WATER POLLUTION CONTROL RESEARCH SERIES
17O2ODUEO9/7O
LIGHT • CATALYZED CHLORINE OXIDATION
FOR
TREATMENT OF WASTEWATER
ENVIRONMENTAL PROTECTION AGENCY • WATER QUALITY OFFICE
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes
the results and progress in the control and abatement
of pollution in our Nation's waters. They provide a
central source of information on the research, develop-
ment, and demonstration activities in the Water Quality
Office, Environmental Protection Agency, through inhouse
research and grants and contracts with Federal, State,
and local agencies, research institutions, and industrial
organizations.
Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Head, Project Reports
System, Office of Research and Development, Water Quality
Office, Environmental Protection Agency, Room 1108,
Washington, D.C. 20242.
-------
LIGHT-CATALYZED CHLORINE OXIDATION FOR TREATMENT OF WASTEWATER
by
Midwest Research Institute
Kansas City, Missouri 64110
for the
WATER QUALITY OFFICE
ENVIRONMENTAL PROTECTION AGENCY
Project #17020 DUE
Contract #14-12-531
September 1970
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, B.C. 20402 - Price $1.00
Stock Number 5501-0076
-------
EPA Review Notice
This report has been reviewed by the Water
Quality Office, EPA, and approved for publication.
Approval does not signify that the contents
necessarily reflect the views and policies of
the Environmental Protection Agency, nor does
mention of trade names or commercial products
constitute endorsement or recommendation for
use.
ii
-------
ABSTRACT
The purpose of this research effort was to determine the feasi-
bility of applying the light-catalyzed chlorine oxidation process to the
treatment of effluents from secondary waste-treatment plants.
(a) Studies were made of wavelength effects and intensity-time
effects for the purpose of selecting the most practical source of radiant
energy for the process.
(b) A batch-recycle reactor was constructed to establish cer-
tain plant-design parameters for low-pressure mercury arcs under conditions
which simulated large-scale operations.
(c) Cost estimates were made for the application of this process
to a plant producing 10 million gallons of effluents per day.
The wavelength-effect studies, mentioned in (a), indicated that
short-wavelength radiation (below 300 imj,) is more effective than long-wave-
length radiation in promoting the chlorine oxidation process. Analysis of
the wavelength-effect experiments indicated that mercury-arc lamps are the
most practical sources of radiant energy for this process. However, the
ideal mercury-arc lamp is probably not commercially available. Of presently
available sources, low-pressure mercury arcs are probably the most practical.
The intensity-time studie s, mentioned in (a), indicated that the
rate of organic oxidation, achieved in a given irradiation time, is in-
creased by increased radiation intensity; however, for a specified amount
of absorbed radiant energy, lower intensities produce more overall organic
oxidation than do higher intensities. These studies also established that
the chlorine consumption is directly proportional to the amount of radiant
energy absorbed, regardless of intensity.
Experiments with the batch-recycle reactor, mentioned in (b),
provided data for the precise determination of the quantum efficiency which
is defined as the amount of organic oxidation obtained from a given amount
of absorbed radiant energy. Higher quantum efficiencies were observed at
low intensities, and the quantum efficiency increased as the oxidation
proceeded.
Increases in chlorine concentration above 5 mg/liter produced
no significant increase in the oxidation rate; increased chlorine concen-
trations, however, resulted in inefficient utilization of chlorine and
radiant energy.
iii
-------
Thus, for optimum organic oxidation rate, optimum utilization
of radiant energy and most efficient use of chlorine, (l) low-intensity
short-wavelength radiation should be employed, and (2) the chlorine con-
centration should be at a minimum. The optimum chlorine concentration has
not been established, but it is probably below 5 mg/liter.
In the presence of relatively large amounts of chlorine, oxygen
has no significant effect on either the rate or the extent of oxidation.
Under the experimental conditions investigated, it was not possible to de-
tect whether or not any oxygen had participated in the organic oxidation.
A large-scale tertiary-treatment plant was designed to utilize
commercially available, low-pressure, mercury-arc lamps. The reactor design
allowed for long contact time (40 min) and low-intensity radiation. In the
design, the lamps were spaced in the reactor in such a manner that practi-
cally all of the available radiation would be absorbed. Also, the spacing
of the lamps allowed for the decrease in absorption as the reaction pro-
ceeded.
Process-cost estimates, mentioned in (c), were based on a 10-
million gal./day plant representing an investment of $990,000. Fixed charges
based on this investment were estimated at 3.42^/1,000 gal. Operating costs
of 8.35^/1,000 gal. were estimated, of which the major costs were: labor,
2.79^/1,000 gal.; lamp replacement, 1.70^/1,000 gal.; and chlorine, 2.09^/
1,000 gal. Major assumptions were based on experimental results and were
as follows: (l) a required COD decrease of 15 mg/liter, (2) a quantum effi-
ciency of 1.30, (3) a stoichiometric amount of chlorine, and (4) a continuous
addition of caustic.
The estimated overall costs of 11.77^/1,000 gal. could be. reduced
significantly by: (l) the development of more intense and/or more efficient
mercury-arc lamps, (2) the selection of operating conditions which would
permit higher quantum efficiencies, or (3) the requirement of lesser amounts
of chlorine or caustic.
This report was submitted in fulfillment of Project Number
17020DUE, Contract 14-12-531, under the sponsorship of the Water Quality
Office, Environmental Protection Agency.
IV
-------
CONTENTS
Page
Recommendations ix
Introduction 1
Wavelength-Effect Studies 2
Selection of Experimental Techniques 2
Monochromatic Wavelength-Effect Studies 4
Polychromatic Wavelength-Effect Studies 4
Selection of the Most Practical Ultraviolet Source 6
Intensity-Time Studies 7
Intensity-Time Studies Using the Cell Reactor 7
In tensity-Time Studies Using the Batch-Recycle Reactor 9
Batch-Recycle Reactor Studies 14
Investigation of High Intensity and High Flowrate 17
Investigation of High Intensity and Low Flowrate 18
Investigation of Low Intensity and High Flowrate 18
Preliminary Investigation of Minimum Chlorine Requirement .... 19
Oxygen-Effect Studies 20
Estimated Process Costs 22
Radiant Energy Source 22
Effectiveness of the Applied Radiation 24
Plant Description 24
Plant Investment 26
Treatment Cost 27
Experimental 30
Actinometer Procedures and Calculations 30
Calibration of Monochromatic Wavelength Sources 30
Monochromatic Wavelength-Effect Experiments 34
Polychromatic Wavelength-Effect Experiments 48
Intensity-Time Experiments 50
Oxygen-Effect Experiments 58,
Batch-Recycle Reactor Experiments 61
Acknowledgment 81
References 82
-------
CONTENTS (Concluded)
Appendix A - Problems Regarding Source Selection for Large-Scale,
Light-Catalyzed, Chlorine Oxidation 83
Appendix B - Selection of Techniques and Equipment for Wavelength-
Effect Studies 98
FIGURES
1 Spectral Energy Distribution of High-Pressure Mercury Arc,
Showing Regions (A-G) Which Could Be Isolated Using
Interference Filters 5
2 The Effect of Radiation Intensity on TOG Decrease
(Cell Reactor) 8
3 The Effect of Radiant Energy Absorbed on the Extent of
Chlorine Consumption 10
4 The Effect of Radiation Intensity on COD Decrease
(Batch-Recycle Reactor) 11
5 The Effect of Radiation Intensity on TOG Decrease
(Batch-Recycle Reactor) 12
6 Diagram of Batch-Recycle Reactor 15
7 Process Flowsheet for Tertiary Treatment by Light-
Catalyzed Chlorine, Capacity 10-Million Gal./Day 25
8 Apparatus Used in Wavelength-Effect Studies 35
9 Batch-Recycle Reactor Results - High Intensity - High Flow
Rate (Effluent MRI 112, Run No. 66) 68
10 Batch-Recycle Reactor Results - High Intensity - High Flow
Rate (Effluent MRI 112, Run No. 68) 69
11 Batch-Recycle Reactor Results High Intensity - High Flow
Rate (Effluent MRI 304, Run No. 74) 70
12 Batch-Recycle Reactor Results - High Intensity - Low Flow
Rate (Effluent MRI 304, Run No. 78) 71
13 Batch-Recycle Reactor Results Low Intensity - High Flow
Rate (Effluent MRI 304, Run No. 82) 72
14 Batch-Recycle Reactor Results - High Intensity - High Flow
Rate (Effluent MRI 310, Run No. 86) 73
15 Batch-Recycle Reactor Results - Low Intensity - High Flow
Rate (Effluent MRI 310, Run No. 93) 74
16 Batch-Recycle Reactor Results - High Intensity - Low Flow
Rate (Effluent MRI 310, Run No. 101) 75
VI
-------
FIGURES (Concluded)
Figure
17 Batch-Recycle Reactor Results - Low Intensity - High Flow
Rate (Effluent MRI 319, Run Wo. 104) 76
18 Batch-Recycle Reactor Results - High Intensity - High Flow
Rate (Effluent MRI 319, Decreased Volume,. Run No. 108). . 77
19 Batch-Recycle Reactor Results - High Intensity - High Flow
Rate (Effluent MRI 319, Run Wo. 112) 78
20 Batch-Recycle Reactor Results - High Intensity - High Flow
Rate (Effluent MRI 319, Diluted, Run Wo. 117) 79
A-l Wational Sunshine Carbon, with Corex D Filter, Compared
with Watural Sunlight 91
A-2 Spectral Energy Distribution of Carbon Arc with Iron-Cored
Carbons (National B) 92
A-3 Spectral Energy Distribution of Typical 5-KW Xenon Lamp . . 95
B-l Quantum Yields of Chemical Actinometers 102
TABLES
Table
I Results of Representative Experiments in Wavelength-Effect
Study 3
II Effluent-Oxidation Experiments Using Batch-Recycle Reactor. 16
III Characteristics of Interference Filters 36
IV Monochromatic Wavelength-Effect Studies at 366 m|j, 39
V Actinometric Calibration of Radiation Sources for Mono-
chromatic Wavelength-Effect Studies 40
VI Characteristics of Effluents Used in the Monochromatic
Wavelength-Effect Studies 40
VII Monochromatic Wavelength-Effect Studies at 436 mp, 41
VIII Monochromatic Wavelength-Effect Studies at 313 mjj, 41
IX UV-Absorption Characteristics of Chlorinated Effluents Used
in Monochromatic Wavelength-Effect Studies 42
X Monochromatic Wavelength-Effect Studies at 405 mjj,
and 546 mfj, 43
XI Estimated Intensity of 546-mp, Radiation Transmitted
by Filter 44
XII Monochromatic Wavelength-Effect Studies at 254 m|j, 45
XIII Summary of TOG Decreases at 254 mjj, 45
XIV Radiant Energy at 254 mp, Absorbed by Chlorinated Effluents. 46
XV UV-Absorption Characteristics of Chlorinated Effluents
Used in 254-mp, Wavelength-Effect Studies 46
vii
-------
TABLES (Concluded)
Table
XVI Actinometric Calibration of the Radiation Source for
Monochromatic Wavelength-Effect Studies at 254 imj, . . . 47
XVII Spectral Output of Mercury-Arc Spotlamp (lOO-CH-4, SPOT.
H38-4GS) 47
XVIII Data for Actinometric Calculation of Intensity of
Mercury-Arc Spotlamp (Westinghouse 100-CH-4) 49
XIX Actinometer Data for Mercury-Arc Spotlamp (lOO-CH-4). . . 49
XX Results of Polychromatic Wavelength-Effect Experiments. . 51
XXI Estimation of Absorption of Polychromatic Radiation by
Chlorinated Effluent (Run 47-1, Table XX) 51
XXII Intensity-Time Studies (First Series) 52
XXIII Radiant Energy Absorbed (First Intensity-Time Series) . . 53
XXIV UV-Absorption Characteristics of Chlorinated Effluents
(First Intensity-Time Series) 53
XXV Actinometric Calibration (First Intensity-Time Series). . 54
XXVI A Comparison of Effluents MRI 925 and MRI 112 54
XXVII A Comparison of the UV-Chlorine Oxidation Characteristics
of Effluents MRI 925 and MRI 112 55
XXVIII Intensity-Time Studies (Second Series) 56
XXIX Radiant Energy Absorbed (Second Intensity-Time Series). . 57
XXX UV-Absorption Characteristics of Chlorinated Effluents
(Second Intensity-Time Series) 57
XXXI Intensity-Time Studies (Third Series) 59
XXXII Radiant Energy Absorbed (Third Intensity-Time Series) . . 60
XXXIII UV-Absorption Characteristics of Chlorinated Effluents
(Third Intensity-Time Series) 60
XXXIV Analytical Data for Effluents Used in Oxygen-Effect
Studies 62
XXXV Results of First Series of Oxygen-Effect Experiments,
Effluent (S&L 619) 63
XXXVI Results of Second Series of Oxygen-Effect Experiments,
Effluent (MRI 720) 63
XXXVII Actinometric Calibration of Low-Pressure, Mercury-Arc
Lamp 66
XXXVIII Results of Investigation of Minimum Chlorine Requirement. 67
A-1 Spectral Energy Distribution (in w) of Selected,'
Mercury-Arc Lamps 86
A-II Costs of Producing Ultraviolet Radiation from Various
Mercury-Arc Lamps 39
A-III Intensity of Radiation from Bare Carbon Arcs, Without
Reflectors, in Microwatts per Square Centimeter at
Pi stance of 1 Meter from Arc 93
A-IV Solar Energy Distribution 97
A-V Maximum Intensity of Solar Energy 97
VL11
-------
RECOMMENDATIONS
An examination of process-cost estimations indicates that three
major factors must be evaluated before an accurate determination can be
made of the cost of applying the UV-chlorine process to secondary effluents.
These factors are: (l) the conditions which will give maximum quantum effi-
ciency, (2) the minimum amounts of chlorine and caustic that are required,
and (3) the relative utility of radiant energy sources which produce higher
intensities.
The quantum efficiency is a measure of the amount of radiant
energy required to produce a specified amount of oxidation. Higher quantum
efficiencies would permit significantly lower operating and plant-invest-
ment costs. Experiments to date indicate that quantum efficiencies much
higher than those used in the process-cost estimates may be feasible.
Significant reductions in cost would also be possible if less
chlorine and caustic were required. Preliminary experiments indicate that
such reductions might be feasible.
A further significant reduction in plant-investment cost would
be possible if higher intensity, radiant-energy sources could be used.
Higher intensities permit smaller and less expensive reactors, but higher
intensities also result in lower quantum efficiencies. There is a great
difference between the relative intensities of the two types of commercially
available, mercury-arc lamps. Experimental determination must be made of
the relative advantages of each kind of lamp.
Studies designed to investigate maximum quantum efficiency
should consist of: (l) an examination of lower radiation intensities than
previously studied, and (2) a determination of the minimum chlorine con-
centration required to provide the maximum oxidation rate. These studies
should include a determination of the relative effectiveness of radiation
absorbed both by chlorine and by effluent at all stages of the oxidation.
The recommended study of minimum chlorine and caustic require-
ments should also involve relatively low chlorine concentration (perhaps
1 mg/liter or less). Although the most rapid rate of oxidation and the
most efficient use of chlorine are obtained at pH 5, the.most economical
operation may be at ambient pH values, with little or no addition of caus-
tic. The effect of oxygen on the rate and extent of oxidation under these
conditions should be examined since oxygen may participate in the organic
oxidation.
IX
-------
Finally, actual comparison experiments should be performed, using
high- and low-pressure mercury arcs. The effectiveness of radiation from
these lamps throughout the reaction zone should be determined by laboratory
experiments in which the intensity of these lamps is varied. From these
data, the overall rate and extent of the oxidation could be calculated at
any point in the irradiated zone of the proposed, large-scale reactor.
A proposal (MRI Proposal No. C-2837, Revised) incorporating most
of these studies has been submitted to the Federal Water Quality Administra-
tion.
-------
INTRODUCTION
Earlier studies at Midwest Research Institute demonstrated the
technical feasibility of employing chlorine and ultraviolet radiation for
the elimination of dissolved organic material in highly nitrified efflu-
ents from biological, sewage-treatment plants. ("An Investigation of
Light-Catalyzed Chlorine Oxidation for Treatment of Wastewater, Final Sum-
mary Report," FWPCA Contract No. 14-12-72, December 1968.) Chlorine will
slowly oxidize only a small fraction of this organic material in the dark,
but in the presence of ultraviolet radiation, rapid elimination of large
amounts of COD and TOG is possible.
The purpose of the present work has been to provide additional
information required for large-scale application of the process.
The body of this report consists of five major parts and an
Experimental section: (l) The first major part describes wavelength-effect
studies undertaken to determine the most practical source of radiant energy.
(2) The second major part describes intensity-time studies undertaken to
establish the relationship of these two important variables. (3) the third
major part describes a series of experiments performed using a batch-recycle
reactor. A reactor of this kind was very useful in determining certain
plant-design parameters under conditions which simulated large-scale opera-
tion. (4) The fourth major part describes the effect of dissolved oxygen on
the UV-chlorine oxidation when relatively large concentrations of chlorine
are present. (5) The fifth major part is a review of process cost estimates.
There are two appendices to this report: Appendix A outlines
the problems involved in selecting the most practical ultraviolet source
for large-scale application of this process; Appendix B is a discussion of
the problems encountered in selecting the techniques and equipment that
were used in the wavelength-effect studies.
-------
WAVELENGTH-EFFECT STUDIES
The most important factor in determining the feasibility of the
light-catalyzed chlorine oxidation process for wastevater treatment is the
selection of the most useful source of radiant energy from among the wide
variety of types available. No prior information was available concerning
the kind of radiant energy most effective in producing the desired reaction.
On the other hand, many other chlorine reactions—for example, hydrocarbon
chlorination—are known to be catalyzed by all radiation below a threshold
wavelength of about 480 m^,. Thus, wavelength effects are relatively well
understood for many other chlorine processes. Without preliminary informa-
tion of this kind, the selection of the most practical radiant-energy source
is extremely complicated. The problem of source selection for light-catalyzed
chlorine oxidation is discussed in more detail in Appendix A, p. 83.
After consideration of the problems involved in selecting the
most practical radiation source, a study was undertaken to determine which
wavelengths were the most effective in promoting the chlorine oxidation of
wastewater.
In order to determine the effect of wavelength, it was necessary:
(l) to devise a system for providing monochromatic (or nearly monochromatic)
radiation at several different wavelengths, and (2) to observe the extent
of organic oxidation produced by a given amount of absorbed radiation at
each of those wavelengths. Not only was it necessary to establish the
relative effectiveness of each ultraviolet wavelength (or group of wavelengths)
between 220 and 400 mjj,, but it was also important to establish the contribu-
tion, if any, of visible radiation (400-700 rmj,). The effect of visible radi-
ation was of particular interest because there is much unavoidable visible
energy available from most commercially available sources (see Table A-I,
Appendix A, p. 86).
Selection of Experimental Techniques
A number of alternatives were considered in selection of the
monochromatic radiation source, design of the reaction vessel, and method
of determining the amount of radiation absorbed. These alternatives are
discussed in Appendix B, p. 99.
A major factor in the ultimate decision concerning experimental
techniques to be employed in the wavelength-effect studies was the very
recent availability of interference filters for the ultraviolet region. A
series of interference filters, which permitted very effective isolation
of the ^iajor wavelengths available in a high-pressure mercury arc, was
obtained.
-------
TABLE I
RESULTS OF REPRESENTATIVE EXPERIMENTS
IK WAVELENGTH-EFFECT STUDY
Chlorine
0-1
Wavelength
Uu)
546
436
436
405
405
366
366
313
313
254
254
254
Irradiation
Time
(min)
120
30
120
60
120
60
120
60
130
20
40
60
Intensity
(raw)
1
3
3
2
2
2
2
1
1
1
1
1
.20
.61
.61
.20
.20
.70
.70
.04
.04
.15
.15
.15
Radiant Energy
(w rain/gal)
Applied
180
136
544
167
333
195
390
79
171
29
58
90
Absorbed
1.8
1.4
5.4
7.7
15.3
12.0
23.0
13.0
29.0
12.0
23.0
32.0
Concentration
(mg/ „ )
Initial
95
140
109
84
109
89
84
99
118
92
90
68
Final
90
135
108
80
91
79
71
67
43
53
26
4
Difference
5
5
1
4
18
10
13
32
75
39
64
64
TOC (mg/4 )
Initial
8.2
8.1
9.9
8.9
9.9
8.9
8.7
8.4
8.4
8.9
8.5
8.8
Final Difference
8.8
8.0
10.0
9.3
9.4
8.7
9.2
7.4
6.5
8.9
5.9
5.1
0.0*
0.1
0.0*
0.0*
0.5
0.2
0.0*
1.0
2.1
0.0
2.6
3.7
In these experiments, a small increase in TOC was observed. These anomalous results are believed to have been
caused by a lack of precision in determining small differences in TOC.
-------
The spectral energy distribution of a high-pressure mercury arc
is shown in Figure 1, and the regions which could be isolated "by the avail-
able interference filters are indicated.
Monochromatic Wavelength-Effect Studies
A summary of representative experiments in the wavelength-effect
studies is presented in Table I. Additional experiments and details are
presented in the Experimental section, p. 34.
Analysis of these results indicates that short-wavelength radia-
tion (below 300 mp,) is the most effective in producing decreases both in
TOG and in chlorine concentration. A disadvantage of long-wavelength radia-
tion (436-546 m|j,) is that it is not strongly absorbed by secondary effluents.
The inability to deliver energy-per-unit-volume is a significant deterrent
to the application of relatively long-wavelength radiation to this process.
Because of the relatively low energy of long-wavelength radiation, and be-
cause this energy is not strongly absorbed by the effluent-chlorine solution,
higher intensities of incident radiation are required in order to produce a
given amount of absorbed energy. Therefore, in order to be practical, long-
wave-length radiation would have to be relatively more efficient than short-
wavelength radiation in producing oxidation per watt of energy absorbed.
Our results, however, indicate that this is not the case.
Polychromatic Wavelength-Effect Studies
The conclusion that lower wavelengths were most effective was
confirmed by polychromatic irradiation studies in which a radiation source
that did not emit radiation below 300 mji was used. The results of these
experiments are presented in the Experimental section, p. 48.
In the polychromatic irradiation studies, the incident intensities
were much greater than in the monochromatic irradiation studies. However,
only a small fraction of this radiation was absorbed by the chlorinated
effluent. The major portion of the absorbed energy was in the wavelength
range 300-370 mp,.
An absorbed energy-per-unit-volume of 140 w min/gal. produced a
TOC decrease of 2.4 mg/liter (see Tables XX and XXI, p. 5l). However, as
shown in Table I, radiation at 254 m|j, will produce a TOC decrease of 2.6
mg/liter with an absorbed energy-per-unit-volume of only 23 w min/gal.
Thus, the 254-rmj, radiation is about six times more effective than polychro-
matic radiation between 300-370 mp,.
-------
D
CJI
26
24
22
20
18
0 14
LLJ
LLJ 12
0
LLJ
I '°
8
6
4
2
-
_
-
-
A
-
~ ill
i
00 250
|
C
B
l
300
t—
i
—— 1
I- G -•
p
E
i
i i
i
i
350 400 450 500 550 600
Wavelength (mil)
Figure 1 - Spectral Energy Distribution of High-Pressure Mercury Arc, Showing
Regions (A-G) Which Could Be Isolated Using Interference Filters
-------
Selection of the Most Practical Ultraviolet Source
A review of available, radiant-energy sources (Appendix A) indi-
cates that mercury arcs are the most practical of the presently available
sources of ultraviolet radiation below 300 mJJ.. However, as discussed in a
subsequent section (Estimated Process Costs, p. 22), the ideal mercury arc
iay not be commercially available at the present time, and will probably
require special manufacture to particular design specifications. Of the
ultraviolet sources commercially available at the present time, low-pressure
mercury arcs are probably the most practical, and a 10-mgd tertiary treatment
plant has been designed to utilize these lamps (p. 24).
-------
INTENSITY-TIME STUDIES
The relationship of radiant-energy intensity to irradiation time
is a critical factor in pilot-plant design. Because of the nature of this
process, the intensity of radiation will vary greatly throughout the reaction
zone. The intensity will "be greatest near the source, but will diminish
rapidly as the distance from the source increases. The intensity at any
given point is dependent upon the intensity of the source, the distance from
the source and the optical properties (both absorptive and dispersive) of
the reaction mixture. In order to calculate the rate of reaction at any
point or throughout the reaction zone, it is necessary to establish the
reaction rate which would be produced by a given amount of radiant energy.
However, since the amount of radiant energy is the product of intensity
multiplied by time, the relationship between these factors must be established.
For example, it was necessary to determine whether 1 w of radiant energy
applied for 10 min, would produce the same results as 10 w for 1 min.
Knowledge of the intensity-time relationship was also required in
order to select the most effective source of radiant energy. For example,
it must be determined whether one high-intensity source would be more effec-
tive than several low-intensity sources. Furthermore, since intensity de-
creases with distance from the source, intensity-time studies were necessary
to establish the optimum depth of effluent which can be effectively treated.
The intensity-time studies were concerned primarily with radiation
from low-pressure mercury arcs. These lamps are economical sources of short-
wavelength radiation (254 nip,) which was known, from the wavelength-effect
studies, to be highly effective in promoting the oxidation.
Intensity-Time Studies Using the Cell Reactor
Initial experiments were performed with the same equipment and
procedures used in the wavelength-effect experiments. Detailed results
are presented in the Experimental section, p. 50.
The results of the intensity-time studies performed, using the
cell reactor, can be summarized as follows:
1. As indicated in Figure 2, there was very little effect of
intensity on TOG decrease up to an absorbed energy-per-unit-volume of 30 w
min/gal. At higher values of absorbed energy, lower intensities were more
effective in reducing TOC.
-------
6.0 r
CD
10
o
NTENSITY
3 mw
20 30 40 50 60
RADIANT ENERGY ABSORBED (wmin/gal.)
70
80
90
Figure 2 - The Effect of Radiation Intensity on TOG Decrease (Cell Reactor)
-------
2. The decrease in chlorine concentration was dependent only upon
total energy absorbed, regardless of intensity. Figure 3 is a plot of the
decrease in chlorine concentration versus energy absorbed. The data were
obtained from a series of experiments in which the irradiation time was
varied from 7.5 to 101 min and the applied intensity was 1 to 3 mw. The
plot in Figure 3 shows a pronounced change in slope after the absorption of
radiant energy equivalent to about 50 w min/gal. At higher values of absorbed
energy, the chlorine concentration was less than 10-15 mg/liter. Thus, at
low chlorine concentrations the absorbed radiant energy is very much less
effective in reducing chlorine concentration.
The extent of organic oxidation was measured in terms of the
decrease in absorbance. The absorbance due to chlorine was subtracted from
the total absorbance to obtain the absorbance due to the effluent. However,
no relationship could be established between the rate of decrease in effluent
absorbance and the applied intensity. Also, no relationship could be estab-
lished between the effluent absorbance and the total amount of radiant energy
absorbed.
Intensity-Time Studies Using the Batch-Recycle Reactor
Intensity-time data obtained, using the batch-recycle reactor, pro-
vided important information concerning the relationship of these factors (a
complete discussion of the batch-recycle reactor studies is presented in the
next section of this report). The most significant difference between the
cell reactor and the batch-recycle reactor was that in the latter the con-
centration of chlorine was relatively low: 5-10 mg/liter compared to 50-100
mg/liter in the cell reactor. Also, since relatively large amounts of
effluent were treated in the batch-recycle reactor, COD determinations could
be obtained.
The results of two, representative, batch-recycle reactor experi-
ments are presented in Figures 4 and 5. (Additional details are presented
in the Experimental section, p. 61). The results were similar to those ob-
tained using the cell reactor, in that low intensity was more effective than
high intensity. That is, a given amount of absorbed energy would produce a
greater amount of TOG and COD decrease if the energy was applied at low
intensity (and, obviously, for a longer time). However, there were three
important differences between the cell-reactor results and the batch-recycle
reactor results:
1. The differences between low and high intensity were much more
pronounced in the batch-recycle reactor experiments; low intensities were
very much more effective than high intensities.
-------
10
20
30
40 50 60 70 80
RADIANT ENERGY ABSORBED (wmin/gal.)
Figure 3 - The Effect of Radiant Energy Absorbed on the Extent of
Chlorine Consumption
-------
12 -
O
LOW INTENSITY (0.53 w)
(RUN NO. 93)
HIGH INTENSITY (1.78w)
(RUN NO. 86)
6 8 10 12 14
RADIANT ENERGY ABSORBED (w min/gal.)
16
18
20
22
Figure 4 - The Effect of Radiation Intensity on COD Decrease (Batch-Recycle Reactor)
-------
o>
E
K i UJ T
h' oi o
U
LLJ
Q
U
O
2 -
0
LOW INTENSITY (0.53 w)
(RUN NO. 93)
HIGH INTENSITY (1 .78 w)
(RUN NO. 86)
8 10 12 14 16
RADIANT ENERGY ABSORBED (wmin/gal.)
18
20
22
24
Figure 5 - The Effect of Radiation Intensity on TOG Decrease (Batch-Recycle Reactor)
-------
2. The differences between low and high intensity were apparent
even at low values for the absorbed energy.
3. As the oxidation proceeded, the amount of TOG decrease and
COD decrease produced per unit amount of absorbed energy increased markedly
in the batch-recycle experiments, but actually decreased in the cell-reactor
experiments. Thus, the effectiveness of the absorbed.radiation increased
with time in the batch-recycle experiments, but not in the cell-reactor
experiments.
These important differences are evident from a consideration of
the shapes of the curves presented in Figures 2, 4 and 5.
Apparently, the reason for these differences is the large differ-
ence in chlorine concentration used.
The original work on UV-chlorine oxidation had established that
high concentrations of chlorine were wasteful of chlorine. These new ex-
periments have established that high concentrations of chlorine are also
wasteful of radiant energy.
13
-------
BATCH-RECYCLE REACTOR STUDIES
A batch-recycle reactor was constructed in order to evaluate
certain design factors related to UV-chlorine oxidation. A diagram of this
reactor is presented in Figure 6, and a detailed description of the reac-
tor and operating procedures is presented in the Experimental section, p.
61. A low-pressure, mercury-arc lamp was used in this reactor because the
wavelength-effect studies had indicated that this kind of lamp was the most
practical of presently available, radiant-energy sources.
A reactor of this kind is especially suitable to this process
for the following reasons: (l) a batch-recycle reactor is ideal for rela-
tively slow, photochemical reactions such as this; (2) relatively large
quantities of effluent can be processed; thus, analytical determinations
which require relatively large samples, such as COD, are feasible; (3) an
investigation of relatively low chlorine concentrations can be performed;
(4) the pH can be controlled by alkali addition instead of by employing a
relatively high concentration of buffer; (5) the system is dynamic rather
than static, and the problems of unmixed reactors are avoided; (6) the sys-
tem is similar to the ultimate system which will be applied to wastewater
treatment; and (?) some critical design parameters can be established, such
as optimum effluent depth, the optimum intensity of the radiant energy
source, and the amount of organic oxidation which can be obtained from a
given amount of absorbed energy.
A series of effluent-oxidation experiments were performed, and
the results are summarized in Table II.
Additional details are presented in the Experimental section,
pp. 68-79.
In each of the experiments, TOG and COD decreased at an almost
constant (linear) rate. The rate of COD decrease was slightly greater than
the rate of TOC decrease. These results are not greatly different from
those obtained from batch reactions run during the original research on
UV-chlorine oxidation (Contract No. 14-12-72).
The intensity of the radiation impinging on the chlorinated
effluent was carefully measured by actinometric techniques. (See Experi-
mental section, p. 65).
14
-------
CJ1
LOW-PRESSURE
UV LAMP
QUARTZ
IMMERSION WELL
MAGNETIC
STIRRER
ULTRAVIOLET METER -
Figure 6 - Diagram of Batch-Recycle
-------
Run
78
101
82
93
104
TABLE II
EFFIUENT-OXIDATION EXPERIMENTS USING
BATCH-RECYCLE REACTOR
COD Decrease^/ TOG Decrease^/
Effluent (mg/i/min) (mg/i/min)
(High Intensity, High Flow Rate)
66
68
74
86
112
loe^/
117l/
112
112
304
310
319
319
319
1.31
0.94
1.46
1.04
1.30
1.08
0.81
0.38
0.41
0.56
0.35
0.40
0.51
0.33
304
310
304
310
319
(High Intensity, Low Flow Rate)
0.93 0.39
1.11 0.28
(Low Intensity, High Flow Rate)
0.83 0.60
0.69 0.31
0.70 0.25
Quantum-'
Efficiency
Initial Maximum
0.42
0.21
0.47
0.25
0.62
0.61
0.35
0.20
0.43
0.58
0.70
0.33
1.36
1.04
0.55
0.89
1.33
1.95
1.16
0.61
0.66
2.36
2.48
1.47
a/ The rate of decrease in COD and TOG was calculated as mg/^/min of UV
contact time. Contact time is the contact time per cycle, multiplied
by the number of cycles.
b/ The quantum efficiency is the decrease in COD produced by absorbed UV
energy and is expressed in terms of molecules of oxygen utilized per
absorbed quantum. The procedure for calculating quantum efficiency
is presented in the Experimental section, p. 96.
£/ In Run No. 108, the reservoir contained only half the customary amount
of effluent.
df In Run No. 117, the effluent was diluted with an equal volume of distilled
water.
16
-------
The fraction of radiation absorbed was also accurately determined
at various times during the reaction (this procedure is also described in
the Experimental section, p. 64). The fraction of radiation absorbed also
decreased as the oxidation proceeded, but not at a constant rate; the rate
of decrease was greater during the first part of the oxidation than it was
during the last part.
The rate and extent of decrease in the fraction of radiation
absorbed were of particular interest because this phenomenon has important
implications in regard to reactor design and quantum efficiency. The EDP
(effective depth of penetration; the depth at which 90$ of the radiation is
absorbed) can be calculated from the absorbance data (cell length in centi-
meters, divided by absorbance). At the beginning of a typical experiment,
the EDP is about 9 cm and increases to 40-50 cm at the end of the experi-
ment. A commercially feasible reactor must obtain maximum use of emitted
radiation. In order to accomplish this, some provision for the change in
EDP must be made. Either the effluent depth must be greater toward the end
of the treatment, or some method of utilizing the most efficient average
radius must be derived. (Thorough mixing of the reactor, or recycling of
treated effluent, might permit effective use of a constant-depth reactor.)
As a result of both the constant rate of COD decrease and the
decrease in radiation absorbed, the effectiveness of the absorbed radiation
increased as the oxidation proceeded.
Experiments with the batch-recycle reactor permitted very accu-
rate determination of the effectiveness of the absorbed radiation in pro-
ducing a given amount of organic oxidation. By definition, the term
"quantum efficiency," (E), is a measure of the effectiveness of the ab-
sorbed radiation in which the extent of organic oxidation is expressed as
the number of molecules of oxygen which have been utilized, and the ab-
sorbed energy is expressed in quanta. Thus, "quantum efficiency" is anal-
ogous to the term "quantum yield" which is generally used in photochemistry.
The procedure followed in calculating quantum efficiency is presented in
the Experimental section, p. 80.
In each experiment, the quantum efficiency increased as the
oxidation proceeded, although there were very wide differences in the max-
imum quantum efficiencies which could be attained.
Investigation of High Intensity and High Flowrate
The results obtained with maximum lamp intensity (1.78 w) and
high flowrate (665 ml/min) are presented in the first part of Table II.
There was considerable variation in the rates of COD and TOG decrease; the
17
-------
rate of COD decrease ranged from 0.94 to 1.31 mg/liter/min, and the rate
of TOC decrease ranged from 0.35 to 0.55 mg/liter/min.
These variations apparently cannot be correlated with the initial
absorbance of the effluent; in some experiments a higher initial absorbance
was accompanied by a faster rate of COD decrease, and in other experiments
the reverse was true. In Run No. 117, the effluent (MRI 319) was diluted
with an equal volume of distilled water. The decrease in absorbance caused
by the dilution did not produce a proportional decrease in oxidation rate.
In Run No. 108, only one-half of the customary amount of effluent
was placed in the reservoir. This procedure effectively doubled the total
contact time in the reactor for the same flowrate and elapsed time. Also
with half as much effluent in the reservoir, the amount of "dark time" was
decreased by about one-half; that is, for a specified amount of irradiation
time, the reaction mixture spends only half as much time in the dark when
the reservoir is only half filled. If a significant "dark reaction" oc-
curred, then higher oxidation rates would be expected if the reservoir were
completely filled. A comparison of Run No. 112 with Run No. 108 (Table II)
shows that the rate of COD decrease was greater with an increased amount of
dark time. This indicates that some decrease in COD was occurring in the
dark. However, the rates of TOC decrease show the opposite effect; i.e.,
the rate of TOC decrease was considerably greater when the dark time was
minimized. In both experiments, the quantum efficiency increased at a rate
proportional to the extent of COD decrease.
Investigation of High Intensity and Low Flowrate
The results of experiments performed at flowrates less than half
those employed in the high-flowrate experiments are shown in the second
part of Table II. In the low-flowrate experiments the total contact time
(irradiation time) is the same as in the high-flowrate experiments. The
reason is that at the low flowrate a given portion of the reaction mixture
passes more slowly through the reactor, but it passes through a fewer number
of times. Within experimental error, flowrate (as expected) does not appear
to affect the rates of TOC or COD removal, nor does flowrate affect the
quantum efficiency.
Investigation of Low Intensity and High Flowrate
The results of experiments performed at low intensities (0.53 w
compared with 1.73 w) are presented in the last part of Table II. A com-
parison of rates of COD and TOC decrease shows that the large decrease in
intensity did not greatly affect the organic oxidation rate, (in fact, the
rate of TOC decrease in Run No. 82 was the fastest observed.)
18
-------
Obviously, the low-intensity radiation was more efficient, and
this fact was indicated by the quantum efficiency calculations. At all
levels of absorbed radiation, the low-intensity radiation had about twice
the quantum efficiency as the high-intensity radiation.
Preliminary Investigation of Minimum Chlorine Requirement
Toward the end of the experimental period, a preliminary investi-
gation was made of the possibility of using less-than-stoichiometric amounts
of chlorine. In these experiments, chlorine was added to bring the initial
chlorine concentration to about 5 rag/liter, but the chlorine concentration
was allowed to decrease to zero before additional chlorine was added. The
results of these experiments are described in the Experimental section,
p. 67. Less-than-stoichiometric amounts of chlorine were required under
these conditions, but some loss in overall, organic, reaction rate was ap-
parent. Possibly an optimum chlorine concentration below 5 rag/liter could
be established at which maximum oxidation rates could be maintained with
minimum amounts of chlorine consumed.
The effect of oxygen under these conditions has not yet been
established. However, the preliminary experiments have indicated that a
surplus of oxygen does not increase the reaction rate, nor does it affect
the amount of chlorine consumed. The role of oxygen in the UV-chlorine
process is discussed in the next section of this report.
19
-------
OXYGEN-EFFECT STUDIES
The results of numerous photochlorination experiments published
in the literature suggest that dissolved oxygen might have a profound in-
fluence on the catalyzed, chlorine oxidation. For example, traces of oxygen
produce a strong inhibiting influence on organic photochlorinations. Further-
more, there are some well-known photocatalyzed reactions of chlorine with
oxygen.
Since the photochemical chlorine oxidation takes place in an
effluent that has been subjected to extended aeration and is thus nearly
saturated with air, it is possible that oxygen (or products of the reaction
of oxygen with chlorine) are vitally involved in the organic oxidation.
Oxygen, or oxygen-chlorine reaction products, could act as inhibitors or
promoters of the organic oxidation. Perhaps dissolved oxygen could act as
a major oxidizing agent, and effective organic oxidation could be achieved
using oxygen in the presence of only small smounts of chlorine. Conversely,
if oxygen is an inhibitor of the chlorine oxidation, then perhaps more rapid
oxidation and more efficient use of chlorine could be achieved if all oxygen
were excluded from the system.
A series of experiments was performed in order to check these
possibilities. The equipment used was the 5-liter reactor used in the
original UV-chlorine work.
In the first experiment, the effluent was saturated with oxygen,
and oxygen was continuously supplied to the effluent-chlorine mixture during
the additions. In the second experiment, the dissolved oxygen was purged
with nitrogen, and a positive nitrogen pressure was maintained during the
oxidation. In the third experiment, no treatment of the reaction mixture
with either oxygen or nitrogen was employed.
In all three experiments, the rates and extents of COD and TOG
decreases were about the same (details are presented in the Experimental
section, p. 58). Apparently, oxygen had no effect on the rate of oxidation.
Certainly, the results show that oxygen had no inhibitory effect.
In subsequent work employing a batch-recycle reactor, a question
arose: is it possible to use less-than-stoichiometric amounts of chlorine
in this process? Preliminary experiments indicated that the answer was
affirmative. If so, then molecular oxygen is very probably the oxidizing
agent which accounts for the organic oxidation achieved in excess of that
accomplished by chlorine.
20
-------
The experiments described in this section do not contradict this
possibility. Under the conditions investigated, a large concentration of
chlorine and a relatively small concentration of oxygen were involved.
Under these conditions, it would not be possible to detect whether or not
significant amounts of oxygen were actually consumed during the experiments.
To verify this possibility, additional experiments are required
in which (l) minimum concentrations of chlorine are employed and (2) careful
measurement is made of the amount of chlorine consumed and of the organic
oxidation accomplished.
21
-------
ESTIMATED PROCESS COSTS
Process costs have been estimated for the large-scale application
of the UV-chlorine process to wastewater. Two major factors affect these
estimates: (l) the kind of ultraviolet source used, and (2) the amount of
oxidation which can be accomplished by a given amount of applied radiation.
As discussed below, the wavelength-effect studies have shown that low-
pressure, mercury-arc lamps are most probably the best, presently available,
radiation sources. Also, preliminary data from the batch-recycle reactor
studies have provided a useful estimate of the effectiveness of the applied
radiation.
Radiant Energy Source
A primary object of this research effort was to obtain informa-
tion which would permit the selection of the most practical, radiant-energy
source. The wavelength-effect studies have indicated that wavelengths
shorter than 300 mM- are the most effective. A consideration of available
sources (see Appendix A) indicated that the most practical sources of this
kind of radiation at the present time are mercury-arc lamps.
There are two general types of mercury-arc lamps: low-pressure
and high-pressure arcs. High-pressure arcs emit a broad spectrum of UV
wavelengths between 220 and 366 mp,. Low-pressure lamps emit radiation al-
most exclusively in the region of 254 mp,. High-pressure arcs emit much-
more-intense radiation than low-pressure arcs of about the same size. Also,
the high-pressure arcs are usually short, typically less than 1 ft in length.
Low-pressure arcs can be of almost any length, and lamps 5 ft in length are
commercially available. The useful life of high-pressure arcs is relatively
short—usually less than 1,000 hr; low-pressure lamps have long lifetimes —
usually 7,500 hr or longer. High-pressure arcs are more expensive than
low-pressure arcs, but do not have the relatively high electrical efficiency
of the low-pressure arcs.
It should be emphasized that there are many variables in the con-
struction of mercury-arc lamps and that certain design changes are possible
which can maximize the usefulness of a mercury arc for specific purposes.
For example, higher intensities can be produced from low-pressure arcs by
the application of higher voltages. The higher intensity is not obtained
without some concessions; for example, shorter lifetimes or lower electri-
cal efficiency is usually the result of the application of higher voltages.
22
-------
Low-pressure mercury arcs are commonly used for the sterilization
of water and are designed to emit the germicidal wavelengths with high effi-
ciency. Prominent manufacturers of this kind of equipment were contacted
when it became evident that low-pressure arcs might also "be useful for the
UV-chlorine oxidation process. It was somewhat surprising to learn that
most of the manufacturers were dissatisfied with commercially available
arcs; for example, one company imports specially made arcs from Europe, and
another company 'manufactures its own arcs.
Because of the present state of the art of UV lamp manufacture,
it is highly probable that the ideal mercury arc for the UV-chlorine
process is not presently commercially available, but will require special
manufacture to particular design specifications.
However, for the current review of process costs, commercially
available low-pressure mercury arcs have been selected because they repre-
sent the nearest approximation to the kind of UV source that will be needed.
The main disadvantage of presently available, low-pressure arcs is
their relatively low intensity which, in turn, necessitates the use of a very
large number of lamps. Similar lamps of much higher intensity can probably
be produced. -Lamps of two or three times the intensity of presently avail-
able lamps would permit significant reduction in plant investment costs--
perhaps as much as 3.00^/1,000 gal. However, the increased intensity could
probably not be obtained without an offsetting loss of lamp lifetime or of
electrical efficiency. Therefore, the estimated lamp costs and power costs
probably represent a close approximation of the actual operating costs.
The experimental results show that low-intensity radiation produces
higher quantum efficiencies. The advantage of low-intensity radiation would
not be seriously offset by a substantial increase in the intensity of the
source. This apparent contradiction is explained by the fact that regard-
less of the intensity of the source, most of the effluent in a reactor which
is designed for optimum utilization of the radiant energy is actually treated
with low-intensity radiation because of attenuation factors. For -example,
for a typical effluent in a cylindrical, concentric reactor, over half of
the incident radiation would be absorbed by only 10$ of the liquid located
nearest to the radiant-energy source. The next 10$ fraction of effluent
would be exposed to radiation of less than one-half the incident intensity.
Thus, the bulk of the effluent affected by the radiation would be absorbing
radiation at relatively, very low intensity.
23
-------
Effectiveness of the Applied Radiation
The batch-recycle reactor studies provided the first realistic
estimate of the effectiveness of radiation from a low-pressure mercury
arc. These experiments are described in detail in the section of this
report entitled "Batch-Recycle Reactor Studies/' p. 61. The effectiveness
of the applied radiation was determined as the "quantum efficiency," de-
fined as the number of molecules of oxygen utilized per quantum absorbed.
Two surprising observations were made: first, the quantum efficiency in-
creased as the oxidation proceeded, and second, the quantum efficiency was
greater if the applied intensity were decreased.
In the present process-cost estimation, it was assumed that the
quantum efficiency is equal to the average quantum efficiency (1.30) ob-
tained in two of the low-intensity experiments (Table II). This assumption
is a conservative one because most of the reaction mixture in a practical
reactor will be exposed to very low intensities. In the proposed large-
scale reactor design, low intensities can be employed efficiently because
of allowance for long contact time; the contact time in the proposed reactor
is about 40 min, or almost four times longer than the contact times examined
in the batch-recycle reactor studies.
Plant Description
A process flowsheet for a plant handling 10-million gal./stream
day is presented in Figure 7. The plant has been designed to reduce the
COD from 25 to 10 mg/liter using the G64T6, low-pressure, mercury-arc lamp.
This lamp is in commercial production. The UV output is 18.5 w for an
electrical input of 65 w. A plant treating 10-million gal./day would require
4,020 lamps.
To reduce the COD from 25 to 10 mg/liter, the stoichiometric amount
of chlorine required is 67 mg/liter. This chlorine would be added contin-
uously to the wastewater at four points in the'reactor flume." Since the
effectiveness of the UV-catalyzed chlorine oxidation is influenced by pH,
provision was made for the addition of caustic at two points in the
"reactor flume."
The plant would provide 10 reactor flumes with a depth of 4.5 ft,
a width of 5.5 ft and a length of 160 ft. Each flume would contain 402
lamps enclosed in quartz tubes. Four horizontal lamps are placed in verti-
icalrows 12 in. apart in the front section of the flumes. Only three lamps
in vertical rows 20 in. apart would be needed in the last part of the flumes.
This spacing of lamps permits absorption by the effluent of practically all
the radiation and also allows for the decrease in absorption as the reaction
proceeds.
24
-------
CHLORINE 5,580 UBS/DAY
—Q-
CHLORINE STORAGE TANK
19,000 GALLON CAPACITY
WASTE WATER FROM
ro
cn
SECONDARY *=*
TREATMENT 5 PUMPS
1,400 CPU
COD 25MS/L
—Q-
O
O
o
TERTIARY TREATED WASTE KATER
10 MILLION GALLONS/STREAM DAY
COD IOMG/L
10 REACTOR FLUMES
402 LAMPS PER FLUME
SIZE: 4.5 X 5.5 X 160 FT
CAUSTIC STORAGE TANK
12.000 GALLON CAPACITY
CAUSTIC PUMP
MIDWEST RESEARCH INSTITUTE
KANSAS crnr. MISSOUM
PROCESS FLOWSHEET
TERTIARY TREAMfT Of LIGHT-CATALY2H) OLORINE
CAPACITY: 10 MILLION GALLONS/DAY
APPROVED
SCALE •
DATE 5/23/70
DWG. NO.
M55 - I
Figure 7 - Process Flowsheet for Tertiary Treatment by Light--Catalyzed Chlorine,
Capacity 10-Million Gal./Day
-------
One feed pump would serve two flumes. Figure 7 shows two flumes
with the one pump.
Chlorine would be supplied as vapor from a 19,000-gal., liquid-
chlorine tank. Automatic, flow-control equipment would regulate the chlorine
flow to each injection point. Caustic from a storage tank would be con-
trolled by pH measurement near the injection points.
An alternate reactor design was considered. It consisted of
annular flow around several lamps. This heat-exchanger type of design
would require arrangement for parallel and series flow of several hundred
units. Thus, piping and valve arrangements would become quite extensive.
It is felt that the "flume design" has more advantages from a simplicity
and cost standpoint.
Plant Investment
The plant investment has been estimated for a plant capacity of
10-million gal./stream day. Design basis involves decreasing the COD from
25 to 10 mg/liter. Figure 7 shows the major equipment involved which in-
cludes the following items:
10 Reactor flumes - each 4.5 x 5.5 x 160 ft
4,020 Quartz tubes
4,020 Ballasts
5 Wastewater pumps - 1,400 gpm
Chlorine storage tank - 19,000 gal.
Caustic storage tank - 12,000 gal.
Chlorine pump - 100 gpm
Caustic pump - 100 gpm
Estimated cost of installed major equipment is $257,500.
Other plant costs:
Piping $ 54,300
Electrical 144,000
26
-------
Instrumentation $102,000
Painting 17,000
Subtotal $574,800
Contingency 57,480
Contractor and
engineering 86,220
$718,500
A building 130 x 190 ft would "be required to house all 10 reactor flumes,
control equipment, office and lab. Using two acres for the plant site, the
building and site costs have been estimated as
Site $ 24,000
Building
130 x 250 ft 247,000
$271,000
This gives a total plant investment of $989,500, excluding lamp costs. The
G64T6 lamp has a life expectancy of one year. Therefore, it is felt this
cost item of $56,200 is better treated as a yearly operating cost under the
next section.
Treatment Cost
Treatment cost was estimated as follows:
Annual ^/1,000 gal.
Labor: 2 men/shift at $350 $ 61,320 1.86
Labor overhead: 50$ 30,680 0.93
Raw materials
Chlorine: 5,580 Ib/day at 3.750 69,000 2.09
Caustic: 2,000 Ib/day at 3.300 21,780 0.66
27
-------
Utilities
Fuel gas: 25^/MM Btu 600 0.02
Electricity: 456 kw at 1.0^/kw-hr 36,100 1.09
Lamps: 5,680 at $14 56,200 1/70
$275,680 8.35^
Fixed charges on the investment of $989,500 have been estimated as
Annual ^/1,000 gal.
Depreciation over 20 years $ 49,475 1.50
Interest at 4$ 23,800 0.72
Insurance at 1$ 9,895 0.30
Maintenance at 3$ 29,685 0.90
$112,855 3.420
The total treatment costs are $388,535 annually, including fixed charges
on investment and operating costs. The unit-cost total is 11.77^/1,000 gal.
of wastewater treated.
By placing the 10 reactor flumes outside the building, costs could
be reduced an estimated $137,000. The plant investment would drop to
$852,500, and the fixed charges would be reduced by $15,600 annually. Total
unit costs would then be 11.30^/1,000 gal.
Substantial reductions in cost would be possible if (l) more in-
tense, longer-lived, or more efficient mercury-arc lamps were available,
(2) higher quantum efficiencies could be obtained, or (3) less-than-
stoichiometric amounts of chlorine were required.
The estimate of the cost of caustic was based on the experimentally
observed amount required to maintain the optimum pH of 5.0. Considerabley
less caustic would be required if the pH did not require adjustment. Ex-
perimental determination of the economic consequences of operating at
ambient pH has not been made.
28
-------
Consideration was also given to the cost of treating effluents of
higher COD content. Because of the longer reaction time required for
effluents of this kind, a higher quantum efficiency (3.72) could "be reason-
ably presumed. An adjusted estimate of the cost of reducing the COD of an
effluent from 45 mg/liter to 10 mg/liter is presented below:
Cost
(1/1,000 gal.)
Labor 1.86
Labor overhead 0. 93
Chlorine 4.88
Caustic 1.54
Fuel gas 0.02
Electricity 0.96
Lamps 1.22 11.40
Investment: $989,500
= 860,000 2'97
14.38
The above estimate clearly shows the relatively high contribution
to cost made by chlorine and caustic, and it suggested further study. These
factors were investigated toward the end of the research program using the
"batch-recycle reactor- Preliminary data (see Experimental section, p. 67)
indicated that up to one-third less than the stoichiometric amount of
chlorine could be employed. This savings in chlorine would amount to
1.6^/1,000 gal. in the above cost estimate.
29
-------
EXPERIMENTAL
Actinometer Procedures and Calculations
All of the actinometer experiments were performed in a darkroom
equipped with a red safelight. Ultraviolet and visible absorbances were
obtained using a Beckman DU spectrophotometer which was located in the
darkroom. Most of the procedures and calculations are similar to those
recommended by Calvert and Pitts.!/
Calibration of Monochromatic Wavelength Sources
Solutions of 0.00600 M ferric chloride and 0.01800 M potassium
oxalate were prepared. A 10-ml portion of each was placed in a brown
bottle protected from light with aluminum foil. The ferrioxalate solution
was pipetted into a cylindrical quartz cell of 1.0-cm lightpath and 2.2-cm
diameter (volume 2.65 ml). The cell was placed in the irradiation chamber,
as near to the interference filter as possible, and was irradiated. The
high-pressure, mercury-arc lamp had been allowed to warm up prior to placing
the cell in the holder.
The number of ferrous ions (NFe++) produced can be calculated from
the following equation:
6.023 x 102° ViV3 log (I0/I)
NFe++ =
where V^ = volume of actinometer solution irradiated (2.65 ml).
V2 = volume of aliquot taken for analysis (2.00 ml).
V3 = final volume to which aliquot, V2, is diluted (5.00 ml).
log (lo/l) = absorbance of solution at 510 mp,.
i = path length of spectrophotometer cell used (l.OO cm).
e = experimental value of the molar extinction coefficient of
the ferrous complex (l.ll x 10^ liters/mole-cm).
Thus, for this system, NFe++ = 3.67 x 1017 x A, where A is the difference
between the absorbance of the irradiated and nonirradiated actinometer
solutions.
30
-------
The intensity of the light beam, incident just within the photo-
cell reactor front window (l^), can be calculated from the following
equation:
i HFe++
I0 = —. quanta/sec, (l)
cp t
where cp = the quantum yield of the actinometer reaction at the par-
ticular wavelength.
t = the irradiation time (sec).
e = the molar extinction coefficient of the actinometer solution
at that wavelength.
[A"] = the concentration of the actinometer solution.
H = length of the cell.
The quantity, (l-10-eLA.U), is the fraction of the incident light
absorbed by the actinometer and can be measured experimentally using a
spectrophotometer (l-T, where T = transmittance), or it can be calculated
from known values of e, A and i,.
The selected procedure for determining the amount of ferrous iron
formed by the oxidation of oxalate by ferric iron was similar to that of
Hatchard and Parker£/ and is described by Calvert and Pitts.i/ An excep-
tion is that smaller volumes have been employed. A solution of ferrous
sulfate (5.06 x 10"4 M) in 0.1 N HgSO^ was freshly prepared from 0.1012 M
ferrous sulfate standardized (by titrating an aliquot of standard dichromate)
just prior to use. Portions of the dilute ferrous sulfate solution (0, 0.1,
0.2, 0.4, 0.6, and 0.9 ml) were mixed with 2 ml of 0.1 N H2S04, 8 drops of
0.1$ 1,10-phenanthroline in water, 1 ml of buffer (600 ml of 1 N sodium
acetate and 360 ml of 1 N H2S04 diluted to 1 liter), and the resulting solu-
tion was diluted to 5.0 ml in a volumetric flask and allowed to stand in the
dark for 1 hr. The absorbance of these solutions was determined, and a
linear plot of absorbance versus concentration was obtained having a slope
in excellent agreement with the 1.11 x 104 liters/mole-cm reported by
Calvert and Pitts.
Calibration of Polychromatic Wavelength Sources: The following
is a derivation of the equation used to calculate the total intensity of a
polychromatic source using the ferrioxalate actinometer.
31
-------
The intensity of radiation just within the cell-reactor front
window is determined from Eq. (l).
Conversion from quanta/sec to watts results in the following:
FO++ / ^
I, . = £S watts (2)
tot cp t (l-T)X
X = wavelength in mp,.
k = constant (19.86 x 10"1? mp,'w sec/quantum),
T = transmittance.
Transposition of the equation gives the following:
. itot
The number of ions formed at each separate wavelength can be cal-
culated using the following equation:
where \ = the discrete wavelength considered.
Bie total number of ions is the same sum of the ions formed at
each separate wavelength.
The intensity at each wavelength can be calculated from the fol-
lowing equation, provided that information is available concerning the out-
put as a function of wavelength:
(FX/Ftot)
32
-------
Substitution into Eq. (5) gives:
Removing the constant term (ifcot) Sives:
E
\
Solving for the total intensity yields the following equation:
itot = -= ^^ 0 )
The equation for N™ ++ taken from Calvert and Pitts—/ is:
6.023 x 1020 VjVg log (Ip/l)
V-T = 3.00 ml = volume of actinometer solution irradiated.
VQ = 2.00 ml = volume of aliquot taken for analysis.
Vrz = 5.00 ml = final volume to which the aliquot, ¥2, is diluted,
o
log _£ = A = absorbance of actinometer solution at 510 nip,.
I 510
& = 1 cm = path length of spectrophotometer cell used.
e = 1.11 x 10 liter/mole-cm = experimental value of the molar
extinction coefficient of the Fe++ complex as determined
from the slope of the calibration curve.
Substituting the known values for k, V-j_, V2, V3, & and e into
Eq. (9) yields the expression used to calculate the intensity of poly-
chromatic radiation:
33
-------
(oi)
X (Fx/Ftot)
Monochromatic Wavelength-Effect Experiments
Equipment: All equipment for the wavelength-effect studies was
located in a separate laboratory which was converted to a darkroom and was
equipped with a red safelight.
A diagram of the apparatus constructed for use in the monochromatic
irradiation studies is presented in Figure 8. A 22-mm-diameter circular
hole was cut in one end of a 4-in. x 6-in. metal box. A holder for the
2-in. x 2-in. interference filters was attached to the outside of the box.
A cell holder was placed inside the box in such a manner that a quartz cell
could be located very near to the interference filter or could be located
up to 5 in. away from the filter. The inside of the box was covered with a
black nonreflecting cloth which provided a light-tight seal when the box was
closed.
The ultraviolet source used in most of the experiments was a
"Universal Point Source", Model 401, supplied by PEK Labs, Inc., of Sunnyvale,
California. This source is equipped with a power supply which permits
adjustment of the current. The high-pressure mercury arc was enclosed in
a housing equipped with a photographic-type variable aperture to permit ad-
justment in radiant-energy output. A new opening in the back of the lamp
housing was made in order to locate the source closer to the interference
filter. To obtain precise timing of the actinometer exposures, a simple
shutter was devised to cut off all radiation from the cell before and after
irradiation.
The optical characteristics of the interference filters are pre-
sented in Table III.
34
-------
IRRADIATION CHAMBER
APERTURE
INTERFERENCE FILTER
UV SOURCE
PROPER POSITION FOR CELL
APERTURE
Figure 8 - Apparatus Used in Wavelength-Effect Studies
35
-------
TABLE III
CHARACTERISTICS OF INTERFERENCE FILTERS
Mercury-Arc
Line (mp.)
313
334
366
405
436
546
Beak Wavelength
of Filter (tmj.)
314
334
363
407
435
541
Optical Transmission
at Peak (%)
24
10-20
29.5
35-40
35-40
35-40
Half-Beak
Width (mp.)
6.5
11
7
12
10
10
A low-pressure, mercury-arc lamp was used in the experiments at
254 mjj,. Using a suitable shield, the lamp was located at a point near the
aperture of the irradiation chamber, and no filter was employed.
To determine the intensity of the monochromatic irradiation en-
tering the cell, the ferrioxalate actinometer was employed. The actinometer
procedures are described in the preceding section.
Buffer Systems: Calculations indicated that, for 4.0 liters of
effluent and 190 ppm of chlorine, a buffer charge of 102 g of KH2P04 and
2 g of Na2HP04 is required in order to maintain the pH between 4.5 and
5.5 as all of the chlorine is converted to HC1. A UV-catalyzed experiment
of this kind was carried out using effluent MRI 730, and the initital pH
of 5.0 decreased to only 4.6 after all the chlorine had been consumed.
A similar experiment was performed in which the pH was maintained
using added alkali. The rates of chlorine consumption and organic oxidation
were about the same in each experiment and were comparable to those observed
using other effluents. Thus, phosphate buffers do not interfere with either
the rate or extent of the light-catalyzed chlorine oxidation nor with the
analytical procedures employed.
Microanalysis for Chlorine: It was necessary to develop a reliable
analysis for chlorine to be applied to very small samples taken from the
irradiated cell containing effluent and chlorine. For this purpose a cali-
bration curve was prepared using an orthotolidine-arsenite procedure capable
of detecting 1-40 mg/liter of chlorine in 50 p,l of sample. The sample
(50 (jJL) was taken from a chlorine solution which had been analyzed ampero-
metrically (4 ml diluted to 200 ml). The sample was mixed with 2.5 ml of
orthotolidine reagent (0.0675 g in 500 ml of water; a 1:20 dilution of the
conventional reagent) in a 1-cm cylindrical quartz cell and was shaken for
36
-------
exactly 30 sec. At this time, 3 drops of sodium arsenite (5.0 g of NaAs02
in 1 liter of water) was added, the cell was shaken and the absorbance at
440 mn (0.01-mm slit width) was determined immediately. A straight line
passing through the origin was obtained when the absorbance was plotted
concentration from 0 to 40 mg/liter.
Using this method, the concentration of chlorine was determined
at various stages of the effluent oxidation in the irradiated cell.
Effluent- Irradiation Procedure; Each 50-ml portion of effluent
was treated with 1.25 g of monopotassium dihydrogen phosphate (KHgPO^),
0.025 g of disodium monohydrogen phosphate (RagHPO^ and sufficient chlorine
water to produce the desired initial concentration.
Prior to each series of effluent irradiation experiments, the in-
tensity of the radiation incident on the cell was determined using the
ferrioxalate actinometer. Each actinometer study consisted of a series of
accurately timed irradiation exposures. Two or more exposure times were
employed, and, at each time level, duplicate or triplicate runs were made.
Care was taken to insure that the actinometer solution was not overexposed.
Total Organic Carbon Determinations ; Total organic carbon (TOC)
was determined using the Beckman, Model 915, Total Organic Carbon Analyzer.
Experimental Problems ; Because the wavelength-effect experiments
required small volumes of reaction mixtures, TOC determinations were the
only feasible method for determining changes in the organic content of the
irradiated effluent. The determination of changes in chlorine concentrations
was very precise, but the precision of the TOC determinations was unsatis-
factory for small changes in organic concentration. To overcome this
problem, critical experiments were repeated until an acceptable degree of
confidence in the results could be obtained.
Another factor which influenced the precision of the results is
the determination of UV absorbances of irradiated and nonirradiated
samples at relatively long wavelengths (350 imi, and longer). The absorbance
figures are required in order to calculate the amount of radiant energy
absorbed. Since the absorbance of a chlorinated effluent in a 1-cm cell is
relatively low at these wavelengths (absorbance equals approximately 0.04
at 360 mp,), the error inherent in the spectrophotometer (estimated to be
about 0.005) becomes very significant.
37
-------
Experimental Results: The results of monochromatic wavelength-
effect experiments are presented in the following paragraphs.
At 566 mp,; The results of the effluent oxidation studies
at 366 mp, are presented in Table IV. The actinometer data are presented
in Table V, and a description of the effluents used is presented in Table VI.
The radiant energy absorbed by the effluent-chlorine solution was calculated
from the applied energy (actinometer data), multiplied by the UV absorption
(fraction of incident intensity absorbed). The UV absorption was obtained
by determining the UV absorption spectrum of the effluent-chlorine solution
before and after irradiation. If considerable decrease in absorption
occurred, then an average of the absorption before and after irradiation
was used.
Except for one, apparently anomalous, result, very little
decrease in chlorine concentration or TOG was observed at this wavelength
even when the energy absorbed was 23 w min/gal.
At 456 mp,; The results obtained at 436 mp, are presented
in Table VII. This wavelength appears to be very ineffective in reducing
chlorine concentration or TOG. Because of low absorbance, long irradiation
times were required in order to obtain significant energy absorption by the
solution.
At 515 mp,; The results obtained at 313 mp, are presented
in Table VIII, p. 41. This wavelength is considerably more effective in
lowering chlorine concentration and TOG than are longer wavelengths. (The
chlorine concentration decreased by 75 mg/liter, but the TOG decreased
only 1.9 mg/liter upon the absorption of 29 w min/gal. of radiant energy.)
At 405 mp,: The 405-mp, interference filter used with the
high-pressure, PEK, mercury-arc lamp produces adequate monochromatic radiation
intensities for the photochemical study (Table v). However, the low absorp-
tion of this radiation by the effluent-chlorine mixture (Table IX) results
in relatively small amounts of radiant energy absorbed (only 7.7 w min/gal.
after 1 hr). Under these conditions, very little decrease in chlorine
concentration or TOG was achieved (Table X).
The intensity of 405-mp, radiation impinging on the cell in
these experiments is 580 pw/cm2, a relatively high value (2.2 x 103 pw of
energy impinge upon the 3.8-cm2 area of the cell). In order to deliver to
the solution an amount of energy equivalent to 15 w min/gal. within 15 rain,
the intensity of radiation impinging on the cell would have to be increased
to about 8 times its present value. In actual practice, incident inten-
sities at 405 mp, of the required magnitude (about 2,000 pm/cm2) would
probably be impractical.
38
-------
TABLE IV
MONOCHROMATIC WAVELENGTH-EFFECT STUDIES AT
O)
CD
Radiant Energy Dark
Entering Cellk/ Chlorination
(mw) Time (hr)
3.2
3.2
3.2
3.2
1.2
1.2
1.2l/
2.7
2.7
2.7
2.7
24
24
1
3
4
1
2
48
48
48
72
366 myj^/
Radiant
Irradiation Energy
Time Applied
(min) (w min/gal.)
10
20
10
30
30
30
30
120
63
60
120
40.5
81.0
40.5
122
47
47
9.4
390
205
195
390
Radiant
Energy
Absorbed
(w min/gal . )
3.5
7.0
3.1
9.3
3.8
3.8
4.2
17.0
12.0
12.0
23.0
Chlorine
Concentration
(mg//)
Before
63
63
72
71
74
72
68
89
89
89
84
After
62
58
63
63
14£/
64
66
65
85
79
71
TOG
Before
-
8.0
8.0
7.5
7.2
7.2
9.1
8.8
8.9
8.7
After
-
8.2
8.1
6.8
7.1
7.1
7.9
9.1
8.7
9.2
a_/ All except one of the effluent oxidations were performed using a cylindrical quartz cell, 1 cm in
length. The first seven experiments employed effluent MRI 731. The last four employed effluent
MRI 925. The characteristics of these effluents are presented in Table VI.
b/ The actinometer results are presented in Table V.
c_/ This anomalous result was not confirmed by subsequent experiments.
d/ This experiment was performed using a 5-cm cell.
-------
TABLE V
ACTINOMETRIC CALIBRATION OF RADIATION SOURCES FOR
MONOCHROMATIC WAVELENGTH-EFFECT STUDIES
Radiation
Source-^/
A
B
C
A
C
C
C
C
C
C
C
Radiation
Wavelength
0-u)
366
366
366
313
313
313
436
405
405
405
546
Exposure
Time
(sec)
30
15
15
60
60
60
30
30
60
90
Increase in
Absorbance
at 510 mu
0.524
0.102
0.210
0.017
0.396
0.301
0.297
0.310
0.717
1.05
Number of
i i
Fe Ions
Produced
(x 10- 16)
21.3
4.15
8.54
0.69
16.1
12.2
12.1
14.6
29.2
42.7
Radiant Energy
Cell
(quanta/ sec
X lO'15)
5.88
2.21
4.85
0.093
2.17
1.65
7.92
4.84
4.82
4.71
Entering
3.20
1.21
2.65
0.059
1.37
1.04
3.61
2.37
2.36
2.31
1
a/ Radiation sources: A, PEK lamp with pyrex lenses; B, same lamp with lenses
removed; C, same lamp, no lens, located nearer to filter.
b/ This value was estimated; see text.
TABLE VT
CHARACTERISTICS OF EFFLUENTS USED IN THE MONOCHROMATIC
WAVELENGTH-EFFECT STUDIES
COD (mg/£)
TOC (mg/£)
Turbidity (JTU)
A.Tinionia-N (mg/l)
Nitrate-N (:?ig/;)
Chlorine demand (mg/£)
MRI 73lS/
23.6
8.4
1
2.5
22
-
MRI 925
30
10.6
1.9
0.8
7.2
8.3
a/The designation MRI 731 indicates that the effluent was prepared at
MRI on July 31 using a laboratory-scale activated sludge facility
starting with effluent from the Gracemore (Missouri) Municipal
Sewage Treatment Plant.
-------
TABLE VII
MONOCHROMATIC WAVELENGTH-EFFECT STUDIES AT 436 mu
Radiant Energy Dark
Entering Cell Chlorination
(mw) Time (hr)
3.61
3.61
3.61
24
2
4
Radiant
Irradiation Energy
Time Applied
(min) (w min/gal.)
30
30
120
136
136
544
Radiant
Energy
Absorbed
(w min/gal . )
1.4
1.4
5.4
Chlorine
Concentration
(mg/l)
Before
140
122
109
After
135
119
108
TOC
(mgA)
Before
8.1^/
10. &k/
9.9^/
After
8.0
10.3
10.0
a/ Effluent MRI 731.
b/ Effluent MRI 925.
TABLE VIII
MONOCHROMATIC WAVELENGTH-EFFECT STUDIES AT 515
Radiant Energy
Entering Cell
(mw)
0.059
0.059
1.37
1.04
1.04
Dark
Chlorination
Time (hr)
48
96
1
1
24
Radiant
Irradiation Energy
Time
(min)
161
157
30
130
60
Applied
(w min/gal . )
11.9
11.6
52
171
79
Radiant
Energy
Absorbed
(w min/gal . )
2.9
2.3
10
29
13
Chlorine
Concentration
(mg
Before
71
53
150
118
99
;/?,)
After
56
47
128
43
67
TOC
(mg/l)
Before After
8.12/ 8.5
8*1 b / DO
• -1 I O . £-
8 . 4k / 6.5
8 . 4k/ 7.4
Effluent MRI 731.
Effluent MRI 925.
-------
Furthermore, the data show that, at the presently available
intensity, the application of 15-16 w min/gal. in 120-135 min does not
produce significant decreases in TOC and does not produce large decreases
in chlorine concentration.
TABLE IX
UV-ABSORPTION CHARACTERISTICS OF CHLORINATED EFFLUENTS USED IN
At 405 tap*
At 546 nip,*
MONOCHROMATIC
Chlorine
Concentration (m|
WAVELENGTH-EFFECT STUDIES
?A)
Initial Final
ip* 84
109
112
80
91
94
Absorbance
(l-cm cell)
Initial Final
0.021 0.021
0.021 0.021
0.021 0.021
Fraction of
Radiation Absorbed
Initial
0.046
0.046
0.046
Final
0.046
0.046
0.046
95
90
0.005
0.005 0.010
0.010
Absorbances were calculated from absorbances determined in a 10-cm cell
compared to the absorbance of ultrapure water in the same cell.
At 546 imi: Additional problems were encountered in de-
termining the usefulness of longer wavelength radiation. 'The 546-imj,
radiation from a mercury arc is absorbed to a much lesser extent by the
chlorinated effluent than the 405-m|i radiation. Moreover, the intensity
of the 546-imx radiation cannot be determined by actinometry because the
quantum yield at this wavelength for the ferrioxalate actinometer is
nearly zero (see Appendix B).
The intensity of radiation at this wavelength can be
estimated by comparing the relative intensities of emitted radiation at
405 mjj, and 546 mp, and the relative transmission of the two wavelengths by
the filters (Table Xl). By multiplying the observed intensity at 405 m\i
(2.2 x 10--^ w) by the calculated ratio of the energy transmitted by the
filters (0.055), an estimate can be obtained for the intensity of 546-mu.
radiation impinging on the sample (l.2 x 10-3 w).
42
-------
TABLE X
MONOCHROMATIC WAVELENGTH-EFFECT STUDIES AT 405 mu and 546
Wavelength
405
405
405
546
Radiant Energy
Entering Cell
(raw)
2.2
2.2
2.2
1.2
Dark
Chlorination
Time (hr)
72
96
1
24
Irradiation
Time
(min)
60
120
135
120
Radiant
Energy
Applied
(w min/gal . )
167
333
374
180
Fraction
of
Radiation ,
Absorbed—'
0.046
0.046
0.046
0.010
Radiant
Energy
Absorbed
(w min/gal . )
7.7
15.3
17.2
1.8
Chlorine
Concentration
(mg//,)
Before After
84 80
109 91
112 94
95 90
TOC
Before
8.9
9.9
8.3
8.2
After
9.3
9.4
8.3
8.8
aj All of the effluent oxidations were performed using a cylindrical quartz cell, 1 cm in length.
(Table VI) was used in all of the experiments.
b/ The fraction of radiation absorbed is the average of initial and final values (see Table IX).
Effluent MRI 925
-------
TABLE XI
ESTIMATED INTENSITY OF 546-mu RADIATION
TRANSMITTED BY FILTER
Fraction Relative Amount
Irradiated Transmitted of Radiant Energy-
Wavelength Intensity.?/ by Interference Transmitted by Filter
(mix) (mw/steradian) Filter£/ (mw/steradian)
405 15 0.37 5.5
546 10 0.30 3.0
a/ These values are supplied by the lamp manufacturer, PER Applications
Information Sheet No. AN-202.
b/ The transmission values were obtained experimentally using a Beckman,
DK-2, spectrophotometer.
As shown in Table IX, only 1$ of the radiation at 546 mp,
is absorbed by a 1-cm cell. Thus, although the applied energy was 180
w min/gal. (Table X), the absorbed energy was only 1.8 w min/gal.
The absorption of this amount of energy at this wavelength
did not decrease the TOG and only slightly decreased the chlorine concen-
tration. The delivery of more than 1.8 w min/gal. of radiation at this
wavelength would probably be impractical.
At 254 imj.; The monochromatic ultraviolet source used in
these experiments was the low-pressure, mercury-arc lamp described earlier
in this section. The lamp was located at point near the irradiation chamber,
and no filter was employed.
Duplicate experiments were run in each series, and corre-
sponding samples were acidified to pH 2, dechlorinated with a slight excess
of sodium bisulfite and shipped to the Taft Water Research Center in
Cincinnati for TOG analyses. (The Beckman TOG analyzer at MRI was not
operating reliably during these experiments.)
A summary of the experimental data is presented in Tables
XII to XVI. In general, the decreases in absorbance, chlorine concentra-
tion, and TOG were proportional to the radiant energy absorbed.
44
-------
TABLE XII
MONOCHROMATIC WAVELENGTH-EFFECT STUDIES AT
Radiant
Irradiation Energy
Time Absorbed^/
(min)
20
20
20
20
40
40
40
40
62
60
60
60
(w min/gal . )
11.8
11.8
12.4
12.0
21.3
20.1
22.5
22.2
32.1
32.4
31.6
31.2
254 mu
Chlorine
Concentration T0c£/
(mg/4) (mg/0)
Initial
70.3
68.2
92.3
96.4
69.0
69.0
89.5
88.9
67.8
68.9
100.9
98.3
Final
42.5
41.0
52.9
54.3
21.5
21.5
25.7
22.7
4.2
4.2
9.2
10.2
Difference
27.8
27.2
39.4
42.1
47.5
47.5
63.8
66.2
63.6
64.7
91.7
88.1
Initial
5.9
(9.2)
8.9
(9.6)
11.5
(9.2)
8.5
(9.2)
8.8
(10.4)
8.9
(11.8)
Final
5.3
(9.4)
8.9
(8.6)
7.8
(7.0)
5.9
(7.8)
5.1
(8.6)
5.9
(7.2)
Difference
0.6
-0.2
0.0
1.0
3.7
2.2
2,6
1.4
3.7
1.8
3.0
4.6
a/ Data for the calculation of Radiant Energy Absorbed are listed in Table XIV.
b/ The parentheses indicate TOC analyses which were performed at the Robert A.
Taft Water Research Center.
TABLE XIII
SUMMARY OF TOG DECREASES AT 254 muP
Irradiation Time
_ (min) _
20 40 60
TOC Decreases (mg/jj) 0.6 3.7 3.7
-0.2 2.2 1.8
0.0 2.6 3.0
1.0 1.4 4.6
Averages 0.4 2.45 3.27
a/ The experiments are described in Table XII.
45
-------
TABLE XIV
ENERGY AT 254 mn, ABSORBED BY CHLORINATED EFFLUSNiS
Irradiation
Time
(ndr.)
20
20
20
20
40
40
40
40
62
60
60
60
Radiant Energy
Entering
CelW
(mw)
1.15
1.15
1.15
1.15
1.15
1.15
1.15
1.15
1.15
1.15
1.15
1.15
Radiant
Energy
Applied
(w min/gal. )
29.0
29.0
29.0
29.0
58.1
58.1
58.1
58.1
90.0
87.2
87.2
87.2
Fraction
of Energy .
Absorbed-'
0.408
0.406
0.430
0.415
0.367
0.346
0.387
0.382
0.358
0.372
0.363
0.358
Radiant
Energy
Absorbed
(w min/gal.)
11.8
11.8
12.4
12.0
21.3
20.1
22.5
22.2
32.1
32.4
31.6
31.2
a/ From average of Actinometer Studies, Table XVI.
b/ From UV Absorption Characteristics, Table XV.
TABLE XV
UV-ABSORPTION CHARACTERISTICS OF CHLORINATED EFFLUENTS USED
IN 254- mg,
Irradiation
Time
(min)
20
20
20
20
40
40
40
40
62
60
60
60
WAVELENGTH-EFFECT
STUDIES
Absorbance
(l-cm
cell)
Initial Final
0
0
0
0
0
0
0
0
0
0
0
0
.272
.263
.343
.317
.260
.252
.345
.320
.282
.29C
.307
.302
0
0
0
0
0
0
0
0
0
0
0
0
,
,
,
•
,
•
.
.
.
.
.
188
185
193
182
146
126
140
140
120
129
125
115
Fraction of
Radiation Absorbed
Initial
0
0
0
0
0
0
0
0
0
0
0
0
.465
.463
.500
.479
.450
.440
.500
.490
.476
.487
.475
.484
Final
0
0
0
0
0
0
0
0
0
0
0
0
.351
.348
.359
.351
.284
.251
.274
.274
.240
.257
.250
.231
Average
0
0
0
0
0
0
0
0
0
0
0
0
.408
.406
.430
.415
.367
.346
.387
.382
.358
.372
.363
.358
-------
TABLE XVI
ACTIMOMETRIC CALIBRATION OF THE RADIATION SOURCE FOR MONOCHROMATIC
WAVELENGTH-EFFECT STUDIES AT 254
Radiant Energy
Exposure Increase in Wo. of Fe++ Ions Entering Cell
Time Absorbance Produced /quanta/secN
(sec) at 510 mu (x 10"16) \x 1Q-15 / (mw)
b/
60 0.271-' 11.0 1.47 1.15
120 0.572^7 23.0 1.55 1.21
a/ The monochromatic UV source was a lov-pressure mercury-arc lamp.
b/ Average of two experiments.
TABLE XVII
SPECTRAL OUTPUT OF MERCURY-ARC SFOTLAMP (lOO-CH-4, SPOT. H58-4GS)
Wavelength
Range (mg,)
300-310
310-320
320-330
330-340
340-350
350-360
360-370
370-380
380-400
400-410
410-430
430-440
440-540
540-550
550-570
570-580
580-760
Watts
Radiated
0.01
0.13
0.04
0.09
0.05
0.06
0.16
0.06
0.14
0.55
0.09
1.01
0.35
1.55
0.14
1.53
0.96
Fraction of Total
300-540 mu
0.004
0.048
0.015
0.033
0.018
0.022
0.058
0.022
0.051
0.201
0.033
0.369
0.128
—
--
--
--
Output (F-^/F)
300-550 mjj,
0.002
0.030
0.009
0.021
0.012
0.014
0.037
0.014
0.033
0.128
0.021
0.236
0.082
0.361
--
--
—
47
-------
The TOO data obtained from the Taft Water Research Center
essentially confirmed the TOG data obtained at MRI in duplicate experiments.
However, the inherent lack of precision, at this level of TOG difference,
was apparent.
Iblychromatic Wave length- Effect Experiments
The effect of the radiation from a polychromatic, mercury-arc
lamp on the chlorine oxidation of effluents was investigated. The tech-
niques used were similar to those used in the monochromatic, wavelength-
effect studies (l-cm quartz-cell reactors). The spectral ouput of the
lamp is presented in Table XVII. (Note that there is no significant
emission below 300 m^.)
Actinometric studies were employed to provide an estimate of
the intensity of the radiation entering the cell.
Data required for polychromatic actinometer calculations are
presented in Table XVIII. As indicated in Table XVIII, the value of the
desired summation is as shown below:
= 322'47
Thus, from Eq. (ll) derived in the first part of this Experimental section:
_
300-540 ~ 1 - X °-251
sec
Actinometer data for the mercury- arc spotlamp are presented in
Table XIX. The total intensity between 300 m\j. and 540 nip,, as calculated
according to the above equation, was 3.9 x 10" 3 w.
As indicated in Table XVII, there is a certain amount of flexi-
bility in selecting the wavelength range for determining "total" intensity.
In our opinion, the wavelengths which must be considered are those which
influence the actinometer. Wavelengths which do not affect the actinometer
cannot be measured. There are two factors which can influence the extent
of actinometer response: the quantum yield and the fraction of the radiation
absorbed. For the ferrioxalate actinometer, both the quantum yield and
the fraction absorbed become very small at about 540 mp,. (See Table XVIII.)
However, there is a very intense mercury emission at 546 mp,. The inclusion
of this emission line into the actinometer calculations will affect the
48
-------
TABLE XVIII
DATA FOR ACTIHOMETRIC CALCULATION OF INTENSITY OF
MERCURY-ARC SPOTLAMP (WfiRTTTKfflDlJSE 100-CH-4^
Wavelength
305
315
325
335
345
355
365
375
390
405
420
435
490
545
Quantum
Yield (9)
1.24
1.24
1.24
1.23
1.23
1.22
1.22
1.21
1.18
1.15
1.12
1.11
0.88
0.27
Fraction of
Radiation
Absorbed by
Actinometer
(1-T)
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.99
0.95
0.81
0.56
0.10
0.01
Fraction of Emitted
Radiation Which
Affects Actinometer
(FX/F300-540) X ?>
0.004
0.048
0.015
0.033
0.018
0.022
0.058
0.022
0.051
0.201
0.033
0.369
0.128
—
t(l-T)x (FX/F300_540)
1.40
18.57
5.88
13.51
7.76
9.48
25.99
9.94
23.25
89.04
12.50
99.63
5.52
--
322.47
TABLE XIX
ACTINOMETER DATA FOR MERCURY-ARC SPOTLAMP (100 -CH-4)
Exposure Time
(sec)
30
30
30
15
15
15
0.452
0.447
0.497
0.217
0.239
0.248
a/ A510 is the difference in absorbance, at 510 m^,
between the exposed and nonirradiated actinometer
solutions.
49
-------
calculated fraction of total output (F^/Fj- t) and the calculation of total
intensity. However, the "total" intensity can be calculated with or without
this wavelength because of its relatively small effect on the actinometer
solution. Because of this flexibility, it is necessary to distinguish
exactly which wavelength range constitutes the "total" energy output.
Results of the effluent irradiation experiments employing poly-
chromatic radiation are presented in Table XX. Note that the applied in-
tensities are very large compared to those used in the monochromatic, wave-
length-effect experiments.
The calculation of total energy absorbed would require determina-
tion of the fraction of energy absorbed at each wavelength. The longer
wavelengths, 400 mp, and above, are only slightly absorbed, and spectro-
photometer data obtained using longer cells would be required in order to
obtain absorption data sufficiently precise to carry out this calculation.
Wavelengths above 400 mp, represent a very large portion (about
75%) of the "total" radiant energy applied, but this energy is almost com-
pletely wasted because of the low absorption. Wavelengths below 400 mp,
are much more strongly absorbed, and an estimate of the amount of energy
absorbed between 300-370 mp, is presented in Table XXI.
A number of significant facts are apparent from a consideration
of Tables XX and XXI. First, of the total of over 7,000 w min/gal. of
applied radiant energy, only a small fraction, about 140 w min/gal.,is
absorbed. Second, the major fraction of energy which is absorbed is in
the wavelength range of 300-370 mp,, and energy absorbed in this range does
not produce the extent of oxidation which can be achieved at lower wave-
lengths. The results of these experiments confirm the results of the
monochromatic, wavelength-effect experiments.
Inte ns i ty- Time Expe r ime nts
The intensity-time experiments were performed using the same
equipment used in the wavelength-effect studies. The ultraviolet source,
however, was the unfiltered radiation from a low-pressure, mercury-arc lamp.
Data supplied by the manufacturer (Nester-Faust) indicate that
96% of the UV output of this lamp is between 245 mp, and 260 mp,.
In the first series of experiments, the intensity of the radiant
energy and the irradiation time were varied. Particular attention was paid
to experiments in which the product of intensity and time was constant. The
results are presented in Tables XXII-XXVI.
50
-------
TABLE XX
RESULTS OF POLYCHROMATIC WAVELENGTH-EFFECT EXPERIMENTS
Run Intensity^/
No. (w)
45-1
45-3
47-1
0.164
0.192
0.187
Total Radiant
Energy
Irradiation Applied^/
Time (min)
Chlorine
Concentration
(mg/j&)
TOG
(w min/gal.) Initial Final Difference Initial Final Difference
5.0 1,040 138 111 27 9.8 8.6 1.2
15.0 3,600 136 83 53 8.6 7.9 0.7
30.0 7,050 138 60 78 7.9 5.5 2.4
a/ The UV source was a Westinghouse (lOO-CH-4) spotlamp.
b/ These figures show the total applied energy within the wavelength range, 300-540
01
H
TABLE XXI
ESTIMATION OF ABSORPTION OF POLYCHROMATIC RADIATION BY
CHLORINATED EFFLUENT (RUN 47-1, TABLE XX)
Fraction of
Radiant Energy
Absorbed^/ (w)
0.250
0.190
0.150
0.010
0.090
0.065
0.065
Wavelength
Range
(mn)
300-310
310-320
320-330
330-340
340-350
350-360
360-370
Fraction of
Total Emitted
Radiation
0.004
0.048
0.015
0.033
0.018
0.022
0.058
Radiant Energy
Entering CellSV
(w)
0.00076
0.00912
0.00285
0.00627
0.00342
0.00418
0.01100
Energy Absorbed
(w min/gal.)
7.2
65.5
16.2
2.4
11.6
10.3
27.1
140.3
a/ The radiant energy entering cell is calculated by multiplying the "total" intensity obtained from
the actinometer data (0.19 w.) by the fraction of total radiation emitted within the particular
wavelength range.
b/ As before, these values are obtained from the UV-absorption curves. Estimates were required in
some cases because of the low absorption by 1-cm. cells.
-------
01
ro
Run
No.
1
2
3
4
5
6
7
8
9
10
11
Intensity
(mw)
3.07
3.07
3.07
3.07
3.07
3.07
3.07
1.34
1.34
1.34
0.94
TABLE XXII
INTENSITY-TIME STUDIES (FIRST SERIES)
Irradiation
Time Cmin)
7.4
7.4
14.8
14.8
14.8
7.4
7.4
16.9
16.9
16.9
24.3
Radiant Energy
Absorbed?/
(w min/gal.)
28.6
28.6
57.1
57.1
57.1
28.6
28.6
28.6
28.6
28.6
28.6
Chlorine Concentration
Initial
68.4
56.4
54.8
52.2
96.4
59.3
94.1
72.0
105.0
93.8
90.9
(mg/jf. )
Final
42.1
25.7
13.6
-
9.3
29.7
53.6
38.0
58.0
53.5
52.5
Difference
26.3
30.7
41.2
-
87.1
29.1
40.5
34.0
47.0
40.3
38.4
a/ Data for the calculation of Radiant Energy Absorbed are listed in Table XXIII .
b Absorbance decrease; Mo-P total observed decrease in absorbance
Absorbance
Decrease^/
AA
AAT
0.122 0.027
0.119 0.032
0.120 0.042
0.120
0.208 0.088
0.141 0.031
0.141 0.041
0.090 0.035
0.113 0.048
0.097 0.041
0.080 0.039
-,.. = absorbance decrease attributable to decrease in chlorine concentration
UJ-o
= a'bsor"':)ance decrease attributable to decrease in absorbance of effluent.
Eff
0.095
0.087
0.078
0.122
0.110
0.100
0.055
0.065
0.056
0.041
(AAT =
AA
Eff
-------
TABLE XXIII
RADIANT ENERGY ABSORBED (FIRST INTENSITY-TIME SERIES)
Radiant Energy Radiant
Run
No.
1
2
3
4
5
6
7
8
9
10
11
Irradiation
Time
(min)
7/4
7.4
14.8
14,8
14.8
7.4
7.4
16.9
16.9
16.9
24.3
Entering
(mw)
3.07
3.07
3.07
3.07
3.07
3.07
3.07
1.34
1.34
1.34
0.94
Energy
Applied
(w min/gal. )
28.6
28.6
57.1
57.1
57.1
28.6
28.6
28.6
28.6
28.6
28.6
Fraction
of Energy
Absorbed^/
0.434
0.413
0.339
0.348
0.425
0.423
0.446
0.330
0.363
0.346
0.328
Radiant
Energy
Absorbed
(w rain/gal. )
12.4
11.8
19.4
19.9
24.3
12.1
12.8
9.44
10.4
9.90
9.38
a/ From average of Actinometer Studies, Table XXV.
b/ From UV-Absorption Characteristics, Table XXIV.
TABLE XXIV
UV-ABSORPTION CHARACTERISTICS OF CHLORINATED EFFLUENTS
Run
No.
1
2
3
4
5
6
7
8
9
10
11
(FIRST INTENSTTY-TD/IE SERIES)
Absorbance
(l-cm cell)
Initial
0.313
0.296
0.244
0.250
0.356
0.316
0.333
0.222
0.257
0.236
0.215
Final
0.191
0.177
0.124
0.130
0.148
0.175
0.192
0.132
0.144
0.139
0.135
Fraction of
Radiation Absorbed
Initial
0.512
0.492
0.429
0.438
0.559
0.516
0.535
0.400
0.446
0.419
0.390
Final
0.355
0.334
0.248
0.258
0.290
0.330
0.356
0.260
0.280
0.273
0.266
Average
0.434
0.413
0.339
0.348
0.425
0.423
0.446
0.330
0.363
0.346
0.328
53
-------
TABLE XXV
ACTINOMETRIC CALIBRATIOMg/ (FIRST INTENSITY-TIME SERIES)
Exposure
Time
(sec)
20
20
30
30
30
30
30
30
120
120
60
60
60
Increase in
Absorbance
at 510 ma
0.235
0.245
0.458
0.278
0.255
0.166
0.169
0.162
0.635
0.619
0.384
0.252
0.232
Radiant Energy
Entering Cell
(mw)
3.00
3.13
3.91
2.36
2.51
1.38
1.40
1.35
1.36
1.32
1.63
1.09
0.98
a/ The Monochromatic UV-source was a low-pressure, mercury-arc lamp.
TABLE XXVI
A COMPARISON OF EFFLUENTS MR I 925 AND MR I 112
COD (mg/jt)
TOG (mg/i)
Turbidity
Ammonia-N
NLtrate-N
a/ Figures in parentheses are data obtained when the effluent was
fresh.
1)
£)
y (JTU)
N (."igAO
N (:-.g/J&)
Demand (mg/c)
MRI 925^/
29.1 (30.0)
- (10.6)
0.95 (1.9)
0.2 (0.8)
5.0 (7.2)
8.9 (8.3)
MRI 112
18.4
11.0
1.0
0.1
27.0
3.6
-------
Complete data could not be obtained because of a breakdown in
the TOG analyzer. However, the preliminary conclusion (Table XXII) was
that the extent of reaction as measured by the decrease in chlorine con-
centration and the decrease in effluent absorbance is not proportional to
the product of intensity and time. That is, at a given level of absorbed
energy, low intensities appear to produce greater decreases in chlorine
concentration and absorbance.
The second series of experiments was performed using a freshly
prepared effluent (MRI 112). This effluent was very similar to the effluent
used in the first series (MRI 925). Analytical comparisons and oxidation
characteristics are presented in Tables XXVI and XXVII.
TABLE XXVII
A COMPARISON OF THE UV-CHLORINE OXIDATION CHARACTERISTICS
OF EFFLUENTS MRI 925 AND MRI 112g/
Before After
Irradiation Irradiation
MRI 925
TOG (mg/je) 9.1 6.6
C12 (mg/jt) 79.0 2.5
UV absorbance (254 mn,
1-cm cell, versus air) 0.220 0.084
MRI 112
TOG (mg/Jt) 10-8 7-8
C12 (mg/jR) 79-° 8'5
UV absorbance (254 mp,,
1-cm cell, versus air) 0.265 0.120
a/Reaction conditions: Effluents buffered to pH 5.0; 1-cm quartz cell;
source, low-pressure mercury-arc lamp (254 mpj; intensity 3 mw;
irradiation time, 30 min.
The results of the second series of intensity-time experiments
are summarized in Tables XXVIII-XXX. Experimental procedures were the
same as in the first series.
55
-------
TABLE XXVIII
INTENSITY-TIME STUDIES (SECOND SERIES)
Irradiation
Run TimeS/
No. (minj
32-1
32-3
32-5
'.2-7
32-9
32-11
CJl
Ol
32-15
36-1
36-3
36-5
36-7
36-9
15.0
7.5
15.0
30.0
15.0
30.0
60.0
50.0
88.0
26.0
50.0
101.0
Radiant
Energy
Intensity*-/ Absorbed^/
(mw) (w min/gal. )
3.0 22.5
3.0 15.2
3.0 21.5
2.0 28.1
2.0 14.8
1.0 14.9
1.0 27.1
3.0 77.1
3.0 122.0
3.0 38.2
2.0 45.0
2.0 96.4
Chlorine
Concentration
(m/i.)
Initial
106
105
105
100
101
106
99
145
143
136
139
142
Final
42
65
43
24
48
50
20
7
0
31
18
3
Difference
64
40
62
76
53
56
79
138
143
105
121
139
Initial
9.6
9.6
8.8
8.5
8.1
9.3
9.1
10.3
9.1
8.7
8.4
10.4
TOC
(mg/j
Final
7.0
8.5
7.8
6.9
7.1
8.3
5.3
4.5
3.2
5.2
6.9
5.5
Absorbance
;) Decrease^/
Difference (%} MT AAci2 AAfiff
2.6
1.1
1.0
1.6
1.0
1.0
3.8
5.8
6.9
3.5
1.5
4.9
27
12
10
19
12
11
42
56
76
40
18
47
0.138
0.100
0.113
0.135
0.090
0.135
0.148
0.153
0.181
0.141
0.171
0.192
0.083
0.062
0.082
0.093
0.073
0.076
0.096
0.148
0.153
0.119
0.134
0.149
0.055
0.038
0.031
0.042
0.017
0.059
0.052
0.005
0.028
0.022
0.037
0.043
a/ The effluent used in these studies was MRI 112. The properties of this effluent are compared to the effluent used in past experiments
(MRI 925; Tables XXVI and XXVII).
b/ The source was an unfiltered, low-pressure, mercury-arc lamp.
c/ Data for the calculation of radiant energy absorbed are presented in Table XXIX.
d/ Absorbance decrease; AAj = total observed decrease in absorbance
= absorbance decrease attributable to decrease in chlorine concentration
= absorbance decrease attributable to decrease in absorbance of effluent
-------
TABLE XXIX
RADIANT ENERGY ABSORBED (SECOND INTENSITY-TIME SERIES)
Radiant Energy^/
Run
Wo.
32-1
32-3
32-5
32-7
32-9
32-11
32-15
36-1
36-3
36-5
36-7
36-9
Irradiation
Time
(min)
15.0
7.5
15 0
30.0
15.0
30.0
60.0
50.0
88.0
26.0
50.0
101.0
Entering
Cell
(mw) (w
3.0
3.0
3.0
2.0
2.0
1.0
1.0
3.0
3.0
3.0
2.0
2.0
Radiant
Energy
Applied
min/gal.)
56.7
37.8
56.7
75.6
37.8
37.8
75.6
189.0
333.0
98.2
126.0
255.0
Fraction
of Energy
Absorbed-/
0.397
0.402
0.380
0.372
0.391
0.394
0.358
0.408
0.365
0.389
0.357
0.378
. Radiant
Energy
Absorbed
(w min/gal. )
22.5
15.2
21.5
28.1
14.8
14.9
27.1
77.1
122.0
38.2
45.0
96.4
a/ Source was an unfiltered, low-pressure mercury-arc lamp.
b/ From UV-Absorption
Characteristics ,
Table XXX.
Run
Mo.
32-1
32-3
32-5
32-7
32-9
32-11
32-15
36-1
36-3
36-5
36-7
36-9
TABLE XXX
UV-ABSORPTION CHARACTERISTICS OF CHLORINATED EFFLUENTS
(SECOND INTENSITY-TIME SERIES)
Absorbance
(1-cm cell)
Initial
0.295
0.277
0.268
0.275
0.263
0.290
0.273
0.310
0.297
0.289
0.285
0.313
Final
0.157
0.177
0.155
0.140
0.173
0.155
0.125
0.147
0.116
0.148
0.114
0.121
Fraction of
Radiation Absorbed
(254 mp.)
Initial
0.491
0.470
0.460
0.469
0.454
0.487
0.466
0.510
0.495
0.487
0.481
0.514
Final
0.302
0.334
0.300
0.274
0.327
0.300
0.250
0.305
0.235
0.290
0.232
0.242
Average
0.397
0.402
0.380
0.372
0.391
0.394
0.358
0.408
0.365
0.389
0.357
0.378
57
-------
The amount of chlorine consumed was directly proportional to
the amount of radiant energy absorbed regardless of the intensity (see
Figure 3, p. 10). However, at absorbed energies greater than about 50 w
min/gal., the consumption of chlorine did not increase as rapidly with in-
creased energy absorbed (note that in the experiments in which the absorbed
energy was more than 50 w min/gal., practically all of the chlorine had been
consumed (Table XXVIIl)).
The extent of TOG decrease appears to be dependent upon intensity
as well as upon energy absorbed. Analysis of the TOG data (Table XXVIIl)
indicated that, for a given amount of absorbed energy, an intensity of 3
mw was more effective than an intensity of 2 mw, but 1 mw was more effective
that 3 mw.
The results of the second series show that there is apparently
no relationship between the radiant energy absorbed and the decrease in
effluent absorbance.
A third series of experiments was performed to confirm the ob-
servation that decreases in TOG were dependent upon intensity, as well as
upon total energy absorbed. The results of these experiments are shown in
Tables XXXI-XXXIII and are presented graphically in Figure 2, p. 8.
The results confirm that, for a selected quantity of absorbed
energy, lower intensities are more effective. However, the difference in
effectiveness was not as great as was indicated in the second series of
experiments.
As indicated in Figure 2, p. 8, there was almost no intensity
effect up to a total absorbed energy of 30 w min/gal. Beyond this level,
the lower intensity (l mw) was significantly more effective than the high
intensity (3 mw).
Oxygen-Effect Experiments
The equipment used in these experiments was that used in the
original UV-chlorine work; a 5-liter flask was equipped with a quartz
immersion well, high-pressure mercury arc, pH electrode, etc. The
equipment was described in detail on p. 89 in the Final Summary Report for
FWPCA Contract No. 14-12-72. The amount of chlorine used in these experi-
ments was 30$ in excess of the stoichiometric amount. In the first series,
the chlorine was added at the beginning of the experiment; in the second
series, the chlorine was added incremently during the irradiation. The warmed-
up lamp was inserted about 1.5 min after the initial quantity of chlorine
58
-------
Ol
<£>
TABLE XXXI
INTENSITY-TIME STUDIES (THIRD SERIES)
Run
No.
38-1
38-3
38-5
38-7
Irradiation
Time^
(rain)
20
60
140
47
Radiant
Energy
Intensity^/ Absorbed^/
(mw) (w.min/gal.)
3.0
1.0
1.0
3.0
33.5
32.8
72.2
66.5
Chlorine Concentration
(ng/jj)
Initial
163
163
163
156
Final
54
53
16
15
Difference
109
110
147
141
Initial
10.1
9.1
9.0
8.4
TOG (mg/4)
Final
7.0
5.9
4.2
4.7
Difference ($>)
3.1 32
3.2 35
4.8 53
3.7 44
Absorbance
Decrease^/
AAT A-Acip
0.153 0.107
0 . 134 0 . 108
0.160 0.145
0.196 0.140
AAEff
0.046
0.026
0.015
0.056
a/ The effluent used in these studies was MRI 112.
b/ The source was an unfiltered, low-pressure, mercury-arc lamp.
cj Data for the calculation of radiant energy absorbed are presented in Table XXXII.
d/ Absorbance decrease; AA^ = total observed decrease in absorbance
~ absorbance decrease attributable to decrease in chlorine concentration
= absorbance decrease attributable to decrease in absorbance of effluent
(AAT =
AAEff)
-------
38-1
38-3
38-5
38-7
TABLE XXXII
RADIANT ENERGY ABSORBED (THIRD INTENSITY-TIME
Run
No.
Irradiation
Time
(min)
Radiant Energy^/
Entering Cell
(raw)
Radiant
Energy Applied
( w min/gal . )
Fraction
of Energy,
Absorbed^/
Radiant
Energy
Absorbed
(w min/gal. )
20
60
140
47
3.0
1.0
1.0
3.0
75.6
75.6
177
178
0.443
0.434
0.408
0.376
a/Source was an unfiltered, low-pressure, mercury-arc lamp.
b/ From UV-Absorption Characteristics, Table XXXIII.
33.5
32.8
72.2
66.5
TABLE XXXIII
UV-ABSORPTION CHARACTERISTICS OF CHLORINATED EFFLUENTS
(THIRD INTENSITY-TIME SERIES)
Run
No.
38-1
38-3
38-5
38-7
Absorbance
(l-cm cell)
Initial Final
0.338
0.320
0.316
0.314
Fraction of Radiation
Absorbed (255.7 ing,)
Initial Final Average
0.185
0.186
0.156
0.118
0.540
0.520
0.516
0.513
0.345
0.347
0.300
0.238
0.443
0.434
0.408
0.376
was added, and the total irradiation time was 20 min. The pH was maintained
in the range 5.0 to 6.0, and the chlorine concentration was monitored at
3- to 5-min intervals.
The first series of experiments was performed using effluent
S&L 527 (the characteristics of this effluent are presented in Table XXXIV).
Little difference in oxidation rate was observed in three experiments in
which, prior to oxidation, the effluent was (l) purged with nitrogen, (2)
purged with oxygen, and (3) given no preliminary treatment. However, rather
poor reduction in COD was achieved (a 40$ decrease after 10 min in the
dark and 7 min irradiation), and the effluent was shown to have an unusually
high, initial, chlorine demand (Table XXXIV).
60
-------
The experiments were repeated using effluent S&L 619, and the
results (Table XXXV) were about the same. The extent of oxidation after
10 min in the dark and 7 min irradiation was not as extensive as might be
expected.
During this period, the available effluents were low in organic
content because the influent was diluted by heavy rains. Considerably
lower organic contents were observed, although the ammonia contents were
usually low. (See S&L 619 to S&L 729, Table XXXIV.)
At this point, a sample of sludge and a sample of effluent
(G 718) were obtained from the Gracemore plant. After the sludge sample
had been aerated for 24 hr, the Gracemore effluent was run through the lab-
scale digester. As shown in Table XXXIV, the resulting effluents (MRI 720,
726, 731, 804 and 806) were of high quality.
Effluent MRI 720 was used in the final series of oxygen-effect
experiments. The results of these experiments are in Table XXXVI and
clearly show that there is no significant oxygen effect. Samples for
COD and TOG determinations were taken every 5 min, and plots of COD and TOC
decrease versus time were almost identical for each experiment.
Batch-Recycle Reactor Experiments
Description of Batch-Recycle Reactor; A diagram of the batch-
recycle reactor is presented in Figure 6, p. 15. The reactor portion con-
sisted of a 250-ml, photochemical reaction vessel (Ace Glass, Inc.). A
water-cooled, quartz immersion well was placed in the reactor by utilizing
the 60/50 glass joint at the top of the reactor. Reactant liquid was pumped
into the bottom of the reactor, up past the irradiated zone and back to the
reservoir through an outlet near the top of the reactor. A Teflon-clad
magnetic stir-bar was located at the bottom of the irradiated zone. The
apparatus was constructed in such a manner that the chlorinated effluent
would not contact any material except glass or Teflon.
The total volume of liquid in the reactor zone was 288 ml. The
irradiated zone was an annular volume, 0.57 cm in thickness and 18 cm in
length.
The recycle pump was an all-Teflon Saturn pump (The Fluorocarbon
Company) equipped with a Teflon throttling valve.
The reservoir was a 5-liter, four-necked, J^rex flask equipped
with graduated addition funnels for addition of chlorine water and sodium
hydroxide. A combination pH electrode was also placed in the reservoir, and
61
-------
TABLE XXXIV
ANALYTICAL DATA FOR EFFLUENTS USED IN OXYGEN-EFFECT STUDIES
Effluent?:/
S&L 527
S&L 619
S&L 703
S&L 716
S&L 717
G 718
MRI 720
MRI 726
S&L 729
MRI 731
MRI 804
MRI 806
COD
(mgA)
24.2
17.0
12.6
13.6
14.1
—
30.7
27.2
15.7
23.6
22.9
—
TOC
(rag A)
8.2
5.8
5.0
5.5
7.5
27.4
12.0
12.6
6.5
8.8
8.8
—
NH3-N
(mg/jt)
5.3
3.2
--
--
11.0
17.0
2.0
1.0
--
2.5
1.0
1.5
NOs-N
Turbidity^/
(JTU)
(mg/,e) Filtered Unfiltered
9.6
10.7
—
0.5
—
—
14.0
28.0
--
22.0
18.5
—
1.0
0.7
0.63
0.85
0.92
—
0.95
—
--
1.0
—
--
15.0
0.75
1.20
1.10
1.95
—
--
--
--
--
—
—
Initial Chlorine
DemandSL/ (mg/£)
20.8
2.7
1.6
--
4.8
--
6.7
--
_-
—
2.3
4.5
a/~~ The initials "S&L/1 "G" and'"MRI" stand for, respect!vely, Smith and Loveless, Gracemore and
Midwest Research Institute. The S&L effluents are from the Smith and Loveless research site
where the influent is sewage from the city of Lenexa. The Gracemore effluents are from the
Gracemore sewage treatment plant. The MRI effluents were prepared in a lab-scale, sewage-
treatment facility starting with Gracemore effluents. The numbers following the initials
refer to the date obtained.
b/ Turbidities were determined using the Hach, Model 2100, Turbidimeter.
£/ Initial chlorine demand was determined by adding chlorine to samples of effluent sufficient
to bring the chlorine concentration to about 50 mg/£. A similar determination was made
using distilled water as a blank. After 10 min, the difference in chlorine consumed
represented the chlorine demand of the effluent.
-------
TABLE XXXV
RESULTS OF FIRST SERIES OF OXYGEN-EFFECT EXPERIMENTS,
EFFLUENT ( S&L 619)
COD
TOG (mg/1)
Effluent Treatment?/ Initial Final Decrease Initial Final Decrease
Purged with oxygen 16.0
Purged with nitrogen 16.6
No treatment 15.6
6.2 61 7.3 4.6 37
7.2 57 6.9 3.6 48
6.8 56 6.9 4.2 39
a/ Analytical data for effluent S&L 619 are presented in Table XXXIV.
TABLE XXXVI
RESULTS OF SECOND SERIES OF OXYGEN-EFFECT EXPERIMENTS,
EFFLUENT (MRI 720)
Effluent Treatmen
.t2/
COD (mg/1)
TOG (mg/jj)
Initial Final Decrease Initial Final Decrease
Purged with oxygen
Purged with nitrogen
No treatment
29.4
29.8
30.2
4.7
3.9
4.4
84
87
85
12.0
12.9
11.8
3.1
3.4
2.2
74
74
81
a/ Analytical data for effluent MRI 720 are presented in Table XXXIV.
a glass tube was used to siphon samples as needed. The liquid in the
reservoir was stirred by means of a magnetic stirrer.
Liquid was pumped from the reservoir through a calibrated flow-
meter (100-850 ml/min) to the irradiated zone. The flowrate was controlled
by means of an all-Teflon throttle valve. A Teflon sample-valve was located
in the line from the pump to the irradiated zone.
The UV source was a Nester-Faust, low-pressure, mercury-arc lamp.
Data supplied by the manufacturer indicate that 96$, of the UV output of this
lamp is between 245 imj, and 260 mjj,. The maximum electrical input is 50 w,
and the lamp is reported to be 80$ efficient. According to the manufacturer,
the UV output should be about 40 w. However, our actinometer studies (des-
scribed below) indicated that the UV output was far below this amount, and
an explanation is presented later in this section.
63
-------
An ultraviolet intensity meter was attached to the outer wall of
the reactor for the purpose of detecting accidental changes in source in-
tensity. Absolute intensities were determined by actinometric procedures
described below.
Batch-Recycle Reactor Procedures; The experimental procedure em-
ployed in the recycle reactor studies was as follows. The reservoir (the
5-liter flask) was charged with 5.0 liters of effluent. The low-pressure,
mercury-arc source was allowed to warm up for 10 min, and chlorine was added
to the reservoir to bring the chlorine concentration to the desired level
(5 to 10 ing/liter). The pH was adjusted, if necessary, to about 5.0 using
concentrated hydrochloric acid or 10$ sodium hydroxide. A sample of the
starting mixture was taken, and a UV spectrum (5-cm cell) was obtained.
The TOG content of the sample was determined, and the sample was then
acidified (l ml concentrated I^SO^/ISO ml) and dechlorinated using a stream
of nitrogen gas. The COD of the dechlorinated sample was obtained.
The pump was started, and the flow rate was set by adjusting the
throttle valve. The sensing portion of the UV meter was attached to the
outer wall of the reactor and provided a means for determining whether or
not large changes in intensity occurred. Chlorine concentration, pH, flow-
rate and intensity were monitored frequently during the run. Additional
chlorine and sodium hydroxide were added as needed.
The contact time was determined by multiplying the contact time
per pass (reactor volume divided by flowrate) by the number of times the
reaction mixture was passed through the reactor.
Determination of Fraction of Radiation Absorbed: A critical
parameter that required determination was the fraction of radiation ab-
sorbed by the chlorinated effluent in the irradiated zone. The UV spectrum
of each sample was taken, prior to dechlorination, using 5-cm quartz cells.
The observed absorbances at 254 mp, were ir the range, 0.571 to 0.120. The
first two figures of these absorbance values are highly reliable, but the
third figure must be estimated (the chart paper range is 0 to 1.000). The
absorbances were run versus air using a Beckman DK-1 spectrophotometer.
The 5-cm cell was filled with distilled water, and this absorbance was de-
termined during each series of spectrum determinations for the purposes of
(l) obtaining a correction factor for cell characteristics (reflection, re-
fraction, etc.) and (2) providing a check on the optical cleanliness of
the cell. It is felt that this procedure was less susceptible to error than
the procedure of using two matched cells with distilled water in the refer-
ence cell. Experience has shown that the spectrum of our distilled water
is usually comparable to "ultrapure" water in the wavelength range 250-360 mu..
If the distilled water produced an absorbance higher (at 254 imj,) than ex-
pected for "ultrapure" water (0.041), then a check was made on both the
purity of the water and the optical quality of the cell.
64
-------
An additional, somewhat unique, factor required consideration in
these determinations: the actual absorbance of pure water. Subtraction
of the absorbance of a cell containing distilled water from the absorbance
of a cell containing effluent not only corrects for the cell characteristics,
but also subtracts the absorbance of pure water. In ordinary UV-spectrum
determinations, this produces no problem. However, at 254 imj,, the absorbance
by water itself is not negligible (A = 0.0225 for -a 5-cm thickness)3/ and
must be added to the observed absorbance to obtain the absolute absorbance
of effluent, plus water. When the absolute absorbance of the 5-cm thickness
of chlorinated effluent is obtained, this absorbance is converted by direct
proportionality to the absorbance of the 0.57-cm thickness of the irradiated
zone of the reactor. This final absorbance is then converted to fraction
absorbed (l-T, where T = transmittance) using the equation: A = Iog10 1/T.
This procedure yields highly accurate values for the fraction or radiation
absorbed by the chlorinated effluent in the irradiated zone.
Actinometric Determination of Intensity of Low-Pressure Lamp:
The reactor portion of the batch-recycle reactor was transferred to a
darkroom and filled with actinometer solution [0.006 M K3Fe(C204)3] to
the same liquid level that would be used in the effluent-oxidation ex-
periments .
The low-pressure mercury-arc lamp, described above, was placed in
a cylindrical light-shield and allowed to equilibrate for at least 10 min,
during which time cooling air was ciruclated through the shield. Shorter
equilibration times produced erratic results. If the lamp was not cooled
by air during the equilibration period, the surface temperature of the
lamp increased. When the lamp was placed in the reactor, it then cooled
causing the emission of varying amounts of UV radiation. (The effect of
surface temperature on the emission of low-pressure lamps is well known.)
When the lamp was emitting maximum intensity (at room temperature),
it was quickly transferred from the shield to the immersion well. The time
period between lamp-in and lamp-off was carefully measured. At the end of
the irradiation period, the actinometer solution was drained from the
reactor and an aliquot was taken and analyzed for ferrous ion, as in the
actinometer experiments. (See Actinometer Procedures, p. 30.)
The results of the reactor actinometer studies are presented in
Table XXXVII. The observed intensities were unexpectedly low. This lamp
was originally believed to have a total output of 40 w of UV radiation
(see Final Summary Report for FWPCA Contract No. 14-12-72, pp. 61 and 92).
When the lamp is placed within the immersion well, practically all of the
emitted radiation is "trapped" by the reactor zone because 100$ of the incident
radiation is absorbed by the actinometer solution. Only a small fraction
of the emitted radiation from the top of the lamp is lost through the
65
-------
TABLE XXXVII
ACTINOMETRIC CALIBRATION OF LOW-PRESSURE, MERCURY-ARC LAMP
Dial Setting^/
High
High
High
High
Medium
Medium
Medium
Low
Low
Low
Low
Low
Low
Exposure
Time
(sec)
5.3
5.2
5.7
5.5
10.4
10.2
10.4
32.6
30.1
15.5
15.4
33.2
30.1
Absorbanc
0.385
0.370
0.410
0.385
0.432
0.445
0.445
0.780
0.651
0.340
0.325
0.720
0.595
Absorbance
Produced
per Second
of Exposure
0.0726
0.0712
0.0719
0.070
0.0415
0.0436
0.0428
0.0239
0.0218
0.0219
0.0211
0.0217
0.0198
Intensity^/
(v)
1.76
1.72
1.74
1.70
1.01
1.06
1.04
0.579
0.528
0.530
0.511
0.525
0.479
jy "High11 intensity was produced with the power-supply dial-setting at the
maximum (100); "medium" intensity was equivalent to a dial setting of
75; "low" corresponded to a dial setting of 50.
b/ The absorbance figures are the differences between the exposed and non-
exposed actinometer solution.
£/ The intensity is the total intensity just outside the quartz immersion
well.
upper portion of the reactor. The observed low intensity could not have
been caused by deposits on the lamp or on the immersion well because these
had been carefully cleaned with concentrated hydrochloric acid. Two
possible explanations for the low absorbance are (l) a decrease in the
transmission properties of either the lamp or the immersion well, or (2) a
decrease in the actual production of UV radiation by the lamp.
Decreases in the UV transmission of glass are produced by a well-
known phenomenon called "solarization." Since the immersion well was al-
most new, its transmission should have been nearly maximum. A consultation
with the manufacturer of the lamp (Nester-Faust, Newark, Delaware) revealed
that other investigators had observed that the emitted UV radiation from
this lamp gradually decreases "to almost zero" because of solarization.
ec
-------
The manufacturer suggests that the intensity can be restored by
an annealing process at 550°C. However, this source had been carefully
calibrated; therefore, the reactor studies were continued.
The actinometer data almost certainly represented the actual
output of the lamp. Additional evidence was available from earlier
actinometer studies in which the actinometer solution was placed in 1-cm
quartz cells. If the observed intensity is divided by the area of the cell,
and if this intensity-per-unit-area is used to calculate the total intensity
impinging on a cylindrical surface (whose radius is equal to the cell-to-
source distance), a comparable value for the total intensity is obtained.
Experimental Results; The results of the batch-recycle reactor
experiments are presented in Figures 9-20 (pp. 68-79) and in Table II (p. 16),
The results of the preliminary investigation of minimum chlorine
requirements are presented in Table XXXVIII.
TABLE XXXVIII
RESULTS OF INVESTIGATION OF MINIMUM CHLORINE REQUIREMENT
Chlorine£/
Consumed
COD Decrease^/ TOG Decrease^./ (percent of
Effluent (mg/Vmin) (mg/Vmin) theoretical)
519 0.81 0.24 87
519 Oo83 0.24 87
519 1.39 0.36 d/
623 0.72 0.24 67
623 1.10 Oo49 d/
a/ In all the runs except Run No. 132, oxygen was bubbled continuously
into the reaction mixture. The low-pressure, mercury-arc source was
operated at maximum intensity (1.78 w).
b/ The rate of decrease in COD and TOC was calculated as mg/£/min of UV
contact time.
c/ The amount of chlorine consumed was calculated as the percent of the
stoichiometric amount required to accomplish the observed COD
decrease. The stoichiometric amount of chlorine is calculated by
multiplying the COD decrease (mg/^) by 4.43.
d/ In Runs 201 and 210, the chlorine concentration was maintained in the
range 5-10 mg/,6, but no measure of the total amount of chlorine
consumed was available.
67
-------
24
22 -
3.4
6.8 10.2
CONTACT TIME (min)
Figure 9 - Batch-Recycle Reactor Results - High Intensity - High Flow Rate
(Effluent MRI 112, Hun No. 66)
68
-------
3.4
6.8 10.2 13.5
CONTACT TIME (min)
Figure 10 - Batch-Recycle Reactor Results - High Intensity - High Flow Rate
(Effluent MRI 112, Run No. 68)
69
-------
24
22 -
3.4
6.8 10.2
CONTACT TIME (min)
Figure 11 - Hatch-Recycle Reactor Results - High Intensity - High Flow Rate
(Effluent MRI 304, Run Ko. 74)
70
-------
3.4
6.8 10.2 13.5
CONTACT TIME (min)
Figure 12 - Batch-Recycle Reactor Results - High Intensity - Low Flow
Rate (Effluent MRI 304, Run No. 78)
71
-------
3.4
6.8 10.2 13.5
CONTACT TIME (min)
16.9
Figure 13 - Batch-Recycle Reactor Results - Low Intensity - High Flow Rate
(Effluent MRI 304, Run No. 82)
72
-------
3.4
6.8 10.2 13.5
CONTACT TIME (min)
16.9
Figure 14 - Batch-Recycle Reactor Results - High Intensity - High Flow Rate
(Effluent MRI 310, Run No. 86)
73
-------
6.8 10.2 13.5
CONTACT TIME (min)
16.9
Figure 15 - Batch-Recycle Reactor Results - Lov; Intensity - High Flow Rate
(Iffluent MRI 310, Run Ko. 93)
74
-------
3.4
6.8 10.2 13.5
CONTACT TIME (min)
16.9
Figure 16 - Batch-Recycle Reactor Results - High Intensity - Low Flow Rate
(Effluent MRI 310, Run Wo. 10l)
75
-------
22 -
Figure 17
6.8 10.2
CONTACT TIME (min)
Batch-Recycle Reactor Results - Low Intensity - High Flow Rate
(Effluent MRI 319, Run No. 104)
76
-------
24
22
20
8
6.8
13.6
O COD
D TOC
A E
2.0
1.5
u
Z
LU
U
O
.5
CONTACT TIME (min)
Figure 18 - Batch-Recycle Reactor Results - High Intensity - High Flow Rate
(Effluent MRI 319, Decreased Volume, Run Wo. 108)
77
-------
24
6.8 10.2
CONTACT TIME (min)
Figure 19 - Batch-Recycle Reactor Results High Intensity
(Effluent MRI 319, Run No. 112)
- 2.0
High Flow Rate
78
-------
3.4
6.8 10.2
CONTACT TIME (min)
Figure 20 - Batch-Recycle Reactor Results - High Intensity - High Flow Rate
(Effluent MRI 319, Diluted, Run No. 117)
-------
Calculation of Qiantum Efficiency: "Die quantum efficiency
is defined as the number of molecules of oxygen utilized per absorbed
quantum. Ihe quantum efficiency can be calculated from the COD decrease,
the fraction of radiation absorbed by the reaction mixture, and the intensity
of radiation impinging on the reactor- The COD decrease is expressed
in terms of milligrams of oxygen per liter (mg/liter). The absorbed radia-
tion is conveniently expressed in terms of watt minutes per gallon. Con-
version of units is accomplished as follows:
mg OO/A 3.785.0 g moles 00 6.023 x 1023 molecules
E = - x x x £ x
w min/gal. gal. 103 mg 32 g mole
x 1-505 x IP"20 w min*
quantum
E = COD decrease (mgAe) x 0.930 = molecules of Og
absorbed radiation (w min/gal.) quantum absorbed
1 Qiantum = hv = ^ = 19.86 x 10"17 erg x 1 w sec x 1 min = 1>305 x 1Q-20
*• 253.7 x 10~7 107 erg 60 sec
h = Planck's constant 6.6255 x 10~27 erg sec
c = velocity of light 3 x 1010 cm/sec
v = frequency of radiation per sec
\ = wavelength in cm
80
-------
ACKNOWLEDGMENT
Dr. Alfred F. Meiners, Project Leader, was assisted by Dr. Evelyn
Murrill, Dr. Kenneth Fountain and Dr. Carl Kruse.
Dr. J. M. Smith, Chairman, Department of Chemical Engineering
at the University of California, Davis, acted as consultant for the photo-
chemical work. Dr. Frank C. Fowler, President, Research Engineers, acted
as consultant for the engineering aspects of the work and prepared the
process-cost estimations.
The project was conducted under the general supervision of
Dr. C. C. Chappelow, Jr., Head, Organic and Polymer Chemistry at MRI.
The Project Officer for the Water Quality Office, Environmental
Protection Agency, was Mr. Robert H. Wise.
81
-------
REFERENCES
1. Calvert, J. G., and Pitts, J. N., Photochemistry, John Wiley and Sons,
Inc., New York (1967).
2. Hatchard, C. G., and Parker, C. A., Proc. Roy. Soc. (London), A235_,
518 (1956).
3. Dawson, L. H. , and Hulburt, E. L., J. Optical Society of America, 24_,
175 (1934).
82
-------
APPENDIX A
PROBLEMS REGARDING SOURCE SELECTION FOR LARGE-SCALE,
LIGHT-CATALYZED, CHLORINE OXIDATION
83
-------
In order tc determine which radiant-energy source is best suited
to the catalytic oxidation process, it is simply necessary to determine
which source produces the greatest amount of oxidation for the least cost.
However, selection of the best radiant-energy source cannot be made easily
by a trial-and-error basis because there are so many possible sources.
Adequate information is available concerning the relative amounts
of each wavelength emitted by radiation sources which are presently avail-
able. The large-volume costs of these sources can also be readily esti-
mated. The problem then is to determine which wavelengths are most capable
of catalyzing the oxidation.
The usefulness of a given wavelength will depend upon three fac-
tors: (1) the relative catalytic effectiveness of the wavelength compared
to other wavelengths; (2) the ability of the selected wavelength to pene-
trate wastewater; and (3) the cost of producing that wavelength compared
to other wavelengths.
The first two factors can be determined experimentally, and the
third factor must be considered in relation to the various sources avail-
able; the ideal source will be the one which produces the greatest amount
of the most useful wavelengths for the least cost.
Many types of radiation sources are commercially available. Cer-
tain modifications of existing sources are also possible which alter the
wavelength output and useful life of the sources. A somewhat over-simplified
statement of this situation is that shorter lifetimes and higher radiant-
energy costs are encountered if shorter wavelengths or higher intensities
are required.
A thorough review of all potential radiation sources is beyond
the scope of this report. Louis R. Koller in his text entitled Ultra-
violet Radiationjy has provided an excellent review of the various sources
of radiant energy. He devotes 68 pages to the discussion of electrical
arcs and 16 pages to a discussion of incandescent sources. Since incan-
descent sources are relatively weak and inefficient sources of ultraviolet
radiation, we have given consideration only to the electrical arcs.
The most efficient and most useful sources of ultraviolet energy
are the arcs. Arcs are produced when a current of electricity is passed
between electrodes separated by a gas or a vapor. The amount of ultra-
violet radiation available depends upon the nature of the arc.
I/ Koller, L. R., Ultraviolet Radiation, 2nd Edition, John Wiley & Sons,
Inc., New York (1965).
84
-------
There are two main types of arcs: open arcs and closed arcs.
The carbon arc is a typical example of an open arc, and it produces radiant
energy with the arc exposed to the atmosphere. In closed arcs, of which
mercury-vapor lamps are typical examples, electrical discharges through a
gas or vapor are contained in an envelope which is usually of quartz or
glass. Arcs may take place through any desired gas or vapor and at pres-
sures ranging from a few microns up to hundreds of atmospheres. The pres-
sure is limited only by the strength of the envelope. The gases and vapors
include such diverse substances as hydrogen, helium, argon, neon, krypton,
xenon, mercury, cadmium, zinc, tellurium, and magnesium.
The spectrum of most arcs is chiefly due to the discharge through
the gas or vapor and so is usually of the line or band type. In high-
pressure arcs there is also a continuous background over a part of the
spectrum. In some cases, such as the carbon arcs, the incandescent elec-
trodes may also contribute a continuous component to the spectrum, so that
line and continuous spectra may be present in various proportions.
Mercury Arcs
Mercury arcs have been used extensively for the generation of
ultraviolet radiation. There are several reasons for the choice of mercury:
(l) its spectrum is rich in lines of the ultraviolet; (2) it is fairly inert
and does not react with electrode materials; (3) it does not attack glass
appreciably; and (4) its vapor pressure lies in a convenient range for lamps
which operate near room temperature.
A wide variety of mercury arcs is available commercially. They
differ from each other in such respects as operating pressure, bulb shape,
reflector, electrodes, starting mechanisms and other features which suit
them to perform specific tasks. Literally hundreds of different kinds have
been developed. The spectral energy distributions of selected mercury-arc
lamps are shown in Table A-I.
Essentially, mercury arcs are characterized by the mercury vapor
pressure. This parameter of pressure is an important one in determining the
wavelengths and intensity of radiation generated by the discharge.
High-Pressure Mercury Arcs; High-pressure mercury arcs have
proved to be useful sources of ultraviolet, as well as of visible, radiation.
They emit intense bands of ultraviolet radiation between 220 mp, and 366 m^.
One of the fundamental characteristics of high-pressure arcs is the high
temperature of the gas and the thermal equilibrium between the gas and the
electrons. The temperatures range from 5500° to 8000°F. High-pressure
arcs require much greater energy input per unit length than low-pressure arcs.
85
-------
TABLE A-I
CD
CD
SPECTRAL ENERGY DISTRIBUTION (IN W)
OF SELECTED, MERCURY-ARC I,AMF:£/
G.E.
Type of Ordering Far UV Middle UV Near UV Visible
Mercury Arc Abbreviation (220-260 mq) (280-520 mq) (520-400 mq) (400-700 mq)
High-pressure UV-37 252 (22$) 279 (25$) 252 (22$) 347 (31$)
Medium-pressure H1500-A23 189 (26$) 157 (22$) 140 (20$) 227 (32$)
Low-pressure G64T6 18.0 (90$) 0.4 (2$) 0.3 (1.5$) 1.3 (6.5$)
Ploodlamp H400-A34-1 0 (0$) 1.5 (2$) 20.5 (26$) 55.7 (72$)
Sunlamp RS 0.004 (0$) 1.4 (12$) 3.1 (27$) 7.0 (61$)
"Black light" F40BL 0 (0$) 0.33 (3$) 7.3 (72$) 2.6 (25$)
Fluorescent
a/ These data were obtained from several sources provided by the General Electric Company.
-------
They are characterized by a very brilliant, small-diameter arc operating
at a relatively large current density and high voltage. High-pressure
arcs are shorter than low-pressure arcs of the same voltage and also have
a smaller tube diameter. Compared to other mercury arcs, high-pressure
mercury arcs have relatively short, lamp-lifetimes (1,000 hr).
Low-Pressure Mercury Arcs; As the vapor pressure of mercury in
a discharge is reduced, the intensity of the 253.7-mu line (known as the
mercury resonance line) increases markedly. This phenomenon is exploited
in the construction of low-pressure lamps which emit this radiation almost
exclusively.
Lamps of this kind are widely used for germicidal purposes be-
cause they are the cheapest sources of sterilizing radiation. Maximum
germicidal effect is produced by wavelengths near 260 mu, and germicidal
effectiveness drops off sharply at lower and at higher wavelengths; wave-
lengths above 320 mu are almost completely ineffective for germicidal
purposes.
Low-pressure mercury arcs emit much-less-intense radiation than
high-pressure mercury arcs of about the same physical dimensions. However,
the useful life of low-pressure lamps is much longer (7,500 hr). Also,
low-pressure lamps are considerably less expensive than high-pressure lamps
of the same ultraviolet-energy output. Low-pressure lamps also convert
electrical energy to radiant energy more efficiently.
Sunlamps; Ultraviolet lamps which are used for therapy are known
as sunlamps. The term sunlamp, however, has a very specific meaning and
cannot be applied to ultraviolet lamps indiscriminately. The Council of
Physical Therapy of the American Medical Association defines a sunlamp as
a lamp that, at a specified distance, emits ultraviolet radiation not dif-
fering essentially with that of natural sunlight in total intensity. Its
spectral range must extend from about 290 mu to and including 313 mjj,.
Wavelengths shorter than 280 my, are undesirable.
Since all mercury-vapor arcs emit considerable radiation of shorter
wavelength than 280 m^,, it is necessary to filter out this radiation in order
to make them conform to the above definition. This can be done by filter-
ing the light from an intermediate-pressure, quartz, mercury lamp using a
glass which does not transmit radiation below 280 ma. Another alternative
is to make the lamp envelope of a filter material.
"Black Light" Fluorescent Lamps; Low-pressure mercury arcs are
used to produce ultraviolet in the region around 366 mu, popularly known
as "black light." "Black light" is very effective in exciting fluorescence
in many substances. The inside wall of a "black light" lamp is coated with
87
-------
a phosphor which absorbs the 253.7-m.(j, radiation and emits in a broad band
around 366 mp,. Thus, the phosphor transforms 253.7 rap, radiation into
longer-wavelength ultraviolet. The glass used in constructing the tube
does not transmit the short-wavelength ultraviolet but allows the longer
wavelengths to pass through. The phosphor commonly used is a cerium-
activated calcium phosphate.
The output of ultraviolet energy between 320 ma and 380 ran is a
greater percentage of the electrical input for these lamps than for any
other kind of mercury discharge. The percent of visible radiation is
greater than that for germicidal lamps, but is less than that for any of
the sunlamps. Consequently, "black light" fluorescent lamps are very use-
ful for producing fluorescent effects, blue printing and other applications,
They emit practically no radiation of wavelengths shorter than 320 ma.
High-Pressure, Mercury-Arc Floodlamps; Special types of high-
pressure, mercury-arc lamps have been developed for illumination purposes.
Lamps of this kind are characterized by a very brilliant, small-diameter
arc operating at a relatively large current density and high voltage.
These lamps are often called "Capillary Type Lamps" since they are con-
structed of small diameter, heavy-walled, quartz tubes. These tubes are
usually enclosed in an outer jacket of glass which serves a threefold pur-
pose of filter, heat insulator, and shield in case of breakage of the arc
tube. Lamps of this kind are widely used for illumination purposes and
are therefore relatively cheap compared to other types of mercury-arc lamps
which are not used in such large quantities. Because of the nature of
their use, exceptionally long operating lifetimes are required. Lamps of
this kind are available which have lifetimes of 24,000 hr, and greater.
Also as would be expected, most of the radiation (70 to 75$) from these
lamps is in the visible region. However, because of their relatively long
life and low cost, lamps of this kind are relatively inexpensive sources
of ultraviolet radiation.
A comparison of the costs of producing ultraviolet radiation
from various types of mercury-arc lamps is presented in Table A-II.
88
-------
TABLE A -II
COSTS OF PRODUCING ULTRAVIOLET RADIATION
FROM VARIOUS MERCURY -ARC LAMPS^/
Lamp
(G.E. Ordering
Abbreviation)
UA-37
G64T6
H1500-A23
H400-A33-1
F40BL
RS
HP
LP
MP
FL
BL
SL
Lamp
Cost£/
(*)
147.00
14.00
112.50
12.40
2.00
9.00
Daily
Costd/
(40
3.68
0.0448
0.450
0.0124
0.004
0.108
Daily Cost of 100 w
of UV Output
0.470
0.242
0.0927
0.0564
0.0888
2.40
sj The data in this table were obtained from General Electric.
b/ HP = High pressure, LP = Low pressure, MP = Medium pressure, FL =
Floodlamp, BL = "Black light," SL = Sunlamp.
£/ These figures are the large-scale replacement costs of the individual
lamps.
d/ The daily cost is the replacement cost, divided by the lamp life in days,
Carbon Arcs
The carbon arc is a typical example of an open arc . It consists
of a discharge between two carbon electrodes in air at atmospheric pres-
sure. The arc is started by bringing the two electrodes into contact and
then separating them slightly. The resulting discharge is intensely bright
and hot.
During operation, the electrodes are consumed. The positive elec-
trode is burned at the rate of about 2 in/hr in low-intensity arcs and as
much as 12 in/hr in high-intensity arcs. The negative electrode is con-
sumed at a somewhat slower rate . In order to maintain the constancy of
the arc, the electrode must be fed (either by hand or by an automatic de-
vice) in order to maintain a constant arc-gap.
The spectrum of the carbon arc consists of the continuous spec-
trum of the incandescent electrodes on which is superimposed the band
spectrum of the luminous gas. The resulting spectrum is relatively con-
tinuous. The cyanogen bands between 380 mp, and 390 rajj, are a particularly
marked feature of this spectrum and add a pronounced peak in this region.
89
-------
This peak is very noticeable in Figure A-l which depicts the spectral
energy distribution of a typical carbon arc.
Increased ultraviolet output can be produced from carbon arcs
by the addition of certain substances, such as iron, to the carbon core.
Figure A-2 shows the spectral energy distribution of a carbon arc with
iron-cored carbons. This spectrum has many lines ranging from the visible
through the ultraviolet to 230 m^,. The overall ultraviolet output is con-
siderably greater than that produced by ordinary carbon arcs, although the
output of visible radiation is considerably less.
The spectral outputs that can be obtained, using carbon arcs of
different compositions, are shown in Table A-III. Compared to mercury
arcs, carbon arcs are very intense, though not necessarily efficient,
sources of ultraviolet. Their usefulness is based on the fact they are
extremely concentrated sources; that is, a larger amount of radiation can
be produced from a smaller area of radiating surface. However, if effi-
ciency is an important consideration, the carbon arcs are less efficient
in many cases than mercury arcs. For example, the intensity of radiation
between 230 m)j, and 280 mu at a distance of 1 m from a National W carbon
Arc is 1970 p.w/cm2. Since the input to the arc is 4800 w, the "yield" of
radiant energy is 0.41 |j,w/cm /w input. The intensity in the same wave-
length range and at the same distance from a 30-w germicidal lamp is
75 (j,w/cm2, or 2.5 p,w/cm2/w input.
In general, the advantages lie with the carbon arc if a high
intensity is required from a single unit. On the other hand, when effi-
ciency is a consideration, the advantage usually lies with the mercury arc.
The relatively short life of a carbon arc (namely, a few hours
before the carbons must be replaced) is a disadvantage in comparison with
the life of an enclosed arc (namely, several thousand hours). However,
devices are available which can automatically feed carbon rods to an arc
on a 24-hr basis. In some applications, the presence of the combustion
products of the arc is also a disadvantage.
High-Pressure Xenon Arcs
Until about 1950, mercury-vapor arcs were the most widely used
arcs both in therapy and in industry. Since then, there has been a notable
increase in the use of high-pressure xenon arcs, particularly for industrial
and technological applications.
90
-------
4000 5000
Wavelength in Angstroms
Figure A-l - National Sunshine Carbon, with Corex D Filter, Compared with
Natural Sunlight. Solid line, with Corex D filter;
dotted line, natural sunlight.
91
-------
2000 3000 4000 5000 6000 7000
Wavelength in Angstroms
Violet Blue Green Yellow Orange Red
Figure A-2 - Spectral Energy Distribution of Carbon Arc With Iron-Cored
Carbons (National B). Upper curve, 60-amp alternating
current, 50 v across arc. Lower curve, 30-amp alternating
current, 50 v across arc.
92
-------
TABLE A-III
INTENSITY OF RADIATION FROM BARE CARBON ARCS, WITHOUT REFLECTORS, IN
CD
MICROWATTS PER SQUARE CENTIMETER AT DISTANCE OF 1 METER FROM ARC*
Type of
"National"
carbon
"National"
Arc
Sunshine
Sunshine
Current
60
Volts
50
2300-
o
2800 A
129
2800-
O
3800 A
187
3200-
0
3800 A
1010
2300-
O
3800 A
1326
carbon with Co rex D
filter
"National"
"National"
"National"
"National"
"National "
"National"
"National "
B carbon
C carbon
D carbon
E carbon
K carbon
U carbon
W carbon
60
60
60
60
60
60
60
80
50
50
50
50
50
50
50
60
2
684
613
80
221
709
746
1970
103
326
413
40
130
262
589
1627
860
1010
1188
296
584
1310
1000
2620
965
2020
2214
416
935
2281
2335
6217
3800-
7000 A
6094
5445
1910
1656
1819
5665
1529
2148
6223
Calculated from National Carbon Company Catalog, Section A-4300.
-------
Xenon arcs, between tungsten electrodes operating at pressures
in the 10- to 30-atm range, are efficient sources of intense visible and
ultraviolet radiation. At these pressures, the xenon spectrum becomes a
continuum extending from the ultraviolet into the visible and then the
infrared. Commercial xenon lamps range in size from 150 to 20,000 w. A
typical spectrum of this type of lamp is shown in Figure A-3. The spectrum
varies somewhat with input to the lamp and with the operating pressure;
however, the general character is not affected. The brightness of these
lamps is comparable to that of the carbon arcs.
Xenon arcs are presently replacing carbon arcs in certain applica-
tions, particularly lithographic processes. The cost of xenon arcs is
usually four to five times the cost of comparable carbon arcs. However,
because of their convenience, and especially because of their cleanliness,
lithographers are in the process of converting to these types of lamps.
It should be mentioned, however, that the cost of illumination in litho-
graphic processes represents a very small fraction of the overall operating
costs.
General Considerations in Arc Selection
The choice of an arc for a particular application depends upon
several factors. Foremost is the wavelength of the desired radiation.
Other factors are the presence of absence of other wavelengths, plus total
yield, efficiency and intensity of the source. The total yield depends
upon the kind of lamp and the wattage. For a given type, the total yield
may sometimes be increased by employing a higher wattage. The yield, how-
ever, is quite distinct from the efficiency. The efficiency depends upon
the kind of lamp; thus the efficiency of production of 253.7-mp, radiation
is about 24$ for a 30-w, low-pressure, germicidal lamp; while it is 3$ for
a high-pressure, mercury-arc lamp. In spite of the higher efficiency, the
total output of 253.7-m|j, radiation from the germicidal lamp is only one-
fourth as great as from the high-pressure lamp. The radiant flux, per unit
area, produced by the germicidal lamp is the lowest of any of the mercury-
arc sources. Extremely high radiant flux, per unit area, is produced by
high-pressure, mercury-arc lamps because of their relatively small dimensions,
The germicidal lamp emits very little radiation other than the
253.7-mu radiation. However, the high-pressure mercury lamp is a rich
source of longer-wave ultraviolet, visible and infrared radiation. High-
pressure xenon arcs, which have practically a continuous spectrum, radiate
only about 9$ of their energy in the region below 400 mM*.
The selection of the ideal, radiant-energy source depends upon
many factors. Where small, compact sources are required, high-pressure
lamps or carbon arcs are favored. Where efficiency is a consideration in
order to avoid unnecessary heat loads, other types of lamps are required.
94
-------
400
350
300
'250
'200
1150
100 -
50 -
0.2
0.4
0.6
0.8 1.0 1.2 1.4
Wavelength-microns
1.6
20
Figure A-3 - Spectral Energy Distribution of Typical 5-KW Xenon
Lamp. Electrode radiation excluded.
95
-------
Ir. ~ther cases, the only consideration may be to obtain the maximum output
it. the desired range, regardless of the number of units or their size or
their efficiency. According to Koller,i/ "No general selection rules can
be given."
Cur, light
Sunlight is rich in ultraviolet wavelengths known to be useful
for catalyzing oxidations by chlorine. Table A-IV shows the intensity of
illumination in the ultraviolet and visible regions produced by sunlight
during midday on a typical clear day in midsummer. Maximum, total, radiant
energy, within specific wavelength ranges under the same conditions, is
shown in Table A-V.
The amount of radiant energy available in. the ultraviolet and
visible regions compares favorably with the energy produced by mercury-arc
or carbon-arc lamps, although no radiant energy below 300 rap, is available
from sunlight.
Although obviously the cheapest source of radiant energy, sun-
light presents the troublesome problem of periodic, as well as random,
fluctuations in occurrence and intensity. However, even if use of sunlight
were technically feasible, standby sources of radiation would probably be re-
quired to prevent unusually large accumulations of untreated or partially
treated wastewater during periods of greatly decreased illumination.
I/ Roller, L. R., Ultraviolet Radiation, 2nd Edition, John Wiley & Sons,
Inc., New York (1965).
-------
TABLE A-IV
SOLAR ENERGY DISTRIBUTION*
Wavelength Intensity of Illumination*
_ (u.w/cm2) _
300 5.2
310 47.5
320 125
330 204
340 233
355 259
370 325
385 333
400 433
410 548
420 600
430 617
440 627
450 669
500 718
550 689
600 621
650 556
700 549
Intensities are calculated assuming a horizontal surface during midday on
a typical clear day in midsummer in Cleveland, Ohio; latitude 41.5°N.
Intensities are calculated for each spectral band of 5 m\s, centered at
the various wavelengths. From "Germicidal, Erythemal and Infrared
Energy," by M. Luckiesh, D. Van Nostrand Company, New York, 1946, p. 48.
TABLE A-V
MAXIMUM INTENSITY OF SOLAR ENERGY*
Radiant Energy
Wavelength Range (ixw/cm )
Shorter than 350 mp, 1,180
350-400 mp, 2,360
400-700 mp, 42,000
Longer than 700 mp, 40,800
TOTAL 86,340
* Intensities are calculated assuming a horizontal surface.during midday
on a typical clear day in midsummer in Cleveland, Ohio; latitude 41.5°N.
From "Germicidal, Erythemal and Infrared Energy," M. Luckiesh, D. Van
Nostrand Company, New York, 1946, p. 35.
97
-------
APPENDIX B
SELECTION OF TECHNIQUES AMD EQUIPMENT
FOR WAVELENGTH-EFFECT STUDIES
98
-------
Prior to the wavelength-effect studies, there were a number of
alternatives to consider in the selection of the monochromatic radiation
source, the design of the reaction vessel and the method of determining
the amount of radiation absorbed. These alternatives are discussed below.
Selection of Monochromatic Radiation Source: There are two
general methods for providing monochromatic, or nearly monochromatic, radia-
tion for studies of this kind. The most direct procedure is to employ a
monochrometer. A monochrometer is an instrument which separates individual
wavelengths, being emitted by polychromatic sources, by means of a prism or
grating. All ultraviolet spectrophotometers are equipped with monochrometers;
however, these monochromatic sources are far too feeble to be of practical
value in photochemistry. Monochrometers which produce fairly intense mono-
chromatic radiation are commercially available; however, these are very ex-
pensive and are not common laboratory equipment. Most photochemical data
involving the use of a monochrometer have been supplied by investigators who
have constructed their own monochrometer- The major problem is to produce
a monochromatic source of sufficiently high intensity. Another problem is
that the rectangularly shaped image of the slit produces a radiation beam
which is not compatible with the reaction vessel and is not homogeneous
in intensity.
Another method for producing nearly monochromatic radiation is
to use filters. There are three principal types of filters: chemical
filters, glass filters, and interference filters. Chemical filters are
simply solutions of certain chemical compounds of known spectral trans-
mission. Many filters of this kind have been used for various photochemical
applications.
Glass filters are the most familiar type of optical filter, and
glass filters are available which will cleanly separate any desired visible
wavelength from a medium-pressure mercury arc. Unfortunately, no such
series of glass filters have been developed for the ultraviolet region.
Interference filters consist of two, semitransparent, evaporated
metal films on glass plates which are separated by a transparent layer of
a thickness comparable to a particular, desired wavelength of light. The
filter reflects all incident light except for that characteristic band.
Interference filters provide several advantages over other types
of filters. Besides allowing a relatively short, optical path between
source and reactor cell, interference filters do not become warm during
use because practically all of the radiation which is not transmitted is
reflected. Interference filters are stable at elevated temperatures
(108°F), and they cannot fade or change their transmission properties since
the nature of their filtering action does not depend on light absorption.
99
-------
A major factor in the ultimate decision concerning experimental
techniques employed in the wavelength-effect studies was the very recent
availability of interference filters for the ultraviolet region. A series
of interference filters was obtained which permitted very effective isola-
tion of the major wavelengths available in a high-pressure mercury arc.
An interference filter capable of isolating the important line
of mercury-arc radiation at 253.7 mg, was not readily available. To produce
monochromatic radiation of this wavelength, a low-pressure mercury arc was
employed. Data supplied by the manufacturer of this lamp (Nester/Faust)
indicate that 96$ of the UV output of this lamp is in the range, 245 to 260
m\i, with a. characteristically sharp maximum at 253.7 rmj,.
Photochemical Reactor Cells; The most common and functional de-
sign of reactor cells for quantitative liquid-phase studies is a cylindri-
cal cell with optically flat, circular windows fused to the cell body at
right angles to its axis. The size of the cylinder and windows should be '•
consistent with the dimensions of the light beam used in the experiments.
There should be as little unradiated cell volume as possible, so that there
can be no question about the actual effective volume of the cell.
In quantitative photochemical work it is desirable to use flat
windows and to have nearly perpendicular incidence of the light beam on
the window. Under these conditions, the fraction of light reflected at
the window interface is a minimum. If a curved surface, such as the wall
of a test tube or flask, is placed in a homogeneous light beam, much of the
incident radiation will be lost, and the distribution of light within the
vessel will be very nonhomogeneous.
For our purposes, the most convenient, photochemical reactor cells
were the quartz, cylindrically shaped cells used to hold samples for ultra-
violet spectrum determinations. Cells of this kind have flat, perpendicular,
quartz windows, and cells of 1, 5, and 10 cm in length were readily avail-
able. The diameter of these cells is about 2 cm, a size which permitted
uniform irradiation of the cell by the radiation transmitted by the inter-
ference filters. A further convenience of these, cells results from the
fact that the ultraviolet spectrum of the substance could be determined be-
fore and after the experiment. These spectral data were required in order
to determine the amount of radiation absorbed by the chlorinated effluent;
the data were also useful in determining the extent of organic oxidation and
in determining the effective depth of penetration of various wavelengths.
Actinometry; In order to compare the effect of radiation at one
wavelength with the effect of radiation of another wavelength, it is
necessary to know the intensity of the radiation at both wavelengths. In
100
-------
the wavelength-effect studies, it was necessary to determine accurately
the intensity of the radiant energy impinging upon the photochemical reactor.
In general, intensity is detected by physical or chemical methods. Devices
which employ physical methods include thermopiles, bolometers, radiometers,
phototubes, photoconductive cells, and photovoltaic cells. The advantages
and disadvantages of these detectors are discussed in considerable detail
by Roller .i/
In our opinion, actinometry (the determination of the intensity
of radiant energy by chemical methods) was the best method for solving this
particular problem.
An actinometer is a radiation-measuring, chemical device which
consists of chemicals that undergo a visible, or easily measurable, change
on exposure to radiation. There exists a very useful and accurate actinometer
which is based upon the rate of photochemical decomposition of oxalic acid in
the presence of uranyl sulfate. The amount .of oxalic acid decomposed is
determined by titration with permanganate.—'
A more sensitive actinometer is based on the photolysis of ferri-
oxalate. This method depends upon the photometric determination of the
photolysis products.V it is hundreds of times more sensitive than the
uranyl-oxalate method. The minimum detectable energy is 3 x lO-1-^ quanta.
This method can be used to investigate weak sources without making unduly
long exposures. The quantum yield is of the order of unity at all wave-
lengths up to 510 m(j,.
The quantum yields of these two actinometer systems are presented
in Figure B-l.
Batch-Recycle Reactor Versus Cell Reactor: Consideration was
given to the application of a batch-recycle reactor to the wavelength-
effect studies. Dr. J. M. Smith, Chairman of the Department of Chemical
Engineering, University of California, Davis, California, pointed out that
a batch-recycle reactor provided a number of advantages for a relatively slow,
photochemical reaction like chlorine oxidation.
I/ Koller, L. R., Ultraviolet Radiation, 2nd Edition, John Wiley & Sons,
Inc., New York (1965).
2/ Forbes, G. S.,and Heidt, L., J. Am. Chem. Soc., 56^2363 (1934).
3/ Hatchard, C. G.,and Parker, C. A., Proc. Roy. Soc. (London), A235, 518
(1956).
101
-------
1.30
1.20
1.10
1.00
0.90
* 0.80
9
LU
> 0.70
=>
Z 0.60
D
° 0.50
0.40
0.30
0.20
0.10 -
0.006 M K3Fe(C2O4)3
A ^^.A A
•
-------
Essentially, a batch-recycle reactor of the type used "by Dr. Smith
is a vertical quartz tube through vhich the reaction mixture flows. The
mixture is pumped from a reservoir through the irradiated, quartz-tube reactor
and is returned to the reservoir. The reactor tube is located at one focus
of an elliptical reflector system, and the radiation source is located at the
other focus of the system. A series of concentric filters are located around
the radiation source to permit variation in the wavelength and intensity of
the radiant energy.
In the batch-recycle reactor system, the optical purity of the
incident radiation is far from monochromatic, but the loss of monochromatic
character is offset by the increased intensity of the available radiation.
The procedure of using a batch (i.e., nonrecycle) reactor, such as
the photochemical reactor cell, for relatively long periods of time (in order
to produce detectable changes in concentration) can cause inaccuracies be-
cause of the lack of mixing. In the absence of mixing there are uncertain-
ties because of temperature and concentration gradients. However, accord-
ing to Dr. Smith, no method has been developed for mixing in a photochemical
batch reactor without influencing the radiation path. Since earlier studies^/
had shown (l) that the UV-chlorine oxidation is not affected by temperature
and (2) that the rate of oxidation is not proportional to chlorine concentra-
tion, it was felt that the lack of mixing would not be a serious problem.
Another disadvantage of the small reactor cell is that the pH
must be controlled by means of buffers, not by the intermittent addition
of sodium hydroxide, as the reaction proceeds. Therefore, it was necessary
to perform experiments in a large reactor in order to demonstrate that there
was no significant difference between the results obtained with buffers and
the results obtained by maintaining pH by alkali addition.
A further disadvantage in using a system in which interference
filters are employed to provide monochromatic radiation is the small volume
of reaction mixture employed. Because the size of the reaction vessel is
limited to a cell about 1 in. in diameter and 1 to 5 cm in depth, determina-
tions of chemical oxygen demand (COD) were not possible because of the
small volumes of samples available.
Because of the small size of a cell reactor, chlorine cannot be
added gradually and the required amount of chlorine must be added at the
4/ "An Investigation of Light-Catalyzed Chlorine Oxidation for Treatment
~~ of Wastewater," Final Summary Report, FWPCA Contract No. 14-12-72,
December 1968.
103
-------
"beginning of the experiment. High chlorine concentrations were necessary
"because relatively large decreases in TOC were required for precise deter-
minations of changes in organic concentration. However, high chlorine con-
centrations were not considered to "be a major problem "because the original
work on UV-chlorine oxidation had shown that chlorine concentration did not
affect the rate of TOC or COD decrease.
Thus, it was felt that reliable data concerning the effect of
wavelength could be obtained using the cell reactor in spite of the necessity
for relatively high chlorine concentrations.
In summary; there are advantages and disadvantages to both types
of reactor systems, and these are summarized below:
2.
3.
Batch-Recycle Reactor
Advantages
Permits large volumes of
reaction mixture; thus, COD
as well as TOC determinations
can be performed.
Permits pH control by alkali
addition.
Avoids problems of unmixed
reactor -
Disadvantages
1. Cannot utilize monochromatic
radiation.
2. Actinometer results are only
approximate.
3. Reactor is relatively
elaborate to construct.
4. More like ultimate reactor system.
5. Can utilize relatively intense
radiation at reactor area.
Cell Reactor
Advantages
1. Can utilize truly monochromatic
radiation.
2. Can use actinometer precisely.
Di sadvantages
1. Buffers must be used for pH
control.
2. Snail volume of reactor pre-
cludes COD determinations.
104
-------
3. Can obtain quantitative wave- 3. Cannot be stirred since that
length effects. might cause temperature and
concentration gradients.
4. Is relatively simple to operate
(using interference filters). 4. Incident radiation is rela-
tively feeble; therefore,
long reaction times may
be required.
The final decision was to construct a cell reactor for the wave-
length-effect studies and to employ a batch-recycle reactor to obtain other
reactor-design factors.
105
-------
Accession Number
Subject Field & Group
05D
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
Title
LIGHT-CATALYZED CHLORINE OXIDATION
FOR TREATMENT OF WASTEWATER
•J Q Authors)
Alfred F. Meiners
16
21
Project Designation
Program No. 17020DUE09/70
Contract No. 14-12-531
Note
22
Citation
Contractor's Final Report
23
Descriptors (Starred First)
Aerobic Treatment, Environmental Engineering, Oxygen Demand, Oxygen Requirements,
Pollution Abatement, Waste Assimilative Capacity, Water Pollution Control,
Colorination*, Ultraviolet Light Catalysis*, Secondary Effluent Treatment*
25
Identifiers (Starred First)
Chlorine, Light Catalysis
Abstract
97
1 Application of light-catalyzed chlorine oxidation to the treatment of effluents
from secondary waste-treatment plants was studied. Wavelength effects, intensity-time
relationships, quantum efficiencies, chlorine concentration and the effect of dissolved
oxygen were studied in both a small batch reactor and a large batch-recycle reactor.
For optimum organic oxidation rate, optimum utilization of radiant energy and
most efficient use of chlorine, (l) low intensity, short-wavelength radiation should be
employed, and (2) the chlorine concentration should be at a minimum (less than 5 mg/liter)
A plant for processing 10 million gallons of effluent per day was designed.
Estimated costs for construction and operation of this plant totaled 11.77^/1,000 gal.
Results to date indicate that raw materials costs (chlorine and caustic) can be reduced,
and that actual quantum efficiencies will be higher than the conservative efficiency
upon which this estimate is based.
Abstractor Alfred F< Meiners
Institution
Midwest Research Institute
WR:102 (REV JULY 19691
WRSI C
SEND TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C 20240
* 6PO: 1989-359-339
------- |