Citation
Autoxidation of industrial wastes with oxygen

Material Information

Title:
Autoxidation of industrial wastes with oxygen
Creator:
Riley, Boyd T., 1939-
Publication Date:
Copyright Date:
1966
Language:
English
Physical Description:
xi, 179 leaves. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Calibration ( jstor )
Chemicals ( jstor )
Free radicals ( jstor )
Heaters ( jstor )
Liquids ( jstor )
Oxidation ( jstor )
Oxygen ( jstor )
Pressure ( jstor )
Reaction kinetics ( jstor )
Steam ( jstor )
Dissertations, Academic -- Environmental Engineering Sciences -- UF
Distillation, Destructive ( lcsh )
Environmental Engineering Sciences thesis Ph. D
Factory and trade waste ( lcsh )
Oxidation, Physiological ( lcsh )
Pyrodigneous acid ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis -- University of Florida.
Bibliography:
Bibliography: leaves 171-178.
Additional Physical Form:
Also available on World Wide Web
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
022045606 ( AlephBibNum )
13433356 ( OCLC )
ACY4480 ( NOTIS )

Downloads

This item has the following downloads:


Full Text








AUTOXIDATION OF

WITH


INDUSTRIAL WASTES

OXYGEN


BOYD T. RILEY, JR.















A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY











UNIVERSITY OF FLORIDA


June, 1966












ACKNOWLEDGMENTS


For their guidance in the preparation of this disserta-

tion, the author wishes to express his sincere appreciation

to the members of his supervisory committees

Professor J. E. Kiker, Jr., chairman; Dr. C. I. Harding,

co-chairman; Dr. H. D. Putnam; and Dr. G. B. Butler.

The cooperation and advice extended by Dr. R. S. Sholtes,

Dr. W. H. Morgan, and Dr. H. A. Bevis greatly assisted

in preparing the experimental equipment. The aid of

Mr. H. McGraw in fabricating special fittings is noted in

a similar manner.

Finally, the author wishes to thank his wife for her

patience and encouragement, as well as her participation

in both the experimental work and manuscript preparation

for this dissertation.

The graduate training leading up to this study was

supported in part by U. S. Public Health Service Training

Grant 8T1-ES-11-04.

This study was supported financially by Cabot

Corporation.












TABLE OF CONTENTS


ACKNOWLEDGMENTS.... ............. ....... .... ..... .

LIST OF TABLES..................................

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

ABSTRACT...... .................... .. ...............

CHAPTER

I. INTRODUCTION ... ......... ............. ...

II. REVIEW OF THE ZIMPRO PROCESS..............

III. INTRODUCTION OF THE THEORY OF AUTOXIDATION
OF ORGANIC SUBSTANCES..........o......

IV. MATERIALS AND TECHNIQUES FOR THE
EVALUATION OF PARTIAL AUTOXIDATION OF
WOOD DISTILLATION AS PRETREATMENT
PROCESS. .... ...... .. .. ... .... .... ..

V. RESULTS OF THE INVESTIGATION OF PARTIAL
AUTOXIDATION AS A PRETREATMENT
PROCESS.. .. ... .. ........ ............

VI. CONCLUSIONS FROM THE STUDY OF PARTIAL
AUTOXIDATION AS A PRETREATMENT
PROCESS.... .. .. .... . ... .... ......

VII. EQUIPMENT FOR THE INVESTIGATION OF
COMPLETE AUTOXIDATION OF WASTES........

VIII. TECHNIQUES AND RESULTS OF THE STUDY OF
AUTOXIDATION OF WASTES...............

IX. CONCLUSIONS FROM THE STUDY OF AUTOXIDATION
OF WASTES... ...... ...... ......... ....


Page

ii

v

vi

ix



1

14


21




34



47



63


65


77


127


APPENDICES

1. CALIBRATION OF PRESSURE AND TEMPERATURE
SENSOR SYSTEMS......................... 131

iii






APPENDICES Page

2. COD TECHNIQUE.... ........ ................ 136

3. SOLUBILITY OF OXYGEN IN WATER............ 137

4. PARTIAL PRESSURE OF STEAM AS TEMPERATURE
VARIES............. .......... ......... 140

5. MAXIMUM THEORETICAL PARTIAL PRESSURE OF
CARBON DIOXIDE........................ 142

6. CALCULATION FOR COMPUTATION OF OXYGEN
PARTIAL PRESSURE IN AN EXPERIMENT..... 143

7. RESULTS OF TRIPLICATE COD DETERMINATIONS
ON RUN 7.............................. 144

8. TRACE ELEMENT FEED SCHEDULE FOR TRICKLING
FILTER........ ........... .... ... .... 145

9. CALIBRATION OF PRESSURE AND TEMPERATURE
SENSOR SYSTEMS FOR PARR REACTOR....... 146

10. INFORMATION ON GAS CHROMATOGRAPH
ANALYSES............ ..... ............. 149

11. STATISTICAL COMPARISON OF COD
DETERMINATIONS............ ............ 153

12. PRELIMINARY DESIGN AND ECONOMIC-
FEASIBILITY STUDY OF AUTOXIDATION AS A
WASTE TREATMENT PROCESS FOR
PYROLIGNEOUS ACID.................... 156

13. RECOMMENDATIONS FOR FUTURE RESEARCH...... 168

LIST OF REFERENCES................................ 171

BIOGRAPHICAL SKETCH............................... 179













LIST OF TABLES


Table Page

1. Summary of Acid Water Characteristics............ 4

2. Compounds Formed by Wood Carbonization............ 6

3. Autoxidation Properties of Several Materials..... 18












LIST OF FIGURES


Figure Page

1. Flow Diagram of Wood Distillation Process......... 2

2. Dilutions of Neutralized and Un-neutralized Acid
Water from Cabot Corporation .................. 5

3. Typical Zimpro Process........................... 15
4. Autoxidation Mechanisms....................... ... 23

5. Exterior View of 67-Liter Reactor ................ 36

6. Interior View of Reactor Top..................... 37

7. Trickling Filter................................. 44

8. COD vs. Time, Huns 2 and 3........................ 51

9. COD vs. Time, Run 4...................... ......... 52

10. COD vs. Time, Runs 5, 6, and 7.................... 53

11. COD vs. Time, Runs 8, 13, 14, and 15.............. 54

12. COD vs. Time, Runs 9, 16, and 17................ 56

13. COD vs. Time, Runs 10, 11, and 12................ 57

14. Typical Samples From Autoxidation Experiments..... 59

15. Trickling Filter Data............................ 61

16. Assembled Parr Reactor....................... ....... 70

17. Disassembled Parr Reactor....................... 71

18. Sampling Tube on Reactor.......................... 72

19. Experimental Equipment............................ 76

20. COD vs. Time, Runs 26, 32, and 38 ................ 80

21. COD vs. Time, Runs 44 and 49...................... 82






Figure Page

22. COD vs. Time, Runs 27, 33, and 39................ 83

23. COD vs. Time, Runs 45 and 50..................... 84

24. COD vs. Time, Runs 28, 34, and 40................ 85

25. COD vs. Time, Huns 46 and 51 .................... 86

26. COD vs. Time, Runs 29, 35, and 41................

27. COD vs. Time, Huns 30, 47, and 52................ 89

28. COD vs. ime, .uns 36, 42, and 48................ 90

29. COD vs. Time, Runs 31, 37, and 43................ 91

30. Samples from Autoxidation Reactions.............. 93

31. COD vs. Time, Run 54 ............................. 96

32. COD vs. Time, Run 55.............. ..... .......... 97

33. COD vs. Time, Run 57............................. 99

34. COD vs. Time, Run 5 .............. ... ......... .. 1 0

35. COD vs. Time, Run 59 ......................... ... 101

36. COD vs. Time, jun 69 ....................... 103

37. COD vs. Time, Run 70 .......................... ..1 4

38. COD vs. Time, Run 63............................. 1 6

39. COD vs. Time, .un 64 ............. ... ..... .... 107
40. COD vs. Time, Run 65............................. 1

41. COD vs. Time, Run 66.......................... 110

42. COD vs. Time, Run 67 ............................. 111

43. COD vs. Time, Run 6 .. ................ ........... . 112

44. COD vs. Time, Run 71............. .... ............. 13

45. COD vs. Time, Run 72............................ 116

46. COD vs. Time, Run 73.............................. 118

47. COD vs. Time, Run 74 .............................. 119

48. COD vs. Time, Run 75 ............................. 120

vil






Figure Page

49. COD vs. Time, Run 76 ............................ 122

50. COD vs. Time, Run 78............................123

51. COD vs. Time, Run 79................ ............124

52. COD vs. Time Run 77 ............................ 126

53. Circuit Diagram for Pressure-Sensing System....132

54. Calibration of Pressure Transducer-Recorder
System ...................................... 133

55. Circuit Diagram of Thermocouple-Aecorder
System.......................... ........ .. 134

56. Calibration of Thermocouple-Recorder System..... 135

57. Solubility of Oxygen in Water.................138

58. Solubility of Oxygen in Water...................139

59. Saturated Steam Partial Pressure vs.
Temperature... ................ ...... ......... 141

60. Calibration of Thermocouple-Recorder Against
Mercury Thermometer.......................... 147

61. Calibration of Pressure Transducer..............148

62. Typical Gas Peaks Obtained From An Analysis.....150

63. Calibration Curves for Oxygen, Nitrogen, and
Carbon Dioxide...............................151

64. F. & M. Model 720 Gas Chromatograph...............152

65. Extrapolation of Reaction Data................... 163

66. Flow Diagram of Autoxidation Waste Treatment
Process.............................. .......... 164

67. Schematic Plot Plan of Autoxidation Waste
Treatment Process.............................165

68. Schematic Profile of Autoxidation Waste
Treatment Process........................... .166

69. Design of Chemical Reactor......................167


viii













Abstract of Dissertation Presented to the Graduate
Council in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy


AUTOXIDATION OF INDUSTRIAL WASTE
WITH OXYGEN

By

Boyd T. Riley, Jr.

June, 1966

Chairman: Professor J. E. Kiker, Jr.
Co-Chairman: Dr. C. I. Harding
Major Department: Bioenvironmental Engineering

Destructive distillation of wood for charcoal manu-

facture produces a liquid waste known as pyroligneous acid,

which is driven out of the wood during the carbonizing

process. Pyroligneous acid is composed of many different

organic substances which are dissolved or dispersed in

the water. Pyroligneous acid has a Chemical Oxygen Demand

(COD) of about 100,000 mg/1, a pH of about 2.5; it is

amber in color, and toxic to microflora. A 1:100 dilution

of pyroligneous acid is necessary before it can be used

as a biological substrate. Such a high dilution makes

waste treatment by chemical rather than biological means a

desirable alternative. Many different chemical treatments,

among which are coagulation, chlorination, and pH adjust-

ment, have been attempted in the laboratory. These

experiments were completely unsuccessful.







After reviewing the information available on pyro-

ligneous acid, autoxidation as a treatment process was

considered. Many of the compounds found in this acid are

toxic to microflora and act as inhibitors of autoxidation

reactions. Pretreatment of the pyroligneous acid by partial

autoxidation might selectively reduce the toxicity of the

waste. If so, this pretreatment could be followed by

conventional biological treatment.

A 67-liter chemical reactor was used to autoxidize

partially the pyroligneous acid at about 1250C and the

equivalent of 100 psi at 0C oxygen partial pressure.

This type of treatment yielded about a 10 per cent COD

reduction in the waste during a 4-hour period. The toxicity

of the treated waste was measured by determining the loading

necessary for optimum response on a laboratory-scale

trickling filter. Results indicated that the toxic

compounds were not selectively oxidized.

Using a bench-scale reactor, complete treatment of

pyroligneous acid by autoxidation was investigated and

found to be technically feasible. Minor variations in

experimental technique were found to cause large variations

in the rate and order of the reaction. The most rapid

reaction rates were obtained when undiluted pyroligneous

acid was rapidly mixed with high pressure oxygen at 1750C

and the equivalent of 150 psi at 0C oxygen partial pressure

Two rates were observable a very fast rate during the first

16 minutes of the experiment which resulted in an 80 per
cent reduction in the COD of the waste, and a slower rate

x






which lasted for the duration of the experiment (1 hour)

and resulted in a total COD reduction of 87 per cent. The

reaction was strongly exothermic during the first minutes

of the experiment, resulting in a 500C temperature rise

of the reactants. Changes in the temperature at which the

reactions were initiated caused large changes in the

reaction rates, with increases in temperature causing

greater than linear increases in reaction rates. The

reaction rates accelerated with increases of the oxygen

partial pressure at which the reactions were initiated.

Partial pressures of 50 psi and less reduced the reaction

rates. Complete mixing was necessary to maintain oxygen

in solution at any partial pressure.

Two other wastes and two pure compounds were also

investigated, but were not found to autoxidize as rapidly

as pyroligneous acid.

A preliminary design was proposed for a 15,000 gal/day

pyroligneous acid treatment plant. An economic-feasibility

study, based on this design yielded a cost of 1.4~/lb. of

COD removed.












CHAPTER I

INTRODUCTION

Description of the Waste


The development of a technically and economically

feasible process for the treatment of pyroligneous acid,

a waste product of the wood distillation industry poses

a challenging problem for the Environmental Engineer.

This dissertation suggests a solution to that problem.

The research conducted in the preparation of this disser-

tation was confined to the treatment of waste produced

by Cabot Corporation's Gainesville, Florida, plant. This

waste results from Cabot's use of the retort operation in

the processing of coniferous woods.

The retort operation is a batch process of destructive

distillation. (See Figure 1.) The stump wood is placed

into rectangular slat-sided iron crates which are mounted

on railcars. A train of several crates is assembled and

moved into a retort by a small engine. After the retort

is sealed, it is heated by oil burners to about 8000F, with

the atmosphere inside the retort kept under reducing

conditions. Essentially all of the liquids present in

the wood are driven out as the wood is carbonized. After

the wood has remained in the retort for 8 hours, the hot

retort is opened and the charcoal is moved to cooling

1


















Hot Charcoal




Cooling
Chamber



Cold Charcoal



Classification


Small Piece
of Charcoal


Liquids


Steam
Distillation


Oil
Sales


s


Sales of
Large Pieces
of Charcoal


Briquet
Sales


Flow Diagram of Wood Distillation Process


Figure 1.







chambers which are similar to retorts in construction;

the hot charcoal ignites as soon as it comes into contact

with oxygen. After the cooling chamber is sealed, the

enclosed oxygen is used up and the flames disappear. The

charcoal cools for 8 hours, after which it is removed from

the cooling chambers and stored. After size classification,

the large pieces are sold as raw material for the production

of activated carbon, and the small pieces are pulverized,

mixed with starch, and pressed into briquets.

The vapors pass out of the retort and through con-

densers. The liquids are pumped to unlined settling lagoons

where the heavy tars and oils are separated from the

aqueous portion of the liquid. The concentrated oils and

tars are pumped from the bottom of the settling lagoons

to a steam distillation column where they are further

fractionated, barreled, and sold. The aqueous portion,

known as acid water or pyroligneous acid by the industry,

seeps through the sandy walls of the lagoons and becomes

surface drainage.

Acid water is either bacteriostatic or toxic, depending

on the circumstances. Veijolo and Mustakallo (1) separated

wood tars into various gross fractions by solvent extractiorn

and tested the effects of these fractions on biological

systems. They concluded that the resin acid fraction was

primarily responsible for the bacteriostasis, and that

these resin acids were effective in either the acid or







salt form. The following table summarizes a few of the

characteristics of acid water as sampled at Cabot Corpora-

tion.


Table 1

Summary of Acid Water Characteristics

Color Dark Amber

COD 100,000 mg/l

Iron 73 mg/l

Orthophosphate 490 mg/1

Hardness 1,500 mg/1 as CaCO3

Phenol 750 mg/1

pH 2-3


Figure 2 shows dilutions of both neutralized and

un-neutralized acid water, also from Cabot Corporation.

Major quantitative differences have been reported

in the composition of acid water obtained from the

destructive distillation of coniferous woods as compared

with acid water obtained from the destructive distillation

of deciduous woods. The qualitative analyses of both

types of acid water are similar (2). Wise and Jahn (3)

have compiled a table from several sources listing 213

different organic compounds which have been isolated

from pyroligneous acid. The names of all the compounds

listed by Wise and Jahn are presented in Table 2.













































Figure 2.


Dilutions of Neutralized and Un-neutralized
Acid Water from Cabot Corporation






Table 2


Compounds Formed by Wood Carbonization


Ammonia Formaldehyde
Formic acid Methanol
Methylamine Glyoxal
Acetaldehyde Acetic acid
Methyl format Glycolaldehyde
Glycolic acid Trimethylamine
Ethyl alcohol Dimethylamine
Acrylic acid Acetone
Allyl alcohol Propionaldehyde
Methyl acetate Hydroxypropanone
Propionic acid Isopropyl alcohol
Propyl alcohol Methylal
Succinic anhydride Furan
Crotonaldehyde Biacetyl
3-Butenoic acid Butyrolactone
Crotonic acid Methacrylic acid
Methacrylic acid, polymer 2-Butanone
2-Buten-1-ol Butyric acid
Isobutyric acid Methyl Propionate
Isobutyl alcohol Acetaldehyde dimethyl acetal
1,2-Dimethoxyethane 2-Furfuraldehyde
2-Furoic acid Pyridine
2-Cyclopenten-1-one 2-Methylfuran
3-Methylfuran Furfuryl alcohol
Cyclopentanone Tiglaldehyde
Angelic acid Methyl crotonate
2,3-Pentanedione 2-Pentenoic acid
3-Pentenoic acid Tiglic acid
e-Hydroxyvalerolactone -Hydroxyvalerolactone
Levulinic acid Acetoxypropanone
ca-Hydroxy-y-hydroxyvalerolactone
Pentene 3-Methyl-2-butanone
Isovaleraldehyde 2-Pentanone
3-Pentanone Pivalaldehyde
Valeraldehyde c-Methylbutyric acid
Methyl butyrate 1,2-Cyclopentanediol
Methyl isobutyrate Isovaleric acid
4-Hydroxy-2-pentanone Valeric acid
Pentane Isoamyl alcohol
Propionaldehyde dimethyl acetal
Benzene Phenol
5-Methyl-2-furfuraldehyde 2-Furyl methyl ketone
Catechol 5-(Hydroxymethyl)-2-
furfuraldehyde
4-Methyl 2-furoic acid Methyl 2-furoate
Maltol Pyrocinchonic anhydride
Pyrogallol 2-Picolene
2-Methyl-2-cyclopenten-l-one 4-Methyl-2-cyclopenten-1-one






Table 2 (Continued)

2,5-Dimethylfuran 1-(3 or 4)-Cyclohexandione
2-Hydroxy-3-methyl-2-cyclopenten-l-one
Cyclohexanone 3-Hexen-2-one
Mesltyl oxide l-Hydroxy-2-butanone acetate
2,3-Hexanedione 4-Methyl-2-pentenolc acid
Hydroxy-2-propanone propionate Levoglucosan
Cyclohexanol Tetrahydro-2,5-dimethylfuran
3-Hexanone Butyl methyl ketone
2-Methyl-3-pentanone Caproic acid
Isocaprolc acid o-Methylvalerlc acid
Methyl valerate Toluene
o-Cresol m-Cresol
p-Cresol Gualacol
l-Methoxy-2,3-dlhydroxybenzene 5-Methyl pyrogallol
Lutidine l-Methyl-2-cyclohexen-5-one
Propylfuran 2,3,5-Trlmethylfuran
1-Heptyne Cyclohexanecarboxaldehyde
2-Ethyl-2,3-dihydro-5-methylfuran
Diemthylcyclopentanone 5-Heptenoic acid
Butyrone Enanthaldehyde
Methyl caproate Enanthic acid
Heptane Benzofuran
m-Xylene o-Ethylphenol
2,3-Xylenol 2,4-Xylenol
3,5-Xylenol Creosol
6-Methylgualacol 2,6-Dimethoxyphenol
Methoxy-4-homocatechol 2,4-Dimethyl-4-cyclohexene-
1-one
Methylcyclopentanone 3-Isopropyl-2-cyclopenten-
i-one
Trimethylcyclopentanone 3,6-Octanedione
Caprylic acid Methyl enanthate
Dihydroxycaprylic acid 4-Vinylguaiacol
Cumene Pseudocumene
3,5-Dimethylguaiacol 4-Ethylgualacol
Homoveratrole 2,6-Dimethoxy-4-methylphenol
5-propylpyrogallol Isophorone
Amylfuran 2,4,--Trimethylcyclohexanone
Cyclohexanepropionaldehyde 3,3,5-Trimethylcyclopentanone
Pelargonic acid Naphthalene
Estragole Eugenol
Isoeugenol Cymene
Durene Thymol
4-Propylgualacol 2,6-Dimethoxy-4-ethylpheno1
A5-propyl-monomethyl ether of pyrogallol
Camphene Limonene
Nopinene Pinene
Sylvestrene r-Terpinene
Terp nolene Camphor
Borneol Cineole
Fenchyl alcohol Isofencyl alcohol
r-Te rpineol Canric acid
1,3,3-Trimethylblcyclo(2.2.2)-5-o ten-2-one
5-Propyl-1,3-dimethyoxy-2-hydroxybenzene
2,5-Difurfuryledine-1-cyclopentanone
Cadinene Pentadecane






Palmitic acid Heptadecane
Chrysene Retene
Oleic acid Stearic acid
Octadecane Nonadecane
Abietic acid Pimaric acid
Arachidic acid Eicosane
Heneicosane Behenic acid
Do cosane Tricosane
Lignoceric acid Melene
c-Methyl-S-ethylacrolein

No single compound accounts for more than a few per

cent by weight of the organic material in the acid water.

Acetic acid is one of the more abundant compounds (2).

The preceding information suggests that either an

insensitive, broad-spectrum, efficient treatment process

must be devised or some use must be found either for the

acid water or for products made from the acid water which

would provide the industry with financial returns.


Review of Applications for Wood Tars and Pyroligneous Acid

Considerable work has been reported in the literature

concerning uses for wood tars, although little information

is available on acid water alone. Hedstrom and Ballin (4)

own a Swedish patent on a process for treating halogenated

wood tars with polysulfides under heat and pressure to

produce a product which resembles rubber. Lindhe (5)

developed and patented a process for oxidizing wood tars

after the light oil fractions had been removed. Air is

used as the oxidizing agent and the temperature of the

process may range from 1000C to 3000C. Pitch of any

desired melting point may be produced by controlling the

length of the oxidation as well as the temperature.

Jonsson (6) suggested that wood tars could be used as






a fuel in piston engines. A mixture of 75 to 80 per cent

pine stump resins and 20 to 25 per cent spruce resins

produces the best results. Eldus (7) proposed that wood

tars be used as motor oils and as motor oil additives. A

commercially owned French patent describes a process for

hydrogenating wood tars (8), which is carried out at from

3000C to 5000C and 20 to 80 kg/sq cm hydrogen pressure.

Various combinations of aliphatic and alicyclic compounds

may be produced by the proper selection of the reaction

conditions. The products can be used as lubricants,

solvents, and transformer oils. Bond (9) has developed a

process for "sweetening" hydrocarbon distillates containing

mercaptans. Wood tars are mixed with the petroleum dis-

tillates and air is bubbled into the mixture while the

temperature of the mixture is held at 540C. The wood

tars aid in the destruction of the mercaptans in the hydro-

carbon distillates. Uranov et al. (10) proposed that wood

tars be used as an adhesive in the manufacture of charcoal

briquets, rather than the commonly used starch.

Mukherjee and Srivostava (11) developed a process for

manufacturing crude plastics using sugar, molasses, phenol,

creosote, and wood tar as raw materials. A somewhat

similar process has been proposed by Lyass et al. (12)

for the production of bakelite-type plastics. The raw

materials for this process are: one part wood tar, three

to five parts formaldehyde, traces of hydrochloric acid,

one hundred parts sand, and 11 per cent acetone.







Kozlov et al. (13) have described a process incorporating

flotation agents made from pyroligneous acid and wood tars

for processing copper sulfide and zinc sulfide ores from

Ural mountains. Chertkov and Zrelov (14) stated that wood

tars were not adequate as antioxidant additives in hydro-

carbon motor fuels and that these additives should be

replaced with amino-phenols. This statement implies that

wood tars have been tried as antioxidants in hydrocarbon

fuels.

The preceding applications for wood tars or acid water

do not appear to offer adequate returns or a technically

feasible method of treating or disposing of pyroligneous

acid. Little work has been reported in the literature

concerning waste treatment processes which are capable of

handling pyroligneous acid in a suitable manner. Only

two references, discussed immediately below, were found

in the literature concerning waste treatment processes

for pyroligneous acid.


Review of Waste Treatment Processes for Pyroligneous Acid

Szulicka and Grendysz (15) reported the treatment of

waste waters from a wood distillation plant by dilution in

a lagoon. The primary waste effluent of the plant contained

2,500 mg/1 of methanol, 5.3 mg/1 of acetic acid, and

200 mg/1 of phenol at the entrance to the lagoon. This

effluent was found to be strongly bactericidal. Other

wastes entering the pond included ground water, cooling

water, and sanitary wastes from the plant. The effluent







from the pond contained 27 mg/1 of methanol and 42 mg/1 of

phenol, and supported a vigorous microflora believed to be

resistant to phenol and methanol. No data were given on

the overall dilution of the primary waste flow or on the

retention time through the lagoon.

Hickerson and McMahon (16) investigated spray irriga-

tion as a supplement to lagooning and dilution as treatment

methods for the wastes of a wood distillation plant which

used hardwoods as a raw material. They determined the

most suitable grasses for spray irrigation and then used

spray irrigation to supplement controlled discharge of

the lagooned wastes into an adjacent watercourse. They

also investigated other possible treatment processes

including pH adjustment, chemical precipitation, and autox-

idation. The results of these tests were not encouraging.

Their autoxidation study consisted of aerating the waste in

a chemical reactor at 1500C and 100 psi. This treatment

produced a black sludge which they believed would create

troublesome problems in any type of filtration process.

The Sanitary Engineering Section at the University of

Florida carried out extensive chemical and biological

investigations from 1949 to 1951 to develop a satisfactory

treatment process for the Cabot Corporation. While some

technological success was achieved, the cost of treatment

was extremely high. The best biological treatment system

involved neutralizing the waste and then treating it by high

rate trickling filtration. At a recirculation ratio of






100:1, the COD of the waste was reduced about 75 per cent.

Both the effluent and influent were highly colored.

Clarification of the effluent was obtained with large

quantities of powdered activated carbon. The chemical

treatments of the waste which were investigated are:

neutralization with lime, coagulation with iron compounds,

coagulation with alum, acidification, chlorination, chlorin-

ation followed by coagulation, and extraction with an

assortment of organic solvents. Direct application of the

waste to unpaved roads showed little promise (17).


Experimental Proposal

Two conclusions were suggested by the preceding

information. First of all, biological treatment of the

waste is unattractive because of the high dilutions

necessary to reduce the toxicity of the waste and because

the effluent from biological treatment requires rather

expensive processing for clarification. Secondly, chemical

treatments must utilize inexpensive chemicals or regenera-

tive processes to be economically feasible. Autoxidation

appears to fulfill best the requirements of either

pretreatment or complete treatment of the waste. As a

pretreatment, autoxidation of the waste might partially

oxidize many of the toxic compounds in the waste, with a

possible reduction in overall toxicity. Certain of the

known toxic compounds in the waste are similar to known

antioxidants in autoxidation reactions, e.g., phenol

(18,19). The waste may be amenable to complete oxidation






by autoxidation in a few hours if sufficiently rapid re-

action rates can be established and if there is no lengthy

induction time preceding the reaction. Either pure oxygen

or air with proportional pressure increases is an inexpen-

sive oxidizing agent. In addition, sufficient heat must

be generated by the reaction to make the process self-

sustaining. Only one autoxidation system is being used in

the sanitary engineering field at present on a significant

scale. The system is known as the Zimpro Process and is

briefly described in Chapter II of this dissertation.













CHAPTER II

REVIEW OF THE ZIMPRO PROCESS

Description of the Zimpro Process


The Zimpro Process is a continuous operation for the

autoxidatlon of wastes, which process was initially de-

veloped in 1944 by F. J. Zimmerman (20). The process

consists of mixing liquid wastes and air under pressures

of 1,000 psi to 2,200 psi, and at temperatures of 5000F

to 6000F. The carbonaceous matter in the wastes reacts

with oxygen, releasing heat for the process. The heat

and pressure of the process effluent are recovered and

transferred to the Influent. Under typical conditions, a

1-hour reactor retention time would allow for a 90 per cent

reduction in COD (20).

An autoxidatlon process similar to the Zimpro Process

could prove to be applicable to the waste of Cabot Corpora-

tion. Some modifications of the process would be necessary

since oxygen rather than air is the proposed oxidizing

agent; but basically, the processes would be quite similar.

Figure 3 is a schematic illustration of the flow through

a Zimpro Process (21).

The waste or aqueous fuel flows into a holding tank,

and is then conveyed to a high-pressure pump. The high-

pressure pump acts as the pressure source for the system,












































































32


ca
P <0
CO 0 (







and as the source of head which creates flow through the

system. The waste, now under high pressure, is mixed

with metered high-pressure air from an air compressor.

The mixture enters heat exchanger number 1, then is

discharged at an elevated temperature, though not suffi-

ciently elevated to cause the autoxidatlon reaction to

proceed. The waste passes through heat exchanger number 2,

in which it reaches the temperature necessary for the autox-

idation to occur at a rapid rate. The waste then enters

the reactor where it is retained long enough for the

reaction to be completed. Since heat is released during

the reaction, the effluent from the reactor is hotter than

the influent. The effluent passes countercurrently through

heat exchanger number 2, or if enough heat has been acquired

in the reactor a portion of the effluent may by-pass heat

exchanger number 2. The effluent stream, recombined with

the portion by-passing heat exchanger number 2, enters a

gas-liquid separator. The gas phase, including steam,

nitrogen, and carbon dioxide, passes through a pressure-

control valve and is expanded through a turbine to the

atmosphere. The turbine powers an air compressor and a

synchronous motor-generator. The power developed can be

estimated from the following expression: power developed

heat released by autoxidation minus heat losses in the

entire system. If heat losses are greater than the heat

from autoxidatlon, additional heat must be added to the

process. The aqueous portion of the reactor effluent is







withdrawn from the bottom of the gas-liquid separator and

passed countercurrently through heat exchanger number 1.

Finally, the effluent passes through a third heat exchanger

located in the aqueous-fuel storage tank. The effluent

then passes through a pressure-control valve and eventually

into a natural drainage channel (21).


Discussion of the Zimpro Process

The most important feature of the Zimpro Process is

the production of heat during the autoxidation of an other-

wise useless waste. Essentially the "waste" is no longer

wasted because it is used as a fuel. The Zimpro Process

must be self-sustaining once started to be economically

attractive. The use of the Zimpro Process offers two

advantages: first of all, it disposes of highly concen-

trated wastes, and secondly, it can produce quantities of

heat beyond the needs of the process, which may be used in

some other unit operation.

The Zimpro Process, however, does entail certain

problems. Oxidation occurs only in the liquid phase

(20,21,22). Thus, low molecular weight organic compounds

would pass through the process without being oxidized.

Since molecular stability increases as the number of carbon

atoms in the molecule decreases, economical operation of

a Zimpro Process dictates that the stable components,

e.g., methanol, methanoic acid, and acetic acid, leave the

reactor in both the liquid and gaseous effluents. The

liquid effluent would ordinarily have a high Biochemical







Oxygen Demand (BOD) (23). Zimmerman (20) suggests that

organic gases leaving the reactor can be oxidized by

passing them over a silver or platinum catalyst. The liquid

effluent contains insoluble inorganic ash and high BOD,

both of which should be treated by conventional biological

methods (21).

Zimmerman (20) stated that the COD of a waste is

approximately equal to the weight of oxygen necessary to

oxidize the waste. He also suggested the following

expression for predicting the heat value of a waste:

Heat value (BTU/gal) = 48.8 x COD (gm/1). Table 3 lists

some of the properties of selected materials which might

be oxidized by a Zimpro Process (20).


Hy
Et
Ca
Ac
Ox
Py
Fu
La
Ca


Table 3

Autoxidation Properties of Several Materials

Material BTU/lb. lb. of 02/lb. lb. of air/lb.
of material of material

drogen 61,000 7.937 34.34
hylene 21,460 3.420 14.80
rbon 14,093 2.660 11.53
etic Acid 6,270 1.070 4.60
alic Acid 1,203 0.178 0.77
ridine 14,950 2.530 10.90
el Oil 19,376 3.260 14.00
ctose 7,100 1.130 4.87
sein 10,550 1.750 7.55


Waste Sulfide
Liquor
Solids
Semichemical
Solids
Sewage Sludge
Primary
Sewage Sludee
Activated


7,900

5,812

7,820

6,540


1.320

0.955

1.334

1.191


5.70

4.13

5.75

5.14


BTU/lb.
of air

1,780
1,450
1,220
1,365
1,565
1,370
1, 380
1, 455
1,395



1,385
1,410

1,365

1,270







Jackson and Brown (24) made several observations

concerning the practical operation of a Zimpro Process for

the treatment and chemical recovery of spent semichemical

pulping liquors. Among the more important observations

were: (a) corrosion was non-existant; (b) 92 per cent of

the COD was removed; (c) 90 per cent of the sodium sulfate

was recovered; (d) the equipment was easy to operate and

the cost of maintenance was low; (e) steam was recovered

economically.

The Zimpro Process has been successfully used in both

treatment and chemical recovery of the following types of

pulp and paper liquors: magnesium-base spent sulfite

liquor; semichemical acid, neutral, and alkaline sodium

sulfite spent liquors; calcium-base spent sulfite liquor;

ammonium-base spent sulfite liquor; sodium-base spent liquor;

and kraft liquor (21).

Koenig (25) has presented a good summary of the

characteristics of wet oxidation of sewage sludge by the

Zimpro Process:

(1) Oxygen pressure has no effect on the rate or the
extent of the oxidation.

(2) The reaction produces C02 and H20 plus low
molecular weight acids, e.g., CH3COOH, HCOOC,
etc.

(3) The reaction has an initial extremely rapid rate
followed by a much slower rate.

(4) The reaction reaches a different ultimate fraction
converted at each temperature.

(5) The reaction reaches approximately the same
ultimate fraction regardless of initial con-
centration at low temperatures.







(6) The reaction shows a decrease of fraction converted
with increases of initial concentration at low
temperatures.

(7) The fraction converted is a probability function
of temperature.

The experiments upon which this dissertation is based

were conducted using a process similar to the Zimpro Process,

differing only in the use of oxygen rather than air as the

oxidant. The use of oxygen in an industrial process appears

to offer two distinct advantages. First of all, the total

operating pressure of the system can be reduced, since no

nitrogen partial pressure contributes to the total pressure,

hence a lower equipment cost; secondly, since the oxygen

can be purchased under pressure, no compressor is necessary.

These advantages are at least partially offset by the cost

of the oxygen, although the results from experiments using

either air or oxygen as the oxidant should be comparable

if the partial pressures of oxygen are equivalent.














CHAPTER III

INTRODUCTION OF THE THEORY OF AUTOXIDATION
OF ORGANIC SUBSTANCES

General Description


Autoxidations are defined as oxidations which can be

brought about by oxygen gas at normal temperatures without

the intervention of a visible flame or an electric spark

(26). The preceding definition includes wet oxidation as

applied in the field of Sanitary Engineering. Bateman (27)

describes the complexities and vagaries of autoxidation

reactions very aptly. "The mechanisms (of autoxidations)

are complex and variable depending on constituent conditions

as well as trace impurities which may inhibit or initiate

the reaction." Since Bateman was speaking only of the study

of "pure" compounds, one can imagine the enormously complex

and dynamic reaction mechanisms which occur during the

autoxidation of a heterogeneous waste.

Ingold (28) describes the autoxidation of a pure

compound. "The normal curve for oxygen uptake during autox-

idation consists of an initial period where very little

oxidation occurs, known as the induction period. This is


"The succeeding discussion is taken largely from the
article cited in reference (28). The author has inserted
comments to aid in bridging the gap between theory and the
application of the autoxidation to heterogeneous wastes.







followed by a rapid increase in the rate due to autocatal-

ysis by chain-branching intermediates that build up during

the induction period. The rate soon reaches a maximum

value and then slowly starts to decrease." Figure 4 lists

the mechanisms which,in general, describe autoxidations.

RH represents the organic substrate, R02* is the correspond-

ing peroxy radical, and ROOH is the hydroperoxide.


Initiation

Reaction [il is a general expression for the initia-

tion process. The activation of organic species to the

free radical state is caused by heat, light, ionizing

radiation, or reaction with other free radicals. Hydrogen

atoms, or possibly protons, are preferentially attacked

in the following order: (1) hydrogens alpha to a double

bond or to a conjugated system, (2) tertiary hydrogens,

(3) secondary hydrogens, (4) primary hydrogens. Ingold (29)

has suggested that thermal initiation may involve the

reaction:

RH + 02 R* + H02'

since this reaction involves less energy than reaction [ .

Autoxidation of wastes, called wet oxidation by

Zimmerman and others, can be thermally initiated. Since

most wastes are heterogeneous mixtures, it is likely that

at least minute concentrations of hydroperoxide are present

in the waste. Ingold (28) states that in the presence of

minute traces of hydroperoxide, reaction 14] is the dominant

mode of initiation; and as the concentration of










Initiation:

RH activation> R- + (H.) [1]

Propagation:
R- + 02 R02* [2]

RO2* + RH ROOH + R* [3]

Decomposition of peroxide:
ROOH -RO* + .OH [4]

2ROOH RO + RO2. + H20 [5]


R02 + ROOH various products [6

ROOH -- -- nonradical products [7]

Induced decomposition of peroxide:
X + ROOH free radicals [8]

Y + ROOH ROH + YO [9]
Z + ROOH inactive products + Z 10o]

M + ROOH free radicals [ii]

Self-termination:

RO2. + RO2 --- inactive products [12]

Chain-breaking termination:

R02* + IH --- RO2H + I. [13]


Figure 4. Autoxidation Mechanisms







hydroperoxide increases, reaction 5] becomes increasingly

important as an initiation mechanism. Thermal initiation of

unsaturated substrates may occur as follows:
H H-O-
--C-C=C-? + 0=0 ----- -=-- (3 ,31,.
H H H H

-is represents the direct formation of a hydroperoxide

without free radical intermediates.


Propagation

Reaction I2] is extremely rapid except at very low

oxygen partial pressures. Consequently only peroxy radicals

are of importance in chain propagation and termination

(27,32).

In the absence of major steric effects, the rate of

reaction 3]1 depends on the carbon-hydrogen bond str-ngth

and on the availability of electrons at the carbon-hydrogen

bond being broken (33,34). Peroxy radicals can add to

olefinic double bonds (35), and tend to react more with the

oxygenated reaction products than with unoxidized species,

because the former frequently contain more reactive carbon-

hydrogen bonds (36).


Decomposition of Peroxide

The thermal decomposition of a hydroperoxide generally

yields several free radicals, which lead to chain branchin-

and account for the autocatalysis observed in many oxida-

tions. In addition, the solvent plays an important role in

the decomposition of hydroperoxide, causing the rate of






decomposition to change during an autoxidation as the
substrate is oxidized (37). At low hydroperoxide concen-
trations, the rate of hydroperoxide decomposition is first
order as exemplified by reaction [4] At higher hydro-
peroxide concentrations there are substantial deviations
from first order kinetics and apparent rate constant
increases (38,39). These phenomena are probably due to the
increasing importance of reactions [5] and [6] on the
overall autoxidation rate. The rate of decomposition of
hydroperoxides is accelerated by initiators which provide
an additional source of free radicals (40). The rate may
be retarded to a reproducible minimum value by the addition
of inhibitors of free radicals, which include polycyclic
aromatic hydrocarbons (41). The effects of reaction [5]
relative to reaction [4] decrease with increasing tempera-
ture (42). The following reaction represents a mechanism
which could be followed by the alkoxy radicals produced by
reactions [4] and [5] .
RO* + RH ROH + R*
Reactions [6] and [7? represent mechanisms which
terminate the chain reactions started by reactions 1
through [5] Reaction [7] may be considered as the reverse
of reactions [1] through [5] The products of reaction [7],
to be further oxidized, must again be activated to the free
radical state represented by reaction [i .






Induced Decomposition of Peroxide

Reactions [8] and [l1] represent catalysis of the
autoxidation while reactions [9] and [10] represent
inhibition. Reaction [8] shows the induced decomposition
of a hydroperoxide by substances not containing a heavy
metal. It results in the production of free radicals,
which accelerate the rate of the overall reaction. Alcohols

(43), ketones and ethers (44), fatty acids (45), and olefins
(46) are examples of this type of compound. A typical

mechanism for an olefin is as follows:

ROOH + R CH=CH2 -- R02. + R'HHCH3

Increasing the surface to volume ratio of the reaction
vessel or adding powdered solids may increase (47),
decrease (48), or have no effect (49) on the rate of
reaction [8] An increase in rate implies that reaction
[8] occurs only on a surface, whereas a decrease implies

that the surface promotes either reaction [o10 or the
destruction of free radicals without starting a chain
mechanism (28).

Reaction [9] describes a stoichiometric mechanism for

the inhibition of autoxidations by decomposing hydroper-

oxides to stable alcohols. Inhibition of an autoxidation

means the induced premature termination of free-radical-

forming chain reactions. Sulfur and selenium compounds
follow this inhibitory mechanism. Monosulfides which con-
tain at least one aliphatic or cycloaliphatic group attached
to the sulfur atom are more effective antioxidants than






mercaptans and disulfides, while diaryl sulfides and

sulfones are inactive (50). Inhibitors containing sulfur

are usually oxidized to strong acids which also act as

inhibitors, as shown in reaction 10o] Sulfur-containing

inhibitors also cause the production of resinous compounds

and sludges in a system undergoing autoxidatlon (51).

Selenides do not cause the formation of sludges and they

are more effective antioxidants than sulfur compounds (52).

As a rule, selenides are not found in wastes.

Reaction [o10 illustrates the mechanism by which

strong acids inhibit autoxidations. The acids are strong

in the Lewis sense (53). Although acids reduce the rate of

oxidation, they cause the production of resins, sludges,

and corrosive compounds (54). Bases are not as effective

as acids in reaction [0] but when used, resinous products

usually are not formed; when they are formed, these products

are selectively oxidized. Corrosive reaction products are

neutralized by the additives to form compounds with good

detergent properties (55).

Reaction [1i represents the induced decomposition of

hydroperoxides by a heavy metal. Metals which possess two

or more valency states with a suitable oxidation-reduction

potential between them can react with hydroperoxides to

form free radicals (56). Examples of suitable metals are

iron, cobalt, copper, manganese, add nickel. The metals

can act either as an oxidant or a reductant.

Oxidant Ce4 + ROOH -Ce3 + R02* + H






Reductant Fe2+ + ROOH Fe3+ + RO* + OH~

Certain metals, such as cobalt and manganese, can act both

as an oxidant and as a reductant. For these metals the

oxidant reaction may be rate-determining, since the metal is

found chiefly in its more highly oxidized state (57).

While ferric chloride acts as an inhibitor of the oxidation

of cumene and tetralin in aromatic solvents, it acts as an

initiator in polar solvents (58). Complexes of metal ions

and hydroperoxides are more active catalysts than metal

ions alone (59). The anion associated with the metal

can affect the catalytic activity of the metal either by

affecting its redox potential or by "blocking" the formation

of a catalyst-hydroperoxide complex. Anion effects are

specific for each metal and anion combination. The solvent

can affect the catalytic activity of the metal-anion

combination also (56).

Heavy metals act as powerful pro-oxidants of organic

autoxidations. Catalysis by metal surfaces is proportional

to their area and therefore is appreciable only in so far

as the metal is colloidally dispersed or dissolved (60).

The activity of the metal catalyst can be reduced by the

addition of corrosion inhibitors or metal passivators which

form a chemisorbed film on the surface of the metal (61).

Oxidation products form a non-corrosive substrate can also

laquer metal surfaces (62).

The catalytic effects of metal salts reach a constant

value at very low metal concentrations. This may be caused

by either chain termination by the catalyst (63) or by the







occurrence of a steady-state concentration of hydroperoxide,

as when reactions [3] and [l1i proceed at the same rate (64).

If reactions [3] and [11] proceed at the same rate, only one

new hydroperoxide is formed to replace each hydroperoxide

that decomposes. Consequently no chains can be started.

Metal catalysts can sometimes be completely precipitated

from solution shortly after the start of an autoxidation,

consequently such a catalyst could have no affect at all on

the rate of the reaction (65).

Metal ions and surfaces may be inactivated by chelating

agents such as ethylenediamine tetraacetic acid or

N,N -disalicylidene-1,2-propanediamine (66). Chelating

agents are effective because their stearic configuration

prevents the formation of metal ion-hydroperoxide complexes,

and also because they modify the redox potential of the

metal ions (58). The metal chelate may be oxidized during

an autoxidation, thus activating or releasing the metal

ion catalyst (67).


Self-termination

Reaction [12 represents the termination of a chain

by the interaction of two peroxy radicals. Primary and

secondary peroxy radicals terminate chains by way of a

cyclic transition state (68). Tertiary peroxy radicals

terminate reaction chains less readily since they lack a

hydrogen atom on thec<-carbon. These radicals are believed

to produce alkoxy radicals, which then dimerize (69). It

is possible to reduce the rate of autoxidation of a compound






which gives tertiary radicals by adding a compound which

produces primary or secondary radicals (34). Mixtures of

compounds can result in higher overall rates as compared

withthe sum of the rates of autoxidation of each of the

pure substituents.


Chain-breaking Termination

Reaction [13] describes the mechanism by which

inhibitors of free radicals reduce the oxidation rate of

organic compounds. Reaction [13] was first proposed to

cover phenolic inhibitors but was later extended to include

substituted analines. The free radical I* is stabilized by

resonance and sometimes by stearic interference by other

substituted groups. Thus I* does not initiate a new

oxidation chain (60). It will be destroyed by reacting

with another free radical. Other free radical inhibitory

reactions which have been identified are:

I. + R02 --- ROOI (70)

I. + R02* ------ ROOH + I (71)

Is + I --- .. ......... .---------- 12 (72)

Is + I. -- IH + I (73)

Many inhibitors whose active site is not protected by

bulky substituents can participate in the following

reactions:

I. + RH IH + R- (74)

102' + RH -- 100H + R" (74)

I. + 02 --- I02' (73)

I02* + I. --- 01 (73)







The inhibitor radical I* can react with hydroperoxides

and add to double bonds (75).

The rates of uninhibited oxidations are independent

of oxygen pressure, except at very low pressures (27). The

rates of inhibited reactions depend on oxygen pressure.

These data suggest that most Inhibitors react directly

with oxygen.

IH + 02 -- I + HO2. (76)

Oxidation of an inhibitor may either increase or decrease

the effectiveness of the inhibitor. The effect is dependent

on the specific conditions of a particular reaction.

Several qualitative comparisons of the effectiveness of

various Inhibitors have been compiled (77). The efficiency

of a given inhibitor type is increased by an increase in

the electron density at its reactive center, that is, by

a decrease in the redox potential of the inhibitor. If

the redox potential is too low, the efficiency of the

inhibitor is reduced because the inhibitor becomes suscep-

tible to direct oxidation (60).

While phenols and amines are the most effective Inhib-

itors known, other compounds such as quinones (78), ali-

phatic alcohols (79), and aromatic hydrocarbons, particu-

larly polycyclic aromatics (80), act as weak Inhibitors.

Two or more different inhibitors may act synergistic-

ally or antagonistically on the rate of an autoxidation.

Synergistic effects give an inhibited rate which is slower

than direct addition of the effects of the individual






Inhibitors might imply. Antagonistic effects give higher

reaction rates than would be reached if only one inhibitor

were present. Synergistic and antagonistic effects are

specific for each substrate and combination of Inhibitors

(28).


General Discussion

The preceding introduction to the theory of autox-

Idation is but a brief summary of volumes of material

available in the chemical literature. From an engineering

point of view, such theoretical explanations are valuable

for planning new research and improving the economic

operation of existing facilities. Each waste must be

studied separately to develop an optimum autoxidation

treatment process, because the reactions which can occur

simultaneously are so complex.

Three important areas of research for applied wet

oxidation processes are suggested by the theory discussed

in preceding portions of this chapter. Most important is

the development of techniques which tend to promote and

sustain autoxidation chain reactions. Since rapid reaction

rates are attained only when chain reactions are dominant,

reactors should be designed to promote and sustain chain

reactions. A faster reaction rate means smaller equipment

and a lower installation cost.

Other authors have pointed out that some autoxidation

reactions are surface reactions. For some wastes a packed

type of reactor might be best, while for others a tank-like





33

reactor would be as effective as one with a high surface

to volume ratio.

Catalysis of wet oxidation offers the promise of high

economic returns. Efficient catalysts should be discovered

for each waste. Furthermore, the most efficient method of

applying catalyst should be found. Some possibilities are:

dissolving or dispersing the catalyst in the waste, lining

the reactor walls with the catalyst, and making propellers

for mixing the reactants of a catalytic material.













CHAPTER IV

MATERIALS AND TECHNIQUES FOR THE EVALUATION OF PARTIAL
AUTOXIDATION OF WOOD DISTILLATION AS PRETREATMENT PROCESS

Introduction


The use of autoxidation as a pretreatment process

was investigated first since it promised to be more econom-

ical than complete autoxidation. In assessing the effects

of partial autoxidation on the waste, the experiments

were conducted on a pilot-plant scale. The kinetics of

the reaction were also studied while producing a substrate

to study biologically on a trickling filter. Considerable

equipment had to be collected and assembled to implement

the research.


Description of Reactor

A steam-heated chemical reactor of 67-liter volume

was obtained. The assembled reactor weighed 350 pounds

and was mounted on a four-wheel cart. The head of the

reactor could be removed by using a portable bomb hoist

mounted on a rotating T boom made of galvanized pipe.

The reaction chamber of the reactor and all fittings were

made of 316 stainless steel. The reactor was rated to

operate up to 250 psi and 1500C. The maximum temperature

which was attained in our laboratory was about 1300C

because of the inaccessibility of saturated steam at

34







pressures greater than 50 psi. The reactor was composed

of three primary sections. Figure 5 shows an overall view

of the assembled reactor.

The cylindrical base section of the reactor had a

dished bottom. The reaction chamber was surrounded by a

steam jacket. Condensate was removed from the bottom of

the steam jacket and water was circulated through the

steam jacket to cool the reactor. The reaction chamber

was sampled and drained through a 2-inch diameter threaded

opening located in the center of the dished bottom.

A 2-inch high pressure needle valve was attached to a

i-inch x i-inch bushing by a !-inch nipple. The female

half of a nut-union was attached to the outlet side of

the needle valve with a !-inch nipple. The male half of

the nut-union was attached to a !-inch pipe about 6 inches

long which served as a sample collector. The lower end

of the pipe was sealed with a -inch cap. The sample

volume was about 30 ml.

The liquids in the reaction chamber were mixed by a

four-bladed propeller, which was attached to a 1-inch

diameter stainless steel drive shaft. The drive shaft

entered the reaction chamber through the top of the reactor.

Figure 6 is an interior view of the top of the reactor

which shows the relative size of the propeller and drive

shaft. A standard asbestor-filled packing box was used

to maintain a pressure-tight seal around the drive shaft.

Silicone stopcock grease was used as a lubricant in the








Drive System
for Mixing
Reactants


Rupture
Disc
Assembly







Condensate
Drain Line


Steam Line
to Reactor


Gas Inlet
Valve,
Cross,
Pressure
Sensor
Fitting,
Gas Exhaust
Valve



Cooling
Water Line


Cart


Sampling Tube


Figure 5. Exterior View of 67-Liter Reactor












































Figure 6.


Interior View of Reactor Top






packing box rather than hydrocarbon grease, which can

react violently with high-pressure oxygen. An electric

motor attached to a speed-reducing gear box drives the shaft

at about 80 RPM. A 4-inch x 8-inch section of stainless

steel sheet metal was formed into a propeller and attached

to the top of the drive shaft, just inside the reactor.

This propeller was intended to mix the gases in the reaction

chamber.

Aside from the packing box and drive system, there

were two ports in the dished top. (See Figure 5.) A 2-inch

diameter port was fitted with a 2-inch, horizontally

oriented tube which was welded in place. The tube was

about 18 inches long and terminated in a machined flange.

This flange and a matching flange were used to hold a

nickel rupture disc, rated to burst at 350 psi and 720C,

to minimize the hazards of experimental explosions. The

atmosphere of the reactor was found to be highly corrosive

to the nickel discs. The corrosion rapidly reduced the

bursting strength of the discs. Coating the rupture discs

with silicone stopcock grease greatly retarded the corrosion.

A i-inch diameter cross was attached to a port in the

top of the reactor. A needle valve was attached to one

horizontal leg of the cross. Oxygen was fed into the

reactor through this valve from a standard high-pressure

oxygen cylinder. An Airco pressure-reducing valve was

used to control the oxygen pressure in the reactor.

The other horizontal leg of the cross was attached








to a tee. The 900 leg of the tee was fitted with a

1/16-inch tubing fitting which served as a connection for

a pressure-measuring device. The horizontal leg of the

tee was connected to a needle valve. This needle valve was

used to exhaust the gases in the reactor at the end of an

experiment or to decrease the pressure in the reactor during

an experiment. The vertical leg of the cross was fitted

with a thermocouple which extended about 12 inches into

the reaction chamber. The thermocouple was removed to

add waste to the reactor before starting a run.

The pressure and temperature inside the reactor were

measured continuously during an experiment. The readout

of the sensor signals was accomplished with a Brown Sixteen

Point Printing Recorder. All the even numbered recorder

inputs were wired together and connected to the pressure

transducer. Likewise all of the odd numbered recorder

inputs were wired together and connected to the thermo-

couple. The recorder read and printed, alternately,

pressure and temperature at 15 second intervals. Appendix 1

gives circuit diagrams and calibration curves for the

sensor systems.

The reactor was insulated with 1-inch thick fiberglass

insulation which was covered on both sides with aluminum

foil. The insulation helped maintain a constant reactor

temperature and prevented burns while working with the

reactor. Since samples had to be taken at known time

intervals, a stopwatch was used as a time monitor.








Experimental Technique for Reactor

The experimental technique is given in a step by step

manner as follows

1. A sample of waste was collected from the waste

lagoons at Cabot Corporation. (Grab samples were

taken.)

2. The dilution of the waste selected for any given

experiment was prepared with distilled water.

3. Twenty liters of the diluted waste were poured

into the reactor.

4. The waste and reactor atmosphere were purged with

nitrogen gas for 15 minutes. This created an

inert atmosphere in the reactor, which prevented

premature initiation of the autoxidation reactions.

5. The reactor was sealed by screwing the thermocouple

into place and closing all valves.

6. Mixing was begun.

7. The reactor was brought to the desired temperature

by opening the valve on the steam line and opening

the valve on the condensate drain line.

8. The temperature was held constant by adjusting the

condensate drain valve until the flow of steam

through the steam jacket gave the desired tempera-

ture. At temperatures around 1000C, it was neces-

sary to open and shut the steam-inlet valve rather

than use a continuous bleed system.

9. Oxygen was added to the reactor first by








pressurizing the oxygen feed line, then by opening

the reducing valve on the oxygen cylinder, and by

opening the needle valve attached to the cross

on the top of the reactor. The pressure in the

reactor was adjusted with the pressure-reducing

valve. If the pressure was accidently raised too

high, the exhaust valve could be opened briefly to

lower the reactor pressure.

10. The needle valve connected to the bottom of the

reactor was cracked open and any nonrepresentative

material was flushed out of the valve. This valve

was then closed. (This needle valve will be

called valve "A" in the rest of the procedure.)

11. The sampling tube was attached to the nut-union

and submerged in a liquid nitrogen bath.

12. Simultaneously, the stopwatch was started and valve

"A" was opened.

13. After the sampling tube was filled, valve "A"

was closed.

14. Within 40 seconds after having entered the

sampling tube, the sample was completely frozen.

The sampling tube was then disconnected from the

reactor.

15. The sample was thawed until it could be removed

from the sampling tube and transferred to a vial

for storage in a refrigerated area.

16. The sampling tube was washed, rinsed with distilled







water, and drained so that the next sample could

be collected.

17. Steps (10) through (16) were repeated for the

collection of successive samples, except that the

stopwatch was started only with the first sample

and let run continuously for the duration of the

experiment.

18. After the last sample was collected, the steam was

shut off. When the pressure in the steam jacket

had bled out through the condensate drain line,

cool water was flushed through the steam jacket

until the reactor returned to room temperature.

19. The pressure in the reactor was vented to the

atmosphere.

20. The partially oxidized waste was drained from the

reactor and stored in jugs to serve as substrate

for a trickling filter.

21. The reactor was washed and rinsed.


Samples were frozen as they were collected for two

reasons. First of all, chilling or freezing essentially

stops the oxidation, and secondly, freezing the sample

made it possible to remove the sampling tube from the

reactor without losing the sample through vaporization.

The best technique for freezing the sample proved to be to

immerse the sampling tube in the liquid nitrogen bath

before the sample was collected. The reaction was stopped

as soon as the liquids contacted the cold tube, before the








sample was frozen. The container for the liquid nitrogen

bath was a wide-mouth vacuum-bottle liner. To reduce the

hazards of flying glass in the event of an accident, the

vacuum liner was wrapped with plastic tape.

The samples collected during the autoxidation were

analyzed by the Chemical Oxygen Demand test (COD) as

described in Standard Methods for the Examination of Water

and Waste Water (81). This test gave a measure of the

strength of the partially autoxidized waste relative to the

strength of the raw waste. Appendix 2 describes some minor

variations from the standard COD procedure which were used

throughout this research.


Description of Trickling Filter

A trickling filter using Dow-Pac medium was selected

because of its resistance to shock loads, as the best means

of evaluating the toxicity of the treated waste. The wall

of the filter consisted of three one-quarter sections of

8-inch diameter vitrified clay pipe. The pipe and medium

were supported by a wooden stand which allowed air to

circulate vertically through the filter. The base of the

stand was in a 67-liter tank. Substrate was cycled from the

tank through the filter and back into the tank. The tank

outlet was about 4 inches from its bottom, and consequently

acted as a sludge storage area. Flow through the filter

was maintained between 21 and 3 liters per minute and a

small centrifugal pump furnished the head necessary to

maintain flow through the system. Figure 7 shows the



















































Figure 7. Trickling Filter







trickling filter. The flow was distributed over the filter

surface with a horizontal section of pegboard. Substrate

flowed over the pegboard from the hose leading from the

pump. The edges of the pegboard were raised so that about

1 inch of head was developed above each of the perforations.

The flow passed equally through each perforation and into

the filter medium. The 1/8-inch diameter perforations in

the pegboard were arranged in a square pattern on 1-inch

centers. The studies were started using a batch process.

If the results had been encouraging, the trickling filter

could have been converted into a continuous system.


Experimental Technique for Trickling Filter

The primary problem to be solved by experiments with

the trickling filter was that of determining the effective-

ness of autoxidation on reducing the toxicity of the waste.

An experimental technique which provided the desired

information is given:

1. The contents of the tank were mixed.

2. A grab sample was collected and filtered, and its

COD determined.

3. Substrate was poured into the tank, and the tank's

contents were throughly mixed.

4. A grab sample was collected and filtered, and

its COD determined.

5. The procedure was repeated at the same time daily.


The samples were filtered through a fiberglass mat.








Filtration of the samples was necessary since, in an

industrial process, most of the suspended matter would be

removed by sedimentation, and only soluble material would

contribute significant amounts of COD to the filter. By

sampling the tank directly and determining the COD of the

sample, as well as the COD of raw Cabot waste, it was

possible to determine the dilutions of partially autox-

idized waste necessary to maintain an active filter relative

to the strength of the raw waste. This value, when compared

to dilution values for the raw waste, indicated the

effectiveness of partial autoxidation as a pretreatment.













CHAPTER V

RESULTS OF THE INVESTIGATION OF PARTIAL AUTOXIDATION
AS A PRETREATMENT PROCESS

Experimental Plan


Prior to this study, no information was available

concerning the autoxidation of pyroligneous acid. Since

explosions are not unexpected when dealing with high-

pressure oxygen systems, it was felt that a cautious

approach should be made in the experiments. The plan was

to start with mild oxidizing conditions and high dilutions

of the waste, and gradually to proceed to the maximum

oxidizing conditions and the highest concentrations of the

waste. Three factors affected the rate of the reactions:

concentration of the waste, temperature, and the partial

pressure of oxygen. The partial pressure of oxygen

determined how much oxygen dissolved in the waste at any

given temperature


Determination of Oxygen Partial Pressure

Appendix 3 gives data concerning the solubility of

oxygen in water at elevated temperatures and pressures.

The rate of mixing in the reactor affected the rate of

oxygen dissolution. This had an effect on the overall

rate of the autoxidation. Each experiment was started at

a different temperature, oxygen pressure, and concentration

47







of waste. The temperature and oxygen pressure were held

as constant as possible while the change in the COD of

the waste was measured.

The partial pressure of oxygen could not be measured

directly, so the total pressure was measured and the partial

pressure of each of the components deducted from the total

pressure. The result obtained was the partial pressure of

oxygen in the reactor at the temperature of the experiment.

The oxygen pressure was converted to a standard temperature

for comparative purposes. Steam pressure was taken from

readily available information on saturated steam pressures

at various temperatures. The assumption that saturated

steam rather than superheated steam was present in the

atmosphere of the reactor must be valid since there was

liquid water in the reactor during each run. The pyro-

ligneous acid was assumed to behave like water as far as

steam pressures were concerned. While the acid water

undoubtedly did not behave like pure water, the assumption

of equality between the two liquids was valid within the

anticipated limits of accuracy in the experiments. Appendix

4 gives the variation of saturated steam pressure with

temperature. The pressure monitoring system was sensitive

to a pressure change of 1 psi.

As autoxidation occurred in the reactor, some carbon

dioxide was produced as an end product of the reaction. The

carbon dioxide produced by the reaction in turn produced

partial pressures great enough to be measured by the







pressure sensing system. This resulted in a decrease in

the effective partial pressure of oxygen, directly propor-

tional to the increase in partial pressure of carbon dioxide.

The build-up of carbon dioxide was not rapid enough to

appreciably affect the oxygen partial pressure during any

experiment for two reasons. First of all, in most

experiments the dilution of the waste was great, and even if

all the carbon atoms had been converted to carbon dioxide,

there were not enough carbon molecules in the waste to

affect the pressure in the reactor to any measurable extent.

Secondly, the gas volume in the reactor was large, about

47 liters. To create a measurable carbon dioxide partial

pressure several grams of carbon dioxide would have to

have been released by the reaction. The only experiments

that might have been affected by carbon dioxide build-up

would have been those using undiluted waste under the most

extreme oxidizing conditions.

Appendix 5 shows a theoretical calculation of carbon

dioxide build-up as a result of autoxidation. If the

partial pressure of carbon dioxide was approximately zero,

then the partial pressure of oxygen was constant. Any

oxygen consumed by the waste was replaced from the oxygen

cylinder, since the cylinder remained open and connected

to the reactor during each run.

Appendix 6 shows a typical calculation for the com-

putation of an oxygen partial pressure in an experiment.

A similar procedure was used to compute oxygen partial







pressures in all other experiments.


Results of Autoxidations

Figure 8 compares the reaction rates with two 1:100

dilutions of acid water under different oxidizing conditions.

The average rates during the first hour of the reaction

indicated that temperature was important in determining the

rate of the reaction. Run 2 followed first order kinetics,

while Run 3 followed kinetics which were greater than first

order. If an 85 per cent COD reduction is considered

complete treatment of the waste, both rates are very

slow.

Figure 9 gives the rate with a 1:50 dilution of acid

water. The reaction followed two different first order

rates. The rate during the first hour of the reaction was

approximately the same as the rate of Run 3, even though

the initial waste concentration of Run 3 was about one-

half the initial waste concentration of Run 4.

Figure 10 compares the reaction rates of three

different dilutions of acid water. All three reactions

followed first order kinetics. Sensitivity to the partial

pressure of oxygen is shown by the reaction rate for

Run 6. Decreasing oxygen partial pressure appeared to

limit the reaction rate, but increasing the pressure of

oxygen did not promote proportional rate increases.

Figure 11 compares the reaction rates of four runs

on the same dilution of acid water. The initial concen-

trations were not identical because of differences in the









100


120


Time (min)

Figure 8. COD vs. Time


Run Symbol Avg.
Temp.
oC


2 (1- ) I10

3 (0-- ) 133


Avg.
PO2 @ 0C
(psi)


117


Waste
Dilution


2.5


1:100


Avg. Rate
For First
Hour
mg/1 min


2.5


1:100 6.0


89 2.5







300



200


100

80


60



40


Ir)


120


Time (min)


Figure 9. COD vs.


Run Symbol Avg.
Temp.
oC


4 (0--0) 132


Avg.
PO2 @ 00C
(psi)


113


pH



2.5


Waste
Dilution


1:50


Avg. Rate
For First
Hour
mg/1 min


6.6


180


240


I-Le


~


r


-=i;~_Q


(^








100

80


60


120


180


Time (min)

Figure 10. COD vs. Time


Run Symbol Avg.
Temp.
oC


5 (-- ) 127

6 (0--0) 129

7 (0----) 129


Avg.
P02 @ OOC
(psi)


116

47


Waste
Dilution


2.5

2.5


1:10

1:25


116 2.5 1:20


Avg. Rate
For First
Hour
mg/1 min


15.7

2.5

12.0


240




















~t+-

.1 <---;--- --- -


80 n-


60 L


120


Time (min)


Figure 11. COD vs. Time


Run Symbol Avg.
Temp.
oC


Avg.
P0 @ 0C
2(psi)


Waste
Dilution


Avg. iate
For First
Hour
mg/1 min


8 (0-0)

13 (----)
14 (0----0)


46 2.5 1:20


600


400


200


100


180


240


90
112

129


148

135
81


2.5

2.5
2.5


1:20
1:20
1:20


2.7
4.2

9.5
5.7


15 (0) 129







samples of acid water. The importance of temperature was

clearly shown. The data suggest that reaction rates were

primarily a function of temperature. A comparison of the

reaction rates for Runs 7, 14, and 15 supports the assump-

tion that, after a certain oxygen partial pressure is

attained, further increases in oxygen partial pressure

contribute little toward increasing the rate of the

autoxidation. The reactions followed first order kinetics.

Figure 12 gives the reaction rates of three runs at

higher concentrations. Run 9 represented a concentrated

sample, but because of dilution of the acid water in the

lagoon at the time the sample was collected, the COD of

the sample was about the same as that of a 1:1 dilution

of waste collected on another date. The reaction rate

followed first order kinetics. The rates of Runs 9 and

17 were substantially greater than any other runs. This

demonstrates that a significant increase in the concen-

tration of the waste results in an increase in the rate of

the reaction.

Figure 13 shows the effects of altering the pH of

the waste before the autoxidation is attempted. A cal-

careous sludge was used to adjust the pH of the acid water.

The sludge was obtained from the Gainesville water

softening plant. The rates of all three runs were inhibited

at a COD of about 2,100 mg/1. Run 12 showed an increase

in COD during the first 30 minutes of the experiment. The

data supplied no explanation for this phenomenon, although

such a phenomenon would not be desirable in a waste






600




400








200







100


120 240


Time (min)


Run Symbol Avg.
Temp.
OC


9 (0--- ) 133

16 (---A) 126

17 (E--- ) 125


Figure 12.

Avg.
PO @ OC
(psi)


93

101


COD vs.


Time


Waste
Dilution


2.5 none


2.5


1:5


81 2.5 1:1


Avg. Rate
For First
Hour
mg/1 min


25.0

11.7

45.0


360








500





400


/


4 + 4 4- -


---- I I--g


0 30 60 90


120


150 180


Time (min)


Run Symbol Avg.
Temp.
oC


Figure 13.


Avg.
P02 @ OOC
(psi)


COD vs. Time


Waste
Dilution


Avg. Rate
For First
Hour
mg/1 min


10 (0--0)

11 (---l)


108 8.0 1:20


]


300


200






100


210


127

125


92

110


6.5

6.5


1:20

1:20


8.8

10.8

2.2


i


I


I


i J


i-


I


)- - -


b


1---


12 (r---) 130







treatment process. Since neutralization of the acid

water prior to autoxidation did not accelerate the rate

of the reaction, no additional preneutralization experi-

ments were attempted.

The autoxidation experiments at a pH of about 2.5

all followed a pattern of changes which were visually

perceptible in samples removed during a run. The 1:20

dilutions of the waste resulted in the most clearly

perceptible of these typical visual changes in the waste,

as seen in Figure 14. Sample 1 was collected just before

oxygen was added to the reactor, while sample 2 was

collected just after oxygen was added to the reactor. The

remainder of the samples were collected at thirty minute

intervals. The first visual change in the waste appeared

as the waste was heated prior to the addition of oxygen.

Heating partially broke the emulsion of tars in the waste

and caused these tars to adhere to the walls of the reactor.

This effect caused the COD of the waste to drop slightly

before any oxidation occurred. As soon as oxygen was added

to the reactor, the oils and tars in the acid water under-

went an extremely rapid polymerization reaction. The

particulate polymers which were formed caused the waste to

appear to be very dark in color. In addition, the waste

acquired strong surfactant characteristics which resulted

in stable foams. As the autoxidation continued, the waste

gradually acquired a yellow color. The polymers which

were formed at the start of the reaction were oxidized,

















1 5 q 7 q
S11 1'if N1o a1 I* Iv :

iIIufl"E


I


Figure 14. Typical Samples From Autoxidation Experiments






and the waste gradually lost its surfactant properties.

The presence of solids and stable foams in the samples

caused some difficulty in obtaining reproducible COD

values.

Appendix 7 gives triplicate COD values on Run 7.

The samples in which the initial COD was above 10,000 mg/1

did not become clear during the experiment.


Results of Trickling Filter Experiments

The substrate for the toxicity experiments was the

final reactor effluent in the autoxidation experiments.

In addition to the carbonaceous substrate (acid water),

several salts were added to the filter daily. These

salts were added because they are generally acknowledged

to be essential to bacteriological growth. The trace

elements insured that the toxicity of the acid water, and

not the lack of some essential element, would limit the

response of the trickling filter. Appendix 8 lists the

salts used and the amounts which were fed.

The trickling filter was acclimated to the partially

oxidized acid water for about one month prior to collection

of any data. Figure 15 gives the results of feeding the

partially oxidized acid water to the trickling filter.

The vertical lines on the graph represent the COD increases

in the substrate which resulted from the addition of acid

water. The diagonal lines represent the response of the

filter during a 24-hour period. The dashed line indicates

that the filter was still "alive" ten days after the last








O
14


/
/

/

/

/
/
/


J















I _
~^_







f7


I








04-) 0
o o



r-





-4
b3 'd



H -
C) r-4
SGO
iD CC)









-r4
1 II
oC


(t/2") GOD




62

daily feeding, on day thirty. The optimum response of the

filter occurred at a COD of about 1,000 mg/1. Since a 1:100

dilution is required to treat raw acid water by trickling

filtration, and the ratio of the partially oxidized acid

water is 1,000:90,000 or 1:90, no substantial reduction

in toxicity has been attained by partial autoxidation of

the pyroligneous acid.













CHAPTER VI

CONCLUSIONS FROM THE STUDY OF PARTIAL AUTOXIDATION
AS A PRETREATMENT PROCESS


The experiments already described in this disserta-

tion provide considerable information on the relationships

of pyroligneous acid to autoxidation. The main purpose

of this phase of the research was to determine if partial

autoxidation of pyroligneous acid would greatly reduce the

toxicity of the waste so that complete treatment of the

waste could be effected by conventional biological methods.

The data indicate that the toxicity of the waste was not

reduced any more than the COD of the waste was reduced.

Consequently, partial autoxidation of the waste does not

appear to be effective or desirable as a pretreatment for

pyroligneous acid. Either the toxic components of the

waste are not oxidized or they are partially oxidized to

other compounds which are equally as toxic as their

predecessors.

The data show that the acid water is autoxidized

during all experiments, although the reaction rates were

slow compared to the strength of the waste. The data

imply that if the proper conditions were discovered,

complete autoxidation would be possible. Succeeding

chapters of this dissertation support this contention.






No induction period was observed at the start of the

experiments. Consequently, the absence of an induction

period caused the complete oxidation of pyroligneous acid

to appear feasible by autoxidation techniques.

These data suggest that the reaction rates are

dependent primarily on temperature, although the partial

pressure of oxygen also affects the rate at any given

temperature. The reaction rates increased directly with

the temperature and oxygen partial pressure, although the

increases do not appear to be linear with either of these

variables. An increase in the concentration of the waste

caused a relatively large increase in the reaction rate.

The increase in rate was considerably less than the increase

in the concentration. The rate of oxygen transfer from the

gas phase into solution limited the increase of the reaction

rate as the COD or concentration of the waste was increased.

Neutralizing or raising the pH of the waste did not appear

to be desirable.

The preceding conclusions resulted in the study of

autoxidation as a technique for the complete oxidation of

the wood distillation waste. The following chapters of

this dissertation explain the study in detail.












CHAPTER VII

EQUIPMENT FOR THE INVESTIGATION OF COMPLETE
AUTOXIDATION OF WASTES

Introduction


The reaction rates obtained in the 67-liter reactor

under the most extreme oxidizing conditions possible with

the reactor, were not rapid enough to be used for the

complete autoxidatlon of the waste in an industrial

process. The data suggested that the reaction rates

could be accelerated greatly by increasing the temperature

at which the reaction was carried out and by increasing

the rate of oxygen transfer from the gas phase to the

liquid phase during the reaction. Since pressure is

dependent on the temperature, an increase in the temperature

capabilities of the reactor necessitated the use of a

high pressure reactor.


Description of Reactor and Heater

A 2-liter, electrically heated Parr reactor, capable

of operating at temperatures up to 3500C and pressures up

to 1,000 psi, was obtained. The aparatus consisted of

two sections, a 220-volt heater and controls, and the

reaction chamber, or bomb, and controls.

The 220-volt heater was composed of two concentric

cylinders. The opening between them was enclosed on both

65







ends with annular aluminum plates and was filled with

vermiculite insulation. Heat was supplied by six 250-watt

strip heaters, and the heat output was controlled by a

General Radio Co. Type W2HM Autotransformer.

A Sargent Model T Thermomonitor was included in the

heater circuit to obtain automatic temperature control.

The thermomonitor used a thermistor sensor which was

sensitive to 1/10 of 10C temperature change between 110C

and 2350C. Since the thermomonitor operated on 110 volt

AC, a relay was used in the heater circuit rather than a

direct electrical connection through the thermomonitor.

In addition, a switch was included so that the thermomonitor

might be shorted out of the heater circuit, and detached

and used elsewhere without rewiring the heater circuit.

The reaction chamber was composed essentially of two

sections, a cylinder with one end closed, and a top for

sealing the other end of the cylinder. The two pieces

together, which will be referred to as the reaction chamber,

or bomb, were constructed of 316 stainless steel. The

top was fitted with a teflon gasket to provide a pressure

tight seal when the bomb was assembled. The center of the

top housed the drive mechanism and packing box. The drive

mechanism consisted of a V-belt-powered hub which was

mounted on needle bearings. The drive shaft passed through

the packing box and attached to a lower drive shaft on

which two six-blade propellers were mounted. The propellers

were moved up and down on the drive shaft, which rotated








at 600 RPM. A pressure tight seal was maintained around

the rotating drive shaft with two "Rulon" cones. The

hollow cones were mounted back to back on the drive shaft

through openings in the narrow ends of the cones. The

space inside the cones was filled with silicone stopcock

grease. The grease acted as a lubricant and aided in

forming a tight seal around the drive shaft.

The life of the "iulon" cones -as about 40 operating

hours when the silicone grease was replaced at the start

of each experiment. Pressure was placed on the packing

cones before the bomb was pressurized. If this had not

been done, the lubricant would have been forced out of the

cones, resulting in the deterioration of the cones in 1 or

2 operating hours. The packing box was cooled during an

experiment by circulating tap water around the outside of

the box. The packing system of the reactor performed very

satisfactorily.

The gas exhaust for the bomb, a needle valve fitted

into a 1/8-inch diameter opening in the top of the bomb,

was used to lower the pressure in the bomb during an

experiment if by accident the reactor was over pressurized,

and to vent the gases in the reactor at the end of an

experiment. The same pressure transducer used with the

67-liter reactor was attached to a 1/8-inch opening in

the reactor top. The temperature sensor was the same

thermocouple arrangement used with the 67-liter reactor,

with the exception that the functions of the two junctions






were reversed. The sensing thermocouple was inserted into

a thermowell through the top of the reactor and extended

to the bottom of the reactor. Both pressure and temperature

were continuously recorded as in the earlier research.

Appendix 9 gives the calibration curves and other pertinent

information concerning the calibration of the pressure and

temperature sensing systems.

A silver rupture disc was used to minimize the danger

of an explosion during the course of an experiment.

Galvanized pipe was connected to the exhaust side of the

rupture-disc fitting to convey the reactants out the nearest

window. The first experiment conducted in the reactor with

pyroligneous acid demonstrated that silver was as sensitive

to the corrosive properties of the waste as nickel. A

teflon sheet was placed on the reaction-chamber side of the

rupture disc to minimize the corrosion of the rupture disc

and thus prevent premature breakage. The bursting pressure

of the rupture disc was rated as 1,800 psi at room tempera-

ture.

Oxygen was added to the reactor from a standard

cylinder and reducing valve. The oxygen line from the

cylinder attached to a needle valve, called the gas-inlet

valve, which connected to the top of the reactor by a tee.

The long axis of the tee joined the needle valve to the top

of the reactor. A second valve, the liquid-sampling valve,

was mounted on the 900 leg of the tee. A tube inside of the

reaction chamber led from these valves almost to the bottom







of the bomb. This arrangement made it possible to withdraw

liquid samples from the reactor by closing the gas inlet

valve and opening the liquid sampling valve. The tube was

then purged by oxygen when the gas inlet valve was reopened

after the liquid sampling valve was closed.


Collection of Liquid Samples

A coiled tube with outlets through the top of the

reactor was used to cool the reactants during the reaction

or at the end of an experiment. Both air and water were

used as coolants. The cooling coil reduced the volume of

the reactor from 2 liters to 1.362 liters. The top was

held on the cylindrical portion of the reactor by a split-

ring closure mechanism. Figure 16 shows the reactor

completely assembled and placed in the heater. Figure 17

shows the disassembled reactor with the lower drive shaft

disengaged so that the mixing propeller may be seen.

Liquid samples were collected by a technique similar

to that used in earlier research; that is, the sample

filled a sample-collection tube and was frozen with liquid

nitrogen. The sampling tube was made of 3/16-inch tubing

formed into a coil. Figure 18 shows the sampling tube

connected to the reactor just after the liquid nitrogen

bath has been removed. The lower end of the sampling tube

was closed with a Swagelock cap fitting, while the other

end was attached to the liquid sampling valve on the reactor.

The sample volume was approximately 7 ml.









Pressure
Transducer


Mixing
Drive
Motor



Brine Bath


Rupture
Disc Vent
Pipe







Oxygen
Inlet and
Liquid
Sampling
Valve





Heater
Casing


Auto-
transformer


Figure 16. Assembled Parr Reactor











Ru
Di
As


Oxygen Inlet
Valve


Drive Mechanism
and Packing Box

ipture Pressl
*sc Trans(
1sembly Conne



\/


ure
ducer
action


Cylindrical
Portion of
Reactor


Cooling
Coil






Mixing
Propeller


Closure Rings
(Foreground)


Figure 17. Disassembled Parr Reactor












AtSL


Figure 18. Sampling Tube on Reactor







Temperature Control During Experiments

Precise and automatic control of the temperature of

the reactor proved to be difficult to attain for a number

of reasons. First, the thermowell in the reactor was not

large enough to accommodate both the thermistor and

thermocouple sensors. Since the thermocouple gave a direct

measure of the temperature of the reactants, it was used

in the more desirable thermowell position. The thermistor

sensor was located beneath the reactor and held against the

bottom and side of the reactor by fiberglass insulation.

This technique gave a temperature variation of about 150C

after the system had stabilized. The thermistor was

inserted into the vermiculite insulation inside the heater

casing through an opening in the lower annular aluminum

plate. The sensor element was placed at various positions

relative to the strip heaters. However, the temperature

differential between the heater and the bomb was considera-

bly larger than anticipated. The large temperature

differential was caused by a 1/8-inch air gap between the

wall of the bomb and the heater. This air gap provided

very good insulation for the reactor. Consequently, the

temperature inside the vermiculite insulation in the heater

casing was considerably greater than the temperature of the

reactor. The strip heaters were capable of heating the

vermiculite insulation much more rapidly than they heated

the reactor. Because of these factors the temperature

fluctuation in the reactor was about t30C when the







temperature was controlled only by the thermomonitor.

The temperature could have been held closer to a constant

value manually if it had not been necessary to perform

other tasks during an experiment.


Collection of Gas Samples

During the studies that were conducted with the Parr

reactor, it became apparent that some information on the

composition of gases in the reactor was needed. The gas

samples were collected by bleeding a sample from the gas

exhaust valve on the reactor, through a brine bath to

dry the gases, and into a balloon. The moisture trap in

the brine bath was made of a capped section of galvanized

pipe about 8 inches long. The gases entered the trap

through a section of i-inch copper tubing which was soldered

into a hole drilled into the galvanized cap. This section

of the copper tubing which acted as the gas inlet extended

almost to the bottom of the trap. The gases then passed

vertically through coarse brass wool filtering material and

out of the trap through a second section of copper tubing.

The gaseous sample was collected in a rubber balloon which

was attached to one leg of a !-inch stainless steel cross.

The two legs of the cross, which formed 900 angles with the

leg to which the balloon was attached, were used for

evacuating the sampling train and for conveying the gases

from the reactor into the balloon. Tubing and pinch clamps

were arranged on these two legs so that the cross and

balloon were detached from the sampling train without loss








of the sample. The fourth leg of the cross was fitted with

a rubber septum so that portions of the sample were with-

drawn with a hypodermic syringe. Figure 19 shows all of

the equipment used to conduct the autoxidation experiments.

The gases were analyzed for oxygen and other components

on an F. and M. Model 720 thermal conductivity gas chromato-

graph. Appendix 10 describes the conditions used for the

analyses and shows typical peaks and calibration curves.

Several different experimental techniques were used

in the course of the research. The following chapter

presents each technique and the results obtained.
















W-
_ _=
i - -
S


Figure 19. Experimental Equipment


The equipment shown in the above figure is from left
to right as follows: the Sargent Thermomonitor, gas sampling
balloon (on top of the Thermomonitor), brine bath moisture
trap and gas sampling line, the Thermomonitor shorting
switch, autotransformer, heater with reactor in place,
stirrer motor, liquid nitrogen storage dewar, liquid
nitrogen bath dewar, recorder, ice water thermocouple
reference bath (on top of recorder), and pressure reducing
valve attached to the oxygen cylinder.













CHAPTER VIII

TECHNIQUES AND RESULTS OF THE STUDY OF
AUTOXIDATION OF WASTES

Introduction


The purpose of the following research was to investi-

gate the complete oxidation of wood distillation waste

by autoxidation techniques. The variables which were

expected to exert the greatest effect on the reaction

were temperature and oxygen partial pressure. During the

experimentation, it became obvious that minor changes in

experimental technique had profound effects on the rate of

the reaction. Consequently, the final experimental

technique was the result of an evolutionary process.


Source of Waste

Only the undiluted waste was used, to simulate most

realistically an industrial process. One problem already

described was the collection of reproducible samples from

the lagoons at the Cabot plant. To overcome this difficulty,

Cabot Corporation collected all the liquors from one day's

retort operation in a single tank. The settlable oils

were separated and a portion of the acid water was stored

in wooden barrels. The pyroligneous acid thus collected

gave satisfactory reproducibility during the remainder of

the study. The COD of the acid water was about 99,100 mg/1

77







when originally collected and about 90,000 mg/1 after three

months of storage. This loss of COD was probably a result

of the formation by the suspended tars of a coating on the

inside of the barrels.


First Experimental Technique

The experiments were planned to start with the study

of the reactions at a low temperature (1250C) and a low

partial pressure of oxygen (25 psi at 0C), and to proceed

stepwise by increasing the oxygen partial pressure up to

400 psi at 0C. The temperature was increased in 250C

steps up to 2250C, and a range of partial pressures was

investigated. This process was continued until the maximum

temperature and pressure capabilities of the system were

reached. During each experiment the temperature and oxygen

partial pressure were held as constant as possible.

The experimental techniques were similar to, but not

identical with, those used with the 67-liter reactor. The

technique used is given:

1. 1 liter of waste was poured into the reaction

chamber.

2. The reactor was assembled and placed in the heater.

3. Mixing was begun.

4. The gases in the reactor were evacuated and the

reactor was sealed.

5. The reactor was heated to the selected temperature

and the temperature was stabilized.

6. oxygen was gradually added to the reactor over a







period of approximately 3 minutes until the desired

partial pressure was reached.

7. The first sample was forced into the sampling

tube and a stopwatch was started simultaneously.

8. The sample was frozen in the sampling tube with

a bath of liquid nitrogen.

9. The sampling tube was disconnected from the reactor

and the sample was gradually thawed. As the sample

thawed, the expansion of the entrained gases forced

the liquid and particles of ice out of the sampling

tube. They were collected in a vial and analyzed

for COD determination (See Appendix 2).

10. Steps (7), (8), and (9) were repeated for the

collection of the succeeding samples. The stop-

watch ran continuously for the duration of the

experiment.

11. After the final sample was collected, the oxygen

cylinder was closed and the reactor was cooled

by passing water through the cooling ooil.

12. After the reactor was cooled to 500C or less, the

gases were vented to the atmosphere. The reactor

was disassembled and washed.


Results of the First Experimental Technique

The data shown in Figure 20 compare the reaction rates

at three different temperatures and approximately the same

very low partial pressure of oxygen. The rates of all three

reactions were quite slow and were approximately first order.







100


0 30 60 90


Time (min)


Figure 20.


COD vs. Time


Run Symbol


Avg.
Initial
Temp.
oC


Avg.
Initial
PO2 @ 0C
(psi)


Initial
Vol. of
Waste
(ml)


Avg. Rate
For First
Hour
mg/1 min


26 (0--O)

32 (E--])

38 (A--)


132

154


1000

1000


97

150


500 228


120


150


177 21







The rates during the first 20 minutes of the reaction were

substantially greater than the rates for the remainder of

the experiment. This was probably caused by the polymeri-

zation of the oils in the acid water, which occurred as

soon as the oxygen entered the reactor. The solid polymers

stuck to the walls of the reactor and were not measured by

the liquid sampling technique. The initial COD decreased

as the initiation temperature of the reaction was increased.

Heating the acid water partially broke the emulsion of

tars present in the acid water, and these particles adhered

to the walls of the reactor.

The information shown in Figure 21 compares the

reaction rates of two experiments at temperatures and

partial pressures higher than those presented in Figure 20.

A considerable increase in the reaction rate was readily

apparent. As a result of this faster rate, the COD was

reduced to a much lower value than in previous experiments.

Run 49 represented a 75 per cent reduction in the COD of

the waste in 75 minutes. Both reactions followed first

order kinetics.

The reaction rates shown in Figures 22 and 23 compare

the effects of five different initiation temperatures at

low oxygen partial pressures. The kinetics of all the

experiments were first order, and the rates of the two

experiments initiated at temperatures above 2000C were

considerably faster than those from experiments initiated

below 2000C. The results given in Figures 24 and 25

corroborate the results presented in Figures 22 and 23.







100


80


60


15 30 45

Time (min)

Figure 21. COD vs. Time


Run Symbol


Avg.
Initial
Temp.
oc


Avg.
Initial
P02 @ 00C
(psi)


Initial
Vol. of
Waste
(ml)


Avg. date
For First
Hour
mg/l min


44 (--O)

49 (0---GC)


203


400


226 79 300


400

903
































3 40 60
Time (min)
Figure 22. COD vs. Time


Run Symbol


Avg.
Initial
Temp.
oC


Avg.
Initial
PO @ OOC
(psi)


Initial
Vol. of
Waste
(ml)


Avg. Rate
For First
Hour
mg/1 min


27 (0-0)
33 (ED--- )
39 ( )


141
151


1000
500


181 39 500


100

80


60


I
o

r-


0


20






10
0


100


178
334
292


ii


I






100


15 30 45 60 75


Time (min)

Figure 23. COD vs. Time


Run Symbol


Avg.
Initial
Temp.
OC


Avg.
Initial
PO @ 0C
(psi)


Initial
Vol. of
Waste
(ml)


Avg. late
For First
Hour
mg/1 min


45 (0--0)

50 ([D---0)


207


400


455


350 692


I I


227 26







100


80


60




40







20







i_-


100


Time (min)


Figure 24. COD vs.


Run Symbol


Avg.
Initial
Temp.
OC


Avg.
Initial
PO2 @ 00C
(psi)


Initial
Vol. of
Waste
(ml)


Avg. Rate
For First
Hour
mg/l min


28 (0---0)

34 (0--0)
40 (--a)


136

160


52

112


1000

500


177 60 500


Time


110

365

392






100

80


60


Time (min)


Figure 25.


COD vs.


Run Symbol


Avg.
Initial
Temp.
OC


Avg.
Initial
PO2 @ Ooc
(psi)


Initial
Vol. of
Waste
(ml)


Avg. Rate
For First
Hour
mg/1 min


46 (0--0)

51 (E---)


205


400


226 59 350


Time


445

678


I







The data shown in Figures 26 and 27 compare the rates

of reactions initiated at higher partial pressures of

oxygen. Runs 41, 47, and 52 suggest that two first order

reaction rates were followed by the reaction at temperatures

above 1750C and partial pressures of about 100 psi. Run 30,

as shown on Figure 27, illustrates the importance of tem-

perature in achieving rapid reaction rates.

Support for the statements concerning the data shown

in Figures 26 and 27 is given by the data shown in Figure 28.

Higher temperatures affected the reactions in two different

ways. First of all, the higher temperatures caused more

rapid reaction rates, because the percentage of molecules

which reacted with oxygen was greater. This implies that

the rate of oxygen dissolution becomes rate-limiting as

the temperature of the reaction is increased. Secondly,

higher temperatures made a greater portion of the molecules

susceptible to oxidation; hence, a greater percentage of

the COD can be satisfied by the autoxidation.

Data comparing the rates of three experiments carried

out at intermediate temperatures and very high oxygen

partial pressures are given in Figure 29. Only Run 43

exhibits a very rapid rate.

Autoxidation of the waste caused several characteristic

changes to occur in the waste, which could be observed

visually during an experiment. These changes were similar

to those observed during experiments with the 67-liter

reactor. First of all, oils which were emulsified in the

acid water were polymerized into solid particles as soon as






100

80


60


I --


I+


100


0 20 40 60 80

Time (min)
Figure 26. COD vs. Time


Run Symbol


Avg.
Initial
T8mp.
C


Avg.
Initial
PO @ 0C
2(psi)


Initial
Vol. of
Waste
(ml)


Avg. Rate
For First
Hour
mg/1 min


(0--0)

(O---)


C---~

c- - ---- -


126

151

186


120

120

97


1000

500

500


280

275

490


I


1




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E5JF2DI23_N9UY6G INGEST_TIME 2017-07-19T20:28:43Z PACKAGE UF00097873_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

AUTOXIDATION OF INDUSTRIAL WASTES WITH OXYGEN By BOYD T. RILEY, JR. A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA June, 1966

PAGE 2

UNIVERSITY OF FLORIDA 3 1262 08552 2745

PAGE 3

ACKNOWLEDGMENTS For their guidance in the preparation of this dissertation, the author wishes to express his sincere appreciation to the members of his supervisory committee > Professor J. E. Kiker, Jr., chairman; Dr. C. I. Harding, co-chairman; Dr. H. D. Putnam; and Dr. G. B. Butler. The cooperation and advice extended by Dr. H. S. Sholtes, Dr. W. H. Morgan, and Dr. H. A. Bevls greatly assisted in preparing the experimental equipment. The aid of Mr. H. McGraw in fabricating special fittings is noted in a similar manner. Finally, the author wishes to thank his wife for her patience and encouragment , as well as her participation In both the experimental work and manuscript preparation for this dissertation. The graduate training leading up to this study was supported in part by U. S. Public Health Service Training Grant 8T1-ES11-04. This study was supported financially by Cabot Corporation. 11

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGMENTS li LIST OF TABLES v LIST OF FIGURES vi ABSTRACT lx CHAPTER I. INTRODUCTION 1 II. REVIEW OF THE ZIMPRO PROCESS Ik III. INTRODUCTION OF THE THEORY OF AUTOXIDATION OF ORGANIC SUBSTANCES 21 IV. MATERIALS AND TECHNIQUES FOR THE EVALUATION OF PARTIAL AUTOXIDATION OF WOOD DISTILLATION AS PRETREATMENT PROCESS Jk V. RESULTS OF THE INVESTIGATION OF PARTIAL AUTOXIDATION AS A PRETREATMENT PROCESS 4? VI. CONCLUSIONS FROM THE STUDY OF PARTIAL AUTOXIDATION AS A PRETREATMENT PROCESS 63 VII. EQUIPMENT FOR THE INVESTIGATION OF COMPLETE AUTOXIDATION OF WASTES 65 VIII. TECHNIQUES AND RESULTS OF THE STUDY OF AUTOXIDATION OF WASTES 77 IX. CONCLUSIONS FROM THE STUDY OF AUTOXIDATION OF WASTES 127 APPENDICES 1. CALIBRATION OF PRESSURE AND TEMPERATURE SENSOR SYSTEMS 131 ill

PAGE 5

APPENDICES Page 2. COD TECHNIQUE 13^ 3. SOLUBILITY OF OXYGEN IN WATER 137 4. PARTIAL PRESSURE OF STEAM AS TEMPERATURE VARIES 140 5. MAXIMUM THEORETICAL PARTIAL PRESSURE OF CARBON DIOXIDE 142 6. CALCULATION FOR COMPUTATION OF OXYGEN PARTIAL PRESSURE IN AN EXPERIMENT 143 7. RESULTS OF TRIPLICATE COD DETERMINATIONS ON RUN 7 1^ 8. TRACE ELEMENT FEED SCHEDULE FOR TRICKLING FILTER 145 9. CALIBRATION OF PRESSURE AND TEMPERATURE SENSOR SYSTEMS FOR PARR REACTOR 146 10. INFORMATION ON GAS CHROMATOGRAPH ANALYSES 149 11. STATISTICAL COMPARISON OF COD DETERMINATIONS 153 12. PRELIMINARY DESIGN AND ECONOMIC FEASIBILITY STUDY OF AUTOXIDATION AS A WASTE TREATMENT PROCESS FOR PYROLIGNEOUS ACID 156 13. RECO MEN DAT IONS FOR FUTURE RESEARCH 168 LIST OF REFERENCES 171 BIOGRAPHICAL SKETCH 179 lv

PAGE 6

LIST OF TABLES Table Page 1. Summary of Acid Water Characteristics k 2. Compounds Formed by Wood Carbonization 6 3. Autoxldatlon Properties of Several Materials 18

PAGE 7

LIST OF FIGURES Figure Page 1. Flow Diagram of Wood Distillation Process 2 2. Dilutions of Neutralized and Un-neutralized Acid Water from Cabot Corporation 5 3. Typical Zimpro Process 15 4. Autoxldation Mechanisms 23 5. Exterior View of 67-Liter Reactor 36 6. Interior View of Reactor Top 37 7. Trickling Filter 44 8. COD vs. Time, Runs 2 and 3 51 9. COD vs. Time, Run 4 52 10. COD vs. Time, Runs 5, 6, and 7 53 11. COD vs. Time, Runs 8, 13, 14, and 15 54 12. COD vs. Time, Runs 9, 16, and 17 56 13. COD vs. Time, Runs 10, 11, and 12 57 14. Typical Samples From Autoxldation Experiments 59 15. Trickling Filter Data 6l 16. Assembled Parr Reactor 70 17. Disassembled Parr Reactor 71 18. Sampling Tube on Reactor 72 19. Experimental Equipment 76 20. COD vs. Time, Runs 26, 32, and 38 30 21. COD vs. Time, Runs 44 and 49 82 vi

PAGE 8

igure

PAGE 9

Figure Page 49. GOD vs. Time, Run 76 122 50. GOD vs. Time, Hun 78 123 51. GOD vs. Time, Run 79 124 52. COD vs. Time , Sun 77 126 53* Circuit Diagram for Pressure-Sensing System. .... 132 54. Calibration of Pressure TransducerRecorder System 133 $5, Circuit Diagram of Thermocouple-Recorder System 134 56. Calibration of Thermocouple-Recorder System 135 57. Solubility of Oxygen in V/ater 138 58. Solubility of Oxygen in Water 139 59 • Saturated Steam Partial Pressure vs. Temperature l4l 60. Calibration of Thermocouple-Recorder Against Mercury Thermometer 147 61. Calibration of Pressure Transducer 148 62. Typical Gas Peaks Obtained From An Analysis 150 63. Calibration Curves for Oxygen, Nitrogen, and Carbon Dioxide 151 64. F. & M. Model 720 Gas Chromatograph 152 65. Extrapolation of Reaction Data 163 66. Flow Diagram of Autoxidation Waste Treatment Pro ce ss 164 67. Schematic Plot Plan of Autoxidation Waste Treatment Process 165 68. Schematic Profile of Autoxidation Waste Treatment Process 166 69. Design of Chemical Reactor 167 vlii

PAGE 10

Abstract of Dissertation Presented to the Graduate Council in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy AUTOXIDATION OF INDUSTRIAL WASTE WITH OXYGEN By Boyd T. Riley, Jr. June, 1966 Chairman: Professor J. E. Klker, Jr. Co-Chairman: Dr. C. I. Harding Major Department: Bloenvironmental Engineering Destructive distillation of wood for charcoal manufacture produces a liquid waste known as pyroligneous acid, which is driven out of the wood during the carbonizing process. Pyroligneous acid is composed of many different organic substances which are dissolved or dispersed in the water. Pyroligneous acid has a Chemical Oxygen Demand (COD) of about 100,000 mg/l, a pH of about 2.5; it is amber in color, and toxic to microflora. A 1:100 dilution of pyroligneous acid is necessary before it can be used as a biological substrate. Such a high dilution makes waste treatment by chemical rather than biological means a desirable alternative. Many different chemical treatments, among which are coagulation, chlorlnation, and pH adjustment, have been attempted in the laboratory. These experiments were completely unsuccessful. ix

PAGE 11

After reviewing the Information available on pyroligneous acid, autoxldatlon as a treatment process was considered. Many of the compounds found In this acid are toxic to microflora and act as Inhibitors of autoxldatlon reactions. Pretreatment of the pyrollgneous acid by partial autoxldatlon might selectively reduce the toxicity of the waste. If so, this pretreatment could be followed by conventional biological treatment. A 67-liter chemical reactor was used to autoxidlze partially the pyrollgneous acid at about 125°C and the equivalent of 100 psi at 0°C oxygen partial pressure. This type of treatment yielded about a 10 per cent GOD reduction in the waste during a khour period. The toxicity of the treated waste was measured by determining the loading necessary for optimum response on a laboratory-scale trickling filter. Results indicated that the toxic compounds were not selectively oxidized. Using a bench-scale reactor, complete treatment of pyrollgneous acid by autoxldatlon was investigated and found to be technically feasible. Minor variations in experimental technique were found to cause large variations in the rate and order of the reaction. The most rapid reaction rates were obtained when undiluted pyrollgneous acid was rapidly mixed with high pressure oxygen at 1?5°C and the equivalent of 150 psi at 0°C oxygen partial pressure. Two rates were observable: a very fast rate during the first 16 minutes of the experiment which resulted In an 80 per cent reduction in the GOD of the waste, and a slower rate x

PAGE 12

which lasted for the duration of the experiment (1 hour) and resulted in a total COD reduction of 37 per cent. The reaction was strongly exothermic during the first minutes of the experiment, resulting in a 50° C temperature rise of the reactants. Changes in the temperature at which the reactions were initiated caused large changes in the reaction rates, with increases in temperature causing greater than linear increases in reaction rates. The reaction rates accelerated with increases of the oxygen partial pressure at which the reactions were initiated. Partial pressures of 50 psi and less reduced the reaction rates. Complete mixing was necessary to maintain oxygen in solution at any partial pressure. Two other wastes and two pure compounds were also investigated, but were not found to autoxidize as rapidly as pyroligneous acid. A preliminary design was proposed for a 15.000 gal/day pyroligneous acid treatment plant. An economic-feasibility study, based on this design yielded a cost of 1.4^/lb. of COD removed. xl

PAGE 13

CHAPTER I INTRODUCTION Description of the Waste The development of a technically and economically feasible process for the treatment of pyrollgneous acid, a waste product of the wood distillation industry poses a challenging problem for the Environmental Engineer. This dissertation suggests a solution to that problem. The research conducted in the preparation of this dissertation was confined to the treatment of waste produced by Cabot Corporation's Gainesville, Florida, plant. This waste results from Cabot's use of the retort operation in the processing of coniferous woods. The retort operation is a batch process of destructive distillation. (See Figure 1.) The stump wood is placed into rectangular slatsided iron crates which are mounted on rallcars. A train of several crates is assembled and moved into a retort by a small engine. After the retort is sealed, it is heated by oil burners to about 800°F, with the atmosphere inside the retort kept under reducing conditions. Essentially all of the liquids present in the wood are driven out as the wood is carbonized. After the wood has remained in the retort for 8 hours, the hot retort is opened and the charcoal is moved to cooling 1

PAGE 14

Wood Storage Wood Retort Vapors Hot Charcoal Cooling Chamber Cold Charcoal Classification Small Pieces of Charcoal Briquet Plant Briquet Sales Condenser Liquids Lagoons Seepage Heavy Oils Steam Distillation Y Oil Sales Sales of Large Pieces of Charcoal Figure 1. Flow Diagram of Wood Distillation Process

PAGE 15

chambers which are similar to retorts in construction; the hot charcoal ignites as soon as it comes into contact with oxygen. After the cooling chamber is sealed, the enclosed oxygen is used up and the flames disappear. The charcoal cools for 8 hours, after which it is removed from the cooling chambers and stored. After size classification, the large pieces are sold as raw material for the production of activated carbon, and the small pieces are pulverized, mixed with starch, and pressed Into briquets. The vapors pass out of the retort and through condensers. The liquids are pumped to unlined settling lagoons where the heavy tars and oils are separated from the aqueous portion of the liquid. The concentrated oils and tars are pumped from the bottom of the settling lagoons to a steam distillation column where they are further fractionated, barreled, and sold. The aqueous portion, known as acid water or pyroligneous acid by the industry, seeps through the sandy walls of the lagoons and becomes surface drainage. Acid water is either bacteriostatic or toxic, depending on the circumstances. Veijolo and Mustakallo (1) separated wood tars into various gross fractions by solvent extraction, and tested the effects of these fractions on biological systems. They concluded that the resin acid fraction was primarily responsible for the bacteriostasis , and that these resin acids were effective in either the acid or

PAGE 16

salt form. The following table summarizes a few of the characteristics of acid water as sampled at Cabot Corporation. Table 1 Summary of Acid Water Characteristics Color Dark Amber COD 100,000 mg/1 Iron 73 mg/1 Orthophosphate 490 mg/1 Hardness 1,500 mg/1 as CaC03 Phenol 750 mg/1 pH 2-3 Figure 2 shows dilutions of both neutralized and un-neutralized acid water, also from Cabot Corporation. Major quantitative differences have been reported in the composition of acid water obtained from the destructive distillation of coniferous woods as compared with acid water obtained from the destructive distillation of deciduous woods. The qualitative analyses of both types of acid water are similar (2). Wise and Jahn (3) have compiled a table from several sources listing 213 different organic compounds which have been Isolated from pyroligneous acid. The names of all the compounds listed by Wise and Jahn are presented in Table 2.

PAGE 17

•

PAGE 18

Table 2 Compounds Formed by Wood Carbonization Ammonia Formic acid Methylamlne Acetaldehyde Methyl formate Glycolic acid Ethyl alcohol Acrylic acid Allyl alcohol Methyl acetate Propionic acid Propyl alcohol Succinic anhydride Crotonaldehyde 3-Butenoic acid Crotonic acid Methacrylic acid, polymer 2-Buten-l-ol Isobutyrlc acid Isobutyl alcohol 1 ,2-Dimethoxyethane 2-Furoic acid 2Cy clopenten1-one 3-Methylfuran Cyclopentanone Angelic acid 2 , 3-Pentanedlone 3-Pentenolc acid -j''Hydro xyvalero lactone Levullnic acid «-Hydroxy-r-hydro xyvalero la cto Pentene Isovaleraldehyde 3-Pentanone Valeraldehyde Methyl butyrate Methyl lsobutyrate 4-Hydroxy-2-pentanone Pentane Propionaldehyde dimethyl aceta Benzene 5-Methyl-2-furfuraldehyde Catechol 4-Methyl 2-furoic acid Maltol Pyrogallol 2-Methyl-2-cy clopenten1-one Formaldehyde Methanol Glyoxal Acetic acid Glycolaldehyde Tri methylamlne Dlmethylamine Acetone Propionaldehyde Hydro xypropanone Isopropyl alcohol Methylal Furan Biacetyl Butyro lactone Methacrylic acid 2-Butanone Butyric acid Methyl Propionate Acetaldehyde dimethyl acetal 2-Furfuraldehyde Pyridine 2-Methylfuran Furfuryl alcohol Tiglaldehyde Methyl crotonate 2-Pentenoic acid Tiglic acid £Hydro xyvalero lactone Ace to xypropanone ne 3-Methyl-2-butanone 2-Pentanone Pivalaldehyde <*-Methylbutyrlc acid 1 ,2-Cyclopentanediol Isovaleric acid Valeric acid Isoamyl alcohol 1 Phenol 2-Furyl methyl ketone 5-( Hydro xymethyl )-2furfuraldehyde Methyl 2-furoate Pyrocinchonlc anhydride 2-Plcolene i»— Methy 1-2cy clopenten1-one

PAGE 19

Table 2 (Continued) 2,5-Dimethylfuran 1( 3 °r ^)-Cyclohexandione 2-Hydroxy-3-methyl-2-cyclopenten-l-one Gyclohexanone Mesityl oxide 2 , 3-Hexanedlone Hydroxy-2-propanone propionate Cy clone xanol 3-Hexanone 2-Methyl-3-pentanone Isocaprolc acid Methyl valerate o-Cresol p-Cresol l-Methoxy-2 , 3-dlhydroxy benzene Lutldine Propylfuran 1-Heptyne 2-Ethyl-2,3-dihydro-5-iiethylfuran 3-Hexen-2-one l-Hydroxy-2-butanone acetate 4-Methyl-2-pentenoic acid Levoglucosan Tetrahydro-2 ,5-dimethylf uran Butyl methyl ketone Caproic acid «-Methylvaleric acid Toluene m-Cresol Guaiacol 5-Methyl pyrogallol 1Methyl2cyclohexen-5-one 2,3,5-Trimethylfuran Cyclohexanecarboxaldehyde Dlemthylcyclopentanone Butyrone Methyl caproate Heptane mXylene 2,3-Xylenol 3,5-Xylenol o-Methylguaiacol Me t ho xy-£homo catechol Methylcyclopentanone Trimethylcyclopentanone Capryllc acid Dl hydro xycapry lie acid Cumene 3 » 5-Diraethylguaiacol Homoveratrole 5-propylpyrogallol Amy If uran Cyclohexaneproplonaldehyde Pelargonic acid Estragole Isoeugenol Durene ^-Propylgualacol A5-propyl-monomethyl ether of Camphene Nopinene Sylvestrene Terplnolene Borneol Fenchyl alcohol «-Terplneol . , . Gaprlc acid 1 ,3, 3-Trlmethylbicyclo (2.2.2 )-5-o6ten-2-on< 5-Propyl-l,3-dlmethyoxy-2-hydroxy benzene 2,5-Difurfuryledine-l-cyclopentanone Cadlnene Pentadecane 5-Heptenoic acid Enanthaldehyde Enanthic acid Benzofuran o-Ethylphenol 2,4-Xylenol Creosol 2,6-Dimethoxyphenol 2,*4— Dimethyl-4-cyclohexene1-one 3-Isopropyl-2-cyclopenten. 1-one 3 , 6-0ctanedione Methyl enanthate 4-Vinylguaiacol Pseudocumene 4-Ethylguaiacol 2 , 6-Dimethoxy-^-methylphenol Isophorone 2,^, — Triraethylcyclohexanone 3i3»5-Trlmethylcyclopentanone Naphthalene Eugenol Gymene Thymol 2 ^-Dimethoxy-^ethylphenol pyrogallol Limonene Plnene y-Terpinene Camphor Gineole Isofencyl alcohol

PAGE 20

Palmitic acid Heptadecane Chrysene Retene Oleic acid Stearic acid Octadecane Nonadecane Abietic acid Plmarlc acid Arachidlc acid Eicosane Heneicosane Behenic acid Docosane Tricosane Lignoceric acid Melene *-;:ethyl-/9-ethylacrolein No single compound accounts for more than a few per cent by weight of the organic material in the acid water. Acetic acid is one of the more abundant compounds (2). The preceding information suggests that either an insensitive, broad-spectrum, efficient treatment process must be devised or some use must be found either for the acid water or for products made from the acid water which would provide the industry with financial returns. Review of Applications for Wood Tars and Pyro ligneous Acid Considerable work has been reported in the literature concerning uses for wood tars, although little information is available on acid water alone. Hedstrom and Ballin (*0 own a Swedish patent on a process for treating halogenated wood tars with polysulfides under heat and pressure to produce a product which resembles rubber. Lindhe (5) developed and patented a process for oxidizing wood tars after the light oil fractions had been removed. Air is used as the oxidizing agent and the temperature of the process may range from 100°C to 300°G. Pitch of any desired melting point may be produced by controlling the length of the oxidation as well as the temperature. Jonsson (6) suggested that wood tars could be used as

PAGE 21

a fuel in piston engines. A mixture of 75 to 80 per cent pine stump resins and 20 to 25 per cent spruce resins produces the best results. Eidus (7) proposed that wood tars be used as motor oils and as motor oil additives. A commercially owned French patent describes a process for hydrogenating wood tars (8), which is carried out at from 300°C to 500°C and 20 to 80 kg/sq cm hydrogen pressure. Various combinations of aliphatic and alicyclic compounds may be produced by the proper selection of the reaction conditions. The products can be used as lubricants, solvents, and transformer oils. Bond (9) has developed a process for "sweetening" hydrocarbon distillates containing mercaptans. Wood tars are mixed with the petroleum distillates and air is bubbled into the mixture while the temperature of the mixture is held at 5^°C. The wood tars aid in the destruction of the mercaptans in the hydrocarbon distillates. Uranov e_t a_l. (10) proposed that wood tars be used as an adhesive in the manufacture of charcoal briquets, rather than the commonly used starch. Mukherjee and Srivostava (11) developed a process for manufacturing crude plastics using sugar, molasses, phenol, creosote, and wood tar as raw materials. A somewhat similar process has been proposed by Lyass et al. (12) for the production of bakelite-type plastics. The raw materials for this process are: one part wood tar, three to five parts formaldehyde, traces of hydrochloric acid, one hundred parts sand, and 1^ per cent acetone.

PAGE 22

10 Kozlov e_t al . (13) have described a process incorporating flotation agents made from pyrollgneous acid and wood tars for processing copper sulfide and zinc sulfide ores from Ural mountains. Chertkov and Zrelov (14) stated that wood tars were not adequate as antioxidant additives in hydrocarbon motor fuels and that these additives should be replaced with amino-phenols. This statement Implies that wood tars have been tried as antioxidants in hydrocarbon fuels. The preceding applications for wood tars or acid water do not appear to offer adequate returns or a technically feasible method of treating or disposing of pyrollgneous acid. Little work has been reported In the literature concerning waste treatment processes which are capable of handling pyrollgneous acid in a suitable manner. Only two references, discussed immediately below, were found in the literature concerning waste treatment processes for pyrollgneous acid. Review of Waste Treatment Processes for Pyrollgneous Acid Szulicka and Grendysz (15) reported the treatment of waste waters from a wood distillation plant by dilution In a lagoon. The primary waste effluent of the plant contained 2,500 mg/1 of methanol, 5.3 mg/1 of acetic acid, and 200 mg/1 of phenol at the entrance to the lagoon. This effluent was found to be strongly bactericidal. Other wastes entering the pond included ground water, cooling water, and sanitary wastes from the plant. The effluent

PAGE 23

11 from the pond contained 2? mg/1 of methanol and h2 mg/1 of phenol, and supported a vigorous microflora believed to be resistant to phenol and methanol. No data were given on the overall dilution of the primary waste flow or on the retention time through the lagoon. Hickerson and McMahon (16) investigated spray irrigation as a supplement to lagooning and dilution as treatment methods for the wastes of a wood distillation plant which used hardwoods as a raw material. They determined the most suitable grasses for spray irrigation and then used spray Irrigation to supplement controlled discharge of the lagooned wastes into an adjacent watercourse. They also investigated other possible treatment processes Including pH adjustment, chemical precipitation, and autoxidation. The results of these tests were not encouraging. Their autoxidation study consisted of aerating the waste in a chemical reactor at 150°C and 100 psl. This treatment produced a black sludge which they believed would create troublesome problems in any type of filtration process. The Sanitary Engineering Section at the University of Florida carried out extensive chemical and biological investigations from 19^9 to 1951 to develop a satisfactory treatment process for the Cabot Corporation. While some technological success was achieved, the cost of treatment was extremely high. The best biological treatment system involved neutralizing the waste and then treating it by high rate trickling filtration. At a recirculation ratio of

PAGE 24

12 100:1, the GOD of the waste was reduced about 75 per cent. Both the effluent and Influent were highly colored. Clarification of the effluent was obtained with large quantities of powdered activated carbon. The chemical treatments of the waste which were investigated are: neutralization with lime, coagulation with iron compounds, coagulation with alum, acidification, chlorinatlon , chlorlnation followed by coagulation, and extraction with an assortment of organic solvents. Direct application of the waste to unpaved roads showed little promise (17). Experimental Proposal Two conclusions were suggested by the preceding information. First of all, biological treatment of the waste is unattractive because of the high dilutions necessary to reduce the toxicity of the waste and because the effluent from biological treatment requires rather expensive processing for clarification. Secondly, chemical treatments must utilize Inexpensive chemicals or regenerative processes to be economically feasible. Autoxidation appears to fulfill best the requirements of either pretreatment or complete treatment of the waste. As a pretreatment , autoxidation of the waste might partially oxidize many of the toxic compounds in the waste, with a possible reduction in overall toxicity. Certain of the known toxic compounds in the waste are similar to known antioxidants in autoxidation reactions, e.g., phenol (18,19). The waste may be amenable to complete oxidation

PAGE 25

13 by autoxidation in a few hours if sufficiently rapid reaction rates can be established and if there is no lengthy induction time preceding the reaction. Either pure oxygen or air with proportional pressure increases is an inexpensive oxidizing agent. In addition, sufficient heat must be generated by the reaction to make the process selfsustaining. Only one autoxidation system is being used in the sanitary engineering field at present on a significant scale. The system is known as the Zimpro Process and is briefly described in Chapter II of this dissertation.

PAGE 26

CHAPTER II REVIEW OF THE ZIMPRO PROCESS Description of the Zlmpro Process The Zlmpro Process Is a continuous operation for the autoxldatlon of wastes, which process was Initially developed In 1944 by F. J. Zimmerman (20). The process consists of mixing liquid wastes and air under pressures of 1,000 psl to 2,200 psi, and at temperatures of 500°F to 600°F. The carbonaceous matter in the wastes reacts with oxygen, releasing heat for the process. The heat and pressure of the process effluent are recovered and transferred to the Influent. Under typical conditions, a 1-hour reactor retention time would allow for a 90 per cent reduction In COD (20). An autoxldatlon process similar to the Zlmpro Process could prove to be applicable to the waste of Cabot Corporation. Some modifications of the process would be necessary since oxygen rather than air is the proposed oxidizing agent; but basically, the processes would be quite similar. Figure 3 is a schematic Illustration of the flow through a Zlmpro Process (21). The waste or aqueous fuel flows Into a holding tank, and Is then conveyed to a high-pressure pump. The highpressure pump acts as the pressure source for the system, 14

PAGE 27

15 u o p o a 0) as •p 08 0J to
PAGE 28

16 and as the source of head which creates flow through the system. I he waste, now under high pressure, Is mixed with metered high-pressure air from an air compressor. The mixture enters heat exchanger number 1, then Is discharged at an elevated temperature, though not sufficiently elevated to cause the autoxldatlon reaction to proceed. The waste passes through heat exchanger number 2, In which It reaches the temperature necessary for the autoxldatlon to occur at a rapid rate. The waste then enters the reactor where It Is retained long enough for the reaction to be completed. Since heat Is released during the reaction, the effluent from the reactor Is hotter than the Influent. The effluent passes countercurrently through heat exchanger number 2, or If enough heat has been aqulred In the reactor a portion of the effluent may by-pass heat exchanger number 2. The effluent stream, recomblned with the portion by-passing heat exchanger number 2, enters a gas-liquid separator. The gas phase, Including steam, nitrogen, and carbon dioxide, passes through a pressure control valve and Is expanded through a turbine to the atmosphere. The turbine powers an air compressor and a synchronous motor-generator. The power developed can be estimated from the following expression: power developed* heat released by autoxldatlon minus heat losses in the entire system. If heat losses are greater than the heat from autoxldatlon, additional heat must be added to the process. The aqueous portion of the reactor effluent Is

PAGE 29

17 withdrawn from the bottom of the gas-liquid separator and passed countercurrently through heat exchanger number 1. Finally, the effluent passes through a third heat exchanger located in the aqueous-fuel storage tank. The effluent then passes through a pressure-control valve and eventually into a natural drainage channel (21). Discussion of the Zlmpro Process The most important feature of the Zimpro Process is the production of heat during the autoxldation of an otherwise useless waste. Essentially the "waste" is no longer wasted because it is used as a fuel. The Zimpro Process must be self-sustaining once started to be economically attractive. The use of the Zimpro Process offers two advantages: first of all, it disposes of highly concentrated wastes, and secondly, it can produce quantities of heat beyond the needs of the process, which may be used in some other unit operation. The Zimpro Process, however, does entail certain problems. Oxidation occurs only in the liquid phase (20,21,22). Thus, low molecular weight organic compounds would pass through the process without being oxidized. Since molecular stability increases as the number of carbon atoms in the molecule decreases, economical operation of a Zimpro Process dictates that the stable components, e.g., methanol, methanoic acid, and acetic acid, leave the reactor in both the liquid and gaseous effluents. The liquid effluent would ordinarily have a high Biochemical

PAGE 30

18 Oxygen Demand (BOD) (23). Zimmerman (20) suggests that organic gases leaving the reactor can be oxidized bypassing them over a silver or platinum catalyst. The liquid effluent contains Insoluble inorganic ash and high BOD, both of which should be treated by conventional biological methods (21). Zimmerman (20) stated that the COD of a waste is approximately equal to the weight of oxygen necessary to oxidize the waste. He also suggested the following expression for predicting the heat value of a waste: Heat value (BTU/gal) = 48.8 x COD (gm/1). Table 3 lists some of the properties of selected materials which might be oxidized by a Zimpro Process (20). Table 3 Autoxldatlon Properties of Several Materials Material BTU/lb. lb. of 2 /lb. lb. of air/lb. BTU/lb, of material of material of air 34.34 1,780 14.30 1,450 11.53 1.220 4.60 1,365 0.77 1,565 10.90 1,370 14.00 1,330 4.87 1,455 7.55 1,395 5.70 1,335 4.13 1,410 5.75 1,365 5.14 1,270 Hydrogen

PAGE 31

19 Jackson and Brown (24) made several observations concerning the practical operation of a Zimpro Process for the treatment and chemical recovery of spent semi chemical pulping liquors. Among the more important observations were: (a) corrosion was non-existant; (b) 92 per cent of the COD was removed; (c) 90 per cent of the sodium sulfate was recovered; (d) the equipment was easy to operate and the cost of maintenance was low; (e) steam was recovered economically. The Zimpro Process has been successfully used in both treatment and chemical recovery of the following types of pulp and paper liquors: magnesiumbase spent sulfite liquor; semichemical acid, neutral, and alkaline sodium sulfite spent liquors; calcium-base spent sulfite liquor; ammonium-base spent sulfite liquor; sodium-base spent liquor; and kraft liquor (21). Koenlg (25) has presented a good summary of the characteristics of wet oxidation of sewage sludge by the Zimpro Process: (1) Oxygen pressure has no effect on the rate or the extent of the oxidation. (2) The reaction produces C0 2 and H2O plus low molecular weight acids, e.g., CHoCOOH, HC00C, etc. (3) The reaction has an Initial extremely rapid rate followed by a much slower rate. (4) The reaction reaches a different ultimate fraction converted at each temperature. (5) The reaction reaches approximately the same ultimate fraction regardless of initial concentration at low temperatures.

PAGE 32

20 (6) The reaction shows a decrease of fraction converted with increases of initial concentration at low temperatures. (7) The fraction converted is a probability function of temperature. The experiments upon which this dissertation Is based were conducted using a process similar to the Zlmpro Process, differing only in the use of oxygen rather than air as the oxidant. The use of oxygen in an industrial process appears to offer two distinct advantages. First of all, the total operating pressure of the system can be reduced, since no nitrogen partial pressure contributes to the total pressure, hence a lower equipment cost; secondly, since the oxygen can be purchased under pressure, no compressor Is necessary. These advantages are at least partially offset by the cost of the oxygen, although the results from experiments using either air or oxygen as the oxidant should be comparable if the partial pressures of oxygen are equivalent.

PAGE 33

CHAPTER III INTRODUCTION OF THE THEORY OF AUTOXIDATION OF ORGANIC SUBSTANCES General Description Autoxldatlons are defined as oxidations which can be brought about by oxygen gas at normal temperatures without the intervention of a visible flame or an electric spark (26). The preceding definition includes wet oxidation as applied in the field of Sanitary Engineering. Bateman (27) describes the complexities and vagaries of autoxidation reactions very aptly. "The mechanisms (of autoxldatlons) are complex and variable depending on constituent conditions as well as trace impurities which may inhibit or initiate the reaction." Since Bateman was speaking only of the study of "pure" compounds, one can imagine the enormously complex and dynamic reaction mechanisms which occur during the autoxidation of a heterogeneous waste. Ingold (28) describes the autoxidation of a pure compound. "The normal curve for oxygen uptake during autoxidation consists of an initial period where very little oxidation occurs, known as the induction period. This Is M The succeeding discussion is taken largely from the article cited in reference (28). The author has inserted comments to aid in bridging the gap between theory and the application of the autoxidation to heterogeneous wastes. 21

PAGE 34

22 followed by a rapid Increase In the rate due to autocatalysls by chain-branching intermediates that build up during the induction period. The rate soon reaches a maximum value and then slowly starts to decrease." Figure 4 lists the mechanisms which, in general, describe autoxldatlons. HH represents the organic substrate, R02* is the corresponding peroxy radical, and ROOH is the hydroperoxide. Initiation Reaction | lj is a general expression for the initiation process. The activation of organic species to the free radical state is caused by heat, light, ionizing radiation, or reaction with other free radicals. Hydrogen atoms, or possibly protons, are preferentially attacked in the following order: (1) hydrogens alpha to a double bond or to a conjugated system, (2) tertiary hydrogens, (3) secondary hydrogens, (4) primary hydrogens. Ingold (29) has suggested that thermal initiation may involve the reaction: RH + O2 s— R' + H02* since this reaction involvss less energy than reaction [l j# Autoxidation of wastes, called wet oxidation by Zimmerman and others, can be thermally Initiated. Since most wastes are heterogeneous mixtures, it is likely that at least minute concentrations of hydroperoxide are present In the waste. Ingold (28) states that in the presence of minute traces of hydroperoxide, reaction J4] is the dominant mode of initiation; and as the concentration of

PAGE 35

23 Initiation: r H activation^ R . + ( H . } r^] Propagation: r. + 02 *«R02» [2] H02* + RH *ROOH + R» [3] Decomposition of peroxide: ROOH =— R0« + »0H [4] 2R00H *R0« + R02« + H2O [5] §§ 2 + ROOH ~various products [6] ROOH *~nonradical products [7] Induced decomposition of peroxide: X + ROOH *free radicals [8] Y + ROOH aROH + YO [9] Z + ROOH — inactive products + Z [lo] M + ROOH *free radicals [ll] Selftermination: R02* + R02» *Inactive products [l2J Chainbreaking termination: R02« + IH *RO2H + I* [13] Figure 4. Autoxidation Mechanisms

PAGE 36

24 hydroperoxide increases, reaction [_5J becomes increasinglyimportant as an initiation mechanism. Thermal initiation of unsaturated substrates may occur as follows: H H-0-0 r-c-c=c-:-? + 0=0 — a-c=c-c-rf (30,31). H H H H This represents the direct formation of a hydroperoxide without free radical intermediates. Propagation Reaction [2] is extremely rapid except at very low oxygen partial pressures. Consequently only peroxy radicals are of importance in chain propagation and termination (27,32). In the absence of major steric effects, the rate of reaction i 3 I depends on the carbon-hydrogen bond strength and on the availability of electrons at the carbonhydro gen bond being broken ( 33 » 3^) • Peroxy radicals can add to olefinlc double bonds (35)t and tend to react more with the oxygenated reaction products than with unoxidized species, because the former frequently contain more reactive carbonhydrogen bonds (36). Decomposition of Peroxide The thermal decomposition of a hydroperoxide generally yields several free radicals, which lead to chain branching and account for the autocatalysis observed In many oxidations. In addition, the solvent plays an Important role in the decomposition of hydroperoxide, causing the rate of

PAGE 37

25 decomposition to change during an autoxldatlon as the substrate is oxidized (37). At low hydroperoxide concentrations, the rate of hydroperoxide decomposition is first order as exemplified by reaction [4] . At higher hydroperoxide concentrations there are substantial deviations from first order kinetics and apparent rate constant increases (38,39). These phenomena are probably due to the increasing importance of reactions |_5j and (_6j on the overall autoxldatlon rate. The rate of decomposition of hydroperoxides is accelerated by initiators which provide an additional source of free radicals (40). The rate may be retarded to a reproducible minimum value by the addition of inhibitors of free radicals, which Include polycyclic aromatic hydrocarbons (41). The effects of reaction |_5j relative to reaction [4 J decrease with Increasing temperature (42). The following reaction represents a mechanism which could be followed by the alkoxy radicals produced by reactions [_4J and [5] • H0» + RH »ROH + R* Reactions |_6J and [7 J represent mechanisms which terminate the chain reactions started by reactions 1 through [5] . Reaction [7] may be considered as the reverse of reactions [l] through [5] . The products of reaction [7] » to be further oxidized, must again be activated to the free radical state represented by reaction |_1J .

PAGE 38

26 Induced Decomposition of Peroxide Reactions [8 J and [llj represent catalysis of the autoxldatlon while reactions [9] and [lOJ represent inhibition. Reaction (_8_| shows the induced decomposition of a hydroperoxide by substances not containing a heavy metal. It results in the production of free radicals, which accelerate the rate of the overall reaction. Alcohols (43), ketones and ethers (44), fatty acids (45), and olefins (46) are examples of this type of compound. A typical mechanism for an olefin is as follows: ROOH + r'cH=CH2 *R02* + R* SHCH3 Increasing the surface to volume ratio of the reaction vessel or adding powdered solids may increase (4?), decrease (48), or have no effect (49) on the rate of reaction |_8J . An Increase in rate implies that reaction |_8J occurs only on a surface, whereas a decrease implies that the surface promotes either reaction (lOJ or the destruction of free radicals without starting a chain mechanism (28). Reaction [_9J describes a stoichiometric mechanism for the inhibition of autoxidatlons by decomposing hydroperoxides to stable alcohols. Inhibition of an autoxldatlon means the induced premature termination of free-radicalforming chain reactions. Sulfur and selenium compounds follow this inhibitory mechanism. Monosulfldes which contain at least one aliphatic or cycloaliphatic group attached to the sulfur atom are more effective antioxidants than

PAGE 39

27 mercaptans and disulfides, while diaryl sulfides and sulfones are inactive (50). Inhibitors containing sulfur are usually oxidized to strong acids which also act as inhibitors, as shown in reaction [lOj . Sulfur-containing Inhibitors also cause the production of resinous compounds and sludges in a system undergoing autoxldation (5D« Selenides do not cause the formation of sludges and they are more effective antioxidants than sulfur compounds (52). As a rule, selenides are not found in wastes. Reaction jlOJ illustrates the mechanism by which strong acids inhibit autoxidations. The acids are strong in the Lewis sense (53). Although acids reduce the rate of oxidation, they cause the production of resins, sludges, and corrosive compounds (5*0. Bases are not as effective as acids in reaction [lOj , but when used, resinous products usually are not formed; when they are formed, these products are selectively oxidized. Corrosive reaction products are neutralized by the additives to form compounds with good detergent properties (55). Reaction [ll] represents the induced decomposition of hydroperoxides by a heavy metal. Metals which possess two or more valency states with a suitable oxidation-reduction potential between them can react with hydroperoxides to form free radicals (56). Examples of suitable metals are iron, cobalt, copper, manganese, add nickel. The metals can act either as an oxidant or a reductant. Oxidant Ce^ + R00H »Ge 3+ + R02* + H +

PAGE 40

28 Reductant Fe 2+ + ROOH ~Fe3+ + RO« + OH" Certain metals, suoh as cobalt and manganese, can act both as an oxidant and as a reductant. For these metals the oxidant reaction may be rate-determining, since the metal is found chiefly in its more highly oxidized state (57). While ferric chloride acts as an inhibitor of the oxidation of cumene and tetralin in aromatic solvents, it acts as an Initiator in polar solvents (58). Complexes of metal ions and hydroperoxides are more active catalysts than metal ions alone (59). The anion associated with the metal can affect the catalytic activity of the metal either by affecting its redox potential or by "blocking" the formation of a catalyst-hydroperoxlde complex. Anion effects are specific for each metal and anion combination. The solvent can affect the catalytic activity of the metal-anion combination also (56). Heavy metals act as powerful pro-oxldants of organic autoxidations. Catalysis by metal surfaces is proportional to their area and therefore is appreciable only in so far as the metal Is colloidally dispersed or dissolved (60). The activity of the metal catalyst can be reduced by the addition of corrosion inhibitors or metal passivators which form a chemlsorbed film on the surface of the metal (6l). Oxidation products form a non-corrosive substrate can also laquer metal surfaces (62). The catalytic effects of metal salts reach a constant value at very low metal concentrations. This may be caused by either chain termination by the catalyst (63) or by the

PAGE 41

29 occurrence of a steady-state concentration of hydroperoxide, as when reactions [3] and [ll] proceed at the same rate (64). If reactions [3] and [ll] proceed at the same rate, only one new hydroperoxide is formed to replace each hydroperoxide that decomposes. Consequently no chains can be started. Metal catalysts can sometimes be completely precipitated from solution shortly after the start of an autoxidation, consequently such a catalyst could have no affect at all on the rate of the reaction (65)» Metal ions and surfaces may be inactivated by chelating agents such as ethylenediamine tetraacetic acid or N,N -disalicylidene-l,2-propanediamine (66). Chelating agents are effective because their stearic configuration prevents the formation of metal lon-hydroperoxide complexes, and also because they modify the redox potential of the metal ions (58). The metal chelate may be oxidized during an autoxidation, thus activating or releasing the metal ion catalyst (67). Selftermination Reaction i_12J represents the termination of a chain by the interaction of two peroxy radicals. Primary and secondary peroxy radicals terminate chains by way of a cyclic transition state (68). Tertiary peroxy radicals terminate reaction chains less readily since they lack a hydrogen atom on the cx-carbon. These radicals are believed to produce alkoxy radicals, which then dimerize (69). It is possible to reduce the rate of autoxidation of a compound

PAGE 42

30 which gives tertiary radicals by adding a compound which produces primary or secondary radicals (3^). Mixtures of compounds can result in higher overall rates as compared with the sum of the rates of autoxidation of each of the pure substltuents. Chain-breaking Termination Reaction [13] describes the mechanism by which inhibitors of free radicals reduce the oxidation rate of organic compounds. Reaction [13] was first proposed to cover phenolic inhibitors but was later extended to include substituted analines. The free radical I* Is stabilized by resonance and sometimes by stearic interference by other substituted groups. Thus I« does not Initiate a new oxidation chain (60). It will be destroyed by reacting with another free radical. Other free radical inhibitory reactions which have been Identified are: I. + R02« — — ROOI (70) I« + R0 2 » — -*— ROOH + i' (71) I* + I* --»— I2 (72) I« + !• *IH + l' (73) Many inhibitors whose active site is not protected by bulky substltuents can participate in the following reactions* I+ RH IH + R* (7*0 I0 2 * + RH — IOOH + R« (7*0 I+ 02 *102* (73) 102* + !• — 1001 (73)

PAGE 43

31 The inhibitor radical !• can react with hydroperoxides and add to double bonds (75)* The rates of uninhibited oxidations are independent of oxygen pressure, except at very low pressures (27). The rates of inhibited reactions depend on oxygen pressure. These data suggest that most inhibitors react directly with oxygen. IH + 02 — ^I" + H0 2 » (76) Oxidation of an inhibitor may either increase or decrease the effectiveness of the inhibitor. The effect is dependent on the specific conditions of a particular reaction. Several qualitative comparisons of the effectiveness of various inhibitors have been compiled (77) • The efficiency of a given inhibitor type is increased by an Increase in the electron density at its reactive center, that is, by a decrease in the redox potential of the inhibitor. If the redox potential is too low, the efficiency of the inhibitor is reduced because the inhibitor becomes susceptible to direct oxidation (60), While phenols and amines are the most effective inhibitors known, other compounds such as qulnones (78), aliphatic alcohols (79)» and aromatic hydrocarbons, particularly polycycllc aromatics (80), act as weak inhibitors. Two or more different inhibitors may act synergistlcally or antagonistically on the rate of an autoxldation. Synergistic effects give an inhibited rate which is slower than direct addition of the effects of the individual

PAGE 44

32 Inhibitors might imply. Antagonistic effects give higher reaction rates than would be reached if only one inhibitor were present. Synergistic and antagonistic effects are specific for each substrate and combination of inhibitors (23). General Discussion The preceding introduction to the theory of autoxidation is but a brief summary of volumes of material available In the chemical literature. From an engineering point of view, such theoretical explanations are valuable for planning new research and improving the economic operation of existing facilities. Each waste must be studied separately to develop an optimum autoxidatlon treatment process, because the reactions which can occur simultaneously are so complex. Three Important areas of research for applied wet oxidation processes are suggested by the theory discussed in preceding portions of this chapter. Most important is the development of techniques which tend to promote and sustain autoxidatlon chain reactions. Since rapid reaction rates are attained only when chain reactions are dominant, reactors should be designed to promote and sustain chain reactions. A faster reaction rate means smaller equipment and a lower installation cost. Other authors have pointed out that some autoxidatlon reactions are surface reactions. For some wastes a packed type of reactor might be best, while for others a tank-like

PAGE 45

33 reactor would be as effective as one with a high surface to volume ratio. Catalysis of wet oxidation offers the promise of high economic returns. Efficient catalysts should be discovered for each waste. Furthermore, the most efficient method of applying catalyst should be found. Some possibilities are: dissolving or dispersing the catalyst in the waste, lining the reactor walls with the catalyst, and making propellers for mixing the reactants of a catalytic material.

PAGE 46

CHAPTER IV MATERIALS AND TECHNIQUES FOH THE EVALUATION OF PARTIAL AUTOXIDATION OP WOOD DISTILLATION AS PRETREATMENT PROCESS Introduction The use of autoxldation as a pre treatment process was investigated first since it promised to be more economical than complete autoxldation. In assessing the effects of partial autoxldation on the waste, the experiments were conducted on a pilot-plant scale. The kinetics of the reaction were also studied while producing a substrate to study biologically on a trickling filter. Considerable equipment had to be collected and assembled to implement the research. Description of Reactor A steam-heated chemical reactor of 67-llter volume was obtained. The assembled reactor weighed 350 pounds and was mounted on a four-wheel cart. The head of the reactor could be removed by using a portable bomb hoist mounted on a rotating T boom made of galvanized pipe. The reaction chamber of the reactor and all fittings were made of 316 stainless steel. The reactor was rated to operate up to 250 psi and 150°C. The maximum temperature which was attained in our laboratory was about 130°C because of the Inaccessibility of saturated steam at 3^

PAGE 47

35 pressures greater than 50 psi. The reactor was composed of three primary sections. Figure 5 shows an overall view of the assembled reactor. The cylindrical base section of the reactor had a dished bottom. The reaction chamber was surrounded by a steam jacket. Condensate was removed from the bottom of the steam Jacket and water was circulated through the steam jacket to cool the reactor. The reaction chamber was sampled and drained through a 2-lnch diameter threaded opening located in the center of the dished bottom. A ^-inch high pressure needle valve was attached to a |-inch x 5-inch bushing by a i-inch nipple. The female half of a nut-union was attached to the outlet side of the needle valve with a |-inch nipple. The male half of the nut-union was attached to a ^-inch pipe about 6 inches long which served as a sample collector. The lower end of the pipe was sealed with a i-inch cap. The sample volume was about 30 nil. The liquids in the reaction chamber were mixed by a four-bladed propeller, which was attached to a 1-inch diameter stainless steel drive shaft. The drive shaft entered the reaction chamber through the top of the reactor. Figure 6 is an interior view of the top of the reactor which shows the relative size of the propeller and drive shaft. A standard asbestor-f illed packing box was used to maintain a pressure-tight seal around the drive shaft. Silicone stopcock grease was used as a lubricant in the

PAGE 48

36 Rupture Disc Assembly Drive System for Mixing Reactants Condensate Drain Line Steam Line to Reactor Cart Gas Inlet Valve , Cross, Pressure Sensor Fitting, Gas Exhaust Valve Cooling Water Line Sampling Tube Figure 5. Exterior View of 67Liter Reactor

PAGE 49

37 Figure 6. Interior View of Reactor Top

PAGE 50

38 packing box rather than hydrocarbon grease, which can react violently with high-pressure oxygen. An electric motor attached to a speed-reducing gear box drives the shaft at about 80 RPM. A ^-inch x 8-inch section of stainless steel sheet metal was formed into a propeller and attached to the top of the drive shaft, Just inside the reactor. This propeller was intended to mix the gases in the reaction chamber. Aside from the packing box and drive system, there were two ports in the dished top. (See Figure 5.) A 2-lnch diameter port was fitted with a 2-inch, horizontally oriented tube which was welded in place. The tube was about 18 Inches long and terminated in a machined flange. This flange and a matching flange were used to hold a nickel rupture disc, rated to burst at 350 psi and 72°C, to minimize the hazards of experimental explosions. The atmosphere of the reactor was found to be highly corrosive to the nickel discs. The corrosion rapidly reduced the bursting strength of the discs. Coating the rupture discs with silicone stopcock grease greatly retarded the corrosion. A 5-inch diameter cross was attached to a port in the top of the reactor. A needle valve was attached to one horizontal leg of the cross. Oxygen was fed into the reactor through this valve from a standard high-pressure oxygen cylinder. An Airco pressurereducing valve was used to control the oxygen pressure in the reactor. The other horizontal leg of the cross was attached

PAGE 51

39 to a tee. The 90° leg of the tee was fitted with a l/l6-lnch tubing fitting which served as a connection for a pressure-measuring device. The horizontal leg of the tee was connected to a needle valve. This needle valve was used to exhaust the gases in the reactor at the end of an experiment or to decrease the pressure in the reactor during an experiment. The vertical leg of the cross was fitted with a thermocouple which extended about 12 inches Into the reaction chamber. The thermocouple was removed to add waste to the reactor before starting a run. The pressure and temperature inside the reactor were measured continuously during an experiment. The readout of the sensor signals was accomplished with a Brown Sixteen Point Printing Recorder. All the even numbered recorder inputs were wired together and connected to the pressure transducer. Likewise all of the odd numbered recorder inputs were wired together and connected to the thermocouple. The recorder read and printed, alternately, pressure and temperature at 15 second Intervals. Appendix 1 gives circuit diagrams and calibration curves for the sensor systems. The reactor was insulated with 1-inch thick fiberglass Insulation which was covered on both sides with aluminum foil. The Insulation helped maintain a constant reactor temperature and prevented burns while working with the reactor. Since samples had to be taken at known time intervals, a stopwatch was used as a time monitor.

PAGE 52

40 Experimental Technique for Reactor The experimental technique is given in a step by step manner as follows i 1. A sample of waste was collected from the waste lagoons at Cabot Corporation. (Grab samples were taken. ) 2. The dilution of the waste selected for any given experiment was prepared with distilled water. 3. Twenty liters of the diluted waste were poured into the reactor. 4. The waste and reactor atmosphere were purged with nitrogen gas for 15 minutes. This created an inert atmosphere in the reactor, which prevented premature Initiation of the autoxidation reactions. 5. The reactor was sealed by screwing the thermocouple into place and closing all valves. 6. Mixing was begun. 7. The reactor was brought to the desired temperature by opening the valve on the steam line and opening the valve on the condensate drain line, 8. The temperature was held constant by adjusting the condensate drain valve until the flow of steam through the steam Jacket gave the desired temperature. At temperatures around 100°C, it was necessary to open and shut the steam-inlet valve rather than use a continuous bleed system. 9. Oxygen was added to the reactor first by

PAGE 53

41 pressurizing the oxygen feed line, then by opening the reducing valve on the oxygen cylinder, and by opening the needle valve attached to the cross on the top of the reactor. The pressure in the reactor was adjusted with the pressure-reducing valve. If the pressure was accidently raised too high, the exhaust valve could be opened briefly to lower the reactor pressure. 10. The needle valve connected to the bottom of the reactor was cracked open and any nonrepresentatlve material was flushed out of the valve. This valve was then closed. (This needle valve will be called valve "A" in the rest of the procedure. ) 11. The sampling tube was attached to the nut-union and submerged in a liquid nitrogen bath. 12. Simultaneously, the stopwatch was started and valve "A" was opened. 13. After the sampling tube was filled, valve "A" was closed. 14. Within 40 seconds after having entered the sampling tube, the sample was completely frozen. The sampling tube was then disconnected from the reactor. 15» The sample was thawed until it could be removed from the sampling tube and transferred to a vial for storage in a refrigerated area. 16. The sampling tube was washed, rinsed with distilled

PAGE 54

42 water, and drained so that the next sample could be collected. 17. Steps (10) through (16) were repeated for the collection of successive samples, except that the stopwatch was started only with the first sample and let run continuously for the duration of the experiment. 18. After the last sample was collected, the steam was shut off. When the pressure in the steam Jacket had bled out through the condensate drain line, cool water was flushed through the steam jacket until the reactor returned to room temperature. 19. The pressure in the reactor was vented to the atmosphere. 20. The partially oxidized waste was drained from the reactor and stored in Jugs to serve as substrate for a trickling filter. 21. The reaotor was washed and rinsed. Samples were frozen as they were collected for two reasons. First of all, chilling or freezing essentially stops the oxidation, and secondly, freezing the sample made it possible to remove the sampling tube from the reactor without losing the sample through vaporization. The best technique for freezing the sample proved to be to Immerse the sampling tube in the liquid nitrogen bath before the sample was collected. The reaction was stopped as soon as the liquids contacted the cold tube, before the

PAGE 55

43 sample was frozen. The container for the liquid nitrogen bath was a wide-mouth vacuum-bottle liner. To reduce the hazards of flying glass in the event of an accident, the vacuum liner was wrapped with plastic tape. The samples collected during the autoxidation were analyzed by the Chemical Oxygen Demand test (GOD) as described in Standard Methods for the Examination of Water and Waste Water (81). This test gave a measure of the strength of the partially autoxidized waste relative to the strength of the raw waste. Appendix 2 describes some minor variations from the standard GOD procedure which were used throughout this research. Description of Trickling Filter A trickling filter using Dow-Pac @ medium was selected because of its resistance to shock loads, as the best means of evaluating the toxicity of the treated waste. The wall of the filter consisted of three one-quarter sections of 8-inch diameter vitrified clay pipe. The pipe and medium were supported by a wooden stand which allowed air to circulate vertically through the filter. The base of the stand was in a 67-llter tank. Substrate was cycled from the tank through the filter and back into the tank. The tank outlet was about 4 inches from its bottom, and consequently acted as a sludge storage area. Flow through the filter was maintained between Z\ and 3 liters per minute and a small centrifugal pump furnished the head necessary to maintain flow through the system. Figure 7 shows the

PAGE 56

44 Figure 7. Trickling Filter

PAGE 57

^5 trickling filter. The flow was distributed over the filter surface with a horizontal section of pegboard. Substrate flowed over the pegboard from the hose leading from the pump. The edges of the pegboard were raised so that about i inch of head was developed above each of the perforations. The flow passed equally through each perforation and into the filter medium. The 1/8-inch diameter perforations in the pegboard were arranged in a square pattern on 1-inch centers. The studies were started using a batch process. If the results had been encouraging, the trickling filter could have been converted into a continuous system. Experimental Technique for Trickling Filter The primary problem to be solved by experiments with the trickling filter was that of determining the effectiveness of autoxidation on reducing the toxicity of the waste. An experimental technique which provided the desired information is given: 1. The contents of the tank were mixed. 2. A grab sample was collected and filtered, and its COD determined. 3. Substrate was poured into the tank, and the tank's contents were throughly mixed. 4. A grab sample was collected and filtered, and its COD determined. 5. The procedure was repeated at the same time daily. The samples were filtered through a fiberglass mat.

PAGE 58

46 Filtration of the samples was necessary since, in an industrial process, most of the suspended matter would be removed by sedimentation, and only soluble material would contribute significant amounts of COD to the filter. By sampling the tank directly and determining the COD of the sample, as well as the COD of raw Cabot waste, it was possible to determine the dilutions of partially autoxidized waste necessary to maintain an active filter relative to the strength of the raw waste. This value, when compared to dilution values for the raw waste, indicated the effectiveness of partial autoxidatlon as a pretreatment.

PAGE 59

CHAPTER V RESULTS OF THE INVESTIGATION OF PARTIAL AUTOXIDATION AS A PRETREATMENT PROCESS Experimental Plan Prior to this study, no Information was available concerning the autoxidation of pyroligneous acid. Since explosions are not unexpected when dealing with highpressure oxygen systems, it was felt that a cautious approach should be made in the experiments. The plan was to start with mild oxidizing conditions and high dilutions of the waste, and gradually to proceed to the maximum oxidizing conditions and the highest concentrations of the waste. Three factors affected the rate of the reactions i concentration of the waste, temperature, and the partial pressure of oxygen. The partial pressure of oxygen determined how much oxygen dissolved in the waste at any given temperature Determination of Oxygen Partial Pressure Appendix 3 gives data concerning the solubility of oxygen in water at elevated temperatures and pressures. The rate of mixing In the reactor affected the rate of oxygen dissolution. This had an effect on the overall rate of the autoxidation. Each experiment was started at a different temperature, oxygen pressure, and concentration 47

PAGE 60

k8 of waste. The temperature and oxygen pressure were held as constant as possible while the change in the COD of the waste was measured. The partial pressure of oxygen could not be measured directly, so the total pressure was measured and the partial pressure of each of the components deducted from the total pressure. The result obtained was the partial pressure of oxygen in the reactor at the temperature of the experiment. The oxygen pressure was converted to a standard temperature for comparative purposes. Steam pressure was taken from readily available information on saturated steam pressures at various temperatures. The assumption that saturated steam rather than superheated steam was present in the atmosphere of the reactor must be valid since there was liquid water in the reactor during each run. The pyroligneous acid was assumed to behave like water as far as steam pressures were concerned. While the acid water undoubtedly did not behave like pure water, the assumption of equality between the two liquids was valid within the anticipated limits of accuracy in the experiments. Appendix 4 gives the variation of saturated steam pressure with temperature. The pressure monitoring system was sensitive to a pressure change of 1 psl. As autoxidation occured in the reactor, some carbon dioxide was produced as an end product of the reaction. The carbon dioxide produced by the reaction in turn produced partial pressures great enough to be measured by the

PAGE 61

k9 pressure sensing system. This resulted in a decrease in the effective partial pressure of oxygen, directly proportional to the increase in partial pressure of carbon dioxide. The build-up of carbon dioxide was not rapid enough to appreciably affect the oxygen partial pressure during any experiment for two reasons. First of all, in most experiments the dilution of the waste was great, and even if all the carbon atoms had been converted to carbon dioxide, there were not enough carbon molecules in the waste to affect the pressure in the reactor to any measurable extent. Secondly, the gas volume in the reactor was large, about 47 liters. To create a measurable carbon dioxide partial pressure several grams of carbon dioxide would have to have been released by the reaction. The only experiments that might have been affected by carbon dioxide build-up would have been those using undiluted waste under the most extreme oxidizing conditions. Appendix 5 shows a theoretical calculation of carbon dioxide build-up as a result of autoxldation. If the partial pressure of carbon dioxide was approximately zero, then the partial pressure of oxygen was constant. Any oxygen consumed by the waste was replaced from the oxygen cylinder, since the cylinder remained open and connected to the reactor during each run. Appendix 6 shox-rs a typical calculation for the computation of an oxygen partial pressure in an experiment. A similar procedure was used to compute oxygen partial

PAGE 62

50 pressures in all other experiments. Results of Autoxldat Ions Figure 8 compares the reaction rates with two 1:100 dilutions of acid water under different oxidizing conditions. The average rates during the first hour of the reaction indicated that temperature was important in determining the rate of the reaction. Run 2 followed first order kinetics, while Run 3 followed kinetics which were greater than first order. If an 85 per cent GOD reduction is considered complete treatment of the waste, both rates are veryslow. Figure 9 gives the rate with a 1:50 dilution of acid water. The reaction followed two different first order rates. The rate during the first hour of the reaction was approximately the same as the rate of Run 3, even though the initial waste concentration of Run 3 was about onehalf the initial waste concentration of Run 4. Figure 10 compares the reaction rates of three different dilutions of acid water. All three reactions followed first order kinetics. Sensitivity to the partial pressure of oxygen is shown by the reaction rate for Run 6. Decreasing oxygen partial pressure appeared to limit the reaction rate, but increasing the pressure of oxygen did not promote proportional rate increases. Figure 11 compares the reaction rates of four runs on the same dilution of acid water. The initial concentrations were not Identical because of differences In the

PAGE 63

51 o a p o o 100 80 60 40 20 10 Run Symbol Avg. Temp . P °G 2 3 (OQ) no -O) 133 30

PAGE 64

52 300 200 o

PAGE 65

53 CM O T-t X \ S Q O o 100 80 60 40 20 10 60 120 180 Time (rain) Figure 10. GOD vs. Time 240 Run Symbol 5 6 (A(O Avg. Temp, -A) 12? -O) 129 (O D) 129 Avg. pH P , @ 0°C 2 (psl) 116 47 116 2.5 2.5 2.5 Waste Dilution lllO 1:25 1:20 Avg. Hate For First Hour mg/1 min 15.7 2.5 12.0

PAGE 66

5h 600 o M bO s Q o o 400 200 60 120 180 240

PAGE 67

55 samples of acid water. The importance of temperature was clearly shown. The data suggest that reaction rates were primarily a function of temperature. A comparison of the reaction rates for Huns 7, 14, and 15 supports the assumption that, after a certain oxygen partial pressure is attained, further increases in oxygen partial pressure contribute little toward increasing the rate of the autoxldatlon. The reactions followed first order kinetics. Figure 12 gives the reaction rates of three runs at higher concentrations. Hun 9 represented a concentrated sample, but because of dilution of the acid water in the lagoon at the time the sample was collected, the GOD of the sample was about the same as that of a 1:1 dilution of waste collected on another date. The reaction rate followed first order kinetics. The rates of Huns 9 and 17 were substantially greater than any other runs. This demonstrates that a significant Increase in the concentration of the waste results in an Increase in the rate of the reaction. Figure 13 shows the effects of altering the pH of the waste before the autoxldatlon is attempted. A calcareous sludge was used to adjust the pH of the acid water. The sludge was obtained from the Gainesville water softening plant. The rates of all three runs were inhibited at a COD of about 2,100 mg/1. Run 12 showed an Increase in GOD during the first 30 minutes of the experiment. The data supplied no explanation for this phenomenon, although such a phenomenon would not be desirable in a waste

PAGE 68

56 600 I o to e Q o o 400 200 100 Hun Symbol Avg. Temp, °G 9 (O O) 16 (A17 (D133 126 125 -A) 120 Time (mln) 2^0 Figure 12. GOD vs. Time Avg. P @ 0°C 2 (psl) 93 101 81 PH Waste Dilution 2.5 none 2.5 1*5 2.5 1:1 360 Avg. Rate For First Hour mg/1 min 25.0 11.7 ^5.0

PAGE 69

57 500 o bO P o o 10 11 12 400 300 200 100 Hun Symbol (O O) (D-Q) -A) 30 60 90 120 150 Time (mln) Figure 13. GOD vs. Time 180 210 Avg. Temp, °G 12? 125 130 Avg. Pn_ ® 0°C o 2 (psi) 92 110 108 pH 6.5 6.5 8.0 Waste Dilution 1:20 1:20 1:20 Avg. Rate For First Hour mg/1 min 8.8 10.8 2.2

PAGE 70

53 treatment process. Since neutralization of the acid water prior to autoxidation did not accelerate the rate of the reaction, no additional preneutralization experiments were attempted. The autoxidation experiments at a pH of about 2.5 all followed a pattern of changes which vrere visually perceptible in samples removed during a run. The 1:20 dilutions of the waste resulted in the most clearlyperceptible of these typical visual changes in the waste, as seen in Figure 14. Sample 1 was collected just before oxygen was added to the reactor, while sample 2 was collected Just after oxygen was added to the reactor. The remainder of the samples were collected at thirty minute intervals. The first visual change in the waste appeared as the waste was heated prior to the addition of oxygen. Heating partially broke the emulsion of tars in the waste and caused these tars to adhere to the walls of the reactor. This effect caused the COD of the waste to drop slightly before any oxidation occured. As soon as oxygen was added to the reactor, the oils and tars in the acid water underwent an extremely rapid polymerization reaction. The particulate polymers which were formed caused the waste to appear to be very dark In color. In addition, the waste acquired strong surfactant characteristics which resulted in stable foams. As the autoxidation oontinued, the waste gradually acquired a yellow color. The polymers which were formed at the start of the reaction were oxidized,

PAGE 71

59 1 j (£ (5 14 W »« »* »*' ImI 5 (o 7 a q io mill Figure 14. Typical Samples From Autoxldation Experiments

PAGE 72

60 and the waste gradually lost Its surfactant properties. The presence of solids and stable foams In the samples caused some difficulty in obtaining reproducible GOD values. Appendix 7 gives triplicate GOD values on Hun 7. The samples in which the initial COD was above 10,000 mg/1 did not become clear during the experiment. Hesults of Trickling Filter Experiments The substrate for the toxicity experiments was the final reactor effluent in the autoxidation experiments. In addition to the carbonaceous substrate (acid water), several salts were added to the filter daily. These salts were added because they are generally acknowledged to be essential to bacteriological growth. The trace elements insured that the toxicity of the acid water, and not the lack of some essential element, would limit the response of the trickling filter. Appendix 3 lists the salts used and the amounts which were fed. The trickling filter was acclimated to the partially oxidized acid water for about one month prior to collection of any data. Figure 15 gives the results of feeding the partially oxidized acid water to the trickling filter. The vertical lines on the graph represent the COD Increases in the substrate which resulted from the addition of acid water. The diagonal lines represent the response of the filter during a 2^-hour period. The dashed line Indicates that the filter was still "alive" ten days after the last

PAGE 73

o 61 / / / / / / o CA CM o CM 00 c3 Q o C3 -p a a u 5 £ Eh o u to O Jvr\ O O O CM O O VA O o o o o H O > »d H •H s a) •p 4-> W ,0 00 f V 2 -) GOO

PAGE 74

62 dally feeding, on day thirty. The optimum response of the filter occurred at a GOD of about 1,000 mg/1. Since a it 100 dilution is required to treat raw acid water by trickling filtration, and the ratio of the partially oxidized acid water is 1,000:90,000 or lt90, no substantial reduction in toxicity has been attained by partial autoxidation of the pyroligneous acid.

PAGE 75

CHAPTER VI CONCLUSIONS FROM THE STUDY OF PARTIAL AUTOXIDATION AS A PRETREATMENT PROCESS The experiments already described in this dissertation provide considerable information on the relationships of pyroligneous acid to autoxidation. The main purpose of this phase of the research was to determine if partial autoxidation of pyroligneous acid would greatly reduce the toxicity of the waste so that complete treatment of the waste could be effected by conventional biological methods. The data indicate that the toxicity of the waste was not reduced any more than the COD of the waste was reduced. Consequently, partial autoxidation of the waste does not appear to be effective or desirable as a pretreatment for pyroligneous acid. Either the toxic components of the waste are not oxidized or they are partially oxidized to other compounds which are equally as toxic as their predecessors. The data show that the acid water is autoxidized during all experiments, although the reaction rates were slow compared to the strength of the waste. The data imply that if the proper conditions were discovered, complete autoxidation would be possible. Succeeding chapters of this dissertation support this contention. 63

PAGE 76

No induction period was observed at the start of the experiments. Consequently, the absence of an Induction period caused the complete oxidation of pyrollgneous acid to appear feasible by autoxldation techniques. These data suggest that the reaction rates are dependent primarily on temperature, although the partial pressure of oxygen also affects the rate at any given temperature. The reaction rates increased directly with the temperature and oxygen partial pressure, although the increases do not appear to be linear with either of these variables. An increase in the concentration of the waste caused a relatively large increase in the reaction rate. The increase In rate was considerably less than the increase in the concentration. The rate of oxygen transfer from the gas phase into solution limited the increase of the reaction rate as the GOD or concentration of the waste was increased. Neutralizing or raising the pH of the waste did not appear to be desirable. The preceding conclusions resulted in the study of autoxldation as a technique for the complete oxidation of the wood distillation waste. The following chapters of this dissertation explain the study in detail.

PAGE 77

CHAPTER VII EQUIPMENT FOR THE INVESTIGATION OF COMPLETE AUTOXIDATION OF WASTES Introduction The reaction rates obtained in the 67-liter reactor under the most extreme oxidizing conditions possible with the reactor, were not rapid enough to be used for the complete autoxidation of the waste in an industrial process. The data suggested that the reaction rates could be accelerated greatly by increasing the temperature at which the reaction was carried out and by increasing the rate of oxygen transfer from the gas phase to the liquid phase during the reaction. Since pressure is dependent on the temperature, an increase in the temperature capabilities of the reactor necessitated the use of a high pressure reactor. Description of Reactor and Heater A 2-liter, electrically heated Parr reactor, capable of operating at temperatures up to 350°C and pressures up to 1,000 psi, was obtained. The aparatus consisted of two sections, a 220volt heater and controls, and the reaction chamber, or bomb, and controls. The 220-volt heater was composed of two concentric cylinders. The opening between them was enclosed on both 65

PAGE 78

66 ends with annular aluminum plates and was filled with vermiculite insulation. Heat was supplied by six 250-watt strip heaters, and the heat output was controlled by a General Radio Go. Type W2HM Auto transformer. A Sargent Model T Thermomonitor was included in the heater circuit to obtain automatic temperature control. The thermomonitor used a thermistor sensor which was sensitive to 1/10 of 1°C temperature change between 110°C and 235°C. Since the thermomonitor operated on 110 volt AC, a relay was used in the heater circuit rather than a direct electrical connection through the thermomonitor. In addition, a switch was included so that the thermomonitor might be shorted out of the heater circuit, and detached and used elsewhere without rewiring the heater circuit. The reaction chamber was composed essentially of two sections, a cylinder with one end closed, and a top for sealing the other end of the cylinder. The two pieces together, which will be referred to as the reaction chamber, or bomb, were constructed of 316 stainless steel. The top was fitted with a teflon gasket to provide a pressure tight seal when the bomb was assembled. The center of the top housed the drive mechanism and packing box. The drive mechanism consisted of a V-belt-powered hub which was mounted on needle bearings. The drive shaft passed through the packing box and attached to a lower drive shaft on which two six-blade propellers were mounted. The propellers were moved up and down on the drive shaft, which rotated

PAGE 79

67 at 600 HPM. A pressure tight seal was maintained around the rotating drive shaft with two "Hulon" cones. The hollow cones were mounted back to back on the drive shaft through openings in the narrow ends of the cones. The space inside the cones was filled with silicone stopcock grease. The grease acted as a lubricant and aided in forming a tight seal around the drive shaft. The life of the "Rulon" cones was about 40 operating hours when the silicone grease was replaced at the start of each experiment. Pressure was placed on the packing cones before the bomb was pressurized. If this had not been done, the lubricant would have been forced out of the cones, resulting in the deterioration of the cones in 1 or 2 operating hours. The packing box was cooled during an experiment by circulating tap water around the outside of the box. The packing system of the reactor performed very satisfactorily. The gas exhaust for the bomb, a needle valve fitted into a l/8-inch diameter opening in the top of the bomb, was used to lower the pressure in the bomb during an experiment if by accident the reactor was over pressurized, and to vent the gases in the reactor at the end of an experiment. The same pressure transducer used with the 67-liter reactor was attached to a 1/8-inch opening in the reactor top. The temperature sensor was the same thermocouple arrangement used with the 67-llter reactor, with the exception that the functions of the two junctions

PAGE 80

68 were reversed. The sensing thermocouple was Inserted Into a thermowell through the top of the reactor and extended to the bottom of the reactor. Both pressure and temperature were continuously recorded as in the earlier research. Appendix 9 gives the calibration curves and other pertinent information concerning the calibration of the pressure and temperature sensing systems. A silver rupture disc was used to minimize the danger of an explosion during the course of an experiment. Galvanized pipe was connected to the exhaust side of the rupture-disc fitting to convey the reactants out the nearest window. The first experiment conducted in the reactor with pyroligneous acid demonstrated that silver was as sensitive to the corrosive properties of the waste as nickel. A teflon sheet was placed on the reaction-chamber side of the rupture disc to minimize the corrosion of the rupture disc and thus prevent premature breakage. The bursting pressure of the rupture disc was rated as 1,800 psi at room temperature. Oxygen was added to the reactor from a standard cylinder and reducing valve. The oxygen line from the cylinder attached to a needle valve, called the gas-inlet valve, which connected to the top of the reactor by a tee. The long axis of the tee joined the needle valve to the top of the reactor. A second valve, the liquid-sampling valve, was mounted on the 90° leg of the tee. A tube inside of the reaction chamber led from these valves almost to the bottom

PAGE 81

69 of the bomb. This arrangement made it possible to withdraw liquid samples from the reactor by closing the gas inlet valve and opening the liquid sampling valve. The tube was then purged by oxygen when the gas inlet valve was reopened after the liquid sampling valve was closed. Collection of Liquid Samples A colled tube with outlets through the top of the reactor was used to cool the reactants during the reaction or at the end of an experiment. Both air and water were used as coolants. The cooling coil reduced the volume of the reactor from 2 liters to 1.362 liters. The top was held on the cylindrical portion of the reactor by a splitring closure mechanism. Figure 16 shows the reactor completely assembled and placed in the heater. Figure 1? shows the disassembled reactor with the lower drive shaft disengaged so that the mixing propeller may be seen. Liquid samples were collected by a technique similar to that used in earlier research; that is, the sample filled a sample-collection tube and was frozen with liquid nitrogen. The sampling tube was made of 3/l6-inch tubing formed into a coil. Figure 13 shows the sampling tube connected to the reactor just after the liquid nitrogen bath has been removed. The lower end of the sampling tube was closed with a Swagelock cap fitting, while the other end was attached to the liquid sampling valve on the reactor. The sample volume was approximately 7 ml.

PAGE 82

70 Mixing Drive Motor Brine Bath Autotransformer Pressure Transducer Rupture Disc Vent Pipe Oxygen Inlet and Liquid Sampling Valve Heater Casing Figure 16. Assembled Parr Reactor

PAGE 83

Drive Mechanism and Packing Box 71 Oxygen Inlet Valve Rupture Disc Assembly \ Pressure Transducer Connection Cylindrical Portion of Reactor Cooling Coil Mixing Propeller Closure Rings (Foreground) Figure 17. Disassembled Parr Reactor

PAGE 84

72 Figure 18. Sampling Tube on Reactor

PAGE 85

73 Temperature Control During Experiments Precise and automatic control of the temperature of the reactor proved to be difficult to attain for a number of reasons. First, the thermowell in the reactor was not large enough to accomodate both the thermistor and thermocouple sensors. Since the thermocouple gave a direct measure of the temperature of the reactants, it was used in the more desirable thermowell position. The thermistor sensor was located beneath the reactor and held against the bottom and side of the reactor by fiberglass insulation. This technique gave a temperature variation of about l5°C after the system had stabilized. The thermistor was inserted into the vermiculite insulation Inside the heater casing through an opening in the lower annular aluminum plate. The sensor element was placed at various positions relative to the strip heaters. However, the temperature differential between the heater and the bomb was considerably larger than anticipated. The large temperature differential was caused by a 1/3-inch air gap between the wall of the bomb and the heater. This air gap provided very good insulation for the reactor. Consequently, the temperature inside the vermiculite insulation in the heater casing was considerably greater than the temperature of the reactor. The strip heaters were capable of heating the vermiculite insulation much more rapidly than they heated the reactor. Because of these factors the temperature fluctuation in the reactor was about +3°C when the

PAGE 86

7^ temperature was controlled only by the thermomonltor. The temperature could have been held closer to a constant value manually if It had not been necessary to perform other tasks during an experiment. Collection of Gas Sample During the studies that were conducted with the Parr reactor, It became apparent that some information on the composition of gases in the reactor was needed. The gas samples were collected by bleeding a sample from the gas exhaust valve on the reactor, through a brine bath to dry the gases, and into a balloon. The moisture trap in the brine bath was made of a capped section of galvanized pipe about 8 inches long. The gases entered the trap through a section of j-lnch copper tubing which was soldered into a hole drilled into the galvanized cap. This section of the copper tubing which acted as the gas inlet extended almost to the bottom of the trap. The gases then passed vertically through coarse brass wool filtering material and out of the trap through a second section of copper tubing. The gaseous sample was collected in a rubber balloon which was attached to one leg of a 4-inch stainless steel cross. The two legs of the cross, which formed 90° angles with the leg to which the balloon was attached, were used for evacuating the sampling train and for conveying the gases from the reactor into the balloon. Tubing and pinch clamps were arranged on these two legs so that the cross and balloon were detached from the sampling train without loss

PAGE 87

75 of the sample. The fourth leg of the cross was fitted with a rubber septum so that portions of the sample were withdrawn with a hypodermic syringe. Figure 19 shows all of the equipment used to conduct the autoxidation experiments. The gases were analyzed for oxygen and other components on an F. and M. Model 720 thermal conductivity gas chromatograph. Appendix 10 describes the conditions used for the analyses and shows typical peaks and calibration curves. Several different experimental techniques were used in the course of the research. The following chapter presents each technique and the results obtained.

PAGE 88

76 Figure 19. Experimental Equipment The equipment shown in the above figure is from left to right as follows: the Sargent Thermo monitor, gas sampling balloon (on top of the Thermomonltor ) , brine bath moisture trap and gas sampling line, the Thermomonltor shorting switch, autotransformer, heater with reactor in place, stirrer motor, liquid nitrogen storage dewar, liquid nitrogen bath dewar, recorder, ice water thermocouple reference bath (on top of recorder), and pressure reducing valve attached to the oxygen cylinder.

PAGE 89

CHAPTER VIII TECHNIQUES AND RESULTS OF THE STUDY OF AUTOXIDATION OF WASTES Introduction The purpose of the following research was to investigate the complete oxidation of wood distillation waste by autoxidation techniques. The variables which were expected to exert the greatest effect on the reaction were temperature and oxygen partial pressure. During the experimentation, it became obvious that minor changes in experimental technique had profound effects on the rate of the reaction. Consequently, the final experimental technique was the result of an evolutionary process. Source of Waste Only the undiluted waste was used, to simulate most realistically an industrial process. One problem already described was the collection of reproducible samples from the lagoons at the Cabot plant. To overcome this difficulty, Cabot Corporation collected all the liquors from one day's retort operation In a single tank. The settlable oils were separated and a portion of the acid water was stored in wooden barrels. The pyroligneous acid thus collected gave satisfactory reproducibility during the remainder of the study. The COD of the acid water was about 99,100 mg/1 77

PAGE 90

73 when originally collected and about 90,000 mg/1 after three months of storage. This loss of GOD was probably a result of the formation by the suspended tars of a coating on the inside of the barrels. First Experimental Technique The experiments were planned to start with the study of the reactions at a low temperature (125°C) and a low partial pressure of oxygen (25 psl at 0°C), and to proceed stepwise by increasing the oxygen partial pressure up to ^00 psi at 0°C. The temperature was increased in 25° C steps up to 225°C, and a range of partial pressures was investigated. This process was continued until the maximum temperature and pressure capabilities of the system were reached. During each experiment the temperature and oxygen partial pressure were held as constant as possible. The experimental techniques were similar to, but not identical with, those used with the 67-liter reactor. The technique used is given * 1. 1 liter of waste was poured into the reaction chamber. 2. The reactor was assembled and placed in the heater. 3. Mixing was begun. k. The gases in the reactor were evacuated and the reactor was sealed. 5. The reactor was heated to the selected temperature and the temperature was stabilized. 6. oxygen war gradually added to the reactor over a

PAGE 91

79 period of approximately 3 minutes until the desired partial pressure was reached. 7. The first sample was forced into the sampling tube and a stopwatch was started simultaneously. 8. The sample was frozen in the sampling tube with a bath of liquid nitrogen. 9. The sampling tube was disconnected from the reactor and the sample was gradually thawed. As the sample thawed, the expansion of the entrained gases forced the liquid and particles of ice out of the sampling tube. They were collected in a vial and analyzed for COD determination (See Appendix 2). 10. Steps (7), (8), and (9) were repeated for the collection of the succeeding samples. The stopwatch ran continuously for the duration of the experiment. 11. After the final sample was collected, the oxygen cylinder was closed and the reactor was cooled by passing water through the cooling ooil. 12. After the reactor was cooled to 50°C or less, the gases were vented to the atmosphere. The reactor was disassembled and washed. Results of the First Experimental Technique The data shown in Figure 20 compare the reaction rates at three different temperatures and approximately the same very low partial pressure of oxygen. The rates of all three reactions were quite slow and were approximately first order.

PAGE 92

80 I o H Q 8 100 80 6o S ^o 20 10 30 60 90 Time (mln) Figure 20. GOD vs. Time 120 150 Run Symbol 26 32 38 (O O) ( ) (A A) Avg. Initial Temp. 132 177 Avg. Initial 0°C Pq 2 C (psl) 17 19 21 Initial Avg. Rate Vol. of For First Waste Hour (ml) mg/1 mln 1000 1000 500 97 150 228

PAGE 93

81 The rates during the first 20 minutes of the reaction were substantially greater than the rates for the remainder of the experiment. This was probably caused by the polymerization of the oils in the acid water, which occurred as soon as the oxygen entered the reactor. The solid polymers stuck to the walls of the reactor and were not measured by the liquid sampling technique. The initial COD decreased as the initiation temperature of the reaction was increased. Heating the acid water partially broke the emulsion of tars present in the acid water, and these particles adhered to the walls of the reactor. The information shown in Figure 21 compares the reaction rates of two experiments at temperatures and partial pressures higher than those presented in Figure 20. A considerable Increase in the reaction rate was readily apparent. As a result of this faster rate, the GOD was reduced to a much lower value than in previous experiments. Run ^9 represented a 75 per cent reduction in the COD of the waste in 75 minutes. Both reactions followed first order kinetics. The reaction rates shown in Figures 22 and 23 compare the effects of five different initiation temperatures at low oxygen partial pressures. The kinetics of all the experiments were first order, and the rates of the two experiments initiated at temperatures above 200°C were considerably faster than those from experiments initiated below 200° C. The results given in Figures 24 and 25 corroborate the results presented in Figures 22 and 23.

PAGE 94

82 l o X to s Q o o 100 80 60 40 20 10

PAGE 95

83 I o K to S Q 8 100 80 60 40 20 10 20 40 60 Time (mln) Figure 22. GOD vs. Time 80 100 Run Symbol Avg. Initial Temp. °C 141 -) 151 -A) 181 27 33 39 (O O) (CD(AAvg. Initial 2 (psl) 28 <0 39 Initial Avg. Hate Vol. of For First Waste Hour (ml) mg/1 min 1000 500 500 178 33^ 292

PAGE 96

34 100 80 60 I o \ s Q o o 40 20 10 Run Symbol 45 50 (O O) ( ) 15

PAGE 97

85 I o P O 100 4o 20 10 20 40 60 Time (mln) Figure 24. GOD vs. Time 80 100 Hun Symbol Avg. Initial Temp. °C Avg. Initial Avg. Rate Initial vol. of For First « 0°C Waste Hour (psi) (ml) mg/1 mln '02 28 40 (O (O -O) -Q) (A A) 136 160 177 52 112 60 1000 500 500 no 365 392

PAGE 98

86 I o 9 Q o o 100 80 60 40 20 10 ^^ -o 15 30 45 Time (mln) Figure 25. COD vs. Time 60 75 Hun Symbol Avg. Initial Temp. °C Avg. Initial P o © 0°C Z (psl) Initial Avg. Rate Vol. of For First V/aste Hour (ml) mg/1 mln 46 51 (O O) (D D) 205 226 56 59 400 350 445 678

PAGE 99

87 The data shown in Figures 26 and 27 compare the rates of reactions Initiated at higher partial pressures of oxygen. Runs kl, 47, and 52 suggest that two first order reaction rates were followed by the reaction at temperatures above 175° C and partial pressures of about 100 psl. Run 30, as shown on Figure 27, illustrates the importance of temperature in achieving rapid reaction rates. Support for the statements concerning the data shown in Figures 26 and 27 is given by the data shown in Figure 28. Higher temperatures affected the reactions in two different ways. First of all, the higher temperatures caused more rapid reaction rates, because the percentage of molecules which reacted with oxygen was greater. This implies that the rate of oxygen dissolution becomes rate-limiting as the temperature of the reaction is increased. Secondly, higher temperatures made a greater portion of the molecules susceptible to oxidation; hence, a greater percentage of the COD can be satisfied by the autoxidation. Data comparing the rates of three experiments carried out at intermediate temperatures and very high oxygen partial pressures are given in Figure 29. Only Run 43 exhibits a very rapid rate. Autoxidation of the waste caused several characteristic changes to occur in the waste, which could be observed visually during an experiment. These changes were similar to those observed during experiments with the 67-liter reactor. First of all, oils which were emulsified In the acid water were polymerized into solid particles as soon as

PAGE 100

88 I o X P o o 100 80 60 S ^ 20 10 20 60 40 Time (mln) Figure 26. COD vs. Time 80 100 Hun Symbol Avg. Initial Tgmp. Avg. Initial P n 3 0°C ° 2 (psl) Initial Avg. Rate Vol. of For First Waste Hour (ml) mg/1 mln 29 35 41 (O O) (A-D) -A) 126 151 186 120 120 97 1000 500 500 280 275 490

PAGE 101

89 0^ I o « a Q o o 100 ~—& 15 Run Symbo 1 47 52 30 (O O) (D ) (A A)

PAGE 102

90 100 l o W to & Q o o 15 30 45 Time (mln) Figure 28. GOD vs. Time 60 75 Run 36 42 48 Symbol Avg. Initial Temp. °C (O O) ( — D) (A &) 153 205 215 Avg. Initial P ? @ 0°C Z (psi) 181 122 152 Initial Avg. Rate Vol. of For First Waste Hour (ml) mg/1 mln 500 500 400 325 520 634

PAGE 103

91 I o Q 8 100 80 60 4o 20 10 20 40 60 80 100 Time (mln) Figure 29. COD vs. Time Run Symbol 31 37 4 3 Avg. Initial Temp. °C 146 175 179 Avg. Initial P , ® o°c Z (psi) 273 243 389 Initial Avg. Rate Vol. of For First Waste Hour (ml) mg/1 min 1000 500 500 200 394 534

PAGE 104

92 oxygen was added to the reactor. This reaction took place within a few seconds. The presence of these particles caused the waste to appear very dark and turbid In appearance. The color changes which characterized the reaction under various conditions are shown in Figure JO, Another characteristic of the reaction was the formation of stable foams in the waste. These foams were usually formed during the first few minutes of an autoxidation. Both the foam and the particles gradually disappeared in the experiments that achieved fast reaction rates. The presence of particles and foam in the samples resulted in poor reproducibility between duplicate GOD determinations. Consequently, all the values plotted from experiments on acid water were averages of duplicate CODs from each sample. The reproducibility of GOD determinations Improved in proportion to the per cent COD reduction. Appendix 11 shows a statistical comparison of replicate COD determinations on one experimental run. Another characteristic of the autoxidation of pyrollgneous acid was the rapid liberation of heat by the reactants during the first 10 to 15 minutes of each run. This exothermlcity caused an average temperature Increase of about 25°G. The temperature in each run was allowed to increase exothermlcally , and upon termination of the exothermic portion of the reaction, the temperature was allowed to decrease to the temperature at which the reaction was Initiated. The total volume of waste placed in the

PAGE 105

93 Figure 30. Samples from Autoxidation Reactions The cards to the left of each row of samples give the Run number, initial temperature, initial total pressure of each experiment. The numbers beneath each column of samples represent the order in which the samples were collected. In each case, sample number 1 was collected immediately after oxygen was added. The intervals between the remaining samples were 15 or 20 minutes, depending on the experiment.

PAGE 106

9^ reactor was decreased as the initiation temperature of different experiments was increased. A decrease in total volume resulted In fewer grams of fuel. Thus, although the reaction rate was faster at higher temperatures, which liberated heat more rapidly, the measured temperature rise was approximately the same in all experiments. The temperature Increase and decrease generally took place during the first 30 minutes of the reaction. A number of inconsistencies are apparent in the data. While increases in temperature and oxygen partial pressure accelerated the reaction rate, certain experiments gave an inconsistently high reaction rate at a relatively low oxygen partial pressure. ( Cfr. Run ^9. ) In addition, the temperature was not held constant, primarily because of the exothermiclty of the reaction. The final product of the oxidation was carbon dioxide, Indicating that, at least in the experiments giving rapid rates of reaction, the oxygen partial pressure must have steadily decreased during the reaction. The effects of decreasing oxygen partial pressure are unknown since the oxygen partial pressure was not directly measured, and thus the computed partial pressures for some of the preceding experiments are not given. The data demonstrate that oxygen partial pressures of at least 50 psi at 0°C are needed to sustain autoxldation reactions at any temperature. Second Experimental Technique The technique for carrying out the experiments was

PAGE 107

95 altered to obtain more valid and consistent data. The major change in technique was to eliminate, to the extent possible, the exothermic temperature increase during the first 30 minutes of a run. This was accomplished by using a small amount of waste (400 ml) in the reactor and bypassing air or water through the cooling coil during the exothermic portion of the reaction. Only temperatures and pressures which had given rapid reaction rates in earlier studies were investigated. Gas samples were taken simultaneously with the liquid samples to follow the changes in oxygen partial pressure, as well as the build-up of carbon dioxide partial pressures in the reactor. Mixing was begun in the reactor Just before oxygen was added to minimize the loss of COD, which x-ras caused by tars that adhered to the walls of the reactor during the time the reactor was heated, and before oxygen was introduced into the reactor. Results of Second Experimental Technique The data in Figure 31 present the results of the modified technique under mild oxidizing conditions. The kinetics of the reaction were either first order or possibly slightly greater than first order. The rate of the reaction was not rapid and the build-up of carbon dioxide was negligible. Consequently, the oxygen partial pressures given with the earlier experiments, which gave rates as low or lower than that shown in Figure 31 were correct. The data shown in Figure 32 substantiate that

PAGE 108

I o X faO S P o o 100 80 60
PAGE 109

97 i o K to Q o o 100 80 60 40 20 10 15 30 45 Time (min) Figure 32. COD vs. Time 60 Hun 55 Interval of Time (min)

PAGE 110

98 shown in Figure 31. Run 57, shown in Figure 33, yielded a more rapid rate than the runs Indicated on Figures 31 and 32. The carbon dioxide build-up became significant toward the end of the reaction. The reaction order was greater than first. The reaction rate and the carbon dioxide build-up given In Figure 34 are comparable to those shown in Figure 33. However, the reaction shown in Figure 34 follows first order kinetics, while that shown in Figure 33 followed kinetics of a higher order. The data suggest that differences in the control of the temperature between the two reactions caused a change in the order of the reaction. For example, the average temperature from to 15 minutes in Run 58 was lower than the average temperature during the 15 to 30 minute interval. The reverse is true for Run 57. The kinetics of the reaction shown in Figure 35 were greater than first order. The temperature of this experiment was highest at the beginning of the run. Runs 54 through 59 demonstrated the importance of experimental technique and, in particular, the important effects of temperature on both the rate and order of an autoxldation reaction. Large carbon dioxide partial pressures were observed, as expected. This increase in carbon dioxide partial pressure resulted in a decrease in the oxygen partial pressure. Since the oxygen partial pressures were relatively high In all experiments, the decrease in the effective partial pressures during the

PAGE 111

99 l o X Q o o 100 80 60 40 20 10

PAGE 112

100 I o x to p o 100 80 60 40 20 10 cx^

PAGE 113

101 I o H to m a o o 100 80 60 ko 20 10 J

PAGE 114

102 course of any run did not appear to have had a great effect on the rate of the reaction. Third gxperlmental Technique A different technique was adopted for better simulation of a treatment process: 1. The reaction was initiated at a certain temperature. 2. The exo therm! city of the oxidation increased the temperature of the reaction. 3. The reaction was maintained at the highest temperature reached during the exothermic portion of the reaction. The more abrupt Initiation of the reaction also was expected to increase the rate of the reaction. This was accomplished by beginning to mix the reactants only after the reactor was brought to the desired oxygen partial pressure. These changes In technique simulated the entrance of the preheated waste into a reactor, where an abrupt mixture of waste and oxygen took place.. This resulted in a temperature rise which would be conserved in a well-insulated reactor. Results of Third Experimental Technique The data shown In Figures 36 and 37 describe two reactions which were initiated at 125°C. The reaction rates shown in these figures are comparatively slow and carbon

PAGE 115

103 I O x to a 8 100 80 60 40 20 10 C)

PAGE 116

104 l o & Q o u 100 80 60 40 20 10

PAGE 117

105 dioxide build-up is negligible. The reaction exhibited sufficient exothermicity to raise the temperature of the reactants about 10°C. Figures 38»39i and 40 present data which describe three experiments initiated at 150°C and different oxygen partial pressures. A considerable increase in rate was apparent as compared to rates in Figures 36 and 37 • In addition, the rates increased proportionately with increases in the partial pressure of oxygen. A portion of the rate increases was caused by Increases In the average reaction temperature, since the exothermicity of the reactions increased with increases in initial partial pressures. The reactions were of a multi-step nature and two hypotheses are suggested by the data. One is that the reaction followed at least two separate mechanisms. The mechanism measured during the first minutes of the reaction was relatively rapid, and the mechanism followed during the latter few minutes was relatively slow. The second hypothesis is that during the first few minutes of the reaction, two different mechanisms cumulatively affected the rate, only the slower of which was sustained throughout the entire experiment. The more rapid mechanism was unable to sustain Itself throughout the entire experiment. The author favors the first hypothesis for the following reasons: 1. At low temperatures (Figures 36 and 37), only a low rate of reaction was measured. This suggests that free radicals were not created as rapidly as

PAGE 118

106 I O X to e Q o o 100 80 60 40 20 10 15 30 45 60 Time (mln) Figure 38. COD vs. Time Run 63 Initiated at 150°C Interval of Time (mln)

PAGE 119

107 I o Q o o 100 80 60 4o 20 10 15 30 45 60 Time (min) Figure 39. GOD vs. Time Hun 64 Initiated at 150°C nterval f Time (min)

PAGE 120

108 I o iC B Q o o 100 80 60 40 20 10 ex

PAGE 121

109 the chain type reactions were terminated. 2. At higher temperatures (Figures 38 through 44), and at high concentrations of waste, free radicals were formed more rapidly than the chain reactions were terminated. This resulted in a greatly increased rate as compared with lower temperature conditions. 3. The rapid rate was terminated because of a lack of material which readily formed free radicals under the reaction conditions. Mot only did the concentration of waste decrease, but the qualitative composition of the waste changed also. As the autoxidatlon occurred, the molecules involved became smaller, hence less easily activated into free radical form. 4. The slower rate mechanism seemed to be almost exclusively temperature dependent. The rate of the slower portions of the higher temperature runs (Figures 4l through 44) was greater than the rate of the low temperature runs (Figures j6 and 37 ) » although the low temperature runs were at much higher concentration of waste. Either hypothesis could be described only by a complex, multi-term kinetic expression. Data from runs initiated at 175° C are shown in Figures 4l through 44. The exothermicity of the reactions was sufficient to raise the temperature as much as 50°G,

PAGE 122

110 « o X to B P 8 100 80 60 40 20 10 15 30 Time (mln) Figure 41, GOD vs. Time ^5 60 Run 66 Initiated at 1?5°G Interval of Time (mln)

PAGE 123

Ill I o bO Q O o 100 80 60 40 20 10 15 30 ^5 60 Time (mln) Figure 42. COD vs. Time Hun 67 Initiated at 175°C Interval of Time (mln)

PAGE 124

112 I o fcO B P o o 100 80 60 40 20 10 15 30 Time (mln) Figure 43. GOD vs. Time 45 60 Hun 68 Initiated at 1?5°C Interval of Time (mln)

PAGE 125

113 I o H bO S Q O o 100 80 60 40 20 10 15 45 30 Time (mln) Figure 44. GOD vs. Time Run 71 initiated at 175°C 60 Interval of Time (mln)

PAGE 126

114 and seemed to be directly proportional to the partial pressure at which the reaction was initiated. Also, increases in the oxygen partial pressure increased the duration of the initial rapid rate. Again, this increase might have been caused by differences among the temperature increases, which were caused by the Increase in partial pressure of oxygen. The curves in Figures 4l through 44 are similar to the curves in Figures 38 through 40, except that much higher reaction rates were attained. Run 71, shown on Figure 44, represents 80 per cent treatment of the waste In about 18 minutes. By approximately tripling the treatment time to 1 hour, only 8? per cent COD reduction was attained. Autoxldatlon of Other Substances Autoxidatlons were performed on two other wastes, raw primary sludge and citrus-processing-plant wash water, and two pure compounds, phenol and pectin, in order to compare these reaction rates with those obtained from pyrollgneous acid. Phenol was selected because it is known to occur in pyrollgneous acid, while pectin was chosen because it Is a troublesome compound found in the waste of citrus-frult-processing plants in Florida. Autoxldatlon of primary sludfi Raw primary sludge from the University of Florida sewage treatment plant was Investigated first. The sludge was prepared for the experiment by homogenlzatlon In a

PAGE 127

115 blender. After homogenization, the sludge was relativelyviscous and approximated the consistency of pancake batter. Four hundred ml of sludge were used in the experiment. The GOD of the raw sludge prior to autoxidation was 15,200 mg/1. The sludge was very difficult to pipet and consequently COD values were not consistent. The autoxidation of the sludge was carried out under conditions which duplicated those giving the highest reaction rates with pyroligneous acid. Figure 45 describes the rate obtained by the autoxidation of the sludge. The rate was quite slow as is shown by the slope of the curve as well as by the absence of carbon dioxide. The data appear quite similar to those obtained with pyroligneous acid when the reactions were initiated at 125°C. Little exothermicity was observed during the experiments. The data suggest that pyroligneous acid was much more susceptible to treatment by autoxidation than was sludge. The sludge underwent little visual change, except that during the final minutes of the experiment a distinct yellow color became apparent in the samples. Since the rate of the reaction was very slow, no further experiments were carried out with sludge. Autoxidation of phenol A stock solution of phenol in distilled water was prepared with a GOD of 49,100 mg/l and a pH of 4.1. This corresponds to a solution of about 19.48 g/1 of phenol.

PAGE 128

116 100 o 2 8 30 45 Time (mln) Figure 45. COD vs. Time (Primary Sludge) Run 72 Initiated at 175°C 60 Interval of Time (mln)

PAGE 129

117 Figure 46 describes an autoxidation of phenol initiated at 150°C. The data fit a first order curve. Very little exothermicity was observed. Figure 47 gives the data from an experiment initiated at 175° C. It is interesting to note that the rate of the reaction Increases with the duration of the reaction. The overall rate was still quite slow, as evidenced by the absence of carbon dioxide. Figure 48 describes the results of an autoxidation reaction which was initiated at 200° G. Figure 48 represents what might be called a typical autoxidation reaction (see Chapter III, pages 21 and 22). The rate of the reaction was inhibited for about the first 15 minutes, then rapidly increased to a maximum. The maximum rate was followed by a rapid decrease to a minimum value. The phenol solution underwent several distinct changes in appearance. The stock solution was colorless, but during the reactions shown in Figures 46 and 47, it changed gradually from colorless to light yellow to dark yellow to brown. In Figure 48 the first sample collected was yellow, the second dark brown, and the remaining samples became successively lighter in color as the reaction proceeded. Autoxidation of pectin Pectin was another compound selected for autoxidation. Pectin is completely insoluble in water; however, by lowering the pH of water to 4, it was possible to cause pectin to hydrollze and go into solution. The autoxidations

PAGE 130

113 I o to s Q 8 100 80 60 40 20 10 15 30 Time (mln) 45 60 Figure 46. COD vs. Time (Phenol) Run 73 Initiated at 150°C Interval of Time (mln)

PAGE 131

119 I o X a Q 8 100 80 60 40 20 10

PAGE 132

120 I o w E P o o 100 80 60 40 20 10 8 6 4

PAGE 133

121 were performed on a stock solution containing about 20 g/1 of pectin. The COD of the stock solution was 8,040 mg/1. A more concentrated solution would have been desirable, however, this solution had a consistency similar to that of syrup. A large increase in the amount of pectin which hydrollzed into the solution would have produced a more exothermic reaction, similar to that of pyroligneous acid. However, such a mixture would probably have resulted in a gelatinous substance. Figures 49, 50, and 51 describe the results of experiments using the stock pectin solution. The rate of the reaction shown in Figure 49 was comparatively slow. The order of the reaction was slightly greater than first. Little carbon dioxide was produced. The experiment given in Figure 50 was initiated at 175 c. The initial rate of the reaction was considerably faster than that of the preceding experiment, and the reaction order was substantially greater than first. The experiment shown in Figure 51 was initiated at 200 C and gave essentially the same results as the experiment shown in Figure 50. None of the three pectin experiments was measurably exothermic. This was probably caused by the low GOD of the stock solution. Autoxldatlon of wash water from a citrus-processing plant The wash water from a citrus-processing plant was selected as a waste for autoxldatlon study. The COD of the raw citrus waste was about 2,100 mg/1 and it had a pH of 4.1. Autoxldatlon of 400 ml of this waste was initiated

PAGE 134

122 CM I o S P 8 100 80 60 40 20 10 \ 15 30 Time (mln) ^5 60 Figure 49. COD vs. Time (Pectin) Run 76 Initiated at 150°C Interval of Time (mln)

PAGE 135

123 CM I o M to S Q o o 100 80 60 40 20 10 15 30 Time (min. ) ^5 60 Figure 50. COD vs. Time (Pectin) Hun 78 initiated at 175°C Interval of Time (min. )

PAGE 136

124

PAGE 137

125 at 175° G. Figure 52 gives the results of this experiment. The reaction rate was quite slow and no carbon dioxide was formed. No exo the rmi city was observed during the experiment. As in the experiments with pectin, this was due to the low COD of the waste. An autoxidation of 1 hour gave a 25 per cent reduction in the COD of the waste.

PAGE 138

CM I O bO Q o o 100 80 60 40 20 Cl=. 10 126 15 30 Time (mln. ) 45 60 Figure 52. COD vs. Time (citrus Waste) dun 77 Initiated at 1?5°C Interval of Time (mln. )

PAGE 139

CHAPTER IX CONCLUSIONS FROM THE STUDY OF AUTOXIDATION OF WASTES Autoxidatlon of wood distillation waste with oxygen has been shown to be a good treatment method from a technological point of view. Data from the studies indicate that an 80 per cent reduction on COD is obtained in 20 minutes and the temperature of the waste is increased to 225°C if autoxidations are initiated at 175°C and at an oxygen partial pressure of 150 psi at 0°C. The cost of reducing the COD of the pyroligneous acid would be about 1.4 ^/lb. of COD removed in an industrial plant, as shown by the preliminary design and economic-feasibility study in Appendix 12. A cost of this order compares favorably with costs given by Babbitt and Jauman (82) for municipal sewage treatment. The data have shown that rapid reaction rates are basically dependent on three conditions: temperature, concentration, and an adequate supply of oxygen. If any one of the conditions had not been satisfied, slow rates would have resulted. However, there is also an order of Importance among the conditions. The most important condition was adequate temperature to initiate and sustain the chain reactions. No matter how concentrated the waste or how high the oxygen partial 127

PAGE 140

123 pressure, slow reaction rates would have been observed if the autoxidation had been Initiated at too low a temperature. The initiation temperature which must be achieved was determined by the qualitative composition of the waste , or the thermal instability of the compounds In the waste to autoxidation. Higher temperatures cause a large percentage of the waste compounds to be thermally unstable so that the temperature also determined the extent of the oxidation. The second factor, concentration or GOD of the waste, affected the rate of the reaction as well as the measurable quantity of heat which was produced by the reaction. Thus, a waste of low GOD will yield heat to the system at a slow rate, and no exothermic temperature rise will be available during a finite time for the system, even though the Initiation temperature would have been satisfactory for the same waste In a greater concentration. The autoxidation rates of both pure compounds and other wastes were much slower than the autoxidation rate of pyroligneous acid, at least partially because the GOD of the pyroligneous acid was much greater than the COD of the other substances. (The difference in qualitative analyses accounts for difference in reaction rate, also.) The third factor, oxygen partial pressure, affected autoxldatlons in several ways. Oxygen must be present for an autoxidation to occur, and since little oxygen dissolves in water, an inter-phase transfer must occur. If the rate

PAGE 141

129 of oxygen transfer from the gas phase to the liquid phase is slower than the rate of the autoxidation reaction, the rate of autoxidation is limited by the rate of the inter-phase transfer. Increasing the oxygen partial pressure and increasing the rate of mixing of the two phases accelerates the rate of gas transfer. The data have shown that even with complete mixing, oxygen is not in excess concentration in the waste, even at the higher partial pressures that were Investigated. The data further suggest that high oxygen partial pressure is most desirable as the reaction is initiated. A high reaction rate was maintained when the reaction was properly initiated, in spite of a 50 per cent decrease in oxygen partial pressure during the reaction. Finally, the data indicate that when the oxygen partial pressure was lower than the equivalent of 50 psl at 0°C at any time, the rate of the autoxidation was greatly decreased. The chain-type autoxidation reaction has the potential of becoming a dominant waste treatment process. This potential exists because the chain reactions have the ability to pass a unit of activation energy from molecule to molecule, rather than returning this energy to the environment as heat. As with any process that is potentially very useful, the problems in learning to utilize and control autoxidations may be difficult to solve. However, the solution of these problems will Improve and reduce the cost of waste treatment.

PAGE 142

APPENDICES

PAGE 143

APPENDIX 1 CALIBRATION OF PRESSURE AND TEMPERATURE SENSOR SYSTEMS The pressure transducer-recorder system was calibrated with an Amthor Instrument Go. dead weight tester. The pressure transducer, Model 46139 C-G100-50, manufactured by G. M. Giannini and Co., has a range of to 1,000 psi. The recorder was manufactured by the Brown Instrument Division of Minneapolis-Honeywell Regulator Go. Model Y153X64pi6-X-4l( V) was used. A 5 mv signal gave full-scale response on the recorder. The thermocouple system used chrome 1-alumel Junctions as voltage sources. A water bath on a magnetic stirrer and heater was used to make the calibration curve. An air bath at 25°G was used as the reference. 131

PAGE 144

132 ft 0) xs u o o as -XT OOO'S > o CM > S > CO to c a c 0) CO I £ 3 (0 CO « o B 2 to 3 O H O cn 0) U 3 b0 H

PAGE 145

133 300 250 200 CO (X 150 100 50

PAGE 146

13^ » CO 0) o o c o a; o c 0) +> 0) Hj 8 & I D & 8 § o § a u bO -p 3 O H o c ^ o O -H CO -P c c CO 3 s 3 bO H fa S o o

PAGE 147

135 140 120 100 o o . 80 60 40 20

PAGE 148

APPENDIX 2 GOD TECHNIQUE The main difference between the CODs run for this research and the specific method given in Standard iMethods for the Examination of Water and Waste Water (81) is in the type of glassware which was used. Three-hundred-ml erlenmeyer flasks with ground-glass necks were used rather than 250-ml round-bottom distilling flasks. Substitution of the erlenmeyer flasks greatly reduced the time involved In running determinations, since each test was made up and refluxed In the same piece of glassware. Graham condensers were used rather than Frledrichs condensers because the author felt that Graham condensers give a more positive condensation of hot vapors than do Friedrlchs condensers. The hot flasks were sometimes removed from the condensers as soon as active boiling had ceased. Whenever hot flasks were handled, the condensers were washed with a little distilled water before the flasks were removed, and the flasks were transferred directly to a trough which was cooled by tap water. The hot flasks reached room temperature after 3 or ^ minutes In the cooling trough. This saved time in each determination, since the flasks were at room temperature when titrated. 136

PAGE 149

APPENDIX 3 SOLUBILITY OF OXYGEN IN WATER The data shown in the following graphs are averaged from data given by Zoss et al. (83) and Pray et al. (84) in the literature. The data of these authors did not agree perfectly, but both techniques appeared to be satisfactory. 137

PAGE 150

138 CVJ I o to e CO o n

PAGE 151

139 8 CM I O x to CM O 315. o°c 8 260. 0°C 204. 5°C 37.8°C 149. 0°C 93.2+°C 10 -2 P (psl x 10"^) Figure 58. Solubility of Oxygen in Water

PAGE 152

APPENDIX 4 PARTIAL PRESSURE OF STEAM AS TEMPERATURE VARIES The curve shown In Figure 59 was plotted from data given in Steam: Its Generation and Use (85), published by Babcock and Wilcox. The points shown on the curve indicate the variations of the partial pressures of distilled water and pyrollgneous acid as measured in the experimental system used in this research. 140

PAGE 153

141 500 400 CO ft

PAGE 154

APPENDIX 5 MAXIMUM THEORETICAL PARTIAL PRESSURE OF CARBON DIOXIDE The following calculation is based on the ideal gas law and on the assumption that all COD reduction results in carbon dioxide formation. This assumption tends to bias the result on the high side. However, it will be adequate for the purposes of this calculation. COD Reduction: 2550 860 = i960 mg/1 Total volume = 67.3 1 Liquid volume = 20 1 Grams of carbon dioxide = (20)U?60) = ??i 8 °Q m § = 16.9 g Molecular weight of carbon dioxide = 44 Volume of carbon dioxide at STP = -^^2. (22.4) = 8.6 1 Pi Vl = nRTi P2 V2 = nRT2 P? = (P)) UllHSZl = (1^.7)(8.6)(402) = 3,9 psl 2 fv2)iTi) w.Ynm) — J y p 142

PAGE 155

APPENDIX 6 CALCULATION FOR COMPUTATION OF OXYGEN PARTIAL PRESSURE IN AN EXPERIMENT Sample calculation of oxygen partial pressure in 67-liter reactor j Data 1 Temperature (T) = 125° C Total Pressure (P) = 224 psi Total pressure = 224 psi partial pressure of steam at 125 C = 35 psi Total pressure of dry gas at 125 C = I89 psi From Ideal Gas Law» lA = Zl P 2 T 2 Use 0°C as reference temperature .% p 2 = 3| p 0°C = !£3U89) = 144 psi Total pressure of dry gas at C = 144 psi partial pressure of 1 atm. of nitrogen at 0°C 14 psi Partial pressure of oxygen at 0°C = 130 psi 143

PAGE 156

APPENDIX 7 RESULTS OF TRIPLICATE COD DETERMINATION ON RUN 7 Sample Number

PAGE 157

APPENDIX 8 TRACE ELEMENT FEED SCHEDULE FOR TRICKLING FILTER

PAGE 158

APPENDIX 9 CALIBRATION OF PRESSURE AND TEMPERATURE SENSOR SYSTEMS FOR PARR REACTOR The temperature calibration curve was obtained bycomparing the response of the thermocouple-recorder system to the response of a mercury-ln-glass thermometer. The thermocouple and mercury reservoir in the thermometer were placed as close to each other as possible and wrapped with glass-wool insulation. The Insulated packet was Inserted in an 8-inch length of 3/4-lnch diameter pipe which was capped on one end. The pipe was wrapped with a standard laboratory heating tape. The heat out-put of the tape was controlled with an autotransformer. A new set of range resistors was Installed in the recorder to achieve a linear response of to 15 mv. (See Figure 60. ) The pressure transducer-recorder system was calibrated with an American Instrument Company Model 47-2161 Dead Weight Tester. The 400 K potentiometer in the pressure sensing circuit was replaced with a 600 K potentiometer to produce a full-scale potential of 15 mv across the pressure transducer. (See Figure 61. ) 146

PAGE 159

147 12 10 > a 8 50 100 150 200 250 300 Figure 60. Calibration of Thermocouple-Recorder Against Mercury Thermometer

PAGE 160

148 700 600 500 400 CO ft 300 200 100 13 5 7 9 mv Figure 6l. Calibration of Pressure Transducer 11

PAGE 161

APPENDIX 10 INFORMATION ON GAS CHROMATOGRAPH ANALYSES The conditions for operating the gas chromatograph to analyse gas samples for oxygen, nitrogen, and carbon dioxide are as follows: Detector temperature Column temperature Injection port temperature Detector cell current Attenuation Carrier Gas flow rate Column Packings (loosely packed) 130° C ambient ambient 150 ma 64 200 ml/min Column A 30 per cent HMPA on 60-80 mesh Co lumnpa ck ( HMPA=Hexamethylphosphophoramlde ) Column B 60-80 mesh molecular sieve Figure 62 shows typical peaks obtained in an analysis. Figure 63 shows typical calibration curves. Figure 64 pictures the Instrument. 149

PAGE 162

150

PAGE 163

151 09 9) s: o c -p to a © >ul C0£ and O2 1000 Ml N 2 2000 1000 3000 Figure 63. Calibration Curves for Oxygen, Nitrogen, and Carbon Dioxide A» C0 2 O2 = N 2

PAGE 164

Figure 64. F. & M. Model 720 Gas Chromatograph

PAGE 165

APPENDIX 11 STATISTICAL COMPARISON OF COD DETERMINATIONS Data from Run 25 have been analyzed using a completely randomized statistical design. The data do not meet all of the assumptions on which a completely randomized design is based; for example, the samples were not collected in a random manner due to the nature of the experiment. The analysis gave a good approximation of the variation among COD determinations, however. Run 25 was performed under mild conditions and gave a low reaction rate. Consequently particulates and foam were present in all samples. As higher rates were obtained, the duplicate COD determinations gave more nearly identical values. Thus, this analysis was biassed toward a high value. 153

PAGE 166

15*

PAGE 167

155 S Z = 10.06 x 10 6 S = 3.18 x 10 3 x = 65,885 P{^> t #05 s + x}= 0.05 5 t .05 s 10 per cent 1

PAGE 168

APPENDIX 12 PRELIMINARY DESIGN AND ECONOMICFEASIBILITY STUDY OF AUTOXIDATION AS A WASTE TREATMENT PROCESS FOR PYROLIGNEOUS ACID This design and feasibility study was Included to estimate the cost of treating pyrollgneous acid by autoxidatlon. The design was based in part on assumptions, since some of the design criteria were not known. The economic-feasibility study is based on estimated costs of specialized equipment. All assumptions were biassed so as to add expense to the process costs. However, the author feels that the final result is a reasonable estimate of the cost of such a process. Design and cost data were obtained from references 86, 87, 88, 89, 90, 91, and from the data presented in this dissertation. Data were extrapolated In Figure 65 to predict the effect of Increasing the initiation temperature of the reactions. These data Indicated that a 10°C Increase in the Initiation temperature of the reaction would result In 100 per cent COD reduction in the waste. The COD reduction actually realized would be, of course, slightly less than 100 per cent. The design of the system la such that it can be operated at more extreme conditions than those used in the laboratory. 156

PAGE 169

157 The cost estimates were based on the use of stainless steel throughout the system. Substantial savings could be effected if some other method of corrosion prevention, such as cathodic protection, could be employed. Design Criteria Daily Flow 15,000 gal Avg. GOD of Waste 100,000 mg/1 GOD to be Removed 80,000 mg/1 Reaction Initiation Temperature 175°G Reactor Operating Temperature 225° G System Operating Pressure 700 psi System Design Pressure 1,000 psi Duration of Reaction 20 min Design of Reactor (Cylindrical) Flow = 625 gph = 83.5 cfh «=* 84 cfh Retention Time = 20 mln /. vol = 28 cf Try 2 Diameter Base Area = (7T)(2) 2 A = 3.14 sq ft Height = 28/3.14 = 8.92 ft ^9 ft Reactor must have steam Jacket or internal coil for starting process. Computation of Gas Flow COD (mg/1) Removed = OgCmg/U Required = 80,000 mg/1 Conversion to Operating Conditions PF2 t? 2 -r??r

PAGE 170

158 V? = (1 ata.HQ.7 l)(4qB°A) _ , 2 (47.5 atii)(273*A) °028 X/ e Flow for One Liter 80 g 20 mln = 4 g/mln = 0.112 1/mln Liquid Flow = 10.4 gal/min = 39.3 1/mln Total Gas Flow 39.3 1/mln x 0.112 1/mln = 4.4 1/mln Computation of Annual Oxygen Requirements 80 g/1 x 15,000 gal/d x 3.78 1/gal x 0.0022 lb./g = 9,990 lb./d fe 5 ton/d For Year = 365 x 5 = 1,325 ton/yr. Selection of Pump and Motor Pump Power Requirements P = U gal/min) (P nsl ) = ( 10. 4) ( 1 ,000 ) = 6 o6 h 1715 I7TT P Use a Centrifugal, Multi-stage, Volute Pump Pump Efficiency Assume 50 per cent Motor 6.06/0.5 = 12.12 hp Use 12 hp 1,800 RPM Motor Selection of Reactor Mixing Motor For a propeller with 12-inch Diameter Blade with 12-inch Pitch Rotating at 600 RPM = 1 hp 5 Propellers 5 hp Use 600 RPM, 5 hp Motor Sizing of Heat Exchanger Assume: Avg. Temperature Differential = 9°F, U = 200 BTU/(hr)(sq ft)(°F) = 3.33 BTU/(hr)(sq ft)(°F) q = (10.5 gal/mln)(8.34 lb./gal ) ( 150°C) (9/5 °?/°<2) = 23,600 BTU/min

PAGE 171

159 A = ttSU = 23400 = ? 88 sq ft t^t 13.33)19) Use 1-lnoh Tubes, Circumference = 0.262 ft No. of Tubes = 788/2.62 = 300 Allow 1 inch between Tubes or 2-inch overall Diameter Area/Tube = TT(0.l67) 2 /4 = 0.022 sq ft End Area of Exchanger = (0.022)(300) = 6.60 sq ft Diameter of Shell = 2.91 ft Total Shell Surface = Tf (2.91) ( 10) = 91. 4 (2)(6.60) = 13.2 1047S sq ft Sizing Waste Holding Tank Assume 4 hrs Storage Capacity Needed Capacity = 4(625 gal/hr) = 2,500 gal = 334 cf Try 8-ft Diameter A = tt(4) 2 = 50.4 sq ft H = 33V50.4 = 6.65 —7 ft Sizing of Pipe Flow = 10.4 gal/min = 0.0232 cfs Try V = 3 ft/sec Q = AV A = 0.0232/3 = 0.00774 sq ft Diameter ss 0.09 ft — 1.08 inches ^ 1 inch Estimate Length of Pipe ^r 77 ft Area of Pipe Surface ~ 2.6 ft/10 ft Length Total Area of Pipe Surface = (7.7)(2.6) = 20 sq ft Size of Gas-Liquid Separator Assume: Diameter = 1 ft, Height = 3 ft

PAGE 172

160 Area to Be Insulated Reactor (2)(tv)(9) = 57.6 sq ft Ends = 6.3 sq ft 63.9 sq ft Heat Exchanger 104.6 sq ft GasLiquid Separator 14.0 sq ft Turbine 10.0 sq ft pl P es 20.0 sq ft Total 212.6 sq ft Selection of Insulation Try Mineral Wool Density 6 lb./sq ft K = 0.26 BTU/hr/sq ft/°F/l-inch thickness Try 6-inch thickness .% K = 0.0433 Temperature Gradient ~ 400°F Heat Loss (400)(0.0433)(212.6) = 3,680 BTU/hr = 6l BTU/mln Selection of Turbine-Generator Total BTU Produced by Reaction (83.5 ou ft/hr)(62.4 lb./cu ft)(90°P) = 468,000 BTU/hr Loss Through Heat Exchange (5,200 lb./hr)(9°F) = 46,800 BTU/hr Loss Through Insulation = 3,680 BTU/hr Net BTU to Turbine = 418,520 BTU/hr H. P. Equivalent (413,520 BTU/hr)(3.929 x 10~V/BTU/hr) = l64 hp (164 hp)( 0.7457 Kw/hp) = 122 Kw Assume Efficiency of 90 per cent Generator Output 110 Kw, Net Generator Output = 97 Kw

PAGE 173

161 Economic Analysis Initial Investment Delivered Equipment Cost (D. E. C. ) Item Equipment Cost $ Reactor 21,000 Heat Exchanger 7,600 Gas-Liquid Evaporator 1,000 Holding Tank 7,000 High-Pressure Pump 300 Pump Motor 200 Turbine-Generator 24,000 Mixing Motor 370 Oxygen Supply ____^zz. Total 61,470 Physical Plant Cost (P. P. C. ) Delivered Equipment Cost 61,470 Equipment Installation (0.43 D.E.C. ) 26,400 Piping (0.36 D.E.C.) 52,900 Instrumentation (0.15 D.E.C.) 9,220 Insulation (0.08 D.E.C.) 4,910 Electrical (0.01 D.E.C.) 6,150 Land and Yard Improvement (0.07 D.E.C.) 4,300 Total 165,350 Direct Plant Cost (D. P. C. ) Engin. and Construction (0.30 P.P.C. ) Total 2: 4Q.600 ?14,950 Engineering News-Record Cost Index 1018/640 = 1.59 D. P. C. 1955 to 1966 Adjusted Total 342,000 Contractors Fee (0.06 D.P.C.) 20,550 Contingency (0.25 D.P.C.) 85.500 Fixed Capital Investment 448,050 Amortization of Initial Investment Use Straight Line Depreciation Interest = 6 per cent

PAGE 174

162 Salvage Value = 10 per cent ( 1. 59 )( 165,350 ) = 25,600 Annual Depreciation = (P-L) k/n +fL±-J ) i + LI = (448,050 25,600) [1/20 + ^0_+_ip^6~| + 25,600(0.06) = 422,450(0.0815]+ 1540 $ 35,940 Operation and Maintenance/Yr Equivalent of 1 Experienced Laborer 6,000 1/3 of Foreman 3,000 Pump $100 Heat Exchanger 590 Total $690 Adjusted Total (1955 to 1966) 1,100 Oxygen (1,325 Ton/Yr ) ( 12. 00) 21,900 Total Operating Cost/Yr 6?,940 Value of Power Produced @ 2£/Kw hr 17,000 Net Operating Cost/Yr $ 50,940 Cost/lb. of GOD ile moved $50, 940/(1, 825)(2, 000) = $0.01395 = 1.40

PAGE 175

163 275 o o u •p s
PAGE 176

164 Waste Influent Holding Tank High Pressure Volute Pump To Drain CounterCurrent Heat Exchanger High Pressure 175° C Waste Oxygen Reservoir Hot Liquid Effluent Uses of Power Output (1) Drive HighPressure Pump (2) Agitate Contents of Reactor (3) Remainder for Other Plant Needs Power Output + Generator Turbine | Steam, C0 2 » and 2 Gas-Liquid Separator 225°C Effluent Completely-Mixed Reactor Figure 66. Flow Diagram of Autoxidatlon Waste Treatment Process

PAGE 177

165 Heat Exchanger Waste Retention Tank To Drain Turbo-Generator Gas-Liquid Separator Oxygen Storage Figure 67. Schematic Plot Plan of Autoxidation Waste Treatment Process Scale 1" —5' -£> = Pressure ControllerRecorder

PAGE 178

C to >» o H -P O CO 166 GO CO
PAGE 179

167 CI o t i c= o — 600 HPM Motor Universal Joint Packing Box (Continuous Feed) Steam Jacket oJ Baffle Plates Propellers 2 Sparger •Waste Inlet Figure 69. Design of Chemical Reactor

PAGE 180

APPENDIX 13 RECOMENDATIONS FOR FUTURE RESEARCH The equipment and techniques used during the research performed quite well. The technique for obtaining liquid samples from a high-pressure, high-temperature system worked satisfactorily and required no elaborate equipment. The main disadvantage of this teohnique was the amount of time necessary to collect one sample, 8 minutes. While this created no problem in this research a study of very fast reactions would necessitate a decrease in the period of time between samples. The use of several sample collection tubes could decrease the amount of time between samples to about 2 minutes. The performance of the heater could be considerably improved by changing its design. The strip heaters should be removed from the mild-steel cylinder and spring loaded In such a way that when the reactor is placed in the heater casing, the strip heaters would be pressed against the walls of the reactor. This modification would greatly Increase the rate of heat transfer from the strip heaters to the reactor and would allow good automatic temperature control If the thermomonltor-temperature sensor were placed between the strip heaters against the outside wall of the reactor. The third technique used In the study of the autoxldatlon 168

PAGE 181

169 of pyroligneous acid minimized the problems of temperature control in the system (p. 102). The system would simulate a full-scale waste treatment process better if continuous flow of both gases and liquids could be maintained through the reactor. The reactor is amenable to such a conversion. Liquid waste could be added and removed from the reactor through the openings which secure the cooling coll in the top of the reactor. A high-pressure pump could be used to maintain liquid flow through the reactor. An alternative would be to place waste in a reservoir outside the reactor. Flow could be maintained by pressurizing the waste reservoir with an inert gas. No changes would be necessary to maintain continuous oxygen flow through the reactor. The loss of steam through the gas exhaust line might cause problems. A continuous system and adequate experimental techniques could probably be developed in about three months after the delivery of the equipment necessary for the modifications. Several areas whioh require future research are apparent to the author. One of the most important aspects of autoxidatlon is the propogatlon of fast chain reactions. The life of these fast reactions might be increased by: (a) increasing the temperature at whioh an autoxidatlon Is performed, and (b) by removing water from the waste as steam. This second method would hold the concentration of the waste at a higher level as the reaction takes place.

PAGE 182

170 The effects of heavy metal catalysts on the autoxidatlon of different wastes should be evaluated. The surface to volume ratio of reactors should be investigated. Finally, the rate of oxygen transfer from the gas phase to the liquid phase should be studied and procedures should be found for accelerating this inter-phase transfer.

PAGE 183

LIST OF REFERENCES 1. Veijolo, V., and Mustakalio, E. , "Bacteriostatic Activity of Wood Tar. " Annales Medlclnae Experimental! s et Blologlae Fennlae , 41 \ '3, 40^ (1963). 2. Stamm, A. J., and Harris, E. E. , "Chemical Processing of Wood." Chemical Publishing Company, Ino. , New York (1955). 3. Wise, L. E. , and Jahn, E. C. , "Wood Chemistry." Relnhold Publishing Corporation, New York (1952) (American Chemical Society Monograph Series. ) 4. Hedstrom, A., and Ballin, A. 0., Swedish Patent 118,788 (May 13, 19^7). 5. Lindhe, J. H. 0.. "Wood Tar Pitch." Swedish Patent 119,284 (July 22, 1947). 6. Jonsson, G. E. V., "Wood Tar as Fuel in Hot Bulb Engines." Ingenlors Vetenskajss Adakemlen . Handllngar . 193. 177 (1947). 7. Eidus, B. R. , "Motor Oil from Wood Tars." Derevopererabatyvoyushchaya 1 Lesokhlmlcheskaya Promyshelennost , 3. 6. 14 (19 54). 8. Compagine Franoaise de Roffinage, "Hydrogenation of Resins, Resin Oils, and Tars." French Patent 973,624 (February 13, 1956). 9. Bond, D. C. , "Oxidation of Mercaptans. " United States Patent 2,535,833 (December 26, 1950). 10. Uranov, I. P. , Gordov, L. V. , and Gusakov, V. N. , "Wood Tar as an Adhesive for Charcoal Briquets." Gldrollznaya 1 Lesokhlmlcheskaya Promyshelennost . lo-ii . 4 mm: — 11. Mukherjee, S. , and Srlvostava, H. C. , "Plastic from Molasses." Prooedlngs of the Sugar Technological Association of India . 1. 50 (19 54). 12. Lyass, A. M. , Medveder, Ya. I., and Snulova, L. D. , "A Substitute for Bakellte in Shell Molding." Lltelnoe Prolzvodstvo . 9, 10 (1956). 171

PAGE 184

17? 13. Kozlov, V. N., Smolenskii, V. B. , and Sokolov, B. A., "Investigation of the Properties of a New Type of Flotation Compounds Obtained from the Waste Products of Wood Chemical Processing Plants." Przemysl Chemlozny , ^Q, 467 (1959). 14. Chertkov, Y. B. , and Zrelov, V. N. , "Wood Tar as an Antioxidant Additive to Hydrocarbon Fuels." Nefryonoe Khozyalstvo . 21* 9, 70 (1954). 15. Szulicka, J., and Grendysz, J., "The Influence of Retention of Waste Waters from the Dry Distillation of Wood In Ponds on the Harmfulness of the Waste Waters." Gaz. Woda i Technlka Sanitatna , 3,4, 346 (I960). 16. Hlckerson, H. C. , and McMahon, E. K. , "Spray Irrigation of Wood Distillation Wastes." Journal of the Water Pollution Control Federation , 21* H 55 (I960). 17. Grantham, G. R. , "Industrial Waste Study for Cabot Carbon Company." Unpublished report of the Engineering and Industrial Experiment Station of the University of Florida (195D. 18. Ludzaak, F. J., and Ettlnger, M. 3., "Chemicals Resistant to Biological Oxidation." Journal of the Water Pollution Control Federation . 21* H?3 (I960). 19. Jew, K. K. , and Alyea, H. N. , "A Comparison of Organic Inhibitors in Chain Reactions." Journal of the American Chemical Society , ^, 575 (1933). 20. Zimmerman, F. J., "New Waste Disposal Process." Chemical Engineering . 6£, 17, 117 (August 25, 1953). 21. Teletzke, G. H. , "Wet Air Oxidation." Chemical Engineering Progress . 65 , 1, 34 (January, 1964). 22. Zimmerman, F. J., "The Zimmerman Process and Its Application to the Pulp and Paper Industry." TAPPI, 4j, 8, 711 (I960). 23. Committee on Sanitary Engineering, Research of the Sanitary Engineering Division, "Sludge Treatment and Disposal by the Zimmerman Process." Journal of the Sanitary Engineering Division. Procedlngs of the American Society of Civil Engineers , Paper 2087, 8£, 3A-4, 15 (July, 1959). 24. Jackson, D. T. , and Brown, R. W. , "Oxidation of Spent Semichemioal Pulping Liquors." Pulp and Paper , 69 (February, 1958).

PAGE 185

173 25. Koenig, L. , "Wet Oxidation. " Ultimate Disposal of Advancedtreatment Waste . Part 1. Environmental Health Series , Water Supply and Pollution Control, U. S. Department of Health, Education, and Welfare, Public Health Service, AWTR-3» 13 (October, I963). 26. Waters, W. A., "The Chemistry of Free Radicals." 2nd ed. , Oxford University Press, London (1948). 27. Bateman, L. , "Olefin Oxidation." Quarterly Reviews (London) . 8, 147 (1954). 28. Ingold, K. U. , "Inhibition of the Autoxidatlon of Organic Substances In the Liquid Phase." Chemical Reviews . 61, 563 (196l). 29. Ingold, K. U., "Kinetics of Oil Oxidation Inhibitors." Journal of the Institute of Petroleum . 45 . 244 (1959). 30. Farmer, E. H. , "Peroxidation in Realtion to Olefinic Structure." Transactions of the Faraday Society . 4£, 228 (1946"n 31. Khan, N. A., "Initiation Processes in the Autoxidatlon of Fatty Acid Esters and Allied Substances. " Journal of Chemical Physics . 22, 2090 (1954). 32. Bolland, J. L. , "Kinetics of Olefin Oxidation." Quarterly Reviews (London) . 2* 1 (1949). 33. Huang, R. L. , and Singh, S. , "The Relative Stabilizing Influences of Substltuents on Free Alkyl Radicals, Part VI." Journal of the Chemical Society . 3183 (1959). 34. Russell, G. A., "The Rates of Oxidation of Aralkyl Hydrocarbons. Polar Effects In Free Radical Reactions." Journal of the American Chemical Society . £8, 1047 (19567: 35. Mayo, F. R. , "Oxidation of Unsaturated Compounds, Part V." Journal of the Ameri can Chemical Society, 80, 2465 {T9WT. 36. Emanuel, N. M. , Maizus, Z. K. , and Provalova, L. G. , "Mechanism of Deoane Liquid-Phase Oxidation." International Journal of Applied Radiation and Isotopes . 7_ T HI (19^9). 37. Stannett, V., and Mesroblan, R. B. , "Studies on the Decomposition of l:4-dimethyl Cyclohexane Hydroperoxide In l»4-dimethyl Cyclohexane in the Absence of Oxygen." Discussions of the Fara day Society, lit, 242 ( 1953"n

PAGE 186

can 174 38. Farkas, A., and Passaglla, E. , "The Decomposition of Cyclohexyl Hydroperoxide and the Peroxide-catalyzed Polymerization of Styrene." Journal of the America Chemical Society , 7_2, 3333 (1950). 39. Bennett, J. E. , "Study of the Oxidation of Hindered Phenols by Electron Spin Resonance." Nature , 186 , 385 (i960). 40. Doering, W. Von E. , Okamoto, K. , and Krauch, H. , "On the Distribution of Oxygen18 In Trlphenylmethyl Benzoate from the Reaction of Trlphenylmethyl and Benzoyl Peroxide carbonyl 0. " Journal of the American Chemical Society , 82 , 3579 (1956). 41. Batten, J. J., "Retardation of the Chain-decomposition of Benzoyl Peroxide by Polycycllc Aromatic Hydrocarbons." Journal of the Chemical Society , 468? (1956). 42. Thomas, J. R. , and Harls, 0. L. , "Substrate Effects on the Decomposition of Alkyl Hydroperoxides and Their Influence upon Autoxldation. " Journal of Physical Chemistry , 6J, 1027 (1959). 43. Shushunov, V. A. , and Sokolov, N. A. , "The Benzoyl Peroxide Decomposition Kinetics In Alcohol-Benzene Solutions." Zhurnal Fizlcheskoy Khlmll . }2, 1796 (1958). 44. Stannett, V., and Mesrobian, R. B. , "The Kinetics of the Decomposition of Tertiary Hydroperoxides in Solvents." Journal of the American Chemical Society , 12, 4125 ( 19507: 45. Tsunoda, Y. , Matsumoto, K. , and Kato , T. , "Autoxldation of Isopropyl Benzene, Part V." Tokal Dendyoku Glho , 1£, 41 (1958). 46. Walling, C. , and Chang, U. W. , "Chain Transfer in the Hydroperoxide Initiated Polymerization of Styrene." Journal of the American Chemical Society , 76, 4873 (1954). 47. George, P., "The Liquid Phase Oxidation of Hydrocarbons. Part I. The Free Radical Character of the SurfaceCatalysed Oxidation of Tetralin. " Transactions of the Faraday Society . 42, 210 (1946). 48. Swern, D. , "Organic Peraclds. " Chemical Re views , 4£, 1 (1949).

PAGE 187

175 49. Bateman, L. , and Hughes, H. , "The Thermal Decomposition of Gyclohexenyl Hydroperoxide in Hydrocarbon Solvents." Journal of the Chemical Society , 4594 (1952). 50. Denison, G. H. , Jr., and Condit, P. C. , "Oxidation of Lubricating Oils, Mechanism of Sulfur Inhibition." Industrial and Engineering Chemistry , 37 , 1102 (W5). 51. Denison, G. H. , Jr., "Oxidation of Lubricating Oils, Effects of Natural Sulfur Compounds and of Peroxides." Industrial and Engineering Chemistry , 36 , 477 (1944). 52. Denison, G. H. , Jr., and Condit, P. C. , "Oxidation of Lubricating Oils, Dialkyl Selenides as Inhibitors." Industrial and Engineering Chemistry . 4l, 944 (1949). 53. Kharasch, M. S. , Fono, A., and Nuderberg, W. , "The Chemistry of Hydroperoxides, I. The AcidCatalyzed Decomposition of cXj^Dimethylbenzyl (o<-cumyl) Hydroperoxide." Journal of Organic Chemistry . 15 , 748 (1950). 54. Williams, A. L. , and Offenhauer, R. D. , "Acidic Components of Catalytically Cracked Distillate Fuel Oil." Industrial and Engineering Chemistry , 4o_, 1259 (1957). 55* Farnand, J. R. , Mcllhinney, A. E. , Peterson, W. S. , Gishler, D. E. , and Puddlngton, I. E. , "Chemical Bleaohlng of Hydrocarbon Oils." Canadian Journal of Technology . 21* ^26 (1955). 56. Chalk, A. J., and Smith, J. F. , "Catalysis of Cyclohexene Autoxidation by Trace Metals in Non-polar Media." Transactions of the Faraday Society . 53. 1214 ( 19577: 57. Walling, C. , "Free Radicals in Solution." John Wiley and Sons, Inc., New York (1957). 58. Hammond, G. S. , and Nandl , U. S. , Private Communication with Ingold, K. U. 59. Benson, D. , and Sutcliffe, L. H. , "The Decomposition of tert . -Butyl Hydroperoxide by Plumbic Acetate in Acetic Acid." Transactions of the Faraday Society . ££, 2107 (1959)": 60. Bolland, J. L. , and Haave , P. ten, "Kinetic Studies In the Chemistry of Rubber and Related Materials, the Inhibitory Effect of Hydroqulnone on the Thermal Oxidation of Ethyl Linoleate." Transactions of the Faraday Society . 43, 201 (1947).

PAGE 188

176 61. Scanley, S. C. , and Larson, R. , "Stabilization of Additives for Lubricating Oils." S.A.5. Journal , 107C , 13 (1958). 62. Matthews, F. W. H. , "Beaker-corrosion Test for Lubricating Oils." Journal of the Institute of Petroleum. 21* ^36 (1949). 63. George, P., and Robertson, A., "The Liquid Phase Oxidation of Hydrocarbons, II. The Free Radical Character of the Heavy Metal Catalysed Oxidation." Transactions of the Faraday Society . 42 . 217 (1946). 64. Woodward, A. E. , and Mesroblan, R. B. , "Low Temperature Autoxidatlon of Hydrocarbons, the Kinetics of Tetralln Oxidation." Journal of the American Chemical Society . 21* 6I89 (1953). 65. Knorre, D. G. , Maizus, Z. K. , and Emanuel, N. M. , "Initial Macroscopic Stage in the Catalist Conversion During the Oxidation of Decane in the Presence of Mn(CiiH 2 3-COO)2." Dokaady Akad Nauk SSSR . £2, 415 66. Pedersen, C. J., "Inhibition of Deterioration of Cracked Gasoline During Storage." Industrial and Engineering Chemistry . 4l, 924 (194977 67. Mendelsohn, M. , Arnett, E. M. , and Frelser, H. , " "Destructive Autoxidatlon of Metal Chelates, I. Effects of Variation of Ligand and Metal on Initial Rate." Journal of Physical Chemistry . 64. 660 (i960). 68. Russell, G. A., "Deuterium-isotope Effects in the Autoxidatlon of Aralkyl Hydrocarbons. Mechanism of the Interaction of Peroxy Radicals." Journal of the American Chemical Society . 79 , 3871 ( 1957). 69. Blanohard, H. S. , "A Study of the Mechanism of Cumene Autoxidatlon, Mechanism of the Interaction of t-Peroxy Radicals." Journal of the American Chemic al Society . 7_8, 1047 ( 1956T 70. Boozer, C. E. , Hammond, G. S. , Hamilton, C. E. , and Sen, J. N.| "Air Oxidation of Hydrocarbons. II. The Stolchiometry and Fate of Inhibitors In Benzene and Chlorobenzene. " Journal of the American Chemical Society . 2Zf 3233 (1955). 71. Dunn, J, R. , Waters, W. A., and Wickham-Jones , C. , "The Retardation of Benzaldehyde Autoxidatlon. Part III." Journal of the Chemical Society . 2753 (1952).

PAGE 189

177 72. Hey, M. E. , and Waters, W. A., "Some Products Formed From the Autoxidation of Cumene , Part II." Journal of the Chemical Society , 2753 (1955). 73. Cook, C. D. , and Norcross, B. E. , "Oxidation of Hindered Phenols. VII. Solvent Effects on the Disproportionatlon of Certain Phenoxy Radicals." Journal of the American Chemical Society , 81 , 1176 (1959). 74. Blokel, A. F. , and Kooyman, E. C. , "Alkylperoxy Radical Radicals. III. Retarded by Aromatic Amines." Journal of the Chemical Society , 22 17 (1957). 75» McGowan, J. C. , and Powell, T. , "Some Reactions of the Radical 2 ,4,6-Tri-t-butylphenoxyl. " Journal of the Chemical Society , 238 (I960). 76. Kennerly, G. W. , and Patterson, W. L. , Jr., "Kinetic Studies of Petroleum Antioxidants." Industrial and Engineering Chemistry . 48, 1917 ( 195677" 77. Pedersen, C. J., "Mechanism of Antioxidant Action in Gasoline." Industrial and Engineering Chemistry , 48, 1881 (1956T 78. Moore, R. F. , and Waters, W. A., "The Retardation of Benzaldehyde Autoxidation, Part IV." Journal of the Chemical Society , 2432 (1952). 79. Frye, F. , Kretschmer, C. B. , and Wiebe, R. , "Effect of Alcohols on Hydrocarbon Autoxidation." Industrial and Engineering Chemistry , 46 , 1516 ( 1954). 80. Zuidema, H. H. , "Performance of Lubricating Oils." Reinhold Publishing Corporation, New York (1952). 81. "Standard Methods for the Examination of Water and Waste Water." Published by American Public Health Association, American Water Works Association, Water Pollution Control Federation, Naw York (i960). 82. Babbitt, H. E. , and Bauman, R. E. , "Sewage and Sewage Treatment." John Wiley and Sons, Inc., New York (1958). 83. Zoss, L. M. , Suciu, S. N. t and Sibbltt, W. L. , "The Solubility of Oxygen in Water." Transactions of the American Society of Mechanical Engineers , 76, 70 (1954). 84. Pray, H. A., Schweickert, C. E. , and Minnich, B. H. , "Solubility of Hydrogen, Oxygen, Nitrogen, and Helium in Water at Elevated Temperatures." Industrial and Engineering Chemistry . 44 , 1149 ( 1952).

PAGE 190

178 85. "Steam: Its Generation and Use." Babcock and Wilcox Company, New York (1963). 86. Brown, A. I., and Marco, S. M. , "Introduction to Heat Transfer." McGraw-Hill Book Company, Inc., New York, Toronto, London (1958). 87. Hicks, T. G. , "Pump Selection and Application." McGraw-Hill Book Company, Inc., New York, Toronto, London (1957). 88. "Handbook of Chemistry and Physics." 4lst Ed., Chemical Rubber Publishing Company, Cleveland, Ohio (i960). 89. "Chemical Engineer's Handbook." 3rd Ed., McGraw-Hill Book Company, Inc., New York, Toronto, London (1950). 90. Thuesen, H. G. , "Engineering Economy." Prentice-Hall, Inc., Englewood Clifts, New Jersey (1957). 91. Aries, R, G. , and Newton, R. D. , '•Chemical Engineering Cost Estimation." McGraw-Hill Book Company, Inc., New York, Toronto, London (1955).

PAGE 191

BIOGRAPHICAL SKETCH Boyd T. Riley, Jr. was born in Nashville, Tennessee, on July 16, 1939. He attended public primary and secondary schools in Nashville, with the exception of one year at Duncan College Preparatory School for Boys. He was graduated from Hillsboro High School in June, 1956. He entered Vanderbilt University, located in Nashville, in September, 1956, and was awarded a Bachelor of Engineering in June, i960. During his four undergraduate years he worked as a student Instructor at three summer engineering camps for Vanderbilt. In addition, he worked part time and summers in the Bidding and Estimating Department of Oman Construction Company of Nashville. After receiving a degree, he worked for one year for Oman Construction Company. He entered Texas A. & M. College at College Station, Texas, in September, 1961 and was awarded a Master of Science in June, 1962. Hydrology and Hydraulics were his major areas of study. From June, I962, until September, I962, he worked for Scarritt College in Nashville, Tennessee. In September, I962, he entered Oklahoma State University in Stillwater, Oklahoma, to undertake a program of study leading to the degree of Doctor of Philosophy in Sanitary Engineering. In August of 1963* he transfered to a similar program at the University of Florida. He is married to the former Barbara Layne Parks of Nashville. 179

PAGE 192

This dissertation was prepared under the direction of the chairman of the candidate's supervisory committee and has been approved by all members of that committee. It was submitted to the Dean of the College of Engineering and to the Graduate Council and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy. June 21, 1966 <*-
PAGE 194

44 IT