5,100
1974
1,940
180
1,100
3,600
Raw Water Intake
(kg/day)

14,940

13,810
16,800
24,570
15,240
13,990
21 ,600
17,700
16,490
29,640
29,630
29,650
22,750
14,670
23,800

12,830
13,620
13,200
19,400
(Ib/day)

32,930

30,440
37,030
54,160
33,600
30,850
47,620
39,000
36,360
65,340
65',320
65,370
50,140
32,350
52,500

28,280
30,020
29,200
42,700
Total Solids
Combined
(kg/day)

32,750

45,930
33,370
33,890
29,680
34,470
36,920
35,700
23,900
44,940
40,370
44,170
38,010
36,000
37,900

29,100
25,800
27,400
35,300
Effluents
(Ib/day)

72,200

101,260
73,570
74,720
65,430
76,000
81 ,400
78,730
52,690
99,070
89,000
97,380
83,800
79,370
83,600

64,160
56,870
60 , 500
77,800
Net Increase
(kg/day)

17,830

32,110
16,560
9,340
14,420
20,500
15,330
18,050
7,410
15,290
10,750
14,380
15,290
21,320
14,100

16,280
12,180
14,200
15,900
(Ib/day)

39,300

70,800
36,500
20,600
31,800
45,200
33,800
39,800
16,330
33,700
23,700
31,700
33,700
47,000
31,000

35,880
26,850
31,400
35,100

-------
                                                        Table F-4
                                                                               .a/
                                        Amnom'a Nitrogen  and  Phosphorous  Content-
                                  Merck and Co., Cherokee Plant,  Danville,  Pennsylvania
                                             October 30  - November 15,  1973
Sample Point
Treatment Plant
Influent
Treatment Plant
Effluent
Combined Total
Plant Discharges
Raw River
Water Intake
Flow
(m
3
3
90
90
3/day)
,370
,370
,500
,500
(mgd)
0.89
0.89
23.9
23.9
(mg/1)
451
454
23
0.30
NH3 - N
(mg/1)^
165-620
335-590
16-31
0.10-0.65
(kg/day)
1,530
1,520
2,070/
24
(Ib/day)
3
3
4

,370
,360
,570/
53
(mg/1)
59
34
2.2
0.2
Phosphorous
(kg/day)
200
116
195^
21
(Ib/day)
440
255
46
a/  All values shown are averages except for NH, -  N  range.
b/  Range
c_/  Incremental increases over and above treatment  plant effluents  shown of 550 kg  (1,210  lb)/day  NH3-N  and
    79 kg (175 lb)/day phosphorous.   These incremental  increases  attributable  to spent  cooling  water  discharges,
    overflows from the small  settling pond,  or other  miscellaneous  unrecorded  discharges.

-------
F-12
     Waste load data were available from the Company for the process
waste stream before treatment, the raw water intake, and the combined
plant effluent including spent cooling water.  These data were evaluated
for the December 1972 through February 1974 period of record.  Monthly
BOD data for this period are summarized in Table F-2.  Total suspended
solids and total solids data are summarized in Table F-3.  Nitrogen and
phosphorous data for the period October 30 through November 15, 1973 are
summarized in Table F-4.

     A review of the data in Table F-2 indicates that treatment perfor-
mance improved by mid-1973 following completion of additional treatment
facilities in December 1972 and full operation of the sludge incinera-
tion process.  Monthly BOD treatment removal efficiencies ranged from 63
to 90 percent with an average of 80 percent.  Because about 95 percent
of the wastewater discharged by the Cherokee Plant is untreated or
partially treated, the overall system BOD reduction is considered more
important than the treatment plant reduction per se.  The overall re-
duction averaged only 78 percent.  In discussions with the Company
during 1974, they indicated that conceivable theoretical BOD reductions
as high as 99 percent could be attained, except that costs would be
entirely prohibitive.

     A comparison of observed waste loads [Table F-2] and plant design
criteria [Table F-l] indicates that current process waste flows averaging
3,780 m /day (1.0 mgd) are about 25 percent higher than^design hydraulic
loading.  The July 1973 monthly average flow of 6,400 m /day (1.69 mgd)
was 111 percent over design.

     With respect to BOD loads, the long-term average of 7,080 kg
(15,600 lb)/day for raw process wastes is slightly below design. The
design criteria [Table F-l] for the cooling water stream assumed a BOD
load of 590 kg (1,300 lb)/day present in the raw water intake.  This
load has averaged only 160 kg (360 lb)/day.  Adjusting the design criteria
for this difference effectively reduces the design combined effluent BOD
load to 2,000 kg (4,400 lb)/day.  Average effluent BOD was 1,600 kg (3,500
lb)/day, somewhat below design levels.

     The Company no longer analyzes for COD.  However, based upon past
experience, they estimate that the CODiBOD ratio for treatment plant in-
fluents should be about 2.0, and for final effluents it should be between
8 and 10.

     Observed total  suspended solids and total solids loads in the com-
bined effluents [Table F-2] averaged substantially above design criteria
[Table F-l].  The TSS loads were unexpectedly high.

-------
                                                                    F-13


     The nitrogen and phosphorous data collected by the Company in
October-November 1973, [Table F-4] demonstrate high'levels of N and P
both before and after treatment.  Ammonia nitrogen concentrations are
exceptionally high.  Phosphorous removals for this period were about 40
percent through the treatment works whereas ammonia nitrogen removals
were virtually zero.  Ammonia-N loads leaving treatment averaged 1,540
kg (3,400 lb)/day, and 2,090 kg (4,600 lb)/day in the combined effluents.
These levels are substantially above the Company design specifications
[Table F-l].  Incremental increases of 550 kg (1,200 lb)/day ammonia-N
and 80 kg (180 lb)/day phosphorous are contributed by the spent cooling
water stream.  It is curious to note that final ammonia-N discharges are
appreciably greater than effluent BOD loads.  In the past, ammonium
salts had been used for waste neutralization purposes.  According to the
Company, this practice is no longer employed.

Future Production Levels

     The large-scale expansion of the Cherokee Plant scheduled for
completion during 1975-1976 will substantially alter production levels
and associated waste loads.  Waste treatment facility improvements will
be needed to cope with the increased waste load.  Merck and Co. has sub-
mitted waste treatment plans to the State of Pennsylvania.  During 1973-
1974, in various materials submitted to the State, the Company indicated
a 35 percent expansion of plant facilities at Cherokee, but also requested
the State for much more than a 35 percent increase in effluent waste
load allocations for various parameters.  The Company not only assumed
that their existing treatment was adequate but also that their plans for
future treatment were adequate.  A number of important questions have
been subsequently raised on this subject.

     The Merck (1975) treatment plant design contemplates future full-
scale process raw wastes of 4,540 m /day (1.2 mgd) containing 17,700 kg
(39,000 lb)/day BOD; 12,300 kg (5,100 lb)/day TSS; 3,650 kg (8,050
lb)/day NH-j-N; 55,650 kg (122,700)/day TSS; and 4.5 kg (10 lb)/day of
CN.  The 4,540 m /day (1.2 mgd) flow figure already requires correction
because the Company assumed a 1,500 m /day (0.4 mgd) increase due to 1-
methyldopa production over and above an average 3,000 m /day (0.8 mgd)
for existing production.  However, the latter flow figure should be at
least 3,780 m /day (1.0 mgd).  The plans and specifications of the
Company indicate their "to-be-expanded" waste treatment facilities would
provide effluent loads of 2,500 kg (5,520 lb)/day BOD (a 85.8 percent
reduction); 920 kg (2,025 lb)/day TSS (only a 50 percent reduction);
1,860 kg (4,100 lb)/day NH3-N; 55,650 kg (122,700 lb)/day TDS (zero reduction);
and 0.4 kg (1 lb)/day CN (a 90 percent reduction).  The Company has in the
past stated they will be achieving 90 percent or better removals of BOD,
but the design plans themselves show only a 85.8 percent BOD removal to
be expected.  Comparing the above figures to design of the existing
treatment plant, future raw waste load increases associated with the 1-methyldopa

-------
F-14


expansion are calculated as follows:  Flow - 20 to 50 percent; BOD - 105
percent; TSS - 17 percent; NH--N - 75 to 220. percent; and IDS - a 115
percent increase.  It is quite evident that "more than 35 percent increases
are involved, and the Company plans are not adequate to meet the effluent
limitations that have been written into the Cherokee Plant NPDES permit
as discussed below.

     Based on an e/tensive data analysis, raw waste loads for current
production levels and future levels for full 1-methyldopa production
have been determined as follows:
Parameter         1974-1975 Full Production    1975-1977 Full Production
                  (kg/day)(Ib/day)    (kg/day)(Ib/day)
BOD
COD
TSS
NH3-N
8,600
17,200
2,100
"
19,000
38,000
4,700

17,700
28,200
2,300
3,650
39,000
62,200
5,100
8,050
DEVELOPMENT OF NPDES PERMIT LIMITATIONS

     Because of the severity of the nitrogen problem at Cherokee, limi-
tations on ammonia nitrogen have been necessarily established for the
NPDES permit.  With the 1-methyldopa production expansion, the Company
predicts a total equivalent raw ammonia-N waste load of around 3,670 kg
(8,100 lb)/day).  The permit has specified future anmonia-N discharges
of 1,040 kg (2,300 lb)/day, representing ammonia-N reductions of 70 to
75 percent.  Available technology strongly infers that ammonia nitrogen
loads could be reduced even below these prescribed levels.  Phosphorous
discharges have not been constrained by the permit, although eventually
this issue will receive increased attention.

     In the NPDES permit, consideration has been given to limitations on
metallic and trace ions for the-combined Cherokee effluents in the
future.  Selected constituents are of real and potential concern pre-
dicted upon limited analyses made available by the Company on effluents
and/or because of relatively high concentrations of these materials in
the incoming plant water supply.  Cyanides and phenolics can be of con-
siderable importance in wastewaters associated with 1-methyldopa pro-
duction, and NPDES future limits have been set at around 0.15 mg/1.
Iron and manganese levels are quite high in the raw water supply and
because of relatively stringent State receiving stream criteria on these
parameters, suggested limits have been given as 0.5 mg/1 for both
parameters.  Iron contributions can also be significant in the waste-
waters coming directly from the production of 1-methyldopa.  Sulfide

-------
                                                                    F-15


limits have been utilized due to appreciable sulfate loads in the Cherokee
effluents and this constituent has been set at a maximum of 0.5 mg/1.
Additionally, maximum levels of chromium and aluminum were tentatively
prescribed at 0.25 mg/1 and 0.75 mg/1, respectively.  These various
limits are to be attained by 1967-1977.  It is also noted from past data
that sodium values are quite high in the Cherokee waste flows although
no action has been taken on this parameter.

     Temperature limits were considered but were not established because
of the ability of the River to adequately absorb existing thermal loads.
Fecal coliform limits of 1,000 organisms/100 ml were tentatively prescribed.

     The future plans also predicted incremental increases in the spent
cooling water waste streams of about 820 kg (1,800 lb)/day BOD and 320
kg (700 lb)/day NH3-N.  In view of the reported presence of only non-
contact waters, these incremental loads were considered unacceptable for
purposes of the NPDES permit.  Accordingly, the Company promised corrective
action.  With these incremental loads, the Company had previously predicted
future total effluent loadings of around 3,320 kg (7,320 Ib) BOD/day and
2,180 kg (4,800 Ib) NFL-N/day.  These loads have been substantially
lowered in the NPDES permit.

     Based upon the tentative future criteria currently being developed
for the bulk pharmaceutical industry, together with other considerations,
average daily load limits to be achieved by 1976-1977 for major para-
meters in the Merck, Cherokee Plant permit are as follows:
Parameter                	Load	             Reduction
                        (kg/day)(Ib/day)            (percent)

BOD
   Gross                 1,200          2,700                 94
   Net                   1,050          2,310
TSS (Net)                  690          1,530                 70
COD                      5,440         12,000                 81
Ammonia-N                1,040          2,300                 72

-------
               PHARMACEUTICAL INDUSTRY
                  CASE HISTORY  G

                WYETH LABORATORIES
              MARIETTA, PENNSYLVANIA

(BIOLOGICALS PRODUCTION AND DRUG FORMULATION PLANT)

-------
                                                                     G-l
              WYETH LABORATORIES, MARIETTA,  PENNSYLVANIA
          (BIOLOGICALS PRODUCTION AND DRUG FORMULATION PLANT)


BACKGROUND AND PROCESSES

     The Wyeth Laboratories Pharmaceuticals  manufacturing installation at
Marietta, Pa., specializes in the production of vaccines and serums,  but
also formulates and packages a number of synthetic organics.  A large ma-
jority of final product is packaged as parenteral  (injectable) ampules.
The facility, opened for manufacturing in May 1965, basically operates
five days a week with a single-shift operation.  Total employment is
around 400 persons.  The plant was visited by EPA, NFIC-Denver, in April
1974 to obtain information for use in drafting a NPDES discharge permit
for the Company at the request of EPA, Region III.

     Processes used to produce vaccines and serums consist of specialized
culturing and growing of bacteria and viruses, converting these bacteria
and virus into vaccines, immunization of animals,  repeated testing, and
packaging of the final products.  Formulation of organic compounds (e.g.
atropine sulfate) involves processes such as solubilization, filtration,
testing, and packaging.

     Types of vaccines most commonly manufactured at Marietta include flu,
tetanus and smallpox vaccines.  Cholera and other "exotic" vaccines are
also produced by the facility on special request.   Snake-bite serums
used against a wide variety of North American snakes also constitute  a fairly
well-known manufacturing line.  Synthetic organic compounds manufactured
include atropine sulfate (an anticholinergic and mydriatic); the anti-
histamines; selected narcotics; heparin (an anticoagulant); phenobarbital,
et. al.  Marietta turns out extremely large numbers of ampules (1 to  2 ml.
used as injectables) amounting to tens of millions annually.

     During late winter and early spring, the plant manufacturers large
batches of flu vaccine.  This manufacturing is highly dependent upon  pur-
chase orders placed by local, State and Federal governments.  About 25
persons are specifically engaged for 2 to 3 months in production of flu
vaccine.  During the remainder of the year these persons are diverted to
other activities.  Tens of thousands of eggs are used daily in producing
flu vaccine.

     Animal colonies at Marietta are reported to be steadily diminishing.
This is attributed to certain vaccines now being successfully cultured and
grown on beef broth rather than on animal forms as previously was necessary.
Monkey colonies are no longer maintained at Marietta and the horse popula-
tion is currently at a minimum level.  Other animal colonies are main-
tained mostly for blood testing purposes.

-------
G-2


WASTE TREATMENT AND CONTROL

     Special waste handling procedures in the animal holding areas include
the transfer of cow dung to incinerators; composting of horse manure with
the compost eventually given away to nearby farmers; and the dry cleaning
of small-animal cages with the droppings going to incinerators and the
cages then being steamed and washed with an alkaline detergent-disinfectant.
Floors and equipment in the animal holding areas are reported to be fre-
quently washed and sprayed with various disinfectants including "OSWL,"
"Wescidine I" and "Terramine."  Large animal carcasses are disposed of to
an off-site rendering establishment.  Discarded eggs and shells are col-
lected into 190 1 (50-gal.) drums and carted away to sanitary landfill.
Liberal amounts of disinfectant are used in the egg collection operations.
In the event of spills of concentrated vaccine solution or any other
hazardous substance occurring within the animal housing areas or in the
main Pharmaceuticals plant, special isolation procedures can be quickly
implemented.  Isolation is followed by flooding affected floors and
equipment with a strong hypochlorite solution for a minimum period of
one hour.  Routine plant procedures involve extensive autoclaving of
glassware' and associated refuse.

     Wyeth indicates they have three main safety systems:  1) cleaning up
spills immediately after occurrence; 2)  air locks in individual prepara-
tion-packaging rooms together with a separate air conveyance and filtra-
tion system; and 3)  activated sludge treatment of plant wastewaters fol-
lowed by three hours of chlorination to give added guarantee in killing
bacteria and virus.  Some years ago, Wyeth conducted toxicological testing
of the treated final effluents upon newborn mice to discern possible en-
vironmental effects.  The Company reports all such tests were negative.

     Process wastes, animal waste residues,3boiler blowdowns and various
spent cooling water streams averaging 227 m /day (60,000 gpd) are discharged
to an activated sludge treatment facility.  Final discharge is to Evan's
Run, a tributary to the main branch of the Susquehanna River in Western
Pennsylvania.

     Spent process, domestic and cooling waters3are combined, passed
through a comminutor, and discharged to a 150 m  (39,500 gal.) equalization
tank equipped with auxiliary aeration.  The equalizing tank provides
roughly an 18-hour detention time.  Wastewaters are carried through the
weekends and holidays, and consequently maximum level in the equalizing
basin is generally reached late on Friday.  Raw wastes are then passed
into two parallel activated sludge aeration basins of 127 m  (33,600 gal.)
capacity each although it has been observed that usually only a single
basin is in operation at any one time.  Wastewater detention in the
activated sludge basins approximates 8 to 12 hours.  Activated sludge
effluents are settled within a circular clarifier with a capacity of
119 m- (31,400 gal.). ^The clarifier overflow rate is reported as
9.0 m /m  (220 gal./ft )/day based upon a (original) design hydraulic
load of 490 m /day (130,000 gpd).  Sludge is recycled from the secondary
clarifier back to the aeration chambers.  Excess biological sludge is

-------
                                                                     G-3
sent to a small aerobic sludge digester.  Stabilized sludge is taken
to sanitary landfill.  The final unit in the treatment plant consists
of a chlorine contact basin of 64 m  (17,000 gal.) capacity giving 3.2 hours
detention at design flow.  However, there is considerable more detention
at current wastewater flow rates.  Residual chlorine is thought to be
in the range of 0.75 to 1.0 mg/1.

     During the visit of April 2, 1974, a murky and deep purplish-red
color was observed in the chlorine contact chamber.  The Company attri-
buted this color to recent cleanout of solids in the equalizing basin.
Contents of the chlorine contact chamber also appeared relatively stag-
nant.  The degree of mixing in the chlorine contact tank could be im-
proved.  There was an abundance of chicken feathers in the secondary
clarifier which were passing over into the chlorine contact unit to-
gether with other floating and suspended matter.  The MLSS content in the
aeration basin(s) have been reported in the range of 2,300 to 9,100 mg/1
with values most often around 4,000 mg/1.  The presence of liberal
amounts of blood are suspect within the treatment plant.  The treatment
facility has a design BOD load of 127 kg (280 lb)/day equivalent to 150 kg
(330 lb)/day ultimate BOD.  It is noted that existing BOD loads are sig-
nificantly below these figures.
WASTE LOADS

     Performance of the wastewater treatment operations was evaluated for
the January 1973 to February 1974 period of record.  Monthly results for
this period are summarized in the Table 6-1.  Even though raw waste loads
entering the treatment system are relatively small, it was noted that both
the hydraulic and organic waste loads varied widely from month to month.
Flows.,ranged from a low of 127 m /day (33,600 gpd) in January 1973 up to
344 itT/day (91,000 god) in July 1973, a ratio of 2.7 to 1.  The mean waste-
water flow was 205 m /day (54,100 gpd).  Monthly BOD loads incoming to
treatment were quite low averaging 7.5 kg (16.6 lb)/day, but ranging from
2.3 kg (5.0 Ib) all the way up to 33.1 kg (73.0 lb)/day.  The average
COD raw waste load was 26.6 kg (58.7 lb)/day, and the monthly COD loads
as for BOD varied widely, ranging from 11.3 kg (25 lb)/day in January
1973 up to 67.7 kg (149.2 lb)/day in July 1973, a ratio of 6 to 1.
Biochemical oxygen demand concentrations both before and after treatment
were amazingly low signifying the presence of either large amounts of cool-
ing and washing waters in the system or, more probably, serious toxicity
impact upon BOD test results.
     Over the 14-month period of record, overall BOD, COD and TSS loadings
leaving the treatment plant in the final effluents were 0.4, 13.2 and 5.9 kg
(0.9, 29.0 and 13.0 lb)/day, respectively.  Monthly BOD effluent loads varied
from 0.1 to 1.0 kg (0.3 to 2.3 lb)/day whereas TSS effluent loads ranged
from 2.3 to 9.8 kg (5.0 to 21.6 lb)/day.  The COD/BOD ratios for the raw
wastewaters were generally in the range of 4:1 to 8:1.   This ratio for final
effluents was in the range of 15:1 to as high as 100:1.  These ratios were

-------
                     Table G-l

Summary of Monthly Waste Treatment Performance Data
    Wyeth Laboratories,  Marietta,  Pennsylvania
        January 1973 through February 1974
Flow
Month
Jan. 73
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov:
Dec.
Jan. 74
Feb.
Average
(m3/day)
127
184
-
142
202
246
345
265
224
190
218
159
171
189
205
(gpd)
33,600
48,600
-
37,400
53,400
65,000
91,100
70,000
59,200
50,100
57,700
42,000
45,200
49,900
54,100
(mg/1)
18*
17*
-
65
24*
19*
96
16*
17*
17*
25*
34*
42
60
35*

BOD
(kg/day)
2.3
3.1
.
9.2
4.9
4.6
33.1
4.2
3.9
3.2
5.5
5.4
7.2
11.4
7.5*
Raw

(Ib/day)
5.0
6.9
-
20.3
10.7
10.2
73.0
9.3
8.5
7.0
12.1
12.0
15.9
25.1
16.6*
Wastes

(mg/1)
90
68
.
90
110
100
196
109
148
158
126
132
151
151
126

COD
(kg/day)
11.3
12.4
-
12.7
22.5
24.5
67.5
28.8
33.2
30.0
27.6
21.1
25.8
28.6
26.6

(Ib/day)
25.0
27.4
-
27.9
49.7
54.1
149.2
63.5
73.1
66.2
60.9
46.6
56.9
63.0
58.7
                                                                                                        CD
Treated Effluent
Month
Jan. 73
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 74
Feb.
Average

(mg/1)
3
1
.
1
1
3
3
2
1
2
2
1
2
1
2*
BOD
(kg/day)
0.4
0.2
.
0.2
0.2
0.7
1.0
0.5
0.3
0.3
0.5
0.2
0.3
0.1
0.4*

db/day)
0.8
0.5
.
0.4
0.4
1.6
2.3
1.2
0.6
0.7
1.2
0.4
0.7
0.3
0.9*

(mg/1)
49
36
.
37
29
76
63
60
71
51
89
64
87
116
64
COD
(kg/day)
6.2
6.5
-
5.3
5.9
18.7
21.5
15.9
15.8
9.8
19.5
10.1
14.3
21.9
13.2

(Ib/day)
13.6
14.4
.
11.6
12.9
41.2
47.4
35.1
34.8
21.6
42.9
22.3
31.6
48.2
29.0

(mg/1)
48
37
.
16
48
18
17
21
26
21
43
21
28
48
30
TSS
(kg/day)
6.1
6.8
.
2.3
9.8
4.4
5.8
5.5
5.9
4.0
9.3
3.3
9.8
9.0
5.9

( Ib/day)
13.4
15.0
-
5 0
21.6
9.6
12.7
12:1
12.9
8.9
20.6
7.2
10.6
19.8
13.0
Waste Reduction
BOD
(percent)
84.0
92.7
.
98.0
96.5
84.4
96.9
87.2
93.0
90.0
90.1
96.5
95.7
95.2
92.3*
COD
(percent)
45.5
47.5
_
58.4
74.0
23.8
68.1
44.7
52.4
67.4
29.6
52.2
44.5
23.5
48.6

-------
                                                                     6-5
extremely high.  Importantly, during many months, the calculated BOD
loads attributable to domestic sewage alone (from the plant employees)
were higher than the measured total BOD raw loads entering the treatment
plant.  Artificially low BOD results were apparently being obtained both
on the raw and final wastewaters due to toxicity.  The overall BOD
removal through the treatment plant from January 1973 through February
1974 was 92.3 percent, which, however, is considered rather meaningless
in view of the toxicity issue.  The overall COD removal of only 48.6 percent
over this same period is considered far from desirable.
NPDES PERMIT CONDITIONS

     Data on heavy metals supplied by Wyeth Laboratories on the Marietta
w&stewater effluents were less than sufficient to assess the impact of
metals upon biological treatment efficiency and any other interferences
related to potential toxicity.  Some metals were in the marginally ac-
ceptable/objectionable range including aluminum, boron, chromium, iron,
copper and mercury.  Levels of metal ions were such that limits were not
necessary on the NPDES discharge permit; however, more information is
sought in the future.

     Final effluent limitations to be reached by April 1976 as specified
in the NPDES permit in terms of the average daily and maximum daily con-
ditions are respectively: BOD - 3.2 and 5.0 kg (7 and 11 lb)/day; COD -
13.6 and 20.4 kg (30.and 45 lb)/day; and TSS - 5.9 and 9.1  kg (13 and 20 Ib)
/day.  Fecal coliform limitations and bioassay testing have also been speci-
fied in the permit.  With respect to nitrogen and phosphorous, these
appear present in the final effluents as highly-oxidized forms; thus, no
permit limits were specified.

-------
            PHARMACEUTICAL INDUSTRY
                CASE HISTORY H

              WYETH LABORATORIES
              PAOLI, PENNSYLVANIA

(DRUG MIXING,  FORMULATION AND PREPARATION PLANT)

-------
                                                                      H-l
                WYETH LABORATORIES, PAOLI, PENNSYLVANIA
           (DRUG MIXING, FORMULATION AND PREPARATION PLANT)
BACKGROUND
     Around 1956, Wyeth Laboratories decided to relocate its drug
formulation and packaging facilities from Philadelphia to Paoli, Pa.,
some 30 miles west.  This move was completed in 1961.  In contrast
to the availability of a municipal sewerage system in Philadelphia,
Company waste treatment facilities were required at the new location.

     The Paoli facility is a typical Pharmaceuticals formulation
plant that prepares and formulates a wide range of medicinals, drugs and
related compounds.  Compounds comprise mainly synthetic organic materials
together with some naturally derived substances.  Final products are
primarily oral medications in the form of tablets and capsules; however
liquid solutions and suspension-in-liquids are also manufactured.  A
partial listing of final products manufactured, according to the best
information available, is as follows.

     Various cough syrups               Phenacetin (antacid)
     Oral contraceptives                Vitamin C
     Various penicillin formulations    Benzaldehyde
     Analgesics                         Boric acid
     Tranquilizers                      "Bismuth"
     Aspirin                            Caffeine
     Suppositories                      Codeine-containing products
     Eyewashes                          "Phenergan"
     Flavoring agents                   Sodium benzoate
     Food colorings                     Manitol
     Magnesium stearate                 Phenobarbital
     Meprobamate (tranquilizer)         Pentritol
     Oxazepam (tranquilizer)            Diethylstilbestrol
     Cyclospasmol

     Major raw materials reported as being received into the Paoli plant
include sugar, corn syrups, lactose, cocoa butter, gelatin (used in
capsules), calcium, talc, kaolin, diatamaceous earth, thiourea, ethyl
alcohol, glycerine, wine, sorbitol, aspirin, bulk penicillin, and various
analgesics.

     The plant essentially operates over a single shift, five days a
week, and employs 1,450 persons.  A skeleton staff is maintained during
the second and third shifts and over weekends.

     Treated effluent from the plant is discharged to a small creek
tributary to Little Valley Creek in the Schuylkill River Basin.

-------
   H-2


     Information presented herein was obtained from two engineering reports
(29, 50) and a March 1974 plant visit by NFIC-Denver to gather information
for development of a NPDES discharge permit.


WASTE TREATMENT AND CONTROL

Waste Sources

     Waste characterization and treatability studies undertaken to provide
the basis for the design of the new Paoli treatment facility identified
the following probable sources of wastewaters (29, 50):

     a)   The formulation and bottling of liquid preparations.
     b)   The formulation of dry preparations being shaped into coated
          tablets, or made up into capsules, and subsequently packaged.
     c)   The formulation and packaging of various jelled preparations.
     d)   The product development laboratory.
     e)   The analytical laboratories.
     f)   Sanitary sewage, cafeteria residues and other domestic needs.

     Spent organic materials were expected to consist chiefly of lactose,
corn syrups, wine, sucrose, gelatin, cocoa butter, alcohol, together
with contributions from the cafeteria, and sanitary wastes.

     Cooling water is recycled through a large double unit cooling
tower.  System blowdown is discharged to the waste treatment system.


Initial Treatment System Design and Operation

     Predicted wastewater flows and characteristics for the 1961 initial
period of operation and for the 1966 design year as developed by the
waste characterization study are summarized in Table H-l.

                               Table H-l
               Projected Waste Loadings, Wyeth Labs (29)
Initial (1961)
Type Waste
Volume
BOD
(m3/day)(gpd) (kg/day )(lb/day)
Process
Wastes
Sanitary
Wastes
Total
121
74
195
32,200
19,500
51,500
38
17
55
83
38
121
Design
Volume
(m3/day)(gpd)
193
114
307
51,000
30,000
81 ,000
(1966)

BOD
( kg/day )(lb/day)
63
27
90
138
60
198

-------
                                                                     H-3
A summary of design criteria for the 1966 average conditions was given
as:,

     COD strength                       -  450 mg/1
     Flow rate                          -  371 1/min (98 gpm)
     Sludge recycle rate                -  136 1/min (36 gpm)
     Aeration basin detention           -  8.3 hr
     Sludge stabilization time          -  4.2 hr.,  ~                ?
     Clarifier rise rate                -  18.3 rrT/nT/day (450 gal/ft /day)
     Clarifier detention                -  2.4 hr

     Laboratory waste treatability evaluations showed that at certain
F/M ratios the wastes exhibited toxic effects.  To preclude toxicity to
sludge micro-organisms, the preliminary studies indicated that an
aeration time greater than 4 hr was necessary.  An acclimated sludge was
eventually developed to counteract toxicity.  Water quality limits for
the receiving stream necessitated that treatment should provide a
minimum of 85 percent reduction of organic matter and removal of sub-
stantially all TSS, and toxic, odor and taste-producing materials.
Since both process and sanitary wastes were generated from 8 AM to 5 PM
on weekdays only, it was decided to segregate the process and sanitary
flows, to accept the sanitary wastes on a demand basis, but to provide
extensive equalization and holding of the process wastes.

     Initial operation of the full-scale treatment facilities began in
March 1961.  This system provided separate conveyance of sanitary and
cafeteria wastes into the treatment plant through a comminutor, with
these wastes entering directly into the activated sludge aeration tanks.
Process wastes were passed through a primary skimming basin, into one of
two equalization tanks, and then pumped at a controlled rate to the
aeration basins.  Capacity of the equalizing tanks was three times the
average daily process wasteflows.  Four aeration basins were construct-
ed, arranged in two parallel sets of two basins each.  Overflows from
the aeration basins passed to two final clarifiers operated in parallel.
Settled effluents were chlorinated before final discharge.  Biological
sludges from the final clarifiers were bled to two aerated sludge
stabilization tanks.  Stabilized sludges were recycled to the activated
sludge aeration basins.

     In the initial months of treatment plant operation, operation of
all aeration basins was found to be unnecessary; consequently, one of these
tanks was converted to an aerobic sludge digester for handling excess
biological solids.  Sludges were ultimately removed from the treatment
circuit via chemical conditioning, vacuum filtration and landfill disposal.

-------
H-4


     During.,1961-1962, average monthly process waste flows ranged from
76 to 106 m /day (20,000 to 28,000 gpd).  Sanitary flows approximated 95
to 132 nr/day (25,000 to 35,000 gpd).  From July 1961 through April
1962, monthly influent loads to the treatment facility ranged from 30 to
70 kg (67 to 154 lb)/day BOD and 46 to 98 kg (101 to 216 lb)/day COD.  Treated
effluent for this same 1961-62 period contained 1.5 to 2.7 kg (3.2 to
5.9 lb)/day BOD and 7.3 to 16.7 kg (16.0 to 36.8 lb)/day COD.
Overall average waste removals were 96.1 percent for BOD and 83.7 percent
for COD.  Effluent TSS concentrations, in terms of monthly averages,
ranged from 7 to 34 mg/1.  The TSS level in the final effluent averaged
21.5 mg/1 during this 1961-62 period.

     The Company reported that somewhat lower BOD removals (i.e. 93 and
94 percent) were recorded when the strength of the wastewater entering
the treatment facility was abnormally low.  Under normal conditions, BOD
removal was in the range of 97 percent and COD removal around 85 percent,
which are impressively high (29).  A later report gives treatment
performance results covering the period January 1961 through February
1963 (50).  This later record of 26 months shows an average hydraulic
load of 190 nT/day (50,000 gpd) with about 54 kg (120 Ib) BCD'/day
entering treatment.  The final effluent was said to be averaging about
12 mg/1 BOD equivalent to 2.3 kg (5.0 Ib) BOD/day and 95.8 percent BOD
removal.  Effluent COD was averaging around 80 mg/1.

     Foaming problems attributed to the presence of detergents were
experienced in the wastewater treatment plant.  A number of foam-depressant
chemicals were employed with only limited success including a waste
silicone emulsion, sperm oil, isodecanol,.and commercial antifoams.
Next, surface sprays were installed at the head end of the final settlers
to assist in breaking down the persistent foam layer, which solved the
foaming at least on the clarifiers.  A better solution was later devised,
however, which involved converting one of the aeration basins into a
foam collecting tank.  Excess foam was collected from within the treatment
circuit and broken down by mixing without aeration, and then fed to the
aerobic digester.  The foam was not as intense in the digester because
of the higher solids levels.  The digester was dewatered daily.   Although
detergents in the effluent were not reduced, nevertheless, operational
difficulties were reported significantly lowered by these procedures
(29, 50).

Current Waste Treatment

     When inspected by NFIC-Denver in March 1974, several changes in
waste treatment from the original design were noted.  Sanitary wastes
are now discharged to the equalizing basins along with process wastes.
Air spargers have been added to the equalizing basins in the last year
and a half to eliminate septic conditions during periods of low flow,
particularly over weekends.

-------
                                                                     H-5
     Process wastes enter a small collection chamber where flotable
greases are skimmed off and collected.  Sanitary wastes pass through a
bar screen and communitor.  Process and sanitary flows then combine
before entering two equalization basins operated on a fill-and-draw
basis.  Combined flows are generally split, half into each equalizing
basin.  Each chamber is 280 m (74,000 gal.) in size, providing a total
of around 1.5 to 2.0 days detention.  Although wastewater flows are
generated only over about 10 hr each day and 5 to 5-1/2 days a week, the
treatment system is operated continuously.

     Equalized flows are fed into three activated sludge aeration
basins, presumably operated in parallel, with each basin having a
capacity of 47 m  (12,500 gal.).  The three basins provide about 9_ hr
detention time when maintaining a 1:1 wastewater to sludge return ratio.
A fourth 47 m  (12,500 gal.) tank serves as an aerobic sludge digestion
chamber.  The F/M ratio in the aeration basins is reported around 0.5.
The MLVSS content is kept around 2,400 mg/1 which i% somewhat lower than
expected.  Secondary clarification capacity is 76 m  (20,000 gal.)
equivalent to 4.5 to 5.0 hr detention.  Recycle biological sludges from
the secondary clarifiers are passed through an aerated sludge stabilization
chamber of 38m  (10,000 gal.) size prior to return to the activated
sludge aeration basins.  A baffled chlorine contact tank provides about
30 min detention of final effluent, producing a residual chlorine of
around 1 mg/1.  Excess biological sludges accumulating within the
treatment system are periodically removed by outside scavengers, presumably
to public landfill.  For the last month of available record (January
1974), process and sanitary flows averaged 243 and 93 m /day (64,100 and
24,700 gpd), respectively, for a total of 336 m /day (88,800 gpd), about
10 percent above design flow.


CURRENT WASTE LOADS

     The Company has compiled sampling results from the Paoli treatment
plant in the form of monthly reports made available to the regulatory
agencies.  The  period of record from January 1973 through January 1974 was
used in evaluating the treatment plant performance for NPDES permit
preparation. These data are summarized in Table H-2.

     A comparison of the 1973-74 data [Table H-2] with the waste loads
given previously for 1961 shows that BOD loads and wastewater volume have
roughly doubled in the past 12 years.  Treatment plant capacity, believed
based on predicted 1966 waste loads [Table H-l], was being exceeded by
about 10 to 20 percent in 1973-74.

     Biochemical  oxygen demand and COD removals averaged 94.5 and 85.0
percent, respectively, during 1973-74 [Table H-2].  Raw waste TSS data
were not collected; therefore, removal efficiencies could not be
determined.

-------
                                                                                                               CT>
                      Table H-2
Summary of Average Monthly Wastewater Treatment Data
       Wyeth Laboratories, Paoli, Pennsylvania
             January 1973 - January 1974
Influent Loads
Flow
Month (m3/day)
Jan. 1973
Feb.
Mar.
April
May
June
July
Aug.
Sept
Oct.
Nov.
Dec.
Jan. 1974
Average of
13 months
Average of
1973 Months
260
327
350
365
355
406
456
475
432
363
312
315
285

316


(gpd)
68,700
86,400
92,400
96,400
93,700
107,200
120,600
125,500
114,200
96,000
82,500
83,200
75,400

95,500


BOD
(kg/day)
59
106
93
61
54
80
109
176
134
120
149
103
115

104


(Ib/day)
130
234
204
134
118
177
241
387
295
265
329
227
253

230


COD
(kg/day)
127
252
312
81
184
114
189
222
202
174
206
196
184

188


(Ib/day)
280
556
687
178
405
252
416
489
446
383
454
433
406

414


(kg/day)
2
4
3
6
7
5
6
10
9
4
3
3
7

5.4

5.3
BOD
(Ib/day)
5
9
7
14
16
12
13
Zl
20
8
7
7
16

11.9

11.6
Effluent Loads
.COD
(kg/day)
16
24
23
25
33
29
27
33
33
15
12
28
53

27.0

24.8
(Ib/day)
35
52
50
56
73
64
60
72
72
34
26
62
117

59.5

54.7
TSS
(kg/day)
7
17
8
16
22
23
19
28
25
9
5
23
34

18.1

16.8
(Ib/day)
16
38
18
35
49
50
41
62
56
19
10
50
76

40.0

37.0
Waste Removals
BOD
(percent)
96.2
96.1
96 5
89.4
86.6
93.1
94.8
94.5
93.5
97.0
97.9
97.1
93.7

94.5


COD
(percent)
87.5
90 6
92.7
68.5
82 0
74.6
85.6
85.3
83.9
91.1
94.3
85.7
71.2

85 0



-------
                                                                     H-7
DEVELOPMENT OF NPDES PERMIT CONDITIONS

     Based on a statistical evaluation of the 1973-74 data in Table H-2,
permit conditions specifying average daily and maximum daily limits of
8.2 and 11.3 kg (18 and 25 lb)/day of BOD were recommended for the Paoli
plant to be achieved by June 1, 1976.  Corresponding limtis for COD were
36.3 and 54.4 kg (80 and 120 lb)/day, respectively.   The COD:BOD ratio
of raw wastewater has been averaging around 1.8 while this ratio for
treated effluent has been in the range of 4.0 to 5.0.  In EPA Effluent
Guideline Limitations for other industries where conventional activated
sludge, or particularly extended aeration, is employed for achieving
Best Practicable Control Technology Currently Available, allowances have
been given for somewhat greater TSS effluent loads as compared to BOD
loads.  Consequently, average daily and maximum daily limits for TSS
were recommended for Paoli as 11.3 and 18.1 kg (25 and 40 lb)/day,
respectively.

     The presence of metal lies and toxicity in the wastewaters may or
may not represent continuing problems.  The Company reports 2.5 mg/1
lead in the final effluent together with 1.0 mg/1 tin, 0.8 mg/1 nickel,
and 0.28 mg/1 zinc, but these data are represented by a single analysis
only.  Lead has been limited in the NPDES permit to a maximum daily
level by June 1976 of 0.2 mg/1.  A number of metal lies and other ions
appear in the marginal range including nickel, zinc, tin, copper, iron,
molybdenum, boron, mercury and possibly arsenic.  Because of the unknown
impact of metals and other pharmaceutical waste constituents, the
recommendation has been made that fish survival tests be conducted twice
annually, with further action to be taken if indicated necessary.  Fecal
coliform bacteria in the final effluent are consistently below 400/100
ml and should represent no problem.

-------
             PHARMACEUTICAL  INDUSTRY
                 CASE  HISTORY  I
            MCNEIL  LABORATORIES,  INC.
          FORT WASHINGTON,  PENNSYLVANIA
(DRUG MIXING,  FORMULATION AND  PREPARATION  PLANT)

-------
                                                                      1-1
       MCNEIL LABORATORIES, INC., FORT WASHINGTON, PENNSYLVANIA
           (DRUG MIXING, FORMULATION AND PREPARATION PLANT)
BACKGROUND AND PROCESS DESCRIPTION

     McNeil Laboratories, Inc., a subsidiary of Johnson and Johnson,
Inc., operates a drug formulating facility at Fort Washington engaged in
the mixing, formulating, preparation and packaging of medicinals, drugs
and associated products in dosage form for human consumption and use.
Bulk raw materials (mostly synthesized organics) are received from
domestic and foreign sources.

     The plant has about 600 employees and is believed to be operated
essentially five days a week.  Two shifts are maintained by the manu-
facturing sectors while administrative and research functions employ
only one shift.  About 160 persons are engaged in research and labora-
tory quality control (roughly 20 percent of the total staff).

     Seventy or more products are reported formulated by McNeil including
liquid, semi-solid and solid forms.  No biologicals are produced.  All
final products are intended for human use.  An abbreviated listing of
major types of products includes:

     various barbiturates, some "narcotic type" drugs, analgesics,
     Griseofulvin (an antibiotic), codeine (natural), various cough
     syrups and elixirs, intravenous and intramuscular sterile injection
     packages, anaesthesia solutions, diuretics, many different
     tablets and liquids, iron ox bile and carbolic acid (it is not
     known whether the latter two materials are raw and/or final
     products).  The Company reported that over 1500 batches of
     various products may be formulated annually at this plant.

     Significant colonies of animals (largely guinea pigs and mice) are
maintained at this facility.

     Apparently due to the startup of a new manufacturing plant in
Puerto Rico or for other reasons, the Company has stabilized operations
at the Fort Washington location.  No plans for expansion are apparent in
the forseeable future.  A substantial production capacity increase
occurred during the 1972-1973 period.
                                 o
     Water supply averaging 300 m /day (80.000 gpd) is obtained from a
local water company.  Of this total, 110 m /day (29,000 gpd) is used  for
processing, 114 m /day (30,000 gpd) for cooling, 38 m /day (10,000 gpd) for
boiler feed supply and 38 m /day (10,000 gpd) for sanitary needs.

     Major sources of wastewater include the washing of vessels and
equipment and floor washdowns in the manufacturing area, cleanup of
animal holding areas, and sanitary wastes including cafeteria waste-
waters.  Treated effluent flows to Sandy Run, a very small tributary of
Wissahickon Creek in the Schuylkill River Basin of Pennsylvania.  Sandy
Run has been determined to have high fishing and recreational importance
together with other recognized uses.

-------
1-2


     Background information, waste treatment data and effluent load data
were compiled from two engineering consultant reports prepared in 1969
and 1972, an Industrial Waste Permit issued to the plant by the Sta'te on
22 October 1969, a permit application submitted to EPA in 1971, waste
treatment plant operating data for the January 1973 through January 1974
period submitted by the Company, and other data obtained verbally from
the Company during the plant visitation by NFIC-Denver in March 1974.


WASTE TREATMENT AND CONTROL

Waste Sources

     Major wastewaters from the manufacturing area are reported to
originate from the washing of vessels and equipment and floor washdowns.
Much of the miscellaneous solids and liquids are discarded to landfill.
Mechanical compactors are employed for certain forms of collectable
solids.  McNeil reports an overall chemical loss throughout all of
handling and processing of only 1.0 to 1.5 percent.

     In the animal area, cages are scrubbed dry and the collected materials
including feces and animal carcasses are taken to an incinerator.
Discarded animal bedding and straw are transported to landfill.  The
incinerator fumes are subjected to water scrubbing with effluent discharged
to the central treatment works.  The animal cages, housing assembly and
rooms are given a final thorough cleaning and washing with the cages
then sterilized or pasteurized.  Extraneous hair is reported as a serious
problem in the sewer collection system.  The Company estimates that the
animal holding operations can generate up to 76 m /day (20,000 gpd) of
wastewater and 18 kg (40 Ib) of BOD daily.

     The cafeteria represents another potential area of waste contribution
but these services mainly use "disposables" and kitchen facilities are
minimal.  Waste solids generated in the cafeteria area are usually
bagged and hauled from the plant by a private garbage collector.

     Main boilers operate on natural gas eliminating the need for bottom
ash and fly ash disposal.  The air supply at the plant is continuously
filtered via bag filters for reuse in manufacturing areas.

     The research activities use radionuclides.  Spent radionuclides are
secured and sent to the Atomic Energy Commission for final disposal.

     Process and sanitary wastewaters are collected separately.  Animal-
handling area wastes are combined with the sanitary wastes prior to
treatment.  Cooling water is discharged to the process system.


Waste Treatment

     A 15-year old activated sludge biological treatment facility treats
all wastewaters from the McNeil  installation.   Major treatment units
with their respective capacities and detention times are presented in
Table 1-1.

-------
                                                                     1-3

Unit
Primary Settling Tank
Equalizing Basins
Aeration Tanks
Final Clarifiers
Chlorine Contact Tank
Sludge Thickener
Aerobic Digester
TABLE 1-1.
No. of
Units
1
2
2
2
1
1
1
Treatment

(m3)
76
340
199
36
24
50
81
Units
Capacity
(1,000 gal)
20
90
52.5
9.6
6.3
13.2
21.5

Detention
Time

2.5 days
<24 hours

2-4.5 hour:


     Process wastes are discharged to the primary settling tank.
Flotables are skimmed and sent to the aerobic digester.  Settled sludge
is removed to the sludge thickener that also serves as a storage tank.
The settled process wastes flow to the equalizing basins which are
operated on a batch basis with the inflow being received into one tank
while the contents of the other tank are discharged to the secondary
treatment portion of the plant.  The equalizing tanks enable necessary
carryover of flow in the treatment system during night hours but espe-
cially over the non-processing weekends.  The equalizing chambers are
equipped with air dvffusers to preclude septic conditions during waste
holding.

     The combined sanitary sewage and animal handling wastes pass
through a comminutor for shredding of solids and are then combined with
the effluent from the equalizing tanks and discharged to the aeration
tanks.  One or both aeration tanks can be employed but past operations
have involved use of a single tank providing 24 hours waste detention or
less, predicted upon the rate of flow.  The aeration basins are equipped
with sludge recycling and provisions for nutrient feed, if found necessary.
The mixed liquors leaving aeration receive final settling in a pair of
clarifiers arranged in parallel.  Sludge is either returned to the
aeration basins or the excess is wasted to the aerobic digester.  The
clarified effluents pass through a chlorine contact tank before final
discharge.  Chlorine tank detention times under design conditions are in
the range of about 2 to 4.5 hours.  Design plans call for 1 mg/1 chlorine
residual in the final effluent.

     Primary sludge essentially receives thickening only.  Secondary
sludge is sent to the aerobic digester with excess sludge transferred to
the sludge thickener.  Supernatant from the thickener is displaced to
the primary clarifier.  Sludge is unloaded from the overall system by
means of scavenger takeoff from the bottom of the sludge thickener tank
with ultimate disposal by landfill.

-------
1-4


     Process waste inflow and final treatment plant effluent are measured.
During 1967-68, average total plant influent was only 134 m /day3(35,500 gpd)
consisting of 117 m /day (31,000 gpd) of process wastes and 17 m /day
(4,500 gpd) of sanitary and animal-handling wastewater.  Average plant
inflow in 1973 was 265 m /day (70,000 gpd).  During the prime eight
hours of the working day, the inflow rate was usually 2.5 to 3 times the
average daily inflow.

     As best as can be determined, the criteria used for the design of
the treatment facility were:  flow 348 nf/day (92,000 gpd); influent
BOD, 82 kg (180 lb)/day; BOD reduction, 95 percent; effluent BOD,
15 mg/1; effluent TSS, 30 mg/1.


WASTE LOADS

     Table 1-2 summarizes average final effluent data contained in a
permit application submitted to EPA in 1971.  Effluent loads for the
McNeil formulation plant are very low in comparison to typical bulk
pharmaceutical manufacturing installations.

               Table 1-2.  Permit Application D.ata, 1971
Parameter
Flow
BOD
COD
TSS
TDS
N03-N
Total P
Sodium
Concentration
(mg/1 )
(257 m3/day or 68,000 gpd)
14
66
26
510
14
19
72
Load
(kg/day)

4
17
7
132
4
5
19

(Ib/day)

8
38
15
290
8
11
41
     The June 1972 engineering evaluation of Company sampling data for
the period September 1971 through 1972 indicated that the waste treatment
facility was achieving an average BOD removal of 93.5 percent and was 
achieving a 95 percent reduction about two-thirds to 80 percent of the
time.  The mixed liquor suspended solids concentrations in the aeration
basin were in the range of l,10p to 4,000 mg/1.  Final effluent BOD
averaged about 9 mg/1.  The maximum average rate of flow acceptable into the
McNeil treatment plant was judged to be around 340 to 378 m /day (90,000 to
100,000 gpd).  During March 1972, the average total soluble phosphorous
leaving the plant was found to be about 27 mg/1 (as P0) compared to the
State suggested limit of 0.5 to 1.0 mg/1.  The phosphorous limitations

-------
                                                                     1-5
were recognized as representing a serious constraint and challenge to the
Company.  In-plant controls on phosphorous sources and advanced physical-
chemical treatment were given brief consideration by the Company.

     The most extensive series of waste treatment performance data
available was from the Company's monthly report sheets on the waste
treatment plant.  The data indicated that there was ample reserve in
treatment capacity up through the last plant production expansion.
However, beginning in 1973, treatment system overloads occurred.  The
period of record most appropriate for preparation of the McNeil NPDES
permit was considered to extend from January 1973 through January 1974.
Results of a detailed analysis made on this data [Table 1-3] show that
the activated sludge treatment works at McNeil Labs were doing sur-
prisingly well in spite of incoming waste loads in excess of BOD and
hydraulic design conditions on a number of occasions in 1973.  Overload
conditions were probably most severe during March and August, 1973 but
also occurred on a number od days in May, November and other months
(individual daily values not given herein).  Flow quantities exceeded
379 m /day (100,000 gpd) a significant number of times.  Design capacity
of the activated sludge plant was assumed to be 348 m /day (68,800 gpd);
the remainder was process waste.  The COD/BOD ratio in the incoming feed
to the treatment plant varied from 2.0 to 3.7 averaging 2.5.  The ratio
of COD/BOD in the final effluents varied from 2.2 to 10.9 averaging 4.3.
The overall average COD removal was higher than expected at 87.7 per-
cent.  Without overload, the McNeil system could have likely provided an
average BOD reduction of 93-94 percent and a 89-90 percent COD reduction.
The calculated concentrations of BOD, COD and TSS based upon average
loadings of 3.3, 12.2 and 3.5 kg (7.2, 26.9 and 7.7 lb)/day, yield
respective values of 12.5 mg/1, 47 mg/1 and 13 mg/1.  These concen-
trations would appear quite acceptable.
DEVELOPMENT OF NPDES PERMIT LIMITATIONS

State Effluent Limitations

     State requirements and recommendations for final effluents from the
McNeil, Fort Washington installation applicable in 1973-1974 were de-
scribed as follows:  BOD - minimum of 95 percent removal  and concen-
trations not to exceed an average of 15 mg/1 and a maximum of 30 mg/1 at
any time; total soluble phosphorous (as POJ - the average not to exceed
0.5 mg/1 with an allowable maximum of 1.0 mg/1; and iron  - a maximum
allowable of 7.0 mg/1.  Recent information indicates the  phosphorous
limits no longer apply to the McNeil situation.  These phosphorous
limitations would be extremely difficult and costly to attain, especially
for a plant of this size.

-------
                                                                       Table  1-3
                                                          Monthly Waste Treatment Conditions
                                               McNeil  Laboratories,  Inc., Fort Washington, Pennsylvania
                                                             January  1973 -  January  1974
Influent Loads
Month
Jan. 1973
Feb.
Mar
April
May
June
July
Aug.
Sept.
Oct
Nov.
Dec.
Jan. 1974
Average
90 Percent
Hiqh Value
Flow
(m3/day)
249
271
289
265
251
268
289
295
278
279
286
210
184
260
315

(gpd)
65,900
71 ,700
76,300
70,100
66,300
70,800
76,400
78,000
73,400
73,800
75,500
55,500
48,600
68,800
83,300

BOD
(kg/day)
74
41
67
41
124
52
56
22
50
30
95
70
29
-
_

(Ib/day)
164
90
147
91
273
115
123
48
111
66
210
154
57
-
_

COD
(kg/day)
155
103
197
117
265
106
111
80
116
69
210
185
84
-
_

(Ib/day)
342
226
433
257
584
234
245
177
255
152
462
408
186
-
_

(kg/day)
5.1
6.8
10.0
5.2
2.3
3.7
2.4
1.6
2.2
4.4
2.4
2.0
.8
3.3
6.1

BOD
(Ib/day)
11.2
14.9a/
22.1-'
11.5
5.1
8.1
5.4
3.6
4.8
9.8
5.2
4.4
1.8
7.2
13.6

Effluent Loads
COD
(kg/day)
17.7
23.0
47.5
24 7
7.4
9.6
5.4
10.2
9.7
11.3
6.4
44.8
8.9
12.2
23 0

(Ib/day)
39.0
50 6a/
104.8-'
54.4
16.3
21.2
12.0
22.4
21.3
25.0
14 2h/
98.7^
19 7
26 9
50.7

TSS
(kg/day)
2.6
4.7
10.7
3 5
1.3
2.1
4.5
3 7
5.8
5.1
4 6
2.9
1.0
3.5
6.0

(Ib/day)
5.8
10.4,.
23.6^
7.7
2.9
4.7
9.9
8.1
12 8
11.3
10.1
6 4
2.3
7.7
13.2

Waste Removals
BOD
(percent)
93.0
83.2
. 85.0
87.5
98.3
93 1
95.7
92.4
95.8
85.2
97.4
96 9
96 8
92 4
_

COD
(percent)
88 7
77.3
75.8
75 9
98.1
91.2
95.3
87.5
91.8
84.2
96.8
_
89.6
87.7
_

a/  March values discounted because of critical  hydraulic  and organic  overload on  treatment works during  a  significant portion of the month
             -      -              -
b/  Analytical  results  questionable.

-------