Citation
Biological Denitrification of High Nitrate Industrial Streams

Material Information

Title:
Biological Denitrification of High Nitrate Industrial Streams
Creator:
Peter, Sherin
Place of Publication:
[Gainesville, Fla.]
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (64 p.)

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Chemical Engineering
Committee Chair:
Svoronos, Spyros
Committee Members:
Koopman, Ben L.
Graduation Date:
5/1/2008

Subjects

Subjects / Keywords:
Bioreactors ( jstor )
Carbon ( jstor )
Electrons ( jstor )
Flasks ( jstor )
Flow velocity ( jstor )
Hydroxides ( jstor )
Micronutrients ( jstor )
Nitrates ( jstor )
Potassium ( jstor )
Sodium ( jstor )
Chemical Engineering -- Dissertations, Academic -- UF
biological, denitrification, treatment, water
Genre:
Electronic Thesis or Dissertation
bibliography ( marcgt )
theses ( marcgt )
Chemical Engineering thesis, M.S.

Notes

Abstract:
The purpose of this project was to develop a biological denitrification system for high nitrate industrial streams. High nitrate samples obtained from an industrial source and synthetic nitrate solution prepared using nitric acid and DI water were used in the experiments. The goal was to obtain a denitrification rate of 2 mg NO3-N/L/min. Attached growth systems and continuous stirred tank reactors were used for the study. The reactors were kept in anoxic conditions in order to maintain denitrification conditions. Effects of micronutrient deficiency, volatile contaminants high nitrate concentration, and sodium on denitrification rates were studied. From the study, it was concluded that the industrial stream was treatable; however, volatile contaminants present in the feed solution must be aerated out and metals must be removed by ion exchange prior to denitrification. It was also concluded that micronutrients should be added directly to the reactor in order to avoid its precipitation in the feed and deficiency in the reactor. Studies comparing sodium and potassium neutralized nitrate solutions did not show any significant differences in denitrification rates. Thus, it was concluded that sodium did not affect the denitrification process considerably. In the suspended growth system, denitrification rate of 4 mg NO3-N/L/min was achieved with synthetic nitrate solution as the feed, and a denitrification rate of 2 mg NO3-N/L/min was achieved using industrial nitrate solution. An assay was also developed in order to test the biotreatability of high nitrate solutions. In the assay procedure, a batch test is conducted using the test sample along with a synthetic nitrate solution containing the same NO3-N concentration. The absorbance values and denitrification rates of the test sample and synthetic nitrate solution are obtained after 4, 8, and 24 hours. YNO and ?maxNO of the sample and synthetic nitrate solution are obtained from the absorbance and NO3-N measurements. The biotreatability index, BI, of the test sample is obtained by taking the ratio of ?maxNO of the sample to ?maxNO of a synthetic nitrate solution containing the same nitrate concentration. A BI value of 1 means that the test sample can be successfully denitrified, and values considerably less than 1 show that the denitrification will be very difficult. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2008.
Local:
Adviser: Svoronos, Spyros.
Statement of Responsibility:
by Sherin Peter

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright by Sherin Peter. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
7/11/2008
Classification:
LD1780 2008 ( lcc )

Downloads

This item has the following downloads:


Full Text






Results and Discussion


CSTR1 Results

CSTR1 was started in order to denitrify sample A neutralized with NaOH in the tote.

The reactor was started in batch mode, and was switched to continuous mode at a flow rate of 1

mL/min once the pH and absorbance increased. Micronutrients were added in the feed of the

reactor in the amount of 4 drops/L. The results of the reactor are shown below:


nitrate reduction and feed rate vs. time (Target feed rate: 2 mg NOs-N/Lnmin)


100.00 **
90.00 *
80.00 *
70.00 *
S60.00
S50.00 *-
S40.00
30.00 *
20.00
10.00
0.00 *
8/5 8/15 8/25 9/4 9/14 9/24
time

+ % nitrate reduction feed flow rate


Figure 3-2. CSTR1 nitrate reduction and feed flow rate vs. time


1


0.8


0.6
z

0.4


0.2


0
"o
o









preliminary treatment, and it is removed through physical operation that follows. Thus, the

effluent from the physical operation is relatively clean, and generally no additional treatment is

required (Tchobanoglous, 2003).

Biochemical operations in wastewater treatment may be classified according to the

biochemical transformation, the biochemical environment, and the bioreactor configuration

(Grady, 1999). Biochemical transformations include removal of soluble organic matter,

stabilization of insoluble organic matter, and conversion of soluble inorganic matter. Removal

of soluble organic matter occurs when the microorganisms use the organic matter as carbon and

energy source. Part of the carbon is converted to carbon dioxide during enzymatic reactions, and

the rest is used to produce more biomass. This process can be carried out in aerobic conditions

and anaerobic conditions (Grady, 1999). Stabilization of insoluble organic matter occurs when

the particulates are entrapped within the biomass, and is converted to stable end products.

Conversion of soluble inorganic matter is utilized in biological nutrient removal processes,

generally to reduce phosphorus and nitrogen concentrations in wastewater. Phosphates in

wastewater are converted ultimately to orthophosphates through microbial activity, and are then

taken up by specialized bacteria that store large quantities of phosphates in granules within the

cell (Grady, 1999). Nitrogen can be present as ammonium or nitrate in wastewater. Ammonium

is converted to nitrate in aerobic environments, and nitrate is converted to inert nitrogen gas in

anaerobic environments.

Biochemical environment where microbial activity takes place is classified mainly

according to the terminal electron acceptor during energy production. Three main types of

electron acceptors are oxygen, inorganic compounds, and organic compounds (Grady, 1999).

When dissolved oxygen is present in high concentration, it becomes the primary electron











Cf of exp(az2/2 + DoT) d (a21/2 + DoT) = Cf (exp(cz2/2 + DoT) 1) (3-19)

C(t) = Co exp(-(ca2/2 + DoT)) + Cf exp(-(oz2/2 + DoT)) (exp(aoz2/2 + DoT) 1) (3-20)

C(t) = Co exp(-(ca2/2 + DoT)) + Cf (1- exp(-(az2/2 + DoT))) (3-21)

Applying CSTR1 Results

The results were analyzed over three time periods. As shown in Figure 3-4, time period 1

was taken from to to ti, time period 2 was taken from ti to t2, and time period 3 was taken from t2

to t3. It was assumed that time period 1 and time period 2 had constant dilution rate, and time

period 3 had dilution rate that changed linearly with time. At t = to, Co was taken to be zero; thus

it was assumed that no toxins were present in the system initially. The results of the analysis are

shown in Tables 3-2, 3-3, and 3-4.


100.00 ,

90.00 13

80.00 II

70.00 -

60.00

50.00

40.00

30.00 *
.. *4
20.00

10.00

0.00
8/5 8/10 8/15 8/20 8/25
time


3


2.5


2


1.5 |


1


0.5


0


Figure 3-4. Time periods used for CSTR1 data analysis









(SNO/(KNo + SNO)) is approximated to be equal to 1. Then Equation 4-2 can be rewritten as

shown in Equation 4-5.

dSNo/dt = lmaxNO [Ao + YNo(SNo(O) SNO(t))] (4-6)

(Ao+YNoSNo(o)) = P (4-7)

Sno(On(0t) (dSNO/ P YNOSNO) = lmaxNO of' dt (4-8)

ln(P YNoSNo(t)) = YNO [lmaxNO t + ln(P-YNoSNo(o)) (4-9)

Thus, a plot of ln(P YNoSNo(t)) vs. YNO*t will have a slope of [lmaxNo. Once |JmaxNo of

the test solution is obtained, JmaxNO of the test solution is compared with ,maxNO of the synthetic

nitrate solution. Thus, the ratio BI, given in Equation 4-10, can be considered as a treatability

index. A low BI will indicate that biological denitrification of that sample will be very difficult.

BI=((PJmaxNO) test sample)/((.maxNO) synthetic sample with same N03-N concentration) (4-10)

Development of the Assay

Initially the batch tests were conducted in flasks secured with rubber stoppers with a gas

inlet port and a gas outlet port. The inlet and outlet ports were secured with aluminum foil after

nitrogen gas was bubbled through the flasks. The first batch test was conducted to test the

treatability of the ion-exchanged industrial stream neutralized with potassium hydroxide. This

sample previously had poor denitrification when it was fed to the suspended growth reactor. The

test was conducted alongside sodium hydroxide neutralized industrial stream, which had positive

results after it was fed through the suspended growth reactor. After 24 hours, it was noticed that

all the flasks had high denitrification and absorbance increase. This led to the suspicion that

oxygen had entered the flasks through gas inlet and gas outlet ports, and that the absorbance

increase was mainly due to aerobic growth of the bacteria. When oxygen is present, bacteria can

convert the nitrate to ammonia, which is then used as the nitrogen source for bacterial growth.









At this point, it was concluded that the low denitrification was attributed to some possible

factors related to reactor design and some possible factors related to the feed. In order to

distinguish the causes, a standard suspended growth reactor was started. The suspended growth

reactor used was Continuous Stirred Tank Reactor (CSTR).









stirred continuously. A rubber stopper was used to close the reactor and it was secured to the

reactor using parafilm. The rubber stopper had a feed inlet port, an effluent port, a gas inlet port,

and a gas outlet port. A nitrogen tank was setup beside the reactor and nitrogen was bubbled

through the reactor twice daily for 10 minutes to strip off any oxygen that is present. Initially,

the reactor was started up in the batch mode. Samples for the nitrate tests were taken through the

inlet gas port. The reactor was operated in batch mode until the pH and absorbance increased,

and then it was switched to continuous mode. For the continuous mode, a computer controlled

pump was set up near the reactor with the same inlet and outlet flow rates. The inlet tubing was

secured into the feed port in the rubber stopper, and the outlet tubing was attached to the effluent

port. Cole Parmer L/S 13 tubing was used for feed and effluent connections. The effluent

from the reactor was collected and measured at a specific time period in order to calibrate the

pump.

Sampling and Data Collection

NO3-N, pH, and absorbance measurements were taken twice daily from the effluent of

the reactor. In continuous mode, samples were taken from the effluent. The reactor was purged

with nitrogen gas after each sampling and the gas inlet port was then secured firmly in order to

keep the anoxic conditions in the bioreactor. The gas outlet port was closed only slightly in

order to vent out gas formed in the reactor. Nitrate-nitrogen concentrations were measured using

HACH NitraVer Test'N Tubes and a HACH spectrophotometer. In general, the effluent

sample was filtered and diluted to 20 times before nitrate-nitrogen measurements. Absorbance

measurements were representative of the bacterial concentration in the reactor, and were taken

using a spectrometer. The results from nitrate-nitrogen, pH and absorbance measurements were

used for adjusting the flow rates in the reactor. COD measurements were taken periodically

using HACH COD test kits. The samples were filtered and diluted to 20 times. After the









4 mg NO3-N/L/min was achieved with synthetic nitrate solution as the feed, and a denitrification

rate of 2 mg NO3-N/L/min was achieved using industrial nitrate solution.

An assay was also developed in order to test the biotreatability of high nitrate solutions. In

the assay procedure, a batch test is conducted using the test sample along with a synthetic nitrate

solution containing the same N03-N concentration. The absorbance values and denitrification

rates of the test sample and synthetic nitrate solution are obtained after 4, 8, and 24 hours. YNO

and lImaxNo of the sample and synthetic nitrate solution are obtained from the absorbance and

NO3-N measurements. The biotreatability index, BI, of the test sample is obtained by taking the

ratio of [,maxNO of the sample to [,maxNO of a synthetic nitrate solution containing the same nitrate

concentration. A BI value of 1 means that the test sample can be successfully denitrified, and

values considerably less than 1 show that the denitrification will be very difficult.









LIST OF TABLES


Table page

1-1 Oxidation-reduction reactions for denitrification ....................................................... 17

3-1 Contents of the micronutrient solution (Vishniac and Santer, 1957...............................33

3-2 Concentration after the first time period for CSTR1 ................... ........................ 38

3-3 Concentration after the second time period for CSTR1.......................... ................... 38

3-4 Concentration after the third time period for CSTR1 ................. ............... ............... 38

4-1 B iotreatability assay results ....................................................................... ..................6 1









samples were added to the test tubes with the reactive material, they were placed in a COD

reactor for two hours. COD measurements were then taken using a spectrophotometer.

Feed Preparation

The high nitrate industrial stream obtained was pre-neutralized to appropriate pH using

either sodium hydroxide or potassium hydroxide. Potassium acetate was used as the carbon

source for the bacteria. 17.5 g/L of potassium acetate was used in the feed, along with 0.02 g/L

potassium phosphate. A micronutrient solution was prepared using instructions for trace element

solution by Vishniac & Santer. The contents in the solution are given in Table 3-1. The

micronutrient solution was added in the feed initially in the amount of 4 drops/L. Later in the

experiments, micronutrients were added directly to the reactor rather than in the feed.

Table 3-1. Contents of the micronutrient solution (Vishniac and Santer, 1957)
Ingredient Amount (per 500 mL DI water)
Ethylenediaminetetraacetic acid (EDTA) 25.0 g
ZnSO4 11.2 g
CaCl2 2.79 g
MnCl 2.58 g
FeSO4 2.59 g
Mo7024(NH3)6 0.6 g
CuSO4.5H20 0.785 g
CoCl2.6H20 0.81 g
Magnesium sulfate 2.525 g
Molybdic acid 0.55 g

Artificial nitrate solution used in the experiments was prepared using 69.6 weight percent

nitric acid assay and DI water. The solution was prepared by adding 8.6 mL of HNO3 assay per

one liter of DI water to obtain the same N03-N concentration as sample A. The solution was

neutralized by adding approximately 4 g NaOH pellets per liter of solution. The feed solution

also contained 17.5 g/L potassium acetate as the carbon source and 0.024 g/L potassium

phosphate. The nitrate-nitrogen concentration of the feed was approximately 1450 mg/L, which

was comparable to the nitrate-nitrogen concentration of sample A.













100.00
80.00
60.00
40.00
20.00
0.00


nitrate reduction and feed flow rate vs. time



E

0.5 z
r\ 0
10/ 101 10/ 11 111 111 11/ 111 11/ 11/ 11/ 11/ 12/ 12/ 12/ 12/ 12/ 12/ 12/ 12/
25 28 31 3 6 9 2 15 18 21 24 27 30 3 6 9 12 15 18 21 24

Sample D Sample E Sample A time
synthetic nitrate solution % nitrate reduction feed flow rate
synthetic nitrate solution


Figure 3-16. CSTR4 denitrification and feed flow rate vs. time. Low denitrification rates were
observed for both ion-exchanged samples. High denitrification was observed once the feed was
switched to sample A.


pH and nitrate vs. time


11/3 11/11 11/19 11/27
pH NO3-N(10mg/L)


1300
1200
1100 ,
880 g)_
;8 Ce-

600


188
0


12/13 12/21


Figure 3-17. CSTR4 pH and nitrate concentrations vs. time


Absorbance


0/26


11/11


11/19


11/27


12/13


12/21


* absorbance


Figure 3-18. CSTR4 absorbance vs. time


W1 C
I.5 .2
t ;
UC3


~~ur~~~ e d-,I-~-I, ter-,~~


4
3.5
3
2.5
2
1.5
1
0.5
0
1I


, *, % .<* ** ***@ **C'** **^*, ***4*** ******* *
"-. .

*% *


4111 WL









CHAPTER 2
DENTRIFICATION OF HIGH NITRATE INDUSTRIAL STREAMS USING GROWTH
BIOREACTORS

Initial attempts to denitrify the high nitrate industrial stream were performed in attached

growth bioreactors, Bios2 and Bios3. The submerged attached growth bioreactors allow for

short hydraulic residence times with high solids retention times, and low solids waste after

denitrification. A packed bed bioreactor with upward flow and a recycle stream was used to

conduct the experiments. Rock media obtained from Adventus was used to pack the reactor,

which provided high surface area for the attached growth of bacterial cells. The reactor was kept

anoxic in order to allow for proper denitrification.

Materials and Methods

Setting up Bios2

A 3.78 L HDPE bottle was used to construct Bios2. The set up of the bioreactor is shown

in Figure 2-1.
Effluent tubing Recycle pump








Effluent
tank










Feed pump Bioreactor Inlet orts
Figure 2-1. Attached growth bioreactor









discharging it to the environment. In several operations, flexibility is implemented by placing

several CSTRs in a series. The conditions in various stages of the system can vary, thus

allowing for completion of various transformations. Recycle may also be employed at desired

stage of the system or through the entire chain (Grady, 1999).

Objective of the Project

The objective of this study was to develop a feasible denitrification process for treating

an industrial nitrate stream. Attached growth bioreactors and suspended growth bioreactors were

used for this study. Initial inoculum for the reactors were obtained from the denitrification

basins at University of Florida Water Reclamation Facility. The optimal pH for denitrification

was determined to be between 8 and 9.5 in previous studies. Since industrial nitrate stream does

not contain enough COD to support denitrification, potassium acetate was used as an exogenous

carbon source. The goal of the project was to obtain a denitrification rate of 2 mg N03-N/L/day.









LIST OF FIGURES

Figure page

1-1 D enitrification reaction s............................................................................ ................... 16

2-1 Attached growth bioreactor .......................................... .................... ............... 22

2-2A Bios2 denitrification and feed flow rate vs. time................................... ...... ...............26

2-2B Bios2 pH and nitrate concentrations vs. tim e ........................................ .....................27

2-3A Bios3 denitrification and feed flow rate vs. time.................................... ............... 28

2-3B Bios3 pH and nitrate concentration vs. time................................ ...............29

3-1 Continuous stirred tank reactor......... .................................... ................. ............... 31

3-2 CSTR1 nitrate reduction and feed flow rate vs. time............................................. 34

3-3 CSTR1 pH and nitrate concentrations vs. time....................................... ............... 35

3-4 Time periods used for CSTR1 data analysis................................ ........................ 37

3-5 Dilution rate of CSTR1 with respect to time.............. ...........................................39

3-6 CSTR2 denitrification and feed flow rate vs. time, part I............................................40

3-7 Ratio of C(t)/Cf vs. time for component accumulation in CSTR2...............................41

3-8 Plot of dilution rate vs. time for CSTR2................................... .....................42

3-9 Ratio of C(t)/C(o) vs. time for component deficiency in CSTR2.............................. 42

3-10 CSTR2 denitrification and feed flow rate vs. time, part II. .........................................43

3-11 CSTR2 pH and nitrate concentrations vs. time.....................................44

3-12 C STR 2 absorbance vs. tim e................................ ......................................... ................44

3-13 CSTR3 denitrification and feed flow rate vs. time .................................... .................46

3-14 CSTR3 pH and nitrate concentration vs. time....................................... ............... 47

3-15 C ST R 3 absorbance vs. tim e.......................... ........................................... ....................47

3-16 CSTR4 denitrification and feed flow rate vs. time. ........................... ................................ 49

3-17 CSTR4 pH and nitrate concentrations vs. time.....................................49











pH and nitrate concentration vs. time


10
9-
8 1000

6--

4 500




1/5 115 8125 9/4 9/14 9/24
-U-pH N03-N (100 mg/L) time


Figure 3-3. CSTR1 pH and nitrate concentrations vs. time

As shown in Figure 3-2, the denitrification rate of CSTR1 was initially high. However, it

started to decrease after 8 days and continued to decrease without possibility of recovery. The

pH began decreasing and nitrate concentration started decreasing with time, as shown in Figure

3-3. The reactor failed to recover even after switching to batch mode.

After CSTR1 failed, possible causes that might have contributed to the low denitrification

of the industrial stream were considered. It was hypothesized that a toxic component in the

sample might be accumulating in the reactor and was not consumed during the metabolism of the

bacteria. Possible candidates were sodium buildup, high nitrate buildup, or accumulation of

some other toxic component in the industrial solution. An analysis was done to predict the

concentration of a toxic compound that might be accumulating in the reactor after a certain time

period.

Analysis of CSTR1 Results

After the hypothesis that the accumulation of a constituent is causing low denitrification,

a modeling analysis was done on CSTR1 results in order to predict the accumulation period of

the constituent. A mass balance was done on the reactor as shown in Equation 3-1.









TABLE OF CONTENTS

page

A CK N O W LED G M EN T S ................................................................. ........... ............. .....

L IS T O F T A B L E S ................................................................................. 7

LIST OF FIGURES .................................. .. ..... ..... ................. .8

A B S T R A C T ......... ....................... ............................................................ 10

CHAPTER

1 INTRODUCTION ............... ................. ........... .............................. 12

Biochemical Operations in Wastewater Treatment.....................................................12
B biological D enitrification ......................................................................... ........................ 14
K inetics of D enitrification .................................. ....................................................... 17
Bioreactor Configuration for Denitrification .................... ............ .............. 20
Objective of the Proj ect ............................ .......... .................. ... 21

2 DENTRIFICATION OF HIGH NITRATE INDUSTRIAL STREAMS USING
G R O W TH B IO R E A C TO R S ........................................ .......... .................. ........................22

M materials and M methods ...................................... .. .......... ....... ...... 22
Setting up Bios2 ....................................... ...............22
Feed Preparation and Sampling for Bios2......................... ...............23
S ettin g u p B io s3 ..................... ...........................................................2 4
Feed Preparation and Sampling for Bios3 ............................................ ............... 25
Results and Discussion ..................................... ................. ........ ..... 25
B io s2 R e su lts ..................................................................................................2 5
B io s3 R e su lts ........................................................................................................2 7

3 DENITRIFICATION OF HIGH NITRATE INDUSTRIAL STREAMS USING
SUSPENDED GROWTH BIOREACTORS ..... .................... ...............31

M materials and M ethods ................... ...... .................. .. ..................................... ..... ............. 1
Setting up Suspended Growth Bioreactor ...... ................. ...............31
Sam pling and D ata C collection .............................................................. .....................32
Feed Preparation ................. .. .. ...... ....... ..........33
R e su lts an d D iscu ssio n ..................................................................................................... 3 4
C S T R 1 R esu lts ................................... ...... ............ .. .. ......... ...... ............3 4
A analysis of C STR 1 R results .......... .................................................... .......... ........... 35
Developing Equations ............................... ....... ........ ............... 36
A applying C ST R 1 R results ....................................................................... ................... 37
C ST R 2 R results ........................................................................... 39
C ST R 3 R results ........................................................................... 4 5












nitrate concentration and pH vs. time


-sitched feed to sample C


10

9

8




8 5

4
3

2,#


0 4.. ._
9/13 9/16 9/19 9/22 9/25 9/28 10/1 10/4 10/7 10/1 10/1 10/1 10/1 10/2 12 1 10/2 10/3 11/3
--pH N03-N (100 mg/L) 2 5 8 1



Figure 3-14. CSTR3 pH and nitrate concentration vs. time





absorbance vs. time


1000
.1



8
500





0


/13 9/16 9/19 9/22 9/25 9/28


10/1 10/4 10/7 10/1 10/1 10/1 10/1 10/2 10/2 10/2 10/3 11/3
0 3 6 9 2 5 8 1
time

* absorbance


Figure 3-15. CSTR3 absorbance vs. time


4
3.5
3
2.5
2
1.5
1
0.5
0
9


* *
*. *. .
*

** S.. .***
S. .
.
e
tao~ee eeo e
( eCo e









denitrification. The highest flow rate achieved by the reactor was 5.5 mL/min, with 97.5%

denitrification, as shown in Figure 3-16. This corresponded to a denitrification rate of 1.68 mg

NO3-N/L/day, which was close to the target denitrification rate of 2 mg NO3-N/L/day.

At this point it was concluded that the industrial stream was indeed treatable, but a

component in the stream is acting as an inhibitory factor. Causes for denitrification include

presence of metal ions and presence of volatile contaminants. Thus, a tool was needed to judge

the treatability of the samples. Thus, a batch test was developed in order to test the treatability of

high nitrate streams. It was revealed from the results of the batch tests that volatile contaminants

are indeed present in the industrial stream and they might be the components contributing to the

low denitrification rates and low absorbance values.










inoculum for the reactor. The reactor was initially started with artificial nitrate solution

neutralized with potassium hydroxide pellets as the feed. The feed was later switched to ion-

exchanged industrial stream that was pre-neutralized with potassium hydroxide. The feed

solution contained 17.5 g/L potassium acetate as the carbon source, 0.5 g/L potassium phosphate

as the phosphorus source, and 0.1 g/L magnesium sulfate heptahydrate in order to provide

additional magnesium for bacterial metabolism. The results of the reactor are shown in Figures

3-13, 3-14, and 3-15.

The reactor continued to denitrify at a high rate using sample B (ion-exchanged industrial

stream), as shown in Figure 3-13. The highest flow rate obtained using sample B was 0.8

mL/min with an overall denitrification of 99.97%. This corresponded to a denitrification rate of

0.315 mg N03-N/L-min. The reactor was not able to achieve higher flow rates because of the

shortage of available sample B solution.


CSTR3 nitrate reduction and feed rate vs. time

100.00 0 104, W 1
90.00
C 80.00 0.8
0 70.00 E-
60.00 0.6 "E
50.00 ----------------------_---------- -
40.00 0.42
30.00 Z
S 20.00 0.2 .2
10.00
0.00 0
911: 9123 B 1013 101 3 10123 11/2
A time C

% nitrate reduction -feed flow rate


Figure 3-13. CSTR3 denitrification and feed flow rate vs. time A: HNO3-KOH feed started; B:
Switched feed to sample B, C: Switched feed to sample C









Table 4-1. Biotreatability assay results
Sample A tmaxNO = 0.868
Sample C aerated BI = 0.49
Sample C non-aerated Invalid BI value; bacteria did not grow
Sample D aerated BI = 0.43
Sample D non-aerated Invalid BI value; bacteria did not grow

As shown in Table 4-1, aerated sample C and sample D are treatable biologically. The

non-aerated sample C and non-aerated sample D had negative YNO values, and thus gave invalid

BI values. The absorbance for those samples decreased with nitrate consumption during batch

test.

Conclusions

The objective of the assay is to quantitatively analyze biological treatability of high

nitrate solutions. The biotreatability index, BI is used to determine if the sample can be

denitrified successfully. BI is obtained from the ratio, ([tmaxNO) test sample)/(([lmaxNO) synthetic sample with

same N03-N concentration). YNO values can be obtained from the plots of absorbance vs. SNO observed

during the batch tests, and [tmaxNo is obtained from the plots of ln(3 YNoSNO(O)) vs. YNO*t. If

the value of the ratio is closer to 1, the sample is treatable. A value much lower than 1 shows

that high denitrification will be very difficult to achieve for that sample.









ACKNOWLEDGMENTS

I would like to gratefully acknowledge my advisor, Dr. Spyros Svoronos, for his

guidance and support throughout my graduate studies. I would also like to thank my committee

member, Dr. Ben Koopman, for his valuable advice through the course of the experiments. I

acknowledge Kiranmai Durvasula, for her guidance and advice in experimental work. I am also

grateful to Darrick, for providing industrial samples and materials used in the experiments.

Finally, I thank my parents for their endless support and advice throughout my studies.









d/dt (VC)= FCf FC Vr (3-1)

at constant volume,
V dC/dt= FCf FCo Vr (3-2)

dC/dt = (F/V) Cf (F/V) Co r (3-3)

(F/V) =D (3-4)
(D = dilution rate)

dC/dt= DCf -DC- r (3-5)

Developing Equations

When accumulation is considered, the reaction is neglected; then
dC/dt = D Cf D Co (3-6)

Since dilution rate changes with flow rate and thus with time,
dC/dt + D(t) Co = D(t) Cf (3-7)

Using integrating factor, IF = exp(fD(c)) dc
D(t) = aT + Do (3-8)

exp(J(ac + Do) dz = exp(az2/2 + Do') (3-9)

Multiplying by IF gives
exp(ac2/2 + Do') dC/dt + exp(ac2/2 + Do') Co = exp(ac2/2 + Do') Cf (3-10)
t
I d/dt(exp(ar2/2 + Dor) C) = (exp(ar2/2 + Dor) D(t) Cf (3-11)

C exp(ac2/2 + Do') Co = I D(c) Cf exp(ac2/2 + DoT) d' (3-12)

C exp(ac2/2 + Do') Co = I (ar + Do ) Cf exp(ac2/2 + Do') d' (3-13)

C(t)=Co exp -(at2/2 + Dot)+ exp -(at2/2 + Dot) I (ar + Do ) Cf exp(ac2/2 + Do') d' (3-14)

At constant dilution rate, a = 0, then
C(t) = Co exp -(Dot) + Cf exp -(Dot) I (Do ) exp(Do') d' (3-15)

C(t) = Co exp(-Dot) + Cf (1-exp(-Dot)) (3-16)

When dilution rate is not constant, it is assumed to change linearly with time; then
D = ct + Do. Substituting gives
(ac + Do)d' = d (ac2/2 + Do') (3-17)

Cf oP ((a + Do) exp(acc2/2 + Do') d' = Cf oP exp(acc2/2 + Do') d (ar2c/2 + Do') (3-18)









C S T R 4 R e su lts .......................................................................................................... 4 8
R starting C STR 4 w ith Sam ple A ........................................................ .....................50

4 ASSAY ON THE BIOLOGICAL TREATABILITY OF INDUSTRIAL STREAMS..........52

D ev elop in g E qu action s............................................................. ................................. .. 53
D evelopm ent of the A ssay .................................................................................. ...... 54
Results and Discussion ..................................... ................. ........ ..... 56
C onclu sions.......... ..........................................................6 1

5 CONCLUSIONS AND FUTURE WORK.................................... ................................... 62

L IS T O F R E F E R E N C E S .................................................................................... .....................63

B IO G R A PH IC A L SK E T C H .............................................................................. .....................64







































6











Sample A YNO


0 200 400 600


Sno



Figure 4-7. Absorbance vs. SNO for sample A



Sample A

1



0

-0.5

-1 '

-1.5
Yno*t


800 1000 1200 1400


Figure 4-8. ln(3-YNoSNo(0)) vs. YNO t for sample A




Aerated Sample C YNO


1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000
Sno



Figure 4-9. Absorbance vs. SNO for aerated sample C


S.1 13 1E.03- + 1 3 36E
R- 9 5532E-01


+00


, 0RI-:.R. 00OL4
R: 0= 0.42 "'4


= -1 4015E-C5. .5 0?E -01
S= 1 851 TE-.E










reactor. However, the effect was short-lived and the reactors deteriorated again. This led to the

conjecture that micronutrients added to the feed may have precipitated, as suggested by Ben

Koopman. Thus, most of the micronutrients added to the feed of the CSTR would have been

precipitated out. Taking this into consideration, micronutrients were added directly to the reactor

everyday in proportion to the flow rate of the feed. Generally, 1 mL of micronutrients was added

to the reactor when the feed flow rate is 1 mL/min. Magnesium was also added to the feed

solution in the form of magnesium sulfate heptahydrate in order to ensure its availability.

After the micronutrients were added, an increase in the reactor performance as well as

bacterial growth was observed, as shown in Figure 3-6. These results showed that it was indeed

the deficiency of micronutrients that caused low denitrification in the reactor previously. It was

also determined that the addition of micronutrients to the feed can cause precipitation, and thus,

it must be added directly to the reactor. The amount of potassium phosphate added to the feed

was also increased to 0.5 mg/L in order to avoid phosphorus deficiency.


CSTR2 nitrate reduction and feed rate vs. time Part II

100.00
** < 4 t
80.00 ,
2 60.00
c 40.00 2
20.00
0.00 0
8131 9110 9120 9130 10/10 10120 10130 11/9 11119 11 29 12/9 12/19
time feed rate lowered due to limited
carbon availability
% nitrate reduction feed flowrate


Figure 3-10. CSTR2 denitrification and feed flow rate vs. time, part II. As shown in the figure,
a deficiency in carbon occurred during the operation, and the flow rate to the reactor was set at
the minimum for 2 days. The flow rate was increased rapidly after carbon was available to reach
appropriate steady state.










denitrification rate was also observed after the addition of micronutrients to the feed and to the

reactor. 40 drops of micronutrients were added to the feed and 5 drops were added directly to

the reactor. The denitrification briefly increased after the addition of micronutrients; however,

the results did not stay positive for long. The highest flow rate achieved by Bios2 was 0.8

mL/min, with a denitrification rate of 95.96%. This corresponded to a denitrification rate of

0.354 mg NO3-N/L/day. Thus, bioreactor failed to denitrify the industrial stream at the desirable

rate, which was 2 mg NO3-N/L/min.


Nitrate reduction and feed rate vs. time

100 P 1.0

90
$6o. .
80 0.8
t. iC

70 0
C
t 60 60 0.6




30
E








6/7Y 6/17 6/27 7/7 7/17 7/27 8/6
New Bios2 Feed pH agitated
Me Bleia stated
Begin decrease ime (Days)Feed w/ K-Acetate % Nitrate Reduction Feed Flow


Figure 2-2A. Bios2 denitrification and feed flow rate vs. time. As shown in the figure, a slight
increase in denitrification was observed when the media was agitated. However, it still
continued to perform poorly. An increase in denitrification was also shown immediately after
the addition of micronutrients. Also, as shown in the figure, when the pH of the effluent was
measured to be higher than 9.5, the pH of the feed was decreased by the addition of concentrated
HC1 to the feed tank. Target feed rate for the reactor was 2 mg N03-N/L/min.










dilution rate vs. time


S0.0006
0.0005 5E-08 + 0 0002
0.0004
0.0003
0.0002
o 0.0001
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
time (minutes)




Figure 3-5. Dilution rate of CSTR1 with respect to time. From the linear regression, ao was
calculated to be 5E-08 and Do was obtained to be 0.0002.

The analysis showed that after 57 hours, 57.5% of the initial toxic component would have

accumulated in the reactor. The model predicts that at the time of failure (8/15/2007),

approximately 94.6% of the feed concentration of the toxin would have accumulated in the

reactor, which can cause significant impact on the denitrification rates. Therefore, a new reactor

was started using synthetic nitrate solution to investigate whether the accumulating toxin is a

component present in the industrial feed or the sodium ions added to the feed for neutralization.

CSTR2 Results

CSTR2 was started using synthetic nitrate solution prepared using nitric acid and DI

water. Cole Parmer L/S 13 tubing was used to pump the feed into the reactor, and Cole

Parmer L/S 14 tubing was used to pump out the effluent. Once the pH and the absorbance of

the reactor increased, the reactor was switched to continuous mode. Two computer controlled

pumps were placed near the reactor; one pump was used for feed and the second pump was used

for the effluent. Two pumps were required because of the difference in feed and effluent tubing

sizes. Micronutrients were added to the feed in the amount of 4 drops/L. The results of the

reactor are shown in Figure 3-6.









The overall molar-based equation for bacterial growth can be obtained by using the

Equations 1-2a and 1-2b (Grady, 1999).

R = Rd fe*Ra fs*R (1-2a)

fs+ fe = 1 (1-2b)

Rd is the equation for electron donor, Ra is the equation for electron acceptor, Re is the

equation for bacterial cell synthesis, fs is the fraction of transferred electrons used for cell

synthesis, and fe is the fraction of electrons used for energy. The negative terms are inverted

when used in the equation. The fraction of electrons used for synthesis can be obtained from the

growth yield. The carbon for synthesis is shifted to biomass and the carbon used for energy

converted to carbon dioxide during the oxidation-reduction reactions. When industrial carbon

source is used, the formula for the electron donor can be estimated by analyzing the elemental

compositions. Also, when nitrate is the predominant nitrogen form, the nitrogen must be

reduced from +V state to -III state before it can be used as the nitrogen source for cell synthesis

(Grady, 1999).

The equation for microbial growth is often written also with COD basis in terms of true

growth yield. The general formula is given in Equation 1-3,

Ss + (-(1-YH)) SN -----------> YHXBH (1-3)

where Ss is amount of substrate COD, SN is the amount of terminal electron acceptor in COD

units, YH is the growth yield of active heterotrophic bacteria, and XBH is the active heterotrophic

biomass in COD units (Grady, 1999).

Bacteria grow exponentially in favorable conditions. When there is a limiting nutrient

present, the specific growth rate coefficient of the bacteria, [t, will depend on the concentration

of the nutrient. The limiting nutrient can be the carbon source, nutrients, or other factor needed









Start-Up 2/5/2007; Feeding Begins 2/12/2007


10

6 o
S ,. U

0 ~ *. M ronutrlents
0 added pd
7/9 7/14 7/19 7/24 7/29 8/3 8/8 8/13 8/18 8/23
N03-N (100 mg/L) pH


Figure 2-3B. Bios3 pH and nitrate concentration vs. time

After Bios2 and Bios3 failed to denitrify industrial nitrate streams properly, several

factors were considered that might have resulted in low performance of the reactors. One

possibility was the presence of metals in high concentrations in the feed, which can affect

bacterial metabolism. A metals analysis of Bios2 feed, Bios3 feed, and Bios3 effluent was

conducted using Inductively Coupled Plasma (ICP). However, very low concentrations of

metals were found in all samples. Another factor was the high nitrate concentration of the

industrial stream. The effects of high nitrate concentration of the feed on denitrification were not

studied. Also, clogging of the tubing could have occurred due to various inlet ports and low flow

rate of the feed and recycle. Clogging in the pores of the rock media could have occurred also,

due to high concentration of biomass, which can lower the surface area for the attached growth

of bacterial growth. Channeling and short-circuiting inside the reactor is also possible, and

proper contact of the feed stream with the bacteria can be diminished. Another factor considered

was the presence of toxic materials in the feed affecting bacterial growth. It was also possible

that toxic substances present in the feed accumulated on the rock media, which can affect the

metabolism of the attached bacteria and cause low denitrification.









Kinetics of Denitrification

Bacterial metabolism can be described as a series of oxidation-reduction reactions. The

substrate is consumed during growth, and additional cells are produced by binary fission, asexual

mode, or by budding (Tchobanoglous, 2003). In binary fission, the cell becomes two organisms.

The time required for binary fission is called generation time, which can range from less than 20

minutes to several days. In water treatment systems, the generation of new cells is limited by

availability of substrate and nutrients. The generalized equation for microbial growth is given in

Equation 1-1 (Tchobanoglous, 2003):

Carbon source + energy source + electron acceptor + nutrients ------- biomass +
CO2 + reduced acceptor + end products (1-1)

For a balanced growth equation, carbon utilization and biomass growth are coupled.

Therefore, the growth yield, Y, is expressed as units of biomass produced per unit of substrate

removed. When the reactions are coupled, the rate of substrate utilization for synthesis and the

rate of biomass growth are proportional, with Y as the proportionality constant. The actual

amount of biomass formed always less than Y, since a portion of energy is used for maintenance

purposes. This yield is referred to as the observed yield (Yobs) (Grady, 1999).

Bacterial growth in anoxic conditions, as in denitrification, with nitrate as the nitrogen

source and terminal electron acceptor, and acetate as the carbon source, follows the oxidation

reduction reactions as given in Table 1-1 (Grady, 1999). The general formula used for biomass

is CsH702N.

Table 1-1. Oxidation-reduction reactions for denitrification
Reaction for bacterial cell synthesis with nitrate 1/28 CsH702N + 11/28 H20 = 1/28 NO3s +
as the nitrogen source: 5/28 CO2 + 29/28 H+ + e-
Reaction for electron acceptor (with nitrate as 1/10 N2 + 3/5 H20 = 1/5 N03- + 6/5 H+ + e
the terminal electron acceptor):
Reaction for electron donor (with acetate as the 1/8 CH3COO- + 3/8 H20 = 1/8 CO2 + 1/8
carbon source): HCO3- + H+ + e









Table 3-2. Concentration after the first time period for CSTR1
After the first time period:
C(t) = Co exp(-Dot) + Cf (1-exp(-Dot))
Co= 0
Ci(t) = Cf(1- exp(-Dot))
t = 3420 minutes
Do (dilution rate) = F/V = (1 mL/min)/(4000 mL) = 0.00025 min1
Ci(t) = Cf (1- exp((-0.00025/min)(3420 min))
Ci(t) = 0.575 Cf
Thus, after the first time period, 57.5% of the initial toxin concentration will have accumulated

Table 3-3. Concentration after the second time period for CSTR1
After the second time period:
C2(t) = Ci(t) exp(-Dot) + Cf(1-exp(-Dot))
t = 3365 minutes
Do = F/V = (0.5 mL/min)/(4000 mL) = 0.000125 min-
C2(t) = 0.575 Cfexp((-0.000125/min)(3365 min))+Cf (1- exp(-0.000125/min)(3365 min))
C2(t) = 0.721 Cf
Thus, after the second time period, 72.1% of the initial toxin concentration will have
accumulated in the reactor.

Table 3-4. Concentration after the third time period for CSTR1
After the third time period:
Dilution changes linearly with time. Thus D(t) = ct + Do
t = 4500 minutes
a and Do were determined from the plot of dilution rate vs. time, as shown in Figure 3-5
From the plot, a = 5.4E-8
Do = 0.0002
C3(t) = Co exp(- at2/2 + Dot) + Cf (1- exp(- at2/2 + Dot))
Co = C2(t)= 0.721 Cf
exp(-at2/2 + Dot) = exp{(-5.4E-8*(4500 min)2/2)+((0.000241 min1-)(4500 min))} =
0.196
C3(t) = 0.721 Cf (0.196)+ Cf (0.804)
C3(t) = 0.946 Cf
After the third time period, 94.6% Co will have accumulated in the reactor.










significant changes were observed from Bios2 results. An increase in denitrification was

observed after adding micronutrients to the feed solution (4 drops/L) and directly to the reactor

(5 drops). However, the high denitrification rate was temporary. Gas bubbles were also

observed in the walls of the reactor. Agitation of the reactor failed to increase its performance.

The highest flow rate that was achieved by Bios2 was 1.2 mL/min, with an overall denitrification

of 98.7%. This corresponded to a denitrification rate of 0.612 mg NO3-N/L/day. Therefore,

Bios3 also failed to meet the target denitrification rate of 2 mg NO3-N/L/day. The results

showed using potassium salts did not have any significant impact on denitrification.


nitrate reduction and feed rate vs. time

100 1.00

**
90

*
80 0.80

70 > #
**

S60 0.6
i 50 n I *

S40 0.40

30

20 p 0.20

10

0 i 0.00
7/9 f 7/14 7/19 7/24 7/29 8/ 8/8 813 8/18 8/23
Bios3 Begin Micronutrients Flushed out with water
Time (Days) added % Nitrate Reduction Feed Flow


Figure 2-3A. Bios3 denitrification and feed flow rate vs. time. Bios3 results did not vary
significantly from Bios2 results, even though potassium hydroxide was used to neutralize the
feed rather than sodium hydroxide. Notice that an increase in denitrification was also observed
in Bios3 immediately after the addition of micronutrients.










confirmed the theory that volatile contaminants were causing the low denitrification in the

CSTR.

Results and Discussion

The assay was applied to test the treatability of aerated and non-aerated industrial stream

samples neutralized using sodium hydroxide. Inoculum for the batch test was obtained from

CSTR4 effluent. Measurements were taken after 4, 8, and 24 hours. The analyses from batch

test done non ion-exchanged industrial sample are shown in Figures 4-1 through 4-6.


Aerated industrial sample

2.5
S= -87571E- 04. 20 iE-OO
,| 1.5 ^ ^ ^ ^ ^R- = 447E.01
S 1.5 -
0 1
2. 0.5
< 0
0 500 1000 1500 2000 2500


Sno



Figure 4-1. Absorbance vs. nitrate concentration, aerated sample



non aerated industrial sample


1

0.8

0.6

8 0.4

" 0.2

0
1


500


1550 1600 1650 1700 1750 1800 1850 1900


Sno



Figure 4-2. Absorbance vs. nitrate concentration, non-aerated sample


y = -2.0796E-03x 4.0101E+00 **
R- = 7.8896E-01









as shown in Figure 3-11. The absorbance decreased generally with an increase in feed flow rate,

as expected since the effluent flow rate was also increased. The absorbance stabilized once the

reactor reached steady state for each flow rate. The flow rate of the reactor was continuously

increased afterwards up to a flow rate of 11.5 mL/min, as shown in Figure 3-10, which was close

to twice the target flow rate, and the highest denitrification rate observed in CSTR2 was 4 mg

N03-N/L-min. From these results, it was confirmed that high nitrate concentration did not affect

the metabolism and denitrification rate of the denitrifying bacteria.

It was furthermore speculated in the beginning that high amount of sodium can lower

denitrification rates since it will lead to high osmotic pressure across the cell membrane. The

effect of sodium was also tested previously using attached growth bioreactors-by operating a

reactor containing sodium hydroxide neutralized industrial stream and another reactor containing

potassium hydroxide neutralized industrial stream simultaneously. However, the results from the

attached growth bioreactors were not reliable since the reason for poor denitrification could have

been micronutrient deficiencies. Therefore, a new CSTR was started alongside CSTR2 in order

to test the effects of sodium on denitrification rate. The new reactor was started using synthetic

nitrate stream using DI water and 69.6% nitric acid assay, and the solution was neutralized using

potassium hydroxide pellets instead of sodium hydroxide pellets. There weren't any noticeable

differences between the results obtained from sodium hydroxide neutralized solution and

potassium hydroxide neutralized solution. Therefore, it was concluded that sodium does not

have any significant effects on denitrification.

CSTR3 Results

After it was confirmed that high nitrate concentration and sodium did not have

significant effect on denitrification rate, another suspended growth reactor, CSTR3, was started

again with the industrial stream as the feed solution. Effluent from CSTR2 was used as the










nitrate reduction and feed rate vs. time: Part 1


100.00 T ----- -*---* --*--"* - '-*-----------
** 4
100.00 -
I 80.00 *
60.00
40.00 : 1
20.00
0.00 0
8131 915 9110 9115
time

% nitrate reduction feed flow rate


Figure 3-6. CSTR2 denitrification and feed flow rate vs. time, part I

Initially on continuous mode, the reactor was started at 0.5 mL/min. The reactor was

denitrifying very well at this flow rate. The flow rate of the feed was increased up to 2 mL/min

and the reactor continued to denitrify at high rates, as shown in Figure 3-6. However, the flow

rate was increased without allowing the reactor to reach stable conditions. Increasing the feed

flow required an increase in effluent flow rate as well, and increasing the flow prematurely can

lead to washout conditions in the reactor. Therefore, the feed flow rate was decreased to 1.2

mL/min in order for the reactor to stabilize. The initial positive results seen in the reactor were

brief, however, and the nitrate concentration began to increase again in the reactor. The reactor

was switched to batch mode after very low denitrification rates were observed.

It was noticed that denitrification rate was high in the reactor for approximately 265

hours (11 days). The low denitrification rates and low absorbance values were observed later on.

An analysis was done on the accumulation of toxins in the reactor using Equation 3-21. The

initial concentration of the toxin, Co, is assumed to be zero. The ratio of C(t)/Cf was plotted

against time. The results are shown in Figure 3-7









was added to the feed solution. In order to increase the denitrification rate of the bioreactor,

approximately 0.01 g/L of yeast extract and 0.3 ml/L molybdenum (as 1.1 g/L molybdic acid)

were also added to the feed stream, providing nutrients to the bacterial culture.

The initial feed flow for Bios2 was turned on to 0.57 mL/min and the recycle was turned

on to 6.67 mL/min. The recycle flow to feed flow ratio was 11.57. The inlet pH of the feed

solution was maintained at 5. Concentrated hydrochloric acid was added to the feed tank in

increments of 10 mL if the feed stream was measured to be high.

During sampling, it was noticed that the sample port was in fact allowing a mixture of the

feed stream and recycle stream to flow through the valve, and thus the nitrate concentrations

measured through the valve were not a true representation of the NO3-N of the effluent stream of

the bioreactor. Thus the sample port was moved to the recycle outlet, and sample was obtained

by disconnecting the recycle tubing. This allowed for measuring the pH and N03-N of both inlet

stream and outlet stream of the bioreactor. A small volume of the sample was discarded initially

each time in order to flush out the tubing and the sample port valve. The pH and NO3-N tests

were conducted three times a day, and the flow rates were adjusted accordingly. N03-N was

measured using HACH NitraVer test kits and measured using a spectrophotometer. In general,

the sample was diluted to 20 times in order to obtain the nitrate concentration.

Setting up Bios3

The main purpose of Bios3 was to compare the effects of sodium and potassium on

denitrification. The construction and set-up for Bios3 was identical to that of Bios2. The reactor

was filled with rock media from Adventus and 1.1 L of inoculum from UF WRF was added.

The working volume of the reactor was 3 L. Eight inlet ports were still used for feed flow, along

with a recycle outlet, gas outlet, and a sample port for the inlet stream. Computer Controlled









comparatively stable and reliable, and can have higher potential removal efficiency and easy

process control (Barber, 2000).

In domestic wastewater plants, biological denitrification is coupled with nitrification.

Nitrification converts ammonia to nitrate, and in denitrification, nitrate is converted ultimately to

nitrogen gas. Two modes of nitrate removal can occur in effluent streams: assimilating and

dissimilating nitrate reduction (Tchobanoglous, 2003). In assimilating nitrate reduction, nitrate

is converted to ammonia for cell synthesis. This usually occurs when low NH4+-N is available,

and the process does not depend on dissolved oxygen concentrations. In dissimilating nitrate

reduction, the nitrate reduction is associated with respiratory electron chain, where nitrate or

nitrite ions are used as terminal electron acceptor for the oxidation of electron donors

(Tchobanoglous, 2003). The bacteria that are used for denitrification are facultative aerobic;

they are able to switch from aerobic mode to anoxic mode (Grady, 1999). In aerobic conditions,

bacteria use oxygen as the terminal electron acceptor. In the presence of nitrate in anoxic

environments, nitrate is used as the terminal electron acceptor. In the absence of dissolved

oxygen, nitrate reductase enzyme in the electron transport chain is induced, which leads to the

transfer of electrons to nitrate. Presence of oxygen can suppress the activity of the enzyme

(Tchobanoglous, 2003). Therefore, low oxygen concentrations and high nitrate concentrations

are preferable for high denitrification rates.

Heterotrophic bacteria such as paracoccus denitrificans, thiobacillus denitrificans, and

other pseudomonas are usually used for denitrification processes (Tchobanoglous, 2003). The

mechanism for biological denitrification is given in Figure 1-1 (Wasser, 2002). The nitrate

content in the waste stream is converted to an inert form. The enzymes involved in the









CHAPTER 4
ASSAY ON THE BIOLOGICAL TREATABILITY OF INDUSTRIAL STREAMS

A batch test procedure was developed to test the biological treatability of high nitrate water

streams. The assay was inspired by studies on Daphnia by Gasith et. al. Since the parameter to

be tested was denitrification, Daphnia was replaced with denitrifying bacterial culture. The

denitrification results of the test solution are compared with those of the synthetic nitrate

solution. The experiments were mainly conducted on ion-exchanged high nitrate industrial

stream neutralized with potassium hydroxide, and non-ion exchanged industrial stream

neutralized with sodium hydroxide. The batch tests primarily investigated the effects of volatile

contaminants and settled particles on denitrification and bacterial growth. The general procedure

developed to test the treatability of the high nitrate streams is as follows:

1. Grow denitrifying bacteria in a high nitrate stream so that they can be used for the batch
tests.

2. Measure the nitrate-nitrogen concentration and pH of the test solution using HACH
NitraVer Test 'N Tube test kits

3. Prepare synthetic nitrate solution using nitric acid with the same nitrate-nitrogen
concentration as the test solution.

4. Add appropriate chemicals (given as follows) to variable and control solutions in order
to support bacterial growth and metabolism:
a. 17.5 g/L potassium acetate as carbon source
b. 0.5 g/L potassium phosphate as phosphorus source
c. 0.1 g/L magnesium sulfate heptahydrate as magnesium source
d. 0.3 g/L ammonium chloride as the ammonium source

5. Neutralize the test solution and synthetic nitrate solution to pH 8 using sodium hydroxide
pellets

6. Filter the solution using 2 |tm filters.

7. Aerate approximately 200 mL of test solution, and keep the rest non-aerated.

8. Obtain 9 125-mL Erlenmeyer flasks; fill three flasks with 60 mL of aerated test solution,
three flasks with 60 mL of non-aerated test solution, and three flasks with 60 mL of
synthetic nitrate solution.










9. Add 10 mL of bacterial culture and 0.2 mL micronutrient solution-prepared using
instructions for trace element solution by Vishniac & Santer to each flask.

10. Measure the nitrate-nitrogen concentration, pH, and absorbance of each flask. The
nitrate-nitrogen concentration is measured using HACH test kits after filtering the
sample from each flask using 0.45 |tm filters, and diluting the samples to appropriate
concentrations.

11. Secure all flasks with rubber stoppers and place them in an incubator at 37C

12. Measure NO3-N, pH, and absorbance of one aerated flask and one non-aerated flask
three times: first measurement after 3-4 hours, second measurement after 7-9 hours, and
third measurement after 22-24 hours.

13. Obtain ([ max/Y'No)sample / (([ max/'No)synthetic solution from the data obtained in order
to establish the treatability of the high nitrate stream; treatable nitrate solutions will have
a value close to 1.

Developing Equations

Bacterial growth in the test flasks can be quantified since the biomass concentration is

proportional to absorbance values. Biomass growth kinetics is developed assuming the presence

of excess carbon and neglecting decay. The change in absorbance can be characterized by

Monod equation.

dA/dt = maxNo (SNo/(KNo + SNO)) A, nitrate is the limiting component (4-1)

dSNo/dt = maxNO (SNo/(KNo + SNO)) A (4-2)

dSNo/dt = (-l/YNo) (dA/dt) (4-3)

A(t) = Ao + Y'NO (SNO(O) SNo(t)) (4-4)

Thus, a plot of absorbance vs. nitrate concentration will have a slope of -YNO and y-

intercept of Ao. This plot can be obtained from the measurements of absorbance and nitrate

concentration after 0, 4, and 8 hours. It was noted that KNO is typically less than 1, usually for

wastewater streams containing nitrate concentration around 100 mg/L. Since the nitrate

concentration is much higher than 100 mg/Lin the experiments done with this assay, the ratio













synthetic nitrate solution


2000


2500


Sno





Figure 4-3. Absorbance vs. nitrate concentration, synthetic nitrate solution


Aerated industrial sample


0.2
0
-0.2
-0.4
-0.6
-0.8
-1
-1.2
-1.2


Yno*t




Figure 4-4. ln(3-YNoSno(o)) vs. YNo*time, aerated sample


Figure 4-5. ln( P-YNoSno(o)) vs. YNo*time, non-aerated sample


S= -9 172E-04. 2' 502E-00
R' = 9 94S83E.01


S=2 028,. 0 8511
R- = 0 0


S=5Ii8 21E-01. 1 0 ',i EOC(
R = I:.'""21E-0 1









by the organisms for growth. The growth rate coefficient increases initially with increasing

substrate concentration, and asymptotically approaches a maximum called the maximum specific

growth rate, _tmax (Grady, 1999). A general equation used to characterize bacterial growth is

the Monod equation (Equation 1-4). This equation is developed strictly on empirical basis

(Grady, 1999).

t = tmax (Ss/(Ks + Ss)) (1-4)

In the Monod equation, Ss is the substrate concentration and Ks is the half-saturation

coefficient. Ks is the substrate concentration where the specific growth rate of the bacteria is

equal to half of the maximum growth rate. It can used to predict how fast pmax is obtained.

The Monod equation can be extended to consider the effects of various substrates when multiple

limiting nutrients are present, as shown in Equations 1-5 and 1-6 (Grady, 1999). Equation 1-5 is

used when both growth limiting nutrients are required for biomass growth and they can influence

the specific growth rate at the same time. Then the specific growth rate is calculated by taking

the product of the Monod terms for the two nutrients. Equation 1-6 is used when both nutrients

are required for biomass growth, but only one nutrient will limit the growth at a given time.

Then the specific growth rate is obtained by taking the lowest value obtained from separate

single-substrate models.

[i = max (Ssi/(Ksi + Ssi)) (Ss2/(Ks2 + Ss2)) (1-5)

[ = min (pmax (Ssi/(Ksi + Ssi))), (pmax(Ss2/(Ks2 + Ss2))) (1-6)

The pH range of the system should also be considered during the denitrification. The pH

of the system is generally increased during denitrification since OHf ions are produced. The

optimal pH range for the bacterial culture can be determined using various batch tests, and the









bacterial growth can be recorded. Generally, mild alkaline conditions are preferred for

denitrification (Tchobanoglous, 2003).

Bioreactor Configuration for Denitrification

Two general anoxic models are used for denitrification processes: suspended growth

systems and attached growth systems (Grady, 1999). In attached growth systems, microbes are

grown attached to a packing material, generally rocks, sand, or other synthetic materials. The

wastewater flows upward or downward, coming into contact with the attached biofilm, and the

treated water flows out of the system. In anoxic systems, the packing material can be completely

submerged, with gas space above the biofilm liquid layer (Grady, 1999). Excess biomass that

may exit with the effluent can be separated using a clarifier and can be removed for further

processing.

In suspended growth systems, microbes are kept in a suspended state by appropriate

mixing. Suspended growth systems can be operated in batch mode or continuous mode. In

batch systems feed is added into the reactor with suspended biomass and is allowed to react to

completion. Reaction conditions and growth environment change with time. In continuous

systems feed solution is continuous pumped into the reactor and an effluent is pumped out at the

same flow rate. The concentrations of components in effluent are the same as the concentrations

of components in the reactor. One example of a continuous suspended system is the continuous

stirred tank reactor (CSTR). In a CSTR, the liquid volume is kept constant and sufficient mixing

is added to keep the conditions uniform throughout the reactor. A physical operation, such as

sedimentation, may be implemented in order to separate biomass from the effluent. The

overflow from the sedimentation basin will have very small concentration of biomass. The

underflow will have very high concentration of biomass, most of which is recycled to the

bioreactor. A portion of the concentrated slurry is wasted, and is further treated before









In order to limit the presence of oxygen in the flasks, another batch test was conducted

using flasks that were secured tightly with solid rubber stoppers. Prior to securing the flasks,

nitrogen gas was bubbled through the flasks in order to purge out any oxygen. The ion-

exchanged stream neutralized with potassium hydroxide was tested against artificial nitrate

solution prepared using 69.6% HN03 assay. The results were similar to previous test, and high

denitrification and absorbance increase were observed in all flasks. This did not correspond to

the results obtained from suspended growth reactors. Since oxygen was purged out using

nitrogen gas, the possibility of presence of any oxygen in the flasks was discarded.

It was then suspected that volatile contaminants present in the industrial stream might

have caused low denitrification rates in the suspended growth reactor. It was possible that

volatile contaminants may have been removed when nitrogen gas was bubbled through the

flasks. Thus, another batch test was carried out using ion-exchanged sample. The flasks were

again secured with solid rubber stoppers; however, nitrogen gas was not bubbled through the

reactors. At the end of the batch test, it was noticed that little or no denitrification occurred in

the flasks, and no bacterial growth was observed. This supported the hypothesis that volatile

contaminants were causing the low denitrification rates observed previously in the CSTR, and

the high denitrification in the previous batch tests were the result of removal of volatile

contaminants through aeration.

In order to confirm the presence of volatile contaminants in the industrial stream, a batch

test was carried out using aerated and non-aerated ion-exchanged samples. In order to avoid

assimilatory denitrification, 0.3 g/L ammonium chloride was added to the flasks. The flasks

containing aerated samples showed high denitrification and absorbance increase, while the flasks

containing non-aerated samples had low denitrification and no absorbance increase. This











Synthetic nitrate solution


0.4 -
0.2 = 3.0266x 1.1197
0.2 .
R- = 0.961
C 0
0 -0.2 A 0.1 0.2 0.4 05
o -0.4 -
-0.6
S-0.8
= -1
-1.2
-1.4
Yno*t




Figure 4-6. ln(3-YNoSno(o)) vs. YNo*time, synthetic nitrate solution

Figures 4-1, 4-2, and 4-3 show the plots of absorbance vs. nitrate concentration for

aerated high nitrate industrial stream, non-aerated high nitrate industrial stream, and synthetic

nitrate solution respectively. From the plots, the YNO was obtained to be 8.76E-4 for aerated

stream, 2.08E-3 for non-aerated stream, and 9.19E-4 for synthetic nitrate solution. The plots in

Figures 4-4, 4-5, and 4-6 were used to obtain ,maxNo values for aerated sample, non-aerated

sample, and synthetic nitrate solution respectively. JmaxNo values were obtained to be 2.073 for

aerated sample, 0.569 for non aerated sample, and 3.027 for synthetic nitrate. From [lmaxNO

values, the biotreatability index, BI was calculated. BI for aerated sample was 0.635 and for

non-aerated sample was 0.188. The assay showed that the aerated industrial stream can be

treated biologically. The results were verified by using aerated industrial stream as the feed for a

suspended growth reactor. High denitrification rates and absorbance values were observed. The

assay was done on other industrial samples as well. The results obtained from the industrial

samples are given in Figures 4-7 through 4-12. The BI values are summarized in Table 4-1.













aerated Sample C


- 0

C -0.2

S -0.4
ci
g -0.6

-0.8


Yno*t




Figure 4-10. ln(3-YNoSNo(0)) vs. YNO t for aerated sample C





aerated sample D, YNO


Sno




Figure 4-11. Absorbance vs. SNO for aerated sample D


Sample D, aerated


_0 .O-O 4 C0E- 1 8C00

-04
S= 02 'E-01. 1 C001 E+0
-0 CI, h = 9.rI':lEd 01

.1-

-1.2



Figure 4-12. ln(3-YNoSNo(O)) vs. YNO t for aerated sample D


= 1444. 0 5'01
W =CJ


, = -1 4607E.-', 1i33E.01
R: = 8 748-E.01 *









3-18 CSTR4 absorbance vs. tim e....... .. .......... .. ......... .............................. ............... 49

4-1 Absorbance vs. nitrate concentration, aerated sample .................................................56

4-2 Absorbance vs. nitrate concentration, non-aerated sample.................................... 56

4-3 Absorbance vs. nitrate concentration, synthetic nitrate solution ....................................57

4-4 ln(P-YNoSno(o)) vs. YNo*time, aerated sample ........................................... ............... 57

4-5 ln( P-YNoSno(o)) vs. YNo*time, non-aerated sample.................... ........................57

4-6 ln(P-YNoSno(o)) vs. YNo*time, synthetic nitrate solution............... ......................... 58

4-7 A bsorbance vs. SNO for sam ple A ........................................................... ............... 59

4-8 ln(P-YNoSNo(0)) vs. YNO t for sample A .................. .......... .................... 59

4-9 Absorbance vs. SNO for aerated sample C....................................... ....................... 59

4-10 ln(P-YNoSNo(0)) vs. YNO t for aerated sample C .......................... ............... 60

4-11 Absorbance vs. SNO for aerated sample D .............................................. ............... 60

4-12 In(p-YNoSNo(O)) vs. YNO t for aerated sample D.....................................................60



































2008 Sherin Peter









denitrification process are nitrate reductase, nitrite reductase, nitric oxide reductase, and nitrous

oxide reductase (Grady, 1999).


Nitrate reductase
I. N03------ ------------------ NO2
Nitrite reductase
II. NO2- ------------------------------4 NO
Nitric oxide reductase
III. NO ------------------------------ ----- N20
Nitrous oxide reductase
IV. N20 ------------------------------------N2


Figure 1-1. Denitrification reactions

Heterotrophic bacteria use organic carbon as the energy source for metabolism. The

amount of carbon in the stream is measured using chemical oxygen demand. Chemical oxygen

demand measures electrons available in an organic compound. The COD test is based on the

fact that organic compounds can be fully oxidized into carbon dioxide and water with a strong

oxidizing agent. The COD is expressed as the amount of oxygen required to accept the electrons

from the organic compound after complete oxidation (Grady, 1999). Three sources of carbon

may be used in denitrification: COD in the influent wastewater, COD produced during

endogenous decay, or COD from an exogenous source (Tchobanoglous, 2003). Generally,

methanol or acetate is used when exogenous carbon source is needed. Along with carbon, other

nutrients are also required for bacterial metabolism and growth, including phosphorus, sulfur,

potassium, calcium, iron, sodium, chlorine, and magnesium (Tchobanoglous, 2003). Domestic

waters often contain necessary nutrients; for industrial wastewaters, nutrients are added

externally. The energy source used for growth is called the substrate due to the extensive role of

enzymes in microbial metabolism.









CHAPTER 1
INTRODUCTION

Biochemical Operations in Wastewater Treatment

Biochemical operations are used widely in wastewater treatment systems to pollutants

that can cause harm to aquatic environments after discharge (Tchobanoglous, 2003). Pollutants

in aquatic systems can cause low dissolved oxygen concentrations, eutrophication, and increased

toxicity due to organic chemicals. Pollutants in wastewater may be classified by their physical

characteristics, chemical characteristics, by their susceptibility to alteration by microorganisms,

by their origin, and by their effects (Tchobanoglous, 2003). Many of these classifications may

overlap for various components. The goal of wastewater treatment is to remove these pollutants

in an efficient and economical manner by utilizing various unit operations. The unit operations

can be divided into physical, chemical, and biochemical operations. Physical operations are

operated based on laws of physics, chemical operations are operated by utilizing various

chemical reactions to remove toxins, and biochemical operations use enzymatic catalysis of

living microorganisms in order to treat the wastewater (Tchobanoglous, 2003).

Most wastewater containing biodegradable constituents can be treated biologically with

proper control. The primary objective in using biological treatment in industrial wastewater is to

reduce the concentrations of organic and inorganic compounds, as well as remove any nutrients

present, such as nitrate or phosphorus (Tchobanoglous, 2003). Sometimes pretreatment is

needed prior to biological treatment since industrial streams may contain components that are

toxic to bacteria. In biochemical processes, soluble pollutants in the wastewater are converted to

an inert form, such as CO2 or N2, or into new microbial biomass. Generally after the treatment,

excess biomass is removed from treated effluent by a physical operation, such as settling tanks.

Microorganisms may also entrap any insoluble organic matter present in the waste stream after











CSTR2 accumulation of a toxin


1
0.9
0.8 -
0.7
0.6 -
0.5 -* Series1
0 0.4
0.3
0.2
0.1
0
0 1000 2000 3000 4000 5000 6000 7000
time (minutes)


Figure 3-7. Ratio of C(t)/Cf vs. time for component accumulation in CSTR2

As shown in the Figure 3-7, within after a time period of approximately 6400 minutes, the

ratio C(t)/Cf has reached the value of 0.86. Thus, a deterioration of the bioreactor performance

should have been observed by this time. After 188 hours (9/9/2007), the toxic component

concentration would have reached 99% of the feed concentration. However, as shown in Figure

3-6, the reactor continued to perform well for two more days. Thus, accumulation of a toxin was

not causing the low denitrification rates in the reactor. It was then suspected that deficiency of a

key constituent might be causing the low denitrification. Thus, the results were analyzed in

order to predict the time that deficiency of a component might occur.

Analysis was done for the first 6400 minutes of the operation of the reactor, where a linear

increase in flow rate, and thus dilution rate, was observed. The dilution rate was calculated using

the relationship given in Equation 3-8, where D(t) is equal to at + Do. a and Do were obtained

from the plot of dilution rate vs. time, as shown in Figure 3-8.









The bioreactor contained eight feed inlet ports at the bottom and an outlet port at the top.

This allowed for the feed to enter from all sides of the bioreactor, which allowed for better

contact with the denitrifying bacteria inside. An outlet port was connected back to the feed

stream as the recycle, and another outlet port was used to collect the overflow into a receiving

bottle. A sample port was also constructed along the recycle outlet tubing. The feed stream

was pumped from the feed tank and was allowed to mix with the recycle. This allowed for the

regulation of pH of the stream entering the bioreactor. Feed and recycle streams flowed through

peristaltic pumps, and pump flow dials were used to monitor the flow rates. Feed and recycle

pumps were standardized and calibration curves were used to obtain the flow rates with respect

to pump dials.

The mixed stream flowed through the eight inlet ports, and was allowed contact with the

bacteria as it flows upward and out the bioreactor. The top portion of the bioreactor was used as

the gas separator. Tubing was attached to the lid of the bioreactor to allow the gas to escape.

Cole Parmer L/S 13 tubing was used for all the connections in the bioreactor. The bioreactor

was filled with stone media from Adventus and 1.5 L of inoculum from UF WRF. The working

volume of the bioreactor was 3 L.

Feed Preparation and Sampling for Bios2

Acetate was added as the carbon source for the industrial stream that was fed to the

reactor in the form of sodium acetate. The stream was pre-neutralized in order to increase the

pH, and had a concentration of approximately 1300 mg/L N03-N. The carbon source for Bios2

feed was added in order to obtain 8:1 COD-NO3-N ratio. In order to obtain an 8:1 COD-NO3-N

ratio, 10400 mg COD/L was required in the feed stream. The initial COD was assumed to be

6000 mg/L, and thus 4.4 g/L of COD was needed in the feed solution. Thus 6.445 g/L of acetate










Start-Up 2/5/2007 ; Feeding Begins 2/12/2007


10.00
9.00
8.00
7.00
6.00
5.00
4.00 ---- 4 0 44
3.00
2.00
1.00 *
A.00 Vicronutrient
0.00 s 7'added,
t 6/7 6/17 6/7 77 7/17 127 8/6
I media Agitated System Upset
Feed pH decreases Feed w/ K-Acetate begin



Figure 2-2B. Bios2 pH and nitrate concentrations vs. time

One factor that was taken into account after Bios2 failed to denitrify was the effect of

sodium on denitrification. It was hypothesized that sodium contribute significantly to the

osmotic pressure across the cell membrane, which would affect the metabolism and

denitrification of the bacteria. An alternate to using sodium salt for neutralization and as carbon

source was to use potassium salts. Even though potassium ions contribute to the osmotic

pressure also, their effects would be smaller than those of sodium ions. Therefore, a new

bioreactor was started in order to examine if potassium would lead to high denitrification rates of

the industrial stream samples.

Bios3 Results

Bios3 had initial high denitrification rates; however, on the whole, it had the same results as

Bios2, as shown in Figure 2-3A. It was noticed that Bios3 recovered faster from system upsets

than Bios2, possibly due to the presence of potassium instead of sodium. In general, no









CHAPTER 5
CONCLUSIONS AND FUTURE WORK

The objective of the project was to establish high denitrification rate for an industrial high

nitrate stream using biological treatment. At the end of this study, it was concluded that the

industrial stream was indeed treatable. However, volatile contaminants and metals ions present in

the stream must be removed prior to biological treatment. Volatile contaminants may be

removed through aeration, and metal ions can be removed through an ion exchange process.

During the treatment, micronutrients should be added to the reactor directly rather than in the

feed, since precipitation can occur in the feed. It was also established that neutralization with

sodium did not affect denitrification rates, since the industrial stream sample neutralized with

sodium hydroxide and the stream sample neutralized with potassium hydroxide had same results.

CSTR with sodium hydroxide neutralized synthetic nitrate feed was able to reach a

denitrification rate of 4 mg NO3-N/L/min with a flow rate of 11.5 mL/min and an overall

denitrification of 97%. This denitrification rate was twice the target denitrification rate. CSTR

with the feed of sodium hydroxide neutralized industrial stream was able to reach the target

denitrification rate of 2 mg NO3-N/L/min with a flow rate of 5.5 mL/min and an overall

denitrification of 98%.

The denitrification was performed using potassium acetate as the exogenous carbon

source. Thus, effects of carbon limitation on denitrification were not analyzed. For the next

step, an industrial carbon source may be used for the process. Studies using attached growth

bioreactors can also be conducted with micronutrients added directly to the reactor instead of the

feed. Attached growth bioreactors can be more feasible since sludge concentration in the reactor

effluent will be very low.









CSTR4 did not have positive denitrification results initially. The pH and absorbance

decreased steadily as shown in Figures 3-17 and 3-18, and the denitrification was deteriorating

slowly as shown in Figure 3-16. The results were puzzling since the sample was ion-exchanged,

and thus it invalidated the assumption that metal ions may be causing the problem. The reactor

failed to recover after switching to synthetic feed. In order to be certain that the ion-exchange

was effective, an analysis of the samples were done at the industrial site, and it was observed that

one of the samples had high concentration of metal ions that might inhibit the bacterial growth

and lower the denitrification rate. It wasn't positive which of the samples contained the high

metal ion content. It was possible that the low denitrification of the reactor was due to presence

of metal ions in the feed.

It was then decided to restart CSTR4 and observe the denitrification rates of sample E.

Therefore, the reactor was restarted with synthetic feed and inoculum obtained from CSTR2

effluent. Once the reactor was in steady state with a flow rate of 1 mL/min, the feed was

switched to sample E. The denitrification rate was high using this sample; however, the

absorbance started decreasing. This led to the conclusion that metals were not the inhibition

factor in low denitrification rates observed.

Restarting CSTR4 with Sample A

CSTR4 was restarted with sample A since it was suspected that previous low

denitrification rates were caused by deficiency of micronutrients. When sample A was

previously fed into the reactor, micronutrients were added in the feed solution rather than

directly to the reactor. Thus, the reactor was restarted in continuous mode at lmL/min. Effluent

from CSTR2 was used as the inoculum for the reactor.

High denitrification rates were observed throughout the operation of the reactor, and pH

remained steady. The reactor was able to achieve high flow rates with high overall



































To my parents, Peter Korath and Jessy Peter, and to Marun









CHAPTER 3
DENITRIFICATION OF HIGH NITRATE INDUSTRIAL STREAMS USING SUSPENDED
GROWTH BIOREACTORS

Suspended growth bioreactors consist of reactor vessels containing biomass growing in a

suspended state rather than attached to a media. Continuous stirred tank reactor (CSTR) is a

suspended growth bioreactor where the contents in the reactor are stirred continuously in order to

keep the conditions homogeneous within the reactor and the culture suspended. The stream to be

treated is fed into the reactor, and the treated liquid mixed with biomass is pumped out as the

effluent. The fractions of constituents in the effluent are identical to those inside the reactor.

Many processes have a recycle stream where most of the biomass is recycled into the reactor.

Materials and Methods

Setting up Suspended Growth Bioreactor

The assembly of the continuous stirred tank is given in Figure 3-1.

as inlet port Gas outlet port

Influent tubing Effluent tubing







Feed tank .4
Computer controlled
pump




Figure 3-1. Continuous stirred tank reactor

As shown in Figure 3-1, the continuous stirred tank reactor was constructed using a 4.2 L

flask. The flask was filled 10% with feed solution, and 90% with inoculum from UF WRF

denitrification basin. The reactor was then set in a stir plate with the stirrer inside and was









After sample B was exhausted, the feed was switched to sample C, which was non-ion

exchanged industrial stream neutralized with potassium hydroxide. Changes in performance of

the reactor were not speculated due to this action, since it was hypothesized that it was indeed the

deficiency of micronutrients that caused low denitrification of the previous industrial streams.

However, as shown in Figure 3-13, the denitrification rate started to decrease as soon as the feed

was switched to sample C. Figures 3-14 and 3-15 correspond to the pH and absorbance of the

reactor, respectively. It can be seen clearly that the change in feed led to decrease in pH and

absorbance as well. The reactor was then put on batch mode for three days and the feed was

switched back to synthetic feed neutralized with potassium hydroxide. The reactor failed to

recover from the low denitrification rate, however. This led to the belief that metal ions are

interfering with bacterial growth and denitrification process. It was also noticed that sample C

had a tint of green that was not observed in previous samples.

CSTR4 Results

After CSTR3 failed to denitrify sample C properly, other factors that may cause poor

denitrification were considered. The main concern was the presence of metals in the industrial

stream, which can affect bacterial metabolism and lower denitrification rates. This was further

supported by the results obtained from denitrification of sample C, which was ion-exchanged.

After CSTR3 was discontinued, two new batches of ion-exchanged industrial stream neutralized

with KOH were received (sample D and sample E). CSTR4 was started up using sample D. The

feed also contained 17.5 mg/L potassium acetate, 0.5 mg/L potassium phosphate, and 0.1 mg/L

magnesium sulfate heptahydrate. Effluent from CSTR2 was used as the inoculum for the

reactor. Micronutrients were added to the reactor daily in amounts proportional to the feed flow

rate. The results of the reactor are shown in Figures 3-16, 3-17, and 3-18.









BIOLOGICAL DENITRIFICATION OF HIGH NITRATE INDUSTRIAL STREAMS


By

SHERIN PETER



















A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2008









y = 5.3931E-08x + 1.3776E-04
0.0006 R2 9.6752E-01
0.0005
0.0004
0.0003 -
0.0002
0.0001
0
0 -----------------------------
0 1000 2000 3000 4000 5000 6000 7000

Figure 3-8. Plot of dilution rate vs. time for CSTR2

Another plot of C(t)/Co vs. time was generated in order to quantify the depletion of a

component initially present in the system. Analysis done using the data obtained from the

reactor showed that after approximately 107 hours, the concentration of a component added

initially to the reactor would have reached approximately 13% of the feed concentration. The

results are shown in Figure 3-9.

CSTR2 deficiency of a nutrient

1.5

0.5
0-
0 ------------------------------.
0 1000 2000 3000 4000 5000 6000 7000
time


Figure 3-9. Ratio of C(t)/C(o) vs. time for component deficiency in CSTR2

Since the reactor was performing very well up to this point, it was noted that the ingredient

that is exhausted is needed only in very small amounts for denitrification. Further analysis of the

data showed that at the time of the failure (9/11/2007), the amount of the nutrient will have

reached 0.6% of its initial concentration. One component that was required for bacterial growth

in very small amounts was the micronutrients. It was noted that attached film bioreactors had

started to recover previously when micronutrients were added to the feed and directly to the









pumps were used for Bios3. The pumps were calibrated by measuring the effluent accumulated

in the receiving bottle for a certain period of time.

Feed Preparation and Sampling for Bios3

Bios3 was initially started with sodium hydroxide neutralized industrial stream (sample

A)-the same feed as Bios2. The feed was then switched to potassium hydroxide neutralized

industrial stream (sample C). Potassium acetate was used as the carbon source for the feed. The

nitrate-nitrogen concentration of sample C was 2100 mg/L, which was much higher than Bios2

feed. 15 g/L of potassium acetate was added to the feed in order to ensure excess carbon, along

with 0.024 g/L potassium phosphate as a source for phosphorus, 0.01 g/L yeast extract, and 3 mL

of 1.1 g molybdenum in 1000 mL molybdic acid. The feed pH was maintained at 5, and 10 mL

increments of hydrochloric acid were added to the feed if high pH was observed. Nitrate and pH

measurements of the effluent were taken three times daily. Nitrate-nitrogen measurements were

taken using HACH NitraVer Test 'N Tubes and a spectrophotometer. A small amount of the

effluent was flushed out of the tubing before samples were taken. Two samples were collected

for each measurement, one corresponding to the effluent stream collected from the recycle outlet,

the other sample corresponding to the inlet stream collected from the sample port valve. The

bioreactor was started with a feed flow of 0.5 mL/min and a recycle flow rate of 2 mL/min.

Results and Discussion

Bios2 Results

As shown in Figure 2-2A, Bios2 initially had high denitrification rates; however, the

denitrification rate began to decrease eventually. A large amount of gas bubbles were noticed in

the reactor throughout its operation. The reactor was agitated periodically to release the gas

bubbles. It was noticed that when the reactor was agitated, the denitrification rates improved

temporarily. However, the nitrate concentration in the reactor increased again. High









BIOGRAPHICAL SKETCH

Sherin Peter was born on February 23, 1984, in Muvattupuzha, India. She received her

Bachelor of Science degree in chemical engineering from University of South Florida, Tampa, in

2005. She obtained her MS degree in chemical engineering from University of Florida in 2008.









LIST OF REFERENCES


Barber W. P. and Stuckey D. C. (2000) Nitrogen removal in a modified anaerobic baffled
reactor (ABR): 2, Denitrification. Water Research 34, 2413-2422.

Gasith A. J., Jop K. M., Dickson, K. L., Parkerton, T. F., and Kaczmarek, S.A. (1988). Protocol
for the identification of toxic fractions in industrial wastewater effluents. Aquatic Toxicology
and Hazard Assessment: 10th Volume, ASTM STP 971. Adams, W.J., Chapman, G.A., and
Landis, W.G. Editors., American Society for Testing and Materials, Philadelphia. pp. 204-215.

Kanow, P.E., 1993. Denitrification Methods United States Patent 5211847.
http://www.freepatentsonline.com/5211847.html March 25, 2008.

Lee S., Maken S., Jang J., Park K., and Park J. (2006). Development of physicochemical
nitrogen removal process for high strength industrial wastewater. Water Research 40, 975-980

Leslie Grady C.P., Daigger, G.T., and Lim, H.C. 1999. Biological Wastewater Treatment 2nd ed.
CRC Press, Boca Raton, USA.

Tchobanoglous, G., Burton F.L., and Stensel, D.H. 2003. Wastewater Engineering: Treatment
and Reuse McGraw-Hill, NY, USA.

US EPA., 2005. Section 62-302.530, Criteriafor Surface Water Quality Classifications
(Florida) http://www.epa.gov/waterscience/standards/wqslibrary/fl/fl 4_62-302t.pdf March 30,
2008

Vishniac S., and Santer M. (1957). The Thiobacilli. Bacteriol Rev 21, 195-213

Wasser I.M., deVries S., Moenne-Loccoz P., Schroder, I., and Kavlin K. D. (2002). Nitric oxide
in biological denitrification: Fe/Cu metalloenzyme and metal complex NOx redox chemistry.
Chemical Reviews (Review) 102, 1201-1234









acceptor, and the environment is aerobic. When nitrate or nitrite is present in high

concentrations compared to DO levels, the environment is anoxic, and nitrate or nitrite serves as

the terminal electron acceptor. Sometimes organic compounds, carbon dioxide, and sulfate may

act as the terminal electron acceptor, in which the environment is anaerobic (Grady, 1999).

Wide ranges of microbial species thrive in aerobic environments, while anoxic and anaerobic

processes are carried out by a limited number of organisms.

Biological Denitrification

Denitrification is used widely in wastewater treatment plants in order to treat nitrate and

ammonia, in combination with nitrification. Nitrate concentrations in effluent streams are highly

regulated by EPA. Since nitrate can act as a fertilizer, aquatic weeds, grasses and algae can grow

excessively. This can lead to eutrophication and depleted oxygen levels in receiving waters,

causing fish kills (Kanow, 1993). Therefore, unless the water is properly diluted or sufficient

flow of water is employed to prevent accumulation, high nitrate concentrations can cause

ecological imbalance. Nitrate concentrations in contaminated water have also been linked to

outbreaks of methemoglobinemia in infants, and various health problems in animals (Barber,

2000).

The standard for total nitrate concentration in effluent water is < 10 mg/L, or less than the

concentration that exceeds nutrient criteria (US EPA, 2005). Various methods exist for nitrate

reduction in wastewater, including air stripping, break point chlorination, ion-exchange systems,

fluidized bed systems, expanded bed systems, and biological denitrification (Kanow, 1993).

Most of the chemical methods used for denitrification can be costly, and the chemicals involved

may release toxic compounds to the environment (Lee, 2006). Biological denitrification is











pH and nitrate concentrations vs. time


9
8
1000
7 8
6
S5

4 500
3*
2
1
0 0
9/15 9/23 10/1 10/9 10/17 10/25 11/2 11/10 11/18 11/26 12/4 12/12 12/20

-pH N03-N (100 mg/L)


Figure 3-11. CSTR2 pH and nitrate concentrations vs. time



Absorbance vs. time

4
3.5
3
c 2.5
2
*0
2 1.5 "........ .... ." .. ..... -, "" .** *.,
S1.5
1
0.5 ..
0
9115 9123 10/1 10/9 10/17 10125 11/2 11110 11118 11/26 12/4 12/12 12/20
time

absorbance



Figure 3-12. CSTR2 absorbance vs. time

After the addition of micronutrients, increase in denitrification rate and absorbance was

observed in the reactor. The overall denitrification remained high with increasing flow rates, as

shown in Figure 3-10. The pH of the reactor was in between 9 and 9.5 throughout its operation,









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

BIOLOGICAL DENITRIFICATION OF HIGH NITRATE INDUSTRIAL STREAMS

By

Sherin Peter

May 2008
Chair: Spyros Svoronos
Major: Chemical Engineering

The purpose of this project was to develop a biological denitrification system for high

nitrate industrial streams. High nitrate samples obtained from an industrial source and synthetic

nitrate solution prepared using nitric acid and DI water were used in the experiments. The goal

was to obtain a denitrification rate of 2 mg NO3-N/L/min. Attached growth systems and

continuous stirred tank reactors were used for the study. The reactors were kept in anoxic

conditions in order to maintain denitrification conditions. Effects of micronutrient deficiency,

volatile contaminants high nitrate concentration, and sodium on denitrification rates were

studied.

From the study, it was concluded that the industrial stream was treatable; however, volatile

contaminants present in the feed solution must be aerated out and metals must be removed by ion

exchange prior to denitrification. It was also concluded that micronutrients should be added

directly to the reactor in order to avoid its precipitation in the feed and deficiency in the reactor.

Studies comparing sodium and potassium neutralized nitrate solutions did not show any

significant differences in denitrification rates. Thus, it was concluded that sodium did not affect

the denitrification process considerably. In the suspended growth system, denitrification rate of




Full Text

PAGE 1

1 BIOLOGICAL DENITRIFICATION OF HI GH NITRATE INDUSTRIAL STREAMS By SHERIN PETER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORI DA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Sherin Peter

PAGE 3

3 To my parents, Peter Korath and Jessy Peter, and to Marun

PAGE 4

4 ACKNOWLEDGMENTS I would like to grat efully acknowledge m y advisor, Dr. Spyros Svoronos, for his guidance and support throughout my graduate studies. I would also like to thank my committee member, Dr. Ben Koopman, for his valuable advi ce through the course of the experiments. I acknowledge Kiranmai Durvasula, for her guidance and advice in experimental work. I am also grateful to Darrick, for providing industrial samples and materials used in the experiments. Finally, I thank my parents for their endle ss support and advice throughout my studies.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................8 ABSTRACT...................................................................................................................................10 CHAP TER 1 INTRODUCTION..................................................................................................................12 Biochemical Operations in W astewater Treatment................................................................ 12 Biological Denitrification.......................................................................................................14 Kinetics of Denitrification...............................................................................................17 Bioreactor Configuration for Denitrification ................................................................... 20 Objective of the Project....................................................................................................... ...21 2 DENTRIFICATION OF HIGH NITRA TE INDUST RIAL STREAMS USING GROWTH BIOREACTORS.................................................................................................. 22 Materials and Methods...........................................................................................................22 Setting up Bios2.............................................................................................................. 22 Feed Preparation and Sampling for Bios2....................................................................... 23 Setting up Bios3.............................................................................................................. 24 Feed Preparation and Sampling for Bios3....................................................................... 25 Results and Discussion......................................................................................................... ..25 Bios2 Results...................................................................................................................25 Bios3 Results...................................................................................................................27 3 DENITRIFICATION OF HIGH NITR ATE INDUST RIAL STREAMS USING SUSPENDED GROWTH BIOREACTORS..........................................................................31 Materials and Methods...........................................................................................................31 Setting up Suspended Growth Bioreactor....................................................................... 31 Sampling and Data Collection.........................................................................................32 Feed Preparation.............................................................................................................. 33 Results and Discussion......................................................................................................... ..34 CSTR1 Results................................................................................................................ 34 Analysis of CSTR1 Results.............................................................................................35 Developing Equations..................................................................................................... 36 Applying CSTR1 Results................................................................................................37 CSTR2 Results................................................................................................................ 39 CSTR3 Results................................................................................................................ 45

PAGE 6

6 CSTR4 Results................................................................................................................ 48 Restarting CSTR4 with Sample A...................................................................................50 4 ASSAY ON THE BIOLOGICAL TREATABI LITY OF INDUSTRIAL STREAMS .......... 52 Developing Equations........................................................................................................... ..53 Development of the Assay...................................................................................................... 54 Results and Discussion......................................................................................................... ..56 Conclusions.............................................................................................................................61 5 CONCLUSIONS AND FUTURE WORK ............................................................................. 62 LIST OF REFERENCES...............................................................................................................63 BIOGRAPHICAL SKETCH.........................................................................................................64

PAGE 7

7 LIST OF TABLES Table page 1-1 Oxidation-reduction reactions for denitrification ..............................................................17 3-1 Contents of the micronutrient solution (Vishniac and Santer, 1957 .................................. 33 3-2 Concentration after the first time period for CSTR1......................................................... 38 3-3 Concentration after the s econd tim e period for CSTR1..................................................... 38 3-4 Concentration after the third time period for CSTR1........................................................ 38 4-1 Biotreatability assay results.............................................................................................. .61

PAGE 8

8 LIST OF FIGURES Figure page 1-1 Denitrification reactions.................................................................................................. ...16 2-1 Attached growth bioreactor................................................................................................ 22 2-2A Bios2 denitrification an d feed flow rate vs. tim e.............................................................. 26 2-2B Bios2 pH and nitrate concentrations vs. tim e.................................................................... 27 2-3A Bios3 denitrification and feed flow rate vs. tim e...............................................................28 2-3B Bios3 pH and nitrate concentration vs. time...................................................................... 29 3-1 Continuous stirred tank reactor.......................................................................................... 31 3-2 CSTR1 nitrate reduction a nd feed flow rate vs. tim e......................................................... 34 3-3 CSTR1 pH and nitrate concentrations vs. time.................................................................. 35 3-4 Time periods used for CSTR1 data analysis...................................................................... 37 3-5 Dilution rate of CSTR1 with respect to tim e..................................................................... 39 3-6 CSTR2 denitrificati on and feed flow rate vs. tim e, part I.................................................. 40 3-7 Ratio of C(t)/Cf vs. time for component accumulation in CSTR2..................................... 41 3-8 Plot of dilution rate vs. time for CSTR2........................................................................... 42 3-9 Ratio of C(t)/C(o) vs. time for component deficiency in CSTR2...................................... 42 3-10 CSTR2 denitrificati on and feed flow rate vs. tim e, part II................................................ 43 3-11 CSTR2 pH and nitrate concentrations vs. time.................................................................. 44 3-12 CSTR2 absorbance vs. time............................................................................................... 44 3-13 CSTR3 denitrification and feed flow rate vs. tim e............................................................ 46 3-14 CSTR3 pH and nitrate concentration vs. time................................................................... 47 3-15 CSTR3 absorbance vs. time............................................................................................... 47 3-16 CSTR4 denitrification and feed flow rate vs. tim e............................................................ 49 3-17 CSTR4 pH and nitrate concentrations vs. time.................................................................. 49

PAGE 9

9 3-18 CSTR4 absorbance vs. time............................................................................................... 49 4-1 Absorbance vs. nitrate con centration, aerated sample ....................................................... 56 4-2 Absorbance vs. nitrate con centration, non-aerated sam ple................................................ 56 4-3 Absorbance vs. nitrate concentra tion, synthetic nitrate solution ....................................... 57 4-4 ln( -YNOSno(o)) vs. YNO*time, aerated sample................................................................ 57 4-5 ln( -YNOSno(o)) vs. YNO*time, non-aerated sample........................................................ 57 4-6 ln( -YNOSno(o)) vs. YNO*time, synthetic nitrate solution................................................. 58 4-7 Absorbance vs. SNO for sample A......................................................................................59 4-8 ln( -YNOSNO(0)) vs. YNO t for sample A........................................................................... 59 4-9 Absorbance vs. SNO for aerated sample C.......................................................................... 59 4-10 ln( -YNOSNO(0)) vs. YNO t for aerated sample C............................................................... 60 4-11 Absorbance vs. SNO for aerated sample D......................................................................... 60 4-12 ln( -YNOSNO(0)) vs. YNO t for aerated sample D............................................................... 60

PAGE 10

10 Abstract of Thesis Presente d to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science BIOLOGICAL DENITRIFICATION OF HIGH NITRATE INDUSTRIAL STREAMS By Sherin Peter May 2008 Chair: Spyros Svoronos Major: Chemical Engineering The purpose of this project was to develop a biological denitrificat ion system for high nitrate industrial streams. High nitrate samples obtained from an industrial source and synthetic nitrate solution prepared using nitr ic acid and DI water were used in the experiments. The goal was to obtain a denitrific ation rate of 2 mg NO3-N/L/min. Attached growth systems and continuous stirred tank reactors were used for the study. The reactors were kept in anoxic conditions in order to maintain denitrification conditions. Effects of micronutrient deficiency, volatile contaminants high nitr ate concentration, and sodium on denitrification rates were studied. From the study, it was concluded that the industr ial stream was treatable; however, volatile contaminants present in the feed solution must be aerated out and metals must be removed by ion exchange prior to denitrificati on. It was also concluded that micronutrients should be added directly to the reactor in order to avoid its precipitation in the feed and deficiency in the reactor. Studies comparing sodium and potassium neutralized nitrate solutions did not show any significant differences in denitrification rates. Thus, it was concl uded that sodium did not affect the denitrification process considerably. In the suspended growth system, denitrification rate of

PAGE 11

11 4 mg NO3-N/L/min was achieved with synthetic nitrate solution as the feed, and a denitrification rate of 2 mg NO3-N/L/min was achieved using in dustrial nitrate solution. An assay was also developed in order to test the biotreatability of high nitrate solutions. In the assay procedure, a batch test is conducted usi ng the test sample along with a synthetic nitrate solution containing the same NO3-N concentration. The absorbance values and denitrification rates of the test sample and synthetic nitrate solution are obt ained after 4, 8, and 24 hours. YNO and maxNO of the sample and synthetic nitrate solution are obtained from the absorbance and NO3-N measurements. The biotreatab ility index, BI, of the test sa mple is obtained by taking the ratio of maxNO of the sample to maxNO of a synthetic nitrate soluti on containing the same nitrate concentration. A BI value of 1 means that the test sample can be successfully denitrified, and values considerably less than 1 show that the denitrification will be very difficult.

PAGE 12

12 CHAPTER 1 INTRODUCTION Biochemical Operations in Wastewater Treatment Biochem ical operations are used widely in wastewater treatment systems to pollutants that can cause harm to aquatic environments after discharge (Tchobanoglous, 2003). Pollutants in aquatic systems can cause low dissolved oxyge n concentrations, eutrophication, and increased toxicity due to organic chemical s. Pollutants in wastewater ma y be classified by their physical characteristics, chemical characteristics, by th eir susceptibility to alte ration by microorganisms, by their origin, and by their e ffects (Tchobanoglous, 2003). Many of these classifications may overlap for various components. The goal of wast ewater treatment is to remove these pollutants in an efficient and economical manner by utilizi ng various unit operations. The unit operations can be divided into physical, chemical, and biochemical operations. Physical operations are operated based on laws of physics, chemical operations are operated by utilizing various chemical reactions to remove toxins, and bioc hemical operations use enzymatic catalysis of living microorganisms in order to tr eat the wastewater (Tchobanoglous, 2003). Most wastewater containing biodegradable cons tituents can be treated biologically with proper control. The primary objective in using biological treatment in industrial wastewater is to reduce the concentrations of organic and inorga nic compounds, as well as remove any nutrients present, such as nitrate or phosphorus (Tchoba noglous, 2003). Sometimes pretreatment is needed prior to biological treatment since indu strial streams may contain components that are toxic to bacteria. In biochemical processes, soluble pollutants in the wastewater are converted to an inert form, such as CO2 or N2, or into new microbial biomass. Generally after the treatment, excess biomass is removed from treated effluent by a physical operation, su ch as settling tanks. Microorganisms may also entrap any insoluble or ganic matter present in the waste stream after

PAGE 13

13 preliminary treatment, and it is removed through physical operation that follows. Thus, the effluent from the physical operati on is relatively clean, and gene rally no additional treatment is required (Tchobanoglous, 2003). Biochemical operations in wastewater trea tment may be classified according to the biochemical transformation, the biochemical environment, and the bioreactor configuration (Grady, 1999). Biochemical transformations in clude removal of soluble organic matter, stabilization of insoluble organic matter, and co nversion of soluble inorganic matter. Removal of soluble organic matter occurs when the microorganisms use the organic matter as carbon and energy source. Part of the carbon is converted to carbon dioxide during enzymatic reactions, and the rest is used to produce more biomass. This process can be carried out in aerobic conditions and anaerobic conditions (Grady, 1999). Stabiliza tion of insoluble organic matter occurs when the particulates are entrapped within the bioma ss, and is converted to stable end products. Conversion of soluble inorganic ma tter is utilized in biologic al nutrient removal processes, generally to reduce phosphorus a nd nitrogen concentrations in wastewater. Phosphates in wastewater are converted ultimat ely to orthophosphates through microbial activity, and are then taken up by specialized bacteria th at store large quantities of phos phates in granules within the cell (Grady, 1999). Nitrogen can be present as ammonium or nitrat e in wastewater. Ammonium is converted to nitrate in aerobic environments, and nitrate is conve rted to inert nitrogen gas in anaerobic environments. Biochemical environment where microbial activity takes place is classified mainly according to the terminal electron acceptor dur ing energy production. Three main types of electron acceptors are oxygen, inorganic compounds, and organic compounds (Grady, 1999). When dissolved oxygen is present in high con centration, it becomes the primary electron

PAGE 14

14 acceptor, and the environment is aerobic. Wh en nitrate or nitrite is present in high concentrations compared to DO levels, the environment is anoxic, and nitrate or nitrite serves as the terminal electron acceptor. Sometimes or ganic compounds, carbon dioxide, and sulfate may act as the terminal electron acceptor, in which the environment is anaerobic (Grady, 1999). Wide ranges of microbial specie s thrive in aerobic environmen ts, while anoxic and anaerobic processes are carried out by a lim ited number of organisms. Biological Denitrification Denitr ification is used widely in wastewater treatment plants in order to treat nitrate and ammonia, in combination with nitrification. Nitrat e concentrations in effl uent streams are highly regulated by EPA. Since nitrate can act as a fert ilizer, aquatic weeds, grasses and algae can grow excessively. This can lead to eutrophication a nd depleted oxygen levels in receiving waters, causing fish kills (Kanow, 1993). Therefore, unless the water is properly diluted or sufficient flow of water is employed to prevent accumu lation, high nitrate concentrations can cause ecological imbalance. Nitrate conc entrations in contaminated water have also been linked to outbreaks of methemoglobinemia in infants, and various health problems in animals (Barber, 2000). The standard for total nitrate con centration in effluent water is 10 mg/L, or less than the concentration that exceeds nutrient criteria (US EPA, 2005). Various methods exist for nitrate reduction in wastewater, includi ng air stripping, break point chlo rination, ion-exchange systems, fluidized bed systems, expanded bed systems, and biological denitrif ication (Kanow, 1993). Most of the chemical methods used for denitrif ication can be costly, a nd the chemicals involved may release toxic compounds to the environmen t (Lee, 2006). Biologica l denitrification is

PAGE 15

15 comparatively stable and reliable, and can have higher potential removal efficiency and easy process control (Barber, 2000). In domestic wastewater plants, biological denitrifi cation is coupled with nitrification. Nitrification converts ammonia to ni trate, and in denitrification, n itrate is converted ultimately to nitrogen gas. Two modes of nitrate removal can occur in effluent streams: assimilating and dissimilating nitrate reduction (Tchobanoglous, 2003) In assimilating nitr ate reduction, nitrate is converted to ammonia for cell synthesi s. This usually occurs when low NH4 +-N is available, and the process does not depend on dissolved oxyge n concentrations. In dissimilating nitrate reduction, the nitrate reduction is associated with respiratory electron chain, where nitrate or nitrite ions are used as terminal elec tron acceptor for the oxida tion of electron donors (Tchobanoglous, 2003). The bacteria that are used for denitrification are facultative aerobic; they are able to switch from aerobic mode to anoxic mode (Grady, 1999). In aerobic conditions, bacteria use oxygen as the terminal electron accepto r. In the presence of nitrate in anoxic environments, nitrate is used as the terminal electron acceptor. In the absence of dissolved oxygen, nitrate reductase enzyme in the electron transport chain is induced, which leads to the transfer of electrons to nitrat e. Presence of oxygen can suppre ss the activity of the enzyme (Tchobanoglous, 2003). Therefore, low oxygen concen trations and high nitrate concentrations are preferable for high de nitrification rates. Heterotrophic bacteria such as paracoccus denitrificans, thi obacillus denitrificans, and other pseudomonas are usually used for denitr ification processes (Tc hobanoglous, 2003). The mechanism for biological denitrification is gi ven in Figure 1-1 (Wasser, 2002). The nitrate content in the waste stream is converted to an inert form. The enzymes involved in the

PAGE 16

16 denitrification process are nitrate reductase, nitrite reductase, n itric oxide reductase, and nitrous oxide reductase (Grady, 1999). I. NO3 -------------------------------NO2 II. NO2 ------------------------------NO III. NO ------------------------------------N2O IV. N2O --------------------------------------N2 Figure 1-1. Denitrif ication reactions Heterotrophic bacteria use organic carbon as the energy source for metabolism. The amount of carbon in the stream is measured using chemical oxygen demand. Chemical oxygen demand measures electrons available in an or ganic compound. The COD test is based on the fact that organic compounds can be fully oxidized into carbon dioxide and water with a strong oxidizing agent. The COD is expressed as the amount of oxygen required to accept the electrons from the organic compound after complete oxi dation (Grady, 1999). Three sources of carbon may be used in denitrification: COD in the influent wastewater, COD produced during endogenous decay, or COD from an exogenous source (Tchobanoglous, 2003). Generally, methanol or acetate is used when exogenous carb on source is needed. Along with carbon, other nutrients are also required for bacterial me tabolism and growth, incl uding phosphorus, sulfur, potassium, calcium, iron, sodium, chlorine, a nd magnesium (Tchobanoglous, 2003). Domestic waters often contain necessary nutrients; for industrial wastewaters, nutrients are added externally. The energy source used for growth is called the substrate due to the extensive role of enzymes in microbial metabolism. Nitratereductase Nitritereductase Nitricoxidereductase Nitrousoxideredu ctase

PAGE 17

17 Kinetics of Denitrification Bacterial m etabolism can be described as a series of oxidation-re duction reactions. The substrate is consumed during gr owth, and additional cells are pr oduced by binary fission, asexual mode, or by budding (Tchobanoglous, 2003). In binary fission, the cell beco mes two organisms. The time required for binary fission is called gene ration time, which can range from less than 20 minutes to several days. In water treatment sy stems, the generation of new cells is limited by availability of substrate and nutri ents. The generalized equation fo r microbial growth is given in Equation 1-1 (Tchobanoglous, 2003): Carbon source + energy source + elec tron acceptor + nutrients ------biomass + CO2 + reduced acceptor + end products (1-1) For a balanced growth equation, carbon util ization and biomass growth are coupled. Therefore, the growth yield, Y, is expressed as units of biomass produced per unit of substrate removed. When the reactions are coupled, the rate of substrate utilizatio n for synthesis and the rate of biomass growth are proportional, with Y as the proportionality constant. The actual amount of biomass formed always less than Y, si nce a portion of energy is used for maintenance purposes. This yield is referre d to as the observed yield (Yobs) (Grady, 1999). Bacterial growth in anoxic conditions, as in denitrification, with ni trate as the nitrogen source and terminal electron acceptor, and acetate as the carbon source, follows the oxidation reduction reactions as given in Table 1-1 (Grady, 1999). The general formula used for biomass is C5H7O2N. Table 1-1. Oxidation-reduction reactions for denitrification Reaction for bacterial cell synthesis with nitrate as the nitrogen source: 1/28 C5H7O2N + 11/28 H2O = 1/28 NO3 + 5/28 CO2 + 29/28 H+ + eReaction for electron acce ptor (with nitrate as the terminal electron acceptor): 1/10 N2 + 3/5 H2O = 1/5 NO3 + 6/5 H+ + eReaction for electron donor (with acetate as the carbon source): 1/8 CH3COO+ 3/8 H2O = 1/8 CO2 + 1/8 HCO3 + H+ + e

PAGE 18

18 The overall molar-based equation for bacter ial growth can be obtained by using the Equations 1-2a and 1-2b (Grady, 1999). R = Rd fe*Ra fs*Rc (1-2a) fs + fe = 1 (1-2b) Rd is the equation for electron donor, Ra is the equation for electron acceptor, Rc is the equation for bacterial cell synthesis, fs is the fraction of transfe rred electrons used for cell synthesis, and fe is the fraction of electrons used for energy. The negative terms are inverted when used in the equation. The fraction of electr ons used for synthesis can be obtained from the growth yield. The carbon for synthesis is sh ifted to biomass and the carbon used for energy converted to carbon dioxide duri ng the oxidation-reduction reactions. When industrial carbon source is used, the formula for the electron donor can be estimated by analyzing the elemental compositions. Also, when nitrate is the pre dominant nitrogen form, the nitrogen must be reduced from +V state to III state before it can be used as the nitrogen source for cell synthesis (Grady, 1999). The equation for microbial growth is often written also with COD basis in terms of true growth yield. The general formula is given in Equation 1-3, Ss + (-(1-YH)) SN -------------> YHXBH (1-3) where Ss is amount of substrate COD, SN is the amount of termin al electron acceptor in COD units, YH is the growth yield of activ e heterotrophic bacteria, and XBH is the active heterotrophic biomass in COD units (Grady, 1999). Bacteria grow exponentially in favorable cond itions. When there is a limiting nutrient present, the specific growth rate coefficient of the bacteria, will depend on the concentration of the nutrient. The limiting nutrient can be the carbon source, nutrients, or other factor needed

PAGE 19

19 by the organisms for growth. The growth rate coefficient increases initially with increasing substrate concentration, and asymptotically ap proaches a maximum called the maximum specific growth rate, _max (Grady, 1999). A general equation used to characterize bacterial growth is the Monod equation (Equation 1-4). This equati on is developed strictly on empirical basis (Grady, 1999). = _max (Ss/(Ks + Ss)) (1-4) In the Monod equation, Ss is the substrate concentration and Ks is the half-saturation coefficient. Ks is the substrate concentration where the specific growth rate of the bacteria is equal to half of the maximum growth ra te. It can used to predict how fast _max is obtained. The Monod equation can be extended to consider th e effects of various substrates when multiple limiting nutrients are present, as shown in Equa tions 1-5 and 1-6 (Grady, 1999). Equation 1-5 is used when both growth limiting nutrients are required for biomass growth and they can influence the specific growth rate at the same time. Then the specific growth rate is calculated by taking the product of the Monod terms for the two nutrients. Equation 1-6 is used when both nutrients are required for biomass growth, but only one nutrient will limit th e growth at a given time. Then the specific growth rate is obtained by taking the lowest value obtained from separate single-substrate models. = _max (Ss1/(Ks1 + Ss1)) (Ss2/(Ks2 + Ss2)) (1-5) = min ( _max (Ss1/(Ks1 + Ss1))), ( _max(Ss2/(Ks2 + Ss2))) (1-6) The pH range of the system should also be c onsidered during the denitrification. The pH of the system is generally increas ed during denitrif ication since OHions are produced. The optimal pH range for the bacterial culture can be determined using variou s batch tests, and the

PAGE 20

20 bacterial growth can be recorded. Generall y, mild alkaline conditions are preferred for denitrification (Tchobanoglous, 2003). Bioreactor Configuratio n for Denitrif ication Two general anoxic models are used for den itrification processes: suspended growth systems and attached growth systems (Grady, 1999). In attached growth systems, microbes are grown attached to a packing material, generally rocks, sand, or other synthetic materials. The wastewater flows upward or downward, coming into contact with the atta ched biofilm, and the treated water flows out of the system. In anoxic systems, the packing material can be completely submerged, with gas space above the biofilm liquid layer (Grady, 1999). Excess biomass that may exit with the effluent can be separated us ing a clarifier and can be removed for further processing. In suspended growth systems, microbes ar e kept in a suspended state by appropriate mixing. Suspended growth systems can be opera ted in batch mode or continuous mode. In batch systems feed is added into the reactor with suspended biomass and is allowed to react to completion. Reaction conditions and growth envi ronment change with time. In continuous systems feed solution is continuous pumped into the reactor and an effluent is pumped out at the same flow rate. The concentratio ns of components in effluent are the same as the concentrations of components in the reactor. One example of a continuous suspended system is the continuous stirred tank reactor (CSTR). In a CSTR, the liqui d volume is kept constant and sufficient mixing is added to keep the conditions uniform throughout the reactor. A physical operation, such as sedimentation, may be implemented in order to separate biomass from the effluent. The overflow from the sedimentation basin will have very small concentration of biomass. The underflow will have very high concentration of bi omass, most of which is recycled to the bioreactor. A portion of the c oncentrated slurry is wasted, and is further treated before

PAGE 21

21 discharging it to the environment. In several operations, flexibility is implemented by placing several CSTRs in a series. The conditions in various stages of the system can vary, thus allowing for completion of various transformations Recycle may also be employed at desired stage of the system or through the entire chain (Grady, 1999). Objective of the Project The objective of this study was to develop a feasible denitr if ication process for treating an industrial nitrate stream. Attached growth bioreactors and suspended growth bioreactors were used for this study. Initial inoculum for the re actors were obtained from the denitrification basins at University of Florida Water Reclamati on Facility. The optimal pH for denitrification was determined to be between 8 and 9.5 in previo us studies. Since industr ial nitrate stream does not contain enough COD to support denitrification, potassium acetate was used as an exogenous carbon source. The goal of the project was to obtain a denitrification rate of 2 mg NO3-N/L/day.

PAGE 22

22 CHAPTER 2 DENTRIFICATION OF HIGH NITRATE IN DUST RIAL STREAMS USING GROWTH BIOREACTORS Initial attempts to denitrify the high nitrate industr ial stream were performed in attached growth bioreactors, Bios2 and Bios3. The subm erged attached growth bioreactors allow for short hydraulic residence times with high solids retention time s, and low solids waste after denitrification. A packed bed bioreactor with up ward flow and a recycle stream was used to conduct the experiments. Rock media obtained fr om Adventus was used to pack the reactor, which provided high surface area for the attached grow th of bacterial cells. The reactor was kept anoxic in order to allow for proper denitrification. Materials and Methods Setting up Bios2 A 3.78 L HDPE bottle was used to construct Bi os2. The set up of the bioreactor is s hown in Figure 2-1. Figure 2-1. Attached growth bioreactor Inlet ports Bioreactor Effluent tubing Effluent tank Recycle pump Feed pump

PAGE 23

23 The bioreactor contained eight feed inlet port s at the bottom and an outlet port at the top. This allowed for the feed to enter from all si des of the bioreactor, which allowed for better contact with the denitrif ying bacteria inside. An outlet por t was connected back to the feed stream as the recycle, and another outlet port wa s used to collect the overflow into a receiving bottle. A sample port was also constructed along the recycle outlet tubing. The feed stream was pumped from the feed tank and was allowed to mix with the recycle. This allowed for the regulation of pH of the stream entering the biore actor. Feed and recycle streams flowed through peristaltic pumps, and pump flow dials were used to monitor the flow rates. Feed and recycle pumps were standardized and calibration curves were used to obtain the flow rates with respect to pump dials. The mixed stream flowed through the eight in let ports, and was allowed contact with the bacteria as it flows upward and out the bioreactor. The top portion of the bioreactor was used as the gas separator. Tubing was attached to the lid of the bioreactor to al low the gas to escape. Cole Parmer L/S 13 tubing was used for all the c onnections in the bioreactor. The bioreactor was filled with stone media from Adventus and 1.5 L of inoculum from UF WRF. The working volume of the bioreactor was 3 L. Feed Preparation and Sampling for Bios2 Acetate was added as the carbon source for th e industrial stream that was fed to the reactor in the form of sodium acetate. The str eam was pre-neutralized in order to increase the pH, and had a concentration of approximately 1300 mg/L NO3-N. The carbon source for Bios2 feed was added in order to obtain 8:1 COD-NO3-N ratio. In order to obtain an 8:1 COD-NO3-N ratio, 10400 mg COD/L was required in the feed st ream. The initial COD was assumed to be 6000 mg/L, and thus 4.4 g/L of COD was needed in the feed solution. Thus 6.445 g/L of acetate

PAGE 24

24 was added to the feed solution. In order to increase the denitrif ication rate of the bioreactor, approximately 0.01 g/L of yeast extract and 0.3 ml/L molybdenum (as 1.1 g/L molybdic acid) were also added to the feed stream, providi ng nutrients to the ba cterial culture. The initial feed flow for Bios2 was turned on to 0.57 mL/min and the recycle was turned on to 6.67 mL/min. The recycle flow to feed fl ow ratio was 11.57. The inlet pH of the feed solution was maintained at 5. Concentrated hydr ochloric acid was added to the feed tank in increments of 10 mL if the feed st ream was measured to be high. During sampling, it was noticed that the sample port was in fact allowing a mixture of the feed stream and recycle stream to flow through the valve, and thus the nitrate concentrations measured through the valve were no t a true representation of the NO3 -N of the effluent stream of the bioreactor. Thus the sample port was moved to the recycle outlet, and sample was obtained by disconnecting the recycle tubing. This allowed for measuring the pH and NO3-N of both inlet stream and outlet stream of the bioreactor. A small volume of the sample was discarded initially each time in order to flush out the tubing and the sample port valve. The pH and NO3 -N tests were conducted three times a day, and the fl ow rates were adjusted accordingly. NO3-N was measured using HACH NitraVer test kits and measured us ing a spectrophotometer. In general, the sample was diluted to 20 times in order to obtain the nitrate concentration. Setting up Bios3 The m ain purpose of Bios3 was to compare the effects of sodium and potassium on denitrification. The construction and set-up for Bios3 was identical to that of Bios2. The reactor was filled with rock media from Adventus an d 1.1 L of inoculum from UF WRF was added. The working volume of the reactor was 3 L. Eight inlet ports were still us ed for feed flow, along with a recycle outlet, gas outlet, and a sample port for the inle t stream. Computer Controlled

PAGE 25

25 pumps were used for Bios3. The pumps were ca librated by measuring the effluent accumulated in the receiving bottle for a certain period of time. Feed Preparation and Sampling for Bios3 Bios3 was initially started with sodium hydr oxide neutralized industrial stream (sample A)-the same feed as Bios2. The feed was then switched to potassium hydroxide neutralized industrial stream (sample C). Po tassium acetate was used as the ca rbon source for the feed. The nitrate-nitrogen concentration of sample C was 2100 mg/L, which was much higher than Bios2 feed. 15 g/L of potassium acetate was added to the feed in order to ensure excess carbon, along with 0.024 g/L potassium phosphate as a source for phosphorus, 0.01 g/L yeast extract, and 3 mL of 1.1 g molybdenum in 1000 mL mo lybdic acid. The feed pH wa s maintained at 5, and 10 mL increments of hydrochloric acid were added to the feed if high pH was observed. Nitrate and pH measurements of the effluent were taken three ti mes daily. Nitrate-nitrogen measurements were taken using HACH NitraVer Test N Tubes and a spectr ophotometer. A small amount of the effluent was flushed out of the tubing before samples were take n. Two samples were collected for each measurement, one corresponding to the effl uent stream collected from the recycle outlet, the other sample corresponding to the inlet stream collected from the sample port valve. The bioreactor was started with a feed flow of 0.5 mL/min and a recycle flow rate of 2 mL/min. Results and Discussion Bios2 Results As shown in Figure 2-2A, Bios2 initially had high denitrification rates; however, the denitrification rate began to decrease eventually. A large am ount of gas bubbles were noticed in the reactor throughout its operation. The reactor was agitated periodically to release the gas bubbles. It was noticed that when the reactor was agitated, the denitr ification rates improved temporarily. However, the nitrate concentr ation in the reactor increased again. High

PAGE 26

26 denitrification rate was also obs erved after the addition of micronu trients to the feed and to the reactor. 40 drops of micronutrien ts were added to the feed and 5 drops were added directly to the reactor. The denitrification briefly increased after the add ition of micronutrients; however, the results did not stay positive for long. The highest flow rate achieved by Bios2 was 0.8 mL/min, with a denitrification ra te of 95.96%. This corresponded to a denitrification rate of 0.354 mg NO3-N/L/day. Thus, bioreactor failed to deni trify the industrial stream at the desirable rate, which was 2 mg NO3-N/L/min. Nitrate reduction and feed rate vs. time 0 10 20 30 40 50 60 70 80 90 100 6/76/176/277/77/177/278/6 Time (Days)% Nitrate Reduction0.0 0.2 0.4 0.6 0.8 1.0Feed Flow (mg NO3-N/L/min) % Nitrate Reduction Feed Flow New Bios2 Begin Feed pH decreases Media Agitated Feed w/ K-Acetate System Upset Micronutrients added Figure 2-2A. Bios2 denitrification and feed flow rate vs. time. As shown in the figure, a slight increase in denitrification was observed when the media was agitated. However, it still continued to perform poorly. An increase in denitrification wa s also shown immediately after the addition of micronutrients. Also, as shown in the figure, when the pH of the effluent was measured to be higher than 9.5, the pH of the f eed was decreased by the addition of concentrated HCl to the feed tank. Target feed rate for the reactor was 2 mg NO3-N/L/min.

PAGE 27

27 Start-Up 2/5/2007 ; Feeding Begins 2/12/20070.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 6/76/176/277/77/177/278/6 Feed pH decreases Media Agitated Feed w/ K-Acetate begin System Upset Micronutrient s added Figure 2-2B. Bios2 pH and n itrate concentrations vs. time One factor that was taken into account afte r Bios2 failed to denitrify was the effect of sodium on denitrification. It was hypothesized that sodium c ontribute significantly to the osmotic pressure across the cell membrane, which would affect the metabolism and denitrification of the bacteria. An alternate to using sodium salt for neutralization and as carbon source was to use potassium salts. Even t hough potassium ions contribute to the osmotic pressure also, their effects would be smaller than those of sodium ions. Therefore, a new bioreactor was started in order to examine if potassium would lead to high denitrification rates of the industrial stream samples. Bios3 Results Bios3 had initial high denitrificat ion rates; however, on the w hole, it had the same results as Bios2, as shown in Figure 2-3A. It was noticed that Bios3 recovered faster from system upsets than Bios2, possibly due to the presence of potas sium instead of sodium. In general, no

PAGE 28

28 significant changes were observed from Bios2 re sults. An increase in denitrification was observed after adding micronutrients to the feed so lution (4 drops/L) and di rectly to the reactor (5 drops). However, the high denitrificati on rate was temporary. Gas bubbles were also observed in the walls of the reactor Agitation of the reactor failed to increase its performance. The highest flow rate that was achieved by Bios2 was 1.2 mL/min, w ith an overall denitrification of 98.7%. This corresponded to a de nitrification rate of 0.612 mg NO3-N/L/day. Therefore, Bios3 also failed to meet the target denitrification rate of 2 mg NO3-N/L/day. The results showed using potassium salts did not have any significant impact on denitrification. nitrate reduction and feed rate vs. time 0 10 20 30 40 50 60 70 80 90 100 7/97/147/197/247/298/38/88/138/188/23 Time (Days)% Nitrate Reduction0.00 0.20 0.40 0.60 0.80 1.00 % Nitrate Reduction Feed Flow Bios3 Begin System Upset Micronutrients added Flushed out with water Figure 2-3A. Bios3 denitrification and feed flow rate vs. time. Bios3 results did not vary significantly from Bios2 result s, even though potassium hydroxide was used to neutralize the feed rather than sodium hydroxide. Notice that an increase in denitrific ation was also observed in Bios3 immediately after the addition of micronutrients.

PAGE 29

29 Start-Up 2/5/2007 ; Feeding Begins 2/12/2007 0 2 4 6 8 10 7/97/147/197/247/298/38/88/138/188/23 NO3-N (100 mg/L) pH Micronutrients added Figure 2-3B. Bios3 pH and n itrate concentration vs. time After Bios2 and Bios3 failed to denitrify industrial nitrate stre ams properly, several factors were considered that might have result ed in low performance of the reactors. One possibility was the presence of metals in high concentrations in the feed, which can affect bacterial metabolism. A metals analysis of Bios2 feed, Bios3 feed, and Bios3 effluent was conducted using Inductively Coupled Plasma (ICP). However, very low concentrations of metals were found in all samples. Another factor was the high nitrat e concentration of the industrial stream. The effects of high nitrate concentration of the feed on denitrification were not studied. Also, clogging of the tubi ng could have occurred due to various inlet ports and low flow rate of the feed and recycle. Clogging in the po res of the rock media c ould have occurred also, due to high concentration of biomass, which can lower the surface area for the attached growth of bacterial growth. Channeling and short-circuiting inside the reactor is also possible, and proper contact of the feed stream with the bacteria can be diminished. Another factor considered was the presence of toxic materials in the feed a ffecting bacterial growth. It was also possible that toxic substances present in the feed accumu lated on the rock media, which can affect the metabolism of the attached bacteria and cause low de nitrification.

PAGE 30

30 At this point, it was concluded that the low de nitrification was attri buted to some possible factors related to reactor design and some possibl e factors related to th e feed. In order to distinguish the causes, a standard suspended growth reactor was started. The suspended growth reactor used was Continuous Stirred Tank Reactor (CSTR).

PAGE 31

31 CHAPTER 3 DENITRIFICATION OF HIGH NITRATE IN DUST RIAL STREAMS USING SUSPENDED GROWTH BIOREACTORS Suspended growth bioreactors consist of react or vessels containing biomass growing in a suspended state rather than attached to a medi a. Continuous stirred ta nk reactor (CSTR) is a suspended growth bioreactor where the contents in the reactor are stirred c ontinuously in order to keep the conditions homogeneous within the reactor and the culture suspended. The stream to be treated is fed into the reactor, and the treated liquid mixed with biomass is pumped out as the effluent. The fractions of constituents in the effluent are identical to those inside the reactor. Many processes have a recycle stream where most of the biomass is recycled into the reactor. Materials and Methods Setting up Suspended Growth Bioreactor The assembly of the continuous st irred tank is given in Figure 3-1. Figure 3-1. Continuous stirred tank reactor As shown in Figure 3-1, the continuous stirre d tank reactor was constructed using a 4.2 L flask. The flask was filled 10% with feed so lution, and 90% with inoculum from UF WRF denitrification basin. The reactor was then set in a stir plate with the stirrer inside and was Gas outlet port Gas inlet port Effluent tubing Influent tubing Feed tank stirrer Computer controlled pump

PAGE 32

32 stirred continuously. A rubber stopper was used to close the reactor and it was secured to the reactor using parafilm. The rubber stopper had a feed inlet port, an effluent port, a gas inlet port, and a gas outlet port. A nitrogen tank was se tup beside the reactor and nitrogen was bubbled through the reactor twice da ily for 10 minutes to strip off any oxygen that is present. Initially, the reactor was started up in the batch mode. Samples for the nitr ate tests were taken through the inlet gas port. The reactor was operated in ba tch mode until the pH and absorbance increased, and then it was switched to continuous mode. For the continuous mode, a computer controlled pump was set up near the reactor with the same inlet and outlet flow rate s. The inlet tubing was secured into the feed port in th e rubber stopper, and the outlet tubing was attached to the effluent port. Cole Parmer L/S 13 tubing was used for feed and effluent connections. The effluent from the reactor was collected and measured at a specific time period in order to calibrate the pump. Sampling and Data Collection NO3-N, pH, and absorbance measurements were taken twice daily from the effluent of the reactor. In continuous mode, samples were ta ken from the effluent. The reactor was purged with nitrogen gas after each sampling and the gas inlet port was then secured firmly in order to keep the anoxic conditions in the bioreactor. The gas outlet port was closed only slightly in order to vent out gas formed in the reactor. Nitrate-nitrogen concentrations were measured using HACH NitraVer TestN Tubes and a HACH spect rophotometer. In general, the effluent sample was filtered and diluted to 20 times before nitrate-nitrogen measurements. Absorbance measurements were representative of the bacteria l concentration in the reactor, and were taken using a spectrometer. The results from nitratenitrogen, pH and absorbance measurements were used for adjusting the flow rates in the reacto r. COD measurements were taken periodically using HACH COD test kits. The samples were filtered and diluted to 20 times. After the

PAGE 33

33 samples were added to the test tubes with th e reactive material, they were placed in a COD reactor for two hours. COD measurements were then taken using a spectrophotometer. Feed Preparation The high nitrate industrial str eam obtained was pre-neutralized to appropriate pH using either sodium hydroxide or potassium hydroxide. Potassium acetate was used as the carbon source for the bacteria. 17.5 g/L of potassium acetate was used in the feed, along with 0.02 g/L potassium phosphate. A micronutrient solution was prepared using instructions for trace element solution by Vishniac & Santer. The contents in the solu tion are given in Table 3-1. The micronutrient solution was added in the feed initia lly in the amount of 4 drops/L. Later in the experiments, micronutrients were added directly to the reacto r rather than in the feed. Table 3-1. Contents of the micronutri ent solution (Vishniac and Santer, 1957) Ingredient Amount (per 500 mL DI water) Ethylenediaminetetraacetic acid (EDTA) 25.0 g ZnSO4 11.2 g CaCl2 2.79 g MnCl 2.58 g FeSO4 2.59 g Mo7O24(NH3)6 0.6 g CuSO4.5H2O 0.785 g CoCl2.6H2O 0.81 g Magnesium sulfate 2.525 g Molybdic acid 0.55 g Artificial nitrate solution used in the experiments was prepar ed using 69.6 weight percent nitric acid assay and DI water. The so lution was prepared by adding 8.6 mL of HNO3 assay per one liter of DI water to obtain the same NO3-N concentration as sample A. The solution was neutralized by adding approximately 4 g NaOH pelle ts per liter of solution. The feed solution also contained 17.5 g/L potassium acetate as the carbon source and 0.024 g/L potassium phosphate. The nitrate-nitrogen concentration of the feed was approximately 1450 mg/L, which was comparable to the nitrate-nitrogen concentration of sample A.

PAGE 34

34 Results and Discussion CSTR1 Results CSTR1 was started in order to denitrify sam p le A neutralized with NaOH in the tote. The reactor was started in batch mode, and was sw itched to continuous mode at a flow rate of 1 mL/min once the pH and absorbance increased. Micronutrients were added in the feed of the reactor in the amount of 4 dr ops/L. The results of the reactor are shown below: nitrate reduction and feed rate vs. time (Target feed rate: 2 mg NO3-N/L/min)0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00 8/58/158/259/49/149/24time% nitrate reductio n 0 0.2 0.4 0.6 0.8 1feed flow rate (mg NO3-N/L/min) % nitrate reduction feed flow rate Figure 3-2. CSTR1 nitrate reducti on and feed flow rate vs. time

PAGE 35

35 pH and nitrate concentration vs. time0 1 2 3 4 5 6 7 8 9 10 8/58/158/259/49/149/24 timepH0 500 1000nitrate concentratio n pH NO3-N (100 mg/L) Figure 3-3. CSTR1 pH and nitr ate concentrations vs. time As shown in Figure 3-2, the deni trification rate of CSTR1 wa s initially high. However, it started to decrease after 8 days and continued to decrease without possibility of recovery. The pH began decreasing and nitrate concentration st arted decreasing with time, as shown in Figure 3-3. The reactor failed to recover ev en after switching to batch mode. After CSTR1 failed, possible causes that might ha ve contributed to the low denitrification of the industrial stream were considered. It was hypothesized that a toxic component in the sample might be accumulating in the reactor and was not consumed during the metabolism of the bacteria. Possible candidates were sodium buildup, high nitrate buildup, or accumulation of some other toxic component in the industrial so lution. An analysis was done to predict the concentration of a toxic compound that might be accumulating in the reactor after a certain time period. Analysis of CSTR1 Results After the hypothesis that the accumulation of a constituent is causing low denitrification, a modeling analysis was done on CSTR1 results in order to predict the accumulation period of the constituent. A mass balance was done on the reactor as shown in Equation 3-1.

PAGE 36

36 d/dt (VC) = FCf FCo V r (3-1) at constant volume, V dC/dt = FCf FCo V r (3-2) dC/dt = (F/V) Cf (F/V) Co r (3-3) (F/V) = D (3-4) (D = dilution rate) dC/dt = D Cf D Co r (3-5) Developing Equations When accumulation is considered, the reaction is neglected; then dC/dt = D Cf D Co (3-6) Since dilution rate changes with fl ow rate and thus with time, dC/dt + D(t) Co = D(t) Cf (3-7) Using integrating factor, IF = exp( D( )) d D( ) = + Do (3-8) exp( ( + Do) d = exp( 2/2 + Do ) (3-9) Multiplying by IF gives exp( 2/2 + Do ) dC/dt + exp( 2/2 + Do ) Co = exp( 2/2 + Do ) Cf (3-10) t d/dt(exp( 2/2 + Do ) C) = (exp( 2/2 + Do ) D(t) Cf (3-11) C exp( 2/2 + Do ) Co = D( ) Cf exp( 2/2 + Do ) d (3-12) C exp( 2/2 + Do ) Co = ( + Do ) Cf exp( 2/2 + Do ) d (3-13) C(t)=Co exp -( t2/2 + Dot)+ exp -( t2/2 + Dot) ( + Do ) Cf exp( 2/2 + Do ) d (3-14) At constant dilution rate, = 0, then C(t) = Co exp -(Dot) + Cf exp -(Dot) (Do ) exp(Do ) d (3-15) C(t) = Co exp(-Dot) + Cf (1-exp(-Dot)) (3-16) When dilution rate is not constant, it is a ssumed to change linearly with time; then D = + Do. Substituting gives ( + Do)d = d ( 2/2 + Do ) (3-17) Cf 0t ( + Do) exp(2/2 + Do ) d = Cf 0t exp( 2/2 + Do ) d ( 2/2 + Do ) (3-18)

PAGE 37

37 Cf 0t exp( 2/2 + Do ) d ( 2/2 + Do ) = Cf (exp( 2/2 + Do ) 1) (3-19) C(t) = Co exp(-( 2/2 + Do )) + Cf exp(-(2/2 + Do )) (exp( 2/2 + Do ) 1) (3-20) C(t) = Co exp(-( 2/2 + Do )) + Cf (1exp(-( 2/2 + Do ))) (3-21) Applying CSTR1 Results The results were analy zed over three time peri ods. As shown in Figure 3-4, time period 1 was taken from to to t1, time period 2 was taken from t1 to t2, and time period 3 was taken from t2 to t3. It was assumed that time period 1 and tim e period 2 had constant dilution rate, and time period 3 had dilution rate that cha nged linearly with time. At t = to, Co was taken to be zero; thus it was assumed that no toxins were present in the system initially. The results of the analysis are shown in Tables 3-2, 3-3, and 3-4. 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00 8/58/108/158/208/25time% nitrate reductio n 0 0.5 1 1.5 2 2.5 3feed flow rate Figure 3-4. Time periods used for CSTR1 data analysis to t1 t2 t3

PAGE 38

38 Table 3-2. Concentration after the first time period for CSTR1 After the first time period: C(t) = Co exp(-Dot) + Cf (1-exp(-Dot)) Co = 0 C1(t) = Cf (1exp(-Dot)) t = 3420 minutes Do (dilution rate) = F/V = (1 mL/min)/(4000 mL) = 0.00025 min-1 C1(t) = Cf (1exp((-0.00025/min)(3420 min)) C1(t) = 0.575 Cf Thus, after the first time period, 57.5% of the initi al toxin concentration will have accumulated Table 3-3. Concentration after the second time period for CSTR1 After the second time period: C2(t) = C1(t) exp(-Dot) + Cf (1-exp(-Dot)) t = 3365 minutes Do = F/V = (0.5 mL/min)/(4000 mL) = 0.000125 min-1 C2(t) = 0.575 Cf exp((-0.000125/min)(3365 min))+Cf (1exp(-0.000125/min)(3365 min)) C2(t) = 0.721 Cf Thus, after the second time period, 72.1% of th e initial toxin concentration will have accumulated in the reactor. Table 3-4. Concentration after the third time period for CSTR1 After the third time period: Dilution changes linearly with time. Thus D(t) = + Do t = 4500 minutes and Do were determined from the plot of dilution rate vs. time, as shown in Figure 3-5 From the plot, = 5.4E-8 Do = 0.0002 C3(t) = Co exp(t2/2 + Dot) + Cf (1exp(t2/2 + Dot)) Co = C2(t) = 0.721 Cf exp(t2/2 + Dot) = exp{(-5.4E-8*(4500 min)2/2)+((0.000241 min-1)(4500 min))} = 0.196 C3(t) = 0.721 Cf (0.196) + Cf (0.804) C3(t) = 0.946 Cf After the third time period, 94.6% Co w ill have accumulated in the reactor.

PAGE 39

39 dilution rate vs. timey = 5E-08x + 0.0002 R2 = 0.9859 0 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006 0500100015002000250030003500400045005000 time (minutes)dilution rate (min^-1) Figure 3-5. Dilution rate of CSTR1 with respect to time. From the linear regression, was calculated to be 5E-08 and Do was obtained to be 0.0002. The analysis showed that af ter 57 hours, 57.5% of the initial toxic component would have accumulated in the reactor. The model predicts that at the time of failure (8/15/2007), approximately 94.6% of the feed concentration of the toxin would have accumulated in the reactor, which can cause significant impact on the de nitrification rates. Therefore, a new reactor was started using synthetic nitr ate solution to investigate whether the accumulating toxin is a component present in the industria l feed or the sodium ions adde d to the feed for neutralization. CSTR2 Results CSTR2 was started using synt hetic nitrate solution prepared using nitric acid and DI water. Cole Parmer L/S 13 tubing was used to pump the feed into the reactor, and Cole Parmer L/S 14 tubing was used to pump out the effluent. Once the pH and the absorbance of the reactor increased, the reacto r was switched to continuous m ode. Two computer controlled pumps were placed near the reactor; one pump wa s used for feed and the second pump was used for the effluent. Two pumps were required becaus e of the difference in feed and effluent tubing sizes. Micronutrients were added to the feed in the amount of 4 drops/L. The results of the reactor are shown in Figure 3-6.

PAGE 40

40 nitrate reduction and feed rate vs. time: Part 1 0.00 20.00 40.00 60.00 80.00 100.00 8/319/59/109/15time 0 2 4 % nitrate reduction feed flow rate Figure 3-6. CSTR2 denitrif ication and feed flow rate vs. time, part I Initially on continuous mode, the reactor wa s started at 0.5 mL/min. The reactor was denitrifying very well at this flow rate. The flow rate of the fe ed was increased up to 2 mL/min and the reactor continued to denitr ify at high rates, as shown in Figure 3-6. However, the flow rate was increased without allowing the reactor to reach stable conditions. Increasing the feed flow required an increase in effl uent flow rate as well, and in creasing the flow prematurely can lead to washout conditions in the reactor. Ther efore, the feed flow ra te was decreased to 1.2 mL/min in order for the reactor to stabilize. Th e initial positive results seen in the reactor were brief, however, and the nitrate concentration began to increase again in the reactor. The reactor was switched to batch mode after very low denitrification rates were observed. It was noticed that denitr ification rate was high in th e reactor for approximately 265 hours (11 days). The low denitrif ication rates and low absorbance va lues were observed later on. An analysis was done on the accumulation of toxins in the reactor using Equation 3-21. The initial concentration of the toxin, Co, is assumed to be zero. The ratio of C(t)/Cf was plotted against time. The results are shown in Figure 3-7

PAGE 41

41 CSTR2 accumulation of a toxin0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 01000200030004000500060007000 time (minutes)C(t)/Cf Series1 Figure 3-7. Ratio of C(t)/Cf vs. time for component accumulation in CSTR2 As shown in the Figure 3-7, within after a time period of approximately 6400 minutes, the ratio C(t)/Cf has reached the value of 0.86. Thus, a deterioration of the bioreactor performance should have been observed by this time. After 188 hours (9/9/2007), the toxic component concentration would have reached 99% of the feed concentration. However, as shown in Figure 3-6, the reactor continued to perform well for tw o more days. Thus, accumulation of a toxin was not causing the low denitrification rates in the reacto r. It was then suspected that deficiency of a key constituent might be causing the low denitrif ication. Thus, the results were analyzed in order to predict the time that defici ency of a component might occur. Analysis was done for the first 6400 minutes of the operation of the reactor, where a linear increase in flow rate, and thus dilution rate, wa s observed. The dilution rate was calculated using the relationship given in Equati on 3-8, where D(t) is equal to t + Do. and Do were obtained from the plot of dilution rate vs time, as shown in Figure 3-8.

PAGE 42

42 y = 5.3931E-08x + 1.3776E-04 R2 = 9.6752E-01 0 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006 01000200030004000500060007000 Figure 3-8. Plot of dilution rate vs. time for CSTR2 Another plot of C(t)/Co vs. time was generate d in order to quantify the depletion of a component initially present in the system. Analysis done using the data obtained from the reactor showed that after approximately 107 ho urs, the concentration of a component added initially to the reactor would have reached appr oximately 13% of the feed concentration. The results are shown in Figure 3-9. CSTR2 deficiency of a nutrient 0 0.5 1 1.5 01000200030004000500060007000 timeC(t)/Co Figure 3-9. Ratio of C(t)/C(o) vs. time for component deficiency in CSTR2 Since the reactor was performing very well up to this point, it was noted that the ingredient that is exhausted is needed only in very small am ounts for denitrification. Further analysis of the data showed that at the time of the failure (9/11/2007), the am ount of the nutrient will have reached 0.6% of its initial concentration. One component that was required for bacterial growth in very small amounts was the micronutrients. It was noted that attached film bioreactors had started to recover previously wh en micronutrients were added to the feed and directly to the

PAGE 43

43 reactor. However, the effect wa s short-lived and the reactors deteri orated again. This led to the conjecture that micronutrients added to the feed may have pr ecipitated, as suggested by Ben Koopman. Thus, most of the mi cronutrients added to the feed of the CSTR would have been precipitated out. Taking th is into consideration, micronutrients were added directly to the reactor everyday in proportion to the flow rate of the feed. Generally, 1 mL of micronutrients was added to the reactor when the feed flow rate is 1 mL/min. Magnesium was also added to the feed solution in the form of magnesium sulfate heptahydrate in order to ensure its availability. After the micronutrients were added, an incr ease in the reactor performance as well as bacterial growth was observed, as shown in Figure 3-6. These results showed that it was indeed the deficiency of micronut rients that caused low denitrification in the reactor previously. It was also determined that the additi on of micronutrients to the feed can cause precipitation, and thus, it must be added directly to the reactor. Th e amount of potassium phosphate added to the feed was also increased to 0.5 mg/L in or der to avoid phosphorus deficiency. CSTR2 nitrate reduction and feed rate vs. time Part II 0.00 20.00 40.00 60.00 80.00 100.00 8/319/109/209/3010/1010/2010/3011/911/1911/2912/912/19 time% nitrate reduction0 2 4feed flow rate (mg NO3-N/L/min) % nitrate reduction feed flow rate feed rate lowered due to limited carbon availab ility Figure 3-10. CSTR2 denitrification and feed flow rate vs. time, part II. As shown in the figure, a deficiency in carbon occurred during the operation, and the flow rate to the reactor was set at the minimum for 2 days. The flow rate was increas ed rapidly after carbon was available to reach appropriate steady state.

PAGE 44

44 pH and nitrate concentrations vs. time0 1 2 3 4 5 6 7 8 9 109/159/2310/110/910/1710/2511/211/1011/1811/2612/412/1212/20timepH0 500 1000nitrate concentratio n pH NO3-N (100 mg/L) Figure 3-11. CSTR2 pH and ni trate concentrations vs. time Absorbance vs. time 0 0.5 1 1.5 2 2.5 3 3.5 49/159/2310/110/910/1710/2511/211/1011/1811/2612/412/1212/20 timeAbsorbance absorbance Figure 3-12. CSTR2 absorbance vs. time After the addition of micronutrients, increase in denitrification rate and absorbance was observed in the reactor. The overa ll denitrification remained high with increasing flow rates, as shown in Figure 3-10. The pH of the reactor was in between 9 and 9.5 throughout its operation,

PAGE 45

45 as shown in Figure 3-11. The ab sorbance decreased generally with an increase in feed flow rate, as expected since the effluent flow rate was also increased. The absorbance stabilized once the reactor reached steady state for e ach flow rate. The flow rate of the reactor was continuously increased afterwards up to a flow rate of 11.5 mL/min, as shown in Figure 3-10, which was close to twice the target flow rate, and the highest denitrification ra te observed in CSTR2 was 4 mg NO3-N/L-min. From these results, it was confirme d that high nitrate concentration did not affect the metabolism and denitrification rate of the denitrifying bacteria. It was furthermore speculated in the begi nning that high amount of sodium can lower denitrification rates since it wi ll lead to high osmotic pressure across the cell membrane. The effect of sodium was also tested previously using attached growth bioreactors-by operating a reactor containing sodium hydroxide neutralized industria l stream and anothe r reactor containing potassium hydroxide neutralized indu strial stream simultaneously. However, the results from the attached growth bioreactors were not reliable since the reason fo r poor denitrification could have been micronutrient deficiencies. Therefore, a new CSTR was started alongside CSTR2 in order to test the effects of sodium on denitrification rate. The new r eactor was started using synthetic nitrate stream using DI water a nd 69.6% nitric acid assay, and th e solution was neutralized using potassium hydroxide pellets instead of sodium h ydroxide pellets. Ther e werent any noticeable differences between the results obtained from sodium hydroxide neutralized solution and potassium hydroxide neutralized solution. Theref ore, it was concluded that sodium does not have any significant eff ects on denitrification. CSTR3 Results After it was confirmed that high nitrat e concentration and sodium did not have significant effect on denitrifica tion rate, another suspended grow th reactor, CSTR3, was started again with the industrial stream as the feed solution. Effluent from CSTR2 was used as the

PAGE 46

46 inoculum for the reactor. The reactor was init ially started with artificial nitrate solution neutralized with potassium hydroxide pellets as the feed. The f eed was later switched to ionexchanged industrial stream that was pre-neut ralized with potassium hydroxide. The feed solution contained 17.5 g/L potassi um acetate as the carbon sour ce, 0.5 g/L potassium phosphate as the phosphorus source, and 0.1 g/L magnesium sulfate heptahydrate in order to provide additional magnesium for bacterial metabolism. Th e results of the reacto r are shown in Figures 3-13, 3-14, and 3-15. The reactor continued to denitr ify at a high rate using samp le B (ion-exchanged industrial stream), as shown in Figure 3-13. The highest flow rate obtained using sample B was 0.8 mL/min with an overall denitrific ation of 99.97%. This corresponded to a denitrification rate of 0.315 mg NO3-N/L-min. The reactor was not able to achieve higher flow rates because of the shortage of available sample B solution. CSTR3 nitrate reduction and feed rate vs. time 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00 9/139/2310/310/1310/2311/2time% nitrate reduction0 0.2 0.4 0.6 0.8 1feed flow rate (mg NO3-N/L/min) % nitrate reduction feed flow rate A B C Figure 3-13. CSTR3 denitr ification and feed flow rate vs. time A: HNO3-KOH feed started; B: Switched feed to sample B, C: Switched feed to sample C

PAGE 47

47 nitrate concentration and pH vs. time0 1 2 3 4 5 6 7 8 9 109/139/169/199/229/259/2810/110/410/710/1 0 10/1 3 10/1 6 10/1 9 10/2 2 10/2 5 10/2 8 10/3 1 11/3timepH0 500 1000nitrate concentratio n pH NO3-N (100 mg/L) switched feed to sample C Figure 3-14. CSTR3 pH and ni trate concentration vs. time absorbance vs. time 0 0.5 1 1.5 2 2.5 3 3.5 4 9/139/169/199/229/259/2810/110/410/710/1 0 10/1 3 10/1 6 10/1 9 10/2 2 10/2 5 10/2 8 10/3 1 11/3timeabsorbance absorbance Figure 3-15. CSTR3 absorbance vs. time

PAGE 48

48 After sample B was exhausted, the feed was switched to sample C, which was non-ion exchanged industrial stream neutralized with pot assium hydroxide. Changes in performance of the reactor were not speculated due to this actio n, since it was hypothesized that it was indeed the deficiency of micronutrients that caused low denitrification of the previous industrial streams. However, as shown in Figure 3-13, the denitrificati on rate started to decrea se as soon as the feed was switched to sample C. Figures 3-14 and 315 correspond to the pH and absorbance of the reactor, respectively. It can be seen clearly that the change in feed led to decrease in pH and absorbance as well. The reactor was then put on batch mode for three days and the feed was switched back to synthetic feed neutralized w ith potassium hydroxide. The reactor failed to recover from the low denitrification rate, however. This led to the belief that metal ions are interfering with bacterial growth and denitrificati on process. It was also noticed that sample C had a tint of green that was not observed in previous samples. CSTR4 Results After CSTR3 failed to denitrify sample C properly, other factors that may cause poor denitrification were considered. The main concer n was the presence of metals in the industrial stream, which can affect bacterial metabolism and lower denitrification rates. This was further supported by the results obtained from denitrific ation of sample C, which was ion-exchanged. After CSTR3 was discontinued, two new batches of ion-exchanged industrial stream neutralized with KOH were received (sample D and sample E) CSTR4 was started up using sample D. The feed also contained 17.5 mg/L potassium acet ate, 0.5 mg/L potassium phosphate, and 0.1 mg/L magnesium sulfate heptahydrate. Effluent fr om CSTR2 was used as the inoculum for the reactor. Micronutrients were a dded to the reactor daily in amount s proportional to the feed flow rate. The results of th e reactor are shown in Figures 3-16, 3-17, and 3-18.

PAGE 49

49 nitrate reduction and feed flow rate vs. time 0.00 20.00 40.00 60.00 80.00 100.00 10/ 25 10/ 28 10/ 31 11/ 3 11/ 6 11/ 9 11/ 12 11/ 15 11/ 18 11/ 21 11/ 24 11/ 27 11/ 30 12/ 3 12/ 6 12/ 9 12/ 12 12/ 15 12/ 18 12/ 21 12/ 24 time% nitrate reduction0 0.5 1 1.5 2feed rate (mg NO3N/L/min) % nitrate reduction feed flow rate Sample D Sample E synthetic nitrate solution Sample A Figure 3-16. CSTR4 denitrificati on and feed flow rate vs. time. Low denitrification rates were observed for both ion-exchanged samples. High denitrification was observed once the feed was switched to sample A. pH and nitrate vs. time 0 2 4 6 8 1010/26 11/3 11/11 11/19 11/27 12/5 12/13 12/21timepH0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300nitrate-nitroge n concentration pH NO3-N (100 mg/L) Figure 3-17. CSTR4 pH and ni trate concentrations vs. time Absorbance 0 0.5 1 1.5 2 2.5 3 3.5 4 10/2611/311/1111/1911/2712/512/1312/21 timeAbsorbance absorbance Figure 3-18. CSTR4 absorbance vs. time

PAGE 50

50 CSTR4 did not have positive denitrificati on results initially. The pH and absorbance decreased steadily as shown in Figures 3-17 and 3-18, and the denitrification was deteriorating slowly as shown in Figure 3-16. The results we re puzzling since the sample was ion-exchanged, and thus it invalidated the assumption that meta l ions may be causing the problem. The reactor failed to recover after switching to synthetic feed. In order to be certain that the ion-exchange was effective, an analysis of the samples were done at the industrial site, and it was observed that one of the samples had high concentration of meta l ions that might inhibi t the bacterial growth and lower the denitrification rate It wasnt positive which of the samples contained the high metal ion content. It was possibl e that the low denitrification of the reactor was due to presence of metal ions in the feed. It was then decided to restart CSTR4 and obser ve the denitrification rates of sample E. Therefore, the reactor was rest arted with synthetic feed and inoculum obtained from CSTR2 effluent. Once the reactor was in steady state with a flow rate of 1 mL/min, the feed was switched to sample E. The denitrification rate was high using this sample; however, the absorbance started decreasing. Th is led to the conclusion that metals were not the inhibition factor in low denitrification rates observed. Restarting CSTR4 with Sample A CSTR4 was restarted with sample A since it was suspected that previous low denitrification rates were caused by deficien cy of micronutrients. When sample A was previously fed into the reactor, micronutrients were added in the feed solution rather than directly to the reactor. Thus, the reactor was restarted in conti nuous mode at 1mL/min. Effluent from CSTR2 was used as the inoculum for the reactor. High denitrification rates were observed thr oughout the operation of the reactor, and pH remained steady. The reactor was able to achieve high flow rates with high overall

PAGE 51

51 denitrification. The highest flow rate achie ved by the reactor was 5.5 mL/min, with 97.5% denitrification, as shown in Figure 3-16. This corresponded to a denitrif ication rate of 1.68 mg NO3-N/L/day, which was close to the target denitrification rate of 2 mg NO3-N/L/day. At this point it was conclude d that the industrial stream was indeed treatable, but a component in the stream is acting as an inhibito ry factor. Causes for denitrification include presence of metal ions and presence of volatile contaminants. Thus, a tool was needed to judge the treatability of the samples. Thus, a batch test was developed in order to test the treatability of high nitrate streams. It was reve aled from the results of the batch tests that volatile contaminants are indeed present in the industr ial stream and they might be the components contributing to the low denitrification rates and low absorbance values.

PAGE 52

52 CHAPTER 4 ASSAY ON THE BIOLOGICAL TREATABI L ITY OF INDUSTRIAL STREAMS A batch test procedure was deve loped to test the biological tr eatability of high nitrate water streams. The assay was inspired by studies on Daphnia by Gasith et. al. Since the parameter to be tested was denitrification, Daphnia was replaced with denitrifying bacterial culture. The denitrification results of the te st solution are compared with those of the synthetic nitrate solution. The experiments were mainly conducted on ion-exchanged high nitrate industrial stream neutralized with potassium hydroxide and non-ion exchanged industrial stream neutralized with sodium hydroxide. The batch tests primarily invest igated the eff ects of volatile contaminants and settled particles on denitrificat ion and bacterial growth. The general procedure developed to test the treatability of the high nitrate streams is as follows: 1. Grow denitrifying bacteria in a high nitrate stream so that th ey can be used for the batch tests. 2. Measure the nitrate-nitrogen concentration and pH of the test solution using HACH NitraVer Test N Tube test kits 3. Prepare synthetic nitrate solution using nitr ic acid with the same nitrate-nitrogen concentration as the test solution. 4. Add appropriate chemicals (given as follows) to variable and contro l solutions in order to support bacterial growth and metabolism: a. 17.5 g/L potassium acetate as carbon source b. 0.5 g/L potassium phosphate as phosphorus source c. 0.1 g/L magnesium sulfate heptahydrate as magnesium source d. 0.3 g/L ammonium chloride as the ammonium source 5. Neutralize the test solution a nd synthetic nitrate solution to pH 8 using sodium hydroxide pellets 6. Filter the solution using 2 m filters. 7. Aerate approximately 200 mL of test solution, and keep the rest non-aerated. 8. Obtain 9 125-mL Erlenmeyer flasks; fill three flasks with 60 mL of aerated test solution, three flasks with 60 mL of non-aerated test solution, and three fl asks with 60 mL of synthetic nitrate solution.

PAGE 53

53 9. Add 10 mL of bacterial cu lture and 0.2 mL micronutrien t solution-prepared using instructions for trace element solution by Vishniac & Santer to each flask. 10. Measure the nitrate-nitrogen concentration, pH, and absorbance of each flask. The nitrate-nitrogen concentration is measured using HACH test kits after filtering the sample from each flask using 0.45 m filters, and diluting the samples to appropriate concentrations. 11. Secure all flasks with rubber stoppers and place them in an incubator at 37oC 12. Measure NO3-N, pH, and absorbance of one aera ted flask and one non-aerated flask three times: first measurement after 3-4 hour s, second measurement after 7-9 hours, and third measurement after 22-24 hours. 13. Obtain ( _max/YNO)sample / (( _max/YNO)synthetic solution from the data obtained in order to establish the treatability of the high nitrate stream; treatable nitrate solutions will have a value close to 1. Developing Equations Bacterial growth in the test flasks can be quantified since the bi omass concentration is proportional to absorbance values. Biomass growth kinetics is developed assuming the presence of excess carbon and neglecting decay. The change in absorbance can be characterized by Monod equation. dA/dt = maxNO (SNO/(KNO + SNO)) A, nitrate is the limiting component (4-1) dSNO/dt = maxNO (SNO/(KNO + SNO)) A (4-2) dSNO/dt = (-1/YNO) (dA/dt) (4-3) A(t) = Ao + YNO (SNO(o) SNO(t)) (4-4) Thus, a plot of absorbance vs. nitrate concentration will have a slope of -YNO and yintercept of Ao. This plot can be obtained fr om the measurements of absorbance and nitrate concentration after 0, 4, and 8 hours. It was noted that KNO is typically less than 1, usually for wastewater streams containing nitrate concentration around 100 mg/L. Since the nitrate concentration is much higher th an 100 mg/Lin the experiments done with this assay, the ratio

PAGE 54

54 (SNO/(KNO + SNO)) is approximated to be equal to 1. Then Equation 4-2 can be rewritten as shown in Equation 4-5. dSNO/dt = maxNO [Ao + YNO(SNO(o) SNO(t))] (4-6) (Ao+YNOSNO(o)) = (4-7) Sno(0)Sno(t) (dSNO/ YNOSNO) = maxNO 0t dt (4-8) ln( YNOSNO(t)) = YNO maxNO t + ln( -YNOSNO(o)) (4-9) Thus, a plot of ln( YNOSNO(t)) vs. YNO*t will have a slope of maxNO. Once maxNO of the test solution is obtained, maxNO of the test solution is compared with maxNO of the synthetic nitrate solution. Thus, the ratio BI, given in Equa tion 4-10, can be considered as a treatability index. A low BI will indicate that biological denitrif ication of that sample will be very difficult. BI=(( maxNO) test sample)/((maxNO) synthetic sample with same NO3-N concentration) (4-10) Development of the Assay Initially the batch tests were conducted in fl asks secured with rubbe r stoppers with a gas inlet port and a gas outlet port. The inlet and outlet ports we re secured with aluminum foil after nitrogen gas was bubbled through the flasks. The first batch test was conducted to test the treatability of the ion-exchange d industrial stream neutralized with potassium hydroxide. This sample previously had poor denitrification when it was fed to the suspended growth reactor. The test was conducted alongside sodium hydroxide ne utralized industrial stream, which had positive results after it was fed through the suspended growth reactor. Af ter 24 hours, it was noticed that all the flasks had high de nitrification and absorbance increase. This led to the suspicion that oxygen had entered the flasks through gas inlet and gas outlet ports, and that the absorbance increase was mainly due to aerobic growth of th e bacteria. When oxygen is present, bacteria can convert the nitrate to ammonia, which is then us ed as the nitrogen source for bacterial growth.

PAGE 55

55 In order to limit the presence of oxygen in the flasks, another batch test was conducted using flasks that were secured tightly with soli d rubber stoppers. Prior to securing the flasks, nitrogen gas was bubbled through the flasks in order to purge out any oxygen. The ionexchanged stream neutralized with potassium hydroxide was tested against artificial nitrate solution prepared using 69.6% HNO3 assay. The results were sim ilar to previous test, and high denitrification and absorbance increase were obser ved in all flasks. This did not correspond to the results obtained from suspended growth reactors. Since oxyge n was purged out using nitrogen gas, the possibility of presence of any oxygen in the flasks was discarded. It was then suspected that vol atile contaminants present in the industrial stream might have caused low denitrification rates in the su spended growth reactor. It was possible that volatile contaminants may have been remove d when nitrogen gas was bubbled through the flasks. Thus, another batch test was carried out using ion-exchanged sample. The flasks were again secured with solid rubber stoppers; however, nitrogen ga s was not bubbled through the reactors. At the end of the batc h test, it was noticed th at little or no denitrification occurred in the flasks, and no bacterial growth was observed. This supported the hypothesis that volatile contaminants were causing the low denitrification rates observed previously in the CSTR, and the high denitrification in the previous batch tests were the result of removal of volatile contaminants through aeration. In order to confirm the presen ce of volatile contaminants in the industrial stream, a batch test was carried out using aerated and non-aerated ion-exchanged samples. In order to avoid assimilatory denitrification, 0.3 g/L ammonium ch loride was added to the flasks. The flasks containing aerated samples showed high denitrifi cation and absorbance increase, while the flasks containing non-aerated samples had low denitrif ication and no absorbance increase. This

PAGE 56

56 confirmed the theory that volatile contaminants were causing the low denitrification in the CSTR. Results and Discussion The assay was applied to test the treatability of aerated and non-aerated industrial stream samples neutralized using sodium hydroxide. In oculum for the batch test was obtained from CSTR4 effluent. Measurements were taken afte r 4, 8, and 24 hours. The analyses from batch test done non ion-exchanged industrial sample are shown in Figures 4-1 through 4-6. Aerated industrial sample y = -8.7571E-04x + 2.0296E+00 R2 = 9.4477E-01 0 0.5 1 1.5 2 2.5 05001000150020002500 SnoAbsorbance Figure 4-1. Absorbance vs. nitrate concentration, aerated sample non aerated industrial sample y = -2.0796E-03x + 4.0101E+00 R2 = 7.8896E-01 0 0.2 0.4 0.6 0.8 1 150015501600165017001750180018501900 Snoabsorbance Figure 4-2. Absorbance vs. nitrate concentration, non-aerated sample

PAGE 57

57 synthetic nitrate solution y = -9.1872E-04x + 2.2502E+00 R2 = 9.9483E-01 0 0.5 1 1.5 2 2.5 0 500 1000 1500 2000 2500 SnoAbsorbance Figure 4-3. Absorbance vs. nitrate concentration, synthetic nitrate solution Aerated industrial sample y = 2.0728x 0.8511 R2 = 0.8101 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0 0.5 Yno*tln(beta-YnoSno(t)) Figure 4-4. ln( -YNOSno(o)) vs. YNO*time, aerated sample y = 5.6921E-01x 1.0997E+00 R2 = 6.9921E-01 -1.3 -1.2 -1.1 -1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0 0.2 0.4 0.6 0.8 11 2 Figure 4-5. ln( -YNOSno(o)) vs. YNO*time, non-aerated sample

PAGE 58

58 Synthetic nitrate solutiony = 3.0266x 1.1197 R2 = 0.961 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 00.10.20.30.40.5 Yno*tln(beta-YnoSno(t)) Figure 4-6. ln( -YNOSno(o)) vs. YNO*time, synthetic nitrate solution Figures 4-1, 4-2, and 4-3 show the plots of absorbance vs. nitrate concentration for aerated high nitrate i ndustrial stream, non-aerated high nitr ate industrial stream, and synthetic nitrate solution respectively. From the plots, the YNO was obtained to be 8.76E-4 for aerated stream, 2.08E-3 for non-aerated stre am, and 9.19E-4 for synthetic nitrate solution. The plots in Figures 4-4, 4-5, and 4-6 were used to obtain maxNO values for aerated sample, non-aerated sample, and synthetic nitrate solution respectively. maxNO values were obtai ned to be 2.073 for aerated sample, 0.569 for non aerated sample, and 3.027 for synthetic nitrate. From maxNO values, the biotreatability i ndex, BI was calculated. BI for aerated sample was 0.635 and for non-aerated sample was 0.188. The assay showed that the aerated industrial stream can be treated biologically. The results were verified by using aerated industrial stream as the feed for a suspended growth reactor. High denitrification rates and absorban ce values were observed. The assay was done on other industrial samples as we ll. The results obtained from the industrial samples are given in Figures 4-7 through 4-12. The BI values are summarized in Table 4-1.

PAGE 59

59 Sample A YNO y = -1.1341E-03x + 1.7336E+00 R2 = 9.5532E-01 0 0.5 1 1.5 2 0200400600800100012001400 Snoabsorbance Figure 4-7. Absorbance vs. SNO for sample A Sample A y = 0.868x 0.7094 R2 = 0.9042 -1.5 -1 -0.5 0 0.5 1 0 12 Yno*tln(beta-Yno*Sno(t)) Figure 4-8. ln( -YNOSNO(0)) vs. YNO t for sample A Aerated Sample C YNO y = -1.4015E-05x + 5.3903E-01 R2 = 1.8517E-03 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 10001100120013001400150016001700180019002000 SnoAbsorbance Figure 4-9. Absorbance vs. SNO for aerated sample C

PAGE 60

60 aerated Sample C y = 1.1444x 0.5501 R2 = 0.3737 -0.8 -0.6 -0.4 -0.2 0 0.2 0 0.5 Yno*tln(beta-YnoSno(t)) Figure 4-10. ln( -YNOSNO(0)) vs. YNO t for aerated sample C aerated sample D, YNO y = -1.4607E-03x + 6.7163E-01 R2 = 8.7487E-01 0 0.1 0.2 0.3 0.4 0.5 0.6 0 50 100 150 200 SnoAbsorbance Figure 4-11. Absorbance vs. SNO for aerated sample D Sample D, aerated y = 5.0297E-01x 1.0017E+00 R2 = 8.9919E-01 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.00E+00 4.00E-01 8.00E-01 Figure 4-12. ln( -YNOSNO(0)) vs. YNO t for aerated sample D

PAGE 61

61 Table 4-1. Biotreatab ility assay results Sample A maxNO = 0.868 Sample C aerated BI = 0.49 Sample C non-aerated Invalid BI value; bacteria did not grow Sample D aerated BI = 0.43 Sample D non-aerated Invalid BI value; bacteria did not grow As shown in Table 4-1, aerated sample C and sample D are treata ble biologically. The non-aerated sample C and non-aera ted sample D had negative YNO values, and thus gave invalid BI values. The absorbance for those samples d ecreased with nitrate consumption during batch test. Conclusions The objective of the assay is to quantitati vely analyze biological treatability of high nitrate solutions. The biotreatab ility index, BI is used to determine if the sample can be denitrified successfully. BI is obtained from the ratio, ( maxNO) test sample)/((maxNO) synthetic sample with same NO3-N concentration). YNO values can be obtained from the plots of absorbance vs. SNO observed during the batch tests, and maxNO is obtained from the plots of ln( YNOSNO(0)) vs. YNO*t. If the value of the ratio is closer to 1, the sample is treatable. A value much lower than 1 shows that high denitrification will be very difficult to achieve for that sample.

PAGE 62

62 CHAPTER 5 CONCLUSIONS AND FUTURE WORK The objective of the project wa s to establish high denitrificat ion rate for an industrial high nitrate stream using biological tr eatment. At the end of this study, it was concluded that the industrial stream was ind eed treatable. However, volatile contam inants and metals ions present in the stream must be removed prior to biological treatment. Volatile contaminants may be removed through aeration, and metal ions can be removed through an ion exchange process. During the treatment, micronutrients should be added to the reactor directly rather than in the feed, since precipitation can occur in the feed. It was also established that neutralization with sodium did not affect denitrifi cation rates, since the industrial stream sample neutralized with sodium hydroxide and the stream sample neutrali zed with potassium hydroxide had same results. CSTR with sodium hydroxide neutralized synt hetic nitrate feed was able to reach a denitrification rate of 4 mg NO3-N/L/min with a flow rate of 11.5 mL/min and an overall denitrification of 97%. This denitrification rate was twice the target denitrification rate. CSTR with the feed of sodium hydroxide neutralized industrial stream was able to reach the target denitrification rate of 2 mg NO3-N/L/min with a flow rate of 5.5 mL/min and an overall denitrification of 98%. The denitrification was performed usi ng potassium acetate as the exogenous carbon source. Thus, effects of carbon limitation on den itrification were not analyzed. For the next step, an industrial carbon sour ce may be used for the process. Studies using attached growth bioreactors can also be conducted with micronutrients added directly to the reactor instead of the feed. Attached growth bioreactor s can be more feasible since sl udge concentration in the reactor effluent will be very low.

PAGE 63

63 LIST OF REFERENCES Barber W P. and Stuckey D. C. (2000) N itrogen removal in a modified anaerobic baffled reactor (ABR): 2, Denitrification. Water Research 34, 2413-2422. Gasith A. J., Jop K. M., Dickson, K. L., Parker ton, T. F., and Kaczmarek, S.A. (1988). Protocol for the identification of toxic fractions in industrial wastewater effluents. Aquatic Toxicology and Hazard Assessment : 10th Volume, ASTM STP 971. Adams, W.J., Chapman, G.A., and Landis, W.G. Editors., American Society for Te sting and Materials, Ph iladelphia. pp. 204-215. Kanow, P.E., 1993. Denitrification Methods United States Patent 5211847. http://www.freepatentsonline.com/5211847.html March 25, 2008. Lee S., Maken S., Jang J., Park K., and Park J. (2006). Developm ent of physicochemical nitrogen removal process for high st rength industrial wastewater. Water Research 40 975-980 Leslie Grady C.P., Daigger, G.T., and Lim, H.C. 1999. Biological Wastewater Treatment 2nd ed. CRC Press, Boca Raton, USA. Tchobanoglous, G., Burton F.L., and Stensel, D.H. 2003. Wastewater Engineering: Treatment and Reuse McGraw-Hill, NY, USA. US EPA., 2005. Section 62-302.530, Criteria for Surface Water Quality Classifications (Florida) http://www.epa.gov/waterscience/stand ards/wqslibrary/fl/fl_4_62-302t.pdf March 30, 2008 Vishniac S., and Santer M. (1957). The Thiobacilli. Bacter iol Rev 21, 195-213 Wasser I.M., deVries S., Moenne-Loccoz P., Schroder, I., and Kavlin K. D. (2002). Nitric oxide in biological denitrification: Fe/Cu metall oenzyme and metal complex NOx redox chemistry. Chemical Reviews (Review) 102, 1201-1234

PAGE 64

64 BIOGRAPHICAL SKETCH Sherin Peter was born on February 23, 1984, in Muvattupuzha, India. S he received her Bachelor of Science degree in chem ical engineering from University of South Florida, Tampa, in 2005. She obtained her MS degree in chemical e ngineering from Univers ity of Florida in 2008.