gin:

Y

 

v
m?“
3.19

. .o «Vs- 3...
n\.

 

 

2.-.Ob...lv \ rr
» .. .. 5.3-.
at. .4» luv (it...
.‘uch olI-ri213-

 

    

 

 
 

 

      

  

. . o: .. J)” . {51.11.}. . ...t. ... LT! “mlicfis. ... 5 fl h a.
. .. . . . .. . .. , u . ..V. 1, z: , .. .. 7.: ‘ a 1
.‘..ut...‘+...§...$ 73.62. Lain? ..mxéaeé. .m..m....uwu,fi..,..huu .4 ......uw.“ .H

  

 

 

 

 

Immmiiiiiiiiiiiiiiiiiflm

31293017

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

thesis entitled

CONDENSED PHASE CONVERSION OF LACTIC ACID TO
2 , 3-PENTANEDIONE

presented by

Rajesh Baskaran

has been accepted towards fulfillment
of the requirements for

 

 

M.S. degree in Chemical Engineering

5W QIMQM

Major professor

Date 4/2 5/76

0-7 639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE I DATE DUE DATE DUE

 

m: 2 5 25:1»

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/98 :ICIRCIDateDmpGS-ptu

CONDENSED PHASE CONVERSION OF LACTIC ACID TO 2,3-PENTANEDIONE
By

Rajesh Baskaran

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Chemical Engineering

1998

C OXDE

Ti
the homo!
COHdCflSCi
meal we
laCtaIe Wa
CValuaIed.
r€21Ci0l CQL
PTOducts. l
Peflod of ti

the Vapor P

Th
Itrr‘peTElIUre
{”0qu He]
from the Opt
“the PTOdu.
we obtained
ConverSI-On o
HmMWm;

ABSTRACT
CONDENSED PHASE CONVERSION OF LACTIC ACID TO 2,3-PENTANEDIONE
By
Rajesh Baskaran

The continuous conversion of biomass derived lactic acid to 2,3-pentanedione in
the homogenous condensed phase is achieved using alkali metal salt catalysts. Preliminary
condensed phase studies were conducted in batch and semi-batch reactors. Various alkali
metal catalysts were studied at different temperatures and catalyst concentrations. Cesium
lactate was found to be the most effective catalyst. Of the several reactor configurations
evaluated, the most eflicient operation was achieved in a 50ml continuous stirred tank
reactor equipped with an overflow tube for level control and a second port for vapor
products. It was possible to maintain steady state conditions in the reactor for an extended
period of time by operating the reactor at much lower temperatures (ISO-250°C) than in
the vapor phase system. Analysis of condensable vapor products was done using GC.

Three variables were identified as key parameters afi‘ecting reactor performance -

temperature, catalyst concentration in the continuous feed, and feed flow rate. The
product yields of 2,3-pentanedione and acrylic acid, conversion, and selectivity obtained
from the optimization experiments were fit to a quadratic model to generate contour plots
of the product yields, conversion and selectivity. Highest 2,3-pentanedione yield of 20%
was obtained at 240°C. Acrylic acid yields up to 15% were observed at high flow rates.
Conversion of lactic acid increases with temperature and catalyst concentration, the
selectivity to 2,3-pentanedione is highest at low flow rates. A kinetic model to describe

the system was developed.

Thi5 work is .

and my broth

This work is dedicated to my parents, Jayapal and Pn'thi Baskaran,

and my brother Sanjay.

iii

l M
constant en.
helpful dun
Dr. Scranto:

Thar

for being so

ACKNOWLEDGEMENTS

I would like to thank Dr. Dennis Miller, my advisor, for his valuable guidance and
constant encouragement during the course of this work. Dr. James Jackson was very
helpful during numerous group meetings. I would also like to express my appreciation to

Dr. Scranton for serving on my graduate committee.

Thanks to the members of Miller group - Man, Dushyant, Shubham, and Kirthi

for being so helpful during my graduate work.

iv

List of Tab
List of Figt

lntroductior
firm

1.] Liter

1.1.1

1.1.2

” Yap.
1'3 Mec
1'4 Ratit
1.5 Rese

TABLE OF CONTENTS

List of Tables viii

List of Figures x
Introduction 1
Chapter 1 BACKGROUND 2
1.1 Literature Review 2
1.1.1 Lactic Acid 2

1.1.2 Reactions of Lactic Acid 6

1.1.2.1 Thermal Degradation 6

1.1.2.2 Dehydration of Acrylic Acid 6

1.1.2.3 Reduction to Propanoic Acid 7

1.1.2.4 Condensation Reactions 7

1.1.3 2,3-Pentanedione 10

1.1.4 Acrylic Acid ' 12

1.1.5 Propanoic Acid 13

1.2 Vapor Phase Conversion of Lactic Acid 14
1.3 Mechanism 17
1.4 Rationale for PrOposed Work 19
1.5 Research Objectives 22

Chapter 2 LACTIC ACID CONVERSION IN A CONTINUOUS REACTOR 23

2.1 Introduction 23
2.2 Methods 23
2.2.1 Materials 23
2.2.2 Catalyst Preparation 25
2.2.3 Reactor Description 28
2.2.4 Experimental Procedure 29

to
DJ
29

IJ
'JJ
'r-m

3.1 Optir
3.1.1
3.1.2

2.3

2.2.5 Reaction Conditions
2.2.6 Product Analysis
Results and Discussions
2.3.1 Development of Reactor Configuration for Continuous Operation
2.3.2 Effect of Temperature and Catalyst Concentration in Feed
2.3.3 Effect of Feed Rate
2.3.4 Extended Reaction
2.3.5 Material Balance
2.3.5.1 Mass Balance
2.3.5.2 Carbon Balance
2.3.5.3 Hydrogen Balance
2.3.5.4 Comparison of Product Analysis
2.3.6 CO and C02 Production

Chapter 3 OPTIMIZATION AND KINETIC ANALYSIS

3.1

3.2

Optimization of Operating Condition
3.1.1 Box-Benkhen Design of Experiments
3.1.2 Experimental Method
3.1.3 Results and Discussion
3.1.3.1 2,3-Pentanedione Yield
3.1.3.2 Acrylic Acid Yield
3.1.3.3 Conversion of Lactic Acid
3.1.3.4 Selectivity to 2,3-Pentanedione
Kinetic Analysis
3.2.] Continuous Stirred Tank Reactor
3.2.2 Reaction System
3.2.3 Method
3.2.4 Lactic Acid Concentration
3.2.5 Results

3.2.6 Discussion

vi

31
32
35
35
37
40
41
45
45
45
46
47
49

51
51
51
53
55
58
61
63
65
67
67
69
7O
71
75
81

Efifjfgr LIV-g xmretrunmwug

Chapter
4.1 St

4.2 R

Appendi
List of II

Chapter 4 SUMMARY AND RECOMMENDATIONS 85

4.1 Summary 85
4.2 Recommendations 86
Appendices 87

List of References 100

vii

Table 1..
Table 1.:
Table 2.1

Table 2.2

Table 2.3:
Table 2.4:

Table 2.5:
Table 2.6:
Table 2.7:
Table 2.8:
Table 2.9:

Table 2.10:

Table 3.1:
Table 3.2:
Tab1e33;
Table 3.4:

Table 3.5:

Table 3.6:
Table 3].

Table 3.8;
Table 39:
Tibia 310:
Table 3.1 l:

R
St
Rt
Es

Table 1.1:
Table 1.2:
Table 2.1:
Table 2.2:

Table 2.3:
Table 2.4:

Table 2.5:
Table 2.6:
Table 2.7:
Table 2.8:
Table 2.9:
Table 2.10:

Table 3.1:
Table 3.2:
Table 3.3:
Table 3.4:

Table 3.5:

Table 3.6:
Table 3.7:

Table 3.8:
Table 3.9:
Table 3.10:
Table 3.11:

LIST OF TABLES

Physical Properties of Aqueous Lactic Acid

Physical Properties of 2,3-Pentanedione

Chemicals Used as Gas Chromatography Standards

Retention Time and Response Factor for Product Analysis by Gas
Chromatography

Summary of Results for Experiment 50C25, 50C26, and 50C27
Variation of 2,3-Pentanedione Concentration in Condensable
Vapor Product with Reaction Temperature

Summary of Results for Extended Reaction - Experiment 50C50
Mass Balance Calculations for Experiment 50C51

Carbon Balance Calculations for Experiment 50C51

Hydrogen Balance Calculations for Experiment 50C51
Comparison of Product Analysis by GC and CHN Analysis
Comparison of Measured CO and C02 Flow to Theoretical Flow
Rates Estimated from Steady State Product Concentrations
Box-Benkhen Design Values

Results of Optimization Experiments

Regression Statistics and Coefficients for 2,3-Pentanedione Yield
Regression Statistics and Coefficients for Conversion of Lactic
Acid

Regression Statistics and Coefficients for 2,3-Pentanedione
Selectivity

Regression Statistics and Coefficients for Acrylic Acid Yield
Estimate of Free Lactic Acid in the Reactor Using Reaction Data
of Alén and Sjbstrbm (42)

Rate Constants at Different Reaction Temperatures

Kinetic Constants Calculated from Arrhenius Plots

Comparison of Predicted and Experimental Product Concentration

Results of F-Test on 2,3-Pentanedione Concentrations Predicted

viii

11
24
33

39
40

46
47
48
49
50

54
56
56

57

57
73

75
76
82
83

Table 3.1
Table 3.1
Table 8 1:
Table C 1:

Table C2:
Table D1:

Table D2:

 

 

 

Table 3.12:
Table 3.13:
Table Bl:
Table Cl:

Table C2:
Table D1:

Table D2:

Results of F-Test on Acrylic Acid Concentrations Predicted
Results of F-Test on Acetaldehyde Concentrations Predicted
Conversion of Lactic Acid on Heating “ (42)

Comparison of Predicted Product Concentration and Experimental
Concentration for Kinetic Model B

Kinetic constants calculated for Kinetic Model B

Comparison of Predicted Product Concentration and Experimental
Concentration for Kinetic Model C

Kinetic constants calculated for Kinetic Model C

ix

84
84
89
9O

91

96

. I,“
‘. uJyuv ran-unw—L-l'

 

 

 

 

 

 

figure
Figure
Figure
Figure

Figure ,

Figure 1

Figure 1

Figure 1

FlSure 2.

Figure 2.

ENE:

Figure

I.)

FlSure ‘«
“Elite 3

fiflm3_

Figure 1.1:
Figure 1.2:
Figure 1.3:
Figure 1.4:
Figure 1.5:

Figure 1.6:
Figure 1.7:

Figure 1.8:

Figure 2.1:

Figure 2.2:

Figure 2.3:
Figure 2.4:
Figure 2.5:
Figure 3.1:
Figure 3.2:
Figure 3.3:

Figure 3.4:
Figure 3.5:
Figure 3.6:
Figure 3.7:

LIST OF FIGURES

Chemical Synthesis of Lactic Acid

Reaction Pathways of Lactic Acid

Reactions of 2,3-Pentanedione

Oxidation of Propylene to Acrylic Acid

2,3-Pentanedione Yield using Alkali Salt Catalysts for Vapor
Phase Conversion of Lactic Acid (39)

Reaction Mechanism

Effect of Temperature on 2,3-Pentanedione Yield in a Batch

Reactor (5)

Effect of Initial Alkali Concentration on 2.3-Pentanedione Yield in
a Batch Reactor (5)

Reaction Apparatus for the Continuous Condensed Phase
Conversion of Lactic Acid

Overflow Tube to Remove Liquid Product in Continuous
Operation of Reactor

Gas Chromatogram of Condensed Vapor Products

Product Concentrations vs. Time for Extended Reaction

Plot of Reaction Conditions vs. Time for Extended Reaction
Schematic Representation of Box-Benkhen Design

Contour Plot of 2,3-pentanedione Yield at 220°C

Contour Plot of 2,3-Pentanedione Yield at Flow Rate of 0.6
ml/min

Contour Plot of Acrylic Acid Yield at 240"C

Contour plot of Acrylic Acid Yield at a Flow Fate of 1.1 ml/min
Contour plot of Lactic Acid Conversion at 240°C

Contour Plot of Lactic Acid Conversion at a Flow Rate of

11
14
16

18
2O

27

28

34
43
44

59
60

61
62
64
65

Figuf

Figure
F1 gure
figure

Figure

Figure (
Figure C
Figure C
Figure C
Figure D
Figure D:
Figure D3

Figure D4

Figure 3.8:

Figure 3.9:
Figure 3.10:
Figure 3.11:
Figure A1:

Figure C1:
Figure C2:
Figure C3:
Figure C4:
Figure D1:
Figure D2:
Figure D3:
Figure D4:

Contour Plot for Selectivity to 2,3-Pentanedione Formation at a
catalyst Concentration of 7.5g Cs Lactate/ 100g of feed
Schematic Representation of Reactor Model

Primary Reactions '

Schematic Representation of Alternate Reactor Model
Reaction Apparatus for Semi-Continuous Conversion of Lactic
Acid

Arrhenius Plot for Reaction 1 (Kinetic Model B)

Arrhenius Plot for Reaction 2 (Kinetic Model B)

Arrhenius Plot for Reaction 3 (Kinetic Model B)

Arrhenius Plot for Reaction 4 (Kinetic Model B)

Arrhenius Plot for Reaction 1 (Kinetic Model C)

Arrhenius Plot for Reaction 2 (Kinetic Model C)

Arrhenius Plot for Reaction 3 (Kinetic Model C)

Arrhenius Plot for Reaction 4 (Kinetic Model C)

xi

66

68
69
74
88

91
92
93
94
96
97
98
99

”IV—~21

.Q'

h I “m ’ 'mr.
‘ Y'. IDA
W! m

 

acid to
catalyst:
viable fe
conversit
laborator}
phase reac
Semi-batch
temperature
effective car
lacti
TlllS TEdUCeS
Study Was [01
selecmit}. and
find the On e W
“Ere USCd tO idc‘

 

INTRODUCTION

This research investigates the continuous conversion of biomass derived lactic
acid to 2,3—pentanedione in the homogenous condensed phase using alkali lactate
catalysts. Technological developments have made the use of corn derived lactic acid a
viable feed stock alternative for the production of Specialty chemicals (1). The catalytic
conversion of lactic acid to 2,3-pentanedione in the vapor phase was discovered in our
laboratory (2). Several studies were done to screen catalysts and supports for the vapor
phase reaction (3,4). Preliminary condensed phase studies were conducted in batch and
semi-batch reactors (5). Various alkali metal catalysts were Studied at different
temperatures and feed-to-catalyst ratios. Cesium lactate was found to be the most
effective catalyst.

Lactic acid at high temperatures and at reaction conditions forms oligomers (6).
This reduces the yield and eventually stops the reaction. The main challenge of this
study was to minimize the side reaction of lactic acid to polylactic acid and increase the
selectivity and reaction time. This involved evaluating various reactor configurations to
find the one most suitable for the continuous conversion of lactic acid. Contour plots
were used to identify trends in conversion, selectivity, and product yields. The last step

was to develop a kinetic model.

1.1 Lil

1.1.1 Lat

[i131 diSplay:
functional g;
Occurs nature
Chemical 5W]
traditionally t
imeririet’fliate.\
The major pro.
CCA blOChem
electrodi 31551-5 ‘
fennentation (1 j

reeeiting lnCieac

CHAPTER 1

BACKGROUND

1.1 Literature Review

1.1.1 Lactic Acid

Lactic acid (2-hydroxypropanoic acid) is an optically active compound
that displays a large range of reactions due to the presence of the carboxyl and hydroxyl
functional groups. The Swedish chemist Scheele discovered lactic acid in 1780 (7). It
occurs naturally and can be manufactured through fermentation of carbohydrates or
chemical synthesis. It is also a metabolic intermediate in humans. It has been
traditionally used as a food additive and in textile production. Currently it is an
intermediate-volume specialty chemical with an annual production of 50,000 tons/yr.
The major producers of lactic acid are Archer Daniels Midland, Cargill, A. E. Staley, and
CCA biochem b.v. of the Netherlands. The development of separation technologies like
electrodialysis with bipolar membranes, and extractive distillation integrated with
fermentation (1) has lowered the cost of producing lactic acid. Consequently it is
receiving increasing attention as an intermediate and raw material source.

The chemical synthesis of lactic acid produces only the racemic isomer of lactic

acid. The process is based on a byproduct from acrylonitrile synthesis, lactonitrile. The

base ca‘

aeetalde

 

mineral

acid PTOI

 

Willi watt
r":
l g
g; C H_.CF
. CH
C
Flsure 1.1.-
Prod u
stereoisomer_

 

base catalyzed liquid phase reaction involves reacting hydrogen cyanide and
acetaldehyde. The lactonitrile produced is distilled and hydrolyzed using a concentrated
mineral acid like hydrochloric or sulfuric acid to produce lactic acid. The crude lactic
acid produced is esterified to methyl lactate and recovered by distillation and hydrolysis

with water under acid catalysts.

CH3CHO + HCN __> CH3CHOHCN
CH3CHOHCN + H20 +1/2 H2504 _, CH3CHOHCOOH + 1/2(NH4)3804
CH3CHOHCOOH + CH3OH _, CH3CHOHCOOCH3 + H20
CH3CHOHCOOCH3 +HzO _, CH3CHOHCOOH + CH3OH

Figure 1.1: Chemical Synthesis of Lactic Acid.

Production of lactic acid from carbohydrate fermentation can yield the desired
stereoisomer. Current commercial processes use homolactic organisms like
Lactobacillus delbrueckii, L. bulgaricus, L. leichmanii (8,9,10). Selection of
carbohydrate source depends on price, availability and purity. The fermentation process
requires the use of excess calcium carbonate, to neutralize the acid and maintain the
optimum pH for fermentation. The fermentation is a batch process requiring 4 to 6 days,
and yields in excess of 90% (w/w), from a dextrose equivalent, are obtained. The broth
containing the lactate is filtered to remove the cells, carbon treated, evaporated, and
acidified with sulfuric acid to convert the salt to lactic acid and insoluble calcium lactate,

which is then removed by filtration. The filtrate is further purified to food grade lactic

acid by carb
technical grz
distillation. ‘
process is re
The ;
acidulant’tlz
processed It
agent in foo
for delining
ointments, I
lactic acid 2
Sem of sit-n
An i
for the man
Chemicals.
hydroxil an
Mimenat
CatalEZed Se
“likened [C
Elite
alcohols (. l 3

prCiduciion

“Clo

acid by carbon columns and ion exchange. To produce heat stable lactic acid, the
technical grade is esterified with methanol or ethanol. The ester is recovered by
distillation, followed by hydrolysis, and evaporation. The alcohol recovered in the
process is recycled.

The primary uses of lactic acid are in the food industry. It is used as a food
acidulant/flavoring/pH-buffering agent or as an inhibitor of bacterial spoilage in several
processed foods like soups, candy, bread, etc. Another application is as an emulsifying
agent in foods such as bakery goods. Lactic acid finds use in the leather tanning industry
for delining hides. Pharmaceutical and cosmetic applications include use in topical
ointments, lotions, and biodegradable polymers for medical applications. The use of
lactic acid and other 2-hydroxycarboxylic acids have been shown to alleviate or improve
signs of skin, nail, and hair changes associated with intrinsic or extrinsic aging (11).

An increase in the demand for lactic acid is associated with its use as a feedstock
for the manufacture of biodegradable polymers, oxygenated chemicals, and specialty
chemicals. Polylactides are biodegradable therrnoplastics. Lactic acid has both the
hydroxyl and carbonyl group and can be directly converted to polyesters by condensation
polymerization. The preferred route is the preliminary formation of dilactide by metal
catalde selfesterification. Lactide does not undergo esterification and is readily
converted to high molecular weight polymers by ring opening polymerization ( 12).

Extensive research has been done on the conversion of carboxylic acids to
alcohols (13). The availability of cheap lactic acid has led to the possible commercial
production of propylene glycol and ethylene glycol from lactic acid. Low-molecular-

weight polymers of lactic acid can be used to produce devices for the controlled release

of an
const
week
com pC
contrat

vaccine

Table 1.

; .‘rlolecu

l Molecul
Odor
l\

IISIIJV

S?

l

of an active agent in medical applications (14). Such biodegradable polymers are used to

construct the encapsulation matrix. The active agent can be released over a period of one

week to several months. The release rate depends on several factors including polymer

composition, molecular weight, surface to volume ratio, etc. Active agents include

contraceptives, narcotic antagonists, anesthetics, anticancer agents, hormones, enzymes,

vaccines, etc.

Table 1.1: Physical Properties of Aqueous Lactic Acid

 

Molecular structure

CH3-CH(OH)-COOH

 

 

 

 

 

 

 

Molecular Weight 90
Appearance Clear/light brown aqueous solution
Odor Characteristic
Boiling point 230 0F (40% solution)
257 0F (90% solution)
Density 1.20-1.21 g/ml

 

 

1.1.2.1 Tl

L
decatboxt
p},ruvic ac:
Pennangan
acid. formj,
The Catalytj
acid, acetic

Sulfuric acic

1.1.2.2 Deh‘

1.1.2 Reactions of Lactic Acid

The presence of two functional groups (hydroxyl and carbonyl) makes this molecule
very reactive. The primary reaction pathways relevant to this study are listed in Figure

1.2.

1.1.2.1 Thermal Degradation

Lactic acid undergoes disintegration when oxidized. There are several possible
decarboxylation products. In 1884, Beilstein and Wiegland observed the formation of
pyruvic acid when a 2% solution of calcium lactate was treated with potassium
permanganate (15). The acid was extracted with ethyl ether. The formation of oxalic
acid, formic acid, and acetic acid as decomposition products has been demonstrated.

The catalytic oxidation of lactic acid has produced a mixture of acetaldehyde, formic
acid, acetic acid, and carbon dioxide (16). Oxidation of lactic acid using chromic acid in

sulfuric acid yields oxalic acid and carbon dioxide (17).

1.1.2.2 Dehydration to Acrylic Acid

The demand for acrylic acid polymers has led to several studies for conversion of
lactic acid to acrylic acid (18,19). Holmen claimed a process in which an acrylic acid
yield of 68% was obtained at around 400°C using sulfate and phosphate catalysts (20).

The formation of acrylic acid from lactic acid is also reported in supercritical water (21 ).

The n
cataly
of hyd
intermc
major c
comers
l1 was ft“:

the form;

iodide. (24
Pressure of ,

Faun.

1.13.4 Cond

The mechanism proposed for the dehydration of lactic acid to acrylic acid, without
catalyst, involved the leaving of the Ot-hydroxyl group and the carboxyl hydrogen instead
of hydrogen from the methyl group. This resulted in the formation of a lactone as an
intermediate. The decarboxylation of lactic acid to acetaldehyde was identified as the
major competing reaction. Gunter (22) and Langford (23) performed extensive
conversion studies of lactic acid to acrylic acid over various sodium salts and supports.
It was found that sodium metasilicate and bromate exhibit the highest selectivity toward

the formation of acrylic acid.

1.1.2.3 Reduction to Propylene Glycol

Lautemann first reduced lactic acid to propanoic acid in 1860 using hydrogen
iodide. (24). Catalytic reduction of lactic acid, using rhenium black at an average
pressure of 258 atm. and average temperature of 150°C, gave a propylene glycol yield of

84% (25).

1.1.2.4 Condensation Reactions

Lactic acid undergoes condensation reactions, either with itself or other
molecules, to form larger molecules. Several interesting reactions have been reported
(26). Engelhardt (1849) reported that the dry distillation of lactic acid and copper (II)
lactate resulted in the formation of lactoylllactic acid, dilactide, carbon dioxide and

citraconic acid. Another example of this kind of reaction was reported in 1926 by

 

[pat'ex' an
catalysr cc
products c
acid. buty

presence 6

Ipat’ev and Razuvaev. Sodium lactate was hydrogenated in the presence of water and a
catalyst consisting of nickel oxide and clay at 70 atm. and 270C for 12 hours. The
products obtained were methylsuccinic acid (35% yield) and small amounts of propanoic
acid, butyric acid, and 2-methyl butyric acid (27). The condensation of lactic acid in the

presence of a base yields 2,3-pentanedione (28).

Lactic Aci

ENE 12.

O + C0 + H20
~—>
in ......................................... (1)

CO H
Acetaldehyde + 2+ 2

OH
.__, /\ii/ + 1/2 02 ......................................... (2)

Propanoic Acid

 

OH
/ OH ......................................... (3)
OH Ar + H20
0
0 Acrylic Acid
Lactic Acid

0 ..... . ....................
————> /\”/U\ + C02 + ZHZO (4)
O .

2,3-Pentanedione

O O O
___. I I—hbi
O O n

 

Dilactide Poly (Lactic Acid) (5)
OH
._._. |
/\/OH .................................... (6)
Propylene Glycol

Figure 1.2: Reaction Pathways of Lactic Acid.

 

1.1.3 3"

[Q
in

temperarur
solullons' ‘
('29). ll 3]“
solvenl- lfl
Cum
through a rut
natural form 4
560 (1b ('31 1.
catalysts in [lit
to characten' ze
applications. SI
affine chemica
2.3~Pent
environment wi t
Ofa dilute SOlUIlI
Setitration of the
0rEatticsoltent. 1

ll. 1 .

3th

 

lcations as

1.1.3 2,3-Pentanedione.

2,3-Pentanedione (acetylpropionyl) is a high value fine chemical. At room
temperature, it exists as a clear yellow liquid with a distinct buttery odor in dilute
solutions. It is used as a butter flavoring agent and in coffee products to enhance aroma
(29). It also has potential for use as a photopolymerization initiator and biodegradable
solvent. If its cost is reduced, 2,3-pentanedione has possible use as a feedstock.

Currently, 2,3-pentanedione is produced in limited quantities (about 4000 kg/yr.)
through a multi-step chemical synthesis (30) or by extraction from dairy waste. The
natural form of 2,3-pentanedione costs $60 - $80 llb, while the synthetic form costs $50 -
$60 /lb (31). The formation of 2,3-pentanedione from lactic acid over supported
catalysts in the vapor phase was discovered in our laboratory. Several Studies were done
to characterize the reaction and optimize the reaction conditions. Down stream
applications, such as the use of 2,3-pentanedione as an intermediate for the manufacture
of fine chemicals like duroquinone and pyrazine, were studied by Thiel (32).

2,3-Pentanedione undergoes self-condensation to form duroquinone in a basic
environment with heating. The experimental procedure involved the drop wise addition
of a dilute solution of 2,3-pentanedione (1 wt%) into a solution of NaOH. The
separation of the product was done by liquid-liquid extraction using an immiscible
organic solvent. Only low yields of 10% were obtained which makes this pathway
unfavorable when compared to commercial methods (32). Duroquinones find

applications as antioxidants, photosensitive materials, etc.

10

.4 2‘ ... ._~«;-‘.

5!. n.';:'._!..‘—" _‘.4__'.

 

ethylene
quanutn
diamine

oxide an

Table 1.2
Molecula

; Density
1 .

ix
Borhng P

l

I‘\.
l Retractm

\

Sure 1 3.

2,3-Pentanedione also reacts with vicinal diamines(oc,B—diamino) compounds like
ethylenedianrine and propylenediamine. The reaction was achieved by refluxing equimolar
quantities in a solvent. The solvent chosen is usually the ether corresponding to the
diarnine. The dihydropyrazine formed was oxidized to the pyrazine using a copper II

oxide and KOH catalyst (3 2).

Table 1.2: Physical Properties of 2,3-Pentanedione.

 

 

 

 

 

Molecular Weight 100.11 (g/mol)
Density 0.9565 g/ml
Boiling Point 108 °C
Refractive Index 1.4068

 

 

iii

duroquinone

O
2.3-Pentanedione W
W

  

0t,B-diamine 2-ethyl-3-methyl alkylpyrazrnes

Figure 1.3: Reactions of 2,3-Pentanedione.

ll

 

I
-~. ‘-
~:::m 132w}? 'gerWJ

 

 

1.1.4 A.

Althou g
kept prit
esters in

manufac
most do;

acrolein.

mOl}’bde

figure 1_.

arid then

b

1.1.4 Acrylic Acid

The annual worldwide production of acrylic acid is currently at 2 billion lb.
Although the demand for acrylic acid is steadily growing, increased plant capacities have
kept prices low. This growth can be attributed to the increasing demand for acrylic
esters in water based coatings and adhesive applications. Acrylic acid is also used in the

manufacture of superabsorbent polymers.

Of the four commercially viable processes for the manufacture of acrylic acid, the
most dominant is the oxidation of propylene from steam crackers in refineries to
acrolein. Acrolein is further oxidized to acrylic acid at high temperatures over a

molybdenum-vanadium catalyst (33).

propylene aCTOICIU

\— + 02 —’ \jl + H30
0

\ .1/2. __. H
\i 0- W0

O
acrylic acid

Figure 1.4: Oxidation of Propylene to Acrylic Acid.

Other processes include the treatment of ethylene oxide with hydrocyanic acid

and then sulfuric acid. Hydrolysis of acrylonitrile is another possible pathway.

12

 

P
synthesis
exclusive
hydrofom
temperatu
Whey using
cconomica
Propanoic a
consumed 1:
corn Presert
drugs Syntl
is PDCCd SEW
com’mion 01

add fut-[her Va

1.1.5 Propanoic Acid

Propanoic acid (propionic acid) can be manufactured either by chemical
synthesis or fermentation of carbohydrates. Synthetically, propanoic acid is produced
exclusively by the oxidation of propionaldehyde, which is obtained by the catalytic
hydroforrnylation of ethylene with carbon monoxide and hydrogen, under relatively high
temperature and pressure (34). Propanoic acid is also produced from the fermentation of
whey using Propionibacterium acidipropionici (35). This route is however not
economically viable as compared to the synthetic route. The annual production of
propanoic acid is around 400 million pounds worldwide, of which about 60% is
consumed by the U. S. (36). Propanoic acid and its salts find use as animal feed and
corn preservatives, mold inhibitors, herbicides, cellulosic plastics, and anti-arthritic
drugs. Synthetic propanoic acid is priced around $0.41/lb, while natural propanoic acid
is priced several times higher. This has led to the investigation of the possible
conversion of byproduct acrylic acid to propanoic acid by hydrogenation, which would

add further value to the conversion of lactic acid.

13

 

 

catalyst
doneto
conditio
conversi
50dlum ;
meSA
and selem
PhOSphate
SPfictroscoj
COnfirmed l
and NaRPOJ
pOlHDl’iOsphé
PFOIOn exCha
the pregame t
the Catalflic al

acetaldeinde at

  

1.2 Vapor Phase Conversion of Lactic Acid

The continuous formation of 2,3-pentanedione from lactic acid over supported
catalysts in the vapor phase was discovered in our laboratory and several Studies were
done to characterize the reaction over different catalysts and to optimize the reaction
conditions. Studies with various sodium salts showed that the best catalysts, for
conversion of lactic acid to 2,3-pentanedione formation, were group IV and V oxides.
Sodium arsenate gave the best 2,3-pentanedione yield of 25% at a temperature of 300°C
and 0.5 MPa (37). Studies with sodium phosphates demonstrated that both conversion
and selectivity toward 2,3-pentanedione increased with increasing basicity of the
phosphates; in the order of Na3PO4 > NazHPO4 > NaHzPO4. FTIR and 31P-NMR
spectroscopic studies of the phosphate catalysts, before and after lactic acid conversion,
confirmed the formation of tetrasodium pyrophosphate and sodium lactate on NazHPO4
and Na3PO4 catalyst surfaces after reaction (38). With NaHzPO4, sodium
polyphosphates and sodium trimetaphosphate were observed with considerably less
proton exchange between the phosphates and lactic acid. The conclusion drawn was that
the presence of sodium lactate and/or pyrophosphate on the supported catalysts increased
the catalytic activity toward the formation of 2,3-per‘ «ne.

In the vapor phase, lactic acid is converted to 2,.- pentanedione, acrylic acid, and
acetaldehyde at 0.5 MPa over phosphate salts (3). Highest selectivity to 2,3-
pentanedione is achieved at 280-3000C and long residence times, while selectivity to

acrylic acid is best at 350°C and short residence times. The use of low surface area

14

1 , .,_ ._f-.~;_v1 .1-

 

supports
propanoi
materials
Rt

from lacti
MPa. and
was recom
shows the 1

lactic acid 4

supports like silica was found to minimize undesirable Side reactions to acetaldehyde and
propanoic acid, which are favored over high surface area (microporous) or surface acidic
materials.

Reaction conditions were also optimized for the formation of 2,3-pentanedione
from lactic acid, over sodium salts supported on low surface area silica, at 300°C, 0.5
MPa, and 3 second residence time (39). Use of silica support with low surface acidity
was recommended to lower the activity toward acetaldehyde formation. Figure 1.5
shows the percent theoretical yield of 2,3-pentanedione for the vapor phase conversion of

lactic acid over different alkali salt catalysts.

15

 

‘
Er: 3.3

(
at.

.
a .
.2. ..i 3.3;

0
«I.

0:. 13.4.2323

fl- ..v... . c. .

ith/DV

...”...m.

{FINN-ti D. .f 30.6.}- .14.-

tm

260

. S/

are 1.5
Using Allah

9

l

Lactic Acid Conversion Over Alkali Salt Catalysts at P=0.5
MPa, RT=3 s, and loading = 1 mmol/g

 

60
-<.>— LiOH

-0— KCl

+ KOH
+ KNO3
+ C52804

.41“ --—- CSOH
/

-><- CSNO3

50 "" x

    

4o --

2,3-Pentanedione Yield (% Theoretical)

 

 

 

l
I

ni—

260 280 300 320 340 360
Temperature (°C)

0 T

Figure 1.5: 2,3-Pentanedione Yields for Vapor Phase Conversion of Lactic Acid

using Alkali Salt Catalysts (39).

16

‘ l'
..<.. -1.-;~r.s..
+31%—

“ 41.» ~
“——

 

—_'

 

 

Where .\1
mai'K'SlS. r
reflectance
and compu
calculations
Unde-
hidmgen bor

Which then m.

5“ Cantton dih

 

1.3 Mechanism

Craciun (40) proposed the mechanism of catalytic conversion of lactic acid to
2,3-pentanedione. The reaction takes place in the vapor phase over supported alkali
metal phosphates, nitrates, and hydroxides and in the homogenous condensed phase
using alkali metal salt catalysts (5). The mechanism was proposed for conversion of
lactic acid to 2,3-pentanedione in the presence of supported base catalysts (MOH/SiOz,
where M = Na, K, Cs). The possible reaction intermediates are suggested by product
analysis, variable temperature-mass spectroscopy (VT-MS), post reaction diffuse
reflectance infrared Fourier transform (DRIFT) spectroscopy, deuterium labeling studies,
and computational modeling using semi-empirical and ab initio molecular orbital
calculations.

Under high temperature and buffered lactic acid/alkali lactate conditions,
hydrogen bonds and ion pairing may Stabilize the enolate or enol forms of lactic acid
which then may attack a molecule of lactic acid at carbonyl. After the loss of water, the
six carbon dihydroxy-ketocarboxylate (lnterrnediate I) loses CO; to form Intermediate [1.
Loss of hydroxide leads via Intermediate III and subsequent ketonization to 2,3-
pentanedione (40).

Post reaction FT IR studies of the salt catalyzed vapor phase reaction have
revealed that alkali lactate is the primary stable species found on the support during the
reaction (3). The formation of 2,3-pentanedione and acrylic acid are strongly influenced

by the ability of the sodium salt to form sodium lactate on the surface of the support;

17

 

further th
lactic acic
on the sur

reaction 0

figure 1.6-

further the rate of formation of sodium lactate depends on the accessibility of the salt to
lactic acid. Maximum accessibility is expected for salts that form a molten or liquid film
on the surface at reaction conditions (3). This hypothesis led to the development of the

reaction of lactic acid to 2,3-pentanedione in the condensed phase.

 

Figure 1.6: Reaction Mechanism.

18

 

 

1.4

 

 

I J
(J)
I
”D

cold‘f
impI-O‘

salt C31z

phase in .
fixed ini ti
carricr S35-
formed. Th
constant thr<
condenser an

The cc

These results v

  
   

Operating cond

pttduct yield or.

alkali lactate ca

1.4 Rationale for Proposed Work

The initial experiments were done to determine the feasibility of the lactic acid to
2,3-pentanedione conversion in the condensed phase. After verifying the reaction in
cold-finger type reaction flasks, efforts were made to characterize the reaction, to
improve the reactor design, and screen different homogenous catalysts (5). Alkali metal
salt catalysts were found most suitable for this reaction in a batch reactor.

Lactic acid was converted to 2,3-pentanedione in the homogenous condensed
phase in a one step process. The feed lactic acid was mixed with the alkali lactate in a
fixed initial ratio and heated to the reaction temperature using a ramping program. A
carrier gas, like helium, was bubbled through the reaction mixture to remove the product
formed. The reaction mixture was well stirred to keep the concentration and temperature
constant throughout the reactor. The product gases were condensed using a water-cooled
condenser and an ice trap.

The condensed phase conversion was found to be sensitive to several factors (5).
These results were important in catalyst selection and determining the optimum
operating conditions for the continuous reactor. Figure 1.7 shows the percent theoretical
product yield obtained using different maximum temperature setpoints and for different
alkali lactate catalysts. The yield was found to increase with increasing temperature up
to 250°C; the yield dropped beyond this temperature. This drop in yield was

accompanied with an increased formation of polylactides and char in the reactor.

19

 

 

FIN‘UI‘I'I inn“

ntnncdionc Yield

1%

2.3-Po

Figure 1.7:

Figure
acid used. For
”15 alkali salt tr

athlet'ed were 7

 

 

&
U]

b.
Q
%

-x- Na+

(a)
(II
1

-a-K+

-0- Cs+

2,3-Pentanedione Yield
(% Theoretical)
G 8 8’. ‘5

he
a
1

VI
1

 

 

 

200 220 240 260 280 300 320
Temperature (°C)

Figure 1.7: Effect of Temperature on 2,3-Pentanedione Yield in a Batch Reactor (5).

Figure 1.8 shows 2,3—pentanedione yield vs. initial ratio of the alkali lactatezlactic
acid used. For each alkali lactate salt, the yield was a strong function of initial ratio of
the alkali salt to free acid loaded in the reactor. Maximum yields of 2,3-pentanedione
achieved were 20.5% of theoretical using sodium lactate/lactic acid mixture, 28.1%
using potassium lactate/lactic acid mixtures, and 41.5% using cesium lactate/lactic acid
mixtures. Clearly, for any given initial ratio of lactic acid to alkali lactate, the yield
depends on the strength of the bases used. This is consistent with the results of the vapor
phase studies. Thus, cesium lactate gave the highest yields of all the base used and was

chosen for the continuous reactor optimization studies. The strength of the base affected

20

"34.017-

 

 

the

[CIT

prei

 

nltlncdinnc Yield

((7, 1"Inl‘““‘ °

2,3-1’ '

F1é'Ure 1.8:

batch reactor

The prg

lipiESemS all 17]

Th
e Condenged

Frail}; the filth

the 2,3-pentanedione yield. From Figure 1.8, we can further see that for any given

temperature and catalyst concentration the highest yields were obtained from the catalyst

prepared from the strongest base.

 

45

4o ..

35 J’ -x- Na+
w- K+
-0- Cs+

iii 8
1 1

/

/

2,3-Pentanedione Yield
(% Theoretical)

n-I

e

4
U

 

 

 

 

0.01 0.1 1
Fraction Lactate in Initial Mixture (as Alkali Lactate)

Figure 1.8: Effect of Initial Alkali Concentration on 2,3-Pentanedione Yield in a

batch reactor (5).

The production of 2,3-pentanedione from the homogenous condensed phase
represents an improvement over current vapor phase and petroleum based technologies.
The condensed phase conversion has several advantages over these other technologies.

Firstly, the alpha-diketone is produced in a single processing step in good yields and at a

21

 

poten
Secor
’acilit

difficu

 

lOuert
from th
process:
E
favorable

longer per

‘ t
.5

potentially low cost with the technical challenges of feed vaporization avoided.
Secondly, the reaction proceeds without the use of supported catalysts, thereby
facilitating uniform reaction, eliminating mass transport limitations and minimizing
difficulties associated with catalyst deactivation. The reactions can be performed at
lower temperatures and pressures. Further, the need to remove all nonvolatile residues
from the feed prior to conversion may be circumvented and large amounts of feed can be
processed. Ideally, of course this reaction should extend to other potential substitutes.
By choosing a continuous reaction system, it is hoped that the conditions
favorable to the formation of 2,3-pentanedione can be maintained in the reactor for a

longer period of time.

1.5 Research Objectives

The primary objective of this investigation was to demonstrate that the
conversion of lactic acid to 2,3-pentanedione is possible in a steady state continuous
reactor. The second objective was to identify process variables that affect the
performance of the reactor and optimize the reaction conditions. The last goal was to

study the kinetics of the reaction and propose a model to describe the system.

w M
. wwssu-‘vt’. J

 

 

I‘d
t—I

T
complex
the result:
dcpendent
nature of r1
formation (
reactor it vr
PeDOd of III
Vessel alloW

TEdCIOr WOUlt

22

4‘18”"
i 7
Lacrjc at
SOlUIio
I1
. 6 CC

 

CHAPTER 2

LACTIC ACID CONVERSION IN A CONTINUOUS REACTOR

2.1 Introduction

The condensed phase conversion of lactic acid using alkali lactate catalysts is a
complex reaction and designing a continuous reactor presented several challenges. From
the results of the batch experiments it was evident that the yield of 2,3-pentanedione was
dependent on several reaction conditions, especially the temperature. The reactive
nature of this molecule made it difficult to find operating conditions favorable to the
formation of a single product. By conducting the reaction in a steady state continuous
reactor it was possible to maintain the Optimum conditions in the reactor for an extended
period of time. This would minimize the production of polylactic acid in the reaction
vessel allowing the reaction to proceed at a higher selectivity. Further, a continuous

reactor would facilitate scaling up of the reactor system to a pilot plant.
2.2 Methods
2.2.1 Materials

Lactic acid used for all studies was supplied by Purac, Inc. as 88 wt.% aqueous

solution. The cesium hydroxide used to prepare the alkali lactate catalysts was obtained

23

 

H)
In‘
C)

 

”Elli.“ Ell
Acetaldeh:

2‘Pfopanol
\\

from Aldrich as a 50% aqueous solution of 99% purity. The lactic acid was diluted to

34% or 50% with HPLC grade water for continuous studies. High temperature grease

was used to seal the glass-to-glass joints in the reaction vessel. A high temperature

melting point bath oil (SIGMA catalog # M-9389) was used in the oilbath for additional

heating and also in the thermocouple well. The oil is a clear colorless silicone fluid and

is stable to temperatures of 315°C. Chemicals used as GC standards are listed below.

Table 2.1: Chemicals Used as Gas Chromatography Standards.

 

 

Name Supplier Catalog # Purity CAS-Registry
2,3-Pentanedione Aldrich 24,196-2 97% 600- 14-6
Acrylic acid Aldrich 14,723-0 99% 79-10-7
Propanoic acid Aldrich 24,035-4 99% 79-09-4
Lactic acid Purac FCC grade 88% 79-33-4
Methyl ethyl ketone Aldrich 27,069-5 99.5% 78-93-3
Acetaldehyde Aldrich 1 1,007-8 99% 75-07-0
2-Propanol Aldrich 27,0490-0 99.5% 67-63-0

 

24

 

 

 

If,

Set'e

 

2.2.2 Catalyst Preparation

Catalysts were prepared as alkali lactate by the dropwiSe addition of
stoichiometric amounts of the required base to lactic acid. This produced equimolar
quantities of the alkali lactate and water. The cesium hydroxide solution used was 50%
aqueous and 99% pure and the homogenous catalyst prepared contained 44.75% by
weight of cesium lactate. To dissipate the large amount of heat generated, the addition
was done in an ice bath, followed by stirring to ensure homogeneity. The final solution

was stored under helium.
2.2.3 Reactor Description

Several reactor configurations were attempted before arriving at this reactor setup.
Details of reactor development are given in Section 2.3.1. Initial semi-continuous runs
were done in a 3-neck 100-ml round bottom flask. This was changed to a 50-ml
cylindrical reaction flask and a four-neck flask head. Figure 2.1 shows the reaction
apparatus for the continuous conversion of lactic acid in the condensed phase. The flask
head had openings for the following connections: a central 19/38 joint for the
introduction of carrier gas, two straight 14/20 joints to hold the feed inlet and
thermocouple, and a 14/20 elbow for removal of product gases. The inlet for carrier gas
was a glass tube of 6 mm ID. The tube extended below the liquid surface in the reactor.

This caused the carrier gas to bubble through the reaction mixture.

t/r

V85

\ito

fe w

heati

 

using
was p

the rea

filled w.
“'00! and
stirrer hm

heating an.

The continuous feed was pumped from a 125 ml Erlenmyer flask into the reaction
vessel using a Watson Marlow (Model # 505-8) peristaltic pump, with 0.8 mm ID
Viton® tubing. The reactor control thermocouple was placed in a 4 mm glass tube with a
few drops of high temperature melting point bath oil. The reactor was heated using a
heating tape 6 feet long and 0.5 inches wide. The heating of the reactor was controlled
using an Omega CN 2011 programmable temperature controller. A second thermocouple
was placed between the heating tape and the reactor to monitor the surface temperature of
the reaction flask.

Additional heat was supplied to the reactor by placing it into a stirred oilbath
filled with high temperature melting point bath oil. The reactor was insulated with glass
wool and insulating firebricks to minimize heat loss. The reactor was supported over a
stirrer hot plate, which heated the oilbath and also stirred the reactor and oilbath. The

heating and stirring rates were controlled manually.

26

 

 

 

[tilt

- r -211. l.-llill
9:2920 1.29. P =O ....vza Z

 

 

 

E3. 283 Lo 568250 Omega 323500 $055.50 2: SC magma/x :ozomom ”..m DSwE

 

 

 

 

as... 03

DJ

comet: 2:63- l j

 

 

 

 

 

 

 

 

 

 

 

 

 

 

film

5:38 0395

 

 

 

 

Emcee—80 “co—coo .833

   
              

82a waxwom
53:0

  

 

  

        

26930
.anEm coach \ one. wagon
Boas max—O 8:05.80
0.53.595.
x25 meta—.55 r

 

 

$1: mam
2.2—om

0.53.89th cozuaom

 

\. 0. .

contin

and als

40 ml.

 

The CV
liquid p

tubing.

To 1in

5301‘827.

Redc I 01'.

 

The reaction flask was equipped with an overflow tube (Figure 2.2) for
continuous operation. The tube continuously removed liquid product from the reactor
and also acted as a level control for the reactor, resulting in an effective reactor volume of
40 ml. The overflow tube was connected to a flask to collect the liquid overflow product.
The overflow tube and the collection flask were heated with heating tape to prevent the
liquid product, which contained oligomers of lactic acid, from cooling and blocking the

tubing.

Liquid product from Reactor

 

 

 

 

 

 

 

 

10m
_>
IA‘ | ‘ 45mm
\ 2 _x
30mm
—N u——3—
14’

 

To liquid product receiver

Figure 2.2: Overflow Tube to Remove Liquid Product in Continuous Operation of

Reactor.

28

‘ "Q
Ezra": W

vapc

some

 

water
traps f

trap pl;

cooled b
ofthe CC

COandC

”"0 Riken

 

The gases exiting the reactor consisted of the helium carrier gas, condensible
vapor products including 2,3-pentanedione, acrylic acid, acetaldehyde, propanoic acid,
some unreacted lactic acid, and product gases carbon monoxide and carbon dioxide. The
gases were cooled and partially condensed using a water-cooled glass condenser. The
water-cooled condenser was inclined so that any product condensed flowed down to the
traps by gravity. The products were completely condensed and collected in a cold finger
trap placed in ice.

The reaction apparatus for initial experiments included a second cold finger trap
cooled by liquid nitrogen for product collection. This was later abandoned because some
of the CO; would cool in the tubing and block the flow of the carrier gas. The amount of
CO and C02 present in the product gases was determined for selected experiments using

two Riken infra red gas analyzers (model 550 A) obtained from CEA instruments, Inc.

2.2.4 Experimental Procedure

The experimental procedure for the continuous reaction involved the following
steps. First, the reactor flask and flask head were connected and high temperature grease
was used to seal the joints. The reaction vessel was wrapped with heating tape and the
fittings for feed inlet, carrier gas inlet, thermocouple, and product gas outlet were
connected. The reaction vessel was placed into an oilbath that was filled with high
temperature melting point bath oil to a depth of 0.5 inches. Next, the reactor was

insulated with glass wool and insulating firebricks. The overflow tube and collection

29

 

flask We“
was the“ C—
The Produ.
liqUid ”m0
The
catali'5" FC
lactiC acid a
molar r 3210 i
into the reaL‘ ’
(helium) “0";
measure the s
in batches of

feed solution 1

 

 

 

 

Afier c
started using a
hEating of the
increased fro"
.‘ifier the temp
at a rate of 3 r;
temperature Ha.
Ha; increased to
flow rate was 0’0.

{"1 '
...ntrnuowfeed f

flask were wrapped with heating tape and connected to the reaction vessel. The reactor

was then connected in series to the water-cooled condenser and product collection traps.
The product collection traps were placed in Dewar flasks filled respectively with ice and
liquid nitrogen. '

The initial reactor charge consisting of 88 wt% lactic acid and cesium lactate
catalyst. For most continuous experiments, the initial reactor charge consisted of 36 g of
lactic acid and 8 g of cesium lactate catalyst. This corresponded to a lactic-to-lactate
molar ratio of 22: l. The reaction mixture was weighed in a 100-ml beaker and poured
into the reactor. The reaction mixture was stirred with a Teflon® stirbar. Carrier gas
(helium) flow was started and stabilized at 50ml/min. A soap bubble column was used to
measure the gas flow rate. The continuous feed of 50% aqueous lactic acid was prepared
in batches of 100 g and the desired amount of cesium lactate catalyst was mixed with the
feed solution and continuously fed to the reactor.

After completing the initial preparations, the heating of the reaction mixture was
started using a ramping program Similar to the batch reaction. Simultaneously, the
heating of the overflow line and the oilbath was started. The reactor temperature was
increased from room temperature to an intermediate setpoint of 130°C in 30 minutes.
After the temperature reached 130°C, feed addition was begun using the peristaltic pump
at a rate of 3 rpm, which corresponded to a flow rate of about 0.3 ml/min. The reaction
temperature was maintained at 130°C for 20 minutes while the continuous feed flow rate
was increased to the final flow rate of 0.5-1 .5 ml/min. This increase in continuous feed
flow rate was done manually at a rate of about 0.5 ml/min every minute. After the

continuous feed flow rate had reached steady state, the reaction temperature was

30

in:

ICTI

 

coll
were
prodi
pl’O‘CBt

an E X1

to
lo
bu

Cc

reaction is g

25 CC 3.0

\

 

increased to the final reaction temperature setpoint in 40 minutes. Thus, steady state in
temperature and continuous feed flow rate was reached an hour and 30 minutes after the
reaction was started. To establish steady state, it was found necessary to increase the
feed flow rate and temperature gradually, as described. I

After the reaction mixture reached the steady state temperature, the product
collection trap was changed to begin collection of steady state product. The product traps
were weighed and the sample collected for analysis by gas chromatography (GC). The
product collection traps were emptied and replaced periodically as the reaction
proceeded. The product concentrations obtained from the GC analysis were entered into

an EXCEL 7.0 spreadsheet to calculate product yields.
2.2.5 Reaction Conditions

Condensed phase conversion experiments were conducted at a range of reaction
conditions. A typical ramping program for reaction temperature of the continuous

reaction is given below.

25°c—3m—>130°C—20—"“—"—>130°C 40mm >240°c 300“” >240°c

 

 

Quantifiable yields of 2,3-pentanedione were achieved using maximum reaction
temperature setpoints of 200°C to 300°C. Helium gas flow rates of 25 to approximately
2000 ml/min were used in the batch reactor (5); best results had been achieved at 50

ml/min. This flow rate for the carrier gas was used in the continuous reactor studies.

31

2.2.6 Product Analysis

Product analysis was performed using a Varian Model 3700 gas chromatograph,
connected to a Hewlett Packard Model 3394 integrator for determination of product
concentrations. The chromatograph uses a flame ionization detector and a 2 mm 1. D. x
2 m, 4% Carbowax 80/100 Carbopack B-DA glass column (Supelco). Samples were
prepared for injection as a 1:1 mixture of product sample and internal standard. The
internal standard used was a lOg/L aqueous solution of 2-propanol in a 0.06 M solution
of oxalic acid. The oxalic acid was needed to condition the column. Samples that
contain only an aqueous phase are mixed directly; biphasic and neat organic samples
were diluted in a quantity of HPLC grade acetone or water adequate to produce a single
phase. Dilution sizes were recorded and final concentrations from GC analysis were
calculated accordingly. One microliter samples were injected for analysis.

Raw data from the chromatogram, along with feed and product masses, feed flow
rates and reaction conditions were entered into an Excel spreadsheet to calculate product
yields and selectivity. All product yields are calculated from the ratios of
product/internal standard peak areas and detector response factors. Product yields are
reported as a percentage of the theoretical yield based on the lactic acid fed to the reactor;
product selectivity is the percentage of theoretical yield based on lactic acid consumed in
the reactor. Product component identification was done in earlier studies (22,23,39) by
residence time analysis, gas chromatography/ mass spectrometry (GC/MS), and 1H and

13C NMR spectroscopy.

Figure 2.3 shows a chromatogram of the condensed vapor product. The
temperature program used for all gas chromatography analysis was-

So’ . .
100°C ' 0mm >200°C—m—+200°C

 

Using this program 2,3-pentanedione appears at a time of 9.5 minutes. Table 2.2

shows the response factor and retention time for major products.

Table 2.2: Retention Time and Response Factor for Product Analysis by Gas

 

 

Chromatography.

Product Response factor Time (min)
Acetaldehyde 2.00 0.92
2-propanol (Internal Standard) 1.00 3.03
2,3-Pentanedione 1 . 14 9.42
Propanoic acid 1.60 11.03
Acrylic acid 1.98 12.77
Lactic acid 4.10 20.36

 

33

 

 

 

t:—
9.42

 

Figure 2.3: Gas Chromatogram of Condensed Vapor Products. (Numbers to the right

of peaks indicate retention time in minutes)

34

2.3 Results and Discussion

Cesium lactate was chosen as the catalyst for continuous reactor studies as it gave
the best yields in semi-batch studies. Several variables were identified as important in
the optimization of the continuous reactor performance. Reaction temperature and
catalyst concentration in the feed affected the performance of the reactor. It was found
that operating the reactor at lower temperatures allowed the reaction time to be increased
from one hour in semi-continuous operations to 24 hours in continuous Operation. The
lower reaction temperature (ISO-250°C) maintained in the reactor reduced the formation

of tar and oligomers of lactic acid.
2.3.1 Development of Reactor Configuration for Continuous Operation

The reactor configuration had to be adjusted several times to facilitate the
continuous processing of material. Initial semi-continuous condensed phase conversion
was achieved using the same lOO-ml 3-necked flask used in prior batch studies. A figure
of this reaction apparatus is included in Appendix 1. Temperature control was difficult
using this setup as the heating mantle could not supply the heat needed to maintain a
steady state temperature with feed addition. Therefore, a smaller (25 ml) round bottom 3-
neck flask was used to reduce the reaction volume and the heat requirement. This reactor
produced the highest yields in both the semi-batch and continuous operation. In

experiment 25C08, at a reaction temperature of 235°C and 35% lactic acid feed fed to the

35

reactor at 0.1 ml/min, a 2,3-pentanedione yield of 48% was obtained over a run period of
4 hr.

These initial runs were categorized as semi-continuous because only condensible
vapor products were being removed from the reactor. The level of the liquid reaction
mixture in the reactor changed as the reaction progressed because of the accumulation of
lactic acid in the reactor. The greater extent of oligomerization of lactic acid could be
attributed to the long contact time. For true steady state operation, the reactor setup
required the addition of an overflow tube to continuously remove the liquid phase from
the reactor. This would reduce the residence time of lactic acid in the reactor and also act
as an effective level control. This prevented the build up of lactic acid oligomers in the
reactor and extended reaction time. The SO-ml reaction flask described in Section 2.2.3,
allowed the modifications required for the continuous removal of liquid product from the
reactor and also allowed the use of heating tape that provided better temperature control
than the heating mantle used previously.

In experiment 50Cl9, addition of externally generated steam was used as a further
heat source for the reaction. It was also theorized that the use of steam would
continuously reverse the oligomerization of lactic acid. The addition of steam did not
serve its purpose as it lowered the reaction temperature and the reaction temperature
could not be maintained at 240°C . This was because the steam supply available in our
laboratory was at 20 psi. For its possible use as a heat source the steam needed to be
superheated. Instead, the additional heat load was supplied by placing the reactor to a
depth of 0.5 inches in a stirred silicone oil bath heated by the stirrer-hot plate supporting

the reactor. This proved an effective means of heating the bottom of the reactor.

36

Several schemes were tried to operate the reactor as a CSTR. One involved using
another tubing in the peristaltic feed pump to remove the liquid from the reactor
(Experiment 50C21). However, the small diameter tubing used for feed addition did not
permit the free flow of the highly viscous liquid product and the removal tube became
blocked afier a period of time. This scheme was abandoned in favor of removing liquid
from the reactor via an overflow tube using gravitational flow. It was found necessary to
heat trace the overflow tube from the reaction vessel to the flask collecting the overflow.
This lowered the viscosity of the liquid products enough to allow the smooth flow of the
liquid product into a receiver.

Finally, glass beads were used in experiment 50C12 to increase the surface area
for reaction, however the stirring of the reaction mixture with the Teflon stirbar was

ineffective with the glass beads present, and the use of glass beads was thus abandoned.
2.3.2 Effect of Temperature and Catalyst Concentration in Feed

A typical temperature program used in the continuous condensed phase
experiments is given below:
25°C —3"—+ 130°C —‘°“’-—+ 130°C 4‘1» 220°C —""—> 220°C
In experiment 50C25, using an initial reactor charge of 36 g lactic acid and 8 g of
cesium lactate, at a temperatures of 200°C and a feed (50% aq lactic acid) flow rate of
0.55 ml/min, a conversion of 13% was obtained. The addition of feed was begun at a
temperature of 160°C. Less amount of liquid product was recovered from the reactor, as

the liquid level in the reactor did not reach the level of the overflow. In experiment

37

50C26, at similar reaction conditions, the feed addition was begun at a temperature of
130°C. The amount of liquid product collected increased three fold to 30 grams and the
conversion increased to 19%. This showed that the establishment of a steady state in the
feed flow and reaction temperature depends on the continuous removal of liquid product
from the reactor. However a larger amount of acrylic acid was formed relative to 2,3-
pentanedione. The low selectivity to 2,3-pentanedione of 24% could be attributed to the
low catalyst concentration in the reaction mixture because of the absence of catalyst in
the continuous feed. The method of feed addition described above was repeated in
experiment 50C27 with catalyst in the continuous feed. At a temperature of 200°C, with
a similar initial reactor charge, the selectivity increased to 36%. Table 2.3 summarizes

the results of experiments 50C25, 50C26, and 50C27.

38

 

 

Table 2.3: Summary of Results for Experiment 50C25, 50C26, and 50C27.
50C25 50C26 50C27
Date 2/1 1/98 2/21/98 2/24/98
Reaction temperature 200°C 200°C 200°C
Reaction time 2 hr 11 min 3 hr. 3 hr. 4 min
Initial reactor charge
Lactic acid (g) 36.33 36.02 36.02
Cesium lactate (g) 8.022 8.004 8.001
Lactic-to-lactate mole ratio 21.7 21.8 21.8
Continuous feed
Lactic acid (g) 99.9 171.1 176.1
Cesium lactate (g)/ 100 g of feed 0 0 3.75
Feed addition start 160°C 130°C 130°C
Feed time (min) 101 150 154
Feed flow rate (ml/min) 1.0 1.2 1.1
Total Vapor Product( g ) 70.5 141.5 136.3
2,3-Pentanedione (g) 2.00 3.02 1.81
Acrylic acid (g) 2.74 7.53 1.56
Acetaldehyde (g) 2.08 3.25 1.75
Propanoic acid (g) 0.43 0.88 0.18
Liquid product-Overflow (g) 11.2 30.35 36.72
Conversion (%) 14.6 18.9 7.6
Selectivity(23P) 30.1 23.8 35.8
% Theoretical yield (23P) 4.3 4.5 2.7

 

In experiment 50C33, a series of product samples were taken as the reaction

temperature was being increased. Table 2.4 shows the concentration of 2,3-pentanedione

determined from gas chromatography for the product samples. As seen, the presence of

2,3-pentanedione was seen in the first sample taken at a temperature of 170°C. The start

of 2,3-pentanedione production was also confirmed by observing yellow condensate on

the walls of the glass condenser. The increase in 2,3-pentanedione concentrations reflects

drying out of the initial reactor charge to lower the water content in the product.

39

Table 2.4: Variation of 2,3-Pentanedione Concentration in Condensible Vapor

Product with Reaction Temperature.

 

 

Sample Time Reaction Temp 2,3-pentanedione
# (min) (°C) concentration (g/lt)
l 70 173 8.1
2 73 185 13.7
3 78 196 20.8
4 98 197 25.8
5 158 198 44.1

 

2.3.3 Effect of Feed Rate

One of the goals in the development of the continuous reactor was to establish
steady state operation of the reactor. Initial continuous experiments were done using
lactic acid feed concentrations of 34%. However, the large amounts of water present in
this continuous feed increased the heat load on the reactor, and the steady state
temperature could not be reached. For continuous operation, the feed lactic acid
concentration was increased to 50%. This reduced the amount of water that needed to be
evaporated and also increased the concentration of the product. The flow rate was
increased from 0.1 ml/min to 0.5 ml/min in experiment 50C20. This permitted the level
of liquid in the reactor to reach the level of the overflow faster and facilitated recovery of
liquid product. Thus, given the reactor heating limitations, steady state in the reactor

could only be established with higher feed flow rates and lactic acid concentrations.

40

2.3.4 Extended reaction

To determine if the reactor operated at steady state, an extended reaction
(experiment 50C50) of 24 hours was done at 220°C. The catalyst loading used was
7.5g/ 100 g lactic acid feed at a feed flow rate of 1.1 ml/min. The lactic acid conversion
was about 27% and the selectivity to 2,3-pentanedione was 37%. Overall yield of 2,3-
pentanedione was 9.7% and the yield of acrylic acid was 10.1%. Over 1600g of 50%
lactic acid along with 120g of cesium lactate catalyst was fed continuously to the reactor.
This corresponded to a feed throughput of 43 reactor volumes. The total liquid product
collected as overflow was 343g, which corresponded to 8.6 reactor volumes. The product
collection traps used in this run were 500 ml Erlenmeyer flasks. Product concentrations
and reaction temperatures were monitored to observe if there were any changes over
time.

As seen in Figure 2.4, there was a little oscillation in the product concentrations
measured by gas chromatography. Consistent production of 2,3-pentanedione was seen
even at 24 hours, establishing that the condensed phase reaction was stable. The plot of
reaction conditions vs. time shown in Figure 2.5 show that good control of temperature
was maintained through the reaction. Figure 2.5 also shows the manner in which the
flow rate of the feed was gradually ramped to the rate of 1.1 ml/min while keeping the
temperature constant at an intermediate setpoint of 130°C. After the flow reached steady

state, the temperature was ramped to the final set point of 220°C. Table 2.5 summarizes

the results of this experiment.

41

Table 2.5: Summary of Results for Extended Reaction - Experiment 50C50.

 

Date

Reaction temperature
Reaction time

Initial reactor charge
Lactic acid (g)

Cesium lactate (g)
Lactic-to-lactate mole ratio
Continuous feed

Lactic acid (g)

Cesium lactate (g)

Feed addition start
Feed time (min)

Feed flow rate (ml/min)

Product( grams )
2,3-Pentanedione

Acrylic acid

Acetaldehyde

Propanoic acid

Liquid product (Overflow)( g)

Conversion
Selectivity(2,3P)
% Theoretical yield (2,3P)

7/ 17/98
220°C
24 hr

36.10
8.05
21.7

1600
120
130°C
1410
1.22

46.3
69.5
19.6
7.73
343

26.8

37.3
9.7

42

 

80

 

 

 

 

 

 

 

 

 

 

+ Acetaldehyde
70 a -— 2,3-Pentanedione
—I— Propanoic Acid

60 2 +Acrylic Acid m 4 fl
3
01)
:5: 50 ~
‘6
E
3 40 a
C
0
U
8 30 -
‘é’
a.

20 ~

10 i

|
0 i .
0 5 10 15 20 2

LII

Time (hours)

Figure 2.4: Plot of Product Concentrations vs. Time for Extended Reaction.

43

EB: 29 Scam cam 33 59:0

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o
w w m m m w w w m w 0
in:
..... 41-11.. 1h”1!1r . . .. x . . om.P
..-. - - I 4-2.1.: . t;
£1186
) -- -- «No
mm L
mmmm lmwd
mem l .
smog. «we
.mrmm 1 .
Mmmfl 8°
C OI
ummm loco
RSCP
s-++ _ i as:
86
0 O
a a m m m a m a
Gov 9293th

   

Time (hrs)

Plot of Reaction Conditions vs. Time for Extended Reaction.

Figure 2.5:

44

2.3.5 Material Balances

2.3.5.1 Mass Balance

To verify the quality of reactions and analyses an overall mass balance was
calculated for one experiment. Initial and final weights were measured to determine gain
of weight of each component of the reaction apparatus. The weight gain was added to the
mass of products collected. The reactor weight gain represents the amount of product
remaining in the reactor. The effective volume of the reactor was about 40 ml. The
density of the reactor residue, calculated by dividing the amount of reactor residue by the
reactor volume, was approximately 1.1 g/lt. Table 2.6 shows the overall mass balance
calculations. 1R meters measured CO and C02 concentration in the exit gases. The gas
yields were calculated and were added to the product mass. The results show that the

mass balance closes accurately for this reaction.

45

Table 2.6:

Mass Balance Calculations for Experiment 50C51.

 

Mass Balance

 

 

 

 

Initia1(g) Final(g) Gain

Initial charge Lactic 36.07 g Reactor 266.13 309.76 43.63 g
Cesium Lactate 8.05 g Thermocouple 10.44 10.51 0.07 g
Continuous Feed Lactic 400 g Feed Tube 12.97 13.01 0.04 g
Cesium Lactate 30 g Gas Inlet 25.95 26.60 0.65 g
Condenser 96.51 96.73 0.22 g
Tubing #1 81.72 82.13 0.41 g
Tubing #2 36.84 36.87 0.03 g
Total Product(g) 318.31 g
Overflow (g) 108.43 g
CO and CO; 6.99 g

(g)
IN Total Mass 474.127 g OUT Total Mass 478.79 g
Gain 4.66 g
% Recovered 100.98

2.3.5.2 Carbon Balance

The carbon balance calculations for experiment 50C51 is shown in Table 2.7. To
account for all the carbon in the system, CHN analysis of reactor residue, liquid product,
and vapor product was done. The CHN analysis was done at the Department of Civil and
Environmental Engineering at Michigan State University. Carbon enters the reactor as
lactic acid or cesium lactate catalyst. The initial reactor charge consists of 36 g of 88%
lactic acid supplied by Purac, Inc. and 8 g of cesium lactate catalyst prepared by mixing
equimolar amounts of cesium hydroxide and lactic acid. The 88% lactic acid was diluted

to 50% for continuous feed using HPLC grade water. The amount of cesium lactate

46

catalyst in the continuous feed was 7.5g/ 100 g of feed. The carbon balance closes very

satisfactorily to within 3.5%.

The weight percentage of carbon, both in the reactor residue and the liquid

product, was about 38%. This is a validation of the steady state and also shows that the

reactor is well mixed.

 

 

 

 

 

 

 

Table 2.7: Carbon Balance Calculations for Experiment 50C51
Carbon Balance
In Out
Initial reactor charge grams moles g carbon Reactor residue (g) 44.4
Lactic Acid (88%) 36.08 0.35 12.69 %C from CHN analysis 38.98%
Cesium Lactate 8.05 0.016 0.58 Total carbon (g) 17.30
Continuous feed Liquid product (g) 108.43
Lactic Acid (50%) 400 2.22 79.93 %C from CHN analysis 38.12%
Cesium Lactate 30 0.061 2.17 Total carbon (g) 41.33
Vapor Product (g) 318.97
%C from CHN analysis 12.56%
Total carbon (g 40.063
Total Carbon (g) 95.38 Total Carbon (g) 98.70
Difference % -3.49

2.3.5.3 Hydrogen Balance

The CHN analysis data used for the calculating the carbon balance was further

used to calculate the hydrogen balance. These balances are not as accurate as they take

into account the large amounts of water in the system. Table 2.8 shows the hydrogen

balance calculations for experiment 50C51.

47

Table 2.8: Hydrogen Balance Calculations for Experiment 50C51.

 

 

 

 

 

 

 

 

Hydrogen Balance
In Out
Initial reactor charge grams moles g hydrogen Reactor residue (g) 44.4
Lactic Acid (88%) 36.07 0.35 2.117 %H from CHN analysis 4.80%
Cesium Lactate(44%) 8.05 0.016 0.081 Total hydrogen (g) 2.13
Water(acid + catalyst) 8.78 0.48 0.976 Liquid product (g) 108.43
%H from CHN analysis 6.42%
Continuous feed Total hydrogen (g) 6.96
Lactic Acid(50%) 400 2.22 13.32 Vapor Product (g) 318.97
Cesium Lactate(44%) 30 0.06 0.30 %H from CHN analysis 8.40%
Water(acid + catalyst) 216.5 12.03 24.06 Total hydrogen (g) 26.79
Total hydrogen (g) 40.86 Total hydrogen (g) 35.88
Difference % 12.18

2.3.5.4 Comparison of Product Analyses

CHN analysis was done on condensed vapor product samples for several
reactions. Experiment 50C42 was conducted at 240°C, 50C46 at 220°C, and 50C51 at
200°C. Table 2.9 shows the comparison between the carbon and hydrogen content
determined from CC and CHN analysis of the condensed vapor product. The carbon
content determined is similar from these two techniques. It should be noted that the
concentration of lactic acid determined from gas chromatography has an error of 114%.

This could account for the higher carbon % obtained from GC.

48

Table 2.9:

Comparison of Product Analysis by GC and CHN Analysis.

 

 

 

50C42 P4 50C46 P4 50C51 P4

LIQ GC g/lt mol/lt g/lt mol/lt g/lt mol/lt
Acetaldehyde 42.4 0.96 15.1 0.34 3.3 0.08
2,3-Pentanedione 64.3 0.64 27.0 0.27 19.9 0.2
Propanoic acid 10.9 0.15 4.2 0.06 10.2 0.14
Acrylic acid 71.9 0.99 42.9 0.60 97.0 1.36
Lactic acid 203. 2.26 378.1 4.20 215.9 2.4

Water 605.1 33.61 531.8 29.55 652.3 36.24

(difference)

CHN GC CHN GC CHN GC

Carbon % 15.46 18.54 14.43 19.96 12.56 15.43
Hydrogen % 8.54 9.49 8.56 9.07 8.40 9.51

Oxygen 75.05 71.97 76.32 70.97 78.38 75.06

100.00 100.00 100.00

 

 

 

 

2.3.6 CO and C02 Production

The accuracy of the reaction procedure and product analysis can be checked by

the comparison of carbon dioxide and carbon monoxide yields obtained by gas analysis

using IR meters to theoretical yields. The moles of product formed at steady state were

calculated from the product concentration obtained from Gas Chromatography. Dividing

by reaction time gave the average rate of formation of each product in units of moles/sec.

Figure 1.2 shows the reactions of lactic acid. From reaction (1) we can see that

for every mole of acetaldehyde formed either one mole of carbon monoxide or one mole

of carbon dioxide is formed. Further, for every molecule of 2,3-pentanedione formed one

mole of C02 is formed. While the reaction temperature was at steady state, 7.31g of 2,3-

pentanedione was formed. This corresponded to an average CO; production rate of 5.15

x10'° mol/sec.

Table 2.10 shows the calculation of CO and CO; flow rates. The amount of CO
and CO; in the vapor product was calculated from the percentage compositions
determined from the 1R meters. The total gas flow rate was 55 ml/min. Average
percentage of CO and CO; observed was 7.7 % and 22.6 % respectively. This
corresponded to a flow rate of 3.17 x10’°mol/sec for CO and 9.24 x10'° mol/sec for CO;.
By equating the theoretical CO formation from acetaldehyde to the CO flow rates
obtained from the IR meter, the CO; formed from the conversion of lactic acid to
acetaldehyde (6.34 x10'6 mol/sec) can be determined by difference. Adding the amount
of CO; produced from the formation of 2,3-pentanedione from lactic acid (5.15 x10'°
mol/sec) gave the CO; production from the two reactions (1.15 x10'5 mol/sec). This is

comparable to the total flow rate of CO; detected by the IR meter (9.24 x10'°).

Table 2.10: Comparison of Measured CO and CO; Flow to Theoretical Flow Rates

Estimated from Steady State Product Concentrations.

 

 

Theoretical Rate From CO and CO; meters
C0 C02
moles moles/sec moles/sec moles/sec CO CO;
From 2.3? 0.073 5.15 )110'6 0 5.15 1110'6 Avg % 7.7 22.6
Acetaldehyde 0.134 9.51 )110"5 3.17 x10'° 6.34 1110'6 moles/sec 3.17 x10'° 9.24 1110‘
Total 3.17 x106 1.15 1110'5

 

 

50

CHAPTER 3

OPTIMIZATION AND KINETIC ANALYSIS

3.1 Optimization of Operating Conditions

After the design of the continuous reactor system was completed, a study to find
the optimum operating conditions was conducted. Three variables were identified as key
parameters affecting reactor performance - temperature (x1), feed flow rate (x;) and
catalyst concentration in the continuous feed (x3). Optimization involved varying these
three different variables at three different values. This would require 27 experiments.
However, using a two level, three-parameter factorial design designated as a “Box-

Benkhen” design (41), the number of experiments was reduced to 15.

3.1.1 Box-Benkhen design of Experiments

Implementation of the Box-Benkhen experimental design involved identifying
“boundary values” (maxima and minima) for each of the independent variables, and then
conducting experiments in the factor space defined by the boundary values. The range of
these experiments can be described by a cube, the coordinates of the comers representing
the range of the variables. Figure 3.1 shows a schematic representation of the Box-
Benkhen design of experiments. The center of the cube (0,0,0) is the base set of the

conditions. Twelve experiments were conducted at coordinates corresponding to the

51

centers of the cube edges and three at the base conditions. These experiments were
conducted in a random order to prevent any systematic error. The reaction conditions are

listed in Table 3.1.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

X
‘ 3
X2
an 1
«w .»
4 " .K > X1
,fi -' @ .
V

F1 gure 3.1: Schematic Representation of Box-Benkhen Design.

Table 3.1: Box-Benkhen Design Values.

 

Box-Benkhen design Values

 

Parameter Range -1 0 1
Temperature (°C) 200 — 240 200 220 240
Flow Rate (ml/min) 0.6 - 1.5 0.6 1.1 1.5
Catalyst Concentration ‘ 5 - 10 5 7.5 10

 

" catalyst concentration in continuous feed - g cesium lactate/100 g of 50% aqueous

lactic acid feed

52

3.1.2 Experimental Method

All reactions were carried out in the 50 ml reaction flask described in Section 2.2.3
and were started at the same initial conditions. The initial reactor charge consisted of 36 g
of lactic acid and 8 g of cesium lactate catalyst, which corresponded to a lactic-to-lactate
molar ratio of 22: 1. While the continuous feed used in all the experiments was a 50%
aqueous solution of lactic acid, the catalyst concentration in the continuous feed was
varied. The catalyst used was the 1:1 molar cesium lactate prepared as described in
Section 2.2.2. The protocol for the temperature ramp and feed addition described in
Section 2.2.4 was implemented consistently in the entire set of experiments.

The reaction was started at room temperature and the reaction mixture was heated
to 130°C, and maintained at 130°C for 30 minutes while the feed addition-was started.
The feed rate was started at about 0.3 ml/min and gradually increased to the steady state
feed rate (0.5 -—1.5 ml/min.) The length of the experiments varied from 5 to 7 hours. An
extended reaction of 24 hours was also conducted to verify the establishment of steady
state in the reactor system. The results of these experiments are listed in Table 3.2.

The product concentrations were determined using gas chromatography. The
lactic acid used as initial reactor charge was included in the calculation of the yields of
2,3-pentanedione and acrylic acid that are reported. The product concentrations used in
the kinetic analysis were calculated by taking an average of the steady state product

concentrations determined by gas chromatography.

53

 

 

Table 3.2: Results of Optimization Experiments.
ID Temp Flow Cat Time Conv.b Select. ° Yield 2,3P ° Yield
(°C) (ml/min) Cone. " (hr) % % % Acrylic ° %
50C37 200 1.21 10 7 16.7 32.4 5.2 6.1
50C38 220 0.69 5 7 10.1 34.8 3.4 3.6
50C39 220 0.63 10 6 32.2 47.6 14.8 8.5
50C40 200 0.65 7.5 6 17.7 51.1 8.8 4.6
50C4l 220 1.13 7.5 6 25 28.9 7 7.8
50C42 240 1.48 7.5 5 36.6 35 12.4 12.8
50C43 240 1.12 10 6 47.2 42.3 19.2 7.7
50C44 240 0.71 7.5 5 23.5 60.7 13.2 5.0
50C45 220 1.18 7.5 6 15.4 51.8 7.8 4.0
50C46 220 1.49 5 5 15.2 36.7 5.4 5.1
50C47 240 1.13 5 5 31.5 43.9 13.5 7.8
50C48 220 1.47 10 5 26.6 34.2 8.8 10.0
50C49 200 1.18 5 6 13.7 35.6 4.8 5.0
50C50 220 1.22 7.5 24 26.8 37.3 9.7 10.2
50C51 200 1.45 7.5 5 33.4 21.7 6.46 14.9

 

3 Catalyst concentration in continuous feed g cesium lactate/ 100 g of feed (50% aqueous

lactic acid)

°Conversion = (moles lactic acid converted / moles lactic acid fed)*100

° Selectivity = (moles lactic acid converted to 2,3P/ moles lactic acid converted) *100

° % theoretical yield 2,3-pentanedione = (grams 2,3P formed / grams 2,3P formed if all

lactic acid fed was converted to 2,3P)* 100

e % theoretical yield acrylic acid = (grams acrylic formed / grams acrylic formed if all

lactic acid fed was converted to acrylic)* 100

54

3.1.3 Results and Discussion

The product yields of 2,3-pentanedione and acrylic acid, conversion, and

selectivity obtained from these experiments were fit into a quadratic model of the form

Y=bo + brxr + bzxz +b3x3 + b12X1X2 + b13X1X3 +b23X2X3 + 1311102 + b221122 + 1333192.
where x], x;, and x3 are the three independent variables temperature, the feed flow rate,
and catalyst concentration in the continuous feed. The coefficients b.- and bir‘ were
determined using the regression analysis tool in Microsoft Excel 7.0.

The regression was repeated after discounting experiment 50C45. This greatly
improved the fit and much higher R2 values were obtained for regression of acrylic acid
yields and conversion. The results of the regression analysis are listed in Table 3.3 to
3.6. The corrected coefficients reported are the results of the regression analysis after
discounting experiment 50C45. A R2 value of 0.91 was obtained for the regression of
2,3-pentanedione yield. The R2 value for lactic acid conversion was 0.71. A confidence

level of 0.95 was used for all regressions.

55

Table 3.3: Regression Statistics and Coefficients for 2,3-Pentanedione Yield.

 

 

2,3-Pentanedione Corrected
yield
Multiple R 0.95 0.95
R Square 0.91 0.91 '
Adjusted R Square 0.74 0.70
Standard Error 2.25 2.51
Observations l 5 14

 

 

Coefficients

 

 

 

 

 

 

 

Intercept 283.05 270.69
x, -2.67 -2.57
x; -1.78 -1.55
x,*x; 0.10 0.10
X1*X3 0.02 0.02
X2*X3 -2.22 -2.23
X1*x1 0.01 0.01
x;*x; -4.19 -4.79
x3*x3 0.03 0.02
Table 3.4: Regression Statistics and Coefficients for Conversion of Lactic Acid.
- Conversion of lactic acid Corrected
Multiple R 0.85 0.87
R Square 0.72 0.75
Adjusted R Square 0.20 0.19
Standard Error 9.12 9.22
Observations 1 5 l 4
Coefficients
Intercept 718.8 473 .0
x. -6.75 -4.74
x; 16.13 37.72
x, -4.52 0.05
x,*x; 0.07 0.09
x,*x3 0.06 0.06
X;*X3 -3.20 -3.43
x,*x, 0.01 0.01
x;*x; 0.59 -11.37
X3*X3 ~0.15 -O.42

 

56

Table 3.5: Regression Statistics and Coefficients for 2,3-Pentanedione Selectivity.

 

 

 

 

 

Reaction selectivity Corrected
Multiple R 0.77 0.86
R Square 0.60 0.74
Adjusted R Square 013 0.14
Standard Error 10.65 9.06
Observations 1 5 l4
Coefficients
Intercept 150.8 588.5
x, -l.52 -5.10
x; -0.58 -39.03
x; 10.74 2.59
X1*x; 0.09 0.04
x,*x3 0.00 0.00
X;*x3 -3.76 -3.35
x1*x, 0.00 0.01
X2*X2 -4.98 16.31
xfx3 -0.39 0.09

 

Table 3.6: Regression Statistics and Coefficients for Acrylic Acid Yield.

 

 

 

 

 

Acrylic acid yield Corrected

Multiple R 0.75 0.80

R Square 0.56 0.64

Adjusted R Square -0.23 -0.18

Standard Error 3.66 3.56
Observations 1 5 14

Coefficients

Intercept 93 .4 -21 .4

x, -1.04 -0.10

x; -l .74 8.34

X3 5.47 7.61

x1*x; -0.07 -0.06

x1*x3 0.00 0.00

X2*X3 -O.14 -O.25

x1*x, 0.00 0.00

X2*X2 11.79 6.21

X3*X3 -O.26 -O.39

 

57

The regression coefficients obtained were used to generate contour plots of the
product yields, conversion, and selectivity. These plots helped identify trends in reactor
performance as reaction parameters are varied relative to each other. The contour plots
were extremely important in interpreting the results of the experiments, as trends may be
difficult to Spot when varying three parameters. It is important to note that the trends

identified are only valid for the factor space defined by the Box-Benkhen design.

3.1.3.1 2,3-Pentanedione Yield

Figure 3.2 shows the contour plot for the yield of 2,3-pentanedione at a
temperature of 220°C. As expected, the yield increased as the concentration of catalyst
in the feed was increased. At low catalyst concentrations, 2,3-pentandione yield
increased as feed flow rate was increased. Experiments 50C38 and 50C46 were
conducted at 220°C and feed catalyst concentrations of 5 g Cs Lactate/100g of feed.
Yield of 2,3-pentanedione increased from 3.4% at a flow rate of 0.6 ml/min to 5.4% at a
flow rate of 1.5 mein.

This trend was reversed at higher catalyst concentrations. Experiments 50C39
and 50C48 were conducted at 220°C and feed catalyst concentration of 10 g Cs
Lactate/ 100g of feed. A yield of 14.8% was obtained at a feed flow rate of 0.6 mllmin
and a lower yield of 8.8% was obtained at a flow rate of 1.5 mllmin. This lowering in
yields at high flow rates can be explained by the lower residence time in the reactor.

More feed lactic acid may be flowing out unreacted as liquid product.

58

Temperature = 220 °C

% Theoretical yield

 

Flow rate(ml/min) b 5
0' Catalyst Concentration in feed

(g CsLactate/lOOg feed)

Figure 3.2: Contour Plot of 2,3-Pentanedione Yield at 220°C.

Figure 3.3 shows the variation in 2,3-pentanedione yields at a feed flow rate of
0.6 ml/min. For a given catalyst concentration, the yield increased with increase in
temperature. Experiments 50C40 and 50C44 were conducted at a catalyst concentration
of 7.5 g Cs Lactate/ 100 g of feed and a feed flow rate of 0.6 mllmin. As reaction
temperature was increased from 200°C to 240°C, the yield of 2,3-pentanedione increased
from 8.8% to 13.2%.

Similarly, at any temperature, the yield increased with increase in catalyst

concentration in the feed. Experiments 50C38 and 50C39 were conducted at a

59

temperature 220°C and a feed flow rate of 0.6 ml/min. Yield of 2,3—pentanedione

increased from 3.4% to 14.8%, as the catalyst concentration was increased from 5 to 10

g Cs Lactate/100 g of feed.

Feed Flow rate = 0.6 mllmin

 
 
  
  

 

 

251
g 20»
3.
E 15.
1.3
O
Q.)
.C.‘
1-
15°

 

 

Catalyst Concentration in feed
(g CsLactate/ 100g feed)

Figure 3.3: Contour Plot of 2,3-Pentanedione Yield at Flow Rate of 0.6 ml/min.

60

3.1.3.2 Acrylic Acid Yield

Figure 3.4 shows the contour plot of acrylic acid yield at a temperature of 240°C.
The plot is in the shape of a saddle. The yield of acrylic acid increased with increase in
flow rate for any given temperature and catalyst concentration. This trend continued
even as the temperature was increased. Experiments 50C42 and 50C44 were conducted
at a temperature of 240°C and 7.5 g Cs Lactate/ 100g of feed. The yield of acrylic acid
increased from 5% to 12.8%, as the flow rate was increased from 0.6 mllmin to 1.5

ml/min. Clearly, the formation of acrylic acid was favored by shorter residence times.

Temperature = 240 °C

 

 

'0
13
>»
Te
.2
’2'
o
4)
i ‘ * D
0 vi” '1 1
5 2.1 if! 1 \\~\
\\ ~ .0
0 \r\ . \\\\\“\\ \\:\\.r
\ \\\ \ \x"
‘ \b‘ 0: \\\'\\ \l\ § 10
\ \r)’ \ \‘\‘,\\\\\X\Y 8 9
\ \ Qq \\\\//>1 7
Flow rate(ml/min) Q' Q’-\ (55 6
0‘ Catalyst Concentration in feed
(g CsLactate/ 100g feed)

Figure 3.4: Contour Plot of Acrylic Acid Yield at 240°C.

61

The effect of temperature on acrylic acid yield can be seen in Figure 3.5, the
contour plot for the yield of acrylic acid at a feed flow rate of 1.1 mllmin. At a given
catalyst concentration, the yield of acrylic acid increased with increase in temperature.
Experiments 50C37 and 50C43 were conducted at a flow rate of 1.1 ml/min and catalyst
concentration of 10 g Cs Lactate/100 g of feed. The yield of acrylic acid increased from
6.1% at 200°C to 7.7% at 240°C.

A similar trend was seen at a lower catalyst concentration of 5 g cesium
lactate/100g of feed. The yield for acrylic acid increased from 5.6% to 7.8% when
reaction temperature was increased from 200°C (Experiment 50C49) to 240°C

(Experiment 50C47).

Feed Flow rate = 1.1 ml/min

 

  
 

% Theoretical yield

f . " 220
. , ’ r , ’ 230 Temperature °C

 

Catalyst Concentration in feed
(g CsLactate/100g feed)

Figure 3.5: Contour plot of Acrylic Acid Yield at a Flow Fate of 1.1 ml/min.

62

3.1.3.3 Conversion of lactic acid

The conversion of lactic acid was computed from the fraction of lactic acid
converted to condensible vapor products. No attempt was made to account for the lactic
acid converted to polylactic acid by intermolecular esterification. It was assumed that
the formation of oligomers of lactic acid was reversible by hydrolysis, and unreacted feed
recovered as liquid product could be reprocessed.

Figure 3.6 shows the contour plot for conversion of lactic acid at a temperature of
240°C. At a given flow rate, the conversion increased as the catalyst concentration in the
feed increased. Experiments 50C38 and 50C39 were conducted at 220°C and a feed
flow rate of 0.6 ml/min. The conversion increased from 10.1% to 32.2% when the
catalyst concentration was changed from 5 g to 10 g Cs Lactate/ 100 g of feed.

Figure 3.6 shows that for a given catalyst concentration, the conversion increased
as flow rate was increased. Experiments 50C40 and 50C51 were conducted at 200°C
and a catalyst concentration of 7.5 g Cs Lactate/100 g of feed. As the flow rate was
increased from 0.6 ml/min to 1.5 mllmin, the conversion increased from 17.7% to
33.4%. Similarly, experiments 50C44 and 50C42 were conducted at 240°C and a
catalyst concentration of 7.5 g Cs Lactate/ 100 g of feed. The conversion increased from

23.5% to 36.6% as the flow rate was increased from 0.6 to 1.5 ml/min.

63

Temperature = 240 °C

 

 

 

::
.9
12
G)
>
c:
o
D
88
\‘ p ‘7 ’7.-
Flow rate(m1/min) 9' g2 b 5 6
9' Catalyst Concentration in feed
(g CsLactate/lOOg feed)

Figure 3.6: Contour plot of Lactic Acid Conversion at 240°C.

Figure 3.7 shows the contour plot for lactic acid conversion at a flow rate of 1.1
mein. As expected, the conversion increased with temperature. Experiments 50C37
and 50C43 were conducted at a flow rate of 1.1 mllmin and a catalyst concentration of
10 g Cs Lactate/ 100 g of feed. The conversion almost tripled from 16.7% to 47.2% as

the reaction temperature was increased from 200°C to 240°C.

64

Feed Flow rate = 1.1 ml/min

 

 

 

 

 

% Conversion

 

 

 

 

 

Catalyst Concentration in feed
(g CsLactate/lOOg feed)

Figure 3.7: Contour Plot of Lactic Acid Conversion at a Flow Rate of 1.1 ml/min.

3.1.3.4 Selectivity

The effect of reaction conditions on the yields of 2,3-pentanedione and acrylic
acid has been shown in Section 3.1.3.1 and 3.1.3.2. Figure 3.8 shows that the
conversion of lactic acid is most selective to the formation of 2,3-pentanedione at high
temperatures and low flow rates. The highest selectivity of 60% was obtained for
Experiment 50C44. The reaction was conducted at 240°C, at a feed flow rate of 0.6

mllmin and a catalyst concentration of 7.5 g Cs Lactate/100 g of feed.

65

Catalyst concentration
7.5 g CsLactate/lOOg feed

 

 

DJ
o
%Selectivity

“i ~ \ 20
o - = “
”1° ~ . . 10
q, .. / .- /”'/ \ xi” \Y-
0 ,-/'.’" . /«r"’~i 5~ // 0
Temperature C ,9? / Erie/5" yyzr .
Q i y/ . l/f ,,-// 0 8 0.6
'9 T 1.2 l
g 1.4 Fl ml/ .
“12°“ ow rate nun

Figure 3.8: Contour Plot for Selectivity to 2,3-Pentanedione Formation at a Catalyst

Concentration of 7.5g Cs Lactate/100g of feed.

66

3.2 Kinetic Analysis

3.2.1 Continuous Stirred Tank Reactor

The reactor was modeled as a continuous stirred tank reactor, as shown in Figure
3.9. Stream (0), liquid feed, a homogenous mixture of 50% aqueous lactic acid and
cesium lactate catalyst, was fed to the reactor. The reactor was heated and well stirred.
The reaction apparatus is described in Section 2.2.3. The product leaves the reactor in
two streams; stream (1) consisted of the condensable vapor product including, 2,3-
pentanedione, acrylic acid, propanoic acid, and acetaldehyde. Stream (2), liquid product,
contained unreacted lactic acid, lactic acid oligomers, cesium lactate catalyst, and water.
Product concentrations were determined by gas chromatography. It was assumed that no
condensable vapor product was present in the liquid product stream. Gas

chromatography on diluted samples of the liquid product has validated this assumption.

67

V19 CLI

 

 

 

 

 

Vapor Product
VovCLo
v ) V2, CI-Q
Liquid feed ’ 1:123“:
V = Volume of Reactor (lt.) i=0 Inlet
v, = Volumetric flow rate (lt./sec) i=1 Vapor Product
CL,= Concentration of Lactic Acid (mol/lt.) i=2 Liquid Product (Overflow)

CP,= Concentration of 2,3 Pentanedione (mol/lt.)
CA,= Concentration of Acrylic Acid (mol/lt.)
C8,: Concentration of Acetaldehyde (mol/lt.)
CCs,= Concentration of Cesium Lactate (mol/lt.)

Figure 3.9: Schematic Representation of Reactor Model.

68

 

 

11

i 1.’ 7'
"Way

3.2.2 Reaction System

The primary reactions taking place in the reactor are listed below in Figure 3.10.
Reaction 1, the thermal decomposition of lactic acid to acetaldehyde was modeled as a
first order reaction. The condensation of lactic acid to 2,3-pentanedione in the presence
alkali lactate was modeled as a second order reaction, with first order dependence on the
concentrations of lactic acid and cesium lactate. The formation of acrylic acid from
lactic acid was a first order reaction. The reaction system also included first order
decomposition reactions for acrylic acid and 2,3-pentanedione. However, the results of
vapor phase studies showed that the decomposition of 2,3-pentanedione was not
significant at the reaction temperatures used in the condensed phase experiments (3).
Therefore, the decomposition of 2,3-pentanedione was listed as a possible reaction
pathway but not included in the calculations. Each reaction was assumed to be

irreversible with a rate constant k, i =1 — 5.

Lactic acid —'—" Acetaldehyde + Products (1)

Lactic acid + (Cesium) Lactate—b 2,3-Pentanedione (2)
Lactic acid ——-> Acrylic acid (3)

Acrylic acid————> Products (4)
2,3-Pentanedione——> Products (5)

Figure 3.10: Primary Reactions.

69

3.2.3 Method

The steady state molar balance for each species is given by the general
expression-
Moles in = Moles out + Disappearance by reaction (3.1)
Using the rate equations for each reaction and molar balances, an expression for

the steady state exit concentration of each species was obtained. These expressions are

 

listed below. The flow rates of the two product streams, v1 and v3, were calculated by
dividing the amount of product collected by total collection time.

k *CL. *V
Acetaldehyde (mol/lt), CB 1 = ’ ‘ (3.2)

v1

 

_k2 *CL2 *CCs, *V

 

 

2,3- ntanedione mol/lt , CP 3.3
W ( ) ‘ v, + k,v ( )
k,*CL,*V
Acrylic acid (mol/lt), CAl = ‘ ' (3.4)
v1 + k4V

The formation of decomposition products of acrylic acid was modeled by a first
order rate expression. The concentration of acrylic acid was adjusted to represent the
concentration inside the reactor, as it was assumed that the decomposition reaction took
place in the condensed phase.

k4 *CA, *V
(v,+v,)

 

Decomposition products of acrylic acid = (3.5)

70

The expressions for product concentration were entered into an Excel spreadsheet.
Using the Solver fiinction, the exit concentrations were calculated by simultaneously
varying k1 through k4, minimizing the sum-of-squares residuals for conversion and species
exit concentration reactions at a given temperature. The concentration residual was the
difference of the concentration predicted by the model from the average steady state
product concentration measured by gas chromatography. This process was repeated for

the reaction data at two other temperatures.

3.2.4 Lactic Acid Concentration

The lactic acid concentration used to compute the product concentrations in the
above expressions was very important to the success of the model. Since there was no
way to accurately measure the lactic acid concentration inside the reactor at steady state,
the reactor model was used to find the lactic acid concentration theoretically.

Figure 3.9 shows a schematic of reactor model A. This method involved using the
lactic acid concentration in the liquid product for the kinetic analysis. The concentration
of lactic acid was calculated by doing a water balance over the system. Knowing the
amount of water in the liquid product and the density, the amount of lactic acid was
calculated by difference. The concentration calculated cannot be used directly because
literature suggests that lactic acid exists in equilibrium with its intermolecular esters.
Moreover, an initial attempt of using the concentration of lactic acid in the liquid feed

(CL;) to predict product concentrations did not fit the experimental data. The results of

71

this model are listed in Appendix D. It was clear that an estimate of the amount of free
lactic acid in the reactor, that was available for chemical reaction, was needed.

Ale’n and SjoStrOm (42) determined the amount of lactic acid that is converted to
its linear and cyclic dimer, trimer etc. upon heating. The results of their experiments are
listed in Appendix B. The amount of dimer present in the solution remained constant after
about 6 hours of heating. This was assumed to be the equilibrium distribution for the
reversible reaction of lactic acid to its linear dimer, lactyl lactic acid. The concentration
data at 100°C and 125°C was used to calculate equilibrium constants and the heat of
formation of the dimer. The equilibrium constant at reaction temperature was calculated
using van’t Hofl‘s relation. Next, the amount of free lactic acid was calculated using
estimates of the water and lactic acid present in the liquid product. These calculations are

summarized in Table 3.7.

72

 

 

Table 3.7: Estimate of Free Lactic Acid in the Reactor Using Reaction Data of Alén
and Sjéstrém (42).
Exp # Coordinates Water; CD; Free Lactic acid
(g/lt) (mol/lt) (g/lt) (mol/lt) (mol/lt) % of CL;
50C37 (-l,0,1) 357.9 19.9 930.7 10.3 8.06 78.0
50C40 (-1,-1,0) 458.9 25.5 880.0 9.7 7.99 81.8
50C49 (-1,0,-1) 510.5 28.4 812.5 9.0 7.57 83.9
50C51 (-1,l,0) 389.6 21.6 1018.0 11.3 8.81 77.9
50C38 (0,-1,-1) 279.9 15.5 1037.4 11.5 7.58 65.7
50C39 (0,-l,1) 184.8 10.3 1093.6 12.1 7.22 59.4
50C41 (0,0,0) 160.0 8.9 1161.7 12.9 7.34 56.9
50C45 (0,0,0) 1065.5 59.2 310.6 3.44 3.24 93.8
50C46 (0,1,-1) 577.9 32.1 757.7 8.41 6.75 80.2
50C48 (0,1,1) 668.7 37.1 688.9 7.64 6.36 83.2
50C50 (0,0,0) 555.6 30.9 827.1 9.18 7.22 78.6
50C42 (1,1,0) 351.4 19.5 937.1 10.40 6.57 63.2
50C43 (1,0,1) 556.2 30.9 799.2 8.87 6.43 72.5
50C44 (1,-1,0) 742.9 41.3 664.8 7.38 5.84 79.1
50C47 (1 ,0,-1) 770.9 42.8 593.2 6.59 5.34 81.1

 

Similarly, the steady state concentration of the catalyst inside the reactor was not

known. However, as the catalyst concentration in the feed is known, and assuming all

the catalyst entering the system at steady state leaves the reactor as liquid product, we

can compute the steady state catalyst concentration in the reactor by the formula

CCs2 =

V2

V0 *CCSO

73

(3.6)

 

V1, CL1

Va or Product
V0,CL0 v’, CLAV p

 

 

Liquid feed ___., Liquid Product
V2, CL2

Figure 3.1 1: Schematic Representation of Alternate Reactor Model.

Another scheme examined is shown in Figure 3.11. At steady state, the
concentration of lactic acid inside the reactor will be the same as CL... CLaw can be
calculated from the average of the concentration of lactic acid in the two product streams.
CLl, the concentration of lactic acid in the condensed product, was obtained from gas
chromatography and CL; is obtained from an overall component balance for lactic acid.
The amount of lactic acid in the liquid product is the difference between the amount of
lactic acid fed to the reactor (CL) and the amount of lactic acid that is converted to
condensable vapor product. The lactic acid present in the vapor product, (CL;) is also
accounted for. From the component balance for lactic acid:

__ vO *CLO —v,(CLI +2*CP, +CAl +CB,)

 

 

CL2 (3.7)
V2
Thus, CLav can be calculated from the expression below.
L , L.
CL... = V‘C ‘+V'C ‘ (3.8)
VI + v2

The results of this model are listed in Appendix C.

74

3.2.5 Results

The rate constants for reactions 1-4 obtained from the EXCEL solver routine for
the kinetic analysis are listed in Table 3.8. After obtaining rate constants at three
temperatures, an Arrhenius plot (1n k, vs. 1/T) was made to determine the activation

energy and preexponential factor.

 

 

Table 3.8: Rate Constants at Different Reaction Temperatures.
T(I() 473.15 493.15 513.15
kl (sec'l) 1.031305 1.351305 3.88E-05
k2 (liter mot1 sec") 1.65E-05 2.18E-05 3 .80E-05
k3 (sec’l) 4.61E-05 4.84E-05 73415-05
k4 (sec'l) 29113-04 2.96E-04 34313-04 .

 

Figure 3.12 —3. 15 show the Arrhenius plots for k1, k;, k;, and k4 respectively. The
kinetic constants shown in Table 3.9 were calculated from the line fit of the rate constants.
Slope of the line equals -E,/R, where Ea is the activation energy of the reaction and K the
gas constant. The intercept of the line fit equals the natural logarithm of the

preexponential factor.

75

Table 3.9: Kinetic Constants Calculated from Arrhenius Plots.

 

 

Reaction Preexponential factor Activation
(kto) energy(J/mol)
Lactic acid ——> Acetaldehyde + Products 1.83E+02 soc“l 66177.4
Lactic acid + (Cs)1actatc——+ 2,3-P 6.26E-01 liter motl sec" 41620.8
Lactic acid ——+ Acrylic acid 3.011302 sec'1 25616.1
Acrylic acid —. Products 6.80E-03 sec'l 12310.2

 

76

 

 

-10 - > .
0.00194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214

1

-10.2 4
y = -7961.8x + 5.2128

R2 = 0.8803
-104 ,

-10.6 *

Ln k1

1

i

1

1

40.8 . 1
1

1

-11 1
1 1
-112. 1

-11.4a

-11.6«

 

-11‘8 J~—~ __.__-_.,-,_,,, ,_ —-- ,-__.--__. —-—-——-—-<—-- -- - ——~-——- —-«—~«-—~--—--~——— ——-——— —— —=.-

Figure 3.12: Arrhenius Plot for Reaction 1.

77

 

-101 . 5 . 5 =
0.00194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 000214
-10.2 ~
y = -5022x - 0.4455
-103 - F12 = 0.9558

 

-10.4 «

-10.5 r

-10.6 ‘

In k2

 

-‘|0.7 ‘

-10.8 r

-10.9 -

-11..

 

-11.1«

 

 

 

-11.2 —~ — ~—
1a

Figure 3.13: Arrhenius Plot for Reaction 2.

78

 

‘9.4 ’ I » ‘ a
0.00194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214

y = ~2789x - 4.1514

1

-9.5 ~ 2 I
R = 0.8074 1

l

-9.7 ~

In k3

-9.8 ~

-9.9 r

-10-

 

 

 

-10.1
1/T

Figure 3.14: Arrhenius Plot for Reaction 3.

79

 

 

-7.95

\

-8.05 r

\

In k4

-8.1 r

-8.15 -

 

 

 

 

-8.2 ----__.__-,-._,- ---, ---— =5 -
1/T

Figure 3.15: Arrhenius Plot for Reaction 4.

80

fl

0.00194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214

y = 977.9411 - 6.0965
R2 = 0.8118

3.2.6 Discussion

The kinetic model proposed for the condensed phase conversion did not predict
the concentration of products as accurately as hoped. The kinetic model reported was
reached after trying several reaction schemes. It is not possible to discuss the results of
the different possible combinations attempted. The kinetic model reported best
represented the reaction data collected. The experimental data used to calculate the

kinetic constants were reasonably accurate as the different balances shown in Section 2.3.5

 

close favorably.

Table 3.10 shows a summary of the results of the kinetic analysis. The difference
in the predicted and experimental concentrations could be explained by the fact only four
reactions were included in the model. Several other competing reactions may be taking
place, such as the formation of coke from lactic acid, the formation of propanoic acid etc.
For simplicity, the reactor was assumed to be an ideal reactor, nevertheless certain non-
ideal behavior such as short-circuiting of feed was possible.

Most importantly, the concentration of lactic acid inside the reactor at reaction
conditions was very difficult to measure or estimate. The extent of polymerization of
lactic acid was also not known. So, various component balances and equilibrium data
were used to get a reasonable estimate of the fiee lactic acid concentration at reaction

conditions.

81

 

 

 

 

—VUOm-# HCQECOQXD .000 Cogomfiuhm mEOENECUUCOU DDS—Up: 00: mQO—v Cotonou LOP—U OWN—00,64 a

 

 

82

0.00 4sm mew 9.04m

0.0m 0.0m 0._ _ Doom-o.

0.04 . 0.mm 0.0 9.00m 80383:.8 5082 0a cote owEo>< a

v.84 fimm 0.0m .0 Sta owEo><

00— 0.4m 0.: 3.00 400.0 m _ 0.0 044.0 004.0 000.0 2.0. _0 54000
~00 ..m0 0.2 34.0 w~40 0050 N400 >050 420.0 A0. _ -. .0 44000
0.40 00— 0.0m 40 _ ._ 000.0 55.0 09.0 000.0 004.0 2.0. .0 3000
0.00 :04 0.04 400. _ 20.. 040.0 _ .40 w _ 0.0 03.0 A0. _ . _0 N480
0.2 N: 55 N0m.0 30.0 0mm.0 _4m.0 3.0.0 _0m.0 8.0.8 00000
0.04 0.04 0.0m 24.0 000.0 momd 03.0 N440 mwmd 2. _ .00 04000
0.0. 0.4 0.0. mmmd 500.0 0m~0 m0m.0 404.0 03.0 2-. _ .8 04000
0.0m mam mm mm: .0 >~m0 0mm.0 m3 .0 04~0 wmmd 8.0.8 04000
0N4 00— 4.0 $00 m _ 5.0 nmmd 40m.0 m00.0 051.0 8.0.8 24000
Wmm >.0_ Em 34.0 000.0 0S0 00.0 05.0 00.0 2. _ -.8 0m00m
0.0m. 00> 0.0mm 0310 004.0 mmm0 >210 0mw.0 49.0 2-. _ -.00 me00
0.00 w.mm 0: wow .0 RN. 48.0 m0~.0 400.0 and A0. _ . _ -0 $000
m.m_ 0.0 0.0 m _ m0 054.0 44nd hm _ .0 400.0 m4m.0 2 -.0. _-0 04000
0.04 000 N0 _mm.0 0 _ 4.0 004.0 0400 N050 0N4.0 A0. _ -. _ -0 04000
0.04 4.m 0.0 N4m.0 _0m .0 >4N0 NON.0 N40 .0 >4m.0 2.0. _ -0 Sum-000

828.869. 6:09.. ..3 8262982. 2.02 and 82838... 6:6... and 226.80 a 3m
665.65 3062.00.50.03:$828 teewstmaam 23:20 0.9.0.603 26:60:88.0

 

 

 

 

000805250 EEoEtomxm 0cm 6080:0250 8:005 022080 00 cemtamEoU ”0. .m 204-.-

The concentration of 2,3-pentanedione at low temperatures was predicted more
accurately. This could be because at low temperatures the conversion of lactic acid was
low and competing side reactions were at a minimum. Also, the self-polymerization of
lactic acid was limited. Table 3.11 shows the results of an F-Test on the 2,3-
pentanedione concentrations predicted from the kinetic model. A confidence level of
95% was used in this test. The F—Test used is part of the Data Analysis tools in

Microsoft EXCEL.

Table 3.11: Results of F-Test on 2,3-Pentanedione Concentrations Predicted.

 

 

Predicted Experimental
Concentration Concentration
Mean 0.429 0.443
Variance 0.037 0.042
Observations 1 5 15
df 14 14
F 0.872
P(F<=t) one-tail 0.400
F Critical one-tail 0.402

 

The predicted concentrations of acetaldehyde and acrylic acid were
comparatively more accurate at low temperatures. A higher concentration of acrylic acid
was predicted for experiments conducted at low feed flow rates. It is possible that acrylic
acid decomposes by a pathway not listed in the reaction system. Another possible
explanation is that in these experiments the concentrations of lactic acid and cesium
lactate in the reactor did not reached steady state values. Table 3.12 shows the results of
the F-test for the acrylic acid concentrations predicted by the kinetic model. Again, a

confidence level of 95% was used in this test.

83

Table 3.12: Results of F-Test on Acrylic Acid Concentrations Predicted.

 

 

Predicted Experimental
Concentration Concentration
Mean 0.573 0.647
Variance 0.0206 0.0577
Observations l 5 1 5
df 14 14
F 0.357
P(F<=f) one-tail 0.032
F Critical one-tail 0.402

 

The kinetic model predicted higher acetaldehyde concentrations for experiments
conducted at low flow rates. Table 3.13 shows the result of the F-Test on concentrations

predicted from the kinetic model.

Table 3.13: Results of F—Test on Acetaldehyde Concentrations Predicted.

 

 

Predicted Experimental
Concentration Concentration
Mean 0.390 0.471
Variance 0.0434 0.0782
Observations 15 15
df 14 14
F 0.554
P(F<=f) one-tail 0.141
F Critical one-tail 0.402

 

84

CHAPTER 4

SUMMARY AND RECOMMENDATIONS

4.1 Summary

This work has shown that the condensed phase conversion of lactic acid, in the
presence of alkali lactate catalyst, can be conducted in a continuous reactor to yield 2,3-
pentanedione, acrylic acid, acetaldehyde and propanoic acid. The reactive nature of
lactic acid, especially its ability to form intramolecular esters and oligomers, makes the
design and operation of a continuous reactor challenging. By operating the reactor at
steady state, it is possible to maintain conditions in the reactor favorable to the formation
of a single product for a longer period of time. To run the reactor in a continuous mode,
it is necessary to continuously remove liquid product from the reactor using an overflow
tube. Further, high feed flow rates and feed concentrations are required to establish
steady state in the reactor. The reactor was operated at a steady state temperature of
220°C, feed flow rate of 1.1 mllmin, and catalyst concentration of 7.5 g cesium
lactate/100g of 50 % aqueous lactic acid feed for over 24 hours, proving the stability of
the catalyst and the reaction system.

The performance of the reactor depends mainly on three parameters, reaction
temperature, feed flow rate, and catalyst concentration. The highest 2,3-pentanedione
yield of 19% was obtained at a temperature of 240°C. The reaction was most selective to

the formation of 2,3-pentanedione at low feed flow rates. Increase in feed catalyst

85

 

concentration increases the yield of 2,3-pentanedione. The formation of acrylic acid is
favored at high feed flow rates. The highest yield of acrylic acid obtained was 15 %.

The kinetic model developed predicted the concentrations of product accurately
only at low temperatures and high feed flow rates. The increase in percentage error for
higher reaction temperature suggests that reaction pathways not included in the model
may be occurring. The accuracy of the kinetic model can be increased if a precise

estimate of free lactic acid concentration in the reactor is available.

4.2 Recommendations

The optimization of 2,3-pentanedione formation from lactic acid in the condensed
phase is mostly complete. Cesium lactate is the best catalyst for this reaction. A greater
yield of 2,3-pentanedione is possible by choosing a higher feed catalyst concentration
than that selected as maxima for the optimization study. The reaction temperature for
continuous operation should be around 240°C, higher temperature will lead to excessive
formation of char and oligomers in the reactor.

Further work needs to be done to investigate the possibility of converting
byproduct acrylic acid to propanoic acid. This reaction can be either done after the
products are condensed or by possibly hydrogenating acrylic acid as it is formed in the
reactor. The recycle of liquid product, which contains lactic acid and cesium lactate
catalyst, to the reactor after dilution and hydrolysis is another option to increase the yield

of the reaction.

86

 

INN Ii. I'll—D: ..llhfa
I....- .

--.-HR

 

 

APPENDICES

87

 

55.8 030.5 %ow

 

 

 

.0_o< 0003 00 5039.80 w=o=:_0:oU-_Eom .8 333034 cozomum “_< BamE

wt 0.5 3:08; :0 an: 6:005

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8:95:00
330.8903.

 

     

is 0.80 2.5:

II.

E
U

 

 

 

03:80:55 75:00

 

 

 

< £050.?

8» his... 537$

88

Appendix B

Table Bl: Conversion of Lactic Acid on Heating 3 (42).

 

 

Time/hr. Lactic Acid Dimer Lactide Lactide Trimer

100°C
1 85.5 14.5 - - -
2 78.4 20.9 — - - ...
4 72.3 27.2 - - + '4
6 66.5 30.2 + + 2.4 1:

125°C
1 74.6 24.4 - - -
2 67.4 31.1 + 0.7
4 55.1 36.1 0.7 1 5 6.2
6 49.8 36.7 1.3 2 8.9 ‘

 

 

3 Figures are given as % of the total observed.

89

M30000 000.0035 00.0 00000020 20000008000 00:35 00: 800 00:000. .000 umfio>< ..

 

 

 

 

 

 

m8 0% 0mm 0.90
0.2 0:“ 0.: 0.08 .
_.0m 0.0m 0.0 0.00m m2:0800=_2:000320 00:0 00835
Now 0.0. 0?. 0. .58 8.82
0: 2: am an: 48.: 2.0.: 3.4.: 04m: 80: 0-0.: 0.0%
:0 . 02- 0mm 34.: was 88 83 $0: 08.. 8.7.0 440%
00. w: 4.2 8: 08.: $3 20: an: an: 0.0.: 9.0%
0? 08 08 48.. :3 98.: £00 :40: 0.4.: 8.30 00%
02 4.2 we 80: 80.: 00...: 42.: own: 54.: 8.0.00 08%
02 0.04 0.8 0%.: $0.: 820 RS 84.: in: 00.00 040% 0
3 0.2 4.3 an: an: 00.3 488 an: :0: 0.0.8 040% 9
a? 03 0.2 m2: 5.: 88 £00 80.: £00 6.0.00 30%
0% EN 3: m8: 220 R20 30: $20 84.: 8.0.00 040%
3 4.0 E E: 0%.: 83 54.: :20 £0: 0.7.00 90%
0:2 :8 Sam 80: 84.: 20: $20 so: 80.: 0-..-.00 08%
0% 08 0.: 80: 80. EN: 83 was 0mm: 6.20 E08
2.. 02 ....m 20: 02.0 3.0: 800 9%.: ~08 0.0.1 80?.
0.04 :0 0w :3 0:3 $4.: $.00 NS: 94.: 8.30 9.0%
0% 0m .4 $0: an: 0.0: 8.0: 9.3 $00 2.0.0-0 06%

8008.82 8:02 .03 2040208.. 0:8... .03 80082902 0:094 03 220.80 0me

00.00.00 3005000000.:000000 00008010040 330:0 0000.00me 3000-0300000

 

 

 

.m .0002 282v. .00 0000008000 fiEoEtuQxM 0:: 00080000000 8:005 022005 00 08009000 ”—0 0.00-.-

0 £0500<

 

0 This model used the average concentration of lactic acid in the two product streams to

predict the concentrations of product.

0 The prediction of reaction data comparable to the results of the kinetic analysis

reported in section 3.2.

Table C2: Kinetic constants calculated for Kinetic Model B

 

 

Reaction Preexponential factor Activation

(km) energy(J/mol)
Lactic acid _’ Acetaldehyde + Products 3.76E+01 sec'I 57601.7
Lactic acid + (Cs)Lactate———> 2,3—P 5.63 liter mol‘lsec'l 47872.5
Lactic acid —+ Acrylic acid 4.90 13-02 sec" 25396.9
Acrylic acid ——> Products 1.61 E02 sec'l 15879.1

 

 

-9.6 . , .
0.00194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214
'93 * y: -6928.3x+ 3.6248
R3=08682
-10 _

-I().2 .

lnkl

~10.4 ~

-10.6 ~

-10.8 1

 

-1] -4

 

 

-ll.2

 

lfl‘

Figure C1: Arrhenius Plot for Reaction 1 (Kinetic Model B)

91

 

-92 . - - . . . 0
j
0.00194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214

y = ~5758.lx + 1.7292
-9.4 4 R2 = 0.9627

-9.6 ~

 

-9.8 -

ln k2

-l().4 .

 

 

-10.6 1L J

l/T

Figure C2: Arrhenius Plot for Reaction 2 (Kinetic Model B)

-8.8

 

 

 

 

 

, _ _ . . j
0.00194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214
-8.9 - y: -3054.7x - 3.0153
R3=0.7526
-9 «
-9.1 l
('-
3; -92-
-9.3
-9.4 " v \‘4
-9.5 4
-9.6
l/T
Figure C3:

Arrhenius Plot for Reaction 3 (Kinetic Model B)

93

 

‘7.7§ 1 - 1 -

- 0
0.00194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214
-7.8 -
y = -l909.9x - 4.1289

R2 = 0.693
-7.85 «

In k4

-8.05 4

-8_l .4

 

 

 

 

Figure C4: Arrhenius Plot for Reaction 4 (Kinetic Model B)

94

 

000000 E05590 :8 00.3005 52.550050 00205 5: 800 00.500: 5:0 0w§0>< a

 

 

 

 

 

95

 

0.00 0.3. 0. _ 0 0.0%

0.00 0.0m 0.2 0.0mm

0.00 0.00 0.0_ 0:00N 80280052500023 500 00295

w?- Qm 33. .055 00.5%

New 0.00 _.0m .000 000.0 m _ 0.0 N000 03.0 000.0 2-0. 3 3000

fax p.00 0. _ 0.90.0 $00 0000 02-0 :00 005.0 A0. _ -. C 3000
0.00 NS 0.0V 00 _ ._ 000.0 $5.0 000 0.0.0 090 2.0. C 00000
0N0 0.? 0 .00 000; 0 _ 0. _ 000.0 0200 $0.0 $0.0 A0. _ . C «0000

0.0 is 0N N000 000.0 000.0 mmmd $0.0 0mm.0 80.00 00000
0. _ 0 0.00 0.00 000.0 000.0 000.0 08.0 .000 00m.0 2. _ .00 00000
0.0m 0 .N _ N.0 0mm .0 >00 .0 00N.0 300 30.0 05.0 2-. _ .00 00000
0. _ 0 ~00 06 m0 _ .0 50.0 000.0 m_ _ .0 00 _ .0 00m.0 8.000 00000

00m 0.0 0.0 000.0 m _ 50 000.0 03.0 55.0 000.0 8.000 3000
0.00 080 0.: 050.0 000.0 03-0 N050 000.0 000.0 2.700 00000
0.0: 0:00. 0.02 00N.0 000.0 mmmd 0000 N000 000.0 2-. _ -.00 wm000
0.00 0.00 5.8 000.0 Em; 08.0 00 _ .0 N000 2 N0 8. _ . _ -0 _ 0000
0.;- w.m :0 0_m.0 030.0 03.0 0000 M000 mmmd £20.: 00000
0.00 0.00 0.2 _0N.0 03.0 000.0 mmmd 000.0 000.0 8.7.: 00000
vdv 0.0 0.0 N000 _00.0 3N0 000 000.0 5000 20.: 5000

00.300300< 0253‘ mfim 00>:00.S00< 0:53.. Axum 00300—8006. 0:056. 00d 0350500 nmxm
005000 03080395050050 0:222:030 23020 080.003 0:000:5050

 

 

 

.0 .0022 000:0— 50 5000:0050 350.5590 0:0 5005:0050 5:005 00.2005 .5 5000950 ”.0 0.00..-

G 50:00.5.

o This model used the concentration of lactic acid in the liquid product stream to
predict the concentrations of product.

0 The Arrhenius plots not as accurate.

 

 

 

 

 

 

 

 

Table D2: Kinetic constants calculated for Kinetic Model C
Reaction Preexponential factor Activation
(km) energy(J/mol)
Lactic acid —' Acetaldehyde + Products 8.04E+01 sec'I 64053.65
Lactic acid + (Cs)Lactate—-—> 2,3-P 0.109 liter mol'lsec'l 35941.14
Lactic acid ——. Acrylic acid 4.79 13-03 sec1 19357.47
Acrylic acid ——> Products 1.53 E-O3 sec'l 6674.889
-102 - . , .
0.0( 194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214
-10.4 4 y = -7704.3x + 4.3869
R3 = 0.7295
-10.6 ~
-10.8 «
-11 d
_ .112 ~
414 4
-11.6 «
-1 1.8 - \
-12
In

Figure D1: Arrhenius Plot for Reaction 1 (Kinetic Model C)

96

 

 

-10.4
-lO.5 i
-lO.6 ~
-10.7 4
-10.8 -

-l().9

1

ln k2

-ll ‘

-11.l‘

-ll.2*

-ll.3 2

0.00194 0.00196 0.00198 0.002

 

-ll.4

0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214

y = -4323x - 2.211
R2 = 0.7295

 

 

Figure D2:

Arrhenius Plot for Reaction 2 (Kinetic Model C)

97

-:x

I ...11'11:

 

 

-9.75 . . . . .
0.00194 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.0C2l4

-9.8 ~
y = -2328.3x - 5.3406

-9.85 ~ R2 = 0.7295
-9_9 2
-995 -

-10 .1

ln k3

-10.05 ‘

-10.l ~

40.15 7

 

-10.2 . fi

/.

-10.25 *

 

 

 

-l0.3
lfl‘

Figure D3: Arrhenius Plot for Reaction 3 (Kinetic Model C)

98

 

 

-3 . .
0.0Cl94 0.00196 0.00198 0.002 0.00202 0.00204 0.00206 0.00208 0.0021 0.00212 0.00214

-8.02 ~
y = -802.85x - 6.4795

-804 - R2 = 0.7295

-8.06 4
-8.08 5

-8.1 ~

In k4

-8.12 -

-8.14 .

-8'16 9 ' \

l/T

 

 

 

Figure D4: Arrhenius Plot for Reaction 4 (Kinetic Model C)

99

 

i-J

4"
.3 .

 

 

10.

ll.

12.

13.

14.

15.

16.

LIST OF REFERENCES

Datta, R., S. Tsai, Fuels and Chemicals from Biomass, 12, 224, American Chemical
Society (1997).

Miller, D. J., J. E. Jackson, R. H. Langford, G. C. Gunter, M. S. Tam, P. B. Kokitkar,
U. S. Patent # 5,731,471 (1998).

Tam, M. S., R Craciun, D. J. Miller, J. E. Jackson, Ind Eng. Chem. Res, 37, 2360-
2366 (1998).

Wadley, D. C., M. S. Tam, P. B. Kokitkar, J. E. Jackson, D. J. Miller, J. Catal., 165,
162-171 (1997).

. Pen'y, S. M., Masters Thesis, Michigan State University (In progress).

Carothers, W. H., G. L. Dorough, F. J. Van Natta, J. Am. Chem. Soc, 54, 761 (1932).

Holten, C. H., A. Muller, D. Rehbinder, Lactic Acid — Properties and Chemistry of
Lactic Acid and Derivatives, Verlag Chemie (1971).

Monteaguo, J. M., L. Rodriguez, J. Rincon, J. Fuertes, J. Chem Tech, and Biotech,
68,271(1997)

Silva, E. M., S. T. Yang, J. Biotech, 41, 59 (1995).

Davidson, B. H., C. D. Scott, Biotechnology and Bioengineering, 39, 365 (1992).
Yu, R. J., E. J. Van Scott, U. S. Patent # 5,561,158 (1996).

Lipinsky, E. S., R G. Sinclair, Chem. Engr. Prog., 82, 26 (1986).

Kitson, Melanie, P. S. Williams, U. S. Patent # 4,777,303 (1988).

Chang, Y., P. Yang, Z. Jin, H. Chellapa, C. Xing, R. Mueller, M. Anderson, A.
Clarke, E. Ianotti, Proc. of Com Utilization Conf 6, National Corn Growers Assn,
St. Louis, MO. (1992)

Beilstein, F ., E. Weigland, Ber. Deut. Chem. Ges., 17, 840 (1884).

Dakin, H. D., J. Biol. Chem, 4, 91, (1908).

100

 

 

18F

l9F

20 1

2]!)

22C

23L

24L

258

36 E1

 

 

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

Erlenmeyer, E., Ber. Deut. Chem. Ges., 10, 635 (1877).

Fisher, C. H., C. E. Rehberg, L. T. Smith, J. Am. Chem.Soc., 65, 763 (1943).
Ratchfor, W. P., C. H. Fisher, Ind. Engr.Chem., 37, 382 (1945).

Holmen, R. E, U. S. Patent # 2,859,240 (1958).

Mok, W. S., M. J. Antal, Jr., M. Jones, Jr., J. Org. Chem, 54, 4597 (1989).
Gunter, G. C., Ph. D. Dissertation, Michigan State University, (1994).
Langford, R.H., Masters Thesis, Michigan State University, (1993).
Lautemann, E., Liebigs Ann. Chem, 113, 217 (1860).

Broadbent, H. S., G. C. Campbell, W. J Bartley, J. H. Johnson, J. Org. Chem, 24,
1847(1959)

Engelhardt, Liebigs Ann. Chem, 70, 241 (1849).

Ipat’ev, V., Razuvaev, G., (translation) Ber. dent. che. Ges., 59, 2031 (1926).
Gunter, G. C., J. E. Jackson, D. J. Miller, J. Catal., 148, 252 (1994).
McCarthy, S. L., J. Am. Soc. Brew. Chem, 53, 178 (1995).

F uria, T. E., N. Bellanca, Fenarolli ’s Handbook of Flavor Ingredients 2"d ed., 2, 454,
CRC Press, Inc. (1975).

Anonymous, Food and Ingredient News, Section: Process Technology, 3 (1996)
Thiel, M. A., Masters Thesis, Michigan State University (1996).
Tanimoto, M., I. Mihara, T. Kawajiri, U. S. Patent #5,739,392 (1998).

Mark, H. F., D. F. Othmer, Eds, Encyclopedia of Chemical Technology 3” ed, 16,
637, John Wiley and Sons, New York (1978).

Hsu, S. T., S. T. Yang, Biotechnology and Bioengineering, 38, 571 (1991).
Mccoy, M., Ed., Chem. Mktg. Rep, 49 (February 24, 1997).

Gunter, G. C., R. H. Langford, J. E. Jackson, D. J. Miller, Ind Eng. Chem. Res, 34,
974 (1995).

101

 

1..)

00

38. Gunter, G. C., R. Craciun, M. S. Tam, J. E. Jackson, D. J. Miller, J. Catal., 164, 207
(1996)

39. Tam, M. S., Ph. D. Dissertation, Michigan State University, (1997).
40. Craciun, R., Ph. D. Dissertation, Michigan State University, (1997).

41. Box, G. E. P., W. G. Hunter, J. S. Hunter, Statistics for Experimenters, John Wiley
and Sons, (1978).

42. Alén, R., Sj65tr6m, E., Acta Chemica Scandinavia, B 34, 633 (1980).

102

 

"11111111111111111111115