Title: Catalytic reactions of aliphatic alcohols over thorium oxide
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00097934/00001
 Material Information
Title: Catalytic reactions of aliphatic alcohols over thorium oxide
Physical Description: vi, 65 l. : illus. ; 28 cm.
Language: English
Creator: Legg, John Wallis, 1936-
Publication Date: 1964
Copyright Date: 1964
 Subjects
Subject: Catalysts   ( lcsh )
Thorium oxide   ( lcsh )
Alcohols   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis - University of Florida.
Bibliography: Bibliography: l. 63-64.
Additional Physical Form: Also available on World Wide Web
General Note: Manuscript copy.
General Note: Vita.
 Record Information
Bibliographic ID: UF00097934
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000421878
oclc - 11020929
notis - ACG9876

Downloads

This item has the following downloads:

PDF ( 2 MBs ) ( PDF )


Full Text












CATALYTIC REACTIONS OF ALIPHATIC

ALCOHOLS OVER THORIUM OXIDE














By
JOHN WALLIS LEGG


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












UNIVERSITY OF FLORIDA


December, 1964 /















ACKNOWLEDGMENTS


The author wishes to express his appreciation to those whose

contributions, both direct and indirect, have aided in the com-

pletion of this research: to Dr. W. S. Brey, Jr., Director of this

research and Chairman of the Supervisory Committee, whose advice,

encouragement and guidance were essential factors in the under-

taking and completion of this work; to Mrs. Kathryn Scott for the

preparation of the nuclear magnetic resonance spectra, Mr. Douglas

Davis for the mass spectra determinations and Dr. Lamar Miller for

the distillations of reaction products; and to the Atomic Energy

Commission for financial support of this research.

A special word of appreciation is due the author's wife,

without whose encouragement and understanding this project would

probably not have been undertaken and could certainly not have

been completed. The typing of this dissertation is just one

example of her many assistance.















TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS . . . . . . . .... ii

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

LIST OF TABLES . . . . . * .* vi

CHAPTER

I. INTRODUCTION AND HISTORICAL REVIEW . . . . 1

II. EXPERIMENTAL METHODS . . . . . . . 6

A. The Catalyst . . . . . . . . 6

1. Preparation .. . . . . . . 6

2. Surface areas .. . . . . . 9

B. Reaction Runs . . . . .. 10

1. Purification of alcohols . . . 10

2. Reaction apparatus . . . 10

3. Procedure for a series of reaction runs. 13

4. Analysis of reaction products . 15

III. PRESENTATION OF DATA . . . . . . 18

A. Explanation of Tables . . . . . 18

B. Results . . . . . . . . . 28

1. Dehydration . . . . . . 28

2. Dehydrogenation . . . . . ... 37

3. Other reactions of 1-propanol and 1-butanol 39

IV. DISCUSSION OF RESULTS AND CONCLUSIONS .. . . 42


iii









TABLE OF CONTENTS CONTINUED

Page

V. SUMMARY .......... . . ...... .51

APPENDIX . . . . . . . . ... . . . 53

BIBLIOGRAPHY ................... .... 65

BIOGRAPHICAL SKETCH ....... . . . . . 65














LIST OF FIGURES


Figure Page

1. Reaction apparatus ....... . . . ... 11

2. Dehydration of ethanol over catalyst II as a

function of reciprocal flow rate . . . . 29

3. Dehydration of ethanol over different catalysts

as a function of reciprocal flow rate . . . 31

4. Dehydration of alcohols over catalyst V as a

function of reciprocal flow rate . . . ... 33

5. Dehydration of alcohols over catalyst VI as a

function of reciprocal flow rate . . . . . 34

6. Dehydration of primary alcohols as a function

of catalyst surface area . . . . . . .. 36

7. Dehydration of secondary alcohols as a function

of catalyst surface area . . . . . . 38

8. Conversion of 1-propanol to 3-pentanol and

1-butanol to 4-heptanone as a function of.

reciprocal flow rate . . . . . . . 41















LIST OF TABLES


Table Page

I. Reaction runs with ethanol ... . .... 19

II. Reaction runs with 2-propanol .. . . . 23

III. Reaction runs with 1-propanol . . .. 25

IV. Reaction runs with 1-butanol .. .. . . 26

V. Reaction runs with 2-butanol .. . . . 27














I. INTRODUCTION AND HISTORICAL REVIEW


Extensive investigations have been made correlating the

chemical properties of catalysts and their activity in promoting

reactions in the gas phase. A number of purely chemical theories

have been proposed, one of the most notable being the multiplet

theory" of Balandin (1). He proposed that, because of its chemi-

cal nature, a catalyst possesses certain sites which cause re-

actant molecules to be arranged in such a way as to promote the

breaking of certain bonds and the formation of others. Although

much of Balandin's theory is almost certainly in error, notably

the absence of chemisorption, such chemical properties as crystal

structure and the availability of bonding electrons at the sur-

face are important factors in determining the activity of a

catalyst.

Although early investigators realized that the physical pro-

perties of the catalyst surface were probably important, only in

relatively recent years have methods been made generally available

to the chemist for investigation of these physical properties.

Thus, alcohols have been reacted over thorium oxide by a number of

investigators since 1908 when Sabatier and Maihle (2) effected

the decomposition of ethanol, but often the data are difficult to

correlate, since little is known of the physical properties of the

catalysts used. One of the purposes of the present investigation





2


has been to correlate one physical property, the surface area,

with the reactivities of a number of alcohols over different

thorium oxide catalysts.

It is well known that thorium oxide, or thoria, effects both

the dehydration and dehydrogenation of alcohols. Adkins (3) was

probably the first to propose that, over catalysts which promoted

both reactions, the physical properties of the catalyst surface

affected the relative rates of the two reactions and that these

reactions occurred on different sites.

Hoover and Rideal (4), reacting ethanol over thoria which

had been prepared by precipitation from thorium nitrate, found

approximately equal rates for the two reactions. Experiments

with water vapor as a poison showed that adsorption of water de-

creased the dehydration to a greater extent than the dehydrogena-

tion. They concluded that there were separate partial surfaces

for each reaction.

G. M. Schwab (5) extended this idea in making a proposal as

to the nature of each of these partial surfaces. His investiga-

tions of the decomposition of ethanol and formic acid over

thorium oxide, as well as a number of other oxides, led him to

conclude that "the active centers for the dehydration of ethanol

and formic acid are identical, and the same is true for the

centers of dehydrogenation." Treatment of the catalyst such as

prolonged heating, which decreased the number of pores, crevices

and channels, decreased considerably the dehydration and increased

slightly the dehydrogenation. In electron microphotographs, at

100 angstroms resolution, of strongly dehydrating and strongly








dehydrogenating catalysts the former appeared compact while the

latter resembled a loose powder. Thus, Schwab reasoned, the

pores in the dehydrating catalyst were of the order of molecular

size. He proposed that dehydrogenation occurs on the flat sur-

faces of the catalyst while dehydration occurs in pores and cre-

vices of such size that the reactant molecule is attached to

opposing sides.

If Schwab's proposal is correct, an increase in size of the

reactant molecule should have an adverse effect on the dehydra-

tion reaction. In the investigation described in this disserta-

tion, ethanol, 1-propanol, 2-propanol, 1-butanol and 2-butanol

have been reacted over thoria in an effort to determine this

effect of molecular size.

Little work has been done until recently on the reactions

of the higher alcohols over thoria. Winfield (6) reacted 2,

3-butanediol over a number of catalysts in an effort to obtain

methyl vinyl carbinol and then butadiene in a.second dehydration.

Alumina, quartz chips, B203, and BeO, among others, yielded

methyl ethyl ketone. Thoria alone effected the reaction with the

less favorable free energy change, that to methyl vinyl carbinol.

Winfield's catalysts prepared from thorium oxalate showed much

more reactivity than those from the hydroxide. Investigations

in this laboratory (7) have shown that thoria prepared from the

oxalate has a much higher surface area than that from the hydro-

xide, especially as prepared by Winfield.

In a recent article Lundeen and Hoozer (8) point out that

Winfield's production of methyl vinyl carbinol is just one








example of the remarkably specific nature of thoria as a dehydra-

tion catalyst. Whereas alumina (9) and numerous other metal

oxides (10) yield mixtures rich in the 2-olefin upon dehydration

of secondary 2-alcohols, thoria and other oxides of group III B

catalyze the dehydration of the 2-alcohols to yield the 1-olefin

almost exclusively. In contrast to the carbonium ion and oxonium

ion mechanisms usually proposed for the elimination of water over

alimina (9) (10), Lundeen and Hoozer propose a concerted mecha-

nism for the reaction over thoria, in which the elements of water

are eliminated as the olefin is formed, in a single step. A

hydrogen from the methyl group alpha to the hydroxyl carbon, in-

stead of a hydrogen from the alpha methylene group, attaches to

the surface and is eliminated with the hydroxyl group. To attach

a hydrogen from the methylene group, the molecule would have to

overcome eclipsing interactions and bring a larger portion of the

molecule nearer the surface. The concerted mechanism is similar

to that proposed by Schwab (5), but these investigators do not

discuss the porosity of the catalyst.

Although alumina and similar catalysts quite easily promote

the dehydration of primary alcohols to ethers (10), dehydration

to ether does not occur over thoria (11) (12).

Most investigators have found that secondary alcohols dehy-

drate much more readily than do the corresponding primary alco-

hols and that, in general, the higher primary alcohols dehydrate

mor, readily than those of lower molecular weight. Bork and

Tols-toyatowa (13), in working with alumina, put the order as

2-propanc1 > 2-butanol > 2-methyl-l-propanol >1-butanol >








1-propanol>ethanol; the same sequence was obtained by Freidlin

and Levit (14) using as a catalyst tricalcium phosphate. As will

be discussed later, these results are in contrast to those found

in the present investigation of thoria.

At the present time, no single comprehensive mechanism of

the reactions of alcohols over thoria has been proposed. Any

mechanism would have to explain both dehydration and dehydrogena-

tion, including the favoring of one reaction over the other with

a given catalyst, the non-formation of ethers, and the specificity

of the production of 1-olefin to that of 2-olefin from secondary

2-alcohols.















II.. EXPERIMENTAL METHODS

A. The Catalyst


:. Preparation

All of the catalysts were prepared by precipitation of

thorium hydroxide from a solution of thorium nitrate by the rapid

pouring of a more-than-theoretical amount of concentrated ammon-

ium hydroxide into the solution. Some of the thorium nitrate was

obtained from the Fisher Scientific Company, the rest from the J.

T. Baker Chemical Company. The source of the ammonium hydroxide

was the Allied Chemical and Dye Corporation.

After precipitation the hydroxide was filtered and washed

with distilled water; the number of washings and volume of each

wash varied between batc.es. -he precipitate was dried overnight

at 120, ground to a 200 mesh, and heated under vacuum at 300

for two hours. Pellets one-half inch in diameter and one-sixteenth

to one-eighth inch thick were made at a pressure of about 92,000

pounds per square inch; these were usually cut in half before

introduction into the reaction tube. The pelleted catalyst was

then activated by heating under vacuum at 6000 for at least four

hours. The heating at 3000 was done to remove enough water so

that the pellets did not crack or crumble during the final

activation.


1. ll1 temperatures are in degrees centigrade.

6







a. Catalyst I

One-hundred twenty grams of Th(N03)4.4H20 was dissolved in

1,500 milliliters of water and 150 milliliters of concentrated

ammonium hydroxide rapidly poured in with stirring. The mixture

was suction filtered and the precipitate washed, while still on

the filter paper, until the filtrate had a pH of seven, as deter-

mined with pHydrion paper. After drying and pelleting, the

catalyst was activated for four hours.

b. Catalyst II

Eighty grams of Th(N03)4.4H20 was dissolved in 100 milli-

liters of water and 60 milliliters of concentrated ammonium hydro-

xide added by rapid pouring with stirring. The mixture was suction

filtered and washed with five portions, 500 milliliters each, of

water while the precipitate was still on the filter paper. The

final pH of the filtrate was six. After drying and pelleting, the

catalyst was activated for four hours.

c. Catalyst III

Ninety grams of Th(N03)4.4H20 was dissolved in 1,100 milli-

liters of water and 110 milliliters of concentrated ammonium hydro-

xide poured in rapidly with stirring. After filtration the pre-

cipitate was removed from the filter paper and thoroughly stirred

with 500 milliliters of water. The precipitate was allowed to

settle and the supernatent liquid pulled off with an aspirator.

Another 500 milliliters of water was added; the process was re-

peated until 7,500 milliliters of water had been used and the pH

was seven. After the.usual drying and pelleting, the catalyst

was activated for four hours.







d. Catalyst IV

One-hundred twenty grams of Th(N03)4.4H20 was dissolved in

enough water to make 525 milliliters of solution, then 300 milli-

liters of concentrated ammonium hydroxide was added by rapid

pouring with stirring. After suction filtration the precipitate

was removed from the filter paper and stirred with 400 milliliters

of water. This mixture was then filtered and the process repeated

until a total of 2,000 milliliters of wash water was used. This

catalyst was activated a total of twenty-five hours prior to

introduction into the reaction vessel.

e. Catalyst V

Ninety grams of Th(NO3)4.4H20 was dissolved in 1,100 milli-

liters of water and 110 milliliters of concentrated ammonia

rapidly poured in with stirring. The resulting mixture was suc-

tion filtered. After being removed from the filter paper the

precipitate was stirred with 400 milliliters of water. This mix-

ture was filtered and the process repeated until a total of 2,000

milliliters of wash water was used. Catalyst V was activated for

a total of twenty-nine hours prior to introduction into the

reaction tube.

f. Catalyst VI

One-hundred twenty grams of Th(N03)4.4H20 was dissolved in

enough water to make 525 milliliters of solution and 300 milli-

liters of concentrated ammonium hydroxide was added by rapid

pouring with stirring. The resulting mixture was filtered and

the precipitate removed from the filter paper. It was stirred

with 400 milliliters of water and this mixture filtered. This








was repeated four times for a total wash volume of 1,600 milli-

liters. Prior to introduction into the reaction tube catalyst

VI was activated a total of twenty-nine hours.

Catalyst I was activated in a Vycor tube, a portion saved

for surface area measurements and 20.0 grams placed in the re-

action tube. Pellets of catalyst II and III were introduced into

the reaction tube prior to activation, the final activation talking

place in the reaction tube itself. The portion for surface area

measurements was activated separately.

It became apparent during runs with the first three catalysts

that it would be desirable to reactivate the catalysts at times

between runs. However, it has been shown (7) that this reactiva-

tion decreases thie surface area of the catalyst. This decrease is

rapid at first; after several hours activation the surface area

decreases less drastically with heating. In order to be able to

reactivate without changing appreciably the surface area of the

catalyst, catalysts IV, V and VI were each activated for several

hours prior to irntoduction in-o the reaction chamber, surface

areas being determined amfer each few hours activation until

little change was noted with further heating.

2. Surface areas

The nitrogen surface areas, in square meters per gram, of the

six catalysts used, determined by the B.E.T. method (15) (18),

are listed below. The activation time is the total time of acti-

vation prior to the first reaction run over the catalyst.

Catalyst Activation Time Surface ALa
I 4 hours 1,2
II 4 hours 6.5









Catalyst Activation Time Surface Area
III 4 hours 17.8
IV 26 hours 34.4
V 30 hours 12.2
VZ 30 hours 28.9



1- Reaction .uns

1. Purification of alcohols

The ethanol was obtained from the Union Carbide Chemicals

Company, the 2-propanol from the Mallinkrodt Chemical Works, the

1-propanol and 1-butanol from the Fisher Scientific Company and

the 2-butanol from Eastman Crganic Chemicals. In order to remove

aldehyde impurities and water all the alcohols except 2-butanol

were stored from several days to several weeks over calcium oxide

which had been heated at 800 for several hours. They were then

distilled until no impurity could be detected with a vapor phase

chromatograph constructed in this laboratory; two distillations

were usually sufficient. The reage-t grade 2-butanol was used as

obtained; no impurities could be detected in this material with

the chromatograph.

2. Reaction apparatus

The flow system shown schematically in Figure 1 was used for

all reaction runs except for some of the initial trial runs. It

permitted the passage of alcohol over the catalyst at various flow

rates and temperatures and the collection of the liquid and gaseous

products.

The reaction vessel was made of a Vycor tube 81 centimet:_-

lorn .xith an inside diameter of 22 millimeters. A Vycor thermo-

ccupi well extended through the top of the reaction tube to the






















1-I






















^ C



M rq
P4




















coo Co
cz 1
4-4




> d ( *lH

0 O 0, 00 r0-
0 P4 0Hop o 0
C 0 C4 ) --*H 0r
4 0 0 0 0QOP4-
C a 0 Q -0 0 4 0 'd Cd 4 )
4-) d r Cd 4- o d 00 WO E-
o > p, o V 3 o 0 a a00
mOro 0 W (D Cd5 540 /CO4
0 rH 0 o 0 0o 0 :o r- 0u M a)



H ( r i O O O tt H (0








center of the catalyst bed. The catalyst was supported by a

perforated porcelain disc which in turn was supported by a spacer

tube resting on indentations in the reaction tube. Reactant alco-

hol entered the reaction tube through a sidearm near the top of

the reaction tube.

The furnace, a porcelain tube with three separate heating

coils and appropriate insulation, has been described by Schmidt

(12). The middle coil, which surrounded the catalyst bed, was

connected through a Variac to a model 402 Wheelco Proportioning

Capacitrol. The chromel-alumel thermocouple which was used in

conjunction with the Capacitrol was inserted in a thermocouple well

outside the porcelain tube to a point just outside the catalyst

bed. Temperature control above and below the catalyst was main-

tained by manual manipulation of two Variacs to which the top and

bottom coils were connected.

A second thermocouple was inserted into the reaction tube

thermocouple well to the center of the catalyst bed. The reaction

temperature was then read from a second Wheelco Capacitrol; this

second Capacitrol was not used as a regulating device but simply

as a meter. The temperature could be controlled and determined to

within two degrees in the range of reaction temperatures used,

3150 to 420.

The reactant alcohol was contained in a graduated reservoir

situated above the reaction vessel. From the reservoir the alco-

hol flowed through a Fischer and Porter Precision Bore Flowrater

No. 08-150 and a Fischer and Porter glass and teflon needle valve

which regulated the rate of flow. Between the needle valve and








the reaction chamber was placed a "T" with vacuum stopcock and lines

to a manometer, vacuum pump and nitrogen tank.

Vycor chips were always packed above the catalyst to a

height of about seven inches. By regulation of the voltage to the

top coil in the furnace the temperature of these chips could be

made the same as that of the catalyst bed, as determined by raising

the thermocouple in the reaction tube well. Thus the reactant

alcohol flowed onto these chips, evaporated, and was heated to

reaction temperature prior to its contact with the catalyst.

Upon leaving the reaction tube the reaction products and

unreacted alcohol passed into the first product collection flask.

Here all materials which were liquids at room temperature were

trapped out. The remaining products passed through a second collec-

tion trap which was cooled with a dry ice-acetone slush. The re-

maining gases passed through a gas sampling bulb and on through a

Wet Test Meter, manufactured by the Precision Scientific Company,

which measured the total volume of these gases. The Wet Test

Meter was calibrated by passing known volumes of nitrogen through

it at approximately the same pressure differentials as those which

existed during reaction runs.

3. Procedure for a series of reaction runs

Twenty grams of pelleted catalyst was placed on the per-

forated disc, about seven inches of preheater chips packed on top

of the catalyst, and the reaction tube placed in the furnace.

Catalysts I, II and III were activated for four hours after

the reaction tube was in place. Although catalysts IV, V and VI

were activated before being placed in the reaction tube, they were








each reactivated for one hour after placement to remove water

adsorbed during the transfer.

The three heaters were turned on and the preheater and

catalyst bed allowed to come to reaction temperature under an

atmosphere of dry nitrogen. Alcohol was introduced, the flow rate

adjusted to that desired, and the system allowed to equilibrate

for ten to thirty minutes before an actual run was started.

After equilibration, the time, temperature, volume of alco-

hol in the reservoir and wet test meter reading were recorded and

the two collection flasks changed as quickly as possible. The

alcohol flowed for thirty minutes to two hours until at least ten

milliliters had passed into the reaction tube, with periodic checks

on the flow rate and reaction temperature. To conclude a run,

final readings were made and the flasks again quickly changed.

Another period of equilibration at a new flow rate preceded

the next run. Usually four runs were made at a given temperature,

three at different flow rates; the fourth run duplicated the first

as a check on any change in catalyst activity.

After a series at a given temperature was completed, the

unreacted alcohol and reaction products still in the reaction tube

were pumped out and the catalyst heated under vacuum about thirty

minutes at the reaction temperature.

A number of times two series of runs, at different tempera-

tures, were made in a single day. When this was done, the alcohol

continued to flow after the first series was completed while the

system was coming to thermal equilibrium at the new temperature.









4. Analysis of reaction products

a. Liquid products

A vapor phase chromatograph, described by Moreland (15), was

used for all quantitative and most qualitative analyses of liquid

products. Attempts were made to analyze reaction products from

1-propanol using nuclear magnetic resonance (NMR) spectroscopy,

but vapor phase chromatography proved more satisfactory. The

chromatograph column, twenty feet of one-fourth inch copper

tubing, was packed with F. and M. Scientific Corporation's 5 per

cent Carbowax 1500 on Haloport F.

During the analysis of ethanol, 2-propanol, and 1-butanol,

the column temperature was 950 and the pressure of the helium

carrier gas was seven pounds per square inch. The temperature was

600 and the helium pressure was five and one-half pounds for the

analysis of 1-propanol and 2-butanol reaction products.

Most of the reaction products were identified by comparing

retention times on the column to those of known compounds or by

adding a small amount of the known to the product mixture and

noting which peak increased in size. This identification is

discussed in more detail in the Appendix.

Although trace amounts of any number of reaction products

are to be expected in catalytic reactions, no more than 1 per cent

of components other than the initial dehydration and dehydrogena-

tion products were obtained with ethanol, 2-propanol or 2-butanol.

However, 1-propanol and 1-butanol yielded in appreciable amounts

products, one for each alcohol, which were not expected from

simple dehydration or dehydrogenation.








In order to identify these two components, reaction mixtures

of both 1-propanol and 1-butanol were distilled using a spinning

band column. The 1-propanol product had a boiling point in the

range ll4*-117 and the 1-butanol product boiled in the range

142o-145o. Infrared and NMR spectra were obtained for both pro-

ducts. It was concluded from this evidence that the 1-propanol

product was 3-pentanol and the 1-butanol product was 4-heptanone.

Solutions of known composition were made of each of the alco-

hols and known reaction products, chromatograms were made of each

solution, and relative areas of the peaks were determined using a

planimeter.

It was found, in working with numerous compounds, that the

relative areas of two peaks on a chromatogram were very nearly

the same as the relative weights of the two components in solu-

tion. For example, a 20 per cent by weight (39 mole per cent)

solution of water in ethanol produced a chromatogram on which the

water peak was 21.5 per cent of the total area. This correlation

improved as the ratio of the molecular weights of the two compo-

nents approached unity.

In view of this, calibration curves were made of the alcohols

and water and the amounts of the other products calculated

assuming that peak area percentages were the same as weight

percentages. Subsequent analyses of known solutions checked with-

in experimental error.

b. Gaseous products

Time-of-flight mass spectra were obtained on samples of gas

from one run with 1-propanol and one with 1-butanol. In both








cases the gases had passed through two condensation traps so that

practically all materials condensable at dry ice temperatures had

been removed.

The sample from 1-butanol was determined to have 73 per cent

hydrogen with little air contamination. That from 1-propanol had

46 per cent hydrogen with considerable air contamination. Al-

though an exact determination was not made for other species, it

could be determined that at least the major portion of the re-

maining gas was carbon monoxide in each sample.

An infrared spectrum of the olefin produced from 2-butanol

was obtained to ascertain whether this was 1-butene or 2-butene.

In agreement with the results of Lundeen and Hoozer (8), the

spectrum matched a standard spectrum of 1-butene.














III. PRESENTATION OF DATA

A. Explanation of Tables


In Tables I through V are recorded the operating conditions

and results of all the runs for which reliable data were obtained.

Runs 1 through 46 with ethanol were exploratory and are not in-

cluded; a few other runs are omitted because circumstances were

such as to make the results unreliable.

In the third column the reaction temperature is that of the

catalyst bed in degrees centigrade. The reciprocal flow rate is

listed as minutes per mole of alcohol. This was calculated by

dividing the time interval of a run by the number of moles of re-

actant which left the reservoir during that time, as determined

from the volume of alcohol feed and known densities.

The mole per cent dehydration was determined as described in

section II-B-4. The per cent dehydrogenation of 2-propanol and

2-butanol was obtained by determining chromatographically the

amount of ketone produced. Per cent conversions of 1-propanol to

3-pentanol and 1-butanol to 4-heptanone were calculated from Wet

Test Meter gas volumes, assuming ideality. Numerous cross-checks

were made between chromatographic data and data obtained from gas

volumes. Examples are given in the Appendix.

Although small amounts of acetaldehyde were detected during

runs with ethanol, dehydrogenation was seldom over 1 per cent and










TABLE I

REACTION RUNS WITH ETHANOL


Per Cent
Run Reaction 1 Dehydration
No. Catalyst Temperature Flow Rate to Ethylene


373
373
383
383
383
383
400
400
400
400
400
417
417
417
417
426
426
426
427
426
426
426
426
426
426
426
426
426
426
418
418
400
400
400
418
426
426
426
426


792
505
814
435
253
441
736
410
172
371
182
441
800
155
434
1003
410
541
138
260
730
394
583
714
279
545
581
589
542
417
247
403
722
416
832
759
390
532
228










TABLE I Continued


Per Cent
Run Reaction 1 Dehydration
No. Catalyst Temperature Flow Rate to Ethylene


86
87
88
89
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
125
126
127
128
129
130


II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III


365
365
365
365
384
384
397
396
396
395
395
395
348
348
348
348
395
387
387
387
387
385
395
365
384
384
384
384
384
360
360
395
395
395
395
395
365
365
365
365
365
385


756
359
200
575
134
496
327
153
562
314
335
331
335
217
603
331
344
327
595
171
328
469
438
422
422
437
196
592
206
382
446
438
432
190
627
192
487
480
184
775
489
541


46
33
16
40
29
60
51
31
62
43
44
47
21
15
18
13
46
37
49
27
37
48
56
28
42
51
28
60
27
20
24
52
50
34
60
30
24
26
14
34
24
45










TABLE I Continued


Per Cent
Run Reaction 1 Dehydration
No. Catalyst Temperature Flow Rate to Ethylene


131
132
133
134
135
136
137
138
139
140
142
143
144
145
147
148
149
150
152
153
154
156
157
158
159
161
162
163
164
165
166
167
168
170
171
172
173
174
175
176
177
178


III
III
III
III
III
III
III
III
III
III
III
III
III
III
IV
IV
IV
IV
IV
IV
IV
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
VI
VI


385
385
385
385
385
385
385
385
385
385
385
385
385
385
385
385
385
385
385
385
385
360
360
360
360
345
345
345
345
385
385
385
385
365
365
365
365
360
360
365
325
325


189
740
499
486
503
489
466
487
510
506
517
525
487
484
446
206
308
476
435
208
364
323
214
436
348
329
443
211
336
322
205
444
319
323
443
207
324
317
331
338
322
433










TABLE I Continued


Per Cent
Run Reaction 1 Dehydration
No. Catalyst Temperature Flow Rate to Ethylene


179
180
181
182
183
184
185
186
187
188
189
190
191
192


325
325
345
345
345
345
365
365
365
365
325
345
345
345


233
312
305
215
389
307
320
421
202
315
313
363
350
342









TABLE II

REACTION RUNS WITH 2-PROPANOL


Per Cent Per Cent
Run Reaction 1 Dehydration Dehydrogenation
No. Catalyst Temperature Flow Rate to Propylene to Acetone


364
364
364
364
352
352
352
352
345
345
345
345
345
352
332
365
375
352
352
352
352
352
352
345
352
365
352
345
345
345
345
325
325
325
325
315
315
315
315


992
483
229
463
530
243
845
499
485
762
263
490
490
517
465
474
458
489
2008
473
507
539
908
539
531
585
546
522
384
715
515
475
785
347
508
515
868
385
486


3.2
3.2
2.0
3.2
2.5
1.4
2.8
2.3
1.8
2.5
1.3
2.1
2.2
2.6
1.3
3.2
3.2
2.2
3.0
2.0
2.1
2.4
2.8
1.7
2.1
2.8
2.1
5.0
4.4
5.8
4.8
3.3
4.3
2.1
3.0
2.4
2.7
1.9
2.0










TABLE II Continued


Reaction
Temperature


345
315
315
315
315
325
325
325
325
345
345
345
345
315


1
Flow Rate


Per Cent
Dehydration
to Propylene


464
479
310
655
548
492
.327
672
524
424
329
547
447
461


Per Cent
Dehydrogenation
to Acetone


4.5
1.6
0.9
2.0
1.7
1.9
1.4
2.9
1.9
2.7
1.9
3.1
2.8
1.5


Run
No.


Catalyst










TABLE III

REACTION RUNS WITH 1-PROPANOL


Per Cent Per Cent
Run Reaction 1 Dehydration Conversion
No. Catalyst Temperature Flow Rate to Propylene to 3-Pentanol


365
365
390
390
390
390
400
400
400
400
420
420
420
420
390
390
365
365
365
364
390
390
390
390
380
380
380
365
365
365
365
385
3850
385
400
400
400


489
1323
519
1226
308
495
511
1235
349
495
506
318
1171
465
456
1642
466
797
323
485
451
323
587
451
573
616
359
456
383
545
305
298
651
403
300
491
579


8
13
12
18

16
15
21
11
17
25
19

18
16
30
14
17
12
14
22
18
25
26
20
19
17
11
11
12
9
11
14
11
13
18
23










TABLE IV

REACTION RUNS WITH 1-BUTANOL


Per Cent Per Cent
Run Reaction 1 Dehydration Conversion
No. Catalyst Temperature Flow Rate to 1-Butene to 4-Heptanone


365
380
380
380
380
390
390
390
390
365
365
365
365
385
385
385
385
395
395
395
395
365


545
566
855
456
557
553
438
805
609
506
391
732
503
497
367
686
497
492
383
694
505
506









TABLE V

REACTION RUNS WITH 2-BUTANOL


Per Cent Per Cent
Run Reaction 1 Dehydration Dehydrogenation
No. Catalyst Temperature Flow Rate to 1-Butene to 2-Butanone


325
325
325
325
325
315
315
315
315
345
345
345
325
315
315
315
315
325
325
325
325
300
300
300
300
315


640
643
855
473
654
702
901
510
710
608
456
330
679
634
472
796
664
695
486
687
645
656
529
799
695
711


8.9
9.4
8.6
5.9
7.4
5.1
6.6
4.4
6.0
7.9
7.3
5.4
6.8
5.8
5.0
6.1
5.0
5.0
4.8
5.3
4.9
2.7
2.6
2.8
2.5
4.7








never over 2 per cent. In most cases this was so low that no

attempts were made to obtain an exact figure; thus, only con-

version to ethylene and water is included in Table I. Since

ethylene was not tra-ped out, the gas volume served as a good

check on the dehydration figure determined with the chromatograph.

Propene and butene from dehydration of the higher alcohols

could be trapped out and a check made on the dehydration of these

alcohols directly. Numerous material balance calculations were

made; the error was seldom more than 5 per cent and often consi-

derably less.

Trials with known solutions indicated that the dehydration

figures given are accurate to within 5 to 10 per cent of the

values listed, the determinations with ethanol and 2-propanol

being somewhat better than those with 1-propanol, 1-butanol and

2-butanol.

Although the amount of ketone produced could not be deter-

mined to the accuracy implied in Tables II and V, differences

of less than one-half per cent were easily detected. Hence,

the values are given to the nearest tenth per cent conversion.



B. Results

1. Dehydration

In Figure 2 are plotted, as examples of the type behavior

shown by all the alcohols, the results of three series of runs

with ethanol over catalyst II. An exception to this behavior was

the reaction of 1-butanol over catalyst VI; for this reaction, the

best curves through all the points including zero were straight



























0
o o o
ir\ c- nP\
oa oo 'o N

S- a% rrj H
0 0 1 4 +.
0 U rC
H > .. 4
0c x cd t
4- 0

o

W >4


\ 0o





o
\ 1
-- 0 0 .-r











Ol Or-
0\ z 0










= S
0
P 0







0













S C,






cl








lines. Although this is interesting, no attempts have been made

to place any real significance in this deviation from the norm,

since the data were collected over such a narrow range of

conditions.

Since the temperature range covered was so narrow in every

case, no attempts have been made to obtain quantitative tempera-

ture correlations, although the qualitative behavior was as ex-

pected. A few attempts were made to determine apparent kinetic

orders of the reactions. No simple order was found and the data

did not justify further attempts.

Catalyst III was accidentally overheated during the initial

activation in the reaction tube. In spite of this, a few series

of runs were made with ethanol. As anticipated, the conversions

were considerably lower than one would expect for the initially

measured surface area, although all the runs were consistent

within themselves.

Catalyst III was replaced by IV and a series of runs made at

385. During the initial warm-up period for the second series of

runs, the same instrumental malfunction which ruined catalyst III

caused the overheating of catalyst IV. A few more runs at 385

verified that the catalyst activity, at least toward ethanol,

had decreased considerably and the catalyst replaced.

The behavior of ethanol over different catalysts is illus-

trated in Figure 3. Since runs over catalyst IV were made only

at 3850, most of the plots are for 385. Runs were not made at

385 over catalyst VI, so plots at 365* over catalysts V and VI

are included for comparison of catalyst VI with the others.

















00 0
o o W o
o t\ lr\ o o u c.
oO 00 oO r trl r-
00 r LO W) D0 K
S> t M



r-i H r-4 r-I M r-4 0 4.)
HHH >>>H>








C O*H ,--H
4-) A 4-) 4-I 4) .) ..
cr1 CO c~1COO C
cd 0f t (df d (d
(t (i ti4() o


Oo

S4- 0


Q 0 o4-o
HO




0)
o 01





31 O
\ r-r CO P









+ 0 -rl

0




P4 (d



r4
\O N\ 10 L H -




\ C- *0













0 '0
PL4 p








Catalyst III is not included since its surface area at the time

runs were made over it was not known.

It is quite apparent that the dehydration of ethanol over

thoria is strongly surface area dependent, that the number of

dehydration sites for ethanol increases as the surface area in-

creases. The number of sites does not increase linearly with sur-

face area; this point will be discussed more fully later.

In Figures 4 and 5 are illustrated the relative dehydration

reactivities of the various alcohols over catalysts V and VI,

respectively. It should be pointed out that the plots for the

primary alcohols are for runs at 385 or 390 over catalyst V,

but for runs at 365 over catalyst VI. The most reliable data for

1-butanol over catalyst V are at 390 while no runs were made with

ethanol at temperatures greater than 365 over catalyst VI. The

general behavior of each alcohol over each of the two catalysts

was essentially the same as that illustrated in Figure 2, with

the exception, discussed above, of 1-butanol.

Since ethanol was the first alcohol run over each catalyst,

one or more additional runs were made with ethanol after all data-

collecting runs had been made to determine any loss of activity

during runs with the other alcohols. The activity of catalyst V

decreased only slightly during use, but the activity decrease of

catalyst VI was almost 40 per cent. This means that the last runs

over catalyst VI gave conversions lower than that characteristic

of the fresh catalyst. The sequence followed over catalyst VI

was: ethanol, 2-propanol, 2-butanol, 1-butanol, 1-propanol. Al-

though corrections for activity decrease could probably be applied




















0 L



S0 0




r. \ oJ oJ
eir\ oe
O a O o o


cMOk O


0









S0 c$-4
O H





S- 0 'O


O O *H
rl o


S0

O 0

0 wH

0 f




0 0

\ *. H ^



\ V'
















0

0*)
Pr\ 0 \ < (




O\




0 0c
*rl7


















0 0 N
L\O aO\
ri L0 P0 UL'
' r \j P'( o k-


o0 r- 0 r-
-C O 0O
Or (O H H
0 p, c PH (5
0 0 0o
c. PoL P COH





0)


x >3 0
\ \ r- O


So




00
O O
El 4



SO O H *H
0 0


S\ \r- -Z
\O O 0
\ 0



ee1\ \ o OM
S \O \ .
9\ 13 \
0 0








11
1 \ \\1



O O O O O O O O O




0
-H




0o








with reasonable reliability, the uncorrected data are plotted in

Figure 5.

In agreement with published results (13) (14) of work with

alumina, the secondary alcohols are more reactive than the corres-

ponding primary alcohols. But more important, the dehydration

reactivity of the primary alcohols decreases as the molecular

size increases, in contrast to these same results.

Perhaps it should be emphasized that 1-propanol was run

after 1-butanol over catalyst VI, but that the dehydration of 1-

propanol was still higher. Any correction to be applied for a

decrease in catalyst activity would tend to make the difference

between 1-propanol and 1-butanol even greater, but still leave

the reactivity of 1-propanol less than that of ethanol.

Per cent conversion of the three primary alcohols to water

and olefin at a reciprocal flow rate of 400 minutes per mole is

plotted as a function of surface area in Figure 6. The points

for 1-propanol and 1-butanol at a surface area of 28.9 (catalyst

VI) are corrected for loss of catalyst activity. As mentioned

previously, the per cent dehydration of ethanol increases with

an increase in surface area. However, the specific dehydration,

or the dehydration per unit area, decreases with a surface area

increase. At very low surface areas the specific dehydration is

almost constant, but at higher surface areas a doubling of the

surface area causes only a fractional increase in dehydration.

Above a certain point the dehydration of 1-propanol and 1-butanol

exhibits very little dependence on surface area. The difference

in the dehydration reactivity over catalysts V and VI was so

slight as to be barely detectable.



















(V



\ <0
00
co



o


\ \ 00



N \) cs
.0 o



%o m) H
H H 0
cO 0 0





to aq
4)
H z 40
o 1 *0

co c 0
0o Vc d4.

o o 0o 0 OO .ad
\ \\ G\ 0 o4 0


\- 0 L- 0- m

00 40 O
\O O a
\ Z 0 4) to


\ m _C 0 04
4-\1 1,
OOH
0 0 0 0 0 0 0 0 0 0
0 D1 NI






4\ -% 0-) U
4o r1) -I 4 -










,04 p o M P 0
4N




O O OO O O OO\ 4 OO H




Cd 0C


) C) 4 C O \\
Mr 4C OO 0
0 0 00 0








Figure 6 portrays even more emphatically th;n -ig Iurot 4o

5 the differences in dehydration reactivity of the primary alco-

hols. An increase in molecular size definitely corresponds to a

decrease in activity, with the greatest difference being between

ethanol and 1-propanol.

In Figure 7 are plotted dehydration reactivities of the two

secondary alcohols as a function of surface area. One of the

ethanol curves from Figure 6 is included for comparisons. It is

apparent that 2-propanol is more reactive than 2-butanol and that

an increase in surface area causes a considerable increase in

dehydration of either alcohol. There is a temptation to place

significance in the difference in shape between the 2-propanol

and the ethanol curves, but, since the runs were made at differ-

ent temperatures, any conclusions would be subject to doubt.

In summary, the dehydration experiments with the five alco-

hols permit two primary conclusions: (a) The dehydration reacti-

vities of the alcohols are in the order 2-propanol>2-butanol$

ethanol>> 1-propanol>1-butanol. (b) An increase in catalyst

surface area causes a considerable but non-linear increase in

conversion of ethanol, 2-propanol and 2-butanol, but little in-

crease in the conversion of 1-propanol and 1-butanol.

2. Dehydrogenation

The only two alcohols which exhibited simple dehydrogenation

were 2-propanol and 2-butanol, both reacting to yield ketones and

hydrogen.

Although 2-butanol was reacted over only two catalysts, the

data are sufficient to permit the observation that 2-butanol de-

hydrogenates more readily than 2-propanol over thoria.

























00





o0 < O
oo


LD 0


1V rq 0


too





4-)
10 0 ( )

O >-,

-" 0
to

So oo


0 00

o o .o 4

O- O <] <
L\ o 4 -)

I I r-
0 OH rs ) -

P4 0 P4 C
O Or1 \

O CO O O O
0 Z 0 4-)
4 C Om 0 c-



O. O *o


a PQ oo
I 4-) I Ia

110 \q











0 C0 0 0 0 0 0 0 0 0
cl



(1) 4- i 4-
0
*rl 0,


S(0 I0

00 0
0 0 4) H 0
r~i~ (X4







The measured dehydrogenation of 2-propanol was greater over

catalyst V than over either II or VI; dehydrogenation of 2-butanol

was also greater over V than over VI. This appears to suggest

that the dehydrogenation of these alcohols increases with surface

area to a point, then decreases with further surface area in-

crease. More experiments need to be done to verify this, but it

can be stated with a reasonable degree of certainty that, at the

least, dehydrogenation of these two alcohols does not increase

with surface area as rapidly as does dehydration.

3. Other reactions of 1-propanol and 1-butanol

The reaction of 1-propanol to yield 3-pentanol and of 1-

butanol to yield 4-heptanone was somewhat surprising. Although

a number of investigators (16) (17) have reacted acids over

thoria to obtain ketones and esters, no report has been found in

the literature of the production of these compounds from alcohols.

The net reactions producing these products were determined

to be

OH
2 CH3-CH2-CH2-OH -- CH3-CH2-9-CH2-CH3 + CO +- 2H2
H


2CH3-CH2-CH2-CH20H -- CH3-CH2-CH2-K-CH2-CH2-CH3 +- CO + 3H2



An effort was made to detect 3-pentanone in 1-propanol re-

action products and 4-heptanol in 1-butanol reaction products.

Although there was a small amount of high boiling-material in the

1-butanol products which may have been 4-heptanol (it appeared

to be an alcohol and the boiling point was in the right vicinity),





40


no positive identification was made. No trace of 3-pentanone

was found.

In Figure 8 are plotted representative results of the pro-

duction of 3-pentanol from 1-propanol and 4-heptanone from 1-

butanol. Catalyst V was more effective in the production of

these compounds than catalyst VI, the same relationship as that

found for the dehydrogenation of the secondary alcohols.

































Per Cent
Conversion


O 1-Propanol, 390
C 1-Butanol, 390
* 1-Propanol, 3850
8 1-Butanol, 3850


10




30 Catalyst VI

20

10


0 1 2 3 4 5 6 7 8 9
1/Flow Rate, minutes/mole




Figure 8. Conversion of 1-propanol to 3-pentanol and
1-butanol to 4-heptanone as a function of
reciprocal flow rate














IV. DISCUSSION OF RESULTS AND CONCLUSIONS


In Chapter I it was pointed out that thoria effects both de-

hydration and dehydrogenation of alcohols. The evidence seems

strong that these two reactions take place on different sites.

This appears to be supported by the work done with 2-propanol and

2-butanol in the present investigations. The conversion to water

and olefin was shown to increase with an increase in surface area;

after an initial increase with surface area, per cent dehydrogen-

ation appeared to decrease with an increase in surface area, a

trend opposite to that of dehydration.

A point made in Chapter III was that per cent dehydration of

ethanol, as well as the other alcohols, does not increase linearly

with surface area; that is, a doubling of surface area does not

cause a doubling of per cent dehydration. A close examination of

Figure 6 will show that a linear relationship apparently does

exist at very low surface areas, but that this relationship fails

as the surface area increases. This means either that the number

of sites per square centimeter of surface chemically able to pro-

mote dehydration decreases as the surface area increases, the

site density decrease being faster than the surface area increase,

or that the fraction of the surface accessible to nitrogen mole-

cules but not to ethanol molecules increases with surface area.

The latter assumption appears to be more reasonable.








If indeed a large part of the surface is accessible to nitro-

gen molecules but not to ethanol, this means that there must exist

many pores, cracks or crevices of molecular size. It will be re-

called that this was one assumption made by Schwab (5) in pro-

posing that dehydration occurs in pores while dehydrogenation

occurs on flat surface regions.

The relative dehydration reactivities of the primary alcohols

over thoria seems to support Schwab's proposal. Many pores

(cracks, crevices, etc.), accessible to ethanol would not be so to

the larger molecules, or at least a special configuration of the

alcohol molecule would be required for it to enter the pore. The

larger the molecule the less likely it would be to get into this

configuration and the less likely dehydration would be to occur.

Since the reactivities of the secondary alcohols are in the

same order over alumina and thoria, little information is gained

from their reactivities relative to each other or to the other

alcohols as to the mechanism of dehydration over thoria. It will

be assumed that these relative reactivities are primarily a reflec-

tion of the chemical nature of the alcohols themselves. Any pro-

posed mechanism, however, would have to allow for this relative

reactivity.

It has been mentioned previously that thoria is quite speci-

fic in a number of reactions, differing from alumina in the non-

formation of ethers, the specificity of the production of 1-olefins

from secondary 2-alcohols and the production of methyl vinyl

carbinol instead of methyl ethyl ketone from 2, 3-butanediol. All

of this evidence points strongly toward a concerted mechanism for








the dehydration of alcohols over thoria; an ionic mechanism, such

as is usually proposed for alumina, would most likely not permit

such specificity.

Assuming that, indeed, dehydration does occur in pores by a

concerted mechanism one is confronted by two questions: Why do the

molecules enter small pores to dehydrate when flat surface is more

accessible? And why a concerted mechanism over thoria in contrast

to that over alumina?

The following is suggested as a possible mechanism which

seems to correlate most of the evidence obtained in the present

investigation as well as that of previous investigators.

Most mechanisms for dehydration over alumina (10) involve

preliminary removal of the hydroxyl group and then removal of a

proton from the resulting carbonium ion. It is suggested that

thoria, being less acidic than alumina, is incapable of promoting

such a mechanism. The elements of water are eliminated from the

alcohol only if the olefin is formed at the same time. The inter-

mediate, as illustrated below, would have half-formed bonds from

the catalyst to both the hydroxyl group and the hydrogen to be

lost, a half-formed 7T bond between two alcohol carbons and half-

broken carbon-oxygen and carbon-hydrogen bonds.

S S
S S
S H S

S HH 0--S
S S
S .C C. S
S S
S---H H H S
S S
S S
S S








There appears to be no opportunity for ether formation. In

2, 3-butanediol dehydration, water is removed from one end of the

molecule while the double bond is being formed, with the other

end of the molecule unaffected, leaving methyl vinyl carbinol.

But why does the dehydration have to be in a pore? One

possible explanation is that the stereochemistry of the interme-

diate is such as to require a pore for the elimination of water.

The original idea of Schwab was that the hydroxyl group was

attached to one side of a pore while the hydrogen to be lost was

attached to the other side. An analysis using molecular models

shows that if the oxygen of ethanol is attached to one side of

the pore and a hydrogen from the methyl group is attached to the

other side in such a way that a straight line can be drawn from

the center of the attached hydrogen through the center of the

carbon-carbon bond to the center of the oxygen (the hydroxyl group

is staggered between two methyl hydrogens), then the centers of

the other four hydrogens and the center of the carbon-carbon bond

all lie in a plane. This situation seems quite favorable for the

formation of ethylene.

The larger alcohols can attach in the same way in the same

size pore, but only if the molecule has a rather special con-

figuration. Thus ethanol should be considerably more reactive

than the other primary alcohols, and the larger the alcohol mole-

cule the less likely it is to arrange itself in just the correct

configuration.

The specificity of production of 1-butene from 2-butanol

and the reactivities of the secondary alcohols relative to the








primary alcohols can be explained by allowing two more assumptions,

that the oxygen is bonded more readily to the catalyst surface

than any of the hydrogens and that there are many more adsorptions

than there are dehydrations.

In 2-butanol there are three hydrogens on carbon number one,

and two on carbon number three. For a concerted mechanism this

reduces the chances for the production of 2-butene. Furthermore,

an analysis with molecular models shows that only one of these

would be at all likely to attach so as to be eliminated in de-

hydration; attachment of the other in a straight-line relationship

would require such a bending of the molecule that it could no

longer fit in the pore. If it is assumed that the oxygen is

attached to the surface of the pore first, one can picture the

molecule twisting itself until a hydrogen is attached in just the

right position to allow elimination of water. The methyl group

alpha to the hydroxyl carbon could spin quite freely while the

rest of the molecule would be restricted in its movement by the

sides of the pore. The probability is quite high that one of the

methyl hydrogens would attach in the correct position before the

one eligible methylene hydrogen. Thus, the probability of the

production of 2-butene is extremely low.

In assuming that many adsorptions do not result in dehydra-

tions, one allows the chemical differences of the secondary and

primary alcohols to explain the differences in reactivities.

With a free-spinning methyl group next to the hydroxyl carbon, the

secondary 2-alcohols can attach in the straight-line relationship

almost as readily as ethanol, and a larger percentage of these

adsorptions result in elimination of water.








Dehydrogenation presumably occurs in a manner similar to that

over ordinary dehydrogenation catalysts, by a mechanism reminiscent

of Balandin's hypothesis (1 ). The two hydrogens to be eliminated

are adsorbed on the surface adjacent to each other. As the

carbon-hydrogen and oxygen-hydrogen bonds are broken, the hydrogen-

hydrogen bond is formed. The decrease in dehydrogenation in going

from catalyst V to VI can be explained either by a loss in flat

surface available for dehydrogenation, or by the fact that a

higher degree of dehydration left less alcohol for dehydrogena-

tion. Perhaps a combination of these factors was involved.

No fully adequate explanation has been found for the pro-

duction of 3-pentanol from 1-propanol and 4-heptanone from 1-

butanol. Previous investigations (16) (17) with acids over thoria

have indicated that apparently salts are formed with the catalyst

surface, then carbon dioxide is eliminated between two salt mole-

cules. In the same manner, one mechanism for the production of

4-heptanone from 1-butanol would involve the adsorption with

elimination of hydrogen of two molecules of propanol, formed from

dehydrogenation of the alcohol, on adjacent sites. The carbonyl

carbon would be attached to an oxygen of the catalyst, forming

a sort of salt molecule. The two molecules would then eliminate

carbon monoxide between them, forming 4-heptanone. Elimination

of carbon monoxide from adjacently-adsorbed 1-propanol and pro-

panal could form 3-pentanol. The existence of small amounts of

the appropriate aldehyde in products from these reactions seems

to favor a mechanism involving initial production of the aldehyde.

However, additional evidence is needed before any real conclusions

can be drawn.








From Figure 8 it can be seen that catalyst V was more active

in the production of 3-pentanol and 4-heptanone than catalyst VI.

This difference in activity could possibly reflect a loss in acti-

vity of catalyst VI with use; it is recalled that the dehydration

activity of catalyst VI decreased with use. However, the dehy-

dration of the secondary alcohols was greater over VI than over

V, while the dehydrogenation was less. This suggests that perhaps

dehydrogenation and production of 3-pentanol and 4-heptanone occur

on the same type of site. Since the bimolecular reaction almost

certainly occurs on a relatively flat surface, the results appear

to give some support to the proposal that dehydrogenation occurs

on flat surfaces.

No mention has been made of the type of bonding occurring

between adsorbed alcohol molecules and the catalyst surface. Al-

though this point is obviously important, little is actually

known of this and any discussion, although interesting, is little

more than speculation.

It appears likely that the hydrogen to be lost is bonded to

an oxygen on the surface. This oxygen could be part of the thorium

oxide lattice or part of a water molecule bonded to the surface.

In dealing with thorium, one should consider the possibility of

complex formation. The oxygen from an alcohol molecule could

complex with a thorium atom on the surface upon dehydration.

Probably more likely, however, is that the removal of the hydroxyl

group is effected through water already on the surface, possibly

completed with the thorium.








Another point to consider is that active sites could possibly

be the result of chemical abnormalities of some type. These

might be defects common to the crystal as a whole or chemical

peculiarities produced by the very existence of a surface.

Colored samples of thoria (usually blue or blue-green) which

have been prepared suggest the possibility of the existence of

unpaired electrons. Thorium oxide catalysts prepared by calcining

the oxalate are usually more catalytically active than those pre-

pared from the hydroxide. Although a higher surface area can

explain the higher activity to some extent, one should consider

the distinct possibility of lattice defects due to incomplete

removal of carbon and the existence of oxidation states of thorium

other than four. Further work is planned in this laboratory in an

attempt to obtain information as to the existence and nature of

these chemical abnormalities in thoria.

Complicating the whole area of heterogeneous catalysis is

the fact that the environment on the catalyst surface during

catalytic reactions is quite different from that during most

other investigations of the nature of the surface. Thus, more

catalytic reaction investigations are desirable, in addition to

investigations as to the nature of the fresh surface.

At the outset of the investigations described herein, the

intention was to obtain kinetic data from the reactions of one or

two alcohols over thoria and to determine the effect of catalyst

surface area on these reactions. When it was found that a decrease

in catalyst activity prevented the reproducibility of results to

the exactitude necessary for reliable determination of kinetic








constants, it was decided to perform the investigations as to the

effects of the variation of alcohol molecular size. These have

proved fruitful in leading to the proposal as to the mechanism

of alcohol dehydration over thoria and in the discovery of a type

of reaction heretofore unreported in the literature. The results

indicate a need for investigations with still other alcohols. If

the problem of catalytic activity decrease can be solved, exact

kinetic investigations should prove fruitful.

In the present investigations it has been assumed that a

difference in activity from one catalyst to the other was due to

a change in the area of the surface available for reaction. It

is entirely possible, however, that some other property of the

catalysts, which is proportional to the surface area for the

catalysts investigated, is responsible for the change in catalytic

activity. Correlations should be attempted between catalytic

activity and other properties of the catalyst, such as lattice

defects, after techniques for the determination of these properties

are developed.

Quite recent investigations (7 ) have resulted in the pre-

paration of catalysts with surface areas around 100 square meters

per gram, very high for thorium oxide. The comparison of reactions

over these catalysts to those over low surface area catalysts

should prove quite interesting.














V. SUMMARY


Six thorium oxide catalysts have been prepared by the pre-

cipitation of thorium hydroxide from thorium nitrate solution

with ammonia. The hydroxide was dried, pelleted and heated under

vacuum at 600* for four to twenty-nine hours. The surface area

of each of the catalysts has been measured.

The products of the reactions of ethanol, 1-propanol, 2-

propanol, 1-butanol and 2-butanol over thoria have been separated

and identified. Each of the alcohols dehydrated to produce water

and the appropriate olefin; 2-butanol produced 1-butene but no

2-butene. 2-Propanol and 2-butanol lost hydrogen to produce ace-

tone and methyl ethyl ketone, respectively. 1-Propanol and 1-

butanol both reacted to lose carbon monoxide and hydrogen; in the

process 3-pentanol was formed from 1-propanol and 4-heptanone

from 1-butanol.

The tendency to dehydrate has been found to be 2-propanol>

2-butanol ;ethanol>> 1-propanol >1-butanol.

An increase in catalyst surface area caused a considerable

but non-linear increase in the dehydration of ethanol, 2-propanol

and 2-butanol, but little increase in the dehydration of 1-

propanol and 1-butanol.

These results have led to a proposal, similar to that of

Schwab's that dehydration occurs by the concerted elimination of





52

water in pores. The alcohol molecule is proposed to attach to

the pore walls in such a way as to make favorable formation of

the olefin. The mechanism permits the explanation of four

peculiarities of thoria as a dehydration catalyst: (a) the non-

formation of ethers, (b) the preferential formation of 1-olefins

from secondary 2-alcohols, (c) the production of methyl vinyl

carbinol from 2, 3-butanediol and (d) the greater dehydration

reactivities of alcohols of smaller size.














APPENDIX

Some Details of the Analyses of Reaction Products

A. Qualitative


Water, a reaction product of all the alcohols, was identified

by chromatographic retention times under various conditions (or

by increase in peak size with addition of water) and by the

characteristic peak shape on a chromatogram.

Schmidt (12) identified acetone and water as the two con-

densable 2-propanol products from reactions under conditions al-

most identical to those in this investigation. He distilled the

products, obtained infrared spectra, and prepared derivatives of

the component suspected of being acetone. In view of this, the

only identification of acetone felt necessary was the verification

of retention times with that of the known. Methyl ethyl ketone, a

product of 2-butanol in a reaction analogous to that of 2-propanol,

was also identified from retention times.

The identification of compounds by chromatographic retention

times leaves some room for doubt, in that two different components

could possibly have the same retention times and produce the same

shape peak. However, the column packing material was designed to

separate components which would be expected from alcohol reactions.

In the absence of evidence to the contrary, it was assumed that

only simple dehydration and dehydrogenation were occurring with

ethanol, 2-propanol, and 2-butanol.

53








The identification of 3-pentanol and 4-heptanone, the only

products other than water from 1-propanol and 1-butanol, was done

as described in section III-A.



B. Quantitative


1. Determination of ner cent dehydration of ethanol

Samples of known solutions of ethanol and water, prepared by

weighing the two components, were put through the chromatograph.

The areas of the peaks on the resulting chromatogram were deter-

mined with a planimeter. Plots were then made of the per cent

contributed by water of the total area of the two peaks against

the mole per cent water in the mixture, and of the peak area per

cent against weight per cent.

Samples of reaction mixtures were put through the chromato-

graph and peak area percentages determined. From the calibration

curves the mole per cent water in the mixture could be determined.

Since one mole of ethanol will produce one mole of water upon

dehydration, the mole per cent water in the mixture could be taken

as the mole per cent conversion.

Periodic mole balance calculations were made, but, since the

method described above gave consistent results, the actual number

of moles of product was not calculated for each run. The follow-

ing data give examples of mole balance calculations of two runs

with ethanol.

Run No. 156 Run No. 159

(a) 0.261 0.247
(b) 0.103 0.088








Run No. 156 Run No. 159

(c) Moles unreacted alcohol 0.166 0.167
(d) Water plus unreacted alcohol 0.269 0.255
(e) Grams liquid product 9.50 9.27
(f) Moles gas 0.082 0.075
(g) Conversion (curves) 38.5% 36.5%
(h) Conversion (Liquid) 39.5% 35.6%
(i) Conversion (gas) 31.4% 30.4%

The number of moles of feed was calculated by multiplying the

volume of alcohol feed times the known density at the temperature

of the reservoir and dividing by the molecular weight. The number

of moles of water and unreacted alcohol ( (a) and (c) above) were

determined by multiplying the grams of liquid product by the

appropriate weight fraction as obtained from peak area percentages

and calibration curves. Trials with nitrogen, at pressure differ-

entials the same as those experienced during reaction, indicated

that the true volume of gas passing through the Wet Test Meter

was very nearly three-fourths of the reading indicated. This

factor was used when calculating the number of moles of gas pro-

duced.

The mole per cent conversion (g) was determined from the

calibration curves as described above. The conversion (h) was

calculated by dividing the number of moles of water by the moles

of feed, and (i) was determined by dividing the number of moles

of gas by the moles of feed.

It is apparent that the simpler determination of mole per

cent of water using the calibration curves directly is as satis-

factory as calculating the number of moles of water produced and

the conversion from this.








Since some of the gas, assumed to be all ethylene, dissolved

in the liquid products, the number of moles of gas was always less

than that of water, but close enough to serve as a convenient

check.

2. Calculation of three-cornon- n.t mixtures; 2-propanol deter-
minations

The calculation of the mole percentages of acetone, 2-propanol

and water in a mixture of these will be considered as an example

of the method used to determine relative amounts of components

in a reaction mixture.

Calibration curves (area per cent vs. mole per cent and area

per cent vs. weight per cent) were made from determinations of

known solutions of water and 2-propanol. The fraction of water in

the mixture containing acetone was obtained from these curves as

if the non-water peak area on the chromatogram was all alcohol.

Since peak areas are proportional to weight, this will hold quite

well in any mixture if the component other than alcohol and water

has a molecular weight close to that of the alcohol, or if this

component is present in small amounts.

For convenience the following symbols will be adopted:

2P = 2-propanol m = number of moles of ...
Ac = acetone wt = weight of ...
H20 = water mw = molecular weight of ...

Assuming area ratios to be equal to weight ratios:



m Ac wt Ac mw 2P area % Ac mw 2P (i)
m 2P ~ wt2P x mw Ac area % 2P x mw Ac









area % Ac mw 2P
m Ac = are% x x m 2P = K x m 2P
area sa 2P mw Ac



Assuming, for convenience, a total of 100 moles:


m Ac +- m 2P


S100 m H20


(iii)


Substituting m Ac from (ii) and m H20 obtained from the

calibration curves, and solving for m 2P, one obtains


100 m H20
m 2P = 1 + K


(iv)


which is then also equal to the mole per cent 2-propanol in the

mixture.

The results of the application of this method to one known

mixture are given below:


Per cent determined
by weight


Acetone
2-Propanol
Water


weight
per cent

8.1
81.4
10.5


mole
per cent

6.6
65.3
28.0


Calculated
per cent

area mole
per cent per cent

7.5 6.4
81.1 66.6
11.4 27.0


Although the figures are given to the nearest tenth per cent,

the error in determination of the peak areas prevents such pre-

cision in most measurements. The calculations listed are probably

somewhat more accurate than those for an average reaction product

determination.


(ii)







An example of the results of mole balance checks of 2-propanol

products is given below; the calculations are for run number 37.

(a) Moles feed 0.134
(b) Moles water 0.046
(c) Moles acetone 0.003
(d) Moles unreacted alcohol 0.095
(e) Grars liquid product 6.31
(f) Moles gas 0.009
(g) Dehydration (curves) 38.7%
(h) Dehydration (liqui_ 34.3%
(i) Dehydrogenation (calculated) 3.0%
(j) Dehydrogenation (liquid) 2.3%
(k) Dehydrogenation (gas) 6.7%

The number of moles of feed was determined in the same way

as for ethanol. The moles of water, acetone and unreacted alco-

hol in the liquid products were calculated by multiplying the

weight of these products by the appropriate weight fraction,

determined as described above.

The dehydration figure (g) and dehydrogenation figure (i),

those listed in Table II, were determined as described above.

The figures (h) and (k) were obtained by dividing the calculated

number of moles of water and of acetone in the liquid products

by the number of moles of feed.

The dehydrogenation figure (k) was obtained by dividing the

number of moles of gas by the moles of feed, assuming all of the

gas to be hydrogen. Because of its high volatility, even at dry

ice temperatures, some propylene most likely was not trapped out.

This probably accounts for the high figure for (k) compared to

(i) and (j).








3. 1-Propanol determinations

The results of a mole balance check of run number 39 with

1-propanol are as follows:

(a) Moles feed 0.139
(b) Grams liquid p-oduct 6.68
(c) holes w-ter 0.026
(d) Moles 3-pent-nol 0.009
(e) Moles unreacted alcohol 0.095
(f) Moles gas 0.049
(g) Dehydration (curves) 23%
(h) Dehydration (liquid) 19%
(i) Conversion to 3-pentanol (calculated) 18%
(j) Conversion to 3-pcntanol (liquid) 13%
(k) Conversion to 3-pe:.tanol (gas) 23%

The number of moles of feed, water and 3-pentanol were deter-

mined in the same manner as for the three components in 2-propanol

liquid products.

The dehydration figure (g) and conversion figure (i) were

calculated by the method described above for a three-component

mixture. The dehydration figure (h) was calculated by dividing

(c) by (a). Since two moles of alcohol were needed to produce

one of 3-pentanol, the figure (j) was obtained by dividing (d)

by (a) and doubling.

The conversion figure (k) was obtained by multiplying the

number of moles of gas by two-thirds, since three moles of gas

(two of hydrogen and one of carbon monoxide) were produced for

every two moles of 1-propanol converting, and dividing by the

moles of feed. Since an unknown amount of propylene was included

in the gas volume measured, this figure is probably somewhat too

high, as comparisons with (i) and (j) indicate. This figure is

listed in Table III, however, since the data calculated this way

were much more precise and probably as accurate as that calculated








from chromatograms. The water peak on the chromatogram tailed

into the 3-pentanol peak slightly. This affected the water peak

little, but made the determination of the base line of the 3-

pentanol peak somewhat uncertain. Since this peak was low and

long, a shift in baseline caused a considerable change in peak

area.

4. 2-Butanol determinations

The ease of separation of 2-butanol products affords an

opportunity to cross-check the analytical methods.

Since 1-butene has a boiling point of -5*, practically all

of it was trapped out. The volume of that trapped in the second

collection flask was measured at its boiling point, and the

number of moles calculated from the known density at this tempera-

ture and the molecular weight. The weight dissolved in the

liquid products was determined from the peak areas of a chromato-

gram, allowing the same assumptions as with a 2-propanol mixture.

This afforded a direct check on dehydration as determined in the

manner previously described.

Since practically all products were trapped out as liquids

except hydrogen, the gas volume afforded a direct check on de-

hydrogenation as determined by obtaining the amount of methyl

ethyl ketone produced.

The results of a mole balance check of run number 26 are

as follows:

(a) Moles feed 0.119
(b) Grams liquid product 8.37
(c) Moles water 0.016
(d) Moles methyl ethyl ketone 0.002
(e) Moles unreacted alcohol 0.107








(f) Moles dissolved 1-butene 0.013
(g) Moles trapped 1-butene 0.003
(h) Moles gas 0.002
(i) Dehydration (curves) 15%
(j) Dehydration (liquid) 13%
(k) Dehydration (1-butene) 14%
(1) Dehydrogenation (calculated) 2.5
(m) Dehydrogenation (liquid) 1.8
(n) Dehydrogenation (gas) 1.8

The determination of figures (a), (c), (d), (e) and (f) was

done as described previously, from chromatographic peak areas.

The dehydration figure (i) was determined from peak areas

and calibration curves, as described previously. The dehydration

figure (j) was obtained by dividing the number of moles of water

by the moles of feed, and the figure (k) was obtained by dividing

the number of moles of 1-butene by the moles of feed.

Dehydrogenation figure (1) was obtained by determining the

mole per cent methyl ethyl ketone in the same way that acetone

was determined in 2-propanol mixtures. The figures (m) and (n)

were obtained by dividing the moles of methyl ethyl ketone and

hydrogen, respectively, by the moles of feed.

5. 1-Butanol determinations

The results of a mole balance check of run number 18 are as

follows:

(a) Moles Feed 0.128
(b) Grams liquid product 8.37
(c) Moles water 0.019
(d) Moles 4-heptanone 0.009
(e) Moles unreacted alcohol 0.107
(f) Moles dissolved 1-butene 0.009
(g) Moles trapped 1-butene 0.009
(h) Moles gas 0.037
(i) Dehydration (curves) 18%
(j) Dehydration (liquid) 15%
(k) Dehydration (1-butene) 14%
(1) Conversion to 4-heptanone (calculated) 17%
(m) Conversion to 4-heptanone (liquid) 14%
(n) Conversion to 4-heptanone (gas) 14%







The determinations of (a), (c), (d), (e) and (f) were done

as described previously.

The dehydration per cent (i) was determined from peak areas

and calibration curves. The dehydration per cent (j) was obtained

by dividing (c) by (a), and the per cent (k) was obtained by

dividing the total number of moles of 1-butene by (a).

The conversion to 4-heptanone (1) was calculated by deter-

mining the mole per cent 4-heptanone in the same way that acetone

was determined in 2-propanol mixtures and doubling this percent-

age. The conversion (m) was obtained by dividing (d) by (a) and

doubling. Since four moles of gas (one of carbon monoxide and

three of hydrogen) were produced for every two moles of 1-butanol

converting to 4-heptanone, (h) was halved and divided by (a) to

obtain (n).












BIBLIOGRAPHY


1. B. M. V. Trapnell, "Balandin's Contribution to Heterogeneous
Catalysis" in Advances in Catalysis III, Academic Press,
Inc., New York, 1951.

2. P. Sabatier and A. Maihle, Compt. Rend., 147, 106 (1908).

3. H. Adkins, J. Am. Chem. Soc., 44, 2175 (1922).

4. G. I. Hoover and E. K. Rideal, J. Am. Chem. Soc., 42, 104,
116 (1927).

5. G. M. Schwab and E. Schwab-Agallidis, J. Am. Chem. Soc., 71,
1806 (1949).

6. M. E. Winfield, J. Council Sci. Ind. Research, 18, 412
(1945).

7. B. H. Davis, Personal Communication.

8. A. J. Lundeen and R. V. Hoozer, J. Am. Chem. Soc., 85, 2180
(1963).

9. H. E. Pines and W. O. Haag, J. Am. Chem. Soc., 83, 2847
(1961).

10. M. E. Winfield, "Catalytic Dehydration and Hydration" in
Catalysis Vol. VII, Reinhold Publishing Corporation, New
York, 1960.

11. J. N. Pearce and M. J. Rice, J. Phys. Chem., 11, 692 (1929).

12. P. G. Schmidt, Doctoral Dissertation, University of Florida
(1959).

13. A. K. Bork and A. A. Tolstopyatowa, Acta. Physicochim,
U.R.S.S., 8, 603 (1938).

14. L. K. Freidlin and A. M. Levit, Isvest. Akad. Nauk S.S.S.R.,
Otdel Khim Nauk, 163 (1952).

15. C. G. Moreland, M. S. Thesis, University of Florida (1962).

16. R. I. Reed, J. Chem. Soc., 4423-6 (1955).





64

17. J. C. Kariacose and J. C. Jungers, Bull. Soc. Chim. Belges,
64, 502-32 (1955).

18. S. Brunauer, P. H. Emmett, and E. Tellor, J. Am. Chem. Soc.,
60, 309 (1938).














BIOGRAPHICAL SKETCH


John Wallis Legg was born September 20, 1936, in Minter City,

Mississippi. After attending the public schools in Drew, Missi-

ssippi, he entered Mississippi College, Clinton, Mississippi, in

September, 1954, and graduated in May, 1958, receiving the degree

of Bachelor of Science, with special distinction.

In September, 1958, Mr. Legg entered the Graduate School of

the University of Florida. He held graduate and research

assistantships until he received the degree of Master of Science

in August, 1960.

From September, 1960, until May, 1962, he held the position

of Assistant Professor of Chemistry at Mississippi College,

Clinton, Mississippi.

In June, 1962, Mr. Legg re-entered the University of Florida

Graduate School and has held a research assistantship since that

time.

He is a member of the American Chemical Society, Alpha Chi

Sigma and Omicron Delta Kappa.

He is married and has two children.













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 Arts and Sciences and to the Graduate

Council, and was approved as partial fulfillment of the require-

ments for the degree of Doctor of Philosophy.


December 19, 1964



Dean, Colligp f Ats and Sciences



Dean, Graduate School



Supervisory Committee:



Chairman \ \



~M7/j




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs