TREATMENT TECHNOLOGY FOR PESTICIDE MANUFACTURING EFFLUENTS: GLYPHOSATE
by
Edward Monnig
Ruth A. Zweidinger
Mary Warner
Rosemary Batten
Dora Livennan
Research Triangle Institute
Post Office Box 12194
Research Triangle Park, North Carolina
27709
Contract No. 68-02-3688
Project Officer: David C. Sanchez
Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
Research Triangle Park, North Carolina 27711
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27709
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental
Research Laboratory, U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
-------
TREATMENT TECHNOLOGY FOR PESTICIDE MANUFACTURING EFFLUENTS: GLYPHOSATE
by
Edward Monnig
Ruth A. Zweidinger
Mary Warner
Rosemary Batten
Dora Liverman
Research Triangle Institute
Post Office Box 12194
Research Triangle Park, North Carolina 27709
Contract No. 68-02-3688
Project Officer: David C. Sanchez
Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
Research Triangle Park, North Carolina 27711
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
ABSTRACT
Various combinations of glyphosate production wastestreams were
subjected to biological treatment following lime-pretreatment to reduce
high levels of glyphosate. Bench scale biological treatment demonstrated
that glyphosate did not appear to interfere with the biological degrada-
tion process at concentrations up to 105 mg/L. On the other hand,
glyphosate itself showed only partial reduction with biological treatment
(28 to 45 percent). The mechanism of this removal is not fully understood
but may include sorption on sludge. No evidence for metabolism of
glyphosate was generated in oxygen uptake studies. While the test does
not provide any evidence for metabolic uptake of glyphosate, it is also
interesting to note that fairly high concentrations of the compound do
not inhibit other microbial processes in acclimated sludge.
Biological treatment significantly reduced the toxicity of these
effluents to algae (Selenastrum capricarnutum) and invertebrates (Daphnia
magna).
Additional treatment options were investigated in an attempt to
reduce glyphosate concentrations in the biologically treated effluents.
These options included ozonation, adsorption, and ion exhange. These
treatment options provided only marginal reduction of glyphosate concen-
trations in biologically treated effluents.
This report was submitted in partial fulfillment of Contract No.
68-03-3688 by the Research Triangle Institute under the sponsorship of
the U. S. Environmental Protection Agency. This report covers a period
from August 1, 1979 to January 30, 1980 and work was completed as of
February 15, 1980.
iii
-------
CONTENTS
Page
Abstract ill
Figures v
Tables vi
1. Introduction 1
2. Conclusions 2
3. Selected Literature Review 3
4. Treatability Studies 12
References 33
Appendices
A. Analytical Method for Determination of Glyphosate .... 35
B. Analytical Procedures for Routine Wastewater
Characterization 37
C. Procedures for Conducting Activated Sludge
Treatability Tests 38
D. Procedures for Algal Assay Tests 40
£. Procedure for Daphnia Bioassay Tests 42
F. Data Summary, for Glyphosate Wastewater Treatment
Conditions 45
iv
-------
Figures
Number Page
1 Wastewater production and treatment schematic for
glyphosate manufacturing facility 5
2 Influent and effluent COD levels for Condition I
biological treatment units 15
3 Influent and effluent COD levels for Condition II
biological treatment units 16
4 Influent and effluent COD levels for Condition III
biological treatment units 17
5 Rate of ozonation of glyphosate 30
-------
TABLES
Number Page
1 Aquatic Toxicity Data for Glyphosate and
Associated Compounds 6
2 Pesticide Manufacturing Wastewater Characterizations ... 8
3 MLVSS Data for Biological Treatment 18
A Oxygen Consumption in Biological Treatment Units 20
5 Effect of Condition I Influent on Algal Growth 21
6 Effect of Condition II Influent on Algal Growth 23
7 Effect of Condition I Effluent on Algal Growth 23
8 Effect of Condition II Effluent on Algal Growth 23
9 Effect of Condition III Influent on Algal Growth 23
10 Effect of Condition III Effluent on Algal Growth 23
11 Effect of Condition I Influent on Daphnia Magna
Survival 24
12 Effect of Condition II Influent on Daphnia Magna
Survival 24
13 Effect of Condition III Influent on Daphnia Magna
Survival 24
14 Effect of Condition III Effluent on Daphnia Magna
Survival 24
15 Comparison of Toxicity of Glyphosate to Daphnia Magna
in Hard and Soft Waters 25
16 Ozonation of Glyphosate 29
17 Adsorption Study with Condition III Effluent 31
18 Estimated Ionic Equivalents and Breakthrough Volumes
of Test Waters 32
vi
-------
SECTION 1
INTRODUCTION
In February 1979 the Research Triangle Institute (RTI) was requested
by the Industrial Environmental Research Laboratory, Research Triangle
Park (IERL-RTP) of the U.S. Environmental Protection Agency (EPA) to
conduct laboratory and pilot studies of the treatability of pesticides
manufacturing wastewaters. The project is designed to investigate the
suitability of individual pesticide manufacturing wastewaters for dis-
charge to biological treatment systems, whether publically owned treatment
works (POTW) or on-site systems.
The approach taken with each pesticide manufacturing wastewater is
hierarchical in nature, that is, less costly, more available methods of
treatment are investigated first. The preferred method of treatment is
assumed to be biological treatment. If the pesticide is judged suitable
to biological treatment based on chemical and toxicological evaluation
of the waste before and after treatment, additional options are not
investigated.
If pesticide manufacturing wastewater disrupts biological treatment
systems, the possibility of pretreating the waste prior to biological
treatment is investigated. Pretreatment options may include pH adjust-
ments, filtration, flocculation, oxidation and others depending on the
nature of the waste and its chemical composition.
If pretreatment does not improve the performance of activated
sludge systems, adsorption techniques may be investigated. These may
involve both carbon and resin systems. Physical-chemical treatability
of wastewaters will again be evaluated as with the biological treatment
system.
This report details a study of the treatability of wastestreams
resulting from the manufacture of glyphosate by Monsanto Agricultural
Products Co., at its Luling, Louisiana facility.
-------
SECTION 2
CONCLUSIONS
Various combinations of glyphosate production wastestreams were
subjected to biological treatment following lime-pretreatment to reduce
high levels of glyphosate. Bench scale biological treatment demonstrated
that glyphosate did not appear to interfere with the biological degrada-
tion process at concentrations up to 105 mg/L. On the other hand,
glyphosate itself showed only partial reduction with biological treatment
(28 to AS percent). The mechanism of this removal is not fully understood
but may include sorption to sludge. No evidence for metabolism of
glyphosate was generated in oxygen uptake studies. While the test does
not provide any evidence for metabolic uptake of glyphosate, it is also
interesting to note that fairly high concentrations of the compound do
not inhibit other microbial processes in acclimated sludge.
Biological treatment significantly reduced the toxicity of these
effluents to algae (Selenastrum capricarnutum) and invertebrates (Daphnia
magna). The test data clearly show that the higher toxicity in influents
versus effluents is not due to glyphosate itself but other wastestream
components, many of which are effectively treated. The toxicity of
glyphosate was found to be dependent on water quality parameters such as .
calcium and magnesium concentration; toxicity decreased as water hardness
increased. Glyphosate was found to be more toxic in soft water than
similar concentrations of glyphosate in the effluents from biological
treatment systems. The decreased toxicity in effluents is probably
related to the addition of calcium in the lime pretreatment step.
Additional treatment options were investigated in an attempt to
reduce glyphosate concentrations in the biologically treated effluents.'
These options included ozonation, adsorption, and ion exbange. These
treatment options provided only marginal reduction of glyphosate concen-
trations in biologically treated effluents.
-------
SECTION 3
SELECTED LITERATURE REVIEW
GENERAL CHEMICAL INFORMATION
Chemical Name
N-(phosphonomethyl) glycine
CAS Number
1071-83-6
Synonyms and Trade Names
Roundup®. MON0573
Structure
0 0
II N
HO - C - CH2 - N - CH2 - P - OH Mol. wt. 169.09
H OH
Chemical Properties
Solubility in water is about 1.9 percent by weight. The pKa's for
the phosphonic acid moiety are <2.0 and 5.6. The pKa for the carboxylic
group is 2.6 and for the amine is 10.6 (Sprankle et al. 1975).
Introduction and Uses
Introduced in 1971 by Monsanto as a postemergent herbicide (Baird
et al., 1971). Registered with the Environmental Protection Agency for
control of annual and perennial weeds before the emergence of agronomic
plants (Folmar et al., 1971). Also effective in controlling ditch bank
vegetation (Comes et al. 1976).
Production
The methods and reactions for production of glyphosate are proprietary.
Production basically involves synthesis and isolation of an intermediate.
The intermediate is then reacted to give the final glyphosate product.
3
-------
Figure 1 presents a waste generation schematic from the production
of glyphosate.
Aquatic Toxicity
Folmar, Sanders, and Julin (1979) conducted extensive testing of
technical grade glyphosate, the isopropylamine salt of glyphosate, the
§ A
formulated herbicide Roundup , and the Roundup Herbicide surfactant on
four aquatic invertebrates and four fish species. All tests were con-
ducted in soft water (40 mg/L as CaCO.). Table 1 summarizes some of
their findings. Generally it was found that technical glyphosate was
g
considerably less toxic than Roundup or the surfactant. The increased
D
toxicity of Roundup Herbicide is likely due to the presence of the
surfactant in this mixture.. Interspecies variation in response to an
individual compound or mixture was relatively small.
Mammalian Toxicity
The acute, oral U5r0 of the mono (dimethylamine) salt of glyphosate
was found to be 9,800 mg/kg on mixed sex rats (Baird et al. 1971). The
acute skin absorption minimum lethal dose of this compound for mixed-sex
rabbits was greater than 7,940 mg/kg (Baird et al. 1971).
The acute oral H>so for rats of glyphosate AS listed in the Registry of
Toxic Effects of Chemical Substances (NIOSH), 1977) is 4,320 mg/kg.
Bababunmi, Olorunsogo, and Bassir (1978) studied the toxicology of
glyphosate in rats and mice. They found acute oral LD_. values of
4,873 mg/kg and 1,568 mg/kg in rats and mice, respectively. The intra-
peritoneal levels were 238 mg/kg for rats and 134 mg/kg for mice.
Fate in Soil
Glyphosate was found to be readily degraded in the soil (Sprankle,
Meggitt, and Penner, 1975). These investigators found that 17 to 45
14 14
percent of C-glyphosate was evolved as CO. in 28 days from various
soil types. Sodium azide treatment of the soil reduced the rate of
degradation of glyphosate indicating that at least part of the degradation
was microbial. Sprankle et al., (1975) also found that glyphosate was
readily bound to clay and organic matter with a corresponding limited
mobility.
-------
GLYPHOSATE COMPLEX
INTER -
MEDIATE
WASH
GLYPHOSATE
WASH
1 LIMEPRECIP.
1
SPILLS
LEAKS
WASHDOWNS
DISTILLATE
RECEIVER
SPI
Lit
WA
SOLIDS TO
LANDFILL
ACID
BASE
INTERMEDIATE
SUMP
COLLECTION TANK
ACID BASE
J L
I NEUT.TANK
TECH. SUMP
EQUALIZATION
BASIN NO. 1
EQUALIZATION
BASIN NO. 2
f
OTHER
PLANT
WASTE
1
AERATED
LAGOON
«UMP
» RIVER
Figure 1. Wastewater production and treatment schematic for glyphosate
manufacturing facility.
-------
TABLE 1. AQUATIC TOXICITY DATA FOR GLYPOSATE AND ASSOCIATED COMPOUNDS
Organism
Daphnia magna
(daphnia)
Gammarus pseudolimnaeus
(scuds)
Chironomous plumosus
(midge larvae)
Sal mo gairdneri
(rainbow trout)
Pimephales promelas
(fathead minnows)
Ictalurus punctatus
(channel catfish)
Lepomis macrochleus
(bluegills)
Tempera-
ture (°C)
22
12
22
12
22
22
22
Roundup
LC50 or
24h
8.3
2.4
13
6.4
® Herbicide Glyphosate
EC50 (mg/L) LC50 or EC50 (mg/L)
48h 96h 24h 48h 96h
3.0
62 43
18 55
8.3 140 140
2.3 97 97
13 130 130
5.0 150 140
Surfactant
LC50 or EC50 (mg/L)
24h 48h 96h
13
2.1 2.0
1.4 1.0
18 13
3.0 3.0
-------
Rueppel, Brightwell, Schaefer, and Marvel (1977) investigated the
14
degradation of C-labeled glyphosate and aminomethylphosphonic acid,
its major soil metabolite, in soil samples. Anaerobic, aerobic and
sterile conditions were studied. Under sterile conditions less than
14
1 percent of the C was released in seven days indicating that chemical
degradation is not a significant mode of degradation. Under anaerobic
14
and aerobic conditions from 30 to 50 percent of the total glyphosate C
14
was evolved as CO- in 28 days depending on soil type. Depending on
soil type the remaining activity could be accounted for primarily as the
major metabolite or as soil bound nonextractable residues. Experiments
14
with C-labeled aminomethylphosphonic acid have demonstrated that 34.8
14 14
and 16.1 percent of the metabolite C was evolved as CO. in 63 days
with two soil types. Slower degradation of aminomethylphosphonic acid
than glyphosate may reflect tighter binding to soil and/or lower cell
permeability of aminomethylphosphonic acid.
-------
TABLE 2. PESTICIDE MANUFACTURING WASTEWATER CHARACTERIZATIONS
Parameter
pH
Cl", mg/L
Acidity, mg/L as CaC03
Total Kjeldahl
nitrogen, mg/L as N
Ammonia, mg/L as N
Oxidized nitrogen mg/L
as N
Total phosphorous, mg/L
Orthophosphate , mg/L
Chemical oxygen demand
mg/L
Total dissolved solids,
mg/L
Total solids, mg/L
Formaldehyde mg/L
Glyphosate mg/L
Glyj>hosate mg/L
intermediate
Total organic carbon
mg/L
Intermediate
wash
<1.00
82,430
190,000
5,690
<0.5
0.6
18,550
1,810
77,000
(15,100)*
present
—
(27,000).
(44,000)
Glyphosate
wash
2.45
4,370
56,250
8,030
213
8.4
14,325
1,060
49,250
...
(81,000)
present
20,300
(20,400)
(27,nno)
Glyphosate
distillate
receiver
2.8
5.5
3,402
4.0
2.0
0.7
2.0
0.4
6,500
...
present
16**
(200)**
f3,100)
*Numbers In parenthesis were furnished by Monsanto.
**Discrepancy discussed on page 10.
-------
and carbon. The COD test typically underestimates the TOC data due to
the less efficient oxidation of some compounds by the COD procedure.
It should be expected from the data provided in Table 2 that all
three wastestreams would provide a substantial challenge for typical
treatment systems. The presence of formaldehyde, glyphosate, glyphosate
intermediate and acidity are potential problems for the biological
treatment of these wastes.
BIOLOGICAL TREATMENT WITH LIME PRETREATMENT AND pH ADJUSTMENT
Three combinations of glyphosate production wastestreams were
tested for biological treatability. The three conditions were designed
to test the treatability of glyphosate and associated compounds in
increasingly complex matrices. These three conditions are discussed
below and a summary is provided in Appendix F for easy references.
Condition I tested the biodegradibility of glyphosate in a rela-
tively simple matrix. Distillate receiver waste was diluted 1 part to 19
parts deionized water. Based on analyses provided by Monsanto of
split samples the influent would contain glyphosate at 10 mg/L and TOC
(COD) at 155(325) mg/L. Subsequent glyphosate analyses of undiluted
distillate receiver waste conducted at RTI gave a lower concentration
than initially indicated (16 mg/L as opposed to 200 mg/L). These results
were confirmed by Monsanto. The initial results apparently reflect an
unknown positive interference which affects one of the two methods used
by Monsanto for the analysis of glyphosate. To compensate for the
reduced levels of glyphosate it was decided on day 35 of biological
treatment to begin spiking Condition I influent with a solution of
technical grade glyphosate (12,300 mg/L as glyphosate) to give a final
glyphosate concentration of 25 mg/L.
Condition II tested a combination of distillate receiver and glypho-
sate wash. To reduce the concentration of glyphosate in the glyphosate
wash an initial lime precipitation treatment was employed. This process
is similar to a system used by Monsanto as noted in Figure 1. Calcium
hydroxide was added to centrifuge wash at a rate of 40 grams per 100 mis
solution which produces a slurry since only a small fraction of the lime
actually dissolves. The addition of this quantity of lime produces a
-------
considerable heat of solution. To duplicate the higher temperatures of
a full scale system, which may improve removal rates, laboratory treatment
was conducted in a water bath at 80°C. Contact time was approximately .
5 min before vacuum filtration through Whatman 1 filter paper. Glyphosate
levels in the glyphosate wash were reduced from 20,300 mg/L to 2,100-
2,600 mg/L. Phosphate levels also showed a large reduction. Orthophos-
phate levels were reduced from 1,060 mg/L to 54 mg/L as P and total
phosphorous levels were reduced from 14,325 to 225 mg/L as P. COD of
the glyphosate wash was reduced from 49,250 to 15,240 mg/L with lime
treatment.
The lime precipitation step will produce a substantial quantity of
solid waste. Currently, Monsanto is treating with lime at a lower rate
with no apparent loss in removal effectiveness. While the quantity of
solid waste would be decreased the actual concentration of glyphosate in .
the solid waste would be increased.
To obtain Condition II influent, 37.5 mis of lime-treatment glypho-
sate wash was added to 962.5 mis of Condition I influent. Influent
glyphosate concentration ranged from 100 to 105 mg/L. COD levels of the
influent averaged 770 mg/L.
Condition III influent tested the treatability of a combination of
distillate receiver waste, lime-treated glyphosate wash, and intermediate
wash. The purpose of Condition III was to test any effect of glyphosate
intermediate and associated compounds on the biodegradability of glypho-
sate. It should be noted that effluent guidelines for the treatment of
wastewaters from the production of pesticide intermediates were not
issued by EPA under the so-called "best practical treatment" (BPT)
standards. The status of pesticide intermediate wastewaters under
future regulations has not been decided. In anticipation of any change
in status it is important to obtain information on the characteristics
of these pesticide intermediates.
The synthesis of glyphosate can be accomplished by several different
methods. The actual method for the commercial synthesis of glyphosate
is proprietary. For this reason the manufacturer was reluctant to reveal
the identity of the intermediate. In addition, a routine method for the
10
-------
analysis of the glyphosate intermediate is not available and could not
be developed within the constraints of this study. Since it would not
be possible to monitor the effectiveness of the lime pretreatment of the
intermediate wash, it was decided to avoid the uncertainty in the inter-
mediate concentration by diluting the intermediate wash with distilled
water to approximate the concentration of intermediate reported by the
manufacturer in the influent to biological treatment. The glyphosate
intermediate concentrations in Condition III were therefore obtained by
calculation based on analyses by the manufacturer rather than independent
analyses.
Given the lack of suitable analytical methods for low concentrations
of glyphosate intermediate, analyses of biologically treated Condition
III effluent were not possible. Various indirect measures could be
taken including the effect of the intermediate on biodegradation of glyphosate,
effect on COD reduction, and effect on toxicity relative to Conditions I
and II.
To obtain Condition III influent 18 mL of intermediate wash were
added to 982 ml of Condition II influent. This dilution was calculated
to give a glyphosate intermediate concentration of 500 mg/L. After
additional discussion with Monsanto personnel and review of effluent
data for the Luling facility, it was decided that 250 mg/L of intermediate
was closer to a typical value which might be expected in the effluent
from glyphosate manufacture. This concentration would allow for a
significant input of intermediate from the "spills, leaks, and washdowns."
From day 5 of biological treatment the input of intermediate wash to
Condition III influent was cut to 9 ml/L. This change reduced the COD
of Condition III influent from 2,150 mg/L to about 1,500 mg/L.
Again it should be noted that Appendix F provides a summary of
these three conditions.
Formaldehyde concentrations of Conditions I, II, and III were lower
than would be seen in any undiluted combination of distillate receiver
waste, glyphosate wash, and intermediate wash. This lower concentration
is meant to reflect the loss of formaldehyde during the approximate four
day residence time in the equalization basins prior to the inlet of the
11
-------
SECTION 4
TREATABILITY STUDIES
CHARACTERIZATION OF WASTEWATER FROM GLYPHOSATE MANUFACTURE
The manufacture of glyphosate results in five wastestreams as is
depicted in Figure 1. Two of these five wastestreams are intermittent
in nature and were not sampled for this study. The two streams which
were not sampled were the spills, leaks, and washdowns from the glyphosate-
intermediate production unit and spills, leaks, and washdowns from the
glyphosate production unit. The three streams sampled for use in the
treatability study include: intermediate washwater from the production
of glyphosate-intermediate; washwater from the recovery of glyphosate
product; and a distillate byproduct from the final glyphosate product
isolation step.
The sampling of these three streams included a three day composite
of glyphosate wash; a two-day composite of glyphosate distillate receiver;
and a one day composite of glyphosate intermediate wash. Over the
sampling period flow rates from the different wastestreams were-fairly
constant and average as follows: intermediate wash, 3.0 gpm; glyphosate
wash, 4.5 gpm; distillate receiver, 112 gpm.
Table 2 presents data on various parameters of these wastestreams.
The numbers in parenthesis represent data supplied by Monsanto on split
samples. The methods of analysis used at RTI are contained in Appendices A
and B. The only large discrepancy in the two sets of numbers was the
measured glyphosate concentration in the distillate receiver. This
discrepancy was resolved in favor of the smaller number. A smaller
discrepancy exists between the total organic carbon (TOC) data and the
chemical oxygen demand (£0D) data. Theoretically the COD should be
32/12 times the TOC which reflects the relative weight of oxygen (0.)
12
-------
biological aeration basin at Monsanto's Luling facility. The formaldehyde
concentration is lower than expected from a simple dilution of glyphosate
effluents which compose about half of the -input to Monsanto's aeration
basin. The actual mechanism of removal has not been confirmed and could
include biological degradation and/or volatilzation. A similar occurrence
was noted during the bench scale treatability studies. Influent to biological
units was mixed daily. At the end of the 18 to 24 hour "shelf life" of the
influent, formaldehyde levels typically dropped 50 to 75%. No formaldehyde
was detected during the entire course of the study in any effluent from the
three treatment conditions (detection limit of 1 mg/L).
Biological treatability of the three combinations of glyphosate waste-
water was tested in bench scale units designed by Swisher (1974, see Appendix C
for details of construction). Two units were run per condition. The units
were innoculated with an activated sludge seed from Monsanto1s biological
treatment system at Luling, Louisiana. The seed was aerated, chilled in a
glass container, enclosed in styrofoam and shipped to RTI by airline. Total
transit time was approximately eight hours.
The initial mixed liquor volatile suspended solids levels (MLVSS) in
the units ranged from 3,700 to 4,400 mg/L. The mixed liquor suspended
solids (MLSS) of the units initially ranged from 10,770 to 11,830 mg/L. The
MLVSS-MLSS ratio averaged about 0.38 which is about half the value of typical
municipal waste treatment plant. This ratio will be discussed more completely
later in the report.
Sludge age of between 15 and 20 days was maintained in Condition II and
III units. Very little sludge was wasted from Condition I due to the long
retention time and correspondence low food to microorganism ratio. Retention
time in all units was initially set at 24 hours. After 34 days Condition II
and III units were extended to 48 hour retention time. The pH of all the
influents was adjusted and buffered between 7.0 and 7.4 by the addition of
0.4 grams NaHC03 per liter of influent and either 10% HC1 or 10% NAOH as
required. Dissolved oxygen was maintained at > 2 mg/L in the aerator section
of the treatment units.
13
-------
Figures 2, 3, and 4 plot influent and effluent COD concentrations
of Conditions I, II, and III, respectively. The Condition I units fed
distillate receiver effluent showed an 89% reduction in COD from an
influent average of 326 mg/L to an effluent value of 35.5 mg/L. The
increase in glyphosate concentration in the influent from <1 to 25 mg/L
appeared to have no effect on COD removal.
During the period in which Condition II units ran at 24 hour reten-
tion time, the units showed a 75% reduction in COD from an influent
average of 770 to an effluent value of 195 mg/L. When these units were
run at a 48 hour retention time, performance deteriorated slightly. COD
removal dropped to 56%. This deteriorating performance reflected in
part a loss of solids from the units. During the 48 hour retention
period, no sludge was intentionally wasted from the aerators. At these
retention times the food to microorganism ratio drops to levels that
favor increased endogenous respiration at the expense of sludge production.
The settling characteristics of the sludge are also increasingly disrupted
to the point that flocculation aids such as cationic polymers may be
required in the sedimentation basins. Since air driven bench scale
biological units have a relatively high energy input per unit volume,
these units with their small settling sections are also more subject to
settling disruptions than full scale systems.
As is noted in Figures 3 and 4, the COD of Conditions II and III
influent showed an increase during the 48 hour retention test. The
cause of this phenomenon is not clear. It is possible that the lime
treatment was not carried out as efficiently during the later part of
the study. Condition I influent, which did not receive centrifuge wash,
showed no comparable increase.
During the period in which Condition III units ran at 24 hour
retention time, COD reduction average 72%. Influent values averaged
1,450 mg/L and effluent values averaged 413 mg/L. COD reduction during
the 48 hour retention time test was 64%. Influent and effluent COD
averages were 1,550 and 557 mg/L, respectively.
Table 3 presents MLVSS data for the biological treatment units.
Condition I units showed a steady drop throughout the study reflecting a
14
-------
BOO
I
O 200
5
u
5
O—O UNIT#1
O—O UNIT #2
INFLUENT
240 81012141010202224262030323430384042444048
TIMCItbyi)
Figure 2. Influent and effluent COD levels for Condition I biological
treatment units.
-------
1WO
O—O UNIT #1
O—O UNIT *2
• INFLUENT
M HIM. RETENTION
-MHRS.RE1
10 » M II
42 44 48 «• M
TIMt Wvyil
Figure 3. Influent and effluent COD levels for condition II biological treatment units.
-------
2000
»
1800
1800
1700
1600
1500
1400
1300
I
J 1200
O
O—O UNIT #1
D—D
A A INFLUENT
1100
iu
O
1000
O
O BOO
i "»
w
0 700
600
600
400
300
200
100
24 MRS. RETENTION
-»*- 48 MRS. RETENTION ••
2 4 6 8 1012141618202224262830323436384042444648
TIME (day*)
Figure 4. Influent and effluent COD levels for
condition III biological treatment units.
17
-------
Day
1
6
13
20
27
34
37
41
48
TABLE 3.
Condition
MLVSS DATA FOR BIOLOGICAL TREATMENT UNITS
I
Unit 1 (g) Unit 2 (g)
4.09
3.36
2.72
2.66
2.14
1.87
1.99
1.71
1.61
3.68
3.22
2.82
2.77
2.14
2.12
1.92
1.68
1.70
Condition
II
Unit 1 (g) Unit 2 (g)
4.36
3.90
4.65
5.02
3.80
3.38
2.81
2.37
2.35
3.99
3.54
3.82
4.28
3.68
3.13
2.75
2.65
2.55
Condition
III
Unit 1 (g) Unit 2 (g)
4.39
4.86
5.55
7.18
6.33
4.76
4.63
4.29
3.75
4.40
4.92
5.37
6.99
6.18
4.65
4.94
4.57
4.10
pattern of endogenous respiration. However, the MLVSS-MLSS ratio improved
to 0.5 percent. The food to microorganism ration .(F/M) initially averaged
0.08 day' which is quite low. With the drop in solids the F/M ratio
increased to 0.2 day . A more efficient way to increase F/M ratio
would be to decrease retention time of the feed solution. However, the
longer retention times were designed to increase the degradation of
glyphosate.
More control over solids levels could be achieved with Condition II
and III units during the 24 hours retention time test. Sludge production
rates were high particularly in Condition III. During the 48 hour
retention time test run, solids production fell off markedly in Condition
II units.
One notable comparison should be made between the solids levels in
Condition II and III units which indicates a significant difference in
response to lime-treated glyphosate wash. Influent to both sets of
units contained a lime-treated glyphosate wash which it can be presumed
was at least saturated with calcium salts. These salts apparently
accumulated in the aerator portion of the biological treatment units.
Eventually, a fine, sand-like, white precipitate collected in the sludge
18
-------
return line of the Condition II units, where it could be removed. No
such precipitate was noted in the Condition III units. Rather these
solids appeared to stay in a type of colloidal suspension, imparting a
lighter hue to the sludge.
Suspended solids in Condition III units increased from initial
levels of 10,000 to 11,000 mg/L to concentrations of about 20,000 mg/L
when corresponding MLVSS levels reached 7,000 mg/L. Increased wasting
of solids brought the MLSS and MLVSS levels back down to about 15,000
and 4,000 mg/L, respectively. Despite the increased sludge wastage the
MLVSS-MLSS ratio in the Condition III units continued at * 0.37. The
ratio for Condition II units on the other hand increased to 0.5 with the
selective removal of calcium precipitates.
Increased sludge age which is typical of long feed retention times
and low F/M will compound the problem of the accumulation of nonbio-
degradable, inorganic solids. The danger of this accumulation is that
inorganic solids levels can reach a level that interferes with biological
activity and the treatment of the waste. In the present example, the
high level of calcium salts might also provide a positive interference
in the determination of MLVSS. In the final ignition step of this
determination which occurs at 550°C (APHA, 1975), CO. can be lost during
the following reaction:
CaC03 •+ CaO + CC^t
Under most circumstances the solubility of CaCCL, is such that the
interference would not be great. However, in the case of Condition III
the calcium salt concentration in suspension may be on the order of
grams per liter which could cause a significant overestimation of the
amount of viable solids in the aerator.
Oxygen consumption levels in the three sets of units was fairly
constant. Oxygen consumption corrected for solids levels and expressed
as mg 0. per gram solids per hour is shown in Table A.
It can be concluded from the operational and performance data
presented above that the biological treatment units were operating well
given the wastewaters under treatment. It could then be concluded that
glyphosate concentrations contained in the influents did not interfere
19
-------
with the biological treatment process. On the other hand, glyphosate
itself did not show a pattern of ready degradation. Glyphosate levels
in the effluent ranged from 56 to 68 mg/L in Condition II during the 24
hour retention time test (100-105 mg/L in the influent). Condition III
effluent glyphosate concentration ranged from 59 to 72 mg/L during the
24 hour retention time test. Condition I effluent values ranged from 10
to 23 mg/L. During the 48 hour retention time test with Condition II
and III there was only minimal reduction in glyphosate concentration
after treatment. Glyphosate values for Condition II and III effluents
varied from 92 to 100 mg/L.
TABLE 4. OXYGEN CONSUMPTION IN BIOLOGICAL TREATMENT UNITS
Day
6
12
19
27
43
Condition I
mg/g/hr
3.21 4.29
4.41 4.26
3.61 4.55
5.89 5.33
3.96 3.32
Condition II
mg/g/hr
4.15 2.71
3.10 3.46
3.59 4.35
6.00 7.01
5.87 4.05
Condition III
mg/g/hr
6.54 5.73
3.89 3.91
4.35 3.86
4.83 4.85
6.35 6.80
It is not clear whether the reduction in glyphosate concentration
during biological treatment was a function of biological degradation or
a physical-chemical sorptive process. In order to distinguish between
these two mechanisms an oxygen uptake study was conducted with sludge
spiked with two different concentrations of glyphosate. Sludge from a
Condition II unit was aerated for 15 minutes to bring oxygen level to
saturation. To 31 mL of sludge from unit 1 was added 1 mL of 125 mg/L
glyphosate solution. To 31 mL of sludge from unit 2 was added 1 mL of
6,000 mg/L glyphosate solution. As a control for each test condition, 1
mL of deionized water was added to 31 mis of sludge from the corresponding
20
-------
units. After addition of test or control solutions, the test vessel was
sealed with an oxygen probe (Yellow Springs Instrument probe and meter)
and stirred with magnetic stir bar at a constant rate. Oxygen readings
were taken until the oxygen levels were reduced to 20% of initial levels.
Little differences was seen in the oxygen consumption rate of either the
test units or their respective controls. Oxygen consumption rate of the
unit fed 125 mg/L glyphosate was 0.1944 og/L/ Din and its control was
0.1910 mg/l/min. The rate for the unit fed 6,000 mg/L glyphosate solution
was 0.1498 mg/L/min and its control was 0.1530 mg/L/min. While the test
does not provide any evidence for metabolic uptake of glyphosate, it is
also interesting to note that fairly high concentrations of the compound
do not inhibit other microbial processes in acclimated sludge.
If sorbtion is the major mechanism for removal, an explanation may
be possible for the lack of reduction during the 48 hour retention time
test. Because the sludge production was low with a corresponding low
rate of wastage the solids may have become saturated with respect to
glyphosate. As a result little sorption of glyphosate would occur and
glyphosate would instead pass through the units.
TOXICOLOGICAL EVALUATION OF GLYPHOSATE WASTEWATERS BEFORE AND AFTER
TREATMENT
Toxicological testing of the various influents and effluents was
conducted with the algal species Selenastrum capricornutum and the
invertebrate Daphnia magna. The methods of testing are detailed in
Appendices D and E.
An overall pattern of diminished aquatic toxicity after biological
treatment was exhibited by all three conditions. As shown in Table 5,
TABLE 5. EFFECT OF CONDITION I INFLUENT ON ALGAL GROWTH
Condition I Influent
(mL/L)
Growth (% of control
on day 14)
0.01
90
0.1
89
1.0
90
10 100
58 ' 0.06
21
-------
the algal EC_Q (concentration effective in reducing growth to 50% of
controls) of Condition I influent was approximately 10 mL/L. Algal
assays on Condition I influents and effluents occurred prior to time
when influents were spiked with glyphosate. Glyphosate concentrations
were less than 1 mg/L. Condition II influent showed a toxicity pattern
very similar to Condition I. As shown in Table 6, the EC., of Condition
II was also approximately 10 mL/L.
All effluents showed reduced toxicity compared to influents. As
shown in Tables 7 and 8, the EC_0 for both Condition I and II effluents
was between 560 and 1,000 mL/L. The similarity in toxic response of
algae to Condition I and II influents and effluents would seem to indicate
that the presence of glyphosate in concentrations of 60 to 105 mg/L has
little effect on algal growth. As will be discussed below, the toxic
effects of glyphosate are somewhat dependent on water quality parameters
such as calcium, magnesium, and carbonate concentrations.
Condition III influent was more toxic than either Condition I or
II. As shown in Table 9, the EC5Q for Condition III influent was between
1.0 and 10 mL/L. It should be noted that the intermediate wash component
of Condition III was not lime treated as normally carried out during
manufacture of glyphosate. Table 10 shows that EC.Q for Condition III
slightly more toxic that Condition I and II effluents. The toxicity of
Condition III effluent was significantly reduced as compared to untreated
influent. As noted above with Condition I and II, glyphosate does not
appear to be the principal reason for the influent toxicity since its
concentration is only marginally reduced in the effluent.
Toxicity testing with daphnia showed a pattern similar to the algal
assays. As shown in Table 11 the LC_. (concentration which is lethal to
50% of the test population) of Condition I influent was between 32 and
56 mL/L. For Condition II the LC_Q was between 18 and 32 mL/L as shown
in Table 12. Biological treatment eliminated the short-term toxicity of
Condition I and II effluents. Undiluted effluents from these two conditions
showed no short term toxicity to Daphnia magna.
22
-------
TABLE 6. EFFECT OF CONDITION II INFLUENT ON ALGAL GROWTH
Condition I Influent
(inL/L)
Growth (% of control
on day 14)
TABLE 7. EFFECT OF
Condition I Effluent
(mL/L)
Growth (% of control
on day 14)
TABLE 8. EFFECT OF
Condition II Effluent
(mL/L)
Growth (% of control
on day 14)
TABLE 9. EFFECT OF
Condition III Influent
(mL/L)
Growth (% of control
on day 14)
0.1 1.0 10
88 95 61
CONDITION I EFFLUENT ON ALGAL GROWTH
56 100 180 320 560
98 84 65 60 58
CONDITION II EFFLUENT ON ALGAL GROWTH
56 100 180 320 560
107 83 67 58 74
CONDITION III INFLUENT ON ALGAL GROWTH
0.1 1 10
105 120 0.7
100
0.43
1,000
8.1
1,000
3
100
0.8
TABLE 10. EFFECT OF CONDITION III EFFLUENT ON ALGAL GROWTH
Condition III Effluent
(mL/L)
Growth (% of control
on day 14)
100 180 320 560
73 64 61 19
1,000
0.39
23
-------
TABLE 11. EFFECT OF CONDITION I INFLUENT ON DAPHNIA MAGNA SURVIVAL
Condition I Influent
oL/L 10 18 32 56 100 control
% Survival at 48 hours 90 100 80 0 0 90
TABLE 12. EFFECT OF CONDITION II INFLUENT ON DAPHNIA MAGNA SURVIVAL
Condition II Influent
ml/L 10 18 32 56 100 control
% Survival at 48 hours 100 100 40 0 0 100
The LC-0 of Condition III influent was between 32 and 56 niL/L as
shown in Table 13. The LC-- of Condition III effluent was between 560
and 1,000 riL/L as shown in Table 14.
TABLE 13. EFFECT OF CONDITION III INFLUENT ON DAPHNIA MAGNA SURVIVAL
Condition III Influent
ml/L 10 18 32 56 control
% Survival at 48 hours 100 100 100 0 100
TABLE 14. EFFECT OF CONDITION III EFFLUENT ON DAPHNIA MAGNA SURVIVAL
Condition III Effluent
mL/L 180 320 560 1,000 control
% Survival at 48 hours 100 100 100 0 90
24
-------
Because Condition III effluent most closely represents the total
effluent from production of glyphosate, it was decided to run a daphnia
reproduction bioassay with this effluent to further test the implication
of the discharge of this type of wastewater on aquatic organisms. The
details of this bioassay are presented fully in Appendix E.
Because of difficulties with the controls this bioassay did not
meet the criteria for absolute validity. However, the results highly
suggest that the wastewater is relatively low in toxicity. Daphnia
raised in 30% concentrations of wastewater showed a reproduction rate
higher than the rate for the typical brood stock raised in our laboratory
under optimal conditions. Daphnia raised at 50% concentrations showed a
reproduction rate approximately 34% that of the daphnia raised in 30%
concentrations. Daphnia raised in 75% concentrations of Condition III
effluents showed good survival over a two week period but almost no
reproduction.
As noted in Appendix E, all daphnia assays are conducted in water
with a hardness of 160 to 180 mg/L as CaCO.. Because of the ionic
character of the glyphosate molecule, it was hypothesized that toxicity
would vary with water hardness. To test this hypothesis separate bioassays
were conducted with reference grade glyphosate (99+% purity) in both
hard and soft water. Appendix E contains the formula for reconstituting
deionized water as either hard or soft water.
As can be seen in Table 15 the LC_. of glyphosate in soft water was
between 32 and 56 mg/L and I-Cc0 in hard water was greater than 100 mg/L.
The results for soft water correspond closely to results obtained by
Folmar et al., (1979) for the effect of technical grade glyphosate on
the invertebrate Chironomous plumosus.
TABLE 15. COMPARISON OF TOXICITY OF GLYPHOSATE TO DAPHNIA
MAGNA IN HARD AND SOFT WATERS
Glyphosate concentration mg/L 100 56 32 18 10 0
% Survival in hard water at
48 hours 100 100 100 100 100 100
% Survival in soft water at
48 hours 0 0 100 100 100 100
25
-------
Algal assays were also conducted to test the effect of glyphosate
on Selenastrum capricornuturn. As shown in Table 15 the EC., of glyphosate
in pure solutions is approximately 10 mg/L.
At least two theories have been developed to account for the decrease
in toxicity of many ionic compounds with increasing water hardness. In
one theory calcium complexes with the toxicant rendering it less available
to the algal cell. According to a second theory carbonate may compete
with the toxicant for available sites on the cell surface. In either
case, it is interesting to note than in the successive dilutions of the
test solutions for the algal assay, as the toxicant is being diluted, so
also are the competing elements being diluted. This may account for the
phenomenon depicted in Tables 7, 8, and 10. As test solution is diluted,
growth seems to plateau at a rate less than control. It may be that in
this middle range of dilution values the actual toxicant concentration
the organism "sees" is constant.
Several conclusions relevant to the discharge of glyphosate production
wastewater can be drawn from the toxicological testing presented above.
Glyphosate in soft water appears to be more toxic than glyphosate in a
biologically treated wastestream. This decreased toxicity in wastewaters
is likely due to the addition of the lime-treated glyphosate wash to the
total effluent mixture with a resulting increase in calcium carbonate in
the effluent. Toxicity of glyphosate in production wastewater could
vary with the characteristics of the receiving water. The toxicity of
glyphosate could be expected to be less in harder receiving waters.
SUMMARY OF BIOLOGICAL TREATMENT RESULTS
A review of the three treatment conditions indicates that the
glyphosate concentration exists rather independently of biological
treatment. Distillation receiver wastes, as represented by Condition I,
were easily biotreated with regard to most organics present with the
exception of glyphosate. The effect of increased concentration should
be investigated. Higher concentrations of glyphosate in Condition II
were not significantly degraded but did not interfere with biodegradation
of other organics. Condition III represents a mixture which is qualitively
quite similar to a total production wastestream. While the glyphosate
26
-------
intermediate concentration was similar to that in the influent of Monsanto's
system this concentration was obtained by dilution rather than lime
treatment. The Condition III mixture, with the addition of intermediate,
was in large part readily biotreatable, though the glyphosate concentration
was only partially reduced as in Condition II. Biological treatment of
all three conditions substantially reduced toxicity of effluents relative
to influents.
It would, however, be difficult to define a typical daily or even
weekly flow from the production of glyphosate. The real source of
variability in any treatment scenario is the input glyphosate and its
intermediate from "spills, leaks, and washdown". A study of wastestream
discharge records indicates that these contributions could be substantial.
It may well be that the most effective form of treatment would be to
minimize this input.
ADDITIONAL TREATMENT OPTIONS
Several other treatment options were investigated in the search for
methods to reduce glyphosate concentration. These options included:
ozonation, adsorption (resin and carbon), and ion exchange.
Ozonation
Theoretically ozone could be used as the sole method of treatment
for many industrial wastewaters. However, the cost of complete ozone
oxidation of waste organics in most mixed wastestreams makes this treat-
ment infeasible.
Ozone is more economically applied in combination with other treat-
ment methods. In its well-studied role as drinking water oxidant-
disinfectant ozonation is used as a final polishing step after the major
removal of organics and pathogens has occurred through coagulation,
flocculation, and sedimentation.
Ozonation has been used in conjunction with activated carbon treatment.
Preliminary ozonation is said to improve activated carbon performance.
The mechanism of improvement includes possible increased adsorbitivity
of oxidized organics and increased biological degradation of adsorbed
organics at the carbon surface (Guirguis et al., 1978; Randtke, 1978).
27
-------
Ozonation has also been used to increase the biodegradability of
complex industrial vastewaters. Several highly toxic and refractory
textile dye waste effluents showed increased biodegradability after
ozonation (Netzer and Miyamota, 1975). Ozone pretreatment may be something
of a two-edged sword since it was found that several more biodegradable
textile waste effluents became less treatable after ozonation. This may
be explained by the fact that the more degradable compounds were oxidized
preferentially by the ozone (Netzer and Miyamota, 1975).
Oxidation of pesticides in pure solutions by ozone with and without
ultraviolet (UV) irradiation has been studied (Prengle and Mauk, 1978).
Ozone used in conjuction with UV increased the destruction of the pesticide
when compared to ozonation alone. Unfortunately it is difficult to make
predictions on the oxidation of pesticides in complex mixtures from data
on pesticides in pure solutions.
Two studies of the ozonation of glyphosate were conducted. The
first study investigated the oxidation of glyphosate in a pure solution.
One liter of technical grade glyphosate solution with a measured glyphosate
concentration of 44.6 mg/L was placed in a one liter graduated cylinder.
Ozone, supplied by a Wellsbach T-23 ozone generator, was pumped through
Teflon tubing and a stainless steel diffuser placed at the bottom of the
cylinder. Ozone production rate was 42 mg/minute as standardized with a
KI solution. Twenty ml aliquots were withdrawn every 30 minutes and
analyzed for glyphosate.
In the apparent presence of a constant or excess concentration of
ozone destruction of glyphosate followed first order kinetics with
respect to glyphosate concentration. Figure 5 presents a plot of log
concentration versus time for the destruction of glyphosate. Table 16
presents the data from which this plot was taken.
TABLE 16. OZONATION OF GLYPHOSATE
Time (hour) Glyphosate mg/L
0 44.6
0.5 27.0
1.0 15.3
1.5 6.5
2.0 3.5
3.0 0.0
28
-------
The time constant K for the reaction, as computed from the slope
of -0.0095, was 0.0219 min" . The half-life of glyphosate under these
conditions was about 32 minutes. Such a long half life does not give a
good prognosis for economical ozonation of this compound.
To test the oxidation of glyphosate in a complex matrix one liter
of Condition III effluent was ozonated under the same conditions presented
in the previous experiment. The original concentration of glyphosate
was 92 mg/L. After 2 hours of ozonation glyphosate concentration was
reduced to 80 mg/L. It can be concluded that ozonation does not provide
a feasible alternative for the reduction of glyphosate. Whether ozonation
might increase the biodegradability of the residual organics in the
biologically treated effluents from glyphosate manufacture could not be
tested within the time constraints of this project.
Adsorption
Due to the ionic character of the glyphosate molecule, carbon
adsorption does not provide a feasible alternative for the removal of
glyphosate from waste streams. On the contrary it is reported that
powdered activated carbon is used during sample "clean up" prior to
analysis for glyphosate without apparent effect on glyphosate recovery
(M. Rueppel, 1980).
To test for adsorption of glyphosate on various synthetic adsorbents,
batch tests were run with the absorbents XAD-4, XE-340, and XE 348 as
manufactured by Rohm and Haas. Of the absorbents used here, XE 348 is
most similar to activated carbon (Rohm and Haas, 1977).
Each resin was soaked in methanol for 30 minutes and then rinsed
well with several volumes of water. To 100 ml aliquots of filtered
(Whatman 2V) Condition III effluent were added either 2 ml or 0.2 ml of
each of the absorbents. In addition, a blank was carried through all
steps of the procedure. All samples were shaken at 110 oscillations per
minute for 90 minutes. At the end of this period the resin was allowed
to settle and the supernatant decanted and analyzed for glyphosate.
Table 17 presents results of this study. Results of these batch tests
indicate relatively little adsorption of glyphosate on these synthetic
adsorbents. It should also be noted that batch tests typically over-
estimate the capacity of large scale column systems.
29
-------
2.0
TJMEJmln)
0
30
60
90
120
logC
1.65
1.43
1.18
0.81
0.54
O
c
1.5
O
o
1.0
0.5
I
30
60
TIME (mini
90
120
Figure 5. Rate of ozonation of glyphosate.
-------
TABLE 17. ADSORPTION STUDY WITH CONDITION III EFFLUENT
Adsorbent Glyphosate (mg/L)
XAD-4
2.0 ml 38.0
0.20 ml 40.2
XE-340
2.0 ml 37.6
0.20 ml 39.0
XE-348
2.0 ml 36.9
0.20 ml 37.4
Blank 45.3
Ion Exchange
Considerable work has been done at RTI on the use of ion exchange
resins to concentrate various compounds from aqueous solutions for
subsequent analysis. It has been found that the concentration efficiency
of ion exchange resins is highly dependent on the ionic strength of the
solution from which the compound of interest is to be isolated. A
compound such as glyphosate will be competing for available sites to
some degree with the other ions in solution. In general, the exchange
capacity of an ion exchange resin and the ionic strength of the water
sample determine the volume of water which can be treated effectively.
By taking conductivity measurements of various samples a prediction can
be made of the volume of effluent which can be passed through an ion
exchange column before breakthrough. Table 18 presents empirically
determined relationships between conductivity and breakthrough volume
for various types of effluents passed through an open tubular chroma-
tography column filled with 10 ml of Biorad A6 1-X8 resin (Pellizzari,
1979). 60 mis of 0.67N NaHSO, in acetonitrile:water (2:1) were required
to regenerate a column of this size. The elution volume of 0.67 N
NaHSO, in pure water would be larger. In an ion exchange the treatment
scenario this elution volume would depend upon the mineral acid and its
31
-------
concentration. In practice H^SO, at concentrations of 1-2 M might be
used.
The conductivity of Conditions I, II, and III effluents was measured
on an Industrial Instruments Conductivity bridge, model RC16BZ. The
conductivity readings in pmhos/co of the biologically treated effluent
from the three conditions were: I, 940; II, 1,620; III, 5,640. Based
on the results presented in Table 18 ion exchange treatment of Conditions
III and II is infeasible. With regeneration of the resin with a strong
mineral acid and no degradation of the glyphosate, the waste disposal
problem would be worse unless large concentration factors can be realized.
Since it is clear that Conditions II and III will experience small, if
any, concentration factors, ion exchange has little applicability to
these waters.
TABLE 18. ESTIMATED IONIC EQUIVALENTS AND BREAKTHROUGH
VOLUMES OF TEST WATERS
Water type
distilled-deionized
drinking water
municipal effluents
energy effluents
1
2
3
4
5
6
7
8
9
10
Conductivity
(pmho/cm)
0.5
136
485
1,530
25
2,720
11,000
2,100
2,700
3,470
1,570
3,830
2,210
Volume before
breakthrough (mL)a
-
2,500
500
166
8,333
100
25
125
100
83
166
70
125
Ion exchange columon filled with 10 mL of Eiorad AG 1-X8 resin (1.4 meq/mL),
32
-------
References
1. APHA, AWA, WPCF. Standard Methods for the Examination of Water
and Vastevater. 14th ed. American Public Health Association,
Washington, D.C., 1975.
2. Baird, 0. D., R. P. Upchurch, W. B. Homesley, and J. E. Franz.
Introduction of a New Broadspectrum Postemergence Herbicide Class
with Utility for Herbaceous Perennial Weed Control. In: Proceedings
of the North Central Weed Control Conference. 26:64-68, 1971.
3. Bababunmi, E. A., 0. 0. Olorunsogo and 0. Bassir. Toxicology
of Glyphosate in Rats and Mice. Toxicology and Applied Pharmacology,
45:319-320, 1978.
4. Comes, R. D., V. F. Bruns, and A. D. Kelley. Residues and Persistence
of Glyphosate in Irrigation Water. Weed Sci. 24:47-50, 1976.
5. Duke, K. M., M. E. Davis, and A. J. Dennis. IERL-RTP Procedures
Manual: Level 1 Environmental Assessment Biological Tests for Pilot
Studies. EPA-600/7-77-043, U. S. Environmental Protection Agency,
Research Triangle Park, North Carolina, 1977. 106 pp.
6. Folmar, I. C., H. 0. Sanders, and A. M. Julin. Toxicity of the
Herbicide Glyphosate and Several of its Formulations to Fish and
Aquatic Invertebrates. Arch. Environm. Contain. Toxicol. 8:269-278,
1979.
7. Guirguis, W., T. Cooper, J. Harris, and A. Ungar. Improved
Performance of Activated Carbon by Preozonation. Journal Water
Pollution Control Federation, 50:308, 1978.
8. Miller, W. E., J. C. Greene, T. Shiroyama. The Selenastrum
Capricornutum Prints Algal Assay Bottle Test: Experimental Design,
Application, and Data Interpretation Protocol. EPA-600/9-78-018,
U. S. Environmental Protection Agency, Corvallis, Oregon, 1978.
126 pp.
9. Ketzer, A., and K. H. Miyamoto. The Biotreatability of Industrial
Dye Wastes Before and After Ozonation and Hypochlorination-Dechlor-
Ihatlon. In: Proceedings of the Thirtieth Industrial Waste Con-
ference, Lafayetta, Indiana, 1975. pp. 804-815.
10. NIOSH. Registry of Toxic Effects of Chemical Substances, U. S.
Department of Health, Education and Welfare, Rockville, Maryland
1977.
11. Pellizzari, E. D. Master Scheme for the Analysis of Organic
Compounds in Water. Technical Progress Report No. 8, EPA Contract
No. 68-03-2704.
33
-------
12. Prengle, H. W., and C. E. Mauk. Ozone/UV Oxidation and
of Pesticides in Aqueous Solution. In: Ozone/Chlorine
Dioxide Oxidation Products of Organic Material, R. G. Rice
and J. A. Cotruso, eds. Ozone Press Int., Detroit, Michigan
1978.
13. Randtke, S. J., Discussion. Journal Water Pollution Control
Federation. 50:2602-2603, 1978.
14. Rohm and Haas, Ambersorb Carbonaceous Absorbents. No. IE-231.
Philadelphia, PA, 1977. 20 pp.
15. Rueppel, M. L., B. B. Brightwell, J. Schaeffer, and J. T. Marvel.
Metabolism and Degradation of Glyphosate in Soil and Water.
Agricultural and Food Chemistry, 25(3):517-528, 1977.
16. Rueppel, M. Research Manager, Monsanto, Inc. Personal
Communication, January 31, 1980.
17. Sittig, M. Pesticides Process Encyclopedia. Noyes Data
Corporation, Park Ridge, N.J., 524 pp. 1977.
18. Sprankle, P., W. F. Meggitt, and D. Penner. Adsorption
Mobility and Microbial Degradation of Glyphosate in the
Soil. Weed Sci. 23;229-234, 1975.
34
-------
APPENDIX A
ANALYTICAL METHOD FOR DETERMINATION OF GLYPHOSATE
The procedure for the analysis of glyphosate involves derivatization
of glyphosate with 4-chloro-7-nitrobenzo-2-oxa-l,3-diazole (NBD-C1),
followed by separation and quantitation of the derivative by
high-performance liquid chromatography (HPLC).
The derivatization procedure is a modification of the method of
Lawrence and Frei (1972). NBD-C1 reacts with alkyl amines in aqueous
solution to form the NBD-amine derivative:
0
0
OH-C-CH0-NH-CH,-P-OH
2 2 |
OH
glyphosate
9
N
,CH2- P-OH
V °H
CH2-COOH
NBD-glyphosate
0.2 ml of 1% NBD-C1 in MeOH 2nd 0.2 ml 0.1 N NaHC03 were added to 0.2 ml
vastewater in a small test tube. The tube was loosely stoppered, shaken,
and heated in a water bath at 80°C for 30 min. The solution was cooled,
filtered through a Mllllpore LS filter, and a 10 Vl aliquot removed for
analysis.
A modular liquid chromatographic system consisting of the following
components was used for analyses: 2 M6000 pumps with a M660 solvent
programmer and a U6K injector (Waters Associates) and a Model SF-770
variable wavelength ultraviolet detector (Schoeffel Instruments).
35
-------
Paired-ion chromatography, utilizing • reverse phase column packed
with LiChrosorb RP-18 packing (Merck) and a CO:PELL ODS pre-column
(Vbatoan) was used to separate and quantify the NBD derivative of
glyphosate. A 40% MeOH/H.O mobile phase with 0.005 H tetrabutylammonium
hydrogen sulfate added as a counter-ion was used. The pH of the mobile
phase was adjusted to 7.0 with 10% NH.OH. Solvent flow was 2.0 ml/min and
the variable wavelength detector was set at 475 run. Minimum detectable
quantity was 50 ng with a standard deviation of 1.2% at 5 mg/liter.
The chromatograras of Conditions II and III influents showed a large
peak preceding the glyphosphate derivative peak. Complete resolution of the
two peaks was not possible. Concentration of glyphosate were verified by the
method of standard additions. Accurate values were determined by plotting peak
areas of samples spiked with glyphosate to give concentrations 2 and 33 times
the estimated glyphosate concentration. Values for unsplked samples could
then be verified by extrapolation.
36
-------
APPENDIX B
ANALYTICAL PROCEDURES FOR ROUTINE WASTEWATER CHARACTERIZATION
Routine wastewater analyses were conducted according to Standard Methods
for the Examination of Water and Vastewater. 14th Edition, (APHA, AWWA,
WPCF, 1975).
pH-
pH was determined electrometrically by Method 424.
Chloride--
Chloride was measured by the mercuric nitrate method (Method 408 B).
Acidity-
Acidity, as CaCO_, was determined by Method 402.
Alkalinity-
Alkalinity, as CaCO., was determined by Method 403.
Nitrogen Forms--
Total Kjeldahl nitrogen was determined after digestion, according to
Method 42. Ammonia (NH.-N) was determined by an acidimetric method as
described in Sections 418 A and 418 D. Nitrite and nitrate nitrogen (N02-N,
NO -N) were determined by the Devarda's alloy method (419 F).
COD—
Chemical oxygen demand (COD) was determined by Method 508.
Residues—
Suspended solids (SS) were determined by Method 208 D. Total solids
(TS) were determined by Method 208 A. Total dissolved solids (TDS) were
determined by Method 208 A. Settleable solids were determined by Method
208 F.
37
-------
APPENDIX C
PROCEDURES FOR CONDUCTING ACTIVATED SLUDGE TREATABILITY TESTS
For biological treatability studies the bench scale, complete-mix,
continuous-feed, activated sludge unit designed by Swisher (1970) was
employed (Figure C-l). This unit has an aerator capacity of 0.4 L and a
settler capacity of 0.075 L. The unit is made entirely of glass, avoiding
the possibility of contamination by organics leaching from the container.
Continuous feed to the units was supplied through Teflon tubing by a
peristaltic pump to give the nominal retention time of 8 hours.
Routine determinations were made of dissolved oxygen, pR, mixed
liquor volatile suspended solids in the aerator and COD, and pesticides
in the influent and effluent. Dissolved oxygen was determined with an
oxygen probe (Yellow Springs Instrument Co.).
38
-------
w
\0
EFFLUENT
FEED
RESERVOIR
EFFLUENT
COLLECTION
AIR
Figure C-l. Diagram of activated sludge pilot unit.
-------
APPENDIX D
PROCEDURES FOR ALGAL ASSAY TESTS
Algal bioassays were conducted according to the freshwater algal assay
procedure described in the report The Selenastrum Capricornutum Printz Algal
Assay Bottle Test (Miller, Greene, and Shiroyama, 1978). The test alga was
Selenastrum capricornutum Printz, obtained from the National Eutrophication
Research Program, EPA, Corvallis, Oregon. This test was designed to measure
algal response to changes in nutrient concentrations and to determine toxi-
city or inhibition.
Wastewaters to be tested were filter sterilized through a sterile
prewashed membrane filter (Millipore Filter, 0.45 MID pore size). Serial
dilutions were then made in sterile algal media to give the appropriate
final concentration. Sufficient inoculum was added to produce an initial
4
cell concentration of 10 cells/mL.
In each set of experiments, algal growth in the presence of a series of
concentrations of the wastewater added to the nutrient medium was compared
to algal growth in the nutrient medium alone. Growth was determined by
direct counts of the algae during the 10-14 day incubation period. Effect
of the wastewater on algal growth was determined in terms of the effect on
the cell yield. Direct cell counts were performed by an automated procedure
utilizing a Fisher Scientific Model FO 16 particle counter.
The tests were conducted in water bath shakers at 24 + 2°C and at
approximately 80 oscillations per minute with constant white fluorescent
lighting at 4300 lux. Test containers were 250 ml Erlenmeyer Pyrex flasks
containing 100 ml of test medium and covered with an inverted Pyrex beaker.
The method of expressing inhibitory or stimulatory effects was that
recommended by Miller et al. (1978), i.e., as the percent growth inhibition
(I) or stimulation (S), as compared to growth in • control culture without
40
-------
the test materials. These authors suggest that, in general practice, the
results be based on the growth at 14 days, i.e., as % I., or % S., at a
given concentration of the effluent being tested. Typical cell counts for
controls on day 14 of a test were between 5 and 6 million cells per ml.
Decreased growth, compared to the control, is evidence of an
inhibitory effect. The manner in which the test is conducted does not
allow determination of whether this inhibition is temporary (algistatic)
or permanent (algicidal). Such a determination would require further
testing by subculturing into fresh medium free of the test material.
-------
APPENDIX E
PROCEDURE FOR DAPHNIA BIOASSAY TESTS
The 48 hour, static daphnia bioassay was adopted from the procedures
outlined by Duke et a_l (1977). Dapbnia magna was selected as the test
species. An original population of Daphnia magna were obtained from V. T.
Waller of the University of Texas at Dallas. The daphnia used in these
bioassays are reared at Research Triangle Institute under conditions which
promote parthenogenetic reduction. Each member of the breeding stock is
reared in isolation. When reproduction occurs young are removed within 24
hours. Age of the breeding stock is staggered and each daphnia is replaced
every 56 days. All daphnia in the present breeding population are the
offspring of a single daphnid.
Daphnia are reared in deionized water which is reconstituted to
standard bard water specifications by addition of 192 mg NaHCO., 120 mg each
of CaSO, • 2H.O and MgSO, and 8 mg of KC1 per liter of water. Daphnia are
fed daily portions of non-steril Selenastrum capricorntum except during
testing periods.
Range finding bioassays may be conducted with a mixed population of
daphnia. All narrow range, definitive bioassays are conducted with first
instar young.
Tests are conducted In 250 ml beaker to which'100 mis of combined
test solution and reconstituted hard or soft water has been added as
noted in the test. When the bioassay required soft water, deionized
water was reconstituted with the addition of the compounds given above
for hard water at 25Z of the concentration given above for hard water.
Ten daphnia are randomly assigned to each condition. Survival of the
daphnia at 24 and 48 hours is judged by careful examination for mobility.
Survival of daphnia in the control condition must be 90% or better for
bioassay to be considered valid.
42
-------
A static bioassay testing the reproduction of daphnia exposed to
Condition III effluent was conducted. Five daphnia were tested at each
concentration. Three concentrations of Condition III biological treated
effluent were tested. These conditions included 30%, 50%, and 75%
effluent diluted in hard water. In each of 5 beakers per concentration
containing 100 ml of the appropriate mixture was placed one first instar
daphnia. Five daphnia placed in hard water alone were run as controls.
Daphnia were fed 1 ml of Selenastrum Capricornutum every other day.
Feed concentration was 10 cells per ml.
Because of difficulties in raising the controls this bioassay did
not meet all provisions for validity. During the early days of the
assay the controls only exhibited a phenomenon we referred to as "floating".
The daphnid becomes trapped at the surface of the water and is held
there by the surface tension of the water. This phenomenon has never
affected the young of the stock breeding population. This "floating"
phenomenon i*s also quite intermittent and occurs only at widely scattered
intervals in our acute testing program. When it does occur it seems
only to affect the control daphnia and the daphnia in the lowest concen-
trations of test solution while higher non-lethal concentrations are
unaffected. When it does occur, most or all daphnia in these low conta-
minant conditions are affected. The problem is not unique to our labora-
tory. W. T. Waller of the University of Texas at Dallas reports using
screens to keep daphnia below the surface during long term assays to
avoid this phenomenon.
Where this phenomenon occurs during acute testing the bioassay is
repeated. During the reproduction assay it was decided to let this
phenomenon, which affected only the controls, run its course. After
several molts the affected daphnia "recovered" but reproduction was
greatly reduced. In the nineteen days of the assay the five controls
produced only six broods of a total of 45 young. Daphnia tested in 30%
Condition III biologically treated effluent produced 18 broods with a
total of 322 young. Daphnia tested in 50% concentrations produced 110
live young and 7 dead young in 10 broods. Daphnia tested in 75% effluent
showed 80% survival through 14 days and the production of only 2 broods
of 4 live and 3 dead young. By comparison to the reproductive rates of
43
-------
the controls used in this assay, the reproductive rate of 5 first instar
young reared in our breed stock during the same 19 days showed a repro-
duction rate of 210 young in 12 broods.
Despite the difficulties with the controls it is obvious that the
daphnia in 30% conditions thrived. This is probably a function of the
increased nutrients in the form of microbial populations contained in a
biologically treated effluent. Because of difficulties with controls
the interpretation of the results of the 50 and 73% conditions more
difficult. Though the daphnia in these conditions did not experience
the same problems as controls, it might be argued that a general weakness
pervading the daphnia population selected for this assay affected the
resistance of these daphnia to toxicants.
44
-------
APPENDIX F
DATA SUMMARY FOR GLYPHOSATE WASTEWATER TREATMENT CONDITIONS
Condition I: 1:20 distillate receiver waste in deionized water.
Spiked with glyphosate to 25 mg/L from day 35 of biological
treatment. COD of the influent averaged 326 mg/L.
Glyphosate averaged < 1 mg/L to day 35; 25 mg/L after day
35.
Condition II: 37.5 mL of lime pretreated glyphosate wash brought up to
1 liter with Condition I influent. COD of influent
averaged 770 mg/L during 24 hour retention time and 977
mg/L during 48 hour retention time. Glyphosate influent
values were 100 to 105 mg/L.
Condition III: 9 mis of concentrated non-pretreated intermediate wash
were brought up to one liter with Condition II influent.
COD influent average 1450 mg/L during 48 hour retention
time test.
As mentioned in the text the purpose of running three conditions was
to test the degradability of glyphosate at increasing concentrations
in increasingly complex waste mixtures. Based on discussions with Monsanto
personnel and examination of effluent data from the manufacture of glyphosate
the following scheme was worked out. Condition I would test the treatability
of glyphosate at low concentration and in a relatively simple mixture.
Because the information originally available indicated that the distillate
receiver wastewater contained about 200 mg/L it was decided to dilute the
stream 1 part in 19 parts deionized water to give a glyphosate concentration
of 10 mg/L. If glyphosate proved to be biologically degradable at these
concentrations could then be increased. Further analyses indicated lower than
expected glyphosate concentrations and it became necessary to spike Condition I
influent to a glyphosate concentration of 25 mg/L.
45
-------
Conditions II and III were meant to test the degradibility of glyphosate
in more complex mixtures with and without the addition of glyphosate-intermediate.
It was decided to hold glyphosate concentrations initially to 100 mg/L which
was determined to be the minimum possible glyphosate concentration which
could be attained from a glyphosate production system. The glyphosate
wash stream was added at a rate of about 3.7% to Condition I influent. If
the glyphosate and associated waste organics were degraded, the glyphosate
concentration in Condnition II influent could be increased by increasing
the concentration of distillate receiver waste in the Condition I and II
wastewater.
Condition III factors in the influence of glyphosate intermediate on
the degradation of glyphosate and associated organics. While the intermediate
wash is a relatively small flow stream, it is very concentrated stream before
lime pretreatment. Intermediate wash prior to lime pretreatment was added at
a rate of 9 mL per liter of Condition II influent. The concentration of the
glyphosate and the glyphosate intermediate in Condition III is fairly repre-
sentative of what might be expected in a glyphosate production waste stream.
46
------- |