United States
Environmental Protection
Agency
ronmental Research f
'78
Cincinnati OH 45268
Research and Development
c/EPA
Autotrophic
Denitrification
Using Sulfur
Electron Donors
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-78-113
July 1978
AUTOTROPHIC DENITRIFICATION USING SULFUR ELECTRON DONORS
by
Alonzo Wm. Lawrence,
James J. Bisogni, Jr., Bill Batchelor and Charles T. Driscoll, Jr.
Cornell University
Ithaca, New York 14853
Grant No. 803505
Project Officer
E. F. Earth
Wastewater Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. 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.
11
-------
FOREWORD
The Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. The complexity of that environment and
the interplay between its components require a concentrated and integrated
attack on the problem.
Research and development is that necessary first step in problem solu-
tion and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention, treat-
ment, and management of wastewater and solid and hazardous waste pollutant
discharges from municipal and community sources, for the preservation and
treatment of public drinking water supplies, and to minimize the adverse
economic, solcial, health, and aesthetic effects of pollution. This publica-
tion is one of the products of that research; a most vital communications
link between the researcher and the user community-
This report summarizes the results of a feasibility study to determine
if various species of sulfur could serve as substrate for biological denitri-
fication of municipal wastewater effluent.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
iii
-------
ABSTRACT
This research project investigated the feasibility of autotrophic de-
nitrification as a nitrate removal process for municipal wastewater. The
overall objective of this project was to evaluate the microbial kinetics,
and to assess the process performance of autotrophic microbially mediated
denitrification using sulfur electron donors.
This study was divided into three experimental phases. Each phase
utilized a different sulfur compound or flow configuration. Included in these
phases were: continuous flow slurry-type with elemental sulfur as the elec-
tron source; semi-continuous flow, complete-mix reactors with thiosulfate or
sulfide as the electron source; and packed bed columnar reactors with elemen-
tal sulfur as the electron source.
Based on theoretical and experimental considerations, kinetic models
and stoichiometric relationships were developed for the autotrophic denitrifi-
cation process.
The results of this study indicate that autotrophic denitrification with
various sulfur species, particularly elemental sulfur, is a feasible scheme
for removal of nitrate from wastewater effluents.
This report was submitted in fulfillment of Grant No. 803505 by Cornell
University, Ithaca, N.Y. under partial sponsorship of the U.S. Environmental
Protection Agency.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Figures viii
Tables x
Symbols , xii
1. Introduction 1
2. Conclusions 3
3. Recommendations 5
4. Background ,.. 6
5. Theoretical Considerations 8
Development of kinetic model 8
Stoichiometry of autotrophic denitrification 15
General concepts of microbial Stoichiometry .... 15
Thermodynamic based predictions of microbial
Stoichiometry 17
Theoretical prediction of autotrophic denitrifi-
cation Stoichiometry 18
6. Analytical Techniques 22
7. Experimental Studies ,. . . 26
Phase I - Slurry reactors using elemental sulfur. ... 26
Cultural characterization experiments ....... 26
Experimental plan 26
Experimental techniques 27
Experimental results 29
Variable sulfur to nitrate-nitrogen feed
ratio 29
Variable ammonia-nitrogen to nitrate-
nitrogen feed ratio 29
Variable temperature study 31
Batch Stoichiometry 33
Continuous culture experiments 34
Experimental plan 34
Experimental techniques 34
Reactor operations 34
Steady state techniques 38
Transient rate study techniques 39
Solids separation study techniques 40
Experimental results 41
Variable solids retention time 41
Variable sulfur to biomass ratio 41
Effect of temperature. 45
Settling and thickening characteristics. . . 45
-------
Discussion of Phase I experimental results 54
Observed yield coefficients 54
Balanced stoichiometric equation 54
Sulfur balance 56
Composition of reactor gas 56
Kinetics 56
Temperature effects 58
Settling and thickening 59
Phase II - Thiosulfate and sulfide experiments 60
Experiment #1 - effect of growth rate - high C . . . . 60
Experimental plan and techniques 60
Experimental results , 60
Experiment #2 - effect of growth rate - intermediate C , 64
Experimental plan and techniques 64
Experimental results 64
Experiment #3 - determination of consumptive ratio ... 64
Experimental plan and techniques 64
Experimental results 64
Experiment #4 - sulfide experiments 64
Experimental plan and techniques 64
Experimental results 64
Discussion of Phase II experimental results 66
Thiosulfate and sulfide as electron donors 66
Observed yield and consumptive ratio 66
Phase III - Packed bed reactor experiments 67
Experiment #1 - dolomitic limestone reactors - sulfide
feed 67
Experimental plan and techniques 67
Experimental results 70
Experiment #2 - elemental sulfur packed bed reactors . . 70
Experimental plan and techniques 70
Experimental results 70
Experiment #3 - sulfur-dolomite packed bed studies ... 70
Experimental plan and techniques 70
Experimental results 74
Experiment #4 - packed bed studies with domestic waste-
water effluent 75
Experimental plan and techniques 75
Experimental results 75
Discussion of Phase III experimental results 77
Elemental sulfur packed bed performance 77
Alkalinity supplementation with dolomite 80
Denitrification of domestic secondary effluent ... 82
8. Engineering Significance 83
Evaluation of sulfur-substrates 83
Technical feasibility 85
Cost of electron donor 86
Cost of supplemental alkalinity 87
Sludge disposal 87
Environmental impact ..... 88
VI
-------
Summary 88
References , 89
Appendices 95
A. Elemental sulfur analysis, ATP analysis . . 95
B. Transient rate test data, results of zone settling test,
results of flocculent settling test 98
vii
-------
FIGURES
Number Page
1 Schematic of sulfur-biofilm system with-zeroth-order nitrate
rate limitation 9
2 Hyperbolic tangent and Monod functions versus ?(S/X) 11
3 Schematic of slurry reactor system 13
4 Proposed electron flow for autotrophic denitrification under
anaerobic conditions with water the electron donor for
synthesis 19
5 Schematic of cadmium reduction column used for nitrate
analysis 24
6 Schematic of semicontinuous culture reactor system 28
7 Rate of gas production versus sulfur to nitrate-nitrogen feed
ratio 30
8 Rate of gas production versus ammonia-nitrogen to nitrate-
nitrogen feed ratio . 32
9 Schematic of continuous culture reactor system 36
10 Observed biomass yield versus solids retention time . 43
11 ATP content of biomass versus solids retention time 44
12 Maximum attainable unit rate of denitrification versus sulfur
to biomass ratio 47
"\
13 Natural Logarithm of maximum attainable unit rate of denitri-
fication versus inverse of absolute temperature. 49
14 Solids flux versus solids concentration for various values of
the sulfur to biomass ratio 50
15 Suspended solids concentration on depth versus time graph with
lines of isoconcentration • 51
viii
-------
Number Page
16 Effluent suspended solids concentration versus overflow rate
for various values of the sulfur to biomass ratio 52
17 Semicontinuous flow sulfide/thiosulfate reactor 62
18 Response of autotrophic denitrifying system to rapid changes
in feed ratio 65
19 Schematic of continuous flow sulfide feed packed bed reaqtor
system 69
20 Schematic of continuous flow sulfur packed bed reactor system . 71
21 Effluent nitrate concentration as a function of hydraulic
retention time for different particle size reactor media . t 73
22 Effluent quality as a function of hydraulic retention time. . . 78
23 Minimum hydraulic retention time for complete denitrification as
a function of estimated sulfur surface area 79
24 Feed alkalinity versus apparent alkalinity consumption for
sulfur packed bed reactors supplemented with dolomite. ... 81
-------
TABLES
Number
1 Summary of Steady-State Material Balances for Autotrophic
Denitrification in a Slurry Reactor System 14
2 Summary of Analytical Techniques .... 23
3 Summary of Experimental Plan . . . , 27
4 Composition of Media for Semicontinuous Culture Experiments. . 29
5 Effect of Sulfur to Nitrate Feed Ratio on Gas Production
Rate 31
6 Effect of Ammonia to Nitrate Feed Ratio on Gas Production
Rate 31
7 Effect of Temperature on Gas Production Rate 33
8 Batch Stoichiometric Coefficients 34
9 Summary of Experimental Plan for Continuous Culture
Experiments 35
10 Composition of Continuous Culture Feed Solution. , 39
11 Summary of Continuous Culture Results at Various 0 T . . . . 42
12 Summary of Continuous Culture Results at Various S/X 46
13 Summary of Continuous Culture Results at Various Temperatures. 48
14 Regression Equations for Results of Solids Separation Tests. . 53
15 Stoichiometric Coefficients for Batch and Continuous
Cultures , 55
16 Experimental Program - Phase II (Steady-State) 60
17 Phase II - Feed Solution Nutrients 61
-------
18 Summary of Steady-State Experimental Results - Phase II. ... 63
19 Phase III Feed Characteristics (Expts. #1, #2, #3) 68
20 Experiment #1 - Phase III Steady-State Experimental Results. . 70
21 Phase III Experiment #2 Reactor Characteristics 72
22 Phase III Experiment #3 Reactor Characteristics 74
23 Experiment #3 - Phase III Results - Sulfur/Dolomite Reactor. . 74
24 Experiment #3 - Phase III Results-Sulfur Reactor 75
25 Experiment #4 - Phase III Sulfur/Dolomite Packed Bed Reactor
Performance with Secondary Effluent Feed (6 = 27.8 Hrs) . . 76
26 Experiment #4 - Phase III Sulfur Packed Bed Reactor Performance
with Secondary Effluent Feed (0 = 21.1 Hrs) 76
27 A Comparison of Measured and Theoretical Sulfate Production
to Nitrate Reduction Ratio 82
28 Comparison of Costs of Sulfur Substrates and Methanol 84
-------
LIST OF SYMBOLS
A,B,C,D = symbols for generalized chemical compounds
A = surface area of clarifier-thickener, [m ]
ATP = adenosine triphosphate, primary energy storage compound
in cells
c = ratio of elemental sulfur to nitrate-nitrogen feed rate,
[mg/mg].
C = ratio of electron equivalents of electron donor (sulfur
compound) in feed to electron equivalents of electron
acceptor in feed
CR = ratio of electron equivalents of thiosulfate or sulfide
consumed to electron equivalents of nitrate reduced
C ,C_ = concentrations of elemental sulfur and nitrate-nitrogen
within the biofilm, respectively; superscripted with o,
', or 6 to denote evaluation at Z=0, Z=Z', or Z=6,
respectively, [mg/£]
CEN = equivalent nitrate-nitrogen concentration, [mg/£]
DO = dissolved oxygen concentration [mg/&]
E = Arrhenius activation energy, [kcal/mole]
f , f = fraction of electron equivalents in observed reaction
allocated to energy and synthesis subreactions,
respectively
AG = change in Gibbs free energy, [kcal/mole]
AG , AG , AG ,
AG , AG = change in Gibbs free energy measured under standard
conditions in general, for oxidation of one electron
equivalent by energy reaction, for conversion of one
electron equivalent of carbon source to pyruvate, for
oxidation of one electron equivalent of sulfur by oxygen,
and for oxidation of one electron equivalent of sulfur
by nitrate, respectively
AG = ATP-energy required to convert one electron equivalent
of pyruvate to cell material, [kcal/electron equivalent!
k = efficiency factor for microbial energy conversion
K = zeroth-order rate constant for removal of CEN during
transient rate tests, [mg/Z-d]
K = saturation coefficient in function expressing dependence
of unit rate of denitrification on nitrate-nitrogen
concentration, [mg/Jt]
m = coefficient indicating whether energy is released (-1)
or (+1) by reaction which converts carbon source to
pyruvate
xii
-------
M = slope of solids flux versus total solids concentration '
0 curve, [m/d]
N,N = concentration of nitrate-nitrogen in reactor and in
influent, respectively [mg/£]
NI = concentration of nitrite-nitrogen, [mg/Jl]
NFN = function expressing dependence of unit rate of denitri-
fication on nitrate-nitrogen concentration
Q = volumetric flow rate for influent, [H/d]
Q/A = surface overflow rate for clarifier-thickener, may be
superscripted with max to denote the maximum value
attainable under specified operating conditions,
[m3/m2-d]
r = ratio of recycle flow rate to influent flow rate
R = universal gas constant, [kcal/mole- K]
R , R , R = volumetric rates of removal of nitrate-nitrogen and
elemental sulfur, and production of biomass, respectively,
o r eff [mg/£-d]
S, S , S , S = concentration of elemental sulfur in reactor, influent,
recycle, and effluent, respectively, [mg/Jl]
S/X ' = ratio of concentration of elemental sulfur to concentra-
tion of biomass, [mg/mgj
T = temperature, [ K]
U = unit rate of denitrification, equal to rate of nitrate-
nitrogen removal divided by biomass concentration,
[mg/mg-d]
U = maximum unit rate of denitrification at a specified
temperature, [mg/mg-d]
U = maximum attainable unit rate of denitrification at a
' specified value of S/X, and temperature, [mg/mg-d]
U = coefficient in Arrhenius equation for temperature
dependence of unit rate of denitrification, [mg/mg-d]
V = volume of reactor, [£]
w . = volumetric flow rate in wastage line, [£/d]
X, X , X = biomass concentration in reactor, recycle, and effluent,
respectively, [mg/Jl]
Xa, xf = total solids concentration in reactor and recycle,
respectively, [mg/£]
y = observed biomass yield, equal to rate of biomass produc-
tion divided by rate of nitrate-nitrogen removal,
[mg/mg]
Z = distance into biofilm from sulfur surface, [m]
ZSV = zone settling velocity measured in batch settling tests,
[m/d]
6 = biofilm thickness, [m]
AAlk = decrease in alkalinity concentration, [mg/£]
ASO.-S = increase in sulfate-sulfur concentration, [mg/£]
0 = hydraulic retention time, equal to reactor volume
divided by influent flow rate, [d]
0 = solids retention time, equal to total amount of biomass
0 divided by rate at which biomass is removed from system
[d]
Xlll
-------
QC' = minimum attainable solids retention time, corresponds
to maximum attainable unit rate of denitrification, [d]
u = stoichionietric coefficient equal to rate of sulfur
removal divided by rate of nitrate-nitrogen removal,
[mg/mg]
TL, ^ = kinetic coefficient which depends on sulfur particle
geometry, biomass density, and intra-film kinetic and
transport coefficients, [mg/mg].
xiv
-------
SECTION 1
INTRODUCTION
Concern with nutrient enrichment of natural waters and safety of drink-
ing water supplies has stimulated recent research and development of bio-
logical denitrification processes. At the present time (1978) the most
highly developed denitrification process employs heterotrophic organisms in
the final stage of a multiple stage reactor system. Because the influent to
the denitrification step contains essentially refractory organics, an exo-
genous supply of organic compounds (typically, methanol) must be added to
supply energy for the microbial denitrification process. However, recent
political-economic events have resulted in rapid increases in the cost of
crude oil and concomitant decreased availability of methanol and other
organic chemicals. Thus, it becomes attractive to consider alternative
methods of denitrification.
An alternative to heterotrophic biological nitrate removal could employ
an enrichment culture of Thiobacillus denitrificans in an autotrophic denitri-
fication process. This organism does not require organic compounds and can
reduce nitrate to nitrogen gas while oxidizing a wide variety of sulfur
compounds (S=, S°, S o!!, SO", SO~) to sulfate. T_. denitrificans is auto-
trophic since it uses inorganic carbon as its source of carbon for cell
synthesis.
Autotrophic denitrification processes can be categorized according to
sulfur source and reactor configuration. Elemental sulfur appears to be the
sulfur compound most likely to be feasible in a full scale process due to its
low cost, ease of storage and handling, and lack of toxicity. Other forms
of sulfur such as thiosulfate or sulfide might also be practical, especially
if industrial wastes containing these compounds were available. Such
soluble sulfur compounds could also be expected to sustain higher rates of
denitrification than elemental sulfur. It is anticipated that slurries,
packed beds, and expanded beds could all be used as reactor configurations
for autotrophic denitrification.
The primary advantage of an autotrophic denitrification process over
heterotrophic processes is the expected cost of supplying electron donors.
Sulfur is now relatively inexpensive (1978) and widespread adoption of sulfur
oxide removal technology for combustion stack gases would mitigate against
any future price increase. It may be possible to link sulfur oxide removal
and autotrophic denitrification either directly by having the sulfur removal
process supply the wastewater treatment plant, or indirectly by the effect on
the sulfur market of increased supplies from stack gas recovery facilities.
-------
The amount of sulfur which could be made available from sulfur oxide control
is significant. In 1968, emissions of sulfur from U.S. power plants was
estimated as 12.2 million tons, while the total U.S. commercial sulfur
production was 10.4 million tons.
\
Autotrophic denitrification processes also may have certain disadvan-
tages. These processes would enrich the wastewater in sulfate and destroy
alkalinity. Sulfate enrichment might be a problem due to deterioration of
water quality caused by the elevated sulfate concentration itself, or by its
stimulatory effect on microbial sulfide production.
Despite the potential attractiveness of autotrophic sulfur-oxidizing
denitrification, little quantitative information was available on which to
base judgments concerning the practical feasibility of the process. Thus,
the overall objectives of this project were to: 1) delineate the kinetic,
stoichiometric and solids separation characteristics of an autotrophic deni-
trification process using elemental sulfur in a slurry reactor system; 2)
determine the feasibility of employing soluble sulfur species, thiosulfate
and sulfide, in completely mixed semi-continuous flow denitrification sys-
tems; and, 3) investigate autotrophic denitrification using sulfur in packed
bed reactor configurations.
-------
SECTION 2
CONCLUSIONS
The results of the three experimental phases of this project indicate
that aUtotrophic denitrification using sulfur electron donors is a feasible
alternative technology for wastewater nitrate removal.
The first experimental phase involved continuous flow, slurry type
reactors, with elemental sulfur as the electron source. From this phase it
was concluded that:
1. Essentially complete nitrate removal (>99.5 percent) can be
attained at steady state.
2. The effect of nitrate concentration on the unit rate of
denitrification can be estimated by a Monod_function with
a saturation constant equal to 0.03 mg/1 NO_-N.
3. The maximum attainable unit rate of nitrate removal is a
linear function of the ratio of reactor sulfur concentration(S)
to reactor biomass organic nitrogen concentration (X) over the
range S/X = 45 - 1S4 mg S/mg organic-N.
4. Temperature dependence of the maximum attainable unit rate of
denitrification over the range 12-30°C can be described by
the Arrhenius equation with an activation energy of 13.2
kcal/mole.
5. Stoichiometry for autotrophic denitrification is relatively
constant over a range of solids retention times (7.6-30 days),
values of S/X (45-194 mg S/mg organic-N), and temperatures
(12-30°C), and can be represented by the following equation:
1.0 NO~ + 1.10 S + 0.40 O>2 + 0.76 H20 + 0.080 NH4
-»• 0.080 C_H 0 N + 0.50 N + 1.10 S0~ + 1.28 H
J / £• £ Q
6. Solids flux is a linear function of solids concentration for
sulfur-biomass slurries, with smaller fluxes at lower values
of S/X.
-------
7. The concentration of effluent suspended solids is a linear
function of overflow rate, with smaller concentrations at
lower values of S/X.
8. Economic feasibility of autotrophic denitriflcation will
depend to a great extent on the relative prices of elemental
sulfur and methanol.
The second experimental phase employed semi-continuous flow reactors,
with thiosulfate or sulfide as the electron source. Prom this phase it was
concluded that:
1. Reliable autotrophic denitrification can be obtained using
thiosulfate or sulfide as electron donors.
2. The consumptive ratio for these systems appears to be close
to 1.35, the thermodynamically predicted value.
3. Thiosulfate systems could be maintained with feed ratios as
low as 0.45 with no apparent inhibition of denitrification*
In addition these systems could be changed between thidsul-
fate and nitrate limiting growth conditions without affect-
ing the stability of the system.
In the final experimental phase, packed bed columnar autotrophic
denitrification was studied. From this phase it was concluded that:
/
1. Autotrophic denitrification is possible in packed bed reactors
using elemental sulfur as an electron source.
2. In reactors packed with elemental sulfut there existed a
strong correlation between sulfur particle size and minimum
hydraulic retention time necessary for complete denitrification.
3. Alkalinity consumption is an inherent characteristic of the
autotrophic denitrification process. Dolomite can be mixed
with elemental sulfur in packing media of the denitrification
reactors to provide alkalinity.
4. Autotrophic denitrification can proceed in the presence of
organics (and hence, heterotrophic denitrification) in packed
bed reactors.
-------
SECTION 3
RECOMMENDATIONS
Investigation of autotrophic denitrification as a nitrate removal
process should continue. The study reported herein showed favorable results
and it appears that autotrophic denitrification is a feasible process.
Primarily, what remains to be investigated is the performance of the
scheme on a pilot scale basis. It appears that elemental sulfur is the most
practical electron donor. Hence, pilot-scale investigations should include
a study of the effect of sulfur particle size (in both slurry and packed
bed configurations) on process performance. The long term effect of organic
matter (and hence heterotrophic denitrification) and suspended solids in
feed streams such as nitrified effluent from municipal plants must alsc- be
investigated.
-------
SECTION 4
BACKGROUND
Thiobacillus denitrif leans, the microorganism responsible for auto-
trophic denitrification, is a gram negative motile rod (0.5 x 1.0 ym) which
does not form spores (1) . A wide range of reduced sulfur compounds (S=, S°,
S-O.,, S.O~, SO^) can be oxidized by this organism to obtain energy (1).
Sulfate is the normal end-product, as intermediates do not accumulate under
optimal growth conditions (2). Oxygen, nitrate, nitrite, nitric oxide, and
nitrous oxide can serve as terminal electron acceptors for sulfur oxidation.
The nitrogen compounds will be reduced completely to nitrogen gas, except
when growth is under conditions of extreme stress (3), such as in the
presence of toxic substances (1). Since this microorganism prefers to use
oxygen rather than nitrate as a terminal electron acceptor, denitrification
should only be expected under anaerobic conditions. The pH range for growth
of Thiobacillus deni tr i f icans is between pH 6 and 8 , and the optimum has been
reported both on the acid and the alkaline side (1,4). Certain strains are
tolerant of high concentrations of metals which would normally be toxic (5) .
Nitrite (1) and pyruvate (6) are inhibitory to growth. Certain keto acids
which are inhibitory to other species of thiobacilli are probably also toxic
to T_. denitrif icans (7). Soluble organic compounds have been found to be
excreted by thiobacilli during growth (6) . As much as 20 percent of the in-
organic carbon fixed by T!_. denitrif icans can appear in the culture media (8).
Although classified as obligately autotrophic (2), there have been some re-
ports of T\ denitrif icans growing on organic compounds (6,9).
Several reviews have been published dealing with the biochemistry of
sulfur oxidation by thiobacilli (10,12). The proposed pathway for elemental
sulfur oxidation involves reduced glutathione (GSH) , which combines with
elemental sulfur to form a polysulfide that is oxidized to sulfite. In
aerobic thiobacilli, the enzyme responsible for this step has been isolated
and found to require molecular oxygen. Therefore, most of the energy from
sulfur oxidation comes from the oxidation of sulfite (12). Two methods of
sulfite oxidation have been observed in T?. denitrif icans . A substrate level
phosphorylation is involved in one of the steps with adenosine phosphosulfate
(APS) as an intermediate (13,14). The other pathway is cytochrome-1 inked
and involves only oxidative phosphorylation (10) .
S° + GSH ->• GSS0H -> SO,
o a j
Sulfite Oxidase
APS Reductase , %
S (1)
'
-------
The physical form of sulfur has a great influence on the rate of its
uptake. Smaller, wettable forms can be expected to be much more available
to the microorganism. Colloidal sulfur is oxidized almost as fast as
soluble sulfur compounds (12).
The exact method by which sulfur is transported into the cell is not
known, but two hypotheses have been made (12) . One scheme postulates a
water-soluble extracellular carrier enzyme which transports the water-insol-
uble sulfur into the cell. Since well washed cells oxidize sulfur at linear
rates without a lag period (15), this hypothesis is probably incorrect. The
second mechanism involves a reaction between sulfur and a cellular component
at the cell wall-sulfur interface. This explanation blends well with the
postulated sulfur oxidation pathway, if a membrane-bound thiol can be sub-
stituted for GSH (12). Examination of the cell surface with an electron
microscope indicates the existence of an intermediate of sulfur oxidation
which apparently contains a thiol group (16).
Intermediates of nitrate reduction by T_. denitrificans have been
reported as nitrite, nitric oxide, and nitrous oxide (17).
N0~ •*• NO~ -»• NO ->• NO -> N° (2)
J £, £, £,
An electron transport system (ETS) with cytochromes is involved in nitrate
reduction (18-21). Inhibition of nitrate reduction by oxygen indicates the
existence of two ETS or a branched system (22,23). The reduction of nitrite
to nitrogen gas, however, does not seem to be linked to cytochromes (17).
The Calvin cycle for carbon dioxide reduction used by photosynthetic
cells is the pathway found in thiobacilli (24). Although some tricarboxylic
acid cycle enzymes are present (25) they are used for biosynthetic
purposes (26).
At the time (1974) work was initiated on this project, there was only
one report in the literature on the application of this microbial phenomenon
to wastewater treatment. Gram (27) had described a laboratory feasibility
study in which a simulated agricultural drainage water was successfully
denitrified using microbially active anaerobic columns packed with elemental
sulfur.
-------
SECTION 5
THEORETICAL CONSIDERATIONS
DEVELOPMENT OF KINETIC MODEL
A mathematical model was developed to describe microbial growth on
solid, water-insoluble substrates at high biomass densities in order to
describe the kinetic behavior of autotrophic denitrification processes using
elemental sulfur. This model was based on a material balance about a differ-
ential section of biofilm attached to a sulfur particle. Figure 1 is a
schematic representation of such a system, where C1 and C_ represent the
concentrations of sulfur and nitrate within the biofilm. Distance into the
biofilm from the sulfur surface is represented by Z, and the total biofilm
thickness is represented by 6. Sulfur is solubilized in the biofilm at
Z = 0 at a concentration of C°, and is biologically oxidized while being
transported through the film. Sulfur does not leave the biofilm and enter
the bulk liquid at Z = 6, since it is insoluble in water. Nitrate enters
the biofilm at Z = 6 at a concentration of C^ , and is similarly removed
while being transported through the film. At the point Z = Z1, the intra-
film nitrate concentration goes to zero, so no reaction occurs in the region
O
-------
Bulk Liquid
0
Figure }.. Schematic of sulfur-biofilm system with zeroth-order nitrate rate
limitation.
-------
The equation resulting from a differential material balance on a sec-
tion of the biofilm was combined with a material balance over the entire bio-
film to produce a kinetic equation for the unit rate of denitrification.
U = U
max
tanh
s/x
N
K +N
n
(3)
U = unit rate of denitrification, [mg/mg-d];
U = maximum unit rate of denitrification, [mg/mg-d];
ItlclX
C = kinetic coefficient which depends on sulfur particle
geometry, biomass density, and intra-film kinetic and
transport coefficients, [mg/mg];
S/X = ratio of reactor sulfur concentration to reactor biomass
concentration, [mg/mg];
N = bulk liquid nitrate-nitrogen concentration [mg/A];
/
K = saturation coefficient in function expressing dependence
of unit rate of denitrification on bulk liquid nitrate-
nitrogen concentration, [mg/£].
It is convenient to separate the right-hand side of Equation 3 into two
functions. One of these functions represents the dependence of U on S/X; the
other the dependence of U on N.
U = U
m,a max
tanh
'Tib'
(4)
NFN =
N
K +N
n
(5)
Um ^ = maximum attainable unit rate of denitrification,
[mg/mg-d];
m,a
NFN = function expressing dependence of U on nitrate concen-
tration (dimensionless). ,
The hyperbolic tangent function
tanh (-
S/X
S/X'
in Equation 4
is similar in form to the Monod function, which is commonly used to describe
microbial kinetics. Figure 2 shows the relationship between the hyperbolic
10
-------
2.5 3.0
Figure 2. Hyperbolic tangent and Monod functions versus £(S/X).
11
-------
tangent function and the Monod function with £(S/X) as the independent vari-
able. Each function has a region of first-order behavior at low values of
CtS/X), and each approaches a limiting maximum at high values. The hyper-r
bolic tangent function, however, approaches its maximum faster and displays
a larger region where the rate is proportional to £(S/X).
The most important aspect of this model is the conclusion that the
primary kinetic variable is neither the biomass concentration nor the sulfur
concentration but their ratio. For constant nitrate concentration and temper-
ature, this ratio determines the rate of denitrification. Under relatively
non-restrictive assumptions, S/X is proportional to the biofilm thickness,
which is the maximum length sulfur or nitrate must travel through the film
before reacting. Decreasing the biofilm thickness increases the average
intra-film sulfur concentration, thereby increasing the observed rate.
Sulfur surface area is the quantity which actually influences the rate, but
sulfur mass can be used as a measure of surface area when the proportionality
between the two remains constant.
Temperature dependence of biological rates is often expressed in an
analogous manner to the temperature dependence of chemical rq.tes, which can
usually be represented by the Arrhenius equation (28,29).
U = U exp(-E /RT) (6)
O a
E = Arrhenius activation energy, [kcal/molej;
a
R = universal gas constant, [kcal/mole- K];
T = absolute temperature, [ K];
U = constant, [mg/mg*d].
Activation energy (E ) is the parameter which incorporates the temperature
dependence. It can be determined by fitting experimental data to a linear
equation relating the natural logarithm of the rate to the inverse of the
absolute temperature.
in (U) = Hn (U ) - (E /R) ^ (7)
O a. 1
Figure 3 shows a schematic of a slurry reactor system similar to the
ones used in Phase I of this study. Material balances on nitrate, biomass,
and sulfur can be made and the results used to apply the kinetic model to
predict behavior of a slurry reactor system. Table 1 summarizes these
material balance equations.
Solids separation in a slurry system is intimately linked with overall
process performance. Solids must be separated from the effluent (clarifi-
cation) and compacted to higher concentrations (thickening) if the process
is to operate. Poor clarification decreases effluent quality by increasing
suspended solids and, in the extreme, can cause system failure due to biomass
12
-------
Reactor
Q,N0,S°=cN0
Clarifier- Thickener
oko
V.X.S.N
rQ.Sr.Xr
(Q-w),N,Seff,Xeff
w.Sr,Xr
N°,N = concentration of nitrate-nitrogen in influent and
reactor respectively, (M/L);
S°,S,Sr,S" = concentration of elemental sulfur in influent,
reactor, recycle line, and effluent respectively, (M/L );
X,X ,X = biomass concentration in reactor, recycle line,
and effluent respectively, (M/L3) •,
Q,rQ,w = volumetric flow for influent, recyle and wastage
flows respectively,(L/T).
Figure 3. Schematic of slurry reactor system.
13
-------
TABLE 1. SUMMARY OF STEADY-STATE MATERIAL BALANCES FOR AUTOTROPHIC
DENITRIFICATION IN A SLURRY REACTOR SYSTEM
Constituent
Nitrate
Biomass
Sulfur
Material Balance Reactor Concentration
i: N0'1* n - n° nex
R. — . N — IN UOA
n 6
e
X o c
Jt\ = rt " 2* ™ JL -i I IN ™W J A
x 0 obs 9
c
o e
cN S , o , , C
R c1 — riLT f *^ »ii i i \ivf i . _ i -
— . — . S — IN (C-u) i UNJ x1
s e 6 6
Sulfur to Biomass
Ratio
S/X =
ObS
ON
R , R , R = volumetric rates of removal of nitrate
and sulfur, and production of biomass,
respectively [mg/fc • d];
6 = hydraulic retention time, equal to V/Q, [d];
8 = solids retention time, equal to
C VX/(w Xr+(Q-w) Xeff), [d];
c = sulfur feed ratio, equal to ratio of sulfur
feed rate to QN°, [rag/rag];
u = stoichiometric parameter equal to R /R ,
[mg/mg].
loss. Poor thickening characteristics require larger areas in the clarifier-
thickener and, in the extreme, cause system failure by inhibiting clarifi-
cation.
The batch flux method is a technique used to determine the thickening
characteristics of a slurry and to estimate conditions under which solids
separation fails (30,31). This method employs a series of batch settling
tests to develop a relationship between zone settling velocity (ZSV) and
initial total solids concentration (X ). This data is then analyzed by
regression techniques to determine a functional relationship between ZSV and
X.. These results are then used to compute the maximum solids flux that can
be achieved in a continuous clarifier - thickener.
14
-------
Batch flocculent settling tests are used to estimate total suspended
solids concentrations in the effluent from a continuous clarifier-thickener.
The test proceeds by adding s. slurry to a settling column and then removing
samples for suspended solids analysis at different times and depths during
quiescent settling. Measured concentrations are plotted on a depth versus
time graph and lines of equal concentration are estimated. This graph is
then used in a standard procedure to predict effluent suspended solids con-
centrations as a function of overflow rate (32).
STOICHIOMETRY OF AUTOTROPHIC DENITRIFICATION
General Concepts of Microbial Stoichiometry
The transformations of compounds involved in microbial growth can be
represented quantitatively by a balanced stoichiometric equation. The co-
efficients in such an equation can be used to determine the relationships
among rates of reaction for all products and reactants.
V A + V.B ->- v C + v D (8)
— R = — R=— R = — R, (9)
v a v. D v c v, d v '
a D c d
v , v, , v , v , = stoichiometric coefficients for components
A, B, C, and D, respectively;
R , R. , R , R, = rates of removal of components A and B and
rates of production of C and D, respectively,
[mg/£-d].
The Stoichiometry of microbial reactions in wastewater treatment has
often been described in terms of observed yield coefficients, rather than
the stoichiometric coefficients shown in Equation 8. An observed yield
coefficient relates the production or removal of one component to the produc-
tion or removal of another component. The component used as a reference is
usually the pollutant of primary concern in the treatment process. For
example, the primary goal in the treatment of sewage is the removal of
oxygen-demanding organics measured as chemical (COD) or biochemical (BOD)
oxygen demand. A major problem of these systems is handling the excess
biomass produced. The stoichiometric relationship between these two concerns
is usually expressed as an observed biomass yield coefficient that relates
the amount of biomass produced per mass of COD or BOD removed. In genera,!,
the observed yield coefficient for some component C using component A as
reference can be defined as follows.
v (molecular weight of C)
YC = -H (10)
obs v (molecular weight of A)
3.
Yc = observed yield coefficient for component C
r\V*e* •*
obs
using component A as reference, [mg/mg].
15
-------
The equation representing the observed stoichiometry of microbial
growth can be considered to be a linear combination of two subordinate
equations (33). These subordinate equations represent the two basic proces-
ses involved in microbial growth—energy conversion and cell synthesis.
Simple element and charge balances are used to construct the sub-reactions.
In all but photosynthetic growth both sub-reactions are oxidation-reduction
reactions. Therefore, it is convenient to express them on a basis of a one
electron transfer to facilitate construction of the overall reaction.
Observed Reaction = f (Energy Reaction) + f (Synthesis Reaction)
6 i S
f , f = fraction of electron equivalents in observed reaction
allocated to energy and synthesis sub-reactions,
respectively. (11)
f + f =1.0 (12)
e s
Microbial growth does not normally display a constant stoichiometry.
This is due to variations in the composition of cell material and differ-
ences in the efficiency with which the microbes couple energy transformation
with cell synthesis. The manner in which these processes are coupled depends
on environmental variables and is expressed in the values of f and f . A
high efficiency (high values of f ) occurs at maximum growth rates when
growth conditions are optimal. As growth rate declines, a smaller fraction
of the energy made available in the energy reaction is effectively used to
produce biomass.
The growth rate of microorganisms in a biological waste treatment
system is the primary variable related to process performance, so it is
easy to relate changes in stoichiometry to process operations (34). The
growth rate in these systems is usually expressed indirectly through the
operational variable called the solids residence time (6 ). This variable
is equal to the reciprocal of the growth rate, and is calculated by dividing
the total amount of biomass in the system by the amount removed per unit
time (34).
Several reports on the effect of growth rate on microbial stoichiometry
are available (35-39). In most instances the effect of 9 on one stoichio-
c
metric parameter is reported. Since the observed stoichiometric equation
is a linear combination of two other balanced equations, specifying any one
stoichiometric parameter determines all others. The parameter most often
used to express stoichiometry in biological wastewater treatment is the
observed biomass yield (Y ), which relates the amount of biomass produced
per substrate utilized. In most instances Y decreases with increasing
V
Microbial stoichiometry can also be used to calculate oxygen uptake
rates (38), and efficiencies of nitrogen and phosphorus removal in waste-
water treatment systems (40). In microbial processes such as nitrification
and denitrification, where hydrogen ions are produced or destroyed, the
stoichiometric equation can be used to estimate process performance from
16
-------
alkalinity measurements.
Thermpdynamic Based Predictions of Microbial Stoichiometry
A method has been developed by McCarty to predict microbial stoi-
chiometry by estimating fg and f using theoretical arguments (41). The
basis of the method is a balance on the primary energy storage component
of the cell, adenosine triphosphate (ATP). Most reactions which release
energy produce ATP. Most reactions which require energy consume ATP.
ATP-energy produced by cell = ATP-energy used by cell
(13)
Estimates of the ATP-energy produced by the cell are made from thermo-
dynamic analysis of the energy available from the energy reaction. The
change in Gibbs Free Energy (AG) for a reaction is the best measure of the
available energy released or required by that reaction. The value of AG for
any given reaction will vary with changing environmental conditions such as
temperature, pH, and relative amounts of reactants and products. For the
range of these parameters usually encountered in microbial systems, the
variation in AG is small. Therefore, a value of AG measured under standard
conditions (AG ) is used in these calculations.
Microorganisms are not completely efficient in producing or in util-
izing ATP. Therefore, a factor representing the efficiency of energy con-
version must be used to determine ATP-energy from thermodynamic energy. This
efficiency factor could vary with changing environmental factors and could
be: different for the energy production and utilization processes. However,
in this analysis it will be considered constant. A value for k of 0.6 has
been found to be the best estimate for the energy efficiency of a variety of
microorganisms (41). The ATP-energy balance incorporating k is:
f k
e
change in Gibbs Free
Energy for one elec-
tron equivalent of
energy reaction
= f
ATP-energy required
for synthesis of one
electron equivalent
of cells
(14)
k = efficiency factor for microbial energy
conversion.
The energy requirement for synthesis is estimated from experimental
data. Several different microorganisms have been found to require approxi-
mately 7.5 kilocalories of ATP-energy to produce one electron equivalent of
cells from appropriate biosynthetic intermediates (41). Pyruvate was chosen
as the synthesis intermediate for this ATP-energy balance because it appears
in both biosynthetic and catabolic pathways in several microorganisms (41).
The ATP-energy required for synthesis consists of the sum of the ATP-energy
necessary to convert the carbon source to pyruvate plus that required to
convert pyruvate to cell material. Conversion of the carbon source to
pyruvate will sometimes release ATP-energy. In this case it should be sub-
tracted from the ATP-energy necessary to convert pyruvate to cell material
to obtain the ATP-energy required for synthesis. If a nitrogen source is
17
-------
used by the microorganism which is not at the oxidation level of ammonia,
the energy required to transform it to that level must also be included in
calculating the energy required for synthesis.
AG° f+1 AG° > 0
fek(-AG°) = fa ( -£ + AG°), m = I _, J < Q (15)
AG — standard free energy change for oxidation of one electron
equivalent by energy reaction, [kcal/electron equivalent];
AG = standard free energy change for cpnversion of one electron
equivalent of carbon source to pyruvate, [kcal/electron
equivalent] ;
AG = ATP-energy required to convert one electron equivalent of
pyruvate to cell material,
= 7.5 [kcal/electron equivalent];
m = coefficient indicating whether energy is released (-1) or
used (+1) by reaction which converts carbon source to
pyruvate.
Solving Equation 15 for the ratio f /f gives:
e s
AG°
-*+ AG°
m c
f /f = (16)
k(-AG°)
Individual values for f and f can be calculated by noting that they are
fractions of a whole.
fe/fs - ^ - 1 (17)
Theoretical Prediction of Autotrophic Denitrification Stoichiometry
A slight modification of Mccarty's method is necessary to apply it to
autotrophic denitrification. Thiobacillus denitrificans reduces carbon
dioxide to cell mass in the same manner as photosynthetic cells. It has been
shown that, from an energetics viewpoint, it is best to assume that water
is the electron donor for this reduction even in non-photosynthetic cells
(41). Photosynthetic cells excrete oxygen as a by-product of this reaction,
but this cannot occur in autotrophic denitrification because it is an
anaerobic process. Thus, oxygen produced in the synthesis reaction during
autotrophic denitrification must be reduced by electrons from the energy
reaction. Figure 4 shows a schematic of the proposed electron flow in
autotrophic denitrification. In the overall reaction Stoichiometry, sulfur
18
-------
f, S04=
Figure 4. Proposed electron flow for autotrophic denitrification ^r
water the electron donor for synthesis
anaerobic conditions with water the
19
-------
is the apparent electron donor for synthesis rather than water, since in-
ternal recycle excludes oxygen from the overall stoichiometry.
The balance on ATP-energy for autotrophic denitrification using the
above assumptions is:
ATP-energy produced by cell = ATP-energy used by cell (13)
AG
k[f (-AG° ) + f (-AG° )] = f I-r2- + AG°] ' (IS)
e sn s so s k c
AG°/k + AG° + kAG°
Vfs = ^ S5 — (19)
6 S k(-AG° )
sn
AG = standard free energy change for oxidation of one elec-
sn
tron equivalent of sulfur by nitrate;
= 21.78 kcal/electron equivalent;
AG = standard free energy change for oxidation of one
electron equivalent of sulfur by oxygen;
k
= 23.33 kcal/electron equivalent.
The energy and synthesis reactions for autotrophic denitrification using
elemental sulfur can be expressed on a one electron equivalent basis.
Energy Reaction
0.200 N0~ + 0.167 S + 0.0667 HO -»• 0.100 N, + 0.167 SO~ + 0.133 H+
J 2 (20)
Synthesis Reaction
0.250 CX>2 + 0.167 S + 0.050 NH* + 0.267 HO -»• 0.050 CgH-0 N
+ 0.167 SO" + 0.383 H+ (21)
These reactions and Equation 11 can be used to calculate Y from the ratio
f /f . This stoichiometric yield coefficient relates the amount of biomass,
measured as organic nitrogen, produced per mass of nitrate-nitrogen removed.
f (1/20) (14) f
- S ~ (0.25) (22)
obs f (1/5) (14) f
e e
An observed yield coefficient of 0.084 mg organic-N/mg NO~-N can be calcu-
lated using Equations 19 and 22 and a value for k of 0.6.
Using a similar analysis balanced stoichiometric expressions ban be
20
-------
derived for other sulfur electron donors. For example, Equations 23 and
24 represent the balanced stoichiometry for thiosulfate and sulfide,
respectively.
Thiosulfate
0.844 S0~ + N0~ + 0.347 CO0 + 0.0865 HCO~ + 0.0865 NH*
^ J j ^ 3 4
+ 0.434 H20 •*• 1.689 SO^ + 0.5 N2 + 0.0865 C^O N
+ 0.697 H+ (23)
Sulfide
0.422 H.S + 0.422 HS~ + NO~ + 0.347 CO,, + 0.865 HCO~ + 0.0865 NH+
2 3234
•*• 0.844 S0~ + 0.5 N- + 0.0865 CrH_O^N + 0.409 H+ (24)
4 ^ 3/2
When soluble electron donors such as sulfide or thiosulfate are
employed two additional stoichiometric parameters become useful. The con-
sumptive ratio, C , and the feed ratio, Cp, are defined in Equations 25
and 26, respectively.
electron equivalents of S_0 or S consumed
CR V (25)
electron equivalents of NO reduced
electron equivalents of electron donor
(sulfur compound) in feed ^ ._g.
F ~ electron equivalents of electron acceptor
(nitrate) in feed
C and CR are defined such that if Cp is greater than C , then growth
will be nitrate limiting. When Cp is less than CR growth will be electron
donor (sulfur compound) limiting.
21
-------
SECTION 6
ANALYTICAL TECHNIQUES
Table 2 presents a summary of the analytical techniques employed in
this project. Detailed discussion of analytical techniques in the subse-
quent paragraphs of this section is restricted to those methods that were
non- routine and required developmental effort by the project staff.
A modification to the cadmium reduction method for nitrate (42) was
required. To standardize flow rates, a modified cadmium reduction column
was constructed by filling a length of 6.3 mm glass tubing with 0.25-0.42 mm
cadmium metal filings to a depth of approximately 80 mm, as shown schematic-
ally in Figure 5. This reductor was connected to another piece of glass
tubing which was bent to facilitate sample collection and the two were
placed in a buret holder on a ring stand. The top of the reductor was
connected to a glass funnel (maximum diameter 100 mm) held on the ring stand
approximately 500 mm above the cadmium. All connections in this apparatus
were made with clear plastic tubing. Samples were prepared for analysis in
the standard manner (42). The height of the funnel above the reductor main-
tained relatively constant flow rates during an analysis. The column was
standardized on a regular basis by passing a sample of known nitrate concen-
tration. Corrections for the reagent blank, column blank and partial re-
duction of nitrite in the column were made when calculating the nitrate
concentration. Experience showed that 10 percent of the nitrate that was
applied to the column would be reduced to some nitrogen compound other than
nitrite.
(-9>J <27>
-- cb
s bl
N = nitrate-nitrogen concentration in sample, [mg/&] ;
N = nitrate-nitrogen concentration in standard, [rag/Jl] ;
S
D.F. = dilution factor, _i.£. , volume of diluted sample divided
by volume of original sample;
A = absorbance of diluted sample passed through column;
A = absorbance of standard; ,
A , = absorbance of column blank, i.e., blank sample passed
cb --
22
-------
TABLE 2. '• SUMMARY OF ANALYTICAL TECHNIQUES
Analysis
Method
Ref.
Comments
00
Nitrogen Species
Nitrate
Nitrite
Organic Nitrogen
Nitrogenous Gases
Sulfur Species
Elemental Sulfur
Sulfate
Thiosulfate
Sulfide
Other Analysis
ATP
Alkalinity
PH
COD
Total Suspended
Solids
cadmium reduction
diazotization
digestion and
1)distillation and
acidometric titration
2)ammonia probe
gas chromatography
iodometric titration
turbidometric
iodometric
titrimetric
luciferin-luciferase assay
Gran acidometric titration
pH meter with glass electrode
dichromate digestion
gravimetric
42
42
43
42
43
43
44
45
43
43
see text
modified for micro-
kjeldahl analysis
see text
see text
see Appendix A2
see text
Accumet 320 pH meter
(Fisher Sci. Co.)
dilute reagents
glass fiber filters
(Whatman GF/C)
-------
Figure 5. Schematic of cadmium reduction column used for nitrate analysis.
24
-------
through reduction column;
= absorbance of diluted sample not passed through column,
i-e_., absorbance due to nitrite;
= absorbance of reagent blank.
0.9 = empirically determined efficiency of reduction of
nitrate to nitrite.
Organic nitrogen was measured by sulfuric acid digestion with mercury
catalyst (43) followed by direct measurement of ammonia with an ammonia probe
(Orion Model 95-10) or distillation and titration of the ammonia (43). In
the initial period of this study, the probe method was used exclusively.
However, this technique developed erratic and insensitive behavior and was
replaced by the distillation and titration procedure.
Elemental nitrogen, nitrous oxide, nitric oxide, carbon dioxide and
oxygen were separated and measured in a two-column gas chromatograph (Varian,
Model 90-P3) with thermal conductivity detector. The first column was
packed with Poropak Q, 0.15-0.18 mm (80/100 mesh), the second with molecular
sieve 13 x, 0.25 - 0.60 mm (30/60 mesh).
£ procedure was developed to analyze elemental sulfur in aqueous
solution (Appendix Al). The procedure consisted of converting the sulfur
to thiosulfate by boiling with a sulfite solution, then analyzing the
thiosulfate by an iodine titration after complexing residual sulfate with
formaldehyde.
High, variable concentrations of carbon dioxide in the reactors caused
variable endpoints in alkalinity titrations. This problem was overcome by
using a Gran titration method for alkalinity which determines the endpoint
from titration data (45). A material balance on hydrogen ions is made by
assuming that all H+ added after the equivalence point is passed, remains
in solution as the free ion.
(V - V.) 10~PH = (V. - V?) C (28)
s t T. T- a
V = volume of sample, [m£];
s
V. = volume titrated, [m £];
Ve = volume titrated at equivalence point, [m.£] ,-
C = concentration of acid in titrant,
..equivalents,
C mlJ *
Volume titrated and pH were recorded at four points below pH = 4.8 and a
least squares regression performed to determine v| .
25
-------
SECTION 7
EXPERIMENTAL STUDIES
The experimental aspects of this project were conducted in three
phases. In the first, and most extensive phase .(Phase I) elemental sulfur
was employed in both semicontinuous and continuous flow complete mix slurry-
type denitrification reactors. The second experimental phase (Phase II)
employed thiosulfate and/or sulfide in completely mixed, semicontinuous flow
denitrification reactors. The final experimental phase (Phase III) employed
sulfur in continuous flow packed bed denitrification reactors. The proced-
ures, results, and discussion of each phase of the experimental study are
presented separately in subsequent subdivisions of this section.
PHASE I - SLURRY REACTORS USING ELEMENTAL SULFUR
A two-part experimental plan was developed to investigate the character-
istics of autotrophic denitrification in slurry reactors fed elemental
sulfur. The first part of the plan was to develop and characterize a micro-
bial enrichment culture that could denitrify with elemental sulfur. In the
second part, these cultures were used in continuous culture experiments to
determine the kinetic and stoichiometric behavior of the process. Solids
separation characteristics of sulfur-biomass slurries from these continuous
cultures were also evaluated. Throughout Phase I, the concentration of bio-
mass in the reactors which is denoted by the symbol X was estimated by
measuring the mixed liquor non-filterable organic nitrogen. Thus, in all
presentations of results and expression of the ratio S/X, the units of X are
mg/JZ, organic-N. It was necessary to use suspended organic nitrogen as a
surrogate parameter for biomass because the high concentrations of inorganic
sulfur contained in the slurry reactors rendered determination of biomass
by the conventional surrogate, i_.e_., volatile suspended solids, a non-repro-
ducible and highly inaccurate exercise. In some of the continuous flow
experiments conducted during this phase of the experimental study, ATP
measurements were performed on the mixed liquor suspended solids. While
such measurements are considered to be correlatable to active bacterial bio-
mass, it was felt that determination of suspended organic nitrogen was a more
reproducible measurement and more .easily related to actual bacterial biomass
through the empirical formula widely used to chemically describe bacterial
protoplasm, ±.e^-, C,-H702N.
Culture Characterization Experiments
Experimental Plan—
The culture characterization portion of Phase I was primarily concerned
26
-------
with estimating process performance so that the continuous culture experi-
ments could be more efficiently executed. The semicontinuous reactors used
during these experiments were operated by periodically removing waste solids
and adding elemental sulfur and a media consisting of tap water enriched with
nitrate and nutrients. One series of experiments estimated the effect of
sulfur feed rate on the kinetics of autotrophic denitrification by varying
the amount of sulfur added to several semicontinuous reactors. Another
series of reactors was operated with varying amounts of ammonia in the feed
to determine if ammonia was strictly required as the nitrogen source for cell
synthesis, as reported for pure cultures of Thiobacillus denitrificans.
Temperature effects on the rate were estimated by operating two reactors at
different temperatures. Three other batch experiments were conducted to
determine reaction stoichiometry. Table 3 summarizes the experimental plan
during culture characterization.
TABLE 3. SUMMARY OF EXPERIMENTAL PLAN
Effect Measured Magnitude of Variable in Experiment
Sulfur feed ratio 5, 25, 100, 500 (mg S/mg NO~-N)
Ammonia feed ratio 0, 0.05, 0.10, 0.20 (mg NH.-N/mg NO~-N)
Temperature 12, 20 (°C)
Stoichiometry (3 replicate experiments)
Experimental Techniques —
Figure 6 shows a schematic representation of the semicontinuous reactor
system used during culture characterization experiments. These reactors
were one-liter, glass bottles mixed by magnetic stirrers. The bottles were
sealed with a rubber stopper and the gas produced during denitrification was
collected in a graduated cylinder inverted in a beaker of water. Reactor
temperature was controlled by placing the reactors in constant temperature
rooms. All experiments were conducted at 20°C except for one reactor oper-
ated at 12°C to estimate temperature effects.
Initial seed for this study was obtained from samples of soil, mud, and
anaerobically digested sludge from a municipal wastewater treatment plant.
Approximately 80 grams of each sample were placed in 300 m£ glass bottles to
which standard media containing thiosulfate instead of sulfur was added. The
bottles were capped and incubated anaerobically at room temperature. All
samples showed increased turbidity after several weeks, but the digested
sludge sample from a nitrifying activated sludge plant was most active.
Supernatants from all the bottles were collected, and mixed together. This
culture was regularly fed standard media and was adapted to growth on
elemental sulfur before being used as the seed for all the reactors used
throughout the entire study.
27
-------
Reactor
O
I I
Magnetic
Stirrer
Gas Collection
Jf2> Apparatus
Figure 6. Schematic of semicontinuous culture reactor systejn.
28
-------
The semicontinuous reactors were operated at a hydraulic retention
time of 5 days and a solids retention time of 20 days by wasting mixed
liquor and supernatant and feeding a nutrient solution every 4 days. At
each feeding, reactor walls were scraped; 200 mJl of the reactor contents were
wasted; and, the solids were allowed to settle. 600 m£ of clear supernatant
were then removed and 800 m£ of feed solution were added. Elemental sulfur
was dried and passed through a 150 micron sieve before addition to the re-
actor. Gas production was measured at various times after feeding and the
rate determined from the slope of the cumulative gas volume-time curve. The
rate of gas production was used to measure the rate of denitrification, since
the gas being produced was almost entirely elemental nitrogen formed by micro-
bial reduction of nitrate. The standard media used in the semicontinuous
reactors was a modification of the media used by Baalsrud (1). Table 4 lists
the components of this media. Tap water that had been dechlorinated by over-
night aeration was used as the basis of the media.
TABLE 4. COMPOSITION OF MEDIA FOR SEMICONTINUOUS CULTURE EXPERIMENTS
Concentration
Constituent (milligrams/liter)
S° 2,500
KNO 721 (100 mg/A as N)
KH-PO 300
f* 4
K,HPO 1,500
NaHCO3 1,000
NH Cl 76 (20 mg/H as N)
• 6H2O 500
FeCl3 . 6H2O 10
Experimental Results—
Variable sulfur to nitrate-nitrogen feed ratio — A major element in
determining the feasibility of autotrophic denitrification using elemental
sulfur was to evaluate how the amount of sulfur available to the micro-
organisms affected their rate of denitrification. This characteristic was
measured in a series of experiments in which four semicontinuous reactors
were operated as described above except that different amounts of sulfur
were added to each. Table 5 and Figure 7 show the results of these experi-
ments .
Variable ammonia-nitrogen to nitrate-nitrogen feed ratio — The effect
of ammonia on the rate of denitrification was measured by operating a series
29
-------
•S 3
o
3
•o
O
u>
s
CE
50
Ratio of Sulfur to Nitrate-Nitrogen in Feed,
100
mg sulfur
mg N03-Nj
O
500
Figure 7. Rate of gas production versus sulfur to nitrate-nitrogen feed ratio.
30
-------
TABLE 5. EFFECT OF SULFUR TO NITRATE FEED RATIO ON GAS PRODUCTION RATE
Nitrogen
Gas Production Rate (m£,
a • hr)
S/NO~-N
Expt.
No.
1
2
^ 3
4
AVG.
5
0.7
1.1
1.3
1.5
1.2
25
3.8
3.5
3.6
3.5
3.6
100
3.7
3.6
5.6
5.4
4.9
500
3.6
3.5
4.9
3.1
3.9
of semicontinuous reactors as previously described except that the amount of
ammonia in the feed to each reactor was different. Feed ratios of 0, 0.5,
0.10 and 0.20 (mg NHv-N/mg NO~-N) were used. The results are tabulated in
Table 6 and presented graphically in Figure 8.
TABLE 6. EFFECT OF AMMONIA TO NITRATE FEED RATIO ON GAS PRODUCTION RATE
Nitrogen Gas Production Rate (m£/&
• hr)
NH3-N/NO~-N
Expt.
No. 0.0
1 1.7
2 2.8
3 2.5
4 3.0
AVG . 2.5
Variable temperature study
0.05 0.10
2.4
3.2
2.5
3.3
2.7 3.3
0.20
3.8
3.5
3.6
3.6
3.6
— A temperature dependence study was per-
formed with semicontinuous reactors operated at 12 and 20 C. Operation of
31
-------
0
J
c
o
o
OL.
in
o
O
•s '
o
cr
0 .05 .10 .15
Ratio of Ammonia - Nitrogen to Nitrate - Nitrogen in Feed,
.20
mgNH3-N
mg N03-N
.25
Figure 8. Rate of gas production versus ammonia-nitrogen to nitrate-nitrogen
feed ratio.
32
-------
a reactor at 30 C was attempted but could not be sustained. Gas production
was initially rapid at 30°C but after a few weeks of operation it decreased
to very low levels t Nitrite-nitrogen accumulated in this reactor to a level
nearly equal to the nitrate-nitrogen concentration of the feed solution.
Table 7 presents the results obtained from the reactors operated at 12 and
20°C. Although the danger of fitting a curve to two data points is acknowl-
edged, quantitative measures of temperature dependence are useful. Therefore,
an Arrhenius activation energy of 13.0 kcal/mole was calculated from the
average gas production rates (Table 7) using Equation 7.
TABLE 7. EFFECT OF TEMPERATURE ON GAS PRODUCTION RATE
Gas Production Rate (mjj,/A • hr)
Expt.
No.
1
2
3
4
5
6
7
8
12°C
1.0
0.9
1.1
1.4
1.4
1.4
1.5
20°C
2.8
2.7
2.7
2.6
2.8
AVG. 1.2 2.7
Batch stoichiometry — An estimate of reaction stoichiometry was
desired for the preliminary characterization of autotrophic denitrification
using elemental sulfur. This data was obtained in a series of batch experi-
ments conducted by taking a seed obtained from the semicontinuous reactors
and mixing it with standard media. When gas production ceased, the reactor
was spiked with a feed solution ten times the concentration of standard
media. This procedure was repeated several times to produce a sufficient
amount of biomass to insure accurate measurement. Organic nitrogen,
nitrate, nitrite, sulfate and alkalinity were measured in the initial and
final reactor media and in the concentrated feed solution. Table 8 presents
the fesults of these experiments expressed as observed yield coefficients.
Observed yield coefficients for biomass (Y ), sulfate-sulfur (yob| ), and
alkalinity (YAjj-k) were calculated by dividing the amount of each compound
produced or destroyed in the microbial reaction by the amount of nitrate-
hitrbgeri removed by the microorganisms. Since these coefficients are based
on measurements made at the end of each experiment, they are average values
33
-------
and cannot represent possible variations in the values of the coefficients
during the course of the experiment.
TABLE 8. BATCH STOICHIOMETRIC COEFFICIENTS
Expt.
No.
1
2
3
AVG.
obs
mg organic-N.
mg NO~-N
0.096
0.075
0.095
0.089
so4-s
mg SO~-S
lmg NO~-N
2.27
2.29
2.49
2.35
yAlk
obs
, meq
vmg NO~-Ny
0.088
0.120
0.132
0.113
obs / obs
mg SO~-S
v meq '
25.8
19.1
18.9
21.3
Continuous Culture Experiments
Experimental Plan—
Two series of continuous culture experiments were conducted to delineate
the kinetics and stoichiometry of autotrophic denitrification at 20°C. Five
reactors were operated at different solids retention times in the first
series, to determine the effect of Gc on steady state reaction stoichiometry
and nitrate removal. The effect of mixed liquor S/X was evaluated in an
additional series of four reactors. The parameter, S/X, of the mixed liquor
was defined as the ratio of suspended elemental sulfur concentration to the
suspended organic nitrogen concentration. After obtaining steady state data
from this second series of reactors, transient rate tests were performed to
determine "the maximum attainable unit rate of denitrification for each of the
four values of S/X studied. One continuous reactor was also operated at each
of two other temperatures to obtain steady state and transient kinetic data.
In addition to the kinetic studies, a series of zone settling tests was
performed on sulfur-biomass slurries from continuous cultures operated at
three different values of S/X. Data from these tests were used in a batch
flux analysis of the settling properties of the slurries. The relationship
between effluent suspended solids and clarifier overflow rate was estimated
from data taken during flocculent settling tests conducted with the same
three slurries. An additional flocculent settling test was performed to
determine the effect of initial solids concentration. Table 9 summarizes the
experimental plan followed during the continuous culture experiments.
Experimental Techniques—
Reactor operations—Figure 9 shows a schematic representation of the
continuous culture reactor system. Six-liter, conical, glass reactors with a
two-liter glass inner cone and a 45 x 255 mm settling cylinder were used
(Biooxidation System, Horizon Ecology Company, Chicago, Illinois). Mixing
34
-------
TABLE 9. SUMMARY OP EXPERIMENTAL PLAN FOR CONTINUOUS CULTURE EXPERIMENTS
U)
ui
Experiment
Variable 0
c
1*
2
3
4
5
Variable S/X
1
2
3
4*
5
Variable T
1
2*
3
S/X
(mg S/mg org-N)
~ 145
^ 145
^ 145
^ 145
i, 145
45
56
100
142
194
145
145
145
0
(days)
10
15
20
25
30
near
maximum
attainable
value
near
maximum
attainable
value
Temperature Transient Microbial Settling
(°C) Assimilation Tests Tests
20 x
20 x
20
20
20 x
20 xx
20 xx
20 x
20 x
20 x
12 x
21 x
30 x
* Indicates single experiment used to describe effect of several variables.
-------
Sampler
Pressure
regulator
Constant k~z>
pressure
device
Water
trap
Figure 9. Schematic of continuous culture reactor system.
36
-------
was accomplished by a recirculation pump which supplied nitrogen gas to the
reactors through fritted glass diffusers at a rate of about 0.3 mVmin. The
upward movement of the gas between the inner and outer cones caused an in-
ternal circulation pattern which kept solids in suspension and reactor con-
tents well mixed. Solids were separated in the Plexiglass cylinder situated
at the top of the inner cone. Clarified effluent was removed from the top
o£ this cylinder and feed was supplied to the reactor by a peristaltic
pump. Influent feed rate was maintained at approximately one liter per hour
which set the clarifier overflow rate at 15 mVm2 • d. The feed system for
the continuous reactors consisted of a peristaltic pump with different heads
for influent and effluent lines, a 110 liter polyethylene feed tank, and
connecting lengths of clear plastic tubing.
Certain modifications were required to adapt the purchased reactors for
autotrophic denitrification experiments. Anaerobic conditions were maintained
within a reactor by fitting it with a Plexiglass cover sealed with a rubber
0-ring. To minimize atmospheric oxygen leaks, a positive gas pressure was
maintained within the reactor. This was done by connecting a pressurized
tank of dry nitrogen to the reactor through a pressure regulator and constant
pressure device. Recycling nitrogen gas within the reactor caused an accumu-
lation of carbon dioxide in the gas stream which equilibrated at a level of
about 2-4 percent. High settling velocities of the sulfur-biomass particles
caused solids to accumulate at the bottom sides of the reactors. This dead
area was eliminated by installing plastic funnels shaped to fit the reactor
bottom. Other dead areas around glass tubing connections to the reactor were
eliminated by installation of rubber plugs. Two traps on the effluent line
were used to take samples and return solids lost from the reactor. The possi-
bility of microbial growth in the lines from the reactor to the feed tanks
was decreased by the addition of an air-break between the reactor and the
feed pump.
Microbial seed for the continuous cultures was obtained from the semi-
continuous reactors. Biomass concentration was increased by initially feed-
ing standard media containing thiosulfate instead of sulfur. When the de-
sired biomass concentration was attained, the cultures were acclimated to
elemental sulfur and continuous flow was begun. Wasting from the reactors
was done once a day according to the following formula:
V -2-*t (29)
W 0Q
V = volume of reactor, [£];
V = volume of mixed liquor wasted, [&];
w
At = time interval between wasting, [d].
Each week the reactors were cleaned by turning off all pumps, scraping
the reactor walls and scouring the inner cone and settling cylinder. Feed
lines were changed and cleaned with a 5 percent solution of sodium
37
-------
hypochlorite. Each day the general condition of the reactors we're noted and
the volumes of feed solution in the feed tanks were recorded. The volumetric
flow rate used in calculating the hydraulic retention time (8) was deter-
mined by dividing the change in feed solution volume by the time interval
between measurements. Influent flow rates were adjusted each day if neces-
sary to maintain constant flows and sufficient sulfur was added to maintain
the desired value of S/X.
Measurement of pH and a spot test for nitrite were performed daily on
effluent samples from each reactor. The spot test for nitrite was considered
an adequate measure of a reactor's performance because nitrate was never
present in significant amounts in the absence of nitrite. Weekly analyses
for alkalinity were made to check the reactors for oxygen leaks. A large
decrease in effluent alkalinity would indicate a significant oxygen leak
because the microorganism can use oxygen to oxidize sulfur (1). Since hydro-
gen ions would be produced by this oxidation (1), alkalinity would decrease
in proportion to the amount of oxygen entering the reactors.
Composition of the feed for the continuous reactors was based on that
used previously in a study of continuous culture heterotrophic denitrifica-
tion (46). Table 10 lists the nutrients added to supplement dechlorinated
tap water. Laboratory-grade, resublimed sulfur (Fisher Chemical Company)
was dried at 60°C and passed through a 150 micron sieve before addition to
the reactors. Sulfur particle size distribution was estimated by micro-
scopic analysis of 200 particles which had been dried and passed through a
74 micron sieve. A mean average dimension of 82 microns was obtained with
10 and 90 percentile points being 42 and 123 microns, respectively. Since
particles larger than 74 microns were observed, it can be concluded that
some agglomeration of sulfur particles occurred.
Steady state techniques—Steady state data were gathered during a samp-
ling period of at least three days which began only after the reactor had
been operating for a period at least as long as three times the solids
retention time. Nitrite and alkalinity were analyzed immediately after the
effluent samples were taken and filtered. Filtration was done with 0.45
micron membrane filters (HAWP, Millipore Company) and a glass fiber pre-
filter (GF/C, Whatman) which had been washed with 240 ml of distilled-deion-
ized water. Sulfate and nitrate analyses were performed on filtered effluent
samples after storage at -10°C. Previous experience showed that filtration
and cold storage was an effective means of preservation for nitrate and sul-
fate. Samples of mixed liquor were taken from reactor wastage and analyzed
for elemental sulfur and total kjeldahl nitrogen. Organic nitrogen was equiv-
alent to total kjeldahl nitrogen for these samples since ammonia concentra-
tions were negligible. Organic nitrogen was used as a measure of biomass
in this study, since elemental sulfur interfered with gravimetric analysis of
suspended solids or volatile suspended solids.
Since samples were taken from the wastage, the measured values of
sulfur and organic nitrogen were adjusted to better represent average reactor
concentration before and after wastage. Calculation of average reactor
38
-------
TABLE 10. COMPOSITION OF CONTINUOUS CULTURE FEED SOLUTION
_ Constituent _ Concentration _
KNO3 30 mg/Jl as N (216 mg/Jl KNO )
NH4C1 1.5 mg/£ as N (6 mg/£ as NH CD
NaHC03 900 mg/S,
K HPO 10 mg/£ as P (56 mg/S, as K HPO )
MgCl • 6H O 1 mg/£
£t £.
FeCl
3 2
• H0 1 mg/£
CaCl2
pH 8.6
sulfur concentration also included consideration of the amount of sulfur
added to the reactor,
2V-V
X = (-)X (30)
2V-V S ,, ,
„ _ / _ ^NC + added
S - ( 2V )Swastage + 2V
X , S = biomass and sulfur concentration measured in
wastage wastage . , .„,
wastage, [mg/&]
S ,, , = amount of sulfur added, tmg] .
added
Transient rate study techniques— Subsequent to steady state operation,
duplicate transient rate tests were performed on some reactors to measure
U under conditions of an excess of nitrate. These reactors were spiked
witt a known amount of nitrate and samples for nitrate and nitrite analysis
were taken at 30 minute intervals for 3-7 hours.
Since nitrate and nitrite are both microbially available electron
acceptors during transient tests, a measure of their combined effect is
39
-------
necessary. The concentration of equivalent nitrate-nitrogen was used for
this purpose. This variable is equal to the concentration of nitrate-nitrogen
in a solution without nitrite, which has the same concentration of electron
equivalents as the solution of nitrate and nitrite in question. Solutions
with equal amounts of electron acceptors are capable of oxidizing equal
amounts of sulfur.
CEN = N + 0.6 NI (32)
CEN = equivalent nitrate-nitrogen concentration, [mg/H]
N = nitrate-nitrogen concentration, [mg/£]
NI = nitrite-nitrogen concentration, [mg/£] .
Data from the transient rate tests were analyzed by assuming that the
rate of removal of equivalent nitrate is independent of its concentration. A
non-steady state material balance equation was used to predict the equivalent
nitrate concentration at any time for a given value of the rate constant.
Q(CEN)° - Q(CEN) - V + VK (33)
dt
CEN = CENtQ e~t/Q + (CEN°-K0) (l-e) (34)
CEN = equivalent nitrate-nitrogen concentration in influent and
in reactor at t = 0, respectively, [mg/&] ,
K = rate constant for zeroth-order reaction [?^— ] .
a— d
The zeroth-order rate coefficient was determined from experimental data
by choosing the value of K which minimized the sum of the squared differences
between the measured equivalent nitrate-nitrogen concentration and the concen-
tration predicted by Equation 34.
min Z [CENi-CENt°e~ti/G- (CEN°-K0) (l-e'^i/0)]2 (35)
i
CEN. = equivalent nitrate-nitrogen concentration measured in
sample, [mg/£] ,
equi
"1
t. = time after addition of nitrate solution when i sample
taken; [d].
U was then calculated using the average steady state biomass concentration.
m, a
Um,a = I <36>
Solids separation study techniques—Zone settling velocity tests were
performed in one liter (60 x 355 mm) or two liter (80 x 385 nm) graduated
cylinders. The settling velocity of a slurry was found to be unaffected by
40
-------
the size of cylinder in which the settling test was conducted. An aluminum
rod, bent at right angles every 60 mm was driven by a 1 rpm motor to pro-
vide stirring during the tests. Tests run without stirring, however, showed
no difference in measured zone settling velocity (ZSV). Interface height was
measured at 15 or 30 second intervals and ZSV calculated by dividing the
average difference in height of the settling zone interface by the time
interval between measurements. Samples were analyzed for total suspended
solids after every test and used to calculate solids fluxes. Linear least
squares regressions were performed on the data to determine the functional
relationship between solids flux and solids concentration.
Flocculent settling tests were performed in a Plexiglass settling
column (45 mm x 2.8 m) with sampling ports spaced at 0.2 m intervals. A test
was begun by adding a slurry from a continuous culture to the top of the
column and keeping it well mixed by a flow of nitrogen gas entering the
bottom of the column. After the gas flow was discontinued and mixing currents
had subsided, a stopwatch was activated. Samples for suspended solids
analysis were taken at each port after the interface had passed and at several
intervals thereafter. The time and liquid level height were recorded for
each sample. This data was analyzed using a standardized procedure for
flocculent settling tests (32).
Experimental Results—
Variable solids retention time—The effect of growth rate (1/Q ) on re-
action kinetics was evaluated by operating five continuous reactors at the
same sulfur feed rate and temperature but different values of 9 . Solids
retention times of 10, 15, 20, 25 and 30 days were evaluated.
Table 11 presents a summary of the results of these experiments.
Organic nitrogen is used as a measure of biomass concentration in Table 11
as well as in all other presentations or discussions of experimental results
in Phase I. Mixed liquor ATP concentration (XATP) is another indicator of
biomass concentration. This parameter was measured during the variable 9c
experiments and results are shown in Table 11. ATP is the major compound
for energy storage within a cell, so it is representative of the amount of
biomass present, if the ratio of ATP to viable biomass remains constant. The
stoichiometric parameters AS04~S and AAlk in Table 11 represent the increase
in the effluent concentration of sulfate-sulfur and decrease in alkalinity
respectively, relative to their influent concentration.
ATP
Graphical presentations of the dependence of YQbs and X /X on GC are
shown in Figures 10 and 11, respectively. Transient rate tests were performed
on the 10 and 15-day reactors and that data is presented in Tables Bl and
B2 in the Appendix. The recirculating gases within the reactors operated at
6 = 25, 30 days were analyzed by gas chromatography. A trace of nitrous
oxide (50 ppm) was detected but no nitric oxide was found.
Variable sulfur to biomass ratio—Four additional continuous culture
reactors were operated at different values of S/X to evaluate this parameter's
41
-------
TABLE 11. SUMMARY OF CONTINUOUS CULTURE RESULTS AT VARIOUS 8
to
Operating & kinetic Mixed liquor
parameters characteristics
0 n S/X X
c ma S N*
10 0.25 142 83
(10)
15 0.25 149 133
(14)
20 0.25 145 171
(32)
25 0.24 139 . 231
(10)
30 0.24 150 234
(19)
XATP
.mg ATP
1
1.05
(0.15)
1.30
(0.39)
_
1.65
(0.25)
0.85
(0.11)
S
(SLA
( a '
11.8
(1.7)
19.7
(6.0)
24.8
(5.6)
32.1
(1.2)
35.1
(1.2)
Effluent
characteristics
NO -N
0.14
(0.13)
0.07
(0.04)
0.01
(0.01)
0.06
(0.01)
0
(n=8)
3"
0.14
(0.11)
0
(n=7)
0
(n=5)
0
(n=8)
0
(n=8)
Stoichiometric
Parameters
obs
mg N*
'mg N*"'
.071
.075
.071
.073
.063
ASO -S
94
(2)
94
(13)
103
(1)
103
. (ID
108
(8)
AAlk
("SSL,
1 a '
3.64
(0.02)
3.67
(0.16)
3.94
(0)
3.81
(0.06)
4.34
(0.38)
Organic-N
**
Nitrate-N
NOTE: Numbers in parenthes.es are standard deviations of measurements.
the number of replicates is given.
If measured value is zero.
-------
.08
.06
o
o>
o
O
.o
10 15 20
0C (days)
30
Figure 10. Observed biomass yield versus solids retention time.
43
-------
16
*t?
u
8
0
10
20
25
30
0e (days)
Figure 11. ATF content of biomass versus solids retention time.
44
-------
effect on reaction kinetics. Solids retention time for each reactor was
chosen to enable operation within 10-20 percent of the estimated minimum
attainable QC for that value of S/X. Table 12 presents a summary of results
from these experiments along with results for the reactor operated at S/X =
142 mg S/mg organic-N and QQ = 10 days in the variable growth rate experi-
ments. Values for the maximum attainable unit rate of nitrate removal (U )
were calculated from results of transient rate tests using a least squares'3
regression technique (Equations 32-35). Data from these tests are presented
in Tables B3-B6 in the Appendix. Figure 12 graphically illustrates the
dependence of U & on S/X. The linear least squares regression line calcu-
lated to describe this relationship is:
U - 0.19 + 0.01 (S/X) (37)
Ui/ C*
U = maximum attainable unit rate of nitrate removal,
m'a mg
r - ± - 1 .
mg organic-N-day '
S/X = ratio of sulfur to biomass concentration in reactor,
[mg S/mg organic-N] .
Filtered (0.45 pm pore size) samples from the reactors operated at S/X =
45, 56, 100, 194 mg S/mg organic-N were analyzed for COD and an average value
of 10 mg/fc was obtained.
Effect of temperature — Continuous feed reactors were also operated at
12 C and 30°C to evaluate the temperature dependence of U . These reactors
were operated at approximately the same S/X and, at solids 'retention times
near the estimated minimum attainable value for each temperature. Table 13
presents the results of these experiments plus the experiment conducted at
21°C and 10-day 0 during the variable growth rate experiments. Results ob-
tained during the transient rate tests used to determine U are presented
in the Appendix in Tables B7 and B8. Results of these transient tests
indicate that nitrite accumulation was much more prnounced at 30°C than at
12°C, Figure 13 is an Arrhenius plot used to determine the apparent acti-
vation energy for U (Equation 7). A least squares regression on the
linearized data yie?dtd an activation energy of 13.2 kcal/mole.
Settling and thickening characteristics — Data from zone settling tests
on slurries taken from continuous reactors operated at three different S/X
values are presented in Appendix Tables B9, BIO, and Bll. Figure 14 shows
the solids fluxes calculated from these data as functions of Xfc.
Four flocculent settling tests were performed on three different con-
tinuous culture slurries. Two tests with different initial solids concen-
tration were performed on the slurry with S/X equal to 150. Appendix Tables,
B12 through B15 present the data from these tests. The results were analyzed
using isoconcentration plots on depth Y§ time graphs such as the one presented
in Figure 15. Figure 16 shows the relationship between effluent suspended
solids and overflow rate for these slurries. Table 14 shows the regression
45
-------
TABLE 12. SUMMARY OF CONTINUOUS CULTURE RESULTS AT VARIOUS S/X
Kinetic
parameters
S/X m,a ^^
.mg S . mg N
I *' ^ * '
mg N mg N -Day
194 2.13
142 1.62
100 1.22
56 0.74
45 0.64
Operating Mixed liquor
parameters characteristics
S
e 0 .q s.
C v '" )
(d) (d) *
8 0.26 17.33
(0.75)
10 0..25 , 11.80
(1.70)
13 0.24 15.20
(1.96)
20 0.25 11.30
(0.50)
30 0.24 12.27
(0.24)
X
mg_N*
1 X. '
90
(8)
83
(10)
152
(12)
203
(7)
275
(14)
Effluent Stoichiometric
characteristics parameters
NO -N
0
(n=5)
0.14
(0,13)
0.02
(0.02)
0.01
(0.004)
0.04
(0.02)
NO -N Y .
2 obs
,mg. .mg N .
(a; ( +*>
)fj XT*
mg N
0 . 095
(n=5)
0.14 .071
(0.11)
0.003 .094
(0.003)
0 .084
(n=5)
0.001 .075
(0.0004)
ASO -S
4
96
(8)
94
(2)
93
(6)
100
(9)
93
(12)
AAlk
3.80
(0.14)
3.64
(0.02)
3.54
(0.26)
3.81
(0.12)
3.33
(0.12)
*
Organic-N
Nitrate-N
NOTE: Numbers in parentheses are standard deviations of measurements.
is .zero, the number of replicates is given.
If the measured value
-------
D
50 100
S/X
150 200
m
-------
TABLE 13. SUMMARY OP CONTINUOUS CULTURE RESULTS AT VARIOUS TEMPERATURES
CD
Kinetic Operating
parameters parameters
U S/X
m, a
T mg N , ,mg S v 0 0
o ( * > ( J c
( C) mg N -day mg N (d) (d)
12 0.97 144 15 0.26
21 1.62 142 10 0.25
30 3.92 141 7.6 0.24
Mixed liquor
characteristics
S
25.14
(1.15)
11.80
(1.70)
10.57
(0.53)
X
(HJL,
175
(10)
83
(10)
75
(8)
Effluent Stoichiometric
characteristics parameters
NOT-N NO~-N Y , ASO -S
3 2 obs ^ 4
,m0, ,mg. ,mg N . /mg S.
\/ \ ) \ &•* o
mg N
0.01 0 .100 92
(0.01) (n=6) (2)
0.14 0.14 .071 94
(0.13) (0.11) (2)
0 0 .080 88
(n=6) (n=6) (6)
AAlk
(Sf)
3.87
(0.11)
3.64
(0.02)
3.16
(0.21)
Organic-N
Nitrate-N
NOTE: Numbers in parentheses are the standard deviations of the measurements.
to zero, the number of measurements is given.
For values equal
-------
1.6
1.2
.4
0 -
0
3.30
3.40
TH X I03
3.50
Figure 13. Natural logarithm of maximum attainable unit rate of denitrifica-
tion versus inverse of absolute temperature.
49
-------
2000
~ 1600
^^%
o
•o
£ 1200
in
i 800
en
400
G
S/X = I50
S/X
ft. =30
mg sulfur
mg organic -N/
/
6C - 30 (days)
S/X=56
ft. =20
40,000 80,000
Solids Concentration (mg/JD
120,000
Figure 14. Solids flux versus solids concentration for various values of the
sulfur to biomass ratio.
50
-------
20
22
40
Time(min)
S/X =56
mg sulfur
24
mg organic-N/
X, = 16,200 (mg/jO
60
80
Figure 15. Suspended solids concentration on depth versus time graph with
lines of isoconcentration.
51
-------
300
250
200
.2
150
u
o
o
S 100
e
H 50
«fr->
«*..
UJ
S/X=56
ec =20
S/X = 45
=30
2,000 4,000 6,000 8,000 10,000 12,000 14,000
A
Overflow rate (gal/day-ft )
Figure 16. Effluent suspended solids concentration versus overflow rate for
various values of the sulfur to biomass ratio.
52
-------
TABLE 14. REGRESSION EQUATIONS FOR RESULTS OF SOLIDS SEPARATION TESTS
Solids flux regression equations
S/X Equation
45 G = 568 - (3.49 x 10~3)X
s t
56 G = 1110 - (9.44 x 10~3)X4.
s t
150 G = 2330 - (1.13 x 10~3)X.
s t
G = solids flux due to subsidence, [Ib/ft -day];
S
X = total solids concentration, [mg/£,].
Effluent solids regression equations
S/X
45 X®ff = 14.8 + (6.34 x 10~3)f
fc c
56 xfff = 2.2 + (1.21 x 10~2)^
t A
150 X = -23.6 + (2.70 X 10~2)J
t c
X6 = total solids concentration in effluent, [mg/Jl] ;
2
= overflow rate, [gpd/ft ]
A
c
53
-------
equations derived from results of solids separation tests.
Discussion of Phase I Experimental Results
Observed Yield Coefficients—
A comparison of values of the observed biomass yield (Y) in the
Phase I continuous culture experiments (Tables 11, 12, and 13? indicates only
minor variation of this parameter with growth rate (i ) or S/X.
c
Table 15 presents average values of stoichiometric coefficients measured
in the Phase I batch and continuous culture experiments. Table 15 also shows
predicted values of the coefficients derived from a balanced stoichiometric
equation incorporating the cell formula C5,H7O2N, and the measured average
biomass yield. There is little difference between the two reactor systems
in observed biomass yields, but there is a large difference in the coeffic-
ient for sulfate production (Yfh|~S^' Tnis is probably due to oxygen leaks
in the continuous reactors which allowed excess sulfur to be microbially
oxidized. The continuous reactors were initially adjusted to keep oxygen
leaks, as measured on an electron equivalent basis, to less than 10 percent
of the equivalent nitrate concentration in the feed. During reactor opera-
tion, however, oxygen leaks increased as indicated by increased alkalinity
destruction. An oxygen leak of approximately 30 mg O2/hr was estimated by
assuming that alkalinity destruction and sulfate production occurred with the
same stoichiometry as measured in the batch reactors. This is equivalent to
an increase in the influent nitrate concentration of 11 mg/£ NO~-N. If a
leak of this magnitude actually were to occur, the observed yield would be
calculated as 0.060 mg organic-N/mg NO~-N by equating oxygen and nitrate
electron equivalents.
Balanced Stoichiometric Equation—
An empirical cell mass formula of C-H-CLN is often used in applying
stoichiometric principles to microbial reactions (33,38,41,47). To confirm
the validity of this formula for autotrophic denitrification, stoichiometric
coefficients ys®4-s and YA^k were calculated using measured values of
Y , from Phase f and a cell synthesis equation incorporating the assumed
protoplasm formula. The values shown in Table 15 indicate that YS°4-S and
yAlk are ^n g00<3 agreement with the corresponding predicted values for batch
experiments. However, observed and predicted values of these parameters do
not agree for continuous cultures. The higher values of YS?4-S and YAlk are
probably a result of the oxygen leaks in the continuous reactors. However,
the observed ratio of these parameters (YS?4-S/YA:~ ) agrees well with the
predicted value, indicating that the use or the cell formula CJELjO N is
reasonable for autotrophic denitrification.
The suggested method for calculating reaction stoichiometry for auto-
trophic denitrification is to use the average value of Y , 0.08 mg
organic-N/mg NO~-N, along with a cell formula of C_H_O?N°€o produce the
following balanced stoichiometric equation.
1.0 NO~ + 1.10 S + 0.40 CO- + 0.76 H-0 + 0.080 Nflt , ,
3 2 2 _ 4 (38)
-»• 0.080 C5H O2N + 0.50 N2 + 1.10 SO + 1.28 H
54
-------
TABLE 15. STOICHIOMETKIC COEFFICIENTS FOR BATCH AND CONTINUOUS CULTURES
Continuous,
average
Batch,
average
Continuous ,
predicted
Batch ,
predicted
Y YS04 ~s
obs obs _
,mg organic-N. m
-------
on _c n
ASO.-S = Y :;4 B (CEN -CEN) (40)
4 obs
AAlk = YA?;k (CEN°-CEN) (41)
obs
Sulfur Balance—
A sulfur balance was performed on the Phase I continuous reactor oper-
ated at S/X =? 100 to determine if there were major sulfur oxidation products
besides sulfate. Sulfide odors were never detected in the effluent but other
sulfur by-products could have been present. The material balance on sulfur
in a slurry reactor can be used to predict reactor sulfur concentrations.
CN° AS°4~S S6ff
S = Qc1^ 1" - V3 (42)
The amount of sulfur added to the reactor each day was calculated
according to the desired feed ratio and the flow since the last feeding. This
amount varied somewhat, so a weighted average daily feed rate was calculated.
The proper weighting function was determined by consideration of the concept
of residence time distribution (RTD) (28,29). This function gives the frac-
tion of particles which entered the reactor together, that have left the
reactor at any given time. The function, 1-RTD, will give the fraction still
within the reactor. This function will weight the variable sulfur feed
rates to best estimate reactor sulfur concentration. The RTD for the solids
in a completely mixed reactor with recycle can be obtained by substituting
0 for 0 in the RTD for a completely mixed reactor without recycle.
c
1 - RTD = e"t/Qc (43)
Effluent elemental sulfur and sulfate concentration were measured at
several times before the steady state reactor sulfur concentration was
measured. The average values for S and ASO -S were 11 mg/& S and 104
mg/JZ. S, respectively. The predicted value of S was 14,900 mg/Jl S, which
compared very well with the average measured value of 15,200 mg/£ S.
Composition of Reactor Gas—
No nitric oxide and only a trace of nitrous oxide (50 ppm) was detected
in the Phase I continuous reactor recirculating gas. A trace of nitrous
oxide was also detected in the nitrogen gas added to the reactors, so the
relative amounts of nitrogenous gases produced during autotrophic denitrifi-
cation cannot be firmly established. Absence of significant amounts of
nitrogenous oxides, however, indicates that elemental nitrogen is the primary
end-product of nitrate reduction. This is consistent with previous analyses
which report no production of nitrous or nitric oxide during heterotrophic
denitrification (48-51). No gas analyses were performed in any of the other
experimental phases.
Kinetics--
Results of Phase I continuous culture experiments can be used to calcu-
late kinetic coefficients. The average steady state effluent nitrate-nitro-
gen concentration measured during these experiments was 0.03 mg/£ NO~-N. An
56
-------
exact value for the saturation constant in NFN (K ) could not be calculated
from these measurements because they were too low"to be accurately analyzed.
However, this value (0.03 mg/£) can be used as a conservative estimate of
KR since most reactors were operated near the maximum attainable unit rate.
Values of Kn measured for heterotrophic denitrification (0.06 (52), 0.08
(53), and 0.16 (54) mg/£ NO~-N) are of the same order of magnitude. Because
the estimated value of Kn is so small, effluent nitrate concentrations from
steady state autotrophic denitrification slurry reactors will be negligible -
for all feasible operating conditions. Any steady state reactor which is
operated such that GC > 9^'a should produce an effluent with negligible
nitrate nitrogen.
Equation 4 represents the dependence of U on the ratio S/X as a
saturation function. The exact form of this function is such that at small
values of S/X there is an extended region in which U is proportional to
S/X (Figure 2). The continuous culture experiments with variable S/X were
designed to operate in this region, since the process requirement for elemen-
tal sulfur is reduced at lower values of S/X. Therefore, values of U
measured in the variable S/X experiments would be expected to be lineaffy
related to S/X with a zero intercept.
The results of the variable S/X experiments presented in Figure 12
show an excellent linear relationship between U and S/X with a small, but
not insignificant, non-zero intercept. This benavior could be due to vari-
ations in sulfur particle size among the various experiments, or to uneven
distribution of biomass over sulfur surface area.
The effect of variation in sulfur particle size would be expected in
the series of variable S/X experiments because those reactors were operated
at different values of 0 . Those reactors operated at higher values of QC
would retain the sulfur particles for a longer time, resulting in increased
microbial oxidation and reduced size. Smaller sulfur particles would present
more surface area for microbial growth per unit sulfur mass. Therefore,
the variable which exerts the primary influence on the unit rate of denitri-
fication—the sulfur surface area—would be underestimated by measurement of
sulfur mass in reactors operated at high values of 0 . This error in measure-
ment would tend to shift points at low values of S/XC (high 0c) to the right
in Figure 12 and move the intercept closer to the origin.
The effect of an uneven distribution of biomass could be explained
using the concept of an "effective sulfur" concentration (SQ). Such a vari-
able would represent the concentration of sulfur actually covered with bio-
mass. The kinetic model for autotrophic denitrification assumes that bio-
mass in the reactor is distributed evenly over the available sulfur surfaces.
This would result in a constant biofilm thickness on every sulfur particle.
In practice, however, it is probable that as the amount of sulfur relative
to biomass increases a larger fraction of sulfur will not be incorporated
into a biofilm matrix. At high values of S/X there will be relatively less
microbial "glue" available to capture sulfur particles. Experimental evidence
Of this behavior was observed in the flocculent settling tests, where more
57
-------
suspended solids were found at higher values of S/X (Figure 16). Thus, a
kinetic model might also be developed using S /X rather than S/X as its
primary variable. If these latter assumptions were applied to the experi-
mental results, one could expect that the slope of the regression line re-
lating U to S/X would increase and its intercept would approach zero.
m,a
The kinetics of autotrophic denitrification at high values of S/X were
investigated in semicontinuous reactors operated with different sulfur to
nitrate-nitrogen feed ratios (c). Unfortunately, the experiments were con-
ducted early in the study, before the importance of S/X as the primary
kinetic variable was recognized. However, it is possible to relate the
sulfur to nitrate-nitrogen feed ratio to S/X. The material balance equation
showing the relationship between S/X and c is shown in Table 1. Inspection
of this equation shows that c will be proportional to S/X when c is signifi-
cantly greater than the stoichiometric ratio (v); effluent nitrate-nitrogen
concentration (N) is low; and, the observed biomass yield (Y , ) is constant.
Since these conditions were generally met during the culture characterization
experiments, c can be used as a surrogate measure of S/X. Since the nitrate
feed rate to each reactor was the same, and since the biomass yield should
be constant, biomass concentrations in the reactors should have been equal.
Therefore, gas production rate which was used to represent the rate of de-
nitrification could also be used as a surrogate measure of D
m,a
The results presented in Figure 7 support the prediction of the model
for the behavior of U at high values of S/X, when considered in terms of
the surrogate variables. The results show that in general the relationship
between gas production rate (surrogate for U ) and c (surrogate for S/X) is
that of a saturation function. In particular^ the hyperbolic tangent function
drawn in Figure 7 to represent the relationship is in good agreement with the
experimental results.
Temperature Effects—
Possible effects of diffusional limitations on the observed temperature
dependence of autotrophic denitrification should be considered when comparing
these results with information from other denitrification or sulfur oxidation
systems. The kinetic model developed for autotrophic denitrification pre-
dicts that the observed rate of denitrification will be limited by transport .
processes whenever the rate is a linear function of S/X. The range of S/X
values used in these experiments resulted in a linear relationship between
U and S/X (Figure 12), so the rates measured in this study were probably
limited by intra-film transfer of sulfur. Rates of heterogeneous chemical
reactions measured under conditions of diffusion limitations are known to
exhibit apparent activation energies equal to half the true value (28,29).
Therefore, the activation energy for the actual microbial reaction in auto-
trophic denitrification should be approximately twice the measured value,
_i.js., 26 kcal/mole. Results of temperature dependence studies on the aerobic
oxidation of sulfur by Thiobacillus thiooxidans under conditions of an excess
of sulfur (55) can be used to calculate an activation energy of 23.6 kcal/mole.
If the temperature dependence displayed by this organism is similar to that
of Thiobacillus denitrificans, then the model's conclusion that observed
58
-------
rates of denitrification are limited by intra-film transport of sulfur appears
valid.
Settling and Thickening—
Several techniques for data analysis have been proposed for use with
the batch flux method to describe the solids separation process (30,31,56).
Results of zone settling tests performed on sulfur-biomass slurries indicate
that a linear function best describes the relationship between solids flux
and solids concentration (Figure 14). These results could also have been
analyzed according to the more frequently used logarithmic or semi-logarith-
mic relationships. However, the linear model was used because it resulted
in a somewhat better fit to the data over the range of experimental observa-
tions .
Applying the batch flux analysis technique to a slurry with a linear
solids flux leads to the conclusion that there are no limitations on the
solids separation process due to thickening (57). The process is limited
only by clarification, ,i.e_. , the ability of the slurry entering the clarifier-
thickener to settle at a rate faster than the overflow rate (Q/A ). The
maximum allowable overflow rate set by clarification limitationsCcan be
calculated from results of zone settling tests according to the following
equation (57):
M(xf - X*)
'S/V1^ = —I a" <44>
Xt <
(Q/A ) = maximum allowable overflow rate, [m /m -d] ;
M = slope of solids flux vs total solids concentration
curve, [m/d];
x£ = total solids concentration in recycle line, [mg/£];
X^ = total solids concentration in reactor, [mg/A].
Although zone settling tests predict limits on the operation of solids
separation processes, they do not predict effluent quality. Results of
flocculent settling tests predict the relationship between the operating
variable for solids separation (Q/Ac) and effluent quality (X® )
(Figure 16). These results indicate improved settling properties at lower
values of S/X. For a given overflow rate, predicted effluent suspended
solids concentration decreases as S/X decreases. This is probably due to
the fact that as S/X decreases the amount of biomass relative to sulfur
increases. This increases the chance that a sulfur particle will become
enmeshed in a biomass matrix and be removed with the larger agglomerates.
This observation is consistent with the postulate of an effective sulfur
concentration used previously to explain kinetic behavior.
Initial solids concentration seemed to have relatively minor effect on
the flocculent settling characteristics of sulfur-biomass slurries. Two
initial solids concentrations were used to obtain settling data for the
59
-------
slurry with S/X = 150 mg S/mg organic-N. The open and filled circles in
Figure 16 represent data obtained at initial solids concentrations of
67,000 mg/£ and 35,600 mg/JZ, , respectively. These results were calculated
using a depth of five feet. There is very little indication that separation
efficiency depends on depth, as indicated by the nearly vertical isoconcen-
tration lines in Figure 15.
PHASE II - THIOSULFATE AND SULFIDE EXPERIMENTS
A series of completely mixed semi-continuous flow reactors were employed
in Phase II, to determine whether autotrophic denitrification with T_.
denitrificans using sulfide or thiosulfate as an electron source was feasible.
This experimental phase was carried out in four experiments, as described
below. Table 16 summarizes the experiments performed under Phase II.
TABLE 16. EXPERIMENTAL PROGRAM - PHASE II (STEADY-STATE)
Experiment
number
1
2
3
4
0 (days)
15
10
5
15
10
5
10
10
10
10
cf
17.7
17.7
17.7
4.1
4.1
4.1
2.0
0.8
0.5
1.7
Feed sulfur form
S2°3=
S2°3=
S2°3=
S2°3=
S2°3=
S2°3=
S2°3=
S2°3=
S2°3=
S~
Experiment #1-Effect of Growth Rate - High C
Experimental Plan and Techniques—
The purpose of Experiment #1 was to determine the effect of growth rate
on denitrification in systems which were fed high feed ratios, C of thio -
sulfate. The basic feed solution components used in this experiment are
given in Table 17. To these basic ingredients Na2S2O3 • 5H2O was added to
60
-------
TABLE 17. PHASE II - FEED SOLUTION NUTRIENTS
Concentration
Constituents (milligrams/liter )
25 (as N)
MgCl2 • 6H20 500
FeSO 10
25
300
1,500
NaHC03 1,000
Tap water to one liter volume
make up the appropriate feed ratio. The pH of the feed and the reactor was
maintained at about 7. The reactor in this experiment was seeded from en-
riched cultures of Thiobacillus denitrificans grown on elemental sulfur in
Phase I.
The reactor was operated in a completely mixed, semi-continuous flow
mode. Biomass solids wasting and feeding was done once per day such as to
control the microorganism net specific growth rate (1/0 ) at levels indicated
in Table 16. A gas collection system was used to indicate denitrification
activity and to keep the systems anaerobic. Figure 17 is a schematic of the
reactor. The reactor volume was four liters.
The reactor was run until steady-state conditions were reached, then
steady-state data was collected for a seven-day period. Reactor sampling
was done daily for thiosulfate, nitrite, nitrate, pH and volatile suspended
solids.
Experimental Results—
The steady state experimental results of Experiment #1 - Phase II are
summarized in Table 18. The values shown in this table are the mean of seven
days of steady-state data.
61
-------
-gas line
feed line
J
V
^^M
-=-
^
flection fluid
eservoir
A
V
{*rS
\r
Gas col lection
tube
I c
V
J.
p
L^
Reactor
Agitation
Figure 17. Semicontinuous flow sulfide/thiosulfate reactor.
62
-------
TABLE 18. SUMMARY OF STEADY-STATE EXPERIMENTAL RESULTS - PHASE II
01
u>
Avg. feed data
(ng/4)
Expt. ®c
No. (d)
15
1 10
5
15
2 10
5
10
10
3 10
10
NO~-N
25
25
25
22
22
22
25
26
25
100
s2o=-s
1246
1246
1246
258
258
258
143
55
33
243*
Avg
Cp NO~-N
17.7
17.7
17.7
4.1
4.1
4.1
2.0
0.8
0.5
1.7
<0.2
<0.2
<0.2
<0. 2
<0. 2
<0.2
0.6
1.6
15.0
21.7
. effluent data -
(mg/Jl )
NO--N
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
0.08
1.4
0.43
Volatile
=5 Y
S-O-j-S suspended obs
solids
1120
1101
1126
191
186
172
66
<2.5
<2.5
0.8*
25
34
20
12
9
24
27
34
-
122
1.
1.
0.
0.
0.
1.
I.
1.
-
1.
0
4
8
5
4
0
1
4
6
Measured
CR
1
2
1
1
1
1
1
1
1
2
.8
.1
.7
.1
.1
.4
.1
.3
.3
.2
Sulfide
-------
Experiment #2 - Effect of Growth Rate - Intermediate C
__^ ^ ^
Experimental Plan and Techniques—
Experiment #2 was designed to determine the effect of 6 on denitri-
fication in systems with intermediate feed ratios of thiosulfate. Reactor
operation was the same as for Experiment #1 except that the feed ratio was
lowered to 4.1.
Experimental Results—
The steady-state results of Experiment #2 - Phase II are summarized in
Table 18. The values shown in this table are the mean of seven days of
steady-state data.
Experiment #3 - Determination of Consumptive Ratio
Experimental Plan and Techniques—
The purpose of Experiment #3 was to determine the consumptive ratio/
CR, and to observe system operation under thiosulfate limiting growth con-
ditions. After the steady-state data for this experiment was collected, the
reactor was used to perform a dynamic study to assess the stability of a
thiosulfate system under fluctuating loads (varying C values). In this
experiment a reactor with thiosulfate limiting growth (C = 0.45) was rapidly
changed to a nitrate limiting growth reactor (C = 3.5) and then rapidly back
to a thiosulfate limiting growth system (C_ = 0.45). Of particular concern
here was the effect of nitrite buildup (at low C values) on the responsive-
ness of the denitrification process.
Reactor operation was as described in Experiments #1 and #2, except
that the feed ratio was maintained at the values indicated in Table 16 for
the steady-state study and was allowed to vary, as described above, for the
dynamic study.
Experimental Results—
The steady-state results of Expeirment #3 are summarized in Table 18.
The values shown in this table are the mean of seven days steady-state data.
The results of the dynamic study are presented in Figure 18.
Experiment #4 - Sulfide Experiments
Experimental Plan and Techniques—
Experiment #4 was a feasibility study to determine if sulfide could be
used effectively as an electron donor in autotrophic denitrification. This
expeirment was conducted in semicontinuous reactors, and operated in a
manner similar to Experiment #1, #2 and #3, with the following modifications.
The reactor volume was reduced to one liter, and Na_S • 9H O was
added to the basic feed solution listed in Table 17. The feed ratio was
adjusted to 1.7. Also, the concentration of phosphate buffer was increased
by a factor of ten to compensate for the caustic nature of Na_S.
Experimental Results—
The results of Experiment #4 are summarized in Table 18. The values
64
-------
ft-0.45
50
40
3 30
x
(9
2
20
10
,Cf = 3.5
Cf=0.45
D S203 -S
A NOg-N
O NOj-N
35 40
Figure 18. Response of autotrophic denitrifying system to rapid changes in
feed ratio.
65
-------
shown in this table are the mean of seven days steady-state data.
Discussion of Phase II Experimental Results
Thiosulfate and Sulfide as Electron Donors—
The results of Experiments #1 through #4 indicate that either thiosul-
fate or sulfide can be used effectively as electron sources for autotrophic
denitrification. Essentially complete denitrification can be accomplished
provided the feed ratio is greater than some minimum value (probably the
consumptive ratio). In the thiosulfate system, denitrification was relatively
stable with respect to rapidly fluctuating feed ratios.
Observed Yield and Consumptive Ratio—
The results of Phase II experiments, as summarized in Table 18 indicate
high variability in Y and CR values. Some of this variability can be
attributed to the inherent chemical instability of thiosulfate and sulfide
in solution.
Consider first the thiosulfate ion in solution. The thiosulfate ion
is composed of two sulfur atoms each of which has a different electronic
structure. The thiosulfate ion is formed by the addition of elemental sulfur
to the sulfite anion. Even after reacting the two sulfur atoms (one in the
elemental state and one in the sulfite anion) remain distinguishable. Given
the proper conditions the thiosulfate anion will decompose back to elemental
sulfur and the sulfite anion. One way to make thiosulfate decompose is to
add acid. Under acidic conditions weakly dissociated sulfurous acid or
bisulfite enhances the decomposition of the thiosulfate because it effec-
tively removes the sulfite ion from the product side of the reaction as
illustrated by Equations 45, 46 and 47.
SO~ -»• S° + SO~ (45)
H+ + S 0^ -*- HSO~ + S° (46)
£* -J J
2H+ + S20~ -4- H2S03 + S° . (47)
In an actively denitrifying .culture of Thiobacillus denitrificans, the
utilization of sulfite by T_. denitrif icans as an electron donor has the same
effect on thiosulfate decomposition as does acid.
Decomposition of thiosulfate causes the calculated Y to be errone-
ously high, because the elemental sulfur that is formed gives positive
values in the volatile suspended solids analysis. Elemental sulfur which
is retained on the glass fiber filter is volatilized easily at 560°C
(elemental sulfur boiling point = 450°C). The inherent inaccuracy of the
volatile suspended solids analysis prevents meaningful confirmation of this
sulfur interference. It is possible to state only that most of the high
observed yield values (greater than the predicted 0.7) were observed in the
excess thiosulfate reactors.
66
-------
It should be noted that Y as measured in Phase II is on a volatile
suspended solids basis, while the Y as reported in Phase I is on
ofes
a
nitrogen content of biomass basis. Ho cellular (biomass) nitrogen determina-
tions were made in the Phase II studies.
The observed yields determined in Phases I and II agree reasonably well
with the calculated theoretical Y of 0.084 mg organic-N/mg NO~-N (0.683 mg
VSS/mg NO--N), 0.087 mg organic-N/mg NO~-N (0.704 mg VSS/mg NCf-N), 0.086 mg
organic-N/mg NO~-N (0.703 mg VSS/mg NO~-N), for elemental sulfur, sulfide,
and thiosulfate, respectively.
The experimental determination of CR is also hindered by the decompo-
sition of thiosulfate. CR is supposed to account for only that thiosulfate
utilized as an electron donor. But chemical analysis does not differentiate
between decomposed thiosulfate and biologically consumed thiosulfate. As a
result the "apparent" utilization of thiosulfate by decomposition causes the
experimentally determined CR to be erroneously high. Decomposition of thio-
sulfate is particularly prevalent when there is an excess of thiosulfate in
solution (such as with high C^ values). Results given in Table 18 are in
accord with this explanation. Reactors with large excess thiosulfate (CL. =
17.7) have experimentally observed C values much higher than the predicted
value of 1.35. It is very significant that the two reactors which were
thiosulfate limiting (no excess thiosulfate) exhibited consumptive ratios
very close to the predicted 1.35 value.
It should be noted that the decomposition of thiosulfate has no effect
on the total number of electrons that can be theoretically transferred from a
quantity of thiosulfate. The decomposition is a disproportionation or auto-
oxidation reaction which involves no external electron transfer.
PHASE III - PACKED BED REACTOR EXPERIMENTS
Because sulfur and sulfide appear to represent the most cost-effective
electron sources on an electron equivalent basis, Phase III studies focused
on their use for autotrophic denitrification in packed bed reactors. This
phase had several specific objectives. One objective was to investigate the
effect of sulfur particle size on the minimum hydraulic retention time requir-
ed for complete denitrification, where elemental sulfur is used as a packing
media. Another objective was to investigate the use of dolomitic limestone
as a source of alkalinity in sulfur packed columns. In addition, dolomitic
limestone was investigated as a packing media for packed bed columns which
were fed suifide as an electron source. The final objective was to assess the
influence of organics in the feed solution on the competition between hetero-
trophic and autotrophic denitrification in packed bed reactors. Phase III
studies were conducted in four experiments as described below.
Experiment #1 - Dolomitic Limestone Reactors - Sulfide Feed
Experimental Plan and Techniques—
In this experiment sulfide Was used as an electron donor in columns
packed with dolomitic limestone. Two dolomite packed bed reactors were
67
-------
operated at a hydraulic retention time (6 = bed pore volume/feed flow rate)
of approximately 9.25 hours.
These reactors were operated at different feed ratios. In the first
reactor (Experiment #la) the 'feed ratio was 3.1 and in the second reactor
(Experiment #lb) the feed ratio was changed to 0.96 after an extended period
of growth under nitrate limiting growth conditions at C = 3.1.
A schematic of the system in Experiments #la and Ib used is shown in
Figure 19. A feed solution containing approximately 25 mg/Jl NO -N (see
Table 19) and the appropriate amount of Na?S was adjusted to a pH of 9.5 With
phosphoric acid and placed in the feed tank. This feed was pumped to a
mixing chamber where the pH was adjusted to 7.8 With a pH controller unit.
TABLE 19. PHASE III FEED CHARACTERISTICS (EXPTS. #1, #2 & #3)
Expt.
no.
#1
#2
#3
concentration (mg/£)
KNO -N 25
3
MgCl2 • 6H20
FeSO4 • 7H2O
CaCl2 • 2H O
MnCl,, • 4H~0
50
1
1
1
1
50
1
1
1
1
H3P°4
adjust pH of stock
solution to 9.5
NaHC0
Tap water
to volume
10 10
1000 variable
to volume to volume
The feed solution was initially adjusted to pH 9.5 to prevent loss of H2S
from the feed solution. The feed was then passed upward through a dolomitic
limestone media reactor (dolomite size range was 3 to 13 mm). Dolomite was
selected as the packing because it provided a relatively cheap source of
alkalinity while serving as the reactor support media.
To decrease start-up time enriched cultures of Thiobacillus denitrifi-
cans were developed on a thiosulfate feed solution and then introduced into
68
-------
FEED
SOLUTION
PH CONTROLLER
ACID BASE
MIXING
CHAMBER
PORT
FEED PORT
FEED
PUMP
DOLOMITE
PACKED BED
REACTOR
MIXING
CHAMBER
EFFLUENT
P6RT
Figure 19. Schematic of continuous flow sulfide feed packed bed reactor
system.
69
-------
the packed bed reactors. The cultures adapted very readily to the sulfide
electron donor. So readily in fact, that growth developed in the mixing
chamber as well as in the dolomite reactor.
Steady-state data was collected for a period of twenty-six days. The
reactor feed, mixing chamber and effluent were sampled daily and analyzed
for nitrate, nitrite, pH, alkalinity, sulfate, and sulfide.
Experimental Results—
The steady-state data for the feed, mixing chamber and effluent solution
of reactors in Experiment #1 of this study phase are summarized in Table 20.
TABLE 20. EXPERIMENT #1 - PHASE III STEADY-STATE EXPERIMENTAL RESULTS
Sample
point
Feed
Expt. #la .
Mixing chamber
(C = 3.1)
Effluent
Feed
Expt. #lb
(Cp = 0.96) Mix±ng Chamber
Effluent
NO;-N
(mg/A)
24.0
19.4
1.1
21.2
13.8
<0.5
S -S
(mg/A)
107.3
93.7
20.4
29.1
4.34
15.6
N02-N
(rag/A)
<0.05
<0.05
<0.05
<0.05
8.2
<0.05
This data is the mean of seven days of steady-state operation.
In Experiment #lb the feed solution was switched to a feed ratio of
0.96 after an extended period of operation under nitrate limiting growth
conditions (C =3.1). In the mixing chamber there was a significant accumu-
lation of nitrite which is indicative of electron donor limiting growth
conditions. In the effluent, however, nitrate and nitrite removals were
essentially complete.
Experiment #2 - Elemental Sulfur Packed Bed Reactors
Experimental Plan and Techniques—
In Experiment #2 elemental sulfur was used as both the packing media
and the electron donor in three continuous flow packed bed reactors. Each
of these reactors contained different size sulfur particles as shown in
Table 21. The reactors were operated at a series of hydraulic retention
times ranging from 2.8 to 18.1 hours. A schematic of the units is shown in
Figure 20. A synthetic waste containing 50 mg/ NO~-N (see Table 19) was
passed in an upflow mode through these columns.
70
-------
FEED
SOLUTION
FEED PORT
SULFUR
PACKED
BED
REACTOR
EFFLUENT
PORT
PERISTALTIC
PUMP
Figure 20. Schematic of continuous flow sulfur packed bed reactor system.
71
-------
TABLE 21. PHASE III EXPERIMENT #2 REACTOR CHARACTERISTICS
Sulfur particle
size range (mm)
0 Range (hr)
13 - 19
8.1 - 18.1
7-13
7.1 - 15.2
2-7
2.8 - 11.3
Appreciable difficulty was experienced in developing cultures that
would adhere to and metabolize the elemental sulfur particles. A solution to
this problem was to introduce the culture to the sulfur column and add thio-
sulfate in the feed solution. With time the microorganism population devel-
oped to the point where attachment to the sulfur particle surface occurred.
The readily available thiosulfate was, however, still being oxidized. When
a visible culture had developed the thiosulfate was gradually eliminated
from the feed solution in order to force the organisms to metabolize the
elemental sulfur. Once the population became established no difficulties
were encountered in metabolizing the elemental sulfur.
Each column was operated until steady-state conditions were reached.
Column influent and effluent were sampled daily and analyzed for nitrate,
nitrite, pH, alkalinity and sulfate.
Experimental Results—
The steady-state data for Experiment #2 of this phase is summarized in
Figure 21. Each data point represents the mean of seven days steady-state
data.
Experiment #3 - Sulfur-Dolomite Packed Bed Studies
Experimental Plan and Techniques—
In Experiment #3 of Phase III, two elemental sulfur columns similar to
those used in Experiment #2 were employed. One of the reactors was supple-
mented with dolomitic limestone. This sulfur-dolomite reactor was operated
at a mean hydraulic retention time of 20.2 hours, in an upflow mode as shown
in Figure 20. The other column was packed with only elemental sulfur to be
used as a control. The sulfur particle size range in this reactor was 7 to
13 mm. This reactor was operated at a mean hydraulic retention time of 15.8
hours. A description of the media used in these reactors is given in
Table 22. The feed solution composition for Experiment #3 is listed in
Table 19.
The reactors were started using the procedure described in Experiment
#2 of this study Phase.
72
-------
20
K
z
H
UJ
-Ch—D
A 1.9 -
O 1.27-
D 0.668
27 cm
0.668 cm
238 cm
6 8 10 12
HYDRAULIC RETENTION TIME Chr)
14
16
18
Figure 21. Effluent nitrate concentration as a function of hydraulic reten
tion time for different particle size reactor media.
73
-------
TABLE 22. PHASE III EXPERIMENT #3 REACTOR CHARACTERISTICS
Sulfur media reactors
Sulfur-dolomite
media reactors
Sulfur particle size
range (mm)
Dolomite particle
size range (mm)
Dolomite
Sulfur
Mean 8 (hrs)
mass ratio
7-13
0
15.8
2-7
2-13
0.357
20.2
The columns were operated until steady-state conditions were reached.
Column influent and effluent were sampled daily and analyzed for nitrate,
nitrite, pH, alkalinity and sulfate.
Experimental Results—
The steady-state data for Experiment #3 of Phase III is given in
Table 23 for the sulfur/dolomite system and in Table 24 for the sulfur system.
Each data point represents the average of seven days steady-state operation.
TABLE 23. EXPERIMENT #3 -PHASE III RESULTS - SULFUR/DOLOMITE REACTOR
FEED
Alk
(mg/£ as CaCO_)
309
220
184
97
37
PH
8.1
7.5
7.6
7.3
7.0
NO~-N
(mg/£)
50.5
49.8
49.3
48.5
47.0
EFFLUENT
Alk
(mg/SL as CaCOp
267
219
204
162
128
pH
7.2
7.1
7.1
7.3
7.0
NO~-N
(mg/t)
<0.5
<0.5
<0.5
<0.5
<0.5
N02-N
(ng/4)
<0.05
<0.05
<0.05
<0.05
<0.05
74
-------
FEED EFFLUENT
AJ.K
(mg/£ as CaCO )
309
257
220
pH
8.1
7.6
7.5
(mgA)
50.5
47.7
49.8
Alk
(mg/Jl as CaCO )
244
131
67
PH
6.8
6.5
6.2
N03-N
(mg/£)
<0.5
28.9
11.0
N02-N
(mg/£)
<0.05
13.7
9.0
Experiment #4 Packed Bed Studies with Domestic Wastewater Effluent
Experimental Plan and Techniques—
In the final experiment, Experiment #4, of Phase III, effluent from a
secondary domestic wastewater treatment plant was used as feed to two packed
bed reactors. The study employed two columns from Experiment #3. One of
the columns contained only sulfur (particle size range 7 to 13 mm). The
second column was packed with sulfur and dolomite (particle size range 7 to
13 mm).
The feed to these columns was gradually changed from the synthetic feed
of Experiment #3 to secondary effluent. The secondary effluent had an
average COD of 68 mg/&, an alkalinity of 188 mg/i as CaCO3/ a pH of 7.2 and
a nitrate-nitrogen of 2.5 mg/£. Since the effluent was not nitrified,
supplemental nitrate was added to bring the total N03~N concentration to
25 to 30 mg/Jl.
The sulfur/dolomite column was operated in an upflow continuous flow
mode with a hydraulic retention time of 27.8 hours. The sulfur column was
operated in a similar manner with a hydraulic retention time of 21.1 hours.
The columns were operated for 26 days. Column influent and effluent
were sampled daily and analyzed for alkalinity, pH, nitrate, nitrite, sulfate
and COD.
Experimental Results—
The steady-state data for Experiment #4 of Phase III xs given in
Tables 25 and 26. Each data point is the mean of 14 days of steady-state
operation.
75
-------
TABLE 25. EXPERIMENT #4 - PHASE III
SULFUR/DOLOMITE PACKED BED REACTOR PERFORMANCE WITH
SECONDARY EFFLUENT FEED (9 = 27.8 HRS)
Parameter
Alkalinity (as CaCO,)
NO~-N
NO~-N
SO^-S
COD
Influent
(mg/£)
205
not detect.
27
55
68
Effluent
(mg/A)
201
0.015
0.1
183
32
TABLE 26. EXPERIMENT #4 - PHASE III
SULFUR PACKED BED REACTOR PERFORMANCE WITH
SECONDARY EFFLUENT FEED (0 = 21.1 HRS)
Parameter
Alkalinity (as CaC03)
NO~-N
NO~-N
SVS
COD
Influent
(mgA)
269
not detect.
27
55
68
Effluent
(mg/4)
261
<.05
0.1
183
40
76
-------
Discussion of Phase III Experimental Results
Packed Bed Performance with Sulfide Electron Source—
The results of Experiment #la of Phase III show that complete denitri-
fication can be accomplished using sulfide as an electron donor, in dolomite
packed beds, under nitrate limiting growth conditions (C =3.1). The data
shown in Table 20 shows, however, that even under electron donor limiting
conditions, nitrate removal can be complete. This table shows that there
was a significant accumulation of nitrite in the mixing chamber which is
indicative of electron donor limiting growth conditions. In the effluent,
however, nitrate and nitrite removals were complete. The ability of the
reactor to maintain complete nitrate removal under apparently sulfide limit-
ing growth conditions was probably due to utilization of elemental sulfur that
accumulated within the packed bed under the previous nitrate limiting growth
conditions. This elemental sulfur was probably the result of a partial oxi-
dation of the feed sulfide in Experiment #la. Such an observation indicates
that this system would have stability under fluctuating feed ratio (Cp) con-
ditions. However, if the system was operated for prolonged periods of time
under electron donor limiting growth conditions the sulfur that accumulated
within the reactor would diminish and nitrate removals would deteriorate.
Elemental Sulfur Packed Bed Performance—
The data for Experiment #2 of this Phase (Figure 21) shows that complete
denitrification was obtained with sulfur as the electron donor provided that
a minimum hydraulic retention time was provided. This minimum hydraulic
retention time appears to be a function of the sulfur particle size in the
reactor.
An example of the data obtained for one of the Experiment #2 reactors
is shown in Figure 22. This figure is a plot of effluent quality vs hydraulic
retention time. At long hydraulic retention times nitrate removal is essen-
tially complete, sulfate concentration in the effluent is high, and alkalin-
ity concentration is low relative to the feed concentration. This character-
izes a system that is functioning properly. As the hydraulic retention time
is decreased, a point will be approached at which the system is stressed.
When a system reaches a minimum hydraulic retention time there will be an
increase in effluent nitrate and nitrite, a decrease in effluent sulfate and
an increase in effluent alkalinity as shown in Figure 22.
It is evident from Figure 21 that each reactor with a different sulfur
particle size has a different minimum hydraulic retention time. The minimum
hydraulic retention time required for complete nitrate removal decreases with
decreasing sulfur particle size.
If the sulfur particles in the reactor are assumed to be spherical, with
an average diameter calculated from the sieve size analysis, it is possible
to estimate the sulfur surface area in the reactor. The estimated sulfur
surface area may then be plotted against the minimum hydraulic retention time
for each reactor, as shown in Figure 23. There appears to be strong corre-
lation between reactor sulfur surface area and the minimum hydraulic reten-
tion time required for complete nitrate removal. This suggests that m the
77
-------
X400
E
\—«
O
fr-
ee
t-
z
UJ
O
o
O
200
— — feed alkalinity
— — — — — — feed nitrate
v— — — — — — —' — — —feed sulfate
D NO3 plus NO2 (asN)
o 304-3
A ALKALINITY as CaCOj
4 6 8 10
HYDRAULIC RETENTION TIME Chr)
12
14
Figure 22. Effluent quality as a function of hydraulic retention time,
78
-------
a
O
51
z z
in O
O u!
II
15
S
12
10
0.2 0.4 0.6 0.8 1.0 1.2
ESTIMATED SULFUR SURFACE AREA OF REACTOR MEDIA (m2>
1.4
Figure 23. Minimum hydraulic retention time for complete denitrification as
a function of estimated sulfur surface area.
79
-------
design of a sulfur media packed bed reactor, sulfur surface area is a major
consideration.
Alkalinity Supplementation with Dolomite—
The intent of Experiment #3 of Phase III was to examine the extent to
which dolomitic limestone could supply alkalinity to the denitrifying cul-
tures. Inorganic carbon, in the form of biocarbonate and carbonate usually
buffer these systems against the metabolic addition of hydrogen ion. In-
organic carbon is also used by the denitrifiers as a cellular carbon source.
From Equations 20 and 21 the amount of inorganic carbon required for denitri-
fication biomass synthesis is computed to be 1.07 mg C per mg of NO -N reduc-
ed. The amount of alkalinity consumed is computed to be 4.38 mg of alkalin-
ity as CaCO, per mg NO -N reduced. From these theoretical calculations it
appears that the system buffer capacity will normally be exceeded long before
inorganic carbon growth limiting conditions will develop.
To verify the predicted alkalinity consumption rates, and determine
how much alkalinity could be supplemented by dolomite, both reactors, in
Experiment #2 were initially supplied with substantial alkalinity (in the form
of sodium bicarbonate) in amounts well above that theoretically required.
The feed alkalinity was then gradually reduced in both systems.
In the system with no dolomite supplement the pH was reduced to 6.5
and 6.2 with average feed alkalinities of 257 mg/Jl as CaCO and 210 mg/£ as
CaCO_, respectively. In both cases denitrification efficiency was greatly
reduced and nitrite accumulation became apparent. Some acclimation of the
microorganisms to these low pH values, however, was noted as the systems were
operated beyond the test period. Baalsrud and Baalsrud (58) suggest that the
lower pH limit for Thiobacillus denitrificans is approximately 6.2.
In the packed bed reactor, with dolpmite supplement, the feed alkalin-
ity was lowered from 309 to 37 mg/£ as CaCO. with no significant pH depres-
sion or denitrification hinderance. It appears that dolomite has the ability
to supply significant alkalinity when the feed alkalinity is very low. The
amount of alkalinity that is supplied by the dolomite in an actively denitri-
fying reactor is a function of the feed nitrate and feed alkalinity concen-
trations. From the experimental data it appears that the dolomite supplies
enough alkalinity to compensate for biological alkalinity comsumption plus an
additional amount which is controlled by solubility phenomena. Apparent
alkalinity consumption is determined by the difference between feed and
effluent alkalinity concentration. When the feed alkalinity becomes low
enough the packed bed reactor with dolomite experienced negative apparent
alkalinity consumption (alkalinity production). This is demonstrated in
Figure 24. An explanation of this phenomena is that in high alkalinity feed
solutions the concentration gradient between the dolomite surface and bulk
solution is low, hence dissolution of limestone is not significant. However,
when the concentration gradient is high, as in the case of low alkalinity
waters, dissolution of limestone is enhanced.
It is not possible to verify the predicted alkalinity consumption in
80
-------
300
200
100
-100
-50
50
100
APPARENT ALKALINITY CONSUMPTION (mg/l as CaCO3)
Figure 24. Feed alkalinity versus apparent alkalinity consumption for sulfur
packed bed reactors supplemented with dolomite.
81
-------
the dolomite supplemented systems. However, in the sulfur column reactors
in both Experiment #2 and #3 the amount of alkalinity consumed (as CaCO ) per
amount of NO~-N reduced averaged 3.7. This compares well with the theoretical
value of 4.4 when the low accuracy and precision of the alkalinity analysis
is considered.
Although the use of dolomite seems to be a cheap and effective method
to compensate for alkalinity consumption in the autotrophic denitrification
process, it has the disadvantage of increasing the hardness of effluents.
In most cases, the incremental hardness is not a significant consideration.
Denitrification of Domestic Secondary Effluent—
In Experiment #3 of this third experimental phase, it was apparent that
heterotrophic denitrification was proceeding simultaneously with autotrophic
denitrification when secondary effluent was used as feed to the packed bed
reactors. The presence of heterotrophic denitrification is indicated by the
relatively low amount of SCT-S produced per amount of NO~-N reduced as shown
in Table 27. This table compares the amount of sulfate produced per amount
TABLE 27. A COMPARISON OF MEASURED AND THEORETICAL SULFATE
PRODUCTION TO NITRATE REDUCTION RATIO
mg SO.-S produced per mg NO^-N reduced
Measured,
Theoretical secondary effluent
Sulfur/dolomite 7.6 4.7
Sulfur 7.6 4.7
of nitrate reduced for a theoretical prediction (no organic matter in feed),
and for the secondary effluent feed system. The amount of S0~-s produced per
amount of NO~-N reduced can be calculated from Equations 20 and 21 to be
7.6. The results of Experiment # 4 show that this parameter averages to be
4.7 mg SO" -S per mg NO^-N reduced. It appears, therefore, that organic matter
was being utilized as an electron donor, as well as sulfur.
No problems were encountered with biological production of sulfide when
secondary effluent was used as the feed to the packed bed systems. No
suspended solids clogging was experienced during any of the experiments.
82
-------
SECTION 8
ENGINEERING SIGNIFICANCE
Autotrophic denitrification using sulfur compounds provides an alter-
native method for biological denitrification. Based on results of these
bench-top studies, the process appears to offer advantages of decreased cost
of electron donor while being independent of rising petrochemical prices.
However, autotrophic denitrification is not without disadvantages. It en-
riches the wastewater with sulfate and normally would require supplemental
additions of alkalinity to offset microbial acid production. \
EVALUATION OF SULFUR-SUBSTRATES
A wide variety of reduced sulfur compounds are potential substrates
for autotrophic denitrification. Evaluation of the relative merits of these
compounds involves estimation of cost and consideration of ease of storage
and handling. The relative cost of various sulfur substrates depends on
their initial cost of purchase and the stoichiometric amount of the compound
required for nitrate removal. The exact amount required depends on unknown
reaction stoichiometry, but since nitrate removal is accomplished by a micro-
bially mediated oxidation-reduction reaction, a reasonable comparison of
the sulfur compounds can be made on an electron equivalent basis. Table 28
shows a comparison of the cost of various sulfur substrates and methanol
expressed on an electron equivalent basis. An electron equivalent is that
quantity of a compound which donates or accepts one mole of electrons in the
oxidation-reduction reaction in question. For example, consider the half-
reaction involved in the oxidation of elemental sulfur by nitrate.
Oxid. : 0.167 S + 0.667 HO -*- 0.167 SO~ + 1.33 H+ + e~
Red. : 0.2 N0~ + 1.2 H + e -> 0.10 NZ + 0.6 H2O
Overall: 0.167 S +0.2 N0~ + 0.067 H20 -*• 0.10 N2 + 0.167 SO~ + 0.13 H
Since these reactions are written on the basis of one electron transfers, it
is simple to calculate the electron equivalents of sulfur and nitrate. One
electron equivalent of nitrate will be 0.2 moles of nitrate, or 0.28 grains
of nitrate-nitrogen. One electron equivalent of sulfur will be 0.167 moles
of sulfur, or 5.33 grams of sulfur. Therefore, the electron content of
elemental sulfur will be 1 electron equivalent/5.33 grams = 0.188 electron
equivalents/gram.
83
-------
TABLE 28. COMPARISON OF COSTS OF SULFUR SUBSTRATES AND METHANOL
**
Compound
Unit cost
($/100 Ibs)
Electron content
(electron-
equivalents/grain)
Cost/electron
equivalent
(^/electron equivalent)
Anhydrous
sodium
sulfite
11
1.58 x 10
-2
1.53
Sodium
thiosulfate 8.30
pentahydrate
Sodium
thiosulfate 10.95
anhydrous
Sodium
sulfide 12.50
flake
3.23 x 10
-2
5.05 x 10
-2
3.33 x 10
•—2
0.567
Q.477
0.826
Hydrogen
sulfide
10
0.235
0.094
Sulfur
crude
flour
Methanol
4.70
7.59
0.188
0.188
0.055
0.089
($/102lbs) x (2.2 x 10~2) = ($/kg).
**
Reference 59.
84
-------
From Table 28, it is apparent that elemental sulfur is the least expensive
sulfur source since it has the lowest initial cost and is relatively highly
reduced. Hydrogen sulfide is more highly reduced but its higher initial
cost makes it less attractive. Sulfite and thiosulfate are expensive due to
high initial cost and low electron content. The high cost of these compounds,
however, might be avoided if industrial wastes containing them are available.
Elemental sulfur has the additional advantage that it is easy to store
and handle.: It is non-toxic, water-insoluble and stable under normal con-
ditions. A major disadvantage of sulfide is its toxicity. Storage and
handling of sulfide would be more difficult than for sulfur since loss of
sulfide through leaks in storage or effluent losses due to overdosing could
cause a serious health or odor problem. Thiosulfate and sulfite should be
relatively easy to store and handle but larger storage volumes would be
required.
Elemental sulfur is produced primarily by Frasch process mining and
recovery of sulfur compounds from natural gas, petroleum and coal (60,61).
Sulfide can be produced by recovery from natural gas and petroleum (60),
hydrogenation of elemental sulfur (61), and reduction of sulfate with coal
(60,61). Sulfide can also be obtained as a constituent of various indus-
trial wastes such as black liquor from Kraft paper mills (60). Methods for
the commercial production of thiosulfate include reaction of elemental
sulfur with sulfite and reaction of sulfide with sulfur dioxide (60). Thio-
sulfate can also be found in some industrial wastes such as those from
petroleum refineries (62). Absorption of sulfur dioxide in alkaline solu-
tion is the primary method of sulfite production (60). Sulfur dioxide for
this, process can be obtained during burning elemental sulfur or roasting
sulfidic ores (60,61). Wastewaters from sulfite-pulping operations or other
bleaching processes also contain sulfite (60).
Another potential source of reduced sulfur compounds is recovery of
sulfur from emissions of sulfur oxides at power plants burning coal. Two
sulfur oxide removal processes, which have been certified as being ready for
commercial use by 1978, produce a stream of concentrated sulfur dioxide
(63). The sulfur dioxide in this stream could be reduced to elemental sulfur
by reaction with coal (64). The potential magnitude of this source can be
seen by comparing the estimated power plant emissions of sulfur dioxide in
1968 (12.2 million tons) to the total U.S. sulfur production for that year
(10.4 million tons) (65,66).
TECHNICAL FEASILIBITY
The ultimate feasibility of any wastewater treatment process depends
on its economic, social, and environmental costs relative to alternative
processes. Accurate feasibility analyses, however, usually require pilot
plant testing of the process. A less detailed analysis of the technical
feasibility of a process can be based on the results of bench-top experiments.
such as those conducted during this study. A technical feasibility analysis
should determine whether a process can be expected to operate under reason-
able technical constraints and whether pilot plant tests are warranted to
85
-------
determine the feasibility of full scale operation.
Because heterotrophic denitrification using methanol is the most common
denitrification process in present use (1978), it is reasonable to use it in
a technical feasibility analysis as the standard of comparison for evalu-
ating autotrophic denitrification using elemental sulfur. The elemental
sulfur-slurry type bench-top studies on autotrophic denitrification reported
here, were conducted using a hydraulic retention time of six hours, which
is slightly higher than that used in heterotrophic processes (67). Hetero-
trophic processes, however, usually required and additional aeration basin
to remove overdoses of organic carbon. Autotrophic processes would not re-
quire .aeration since sulfur is a water-insoluble substrate and is recycled
with biomass. Autotrophic processes can be operated^at somewhat higher over-
flow rates than the value of 49 m /m -d (1200 gpd/ft ) recommended for hetero-
trophic processes (68). Since the values of these parameters (9, Q/A ) are
roughly comparable, autotrophic denitrification can be considered technically
feasible with respect to both reactor and clarifier-thickener sizes. Pilot
plant studies on autotrophic denitrification would be required to more fully
evaluate equipment cost comparisons. Similarity in the requirements for
capital expenditures indicates that the best criterion for comparison of the
two processes is the operating cost of purchasing the electron donor.
COST OF ELECTRON DONOR
The cost of methanol required to treat a wastewater with 30 mg/£ NO -N
can be estimated as 1.2<:/m ($0.046/1000 gal). This calculation is based on
a bulk methanol price of $l.l
-------
long as those used in the slurry reactors. Hydraulic retention time is not
a primary design parameter for slurry systems, but the values used in this
study are probably representative of those that would be adopted for full
scale operation.
Dependence of process economics on the cost of sulfur indicates the
usefulness of a review of recent sulfur market trends. Within the last two
years, the cost of sulfur has ranged from a high of $0.12/kg ($5.60/100 Ib)
to a low of $0.055/kg ($2.50/100 Ib) (70). The cost advantage of a hetero-
trophic slurry system using methanol would be reversed if the lower price
had remained stable at 0.42 vs 1.22
-------
In the packed bed systems sulfur does not leave the columns other than
in the S0~ form. Hence, sludge disposal is not a problem.
ENVIRONMENTAL IMPACT
Sulfur oxidation in the autotrophic denitrification process will enrich
the wastewater in sulfate. Increased sulfate levels might decrease water
quality directly or through the potential for sulfide production. Reaction
stoichiometry predicts an increase of 75 mg/& SCT-S in a wastewater contain-
ing 30 mg/£ NO -N. An average domestic wastewater contains about 25 mg/Jl
SO^-S, (38,71) so most of these wastewaters would not meet the old drinking
water standard of 83 mg/Jl SO^-S (72) after denitrification with elemental
sulfur. The old standard does not appear to be based on taste or on any other
physiological basis, except for a laxative effect on new users (73). Present
U.S. drinking water standards do not contain a sulfate regulation (74).
Therefore, a sulfate concentration of about 100 mg/Jl SO^-S should not cause
serious deterioration of drinking water quality, especially when dilution,
effects are considered.
Certain microorganisms can reduce sulfate to sulfide in natural water
systems if sufficient organic carbon is available under anaerobic conditions.
Production of sulfides cause an odor problem and precipitation of heavy
metals. Insolubilization of some highly toxic metals such as mercury would
be beneficial. Formation of metallic sulfides could theoretically cause
release of phosphorus previously immobilized by the metal. Experimentation,
however, has not shown this to occur (75).
Nitrogen removal is especially important in coastal and estuarine
waters where the effects of sulfate enrichment would be negligible due to
interactions with seawater containing high concentrations of sulfate (900
mg/£ SO=-S) (76). Although sulfate enrichment does not improve water quality,
there does not seem to be sufficient evidence of harmful effects to generally
restrict the application of autotrophic denitrification processes.
SUMMARY
In summary, autotrophic denitrification compares favorably with present-
ly employed heterotrophic denitrification processes. At this time (1978),
elemental sulfur appears to be a lower cost substrate than methanol. The
cost of supplemental alkalinity is not excessive. Another advantage of auto-
trophic denitrification is that sulfur addition does not require precise
control, since overdoses will not appear in the effluent. Development of
new sludge handling techniques does not appear to be necessary for elemental
sulfur-biomass slurries. Further, the problem of sulfate enrichment does not
appear to be serious enough to restrict application of autotrophic denitrifi-
cation. Therefore, autotrophic denitrification using elemental sulfur appears
to be a technically feasible wastewater treatment process and should be
evaluated in pilot plant or full scale studies.
88
-------
REFERENCES
1. Baalsrud, K. , K. S. Baalsrud, "Studies on Thiobac il lus Den itrificans",
Arch. Microbiol., 20:34-62, 1954.
2. Buchanan, R. E. , N. E. Gibbons, eds. , Sergey's Manual of Determinative
Bacteriology, 8th Ed., Williams and Wilkins, Co., Baltimore, 1974.
3. Verhoeven, W. , "Studies on True Dissimilatory Nitrate Reduction: V
Nitric Oxide Production and Consumption by Micro-organisms", Antonie
Van Leeuwenhoek , 22:383-406, 1956.
4. T'yul'panova-Mosevich, M. V., "Denitrifikatsiya na Neorganicheskoi
Sr^de (Denitrification in Inorganic Media)", Arkhiv Biologicheskikle
Nauk, 30(2): --- , 1930, reported in Sokalova, G. A., G. I. Karavaiko,
Physiology and Geochemical Activity of Thiobacilli, Israel Program of
Scientific Translation, Jerusalem, 1968.
5. Kramarenko, L. E. , I. I. Prisronova, "Denitrifying Sulfur Oxidizing
Bacteria in Sulfide Veins and Methods of Demonstrating Them in
Prospecting", Proc. All-Union Geol. Res. Inst. USEGEI, 61:209-230,
1961, reported in Buchanan, R. E. , N. E. Gibbons, eds., Sergey ' s
Manual of Determinative Bacteriology, 8th Ed. , Williams and Wilkins
Co. , Baltimore, 1974.
6. Panf P. C., "Growth of Three Obligate Autotrophs, Thiobacillus
Thioparus, Thiobacillus Neopolitanus , Thiobacillus Denitrificans, on
Glucose", Bacteriol. Proc . , 70:125, 1970.
7. Borichewski, R. M., "Keto Acids as Growth Limiting Factors in Auto-
trophic Growth of Thiobacillus Thiooxidans" , J. Bacteriol. , 93(2):
597-599, 1967.
8, Karavaiko, G. I. , S. A. Moshniakova, "Oxidation of Sulphide Minerals by
Thiobacillus Thiooxidans" , Microbiology, 43 (1) : 156-158, 1974.
9, Taylor, B. F. , D. S. Hoare, S. L. Hoare, "Thiobacillus Denitrificans as
an Obligate Chemolithotroph : Isolation and Growth Studies", Arch.
Migrobiol., 78(3) :193-204, 1971.
10. Suzucki, I., "Mechanism of Inorganic Oxidation and Energy Coupling",
Annu. Rev. Microbiol. , 28:85-119, 1974.
11. Peck, H. D. , "Sulfur Requirements and Metabolism of Microorganisms",
in Muth, O. H., ed., Symposium; Sulfur in Nutrition, AVI Publ. Co.,
Westport, Conn., 1970.
89
-------
12. Trudinger, P. A., "The Metabolism of Inorganic Sulphur Compounds by
Thiobacilli", Rev. Pure Appl. Chem., 17:1-24, 1967.
13. Bowen, T. J., F. C. Happold, B. F. Taylor, "Studies on Adenosine-51
Phosphosulphate Reductase from Thiobacillus Denitrificans", Biochim.
Biophys. Acta, 118:566-576, 1966.
14. Bowen, T. J., P. J. Butler, F. C. Happold, "Some Properties of the
Rhodanase System of Thiobacillus Denitrificans", Biochem. J., 97:651-
657, 1965.
15. Suzucki, I., "Oxidation of Elemental Sulfur by an Enzyme System of
Thiobacillus Thiooxidans", Biochim. Biophys. Acta, 104(2):359-371,
1965.
16. Baldensperger, J.., L. J. Guarraia, W. J. Humphreys, "Scanning Electron
Microscopy of Thiobacilli Grown on Colloidal Sulfur", Arch. Microbiol.,
99(4):323-329, 1974.
17. Ishaque, M., M. I. H. Aleen, "Intermediates of Denitrification in
Thiobacillus Denitrificans", Bacteriol. Proc., 72:175, 1972.
18. Adams, C. A., G. M. Warnes, D. J. D. Nicholas, "A Sulphite-Dependent
Nitrate Reductase from Thiobacillus Denitrificans", Biochim. Biophys.
Acta, 235:398-406, 197.
19. Aminuddin, M., D. J. D. Nicholas, "Sulphide Oxidation Linked to the
Reduction of Nitrate and Nitrite in Thiobaci1lus Denitrificans",
Biochim. Biophys. Acta, 325(1):81-93, 1973.
20. Aminuddin, M., D. J. D. Nicholas, "An AMP Independent Sulphite Oxidase
from Thiobacillus Denitrificans: Purification and "Properties", J. Gen.
Microbiol., 82:103-113, 1974.
21. Aminuddin, M., D. J. D. Nicholas, "Electron Transport during Sulphide
and Sulphite Oxidation in Thiobacillus Denitrificans", J. Gen. Micro-
biol. , 82:115-123, 1974.
22. Peeters, T., M. I. H. Aleem, "Oxidation of Sulfur Compounds and
Electron Transport in Thiobacillus Denitrificans", Arch. Microbiol.,
71(4):319-330, 1970.
23. Peeters, T., M. Ishaque, M. I. H. Aleem, "Tetrathionate Oxidation in
Thiobacillus Denitrificans", Bacteriol. Proc., 73:194, 1973.
24. Aubert, J. P., G. Milhaud, J. Millet, "L1Assimilation de L'Andydride
Carbonique par les Bacteries Chimautotrophies", Ann. Inst. Pasteur,
Paris, 92:515-524, 1957.
25. Peeters, T., M. S. Liu, M. I. H. Aleem, "The Tricarboxylic Acid Cycle
in Thiobacillus Denitrificans and Thiobacillus A2", J_. Gen. Microbiol.,
90
-------
65(1): 29-35, 1970.
26. Kelly, D. P. "Autotrophy: Concepts of Lithotropic Bacteria and Their
Organic Metabolism", Bacteriol. Rev., 25:177-270, 1971.
27. Gram, A. L., "Feasibility of Bacterial Reduction of Nitrates in Sulfur
Columns", Final Report prepared for Federal Water Pollution Control
Administration, Dept. of the Interior, Washington, D. C., Contract No.
14-12-125 (1968).
28. Smith, J. M., Chemical Engineering Kinetics, McGraw-Hill, New York,
1970.
29. Levenspiel, O., Chemical Reaction Engineering, 2nd Ed., Wiley, New York
1972.
30. Dick, R. I., "Evaluation of Activated Sludge Thickening Theories",
J. Sanit. Eng. Div. Am. Soc. Civ. Eng., 93(SA4):9-29, 1967.
31. Vesilind, P. A., "Design of Prototype Thickeners from Batch Settling
Tests", Water and Sewage Works, 115(7):302-307, 1968.
32. Clark, J. W., W. Viessman, M. J. Hammer, Water Supply and Pollution
Control, 2nd Ed., International Textbook Co., Scranton, Pa. 1971.
33. McCarty, P. L., "Stoichiometry of Biological Reactions", in Proceeding
of the International Conference Toward a Unified Concept of Biological
Waste Treatment, Atlanta, Ga., 1972.
34. Lawrence, A. W., P. L. McCarty, "A Unified Basis for Biological Treat-
ment Design and Operation", J. Sanit. Sng. Div. Am. Soc. Civ. Eng. , 96
(SA3):757-778, 1970.
35. Heukelekian, H. , H. E. Orford, R. Manganelli, "Factors Affecting the
Quantity of Sludge Production in the Activated Sludge Process", Sew, and
Ind. Wastes, 23(8):945-958, 1951.
36. Weston, R. F., W. W. Eckenfelder, "Applications of Biological Treatment
to Industrial Wastes: I Kinetics and Equilibria of Oxidative Treat-
ment", Sew, and Ind. Wastes, 27 (7) -.802-820, 1955.
37. van Uden, N., "Kinetics of Nutrient Limited Growth", Annu. Rev. Micro-
biol., 23:473-486, 1969.
38. Metcalf & Eddy, Inc., Wastewater Engineering, McGraw-Hill, New York
1972.
39. Sherrard, J. H., E. D. Schroeder, "Cell Yield and Growth Rate in
Activated Sludge", J. Water Pollut. Control Fed., 45(9):1889-1897,
1973.
91
-------
40. Sherrard, J. H., L. D. Benefield, "Elemental Distribution Diagrams
for Biological Wastewater Treatment", J. Water Pollut. Control Fed.,
48(3):562-569, 1976.
41. McCarty, P. L., "Energetics and Bacterial Growth", in Organic Compounds
in Aquatic Environments, S. D. Faust, J. V. Hunter, eds., Marcel
Decker, Inc., New York, 1971.
42. U. S. Environmental Protection Agency, Methods Development and Quality
Assurance Research Laboratory, NERC, Cincinnati, Methods for Chemical
Analysis of Water and Wastes, 1974.
43. Standard Methods for the Examination of Water and Wastewater, 13th Ed.,
American Public Health Association, Washington, D. C., 1971.
44. Nelson, P- 0., Adenosine Triphosphate as a Measure of Activated Sludge
Viability, M. S. Thesis, Dept. of Civil and Environmental Engineering,
Cornell University, May 1973.
45. Stumm, W., J. J. Morgan, Aquatic Chemistry, Wiley-Interscience, New
York, 1970.
46. Stensel, H. D., R. C. Loehr, A. W. Lawrence, "Biological Kinetics of
Suspended-Growth Denitrification", J. Water Pollut. Control Fed., 45
(2):249-261, 1973.
47. Forges, N., L. Jasewicz, S. R. Hoover, "Principles of Biological
Oxidation", in Biological Treatment of Sewage and Industrial Wastes,
Vol. I: Aerobic Oxidation, J. McCabe, W. W. Eckenfelder, Eds., Reinhold
Pub. Co., New York, 1956.
48. Johnson, W. K., G. J. Schroepfer, "Nitrogen Removal by Nitrification
and Denitrification", J. Water Pollut. Control Fed. , 36(8):1015-1036,
1964.
49. Seidel, D. F., R. W. Crites, "Evaluation of Anaerobic Denitrification
Processes", J. Sanit. Eng. Div. Am. Soc. Civ. Eng., 92(SA2):267-277,
1970.
50. Mechalas, B. J., P. M. Allen, W. W. Matyskiela, A Study of Nitrifica-
tion and Denitrification, Federal Water Quality Administration, Dept.
of the Interior, Washington, D. C., 1970.
51. Sikora, L. J., D. R. Keeney, "Evaluation of a Sulfur-Thiobacillus
Denitrificans Nitrate Removal System", in press, J. Environ. Qual.
52. Requa, D. A., E. D. Schroeder, "Kinetics of Packed-Bed Denitrification",
J. Water Pollut. Control Fed., 45(8):1696-1707, 1973.
92
-------
53. Moore, S. F., E. D. Schroeder, "The Effect of Nitrate Peed Rate on
Denitrification", Water Res., 5:445-452, 1970.
54. Engberg , D. J., E. D. Schroeder, "Kinetics and Stoichiometry of
Bacterial Denitrification as a Function of Cell Residence Time", Water
Res., 9(12):1051-1054, 1975.
55. Starkey, R. L., "Concerning the Physiology of Thiobacillus Thiooxidans,
an Autotrophic Bacterium Oxidizing Sulfur Under Acid Conditions",
J. Bacteriol., 10:135-195, 1925.
56. Dick, R. I., K. W. Young, "Analysis of Thickening Performance of Final
Settling Tanks", Proc. Ind. Waste Conf., 27:33-54, 1972.
57. Batchelor, B., Autotrophic Denitrification Using Sulfur Electron Donors,
Ph.D. Thesis, Dept. of Civil and Environmental Engineering, Cornell
University, September, 1976.
58. Baalsrud, K., K. S. Baalsrud, "Studies on Thiobacillus Denitrificans"
Arch. Microbiol., 20:34-62, 1954.
59. Chemical Marketing Reporter, September 29, 1975, reported in Driscoll,
C. T., Use of Thiosulfate and Sulfide as Electron Donors in Autotrophic
Denitrification, M. S. Thesis, Dept. of Civil and Environmental Engi-
neering, Cornell University, September 1976.
60. Shreve, R. N., The Chemical Process Industries, McGraw-Hill, New York,
1956.
61. Sittig, M. , Inorganic Chemical and Metallurgical Process Encyclopedia,
Noyes Development Corporation, Park Ridge, N.J., 1968.
62. Garrision, W. E., J. G. Kremer, J. Murk, "Improved Hypochlorination
Techniques and Problems in Disinfection of Municipal Wastewaters
Containing Refinery Thiosulfate", Ind. Waste Conf., 28:309-322, 1973.
63. Sulfur Oxide Control Technology Assessment Panel (SOCTAP), Projected
Utilization of Stack Gas Cleaning Systems by Steam-Electric Plants,
Federal Interagency Committee for Evaluation of State Air Implemen-
tation Plans, Washington, D. C., 1973.
64. Bischoff, W. F., P. Steiner, "Coal Converts SO2 to S", Chemical
Engineering, 82(l):74-75, 1975.
65. National Air Pollution Control Administration, Nationwide Inventory
of Air Pollutant Emissions 1968, Washington, D. C.
66. Manderson, M. C., "The Sulfur Outlook", in Sulfur and SO2 Developments,
AIChE, New York" 1971. '"
93
-------
67. U. S. Environmental Protection Agency, Technology Transfer Division,
Process Design Manual for Nitrogen Control, Washington, D. C., 1975.
68. U. S. Environmental Protection Agency, Technology Transfer Division,
Nitrification and Denitrification Facilities Wastewater Treatment,
Washington, D. C., 1973.
69. Chemical Marketing Reporter, March 22, 1976.
70. Chemical Marketing Reporter, July 7, 1975.
71. Schroeder, H. A., "Relation Between Mortality from Cardiovascular
Disease and Treated Water Supplies", J. Am. Med. Assoc., 172(17):1902-
1908, 1960.
72. U. S. Public Health Service, "Drinking Water Standards", Federal
Register, 27(44):2152, 1962.
73. McKee, J. E., H. W. Wolf, Water Quality Criteria, 2nd Ed., The Resour-
ces Agency of California, State Water Quality Control Board, Publi-
cation No. 3-A, 1963.
74. U. S. Environmental Protection Agency, "National Interim Primary
Drinking Water Regulations", Federal Register, 40(248):59566-59588,
1975.
75. Olson, D. M., The Effect of Sulfate and Manganese Dioxide on the
Release of Phosphorus from Lake Mendota Sediments, Water Resources
Center, Wisconsin University, Madison, Wis., abstracted in Government
Reports Announcements S Index, 75(21):66, 1975.
76. Riley, J. P., G. Skirrow, Chemical Oceanography, Vol. 1, Academic
Press, New York, 1965.
94
-------
APPENDICES
APPENDIX Al. ELEMENTAL SULFUR ANALYSIS
A. Reagents
1. Sodium Sulfite, 100 g/£
2. Formaldehyde solution, 37%; with 10-15% methanol as preservative
\
3. 0.25 N Iodine solution
4. 0.025 N Iodine solution
5. 1.0 1) Sodium Thiosulfate
6. Acetic Acid
7. Starch indicator
8. Antifoam spray
I
B. Procedure
1. Pipette an aliquot from a well stirred sample into a 250 ml erlen-
meyer flask using a broken tipped pipette.
2. Add 50 ml sodium sulfite solution. If more than 300 mg sulfur are
in the sample, add 100 ml.
3. Add boiling chips to flask, spray with antifoam, and place flask on
heating apparatus beneath a hood.
4. Boil for 15 minutes past the time when no sulfur particles are
visible.
5. Cool, add 10 ml formaldehyde solution and 10 ml acetic acid. Add
20 ml of both reagents if 100 mi, of the sodium sulfite solution
was added.
6. Titrate contents of flask, or an aliquot of the contents, with 0.25
Ni iodine solution, until the light yellow iodine color begins to
appear. Add a few drops of starch solution and continue titration
until the blue color is stable for a few seconds. If less than
95
-------
10 mg sulfur is expected, titrate with 0.025 IJ iodine solution.
7. Standardize iodine solution by titrating a known volume of a
standard III thiosulfate solution using the same amounts of sodium
sulfite, formaldehyde, and acetic acid as the samples.
8. Calculate sulfur concentration in mg/& from:
= (ml titrated) (normality of iodine solution) (32,000)
(ml of sample)
APPENDIX A2. ATP ANALYSIS
A. Measurement Procedure
1. Add 2 ml cocktail mixture to glass scintillation vial.
a. cocktail mixture
10 ml 0.2 M Sodium Arsenate
20 ml 0.16 M Magnesium Sulfate
10 ml 0.2 M Tris Buffer (pH = 7.75)
^ 5 ml FLE Solution
b. FLE solution preparation
1) remove vile of FLE-50 (Sigma Chemical Co.) from freezer and
rinse contents with five 1 ml portions of distilled de-
ionized water.
2) place in manual homogenizer, put in ice-water bath and
homogenize for five minutes.
3) let sit in bath for 30 minutes, homogenize for 5 minutes,
let sit for 2 hours.
4) filter through 0.45 ym membrane filter.
2. Dilute extracted sample so that concentration is in range of 0.02-
0.20 yg ATP/0.1 ml.
3. Add 0.1 ml of extracted sample (diluted if necessary), to vial,
start stopwatch, swirl, place in scintillation counter.
4. After 15 seconds, switch scintillation counter to "repeat."
5. Use 5th count as measurement.
96
-------
6. Make set of ATP standards and analyze during analysis period to
detect deactivation of enzyme.
7. Make standard curve, adjusting for enzyme deactivation if necessary,
and calculate ATP content in samples.
B. ATP Extraction Procedure
1. Filter sample through 0.45 pm membrane filter.
2. Place filter in test tube containing 7 ml Tris Buffer (0.02 M, pH
7.75) which has been in boiling water bath.
3. Replace in boiling water bath, mix occasionally.
4. After 10 minutes, remove, and cool rapidly.
5. Freeze, if analysis is not immediate.
C. Scintillation Counter Settings
1. Main power- high voltage (h.v.)
2. Preset time - 0.1 minute
3. Mode selector - auto
4. Preset count - 900
5. Gain - 100
6. Window - C-D; C = 5, D = 100
7. Sample changer - "STOP", then "RESET"
8. Coincidence switch - "OFF"
97
-------
APPENDIX B. TRANSIENT RATE TEST DATA, RESULTS OP ZONE SETTLING TEST, RESULTS
OF FLOCCULENT SETTLING TEST.
TABLE Bl. TRANSIENT RATE TEST; 9c = 10, S/X = 142, T = 21°C
Batch Test, N ° = 10.0 mg/£ NO~-N
T(min) NO~-N(mg/£) NO~-N(mg/£)
10
32
50
70
87
107
Batch Test,
T(min) N<
3
17
21
45
60
75
92
7.90
7.10
4.70
2.85
1.60
0.05
to
N
D~-N (mg/J
9.16
8.65
6.96
4.95
4.31
2.44
0.99
0.27
0.22
0.37
0.17
0.20
0.03
20.0 mg/fc NO~-N
I) NO~-N(mg/X,)
0.50
0.37
0.60
0.29
0.24
0.23
0.17
98
-------
TABLE B2. TRANSIENT RATE TEST; 0 =15, S/X = 150, T = 21 C
c
0 = 319
T (min )
13
39
60
79
97
116
146
170
199
0 = 362
T (min)
19
52
84
111
143
179
212
234
260
294
min, Nt0 =9.9
NO~-N(mg/£)
8.6
7.1
4.7
3.8
4.3
3.6
1.1
0
0
min, Nt0 = 19.
NO~-N(mgA)
18.3
14.9
13.0
11.6
9.3
4.5
1.7
0
0
0
mg/£ NO~-N
NO~-N(mg/«,)
0.03
0.07
0.09
0.10
0.12
0.18
1.51
1.34
0.10
81 mg/£ NO~-N
NO~-N(mg/S,)
0.007
0.007
0.009
0.14
0.31
2.27
5.30
5.45
3.30
0.25
99
-------
TABLE B3. TRANSIENT RATE TEST, S/X =45
e = 351
T (rain )
9
39
69
101
130
160
0 = 357
T(min)
9
39
69
99
129
159
189
219
249
279
309
339
369
to
min, N =9.96
NO~-N(mg/£)
5.90
1.30
0.30
0
0
0
to
min, N =19.9
NO~-N(mg/£)
13.2
7.7
1.9
0
0
0
0
0
0
0
0
0
0
_
mg/£ NO -N
NO~-N(mg/£)
5.0
9.10
8.20
6.08
3.52
0.90
mg/£ NO~-N
J
NO~-N(mg/£)
4.72
11.20
15.45
16.90
15.90
15.10
13.15
11.60
9.88
7.35
5.12
2.70
0.60
100
-------
TABLE B4. TRANSIENT RATE TEST DATA; S/X = 56
9 = 359
T(min)
9
40
70
100
129
164
189
219
249
279
308
343
0 = 337
T(iain)
7
38
68
96
126
158
186
218
247
277
306
min, Nt0 = 10.0
NO~-N(mg/«,)
6.8
1.6
0
0
0
0
0
0
0
0
0
0
min, Nt0 = 20.0
NO~-N(mg/£)
12.5
6.65
1.69
0
0
0
0
0
0
0
0
mg/Jl NO~-N
NO~-N(mg/X,)
2.65
9.50
11.90
11.25
10.45
8.95
7.55
6.20
4.82
3.65
2.25
0.71
mg/H NO~-N
NO~-N(mg/fc)
3.80
11.05
16.65
18.60
18.00
15.75
14.35
12.25
10.95
9.20
7.92
101
-------
TABLE B5. TRANSIENT RATE TEST DATA; S/X = 100
0 = 347
T (min )
14
44
71
103
134
162
e = 350
T(min)
10
40
70
103
132
161
190
221
250
280
min , N =
NO~-N(mg/£)
2.10
0.50
0
0
0
0
min , N =
NO~-N(mgA)
11.6
6.6
1.8
0
0
0
0
0
0
0
10.25 mg/£ NO.-N
NO~-N(mg/£)
6.00
10.55
9.75
6.95
4.05
1.10
20.5 mg/SL NO~-N
NO~-N(mg/£)
5.78
13.82
19.80
18.90
17.12
14.01
11.12
7.40
4.30
1.35
102
-------
TABLE B6. TRANSIENT RATE TEST DATA: S/X =194
0 = 332 min
T(min)
11
42
71
101
131
165
190
, Nt0 = 9.75
NO~-Nj[mg/£)
4.26
1.00
0
0
0
0
0
8 = 309 min, NtO = 18.
T (min)
8
39
69
97
127
159
187
219
248
278
301
NO~N(mg/£)
12.50
1.13
0
0
0
0
0
0
0
0
0
mg/£ NO -N
NO~-N(mg/£)
5.42
9.75
8.55
5.95
3.9
1.80
0.37
5 mg/fc NO~-N
NO-—N (mg/&)
2
5.60
14.05
17.75
15.95
13.28
10.15
7.80
4.90
3.62
2.14
1.12
103
-------
TABLE B7. TRANSIENT RATE TEST DATA; T = 12°C
6 = 382 min, NtO =9.45 mg/£ NO~-N
T (min ) NO~-N (mg/£ ) NO~-N (mg/£ )
12
40
71
100
130
162
190
8 = 384
T (min )
10
40
70
100
130
160
190
220
250
280
6.8
6.3
4.7
3.8
2.5
1.6
0.76
to
min, N = 18.9
NO~-N(mg/£)
13.9
14.3
12.2
10.1
8.1
7.1
5.3
3.8
2.6
1.3
0.04
0.06
0.06
0.04
0.04
0.03
0.04
mg/£ NO~-N
NO~-N(mg/S.)
0.04
0.05
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.03
104
-------
TABLE B8. TRANSIENT RATE TEST DATA: T = 30°C
9 = 331 rain, N ° = 20.7 mg/£ NO~-N
T (min) NO~-N (mg/£) NO~-N (mg/Jl)
11 12.1 7.65
41 2.7 16.5
71 0 13.5
101 0 7.9
131 0 2.15
0 = 331 min, NtO = 10.25 mg/£ NC^-N
T (min) NO~-N (mg/£) NO~-N (rag/X,)
13 2.6 7.5
41 0.4 7.8
72 0 3.0
105
-------
TABLE B9. RESULTS OF ZONE SETTLING TEST; S/X = 45
Total suspended solids Zone settling velocity Solids flux
(mg/E) (ft/hr) (Ib/ft2-day)
18,480 18.9 522
19,640 17.2 507
21,140 16.0 506
23,230 14.1 490
24,680 13.0 481
25,660 11.6 446
28,710 10.7 463
29,940 9.74 437
33,290 9.06 452
35,990 8.58 463
40,020 7.09 425
43,520 6.57 429
106
-------
TABLE 10. RESULTS OF ZONE SETTLING TEST; S/X = 56
Total suspended solids Zone settling velocity Solids flux
311
107
-------
TABLE Bll. RESULTS OF ZONE SETTLING TEST; S/X = 150
Total suspended solids Zone settling velocity Solids flux
(mg/A) (ft/hr) (Ib/ft2-day)
36,630 34.4 1892
46,230 27.4 1897
52,200 21.6 1691
63,290 16.9 1602
72,280 13.4 1450
81,920 11.2 1381
95,050 9.45 1347
102,210 7.74 1189
116,360 5.98 1042
118,920 5.16 919
108
-------
TABLE 12. RESULTS OF FLOCCULENT SETTLING TEST; S/X = 45
Average initial solids
Time
(rain)
0
0
0
0
0
3.5
5.5
7
9
13
14
15
18
20
f 25
30
35.5
40
45
50
55.5
60
62
Depth
(ft)
1.33
2.33
3.33
4.33
5.33
1.33
2.25
3.13
4.0
0.88
4.76
1.63
2.49
3.36
4.22
1.07
1.96
2.83
3.70
0.56
1.43
2.31
3.18
concentration = 14,040 mg/&
Total suspended solids
(mg/X,)
12,190
12,790
13,590
14,920
16,700
140
84
84
56
40
68
42
36
42
26
26
30
22
24
16
18
18
18
109
-------
TABLE 13. RESULTS
OF FLOCCULENT SETTLING TEST; S/X = 56
Average initial
Time
(min)
0
0
0
0
0
3.5
4.5
6.0
7.45
9.5
13
14.5
16
18
20
31
35
40
46
50
55
61
67
solids
Depth
(ft)
1.20
2.20
3.20
4.20
5.20
1.20
2.08
2.94
3.81
4.63
0.54
1.41
2.28
3.15
4.01
0.96
1.73
2.60
3.46
0.29
1.18
2.01
2.86
concentration = 16,210 mg/Jl
Total suspended solids
(mg/£)
14,370
15,220
15,640
16,560
19,260
204
160
126
104
84
46
50
42
46
44
24
28
22
32
22
24
26
18
110
-------
TABLE B14. RESULTS OF FLOCCULENT SETTLING TEST; S/X = 150
Average initial solids
Time
(min)
0
0
0
0
0
4
4.5
5
7
8
9.5
10
10.5
11
33
34
35
36
Depth
(ft)
1.25
2.25
3.25
4.25
5.25
1.04
1.93
2.83
3.72
4.62
1.40
2.30
3.19
4.09
0.98
1.87
2.77
3.66
concentration = 35 , 560 mg/H
Total suspended solids
(mg/A)
35,440
35,190
30,610
30,950
32,240
274
222
302
198
238
126
114
100
104
28
22
30
38
111
-------
TABLE B15. RESULTS OF FLOCCULENT SETTLING TEST; S/X = 150
Average initial solids concentration = 67,440 mg/&
Time
(min)
0
0
0
0
0
6
9
12
15
17
18
21
24
28
32
36
40
44
48
Depth
(ft)
1.00
2.00
3.00
4.00
5.00
1.00
1.80
2.60
0.60
3.45
1.33
2.21
3.08
0.96
1.82
2.64
0.46
1.46
2.31
Total suspended solids
(mg/4)
54,220
59,710
66,420
73,640
83,220
138
118
60
58
52
48
66
40
24
36
26
26
26
32
112
-------
TECHNICAL REPORT DATA
(rlease read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-78-115
3. RECIPIENT'S ACCESSIOWNO.
TITLE AND SUBTITLE
AUTOTROPHIC DENITRIFICATION USING SULFUR ELECTRON
DONORS
5. REPORT DATE
July 1978 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
AUTHOR(S) _^__
Alonzo Wm. Lawrence, James J. Bisogni, Jr.,
Bill Batchelor and Charles T. Driscoll, Jr.
8. PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME AND ADDRESS
Dept. of Civil §, Environmental Engineering
Hollister Hall
Cornell University
Ithaca, New York 14853
10. PROGRAM ELEMENT NO.
1BC611, SOS #3, Task C/03
11. CONTRACT/GRANT NO.
Grant #803505
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory--Cin.
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
,OH
13. TYPE OF REPORT AND PERIOD COVERED
4/74 - 4/78 Final
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Project Officer: E. F. Earth, Cincinnati, Ohio.
Telephone: (513)684-7641
16. ABSTRACT
This research project investigated the feasibility of autotrophic denitrificatior
as a nitrate removal process for municipal wastewater. The overall objective of this
project was to evaluate the microbial kinetics, and to assess the process performance
of autotrophic microbially mediated denitrification using sulfur electron donors.
This study was divided into three experimental phases. Each phase utilized a
different sulfur compound or flow configuration. Included in these phases were:
continuous flow slurry-type with elemental sulfur as the electron source; semi-
continuous flow, complete-mix reactors with thiosulfate or sulfide as the electron
source; and packed bed columnar reactors with elemental sulfur as the electron source
Based on theoretical and experimental considerations, kinetic models and
stoichiometric relationships were developed for the autotrophic denitrification
process.
The results of this study indicate that autotrophic denitrification with various
sulfur species, particularly elemental sulfur, is a feasible scheme for removal of
nitrate from wastewater effluents.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Wastewater*
Nitrogen Cycle*
Sulfate Reducing Bacteria*
Various sulfur species
Methanol replacement*
Suspended growth reactor
Packed column reactor
Stack gas sulfur
13B
18. DISTRIBUTION STATEMENT
Release to Public
^^—.———•
EPA Form 2220-1 (9-73)
19. SECURITY CLASS (ThisReport)
Unclassified
127
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
113
« U.S. GOVERNMENT PW1MG OFflCt 1978— 737-140/1420
------- |