Synthesis of Bisphenol A with Heterogeneous Catalysts

by

Liliana Neagu

A thesis submitted to the Deparment of Chernical Engineering in conforrnity with the requirements for the degree of

Master of Science (Engineering)

Queen' s University Kingston, Ontario, Canada August, 1998

Copyright O Liliana Neagu, October 1998

1 * m

National Library of Canada

Acquisitions and Bibliographie Sewices
395 Wellington Street Oftawa ON K l A O N 4 Canada

du Canada

Bibliotheque nationale

Acquisitions et services bibliographiques

395.rue Wellington Ottawa ON K IA ON4

canada

The author has granted a nonexclusive licence dowing the National Library of Canada to reproduce, loan, distn'bute or sell copies of this thesis in microform, paper or electronic formats.

L'auteur a accord une licence non exclusive permettant la Bibliothque nationale du Canada de reproduire, prter, distribuer ou vendre des copies de cette thse sous la forme de microfiche/fh de reproduction sur papier ou sur format lectronique.

The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial emcts fiom it may be printed or otherwise reproduced without the author's permission.

L'auteur conserve la proprit du droit d'auteur qui protge cette thse. Ni la thse ni des extraits substantiels de celle-ci ne doivent tre imprims ou autrement reproduits sans son autorisation.

Abstract
The synthesis of bisphenol A (BPA) with heterogeneous catalysts was uivestigated in a batch system and in a plug flow reactor. Gibbs reactor simulations contributed to a better understanding of the reaction which leads to BPA formation. Experiments were conducted with ~mberlyst@-15, Nafion@ NR-50,Nafione SAC-13. and activated alumina acidified with concentrated hydrochloric acid (AA300/HCl). An expernental design was used to investigate the effects of temperature, catalyst concentration, molar ratio of acetone and phenol in the initial reaction mixture, and the size of the catalyst bead. Al1 the factors significandy innuence some or al1 the aspects of the process of BPA formation.

Al1 three new catalysts: AA300/HC1,Nafion@NR-50,and Nafion@SAC-13,were found suitable to catalyze the production of bisphenol A, using phenol and acetone as starting materials. Both yield and selectivity are significantly higher for the processes that use the newly identined catalysts than the yield and selectivity obtained in the process that uses Amberlyst@15.

Acknowledgments
There are many people 1 want to thank in this section, people who offered me their support and their fnendship, for which 1 am grateful.

I would like to thank my supervisors Dr. Tom Harris and Barrie Jackson for their support, encouragement and understanding throughout the completion of this project. 1 would also like to thank Dr. Whitney and Dr. Brian Hunter for the meaningfid conversations, Steve Hodgson, Lisa Prior and Martin York for the& technical assistance with my research equipment.
1 am grateful to my office mate Shannon Quinn for her sincere fiiendship and her computng knowledge. 1 thank Gregg Logan for his Eendship, for the endess

conversations about food, and for introducing me to graduate student life in the department. Thank to everybody in the department for creating such a pleasant work place.
1 would like to thank my husband for being supportive and understanding sometimes and to rny daughter for being a good and happy child, for sleeping overnight and for not

crying as much as she could have.

Many thanks to the Queen's Day Care Centre staff, Halina, Donna, Pada, Sean, Karen, Lori, Sandra, for taking such a good care of my daughter and for spoiling her more than I did.

As vrea sa multumesc parintilor mei Voica si Aurel Monea pentru ca mi'au dat puterea sa visez si aripi sa zbor.

Table of Contents
1 Introduction.......................................................................... 2 Basic Chemistry and Production of BPA ...................................... 2.1 Preparation of Bisphenol A ................................................. 2.1.1 Acetone Process .................................................. 2.1.1.1 Primary Reaction ...................................... 2.1.1.2 By-Products Formation ............................... 2.1.1.3 Reaction Order ......................................... 2.1.1.4 Equilibrium Data...................................... . ...................... 2.1.1.5 Catalysts..................... . 2.1.1 .5 . 1 Catalyst Enhancers......................... 2.1.1.6 Bisphenol Stabilizers.................................. 2.1.1.7 Solvents................................................. 2.1.1.8 Reaction Mechanism .................................. 2.1.1.9 Reactor Configuration................................ 2.1.2 Alternatives to Acetone as Feedstock......................... 2.2 Purification................................................................... 2.2.1 Catalyst Separation.................. ............................ . 2.2.2 %PASeparation from Cnide .................................... 2.2.2.1 Methods of Separating BPA fiom the 1:1 BPAPhenol Aduct ............................................... 2.2.2.2 By-Products Isomerization to BPA ................. 2.3 Manufacturing ............................................................... 2.3.1 Resin-Catdyzed Process........................................ 2.3.2 Hydrogen Chloride-Catalyzed Process ........................ 2.3.3 Resin-Catalyzed Process II ...................................... 2.4 Physical Properties .......................................................... 2.5 Chernicd Properties.............................. .......................... . 2.6 S u m m q ...................................................................... 3 Gibbs Reactor Simulations........................................................ 3.1 The PRO II@ Gibbs Reactor ................................................ 3.2 Simulation of the Bisphenol A Reaction .................................. 3.2.1 Analysis of the Simulation Results ............................. 3.2.1.1 Effect of Temperature and Acetone:Phenol Molar Ratio on BPA Formation.......................... 3.2.1.2 By-Products Formation ............................... 3 -3 Surnmary of the Sunulation................................................ 3 -4 Conclusions.................................................................. 4 Expermental Investigation........................................................ 4.1 Apparatus-OveMew .........................................................

4.5 Summary ......................................................................

4.2 Materials ....................................................................... 4.2.1 Solid Catalysts..................................................... 4.2.1.1 ~afion@-~erfiuorosulfonated Ionomer .............. 4.2.1.2 High Surface ~ r e NafionBResin................... a 4.3 Procedures .................................................................... 4.3.1 Reactor Loading Procedures .................................... 4.3.1.1 NMR Tube Reaction. ................................. 4.3.1.2 Batch Reactor .......................................... 4.3.1.3 Flow Reactor ........................................... 4.3 -2 Reactor Sarnpling Procedure and Sample Preparation....... 4.3.2.1 NMR Tube Reaction .................................. 4.3.2.3 Batch Reactor .......................................... 4.3.2.3 Flow Reactor........................................... 4.3.3 Reactor Shut-Down and Clean-Up Procedure ................ 4.3 -3.1 NMR Tube Reaction.................................. 4.3.3.2 Batch Reactor.......................................... 4.3 -3.3 Flow Reactor ........................................... 4.4 Sample Analysis.............................................................. 4.4. I Gas Chromatography/ Mass Spectrometry Analysis ........ 4.4.2 NMR Analysis ................................................... 4.4.2- 1 General Introduction to the N M R Procedure Used i this Study ........................................ n 4.4.2.2 Cdculation of the Error Associated with the NMR Analysis .............................................. 4.4.2.3 Procedure for Caiculating the Yield in BPA .......
*

5 Experimental Results and Discussion ........................................... . 5.1 Prelimioary Investigation................................................... 5.1.1 Evaluation of System Reactivity and Blank Reactions...... 5.1-2 Evaluation of Experimental Region ............................ 5.1 -3 Scheme of Reaction .............................................. 5.1.4 Experimental Reproducibility................................... 5.1.5 Validity of Simulation Prediction for Depletion of Acetone.............................................................. 5.2 Investigation of Suitability of New Catalysts............................ 5.3 Performance Comparison of ~ a f i o n @ Amberlysta 15.............. and 5-4 Experimental Design........................................................ 5.4.1 Factors Chosen and Responses................................. 5.4.2 Evaluation of Results corn Experimental Design............ 5.4.3 Precision of Caiculated Effects................................. 5.4.4 Effects Analysis.................................................. 5.4.4.1 Selectivity of BPA Formation ...................... 5.4.4.2 Selectivity of O-pIsomer Formation ............... 5.4.4.3 Selectivity of Chromanes Formation...............

B C

5.4.4.4 Yield in BPA ........................................... 5.4.5 Regression Analysis ............................................. 5.4.5.1 Mode1 for Selectivity of BPA Formation .......... 5.4.5.2 Model for the Selectivity of O-pIsomer Formation................................................... 5 4 - 5 3 Model for the Selectivity of Chromanes Formation................................................... 5.4.5.4 Mode1 for the Yield in BPA ................... -...... 5.4.6 Summary of z3Experimentai Design .......................... 5-5 Additional Runs .............................................................. 5.6 z4 Experimental Design ..................................................... 5.7 Regression Analysis for the z4 Experimental Design................... 5.7.1 Mode1 for Selectiviw of BPA Formation..................... 5.7.2 Mode1 for Selectiviv of Chromanes Formation ............. 5.7.3 Model for Yield in BPA ......................................... 5.8 Summary ...................................................................... Reactions in the Plug Flow Reactor............................................. . 6.1 Reactions with Acidic Activated Alumina................................ 6.2 Reactions with ~ a f i o n NR-50 ............................................ @ 6.3 Reaction with ~ a f i o n @ SAC-13 ........................................... 6.4 S u m m q ...................................................................... ..................... Conciusions and Recommendations....................., , 7.1 Conclusions.............................................................. 7.2 Recommendations....................................................... Health and Safety Regdations .................................................... PRORI@ Input File .................................... .............................. Summary of Simulation Resuits.................................................. The NMR Phenornenon............................................................

List of Tables
-1 Quality characteristics for BPA as raw matenal for polycarbonates ........... .1 Equilibrium constant for the BPA f o.p-isomer transformation............. .2 Results of the reaction of acetone with phenol in the presence of zeolites and cation-exchange resin.......................................................... . . 2.3 Solubilities of bisphenol A in various solvents (g/100g solvent) .............. 2.4 Variation of vapor pressure with temperature..................................... Gibbs reaction simulation with reaction parameters.............................. Equilibrium constant for the BPA o.p4somer transformation based on simulated data......................................................................... Materials used in experiments..................................................... . . Acquisition method on the mass spectrometer.................................... Peak table with retention times and boiling points of the products ............ ... Data acqursrtion parameters......................................................... Summary of the experiments ........................................................ Results of the second set of experirnents........................................... Results of the experiments performed with AmberlystB-15 in the batch reactor ................................................................................... Results of performance cornparison between ~ a f o n @ ~mberlyst@-... and 15 High value, low value, midpoint, range and half range for each factor........ Experimental runs used to investigate the efTect of catalyst concentration (C), temperature (T) and molar ratio of acetone and phenol (R) ............... Responses for the experiments perfomed in the 23experimental design ..... Calculated effects..................................................................... Precision of calculated effects....................................................... 5.10 Results of the regression analysis for the selectivity of BPA formation..... 5.1 1 Results of the regression analysis for the selectivity of O-p isomer formation............................................................................... 5.12 Resdts of the regression analysis for the selectivity of chromanes formation............................................................................... 5.13 Results of the regression analysis for the yield in BPA ........................ 5.14 Additional runs ....................................................................... 5.15 Calculated effects for the additional runs ......................................... 5.16 Cornparison between the calculated effects in the fkst set and the second set of experiments.................................................................... 5.17 Data for the z4experimental design............................................... 5.18 Calculated effects for the 24 design............................................... . . 5.19 Sigmficant effectS. .................................................................. 5.20 Results of the regression analysis for the selectivity of BPA formation.....

5.21 Results of the regression analysis for the selectiviv of chromanes formation. .............................................................................. 5.22 Results of the regession analysis for the yield in BPA.. ...................... 6.1 Summary of the experiments. ....................................................... 6.2 Results of the experiments with AM001 HCl. ................................... . 6.3 Results of the experiments with Nafion@NR-50.. ............................... 6.4 Results of the experiment with Nafion@SAC-13.. ............................... A. 1 Chernicals used in experiments! associated hazards and safety requirements ........................................................................... B. 1PRO/LI@ keyword input file.. ........................................................ C.l Variation of bisphenol A, o,p-isomer, and triphenol formation with the acetone:phenol molar ratio at 323.15 K. The results are presented in mol
%........................................................................................

C.2Variation of bisphenol A, og-isomer, and triphenol formation with the acetone:phenol molar ratio at 333.15 K. The results are presented in mol

C.3Variation of bisphenol A, op-isomer, and triphenol formation with the acetone:phenol molar ratio at 343.15 K. The results are presented in mol

CAVariation of bisphenol A, 07p-isomer, and triphenol formation with the acetone:phenol molar ratio at 353.15 K. The results are presented in mol
%........................................................................................ CSVariation of bisphenol A, og-isomer, and triphenol formation with the

acetone:phenol molar ratio at 363.15 K. The results are presented in mol

%........................................................................................

C.6Variation of selectivity of bisphenol A 0, 07p-isomer (II), and triphend (III) with the temperature at various acetone:phenol molar ratios. ...................................................................................

List of Figures
2.1 Conversion of phenol versus time for various catalysts.......................... 2.2 Product selectivity versus time for the reaction catalyzed by Re-Y ........... 2.3 Product selectivity versus time for the reaction catalyzed by Hmordenite ... 2.4 Product selectivity versus time for the reaction cataiyzed by Amberlyst.15 .. 2.5 Mechanism of condensation of acetone with phenol s a hydrogen bonds ..... 2.6 Reactor configuration................................................................. 2.7 Production of bisphenol A with resin catalyst.................................... 2.8 Production of bisphenol A with hydrogen chlonde catalyst.................... 2.9 Production of bisphenol A with resin catalyst II .................................. 3.1 Variation of BPA formation with molar ratio acetone.pheno1.................. 3.2 Variation of BPA formation with temperature and acetone:phenoI molar ratio .................................................................................... 3.3 Variation of selectivity of BPA with temperature and acetone:phenol molar ratio ................................................................................. 3.4 Variation of og-isomer formation with acetone:phenoI molar ratio ........... 3.5 Variation of o,p&orner formation with temperature at various rnolar ratios acetone.pheno1..................................................................... . . . 3.6 Variation of selectivity of og-isomer with temperature at various molar ratios acetone:phenol ................................................................. 3-7 Variation of triphenol formation with molar ratio acetone.pheno1............. 3 -8 Variation of triphen01 formation with temperature at various molar ratios acetone.pheno1......................................................................... 3.9 Variation of selectivity of triphenol with temperature at various molar ratios acetone.pheno1................................................................. 3.10 Variation of BPA formation with temperature and acetone:phenol molar ratio ................................................................................... 3.1 1 Variation of 0.p 4somer formation with temperature and molar ratio . acetone.pheno1..................... ................................................... 3.12 Variation of triphenol formation with temperature and acetone:phenol molar ratio ............................................................................. 3.13 Variation of BPA, o.p.isomer. and triphenol formation with molar ratio acetone:phenol at 353.15 K .......................................................... 3.14Variation of BPA. opisomer. and triphenol formation with temperature... 4.1 Plug flow reactor ...................................................................... 4.2 Nanon@structure; m = 6 or 7, n 1000. x = 1. 2. or 3 .......................... 4.3 Electron withdrawing effect......................................................... 4.4 Styiized view of polar/ nonpolar microphase separation in a hydrated ionomer................................................................................ 4.5 The Yeager 3 phase mode1 of Nafion@ clusters...................................

viii

4.6 Temperature profile of method used on GC ....................................... 4.7 Chromatogram of the products obtained in the condensation process ......... 4.8 NMR spectrurn of acetone (CDCl, ) ............................................... 4.9 NMR spectrum of phenol (CDCI, ) ................................................. 4.10 NMR spectrum of bisphenol A (CDCl, ) ......................................... -4.11 NMR s p e c t m of the initial mixture of reaction (fiom 0.4 pprn to 3 .0 ppm) ................................................................................... 4.12 NMR spectnun of the initial mixture of reaction (fiom 1. pprn to 3.0 0 ................................................................................... ppm) 4.13 NMR spectrum of the final mi>aure of reaction (firom 0.4 pprn to 3.0 ppm) 5.1 Analysis of the reaction with hornogeneous catalyst (after three days) ....... 5.2 Analysis of the reaction with homogeneous catalyst (after six days) ......... 5.3 Analysis of the reaction with homogeneous catalyst (afier nine days) ........ 5.4 Analysis of the reaction with homogeneous catalyst (after twelve days) ..... 5-5 Crystals of BPA ....................................................................... 5.6 Analysis of the reaction with heterogeneous catalyst (after nine days) ....... 5 -7 Analysis of the reaction with no catalyst (after three days) .................... 5.8 Variation of BPA selectivity in tirne. .............................................. 5-9 Variation of selectivity of O-pisomer in tirne..................................... 5.10 Variation of chromanes selectivity in time ...................................... 5.1 1 Variation of BPA selectivity with temperature .................................. 5.12 Variation of +p isomer selectivity with temperature........................... 5.13 Variation of chromanes selectivity with temperature.......................... 5.14 Variation of BPA yield with temperature ........................................ 5.15 Disappearance of acetone.......................................................... 5.16 Chrornatogram of the products for the process catalyzed by Amberlyst- 15 ...................................................... [6 h] ...................... . .... 5.17 Chromatogram of the products for the process catalyzed by Nafion@ NR50 [3 hl ................................................................................ 5.; Chromatogram of the products for the process catalyzed by AA 300/ HCL S [6 hl .................................................................................... 5.19 Effects of considered factors on selectivity of BPA formation and their significance............................................................................ 5.20 Effects of considered factors on selectivity of O-pisomer formation and their significance ..................................................................... 5.2 1 Effects present in the molecule of phenol and the nucleophilic attack....... 5.22 Effects of considered factors on selectivity of chromanes formation and their significance ..................................................................... 5.23 Effects of considered factors on yield in BPA and their significance ........

List of Abbreviations
US dollars micro litre Angstrom activated alumina activated with hydrochloric acid acetone atomic mass unit amu Analyticd Solution of Groups ASOG atmosphere atm bisphenol A BPA BSPHNOLA bisphenol A Celsius C calibration Cal calorie cal catalyst cat. cubic centimetre centimetre cm carbon number c m Company Co. concentration conc. 1,1,1 -tnchloro-2,2-bis-(p-chloropheny1)-ethane DDT diglycidyl ether bisphenol A DGEBA dimethyl sulfoxide DMSO exempli gratia e.g . experiment exp. Free Induction Decay FID Figure Fig. fmal fin gram G s Chromatography/ Mass Spectroscopy a hour hydrochloric acid mercury Hertz ist est 1.e. International Union of Pure and Applied Chemistry I.U.P.A.C. initial in Ionization ioniz. Infia red IR Kelvin K

kcai

kj
km01 1 b

mega Hertz minimum minute millilitre mi llimetre milli mass unit m u MSDS Matenal Safety Data Sheet NIA not applicable normal boiling point NBP nanometre nnl normal melting point NMP NMR Nuclear Magnetic Resonance non-library NONLIB NRTL non-random two liquid ortho O Para P PFR plug flow reactor Ph phenol PID proportional integral denvative parts per million psi pound per square inch poyvinyl chloride PVC radians rad rf radio fiequency seconds s/n signal to noise SANS small angle neutron scattering SAXS small angle X-ray scattering SI. System International Simulation Sciences SIMSCI SOLUPARA solubility parameter SSE Surn of Squared Errors SSR Sum of Squared Residuals STDPRES standard pressure STDTEMP standard temperature T Tesla Temp. temperature TFE tetrafluoroethylene
min niin ml

LIBID h4AF'P max MHz

kilocalorie kilojoul kilomol pound library methylacetylene and propadiene

maximum

TMS TPPI TSS U.K.

us
wt

UNiFAC

mec

ZNUM

tetramethylsif ane Time Proportional Phase increments Total Sum of Squares United Kingdom universai functionai activity coefficient United States micro seconds weight hydrogen deficiency nurnber

xii

Chapter 1

Introduction

Bisphenol A (BPA) is the commercial name used in the United States for 4,4'isopropylidenediphenol. In Europe I.U.P.A.C. nomenclature and other unsystematic names are still used. Its commercial name indicates the preparation fkom two molecules of phenol and one of acetone. The molecule of BPA c m be described as two phenolic

rings joined together by a bridging isopropylidene group (ChernicalAbstract now calls

the radical 1-methylethylidene) (McKetta and Cunningham, 1976).

Dianin prepared bisphenol A for the first time in 1891 via condensation of acetone and
phenol catalyzed by hydrochloric acid.

The method was not patented until 1917.

Bisphenol A was manufactured on an industrial scale for the first time in 1923 by a

German f-,

Chemishen Fabriken, to be used a s intermediate for producing coating

resins (McKetta and Cunningham, 1976).

Since then, the production of BPA as an intermediate for epoxy resins continued to grow. Some of the first large-scale producers were Firma Resins & Vernis Artificiels in France, Farbenfabriken Bayer in Germany, Dow Chemical Company (since I941), General
Aniline and Film (fiom 1941 to 1954), Shell Chemical Co. (since 1954), Monsanto Co.

(fiom 1956 to 1971): Union Carbide (from 1960 to 1982) and General Electric Co. (since 1967) in the United States, Shawinigan Chemicals in Canada, Esquirn in Mexico, Shell Chernicals U.K. and R.Graesser and Co. in England, Ketjen and Shell in the Netherlands, Mitsui Toatsu Chemicals, Honshu Chemical Industries and Nippon Steel Chemical Co. in Japan, Raghanandan Chemical Indumies in India and others (McKetta and Cunningham, 1976). Bisphenol A is generally used as a reagent for producing polycarbonates, epoxy resins, phenoxy resins, acrylic resins, polysdfone resins, and other polyesters and as an intermediate for semi-synthetic wax (mc.vanderbiIt.edu/vumcdept/derm/contact1008). Halogenated foms are used as flame retardants, and alkylated foms are used as stabilizen and antioxidants for rubber and other plastics, like PVC for example

(essential.org/listproc/dioxin-l/msgO0464.h) It is also used as a component of foodpackaging adhesives, as a fungicide and as a component of dental filling compositions. Recently a toner for developing electrostatic images, that contains BPA, was developed

l, (Unno et a . 1997).

BPA production i the US in 1974 was only 4 5 million lb (McKetta and Cunningham, n 1
1976), compared with 1.65 billion Ib of BPA in 1996 (Hileman, 1997). T i four fold hs

increase of the production over the period of 20 years proves a high demand on the
market for the product in question. The price for BPA in 1974 was by average 0.45$/lb

(McKetta and Cunningham, 1976). Considering the inflation (Consumer Pnce Index),
the BPA pnce in 1998 should have been 1.52$/lb. The actual price for BPA in 1998 was

by average 0.94$/lb (Chemical Market Reporter, 1998). This "decrease" c m be related to the increase in production and interpreted with "The Boston Leamhg Curve" which States that: "Average Unit Selling Pnces, in Constant Dollars, Characteristicaliy Decline

20 to 30 Percent in Real Terrns Each Time Accumdated Experience Doubles" (Jackson,

1997). Considering that the production in 1998 is the same as the production in 1996,

that it doubled twice since 1974, and each t h e it doubled the average unit selling price in constant dollars declined 25% by average, the calculated price of BPA is 0.86 $Ab. This is slightly lower than the actuai price for BPA in 1998.

It is well known that for obtaining light-coloured hi&

rnolecular weight poIymers via

linear condensation, the PLU@ of the monomers m u t be high. Ordinary BPA is adequate

for making most epoxy resins, while BPA of very high puity is needed for polycarbonates (99.8% purity has been mentioned as a minimum requirement (McKetta and Cunningham, 1976)).

The characteristics of BPA used as a raw material for producing polycarbonates are
presented in Table 1.1 (Catana et al., 1993):

Table 1.1: Quality characteristics for BPA as raw materiai fer polycarbonates (Catana et a. 1993) l,
-

Specification Aspect, pallets or crystals Melting rioint, OC Colour of melt, "Hz Light transmission, %, min Water, wtY0, max Ash. wt%. max ~ron. D ~ max D .

Vaiue white 156

50 98 O. 1 0.005

There are several methods of evaluating the quality o f BPA.

The most important

parameter that characterizes the quality of BPA is its colour, and it was found that iron is one of the agents that changes the colour of BPA, due to the coloured complexes that are formed (Wasilewska, 1997). The colour can be estimated by analyMg the percentage transmission of a 50% solution of BPA in methanol or acetone and comparing it to a blank at 350 nm (McKetta and Cunningham, 1976) or 420 MI (Shinohara, 1971).

The technique that is most used for estimating the pur@ of BPA is the melting point (McKetta and Cunningham, 1976). Cnide products have wide-range melting points starting at about 140C. The rnelting point of the pure compound is 157OC. Good commercial grades melt at 154 to 155C. The cryoscopic constant has been reported as

10C (McKetta and Cunningham,1976), and 17C (Challa and Hermans, 1960). Another

simple test is to measure the percent of impurities that easily dissolve in a paraffinic solvent, cyclohexane for example (McKetta and Cunningham, 1976).

Since obtaining high purity BPA is of great importance, improvement of the manufacniring process \vas continuously investigated by researchers. Either the yield or the selectivity of the process, or both, were considered for improvement and several modification of the original method were studied: different catalysts, homogeneous and heterogeneous, alternatives to acetone as feedstock, and alternatives to acetone and phenol as feedstock.

The purpose of this investigation is to outline the b a i s of a search for new solid catalysts
that could be used in a catdytic distillation unit for produchg bisphenol A to improve

yields and selectivities. Catalytic distillation is a process where a reaction takes place simultaneously with a separation process in the same unit (Podrebarac et al., 1997). The major advantage of this type of system over traditional systems are the potential savings

in production costs, since not only one operational unit is eliminated, but also the

nt associated piping and instrumentation that are required to connect the reaction u i with
the separation unit are eliminated.

Catalytic distillation is a process that has the potential of producing bisphenol A at lower production costs. With this purpose in mind, the investigation of more suitable catalysts for the process is of great interest. Prior to the final goal of producing BPA by catalytic distillation, prelirninary investigations m u t be performed to eventually identie new,

more suitable catalysts, and to nd appropriate reaction conditions. The purpose of this thesis is to examine the curent technologies available to produce BPA, to invetigate the fesib-ity of new cataiysts, and to perform experiments to investigate the effects of selected reaction parameters, using these catalysts.

In Chapter 2 a criticai literature review is conducted, which details the existing processes
used in the BPA manufacniring and purification, the alternatives that have been evaluated
with the purpose of improving the process.

Also included are some physical and

chernical properties of bisphenol A.

In Chapter 3 the Gibbs reactor simulations are investigated and the results are compared to the results in the literature.

This simulation provides insight about the reaction

mechanism which leads to BPA formation. The results of these simulations are used to determine the levels, the factors and the responses chosen for the subsequent expenmentd designs.

In Chapter 4 the experimental apparatus and the instrumentation employed to analyze the
products, also the methods used for data processing are described. Safety procedures are detailed as well.

In Chapter 5 the results obtained in the preliminary runs and the results obtained from the
experiments performed in the batch reactor are presented. A two factorial design is used to examine the effects of the chosen factors on the selected respcnses.

In Chapter 6 the results obtained in the expenments performed in the plug flow reactor are presented. This setting was used for the systerns which could not be investigated in

the batch reactor, and also for one of the new identified catalysts, which was investigzted

in the batch reactor as well. Although the nurnber of reactions in the plug flow reactor
was kept to a minimum, important conclusions and lines of fture work emerged.

Finally, i Chapter 7 the concIusions derived fiom the experimental work are presented. n Recommendations for future investigations are given.

Chapter 2

Basic Chemistry and Production Process for BPA

The intent of this chapter is to give an overview of the existing rnethods and reaction schemes for producing crude BPA and to ernphasize the ones that are used industrially. General purification issues will be presented. The physicd and chernical properties of bisphenol A will be summarized. This background material is necessary to expiain process alternatives. Most of the information presented in this chapter is tom McKetta
and Cunningham, 1976.

2.1 Preparation of Bisphenol A

T i subsection describes the chemistry of BPA formation including mechanisms, hs
possible reactions, by-products, and order of reaction.

2.1.1 Acetone Process 2.1.1.1 Primary Reaction

The acid catalyzed condensation of acetone with 2 moles of phenol is the oldest process for producing BPA.

Phenol

Acetone

Bisphenol A

The heat of reaction, for reactants and products in their natural physical state at 25OC, is calcdated fiom heats of formation as +18-4 kcdmol. Severe conditions are not required;
a 1:2 molar ratio mixture of acetone and phenol, in the presence of concentrated

hydrochloric acid or sulfuric acid 70% at room temperature deposits a mass of crude BPA crystals (McKetta and Cunningham, 1976).

The reaction conditions predominantly

favour the formation of the products (Nenitescu, 1980).

Some sources claim that the presence of 10% water in the reaction mixture greatly increases the rate of the reaction catalyzed by hydrochloric acid (Scheibel, 1974). Other sources claim that processes catalyzed by suIfonic acid ion exchange resins rnodified with -1-SH groups are also improved by the presence of 0.6 to 5% by weight water in the

initial reaction mixture @erg and Buysch, 1994). On the other hand, since water is a product of the desired reaction, its presence decreases the yield of BPA. To

counterbaiance this effect, dehydration by various water-binding agents (such as calcium

chlonde or phenyl acetate) or by azeotropic distillation have been suggested (McKetta

and Cunningham, 1976).

The reaction proceeds with an electrophilic attack of the proton fYom the acidic catalyst on the molecuie of acetone. This first step of the mechanism is very similar to the one in
the production of phenolphthalein and DDT and in the akylation of phenol with olefins

(McKetta and Cunningham,

2.1.1.2 By - Products Formation

For reactions involving the substitution of a proton in an aromatic ring, both the rate of reaction and the equilibrium distribution of products are influenced by the density of electrons at the centre of reaction (Nenitescu, 1980). This only applies if there are no

stenc effects. Thus the pp-isorner (BPA) is the most likely to f o m since the density of electrons in the para position of the phenol is higher than in the ortho position. Aiso, the p,p-isomer formation is favoured fiom the thermodynarnic point of view (McKetta and

Cunningham, 1976). Still, opisomer and some o,o-isomer are observed.

OH

It was observed that the o,o-isorner is produced in negligible amounts. Another possible product that can result fiom the reaction of the already formed BPA with the tertiary carbonium ion @-phenyl isopropylidene) (McKetta and Cunningham, 1976) is the so called "triphenol 1" (4,4'-(4-Hydroxy-m-phenyIenediisopropylidene)diphenol):

TnphenolI P-isopropenyl phenol can be obtained when the p-phenyl isopropylidene ion loses a proton. The p-isopropenyl phenol fonned can dimerize and the dimer c m add phenol to yield another triphenol ("triphenol II" or 4,4',4" -(1,1,3-Trimethyl- 1 -propanyl-3-ylidene) triphenol) (McKetta and Cunningham, 1976):

OH
Triphenol II

OH

An irreversible cyclization of the dner to 4'-hydroxy-2,4,4-trimethylflavan (flavan) can

occur if the hydroxyl group in the dimer is in the ortho position relative to the carbon

bearing the methylene group (McKetta and Cunningham, 1976):

Fl avan

If both hydroxyl groups in the dimer are in the ortho position relative to the aliphatic
chain, the 2'-hydroxy isomer is formed (McKetta and Cunningham, 1 7 ) 96:

The acetone can dimerize with itself and form mesityl oxide. The mesityl oxide formed

can M e r react with two molecules of phenol to give a product isomeric with flavan, a
chroman (McKena and Cunningham, 1976):

Acetone

Acetone

Mesrtyl Oxide

Phenol

H~C' ' H C,
chroman 1 chroman il

The dimer resulted from the dimerkation of p-hydroxy-a-methyl styrene, triphenol II and
flavan can be obtained as a result of the reaction between mesityl oxide and phenol as well. The reaction conditions that favor the formation of al1 the by-products presented so

far, are the same as the conditions that favor the BPA formation.

No unsaturated products were observed in the cmde product, leading to the idea that al1 of the unsaturated products formed M e r react to give other by-products. The o,p-isomer,
, C melting point 11l O C , triphenol 1 melting point 191O and chromane, rnelting point

158C were al1 isolated fiom cmde BPA (McKetta and Cunningham, 1976).

Due to the high reactivity of the system, many other components can be produced and are
present in the reaction mixture. A likely one is the spirobiindan (Curtis, 1962), which c m be obtained fkom two molecules of phenol and one molecule of phorone. The phorone is

the resdt of the condensation of three molecules of acetone, which c m occur in the acidic medium provided for the process of BPA formation:

H C' ,
Acetone

phorone

HG,
phorone

Phenol

2. 1.1.3 Reaction Order

The BPA formation is a condensation in two steps. First a molecule of acetone reacts
with a molecule of phenol, then the product, or the corresponding ion, reacts with the

second molecule of phenol. The reaction was reported first order in both acetone and phenol, which indicates that the first step is slower than the second step, therefore it is rate determining (McKetta and Cunningham, 1976). In another study (Kato, 1963), the HC1-catalyzed reaction was second order in phenol. According to de Jong and Dethmers (Dethmers and de Jong, 1965) the activation energy for the overail process is 15 kcdmol. According to Kato (Kato, 1963) the activation energy is19 kcal/mol. These processes are reversible fike most other electrophific substitutions. In the presence of an acid, an equilibrium c m be established between BPA and the main by-product, the o,p-isomer.

2.1.1.4 Equilibrium Data

Using phenol as a solvent for the process, the data presented in Table 2.1 were generated for the equilibrium constant for the BPA+o,p-isomer transformation. 0.067 at 40C,

0.08 at 60C, 0.1 1 at 80C, and 0.16 at 100C (Dethmers and de Jong, 1965).
Table 2.1: Equilibrium constant for the BPA

0.p-isomer transformation
60

Temperature (OC)

40 0.067

80
0.1 1

100 0.16

0.08

The ortho-para ratio increases by increasing the temperature therefore temperatures as low as possible are preferred in order to maximize the BPA formation (McKetta and Cunningham, 1976).

2.1.1.5 Catalysts
For the process catalyzed by gaseous hydrochloric acid, the reaction of BPA formation is reported to be first order in catalyst. E s is the reason why it was recommended to
the process at several atmospheres (Takenaka et al., 1968).
nin

The first catalyst used to produce BPA was concentrated hydrochloric acid. Processes that use gaseous hydrochloric acid or acid ion-exchange resins are also operated in the United States. Aithough the process is slower and the product more difficult to puri@
than in the hydrochlonc acid catalyzed process, sulfuric acid 70% to 75% concentration

can be used as catalyst.

In this case the concentration of the acid m u t not exceed the

above mentioned b i t s in order to minimize the sulfonation. There are some advantages

in using sulfunc acid as catalyst for the process: the apparatus is simpler and the
corrosion is less severe (McKetta and Cunningham, 1976).

Other homogeneous catalysts that can be used but do not seem to have practicai

importance are: hydrogen bromide, boron trifluoride, boric acid, femc chlonde, silicon tetrachloride, phosgene, phosphorus chlorides, phosphorus pentoxide, polyphosphoric acid, hydrogen fluoride, and benzenesulfonic acid (McKetta and Cunningham, 1976). It is mentioned that any acid with an ionization constant y greater than 1O5 is suitable to catalyze the process (McKeta and Cunningham, 1976).

The use of strong acid ion-exchange resins as catalysts for making BPA is widespread. With such catalysts longer reaction times and/or higher temperatures (70 to 90C),both undesirable, are required to attain high conversions compared to soluble catalysts. When

using ion-exchange resin as cataiyst the corrosion is minimal and no recycling or disposal

of the catalyst is required. The acidic zeolites for the production of BPA were tested
(Singh, 1992) in the atternpt of a comparative study of preparation of BPA over zeolites

and cation-exchange resins.

other catalysts.

In principle, zeolites should be more shape selective than

The reaction scheme proposed by Singh (Singh, 1992), considering the reaction products
present in large quantities, is:

Phenol

Acetone

Acetone

2,4 'isopropy lidenediphenoi

4'hydroxyphenyl-2,2,4-trimethyl chroman 1

4chydroxyphenyl-2,4,4-trimethy l chroman II

The results show a strong influence of different catalysts on the total conversion of
phenol (see Table 2.2):

Table 2.2: Results of the reaction of acetone with phenol in the presence of zeolites and cation-exchange resin
Catalyst

Reaction tirne

Conversion of phenol
(wt %) 1.O8 4.35 4.6 1 1.12 2.52 2.88 8.74 19.50 20.14

Product distributiona (wt %) 1 II III+-IV Others

59.26 60.00 57-91 38.13 37.42 36.80 85.37 88.72 89.57 16.02 19.08 19-10 28.59 29.56 3 1.25 3.41 4.44 5.06 4.62 11.49 15.19 8.00 9.35 11.12 2.56 3.54 3.54 20.10 9.43 7.80 25.28 23.67 20.83 8.66 3.30 1.83

(hl
Re-Y
5 17 27 4 16 27 5 17 27

H-mordenite

Amberlyst-15

I is bisphenol A; II is 2,4'-isopropylidenediphenol (ogisorner); III is 4 -hydroxyphenyl2,2,4-trirnethyl chroman 1; N is 4'-hydroxyphenyl-2,4,4-trimethy1 chro~ II; Others are lan compounds found o d y in trace quantities (Singh, 1992).

The data in Table 2.2 are plotted and the graphs are illustrated in Figures 2.1,2.2,2.3, and

2.4. They show that in the case of zeolites, Re-Y gives the highest activity (Fig. 2.1).

This rnight be due to its highest concentration of acid sites compared to the other zeolites
used (H-Y, H-mordenite and H-ZSM-5).

The relative activities of various catalysts decrease in the order: ~ r n b e r l y f 15 > Re-Y > H-mordenite > H-Y > H-ZSM-5

The concentration of the undesired products increase in the order:

Amberlyst" 15 c Re-Y c H-mordenite

Fig. 2.1 Conversion of Phenol versus Time for Various Catalysts

Reaction Time [hl

+ H-mordenite

-e-

Arnberiyst-15

Fig. 2.2 Product Selectiviry versus Time for the Reaction CataIytcd by Re-Y

Reactian Timc [hl

+Chroman

op-isomer [II] 1 and Chroman II [III+IV]

Fig. 2 3 Rodud Seleaivity versu T

i for the Readon C t l z d by H-mordenite aaye

+ BPA m
C

opkamer [lq

+amxrian 1 and Chroman II [m+W

-.-OthcrS

Fig. 2.4 Product Selcctivity venus Time for the Reaction Catalyzed by Amberlyst-15

Rcaction Timc [hl

+ BPA [q
t

o p i s o m e r [lu Chroman 1 and Chroman II [III+IV]

The conversion of phenol increases monotonie with time and the higher the concentration of acid sites in the catalyst the higher the conversion (Fig.2.1). However, the activity of the tested zeolites for the formation of BPA is lower than that of the cation-exchange resins. The data also show that the more acidic the catalyst is, the selectivity of the BPA formation is higher (Fig. 2.2,2.3, and 2.4).

The conversion of acetone and phenol to BPA is catalyzed by bases as well as acids;
sodium phenoxyde (C,H,ONa) is particularly specified (McKetta and C m g h a m ,
1976). However, the method is of no use because both yield and quality of product are

inferior.

2.1.1.5.1 Catalyst Enhancers

Both rate of formation and yield in BPA c m be improved by using 1% or less by weight compounds that contain mercapto groups (McKetta and Cunningham, 1976). Some of the compounds containing mercapto groups are su1fur dichloride, sodium thiosul fate, hydrogen sulfide, iron suifide, alkanethiois, arenethiols, thioglicolic acids,

mercaptoalkanesdfonic acids, alkali alkyl xanthates, 2-mercaptobenzothiazote and others (McKetta and Cunningham, 2 976).

This improvement in rate and yield is possible due to the fact that the carbonium ion containing sdfr (CHJIC+SR is more stable than (CH&2+OH. Being more stable, it can

exist in higher concentration in the reaction rnixtare and consequently dkylate faster the phenol ring (McKetta and Cunningham, 1976).

Sulfonated aromatic organic polymers, such as sulfonated polysiyrene, havng organic mercaptan groups , aminoorgano mercaptan groups (Faler and Loucks, 1981, 1982, 1984), N-alS.laminoorgano mercaptan groups (Faler and Loucks, 1983) attached to backbone sulfonyl radicals by covalent nitrogen-sulfur Iinkages have been used as ionexchange resins for making BPA. Also a sulfonated polystyrene ion-exchange resin

having ionically bound N-allcylaminoorgano mercaptan groups was developed (Pressman

and Willey, 1986). These polymers have been developed with the intention of improving

the degree of activity, selectivity and stability of these sulfonated aromatic organic resins.

In 1994 Rudolph developed a catalyst modified with mercapto amines to be used for BPA

and other bisphenols formation (Rudolph et al., 1994). This continuous search for new and enhanced catalysts demonstrates the serious need for improved yields and selectivities in the process of BPA formation.

2.1.1.6 Bisphenols Stabilizers

Malic, glyceric and lactic acids have been found to be highly efficient for the stabilization

of bisphenols. These hydroxy carboxylic acids or their ammonium or alkali metal salts
c m be added to the feed reactants used to make the bisphenols or to the reaction mixture after the reaction is complete or at any time in between. They are particularly useful when the bisphenol is exposed to high temperatures, such as during the separation of the bisphenol tom the reaction mixture , which, in most cases, involves a melting stage (Dachs et al., 1982).

2.1.1.7 Solvents
The viscosity of the reaction mixture may increase as the process advances. Thus it is preferable to perfonn the reaction in a solvent, which h to be inert in the given reaction a condition, to avoid the formation of even more by-products. Suggested solvents are

chlorinated aliphatic hydrocarbons, acetic acid, or aromatic hydrocarbons (McKetta and

Cunningham, 1976). Excess phenol i preferred since it suppresses the condensation of s

acetone with itself and it is easy to recover and recycle. Feeding acetone at successive stages in multistage or cascade reactors rnawnizes the advantages of excess phenol (McKetta and Cunningham, 1976).

2.1.1.8 Reaction Mechanism

Reinicker and Gates (Catana et al,, 1993) suggested a mechanism for the condensation process, for the reactions catalyzed by sulfonic resins. This mechanism involves the formation of hydrogen bonds between the ketone and the sulfonic resin. These bonds were observed experirnentally by IR spectroscopy.

The proposed mechanism consists of the electrophilic attack of a polar reactive intermediate, which c m be a carbonium ion, on the aromatic ring. In the fus1 step the hydrogen bonds are formed between the carbonyl group of the ketone and the sulfonic group of the resin (1). This intermediate is expected to react with the phenol in the nonpolar surrounding medium, forming a tertiary alcohol (II), which transforms rapidly into a carbonium ion (III). The final step, the formation of the BPA molecule, takes place through hydrogen bonds (Fig.2.5). This type of mechanism also explains the formation

of some of the by-products which can appear during the synthesis or during subsequent
processing of the BPA.

Fig. 2.5 Mechanism of Condensation of Acetone with Phenol via Hydrogen Bonds
(Catana et ai., 1993)

2.1.1.9 Reactor Configuration

If the reaction for producing BPA fiom phenol and acetone is conducted in a fixed bed
reactor containing gel-form or macroporous sulfonic acid ion exchanger resins, the volume/time yield c m be improved by providing the resin as a two-layer bed (Berg et al.,

1995) (Fig.2.6):
the lower layer of the bed comprises an unrnodified resin having a low degree of crosslinking, less than or equal to 2%, and comprises 75 to 85% of the bed volume as a whole; and the upper layer of the bed, which comprises 15 to 25% of the bed volume as a whole, comprises either:

a resin having a higher degree of crosslinking than the lower bed, fiom equal to or greater thm 2% to less than or equal to 4%, in which 1 to
35 mol % of the sulfonic acid groups are optionally covered with

species containing alkyl-SH groups by ionic fixing, or

* a resin having a low degree of crosstirking, less than or equal to 2%,

in which 1 to 25 mol % of the sulfonic acid groups are covered with

species containhg alkyl-SH groups by ionic fixing.

.L
Fig.2.6 Reactor Configuration

2.1.2 Alternatives to Acetone as Feedstock

Compounds that react with acid to generate the isopropylic carbonium ion can be generally used instead of acetone. One of the processes semicommercially applied in Russia used propyne (methylacetylene), or a commercial mixture of propyne and propadiene (MAPP), as an alternative to acetone as feedstock (McKetta and Cunningham, 1976). Other processes clairn the use of isopropenyl acetate or 2-chloropropene instead of acetone (McKetta and Cunningham, 1976):

Use of these, like that of (CH3)2C(SR)I types (from acetone and thiols) (McKetta and

Cunningham, 1976), avoids the formation of water as a by-product.

Industrially, the phenol and the acetone are obtained together in the acid cataiyzed
decomposition of cumyn hydroperoxide (C,H,C(CH3)200H). It is thus namal that cmde reaction mixtures, either enriched in phenol by addition or depleted in acetone by distillation thereof (to produce a more suitable ratio of reactants), were used to make BPA (Kiedik et al., 1993). The simplification achieved in this manner is compensated by inferior yields and selectivities.

BPA can be produced with good yields by adding phenol to p-isopropenyl phenol. The
p-isopropenyl phenol necessary for the process is obtained together with phenol fiorn the by-products of BPA manufacture via alkaline cracking at 220C and 55 mm Hg. This way by-products of the BPA formation process c m be transformed in the desired product,

BPA, for an overall improvement of the yield and the selectivity of the process (McKetta and Cunningham, 1976).

It was reported that BPA is formed in a reaction between phenol and a urea-acetone condensation product (McKetta and Cunningham, 1976). The urea-acetone condensation product is presented below:

2.2 Purification
The process used to produce BPA influences the composition of the mixture fiom the reactor. It is still expected to contain phenol, acid cataiyst (unless an acid ion-exchange resin was used), water, BPA, by-products, a thiol promoter, and sorne acetone (if the reaction was not carried out to depletion of acetone) (McKetta and Cunningham, 1976).

For exampie, a cmde product Stream consisted of 4 1% BPA, 36.2% 07p-isomer, 1.1% o,oisorner, 14.2% phenol, 3.5% chromane, 0.05% flavan, and 12% of unidentified compounds (Verkhovskaya et al., 1973). The ratio of BPA to 07p-isomer to chromane in another crude product meam w 90:7:3 (McKetta and Cunningham, 1976). s
The

composition of the BPA usually available on the market is 94% BPA, 4% og-isomer, 3% triphenol1, and 1% chromanes (Anderson et. al., 1959).

Small differences in the operating conditions may have considerable effect on the process
of BPA formation, and different purification processes may be necessary. This results in purification procedures that are numerous and diverse. Since excess phenol is generally used, its removal and recycling is a step found in most purification processes (McKetta
and Cunningham, 1976).

2.2.1 Catalyst Separation

No catalyst separation is required for the resin catalyzed processes. If a homogeneous
catalyst was used than this has to be neutralized, or washed with water, or distilled out in the case of hydrochloric acid. The hydrochloric acid is the most preferred one among the homogeneous catalysts, because it can be recycled and the waste disposal problems are

thus reduced.

The water has to be removed fiom the system whether homogeneous or heterogeneous catalyst was used. It can be removed by stripping with inert gas such as carbon dioxide

or nitrogen, or with benzene. The addition of benzene facilitates the water removal
without the use of vacuum equiprnent (McKetta and Cunningham, 1976). In 1992 Cipullo announced a more effective way of removing the water fiom the cataiyst bed (Cipullo, 1992). The process involves two steps. In the first step 25 to 90% of the water is removed by vaporization.

In the second step the dehydration is completed by

saturating the catalyst with pllenol.

Sometimes the resin catalyzed processes

nui

to 50% conversion of acetone and in such

cases dong with water the h p p i n g removes acetone and some phenol as well. The
acetone and phenol removal c m be minimized by adding a trace of a metal complexing acid before stripping (oxalic, citric, or tartric acid) (McKeta and Cunningham, 1976).

2.2.2 BPA Separation from Crude

The crude is the mixture of products and unreacted reagents that corne out of the reactor. Most of the BPA produced separates as a 1:1 adduct with phenol afier partially stripping

and cooling the crude. This adduct c m be separated by filtration, centrifugation or both.
The phenol adduct can be M e r subjected to a series of processes with the purpose of

separating the BPA fiom the phenol.

These processes may be remelting,

recrystallization, melting and passing over an ion exchange resin (Faler and CipiifIo,

1988), heating in vacuum to distill out the phenol or heating with excess water (McKetta
and Cunningham, 1976). The product may be M e r refined by soIvent treatment or vacuum distillation.

Strong acids can leach fiom the acidic ion exchange resin catalyst into the reaction mixture during the reaction. These acids can decrease the yield and the selectivity of the overall process by cataiyzing the cracking of BPA during purification and finishing steps. Therefore it is important to remove them before starting the purification of the product, and this can be done effectively by an inorganic oxide (Powell and Uzelmeier, 1991).

Formation of the 1:1 BPA-phenol adduct c m be prevented by: operating the process with very little excess phenol, operating the process with acetone and phenol in a molar ratio close' to stoichiometry in inert solvent or to a less than 100% conversion of acetone, vacuum-stripping phenol fiom the crude, or treating the acid-stripped crude, partiy crystallized or not, with excess water, and
steaming to remove remaining thiol promoter (McKetta and Cunningham, 1976).

2.2.2.1 Methods of Separating BPA from the 1: 1 BPA-Phenol

Adduct
Since most of the modem processes for obtaining BPA operate with a high excess of phenol, the formation of the 1:1 BPA-phenol adduct is inevitable; and so new ways of obtaining high quality BPA fiom the said adduct have been investigated. Such a method

has been reported and consists of fusing the adduct in an atmosphere having a maximum
oxygen content of 0.005% by volume, followed by evaporation of liberated phenol (Asaoka et al., 1994 and 1995).

Selective solvents that dissolve the maximum of by-products and a minimum of BPA are used to separate the BPA fYom the 1:1 8PA:phenol adduct. Such solvents are berizene, heptane, cold ethylene dichlonde, a mixture of an aromatic and an aliphatic solvent, weak aqueous alkalies (NaCo,, W O H ) and organic solvent-water emulsions (McKetta and

Cunningham, 1976).

Recrystallization is another effective procedure. The solvents usually used are aromatic solvents like toluene and chlorobenzene, a mixture of a aromatic solvent with a polar n solvent, methanol or a mixture of methanol and ethylene dichloride, 1,1,2,2tetrachloroethane, acetic acid, and isopropyl alcohol (McKetta and Cunningham, 1976).
A newly developed process purifies the BPA by a two stage crystallization procedure

(Sakashita et al., 1993). A system that uses the combined efTect of a filter and a centrifuge was considered in order to minimize the liquid impurities that rernain on the

crystal cake. The crystals are also washed to reduce the surface adherent impurities on
the final crystals.

The dissolution of cmde BPA in caustic alkali, filtration and precipitation with a strong
acid or carbon dioxide (Flippen et al., 1970) is another possibility. Decoiorizing carbon and inorganic salts c m be added, also a reducing agent (sulfite or hydrosulfite) is advisabIe to add to prevent the BPA f?om becoming coloured, as a result of oxidation by

air (McKetta and Cunningham, 1976). Anhydrous ammonia can be used to precipitate
adduct "salts" that can be isolated and dissociated to yield pure BPA (McKetta and

Cunningham, 1976).

Vacuum distillation has already been mentioned (Kiedik et al., 1993) in spite of the
special equipment required. Another disadvantage of this procedure is the tendency of

BPA to decompose at pot temperatures above 200C, especially if acidic or basic

irnpurities are present (McKetta and Cunningham, 1976). In order to avoid

decomposition, thin-film distillation can be performed instead of vacuum distillation

(Pahl et al., 1965). The decornposition can also be reduced by distilling under a nitrogen
atmosphere and dding polypropylene glycol. a secondary or tertiary aikaline earth phosphate, or diethyl malonate before distillation (McKetta and Cunningham, 1976).

2.2.2.2 By-Products Isomerization to BPA

BPA by-products can be isomerized to BPA in the presence of an acid catalyst (which
can Se an ion-exchange resin or hydrogen chloride) and a fiee mercaptan CO-catalyst (Li,

1989). The alkaline cracking at 220C and 55 mm Hg of the by-products to yield phenol
andp-hydroxy-isopropenynil phenol that c m be recycled to the process has aiready been

mentioned (McKetta and Cunningham, 1976).

This high temperature is necessary

because the chromanes are relatively refractory and tend to build up in recycle strearns (McKetta and Cunningham, 1976). The chroman can also be isolated and purified as a crystalline ciathrate. The BPA can also be regenerated with good yields fiom scrap resins (McKetta and Cunningham, 1976).

2.3 Manufacturing
The most industrially used processes for making BPA in the 'Jnited States and Western Europe are the acetone-phenol ones, in homogenous as weIl as heterogeneous catalysis. Considering the costs involved and the net advantages the heterogeneous catalysis offers,

the resin-catalyzed process is preferred and it has been improved continuously.

A process which considers reacting acetone with very Iittle excess phenol (1:4 to 1:12 molar ratio acetone:phenol in the initial reaction mixture) was reported (Iimun, et al.,

ut 1990). The reaction stage of this process comprises of two steps. In the f s stage the
acetone is reacted with very little excess phenol in the presence of a sulfonated cation exchange resin catalyst modified with a rnercapto goup-containing compound to convert

20 to 60% of acetone. In the second stage the

reaction mixture fiom the first step is

reacted in the presence of hydrochloric acid as catalyst.

Although the literature shows that processes using alternative feeds, such as a postreaction mixture resulting fiom the synthesis of phenol and acetone, are not convenient because of the great variety of by-products and the infenor yields, such a process has been developed and it is now industrially used in the United States.

Accordingly, three flow sheets are presented in this chapter:

a) the resin-catalyzed process using acetone and phenol;
b) the hydrogen chloride-catalyzed process ; and

c) the resin-catalyzed process using a post-reaction mixture of the cumylhydroperoxide decomposition.

2.3.1 Resin - Catalyzed Process

A process catalyzed by a sulfonated cation exchange resin modified with 2-

mercaptoethmol is presented in Fig. 2.7 (McKetta and Cunningham, 1976). A mixture

consisting of 83.4% phenol, 5.1% acetone. 0.1% water, 3.4% recycled BPA and 8.0% recycled by-products are preheated and fed to the reactor. The reactor is operated at about 75C. The residence time is set at one hour. The process runs to a 50% conversion of acetone (McKetta and Cunningham, 1976). Aithough not stated in the reference. the

u i s for product distribution are most likely to be wt'X0. If the units were mol%, the nt
molar ratio of acetone to phenol would be about 1:16, which is undesirable since it would favour the adduct formation.
ACETONE ACETONE WATER

PHENOL

MAKE-UP
ACETONE PHENOL

1 4-

+t

l
2

PHENOL ACE3ONE

ACETONE WATER

PMNOL
\I

,. +

3
BYPRODUCT

<wASH PHENOL
V

'4'~

BPA PHENOL ADUCT

PHENOL. BYPRODUCT, BPA RECYCLE

Fig. 2.7 Production of Bisphenol A with Resin Catalyst (McKetta and Cunningham, 1976) 1-Feed tank; 2-Reactor; 3-Concentrator; 4-Crystallizer; 5-Solid-Liquid separator; 6-Melter; 7-Flaker; 8,9-Distillation columns; 10-Phenol stripper. The reactor effluent, together with some recycled phenol, BPA and by-products go to the
concentrator.

The concentrator is operated at 200mm Hg.

The overhead at the

concentrator is a mixture of acetone, water and phenol (18 to 20%). The boton Stream consists of phenol, BPA and by-products. The overhead passes through a series of

distillation columns to remove the water fiom the acetone and the phenol, which are recycled to the reactor. The bonom Stream from the concenmtor goes to a crystdlizer where it is cooled d o m to separate the BPA as phenol adduct. Afier crystallization the mixture is separated in a centrifge, washed with phenol, and fieed of phenol by melting at 130C, then stripping in a column at 200C and lmrn Hg. The purity of the product obtained with this process is over 90%. The phenol recovered in the sipper is recycled to the centrifuge and the centrifuge liquor is recycled to the reactor (McKetta and
Cunningham, 1976).

2.3.2 Hydrogen Chloride - Catalyzed Process

A process that uses hydrogen chloride as cataiyst is presented in Fig. 2.8 (Pahl et A.,

1965). A version of this is used by Mitsui Chemical in Japan and by General Electric in the United States (McKetta and Cunningham, 1976). A mixture of excess phenol, acetone, BPA and by-products fiom the recycle strearns are saturated with gaseous hydrochlonc acid and fed to the reactor. The reactor is operated at about 50C. The mixture is reacted for several hours under continuous stimng. The effluent fiom the reactor undergoes a preliminary stripping that removes a two-phase mixture of hydrochloric acid, water and some phenol. This overhead goes to a decanter where the two layers separate. The hydrochloric acid is recovered fiom the aqueous phase and recycled. The water goes to the drain. The stripped crude is fed to a senes of separation columns and successively freed of phenol in the phenol still (at about 10 mm Hg) and of o,p-isomer in the isomer still. The phenol and by-products separated in this stage are

recycled to the reactor (McKetta and Cunningham, 1976). The impurities with higher boiling points are separated fiom BPA by vacuum distillation in the BPA still at 1 to 5 mm Hg. The BPA overhead is mixed with some solvent (e.g. benzene) under pressure while molten, then cooIed in the crystallizer to cause crystallization. The purified crystals are separated in a centrifuge and then dried for a high quality product. The liquor fiom the centrifuge goes to a solvent d l . The by-products separated at this stage are recycled to the reactor and the solvent is stored for subsequent uses (McKetta and Cunningham, 1976).
HCL

, ,
/

H a RECYCLE

Fig. 2.8 Production of Bisphenol A with Hydrogen Chloride Catalyst (Pahl et al., 1965) 1-Reactor; 2-HC1 still; 3-Decanter; 4-HC1 recovery column; 5-Solvent still; 6-Solvent storage; 7-Phenol still; 8-Isomer still; 9-BPA still; 10-Crystallizer; 11-Centrifuge; 12-Dryer

2.3.3 Resin - Catalyzed Process II

The resin catalyzed process for obtaining bisphenol A fkom a post-reaction mixture resulting fiom the step of synthesis of phenol and acetone (Kiedik et a. 1993), l, represented by Fig. 2.9, uses a vertical drurn reactor filled up to 70% with a mixture composed of 70% microporous Wofatit-KPS cation-exchange resin and 30%

macroporous Wofatit-PK- 11 O cation exchanger, operated at 85C.

The reactor feed at steady-state operation consists of 55.4 wt% phenol, 6.8 wt% acetone,
0.6 wt% water, 18.9 WWO BPA and 18.3 wt% by-products, including 4.4 wi% o,pisorners.

The process consists of the following steps: 1. Reaction of phenol with acetone, reaction of phenol with p-isopropenylphenol resulting from thermal decomposition of process by-products and recycled to the reaction systern, and isomerizational rearrangement of process by-products to obtain

BPA;
2. The post-reaction mixture together with water (1-4% by weight) and acetone (2-65% by weight) is cooled down to 40C to obtain a precipitate of BPNphenol in phenolic solution;
3. The precipitate is separated by centrifugation into crystalline BPNphenol adduct and

phenolic mother liquor 1. The crystalline BPNphenol adduct is washed with mother

Iiquor II, obtained in step 5, in an amount of 0.2-2.0 parts by weight of the liquor per
1 part by weight of the crystailine adduct;

1 7

~-

H20

A
ACETONE

PHENOL

v
1
L

v
&

&
5

Fig. 2.9 Production of Bisphenol A with Resin Catalyst II (Kiec

4. The BPNphenol adduct is dissolved using the mother liquor II obtained in step 5

andor phenolic solution obtained in step 7; 5. The precipitate obtained in step 4 is separated into t e crystalline BPNphenol adduct h
and mother liquor II to be tumed back to step 3 of the process. The BPNphenol

crystalline adduct is washed with fiesh and regenerated phenol obtained in step 6 and used in a ratio of 1-3 parts by weight of fiesh phenol per 1 part by weight of regenerated phenol;
6 . A high-purity BPA is separated fiom the BPNphenol adduct by distillation at 160C

and 10 mm Hg of a substantial volume of phenol and the phenolic residue is removed

by steam stripping;

7. The phenolic mother liquor 1obtained in step 3 is distilled to remove the acetone, the

water, and part of phenol. The volume of phenol distilled off the rnother liquor I is
; 0.1-0.3 parts by weight per 1 part by weight of mother liquor 1

8. The mother liquor I obtained in step 3 a d o r 7 is exposed to a themal catalytc

decomposition in an amount of 0.05-0.2 parts by weight, resuiting a distillate comprising phenol, isopropenylphenol, and process by-products. n i e catalytic

decomposition is conducted in the temperature range of 200"-300C, and in the absolute pressure range of 1-50 mm Hg in the presence of catdyst selected fiom the group of: Na&IPO,, NaHCO,, NaOH;

9. The cataiytic rearrangement of the reactive components of the distillate obtained in

step 8, while leaving the p-isopropenylphenol contained therein substantially intact, is conducted in the presence of oxaiic acid used in an amount of 0.05-0.5% by weight.
The rearranged distillate is recycled to step 1 of the process.

The composition of the product obtained is: 24% BPA, 16.2%by-products including 4.8%
og-isomers, 52.95% phenol, 5.65% acetone and 1.2% water. The bisphenol A product shows the following properties: crystallization point 156.8"C, coloration of 50% solution
4 APHA, o,p-isomer in trace amounts, codirner in trace amounts, trisphenol 15ppm,

principal product 99.96% by weight.

2.4 Physical Properties

Bisphenol A is a white crystalline solid. appearing like small white to light brown flakes or powder, with mild phendic odor, which sinks in water. !ts specific gravity is given as 1.195 at 25/2S0C. There i no data regarding its vapor density. For the boiling point s records show discordant temperature ranges and imprecise pressures, e.g., 181 to 195OC at 4 mm Hg, 195 to 200C at 6mm Hg, 230C at 7.6mm Hg, 210 to 220C at 4 mm Hg,
and 230C at 5 mm Hg. The value found in the Material Safety Data Sheets for the

boiling point is 220C. Other sources (NTP Chernical Repository, 1991) suggest 250252OC at 13 mm Hg. As it might be suspected, BPA is volatilized only i traces by s t e m n at 1 atm. Pure %PAmelts at about 157OC;no highly precise and reliable value has been published, although many are on record. The heat of fusion is 30.7 c d g (McKetta and Cunningham, 1976).. The density of the monoclinic pnsmatic crystals is given as 1.13
g/ml or 1.195g/ml (McKetta and Cunningham, 1976).. The heat of combustion is 1869

kcdmol and AH,=88.2f O S kcaUmol (McKetta and Cunningham, 1W6).. The flash point is 2 13OC (McKetta and Cunningham, 1976)- Some values of the solubilities are given in Table 2.3 (McKetta and Cunningham, 1976).

Based on the partition coefficients for BPA between water and some organic solvents; it

can be concluded that the alkanes are the poorest extractants, aromatic solvents are much better, and alcohols and esters are the best (Korenman and Goronkhov, 1973). Table 2.4 contains data regarding the variation of the BPA vapor pressure with the temperature.

Table 2.3: Solubilities of Bisphenol A in Various Solvents (g/100g solvent) (Korenman

and Goronkhov, 1973) Temperature "Room Temperature" 18C (except as specified)
0.035' 1.5 1.1 0.05 3 -6 0.2

Solvent Hz0 CH,CI, CHCI, CCI, ClCH,CH,CI ClCH=CCI, CH,OH CZKOH CH,COOH (CH,),CHOH (CHJzCO C6E cyclo-C6H,, a cbCold"
"5C
c"Hotm

Boiling Point (except as specified)

<O. lb
0.86 0.03 7 0.94 0.08

08 .'
6-7

3-4
0.8-1 .O 20

2-3

210b
150 2 1.6
48

200

0.9

Ub; 109 0.61; O.gb 0.0023

7-8

Table 2.4: Variation of vapor pressure with te~perature(McKetta and Cunningham,

1976)

pressure (mmHd

1 1 1 1 1 1 1 1 1 1 1 1
0.2 1.0

50 .

10.0

20.0

40.0

60.0

100.0

200.0

400.0

760.0

25 Chemical Properties .
Bisphenol A reacts as a typical para-substituted phenol. One or both hydroxyl groups, one or both rings can experience modifications. Transformations involving the aliphatic methyl groups of the bndging group can also take place (McKeta and Cunningham,
1976).

BPA is convexted by caustic alkalis into its soluble alkali salts (McKetta and
Cunningham, 1976):

These s d t s are easily alkylated with alkyl halides, such as allyl chloride, to f o m diethers

(McKetta and Cunningham, 1976):

BPA cm undergo cyanoethylation, with basic catalyst, to f o m dinitriles that c m be hydrogenated to diamines (McKetta and Cunningham, 1976):

Dow and ICI Amenca produced ethers for use as components of unsaturated polyesters, (polyesters of fumaric acid for example), and for coatuigs applications by reacting BPA
with epoxides (McKeta and Cunningham, 1976). In this reaction the phenolic groups are

hydroxydky Iated:

BPA reacts with epichlorohydrin to form a bis(chlorohydroxypropyl) ether which yields

the diglycidyl ether (DGEBA), the monomer for most epoxy resins, in a caustic medium (McKetta and Cunningham, 1976):

DGEBA
The phenoxy resins are produced when BPA is condensed in a 1:l ratio with epichlorohydrin, so that the monomer units altemate in a linear polyrner ('McKetta and
Cunningham, 1976):

c H,
Phenoxy resin pattern

Polymers with terminal phenolic groups are obtained when reacting BPA with Iess dian
one molar equivalent of a dihalide such as bis(2-chloroethyl ether) or 1,4-

bis(chloromethy1)benzene.

Commercial polysulfone resins are manufactured when

reacting stoichiometrc amounts of BPA and bis(4-chloropheny1)sulfone (McKetta and Cunningham, 1976; Hill et al., 1992):

Polysulfone resin pattern Polycarbonates are obtained by esterification of BPA with phosgene or its dibenzoate ester (McKeta and Cunningham, 1976). Other diacid chlorides have been also reacted
with BPA to obtain polycarbonates (Shaikh and Sivaram, 1995).

Polycarbonate resin pattern Poly(ary1enecarbonate)s oligomers cm be obtained by carbonate interchange reaction of dimethyl carbonate with BPA (Shaikh et al., 1994):

BPA can be converted to a bis(alky1 carbonate) and fiom there to similar poiyrners by
reacting it with aliphatic esters of the carbonic acid (McKetta and Cunningham, 1976):
/

+ O=C,

CI
Base, R-O-C-

0-R

II O

One of the side reactions that can occur in the melt polycondensation, one of the processes used for manufacturing polycarbonate resins, is generated by the instability caused by the hydroxyl groups. temperatures exceeding 150C: Highly reactive isopropenylphenol is produced at

Aromatic polyesters cari be obtained by transesterification of BPA with dimethyl

terephthalate/isophthalate. The process has two steps. Ln the first step the aromatic
polyester prepolymer is formed (Mahajan et al., 1996):

In the second step the prepolymer eliminates methanol and yields a high molecular weight aromatic polyester (Mahajan et al., 1996):

300-330C O.5 Torr, Catalyst

'
Polyester

The aromatic protons adjacent to the hydroxyl groups in BPA are easily substituted. The
halogenation of the aromatic rings in the ortho positions relative to the hydroxyl groups is usehl for rnaking flame retardants (McKetta and Cunningham, 1976):

The typical catalyst for chlorination is aluminium chloride and the process is performed in chlorinated aliphatic solvents. The solvent used for bromination is acetic acid or a lower alcohol with chlorine added concurrently (McKetta and Cunningham, 1976).

Polyphosphate esters can be also used as flarne retardants.

BPA is reacted with

phosphorodichloridates, prepared from alcohol and POCI, (&shore et al., 1988):

In order to create new useful monomers, bisphenol A was reacted with

tetranuoroethylene (TFE) and carbon dioxide in dimethyl sulfoxide, in the presence of an

aqueous solution of sodium hydroxide to give the salt of a carboxylic acid, which is conveniently isolated as its methyl ester after reaction with dimethyl sulfate (ArnoldStanton and Lemal, 1991). This ester can be M e r tramformed in the correspondhg diol, diamine, diisocyanate and bis(methy1 carbarnate) which can be valuable monorners for tailored polyurethanes, for example.

H O ~ F ~ O H
+~CF+X

CO,

Base DMSO (cH~~)~so:

'

Usefid stabilizers and antioxidants for rubbers and other plastics can be obtained by acidcatalyzed allcylation of BPA with reactive olefins such as isobutylene and styrene. The condensation of BPA with formaldehyde was used in the past to obtain phenolic resins (McKetta and Cunningham, 1976).

By reacting BPA with formaldehyde and methylarnine, using dioxane as solvent, a benzoxazine is formed (Ning and Ishida, t 994):

BPA can participate in other reactions as nitration, sulfonation, aminomethylation, Kolbe

reaction, nitrosation, and diazo coupling (McKetta and Cunningham, 1976).

The hydrogenation of BPA to the isopropylidenedicyclohexanol is described by several

references. BPA is rapidly hydrogenated at 75aC and 365 psi in 2-propanol with 5% rhodiurn/carbon as catalyst. The isopropylidenedicyclohexanol is used as a di01 to improve the chernical resistance of polyester resins.

BPA is decomposed by heating in hydrogen.

If the process is performed at high

hydrogen pressure, it produces only phenol. If the process is performed at low hydrogen pressure, it produces phenol and some isopropylphenol as well. Pyrolysis of BPA yields phenol, p-isopropylphenol, and residual tars. (McKetta and Cunningham, 1976): The acetates of BPA also decompose

However, p-isopropenylphenol is best obtained by cracking BPA in the presence of bases, whereupon this alkenylphenol and phenol are obtained in yields of over 90%. Pisopropylphenol c m be oxidized with hydrogen peroxide in acid solution to yield hydroquuione. By autoclaving the aqueous alkaline solution, the decomposition of BPA
can go as far as obtaining acetone and water (McKetta and Cunningham, 1976).

The electrolysis of a concentrated aqueous solution of BPA conducted on a platinum mesh occurs with total degradation of the aromatic rings, leading in the end to simple short chain aliphatic acids. This procedure is used for BPA removal fiom wastewaters.

BPA forms solid adducts with phenol and cresols. The formation of these products is not
well understood. They are used in the process of BPA purification (McKetta and

Cunningham, 1976).

The synthesis routes available to produce BPA, catalysts, reaction rnechanism, purification issues, physical and chemical properties of the bisphenol A have been reviewed in this chapter. It is clear that the number of synthesis routes available to produce BPA is quite impressive. This study also revealed that the purification process is very cornplex. This is due to the fact that in the given conditions, al1 the compounds involved in the process are very reactive, and they c m interact with themselves or with each other to form a variey of compounds whch are also very reactive. This is the reason why there is a need for new catalysts, which are more selective to the production of BPA.

Another important finding is that acetone and phenol are preferred as reagents for this reaction over some alternative feeds, since a higher purity crude BPA is obtained. Consequently it was decided that the synthesis of BPA in this investigation will be pursued via condensation of acetone and phenol with acidic heterogeneous catalyst. The

reason for considering heterogeneous catalyst is the fact that the final purpose of this

research is to develop a process based on cataiytic distillation.

Another fact the literanire review has revealed is that the higher the acidity of the catalyst,
the better the yield and the selectiviv of the process of BPA formation. This finding

suggested the idea of investigating the suitability of solid superacid catalysts, which have been tried successfully for various reactions, such as alkylations, acylations, isomerizations, hydrations and dehydrations, esterifications, etherifications, nitrations,
and disproportionations.

Chapter 3

Gibbs Reactor Simulations

Gibbs reactor simulations are used to calculate equilibrium yields, compositions and phases of a reaction mixture. Kinetic factors are not considered in the Gibbs reactor simulation. Consequently it is not possible to detemine how long it will take to reach equilibrium for a given systern. The general theory is discussed in many references (i.e.

Smith and Missen, 1991).

The purpose of Gibbs reactor simulations is simulation is to better understand the reaction which leads to BPA formation. One of the interests is to narrow down the experimental
region with respect to the molar ratio acetone:phenol. It is also intended to evaluate the behavior of the process in the range of temperature mentioned by the 1itera-e

as

feasible. The results would be useful in determinhg the levels for the experimental design as well.

3 1 The PRO II@ . Gibbs Reactor

In this study the Pro II" implementation of the Gibbs reactor is used. The particular PRO

II" implementation is discussed in the users manual (Reference Manual, 8 1994 - 1997).

In order to calculate the Gibbs fiee energy of the cornponents it is necessary to estimate or specia activity coefficients for the components (Van Ness, 1982). This requires selection of an appropriate thermodynamic method for the specified mixture. The

thermodynamic method needs to account for the interactions among species. In this study NRTLO l (non-randon two-liquid) thermodynamic method with a UNIFAC fil1 (universal functional activity coefficient), was selected, as being the most appropriate
(Van Ness, 1982). The NRTL equation was developed by Renon and Prausnitz (Smith

and Missen, 1991) to make use of the local composition concept. The UNIFAC rnethod was developed in 1975 by Fredenslund, Jones, and Prausnitz (Smith and Missen, 1991).

This method estimates activity coefficients based on the group contribution concept
following the Analytical Salution of Groups (ASOG) mode1 proposed by Derr and Deal in 1969. Interactions between two molecules are assurned to be a function of group-group interactions. Whereas there are thousands of chemical compounds of interest in chemical processing, the number of functional groups is much smaller. Group-group interaction data are obtained fiom reduction of experimental data for binary component pairs.

In the PRO II@ implementation it is possible to:

a) use library data for components;

b) estimate activity coefficients;

c) input ail data; d) al1 of the above.

If Iiterature or Iibrary data are not available, group contribution methods do not allow
distinction between isomers. For a non-ideai mixture, the results of the simulation are expected to be sensitive to the thermodynamic method and the quality of the input data.

3.2 Simulation of the Bisphenol A Reaction

The following components were present in the Pro II" component library : water, acetone,
phenol, bisphenol A, isophorone, mesityl oxide and mesitylene.

The last three

components, isophorone, mesityl oxide and mesitylene, are potential by-products. Also

as by-products, which were not present in the library, were considered: 2,4'isopropylidenediptienol (o,p-isomer), chroman and triphenol. These nonlibrary

components were individually defined and their thermophysical properties were fiIled f7om structure. The thermophysical properties of the library components required for the Gibbs calculations were provided by the interna1 database of Pro II". In order to

differentiate between the p-p and the O-p isomers of the bisphenol, the normal boiling

point for BPA was cfianged with the value of 493 K, found in the literature (Catana et al.,

1993) (the llbrary value was 633.65 K), also sorne thennodynarnic data were supplied for
the nonlibrary isomer (opisomer). These data were found in the literature (Catana et al.,
1993) and are as follows:

Enthalpy of formation: -3.6928~10~ (kJ/kmoI), Solubility parameter: 9.6034 (caUcc)"O.S, Normal melting point: 383.15 K, Carbon number: 15, Hydrogen deficiency number: 14.

The supplied data for the simulations consisted of reactants, expected products, and reaction conditions. The reaction parameters are summaiized in Table 3.1. The input file
for these simulations is listed in Appendix B.

Since the investigated reaction is carried out in liquid phase, the pressure should not influence the results. After reviewing the data found in the literature about the pressure at
which the reaction is perforrned, the simulations were conducted at atmospheric pressure,

at different temperatures and different acetone:phenol rnolar ratios.

Isothemal and

isobaric operating conditions were selected. The reactor temperature was varied fiom 50 to 90C,the molar ratio acetone:phenoI was varied from 1:19 (0.05) to 1 :1.5 (0.67), and
the pressure was maintained constant, at latm, in order to determine the effect of

temperature and of the molar ratio acetone:phenol on conversion (Table 3.1). Since it is not essential for a full solution, no reaction set was input. In a Gibbs reactor the extent of
the reaction is deterniined by minimizing the overail fiee energy of the reacting species.

Table 3.1: Gibbs reaction simulation with reacuon parameters

Simulation #

Press (atm)
1 1 1 1 1 1 1 1

The extent of the reaction is a measure of how far the reaction goes towards completion and what proportion of the reactants are converted into products. There were also no conmaints imposed on the products.

3.2.1 Analysis of the Simulation Results

The results provided by these simulations are a steady state final value. Therefore, the

time dependence of the reaction is not available. The maximum number of iterations was
lefi at the default value of 50. The Gibbs reactor does not require any information about catalysts, since it does not consider kinetic eEects. Hence, these simulations do not allow the possibility of investigating the effect of various catalysts on either yield or selectivity

of the process. The numencal data for these simulations is tabdated and presented in
Appendix C.

3.2.1.1 Effect of Temperature and Acetone:Phenol Molar Ratio on BPA Formation

Figures 3.1, 3.2 and 3.3 show the variation of bisphenol A formation with the acetoneqhenol molar ratio and the temperature, and the variation of the selectivity with the two factors.

Fig. 3.1 Variation of BPA Formation with MoIar Ratio Acetone:Phenol

Acetone:PhtnoI [molar ratio]

Fig. 3.2 Variation of BPA Formation with Temperature

and Molar Ratio Acetone:Phenol

Temperature [KI

Fig. 3.3 Variation of SeIectivity of BPA with Temperature and MoIar Ratio Acetone:Phenol

Temperature [KI

The results can be summarized as follows:

The yield of BPA formation increases by hcreasing the acetone:phenol molar ratio up to a value of 0.5, and then it decreases (Fig. 3.1).

For acetone:phenol molar ratios between 0.54 and 0.67, the yield of BPA formation
decreases with the temperature (Fig. 3 -2). For acetone:phenol molar ratios between 0.25 and 0.43, the yield of BPA formation presents an optimum at around 343.15 K (Fig. 3.2). For acetone:phenoI molar ratios between 0.05 and 0.18, the temperature has a negfigible effect on the yield of BPA formation (Fig. 3.2).

For molar ratios srnaller than 0.5 and temperatures higher than 333.15 K boh the
temperature and the molar ratio have no significant influence on the selectivity (Fig.
3 -3).

6 . For 0.5 molar ratio the temperature has a negligible effect on the selectivity of the

BPA formation (Fig. 3.3).

7. For molar ratios greater than 0.5, the selectivity decreases with both the rnolar ratio

and the temperature (Fig. 3.3).

The decrease in both yield and selectivity of BPA formation for acetone:phenol molar ratios under 0.5 is due to the fact that these rnolar ratios are smaller than the stoichiometric ratio required by the process (Fig. 3.2 and Fig. 3 -3). Therefore, some BPA is formed d l the phenol is consumed. After this point, no more BPA is produced, and

the BPA aiready formed is probably reacting with acetone or other products formed in the reaction to give heavier by-products. By heavier by-products it is understood products with molecular weights higher than the bisphenol A.

The insignificance of the effect of temperature on selectivity for molar ratios greater than
0.5 can be explained by the fact that the formation of BPA is definitely favored in

comparison with the other cornpetitive reactions, which lead to by-product formation.

3.2.1.2 By-Product Formation

For the simulations, a list of reactants and expected products was supplied. It was observed that the chroman appeared only in traces in the product stream, therefore it was disregarded. Also isophorone, mesityl oxide and mesitylene did not appear in the product stream and they were disregarded as well. Figures 3.4 to 3.9 show the variation of the

yield and the selectivity of the by-product formation with the temperatrue and the molar ratio acetone:phenoI.

The results can be summarized as follows:

1. The o,p-isomer formation shows a maximum at around 0.5 molar ratio acetone:phenol
for temperatures tower than 343.15 K (Fig. 3 -4).
2. For temperatures of 343. L 5 K and higher the formation of op-isomer is uisignificant

regardless of the molar ratio acetone:phenol (Fig. 3 -5).

3. The formation of triphenol is positively influenced by increases of both the

temperature and the acetone:phenol molar ratio (Fig. 3.7 and Fig. 3.8).
4. Some iphenol starts to form only for acetone:phenol molar ratios over 0.54 (Fig.

Fig. 3.4 Variation of 0.p-tsomer Formation with Molar Ratio Acetone:Phenol

Acetone:Phenol [moIar ratio]

Fig, 3.5 Variation o f oq-Isomer Formation with Temperature at Various Molar Ratios Acetone:Phenol

320

330

340

350

360

370

T e m p e n t u r e [KI

Fig. 3.6 Variation of Selectivity o f og-Isomer with Temperature at Various Molar Ratios Acetone:Phenol

Temperature [KI

Fig. 37 Variation o f Triphenoi Formation with Molar Ratio Acetone:Phenol .

Acetone:Phenol [rnolar ratio]

Fig. 3.8 Variation of Triphenoi Formation with Temperature at Various Molar Ratios ~cetone:~henol

Fig. 3.9 Variation o f Selectivity of Triphenol Formation with Temperature at Various Molar Ratios Acetone:Phenol

5. For acetone:phenol molar ratios of 0.54 and 0.67 the formation of triphenol increases

significantly with the temperature over the whole range (Fig. 3.8).

6 . For acetone:phenol molar ratios smaller than 0.54 some triphenol is formed, but the
amounts are not significant even at high temperatures within the considered range (Fig. 3.8).
7. The molar ratio has negligible effect on the selectivity of o,p-isomer formation (Fig. 3 -6).

8. The selectivity of the op-isomer formation decreases with the temperature (Fig. 3.6).
9. At 353.15 K, regardless of the acetone:phenol molar ratio, the quantity of o,p-isomer

formed is negligible and it continues to decrease with increasing temperature (Fig. 3.5

and Fig. 3 6 . .)

10. For acetone:phenol molar ratios greater than 0.5, the selectivity of triphenol formation increases significantly with the temperature (Fig. 3 -9).
11. For acetone:phenol molar ratios smailer than 0.5, the variation of the selectivity of

triphenol formation with the temperature is significant (Fig. 3.9).

12. For acetone:phenol molar ratios smailer than 0.5, the variation of the selectivity of

triphenol formation with the acetone:phenol molar ratio is significant (Fig. 3.9).

Both yield and selectivity of triphenol formation are significant for acetone:phenoI molar ratios over 0.5, because, as shown in the previous chapter, triphenol 1 is formed via condensation of an aiready formed molecule of BPA and a molecule of hydioxy isopropyiidene phenol (the product of the fnst step of condensation in the process of fonnation of BPA). This is also another explanation for the decrease of the yield and the selectivity of the BPA formation for molar ratios of the initial reagents over 0.5.

The o,p-isomer formation is favored by low temperatures because it is an exothermic process. Therefore, the higher the temperature. the lower the yield and the selectivity of
the o,p-isomer formation. Also, the closer the ratio acetone:phenol to stoichiometry, the

lower the amount of op-isomer formed, because given the same conditions, the fonnation of BPA is thermodynamically favored.

3.3 Summary of the Simulation

Both the formation of bisphenol A and oop-isomershow a maximum at around 0.5 molar
ratio acetone:phenol, while the triphenol seems to just start fomiing at the specified molar
ratio. From the analysis presented it appears that there is an optimum at around 0.5

acetone:phenol molar ratio and temperatures somewhere between 343.15 K and 353.15

K. In order to better visualize the optimum, 3-D graphs are presented in Figures 3.10.
3.1 1, and 3.12.

Fig. 3.10 Variation o f BPA Formation with Temperature and Molar Ratio Acetone:Phenol

Fig. 3.11 Variation of o,p-Isomer Formation with Temperature and MoIar Ratio Acetone:Phenol

Fig. 3.12 Variation o f Triphenol Formation with Temperature and Molar Ratio Acetone:Phenol

We are looking for reaction conditions which maximize the amount of bisphenol A and muiimize the amounts of og-isomer and iphenol. The optimum for the system is

positioned somewhere around 353.15 K and 0.5 molar ratio acetone:phenol. Figure 3.13 represents the formation of the three products at 353.15 K over the considered range of acetone:phenol molar ratios. Figure 3.14 represents the formation of the three products at
0.5 acetone:phenol molar ratio over the considered range of temperatures.

Fig. 3.13 Variation of BPA, op-Isomer, and Triphenol Formation with MoIar Ratio Acetone:Phenol at 353.15 K

+BPA

. - O - - Isopmpylidcnc

+Triphcnol

Bisphcnol

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Acetone:Phenol [molar ratio]

The equilibrium constant for the BPA

o,p-isomer transformation calculated fiom the

simulated data are presented in Table 3.2. These data are not in agreement with the data presented in Table 2.1 (page 15). The reason is very likely to be the fact that limited physical property data were available for the o,p-isomer and that the physical properties were estimated with the group contribution methods. As previously mentioned, the group contribution methods are unable to differentiate between isomers.

Table 3.2 Equilibrium constant for the BPA

opisomer transformation based on

simulated data Temperature ( O C ) K

50 0.4348

60 O. 1137

70 0.0322

80 0.0097

90 0.0039

Fig. 3.14 Variation of BPA. o.p-Isomer, and Triphenol Formation with Temperature

3.4 Conclusions
As presented in the previous chapter, the Literature mentions that the process of BPA

formation is favored by excess phenol.

The simulation shows that actually a

stoichiometric ratio of acetone and phenol is a better choice.

The simulation also

concludes that the considered ranges for both the temperature and the molar ratio acetone:phenol are feasible experimental regions.

The Gibbs reactor simulations has been used to speci@ the experimental region with respect to the molar ratio acetone:phenol.

It has dso contributed to a better

understanding of the reaction behavior and should give some indication of what to expect from the experimental program in terms of yields and selectivities.

These simulations have confmed that the species present in the system are remarkably reactive at the given conditions. This was one of the conclusions of the previous chapter
as well. Therefore there is a considerable need for catalytic systems andor reaction

conditions which will produce a cleaner crude, which is translated is less complex and less expensive separation processes.
%y cleaner crude is understood a mixture of

products and unreacted reagents that corne out of the reactor with as few by-products as possible.

The literature review in the previous chapter indicated that the formation of chromanes as by-products of the BPA production process has been observed. The simulations showed
that the chromanes are produced in negligible quantities, and the triphenols are actually

n present in larger quantities i the product stream.

Subsequent work involved experimental investigation of the process simulated in this chapter. The effects of temperature, molar ratio acetone:phenol, and catalyst on BPA and by-products formation and selectivity are analyzed and compared with the results of the simulation and the findings in the literature.

Chapter 4

Experimental Investigation

The purposes of the experimental investigation were to:

1. Identiw new suitable catdysts for BPA formation;

II. Find appropriate reaction conditions; III. Determine the reaction conditions that significantly innuence the process and explain
their influence.

The apparatus and the materials used in the experimental investigation, the analytical

methods employed to analyze the samples, the experimentai procedures employ ed, and
the methods used for data processing are discussed in this chapter. Safety issues are also discussed.

4.1

Apparatus - Overview

Different setups were used for different stages of the experimental investigation. The stage of identi@ing new catalysts was carried out in NMR (Nuclear Magnetic Resonance)

tubes. A brief overview of the theory underlying the NMR phenornenon and associated references is given in Appendix D. This setting provided qualitative information about the formation of the product of interest while using very srnall amounts of catalysts and reagents. Basically at this stage the reaction was studied i a drop of reaction mixture. n Another advantage was the fact that the sampling was automatic. The result was the identification of
IWO

suitable catalysts: ~ a f i o n @ activated alumina acidified with and

hydrochloric acid.

Once this phase was accomplished, a new experimental setup became of interest. This

was a batch reactor. The arrangement was sirnilar to that descnbed by Singh (1992)

(Singh, 1992). A 250 ml, 3 neck round bottom flask with a central opening, imrnersed in a thermostated paraffinic oil bath, was used as batch reactor. A condenser was placed in
the central opening to reflux the vapors of acetone and other vapors which may be

produced during the reaction. A magnetic stirrer was employed for agitation of the catalyst. The temperature was monitored by a thennometer inserted via the second neck of the flask. The third opening remained plugged with a stopper and was used for removuig samples. The batch reactor was used to perform the experiments with ~a.fion@
and ~mberlyst@'15. The setting was not suitable for the reaction with acidified activated

dumina because this catalyst was not robust enough to withstand mixing.

And fnally a third setup was considered, the plug flow reactor (PFR), for the reactions
with acidified activated dumina. A PFR with configuration as shown in Figure 4.2 was

used. This type of reactor was selected to sirnulate reactive distillation conditions. The reactor consisted of volatizkg zone (an insulated circular pipe with length and radius of approximately 19 cm and 1.4 cm respectively) filled with Ioosely packed ceramic saddles followed by a reaction zone filled with the cataiyst (a circular pipe with length and radius of approximately 21 cm and 1 cm respectively). The volatizing and reaction zone were separated by wire rnesh, and the entire reactor was encased in an insulated heating unit. Temperature control was accomplished by means of a thennocouple located in the catalyst bed, attached to a P D (proportionai-integrai-derivative) controller, that regulated power supply to the heater.

2 - Reaction Zone

Fig. 4.1 Plug Flow Reactor

The liquid reagents were fed to the reactor using a 350 ml syringe pump. The nitrogen was fed to the reactor fiom a pressurized gas cylinder with the flow regulated by a

rotameter. The nitrogen and liquid feeds were mixed pnor to addition to the volatizng zone. Since many of the compounds uivolved in the reaction are liquid under the

operating conditions used, it was decided that a downflow configuration would be more suitable for this application. Once through the reactor, the products were collected in a pot. The pressure of the system was maintained using a back-pressure reguiator.

4.2

Materials

The materials used for this study are: acetone, phenol, bisphenol A, nitrogen, N&on@, activated alumina, hydrochloric acid, ~ m b e r l y s t15, deuterated chlorofom, and nitric ~ acid. Supplier information is presented in Table 4.1.

Table 4.1: Materials used in experiments

-

Material Acetone Phenol Bisphenol A Nitrogen ~ a f i o n TYR-50 @ Nafion@SAC- 13 Activated dumina 8x14 mesh Activated alumina 14x18 mesh Hydrochloric acid Amberlyst@15 Deuterated chloroform Nitric acid

Supplier Fisher Scientific Fisher Scientific

CHIMICA BOC GASES DuPont

Catalog # CAS 67-64- 1 CAS 108-95-2

(H20: 7732-18-5) CAS

Purity histological grade 9 1.3% (9%H,O)

DuPont ALCAN CHEMICALS ALCAN CHEMICALS BDH Inc. ROHM and HAAS Cambridge Isotope Laboratones BDH Inc.

CAS 80-05-7 UN 1066 LOT NO. KGO348-349 . LOT NO. 924 SAMPLE NO. 97018 SAMPLE NO. 97018
ACS 393-02 393 89-20-3 (H20: 7732-1 8-5)

97% N/A N/A

NIA NIA

NIA
36.5-38% NIA 99.8% 69-71%

CAS 865-49-6

ACS 579-02

Al1 the materids are hazardous, thcrefore safety regulations must be cl,aictly followed

when operating and analyzing the experiments.

More detailed health and safety

regulations about the materials mentioned above and about the products obtaied in the process are presented in Appendix A.

4.2.1 Solid Catalysts

The solid catalysts used in the experimental investigation were Nafion@NR-50, Naf50na SAC-13, activated aiumina 8x14 mesh and 14x18 mesh, and Amberlyh 15. n i e
are structure and the properties of Nafion@ presented below.

4.2.1.1 ~ a f i o n @ - Perfiuorosulfonated Ionorner

~ a f i o n @ a poly(tetrafluoroethy1ene) based ionomer (Mauritz is

- Nafion@Research), a

registered trademark of E.I. DuPont de Nemours and Co. It is commercially available in

the form of millimeter sized beads known as ~ a f i o n NR-50 resin. Since its development @

in 1960s, it has many applications in liquid and gas separations, fuel cells, and the chlor-

aikali industries.

Because of its thermal and chemical resistance, ion-exchange

properties, selectivity, mechanical strength, and insolubility in water, it has widespread application.

Nafion@ is a copolymer of tetrduoroethene and perfluoro-2-(fluorosulfonylethoxy) propyl Wiyl ether that belongs to the class of solid superacid catalysts. It exhibits acid 1 strength greater than that of 100% H,SO, (scienze.ch.unito.itlch/DipIFM/fisica/eccc /

nafZon.htd). It has hydrophobic (--CF,--CF,-)

and hydrophilic ( 4 0 , H ) regions i n

its polymeric structure, and its superacidity is attributed to the electron withdrawing effect of the perfluorocarbon chah acting on the sdfonic acid group. Figures 4.2. and 4.3 show the chernical structure, and the electron wididrawing effect respectively. The use of

~ a f i o n @ catalyze various reactions, such as alkylations, acy lations, isomerizations, to

hydrations and dehydrations, nitrations, etherifications, disproportionations, and esterifications has been descnbed in detail (Harmer et al., 1996). The study presented in this thesis will show that ~ a f i o n @ able to catalyze condensation as well. is

-(CFI-CF&+CFCFJn(0

I
I

5-1 5% Sulfonic Acid Groups

Fig. 4.2 Nafion@Structure; m = 6 or 7, n n 1000, x = 1,2, or 3 (Harmer et al., 1996).

Structurally, Nafion@is cornplex. Although the exact structure is not known, several models have been proposed since the early 1970s, to describe th- way in which ionic groups aggregate within the Nafion@polyrner. Robertson (1994) has summarized many of these such models: The Mau&-Hopfinger Model, The Yeager Three Phase Model,

The Eisenberg Model of Hydrocarbon Ionomers and The Gierke Cluster Network Model.

flectron withdrawing atom

Fig. 4.3 Electron Withdrawing Effect

(scienze.ch.unito.it/~h/DipIFM/fisica/ecccI/nafion.html)
A common objective of these models is to predict the hdamental feature of unique

equilibrium ionic selectivities, as well as the ionic transport properties of p e r f l u o ~ a t e d ionomer membranes.

As a result of electrostatic interactions, these ionic groups tend to aggregate to form tightly packed regions referred to as clusters (Butler et al., 1994). The presence of these electrostatic interactions between the ions and the ion pairs enhances the intermolecular forces and thereby exert a significant effect on the properties of the parent polymer.

Small angle X-ray scattering (SAXS) (Yeager and Eisenberg, 1982) and neutron
scattering (SANS) experirnents clearly indicate that ionic clustering is present in ~ a f i o n ? However, details on the arrangement of matter within these clusters have not been flly realized. Although no mode1 has been found to provide a complete explanation of the properties and selectivities found, several models base these properties and selectivities on an extensive micro-phase separated morphology (Mauritz - Nafion@ Research). Figure

4.4 shows the sylized view of polar/nonpoIar microphase separation in a hydrated

ionorner.

This over-simplification shows a phase separated morphology of discrete hydrophobic and hydrophilic regions.

The hydrophobic region is composed of the polyrner

fluorocarbon backbone. On the other hand, the hydrophilic region contains the ionic groups and their counter ions.

Fig. 4.4 Stylized View of Polar/Nonpolar Microphase Separation in a Hydrated Ionorner (Mauritz - ~ a f i o n @ Research)

The Yeager Three Phase Mode1 is a phenomenological based model, represented in Figure 4.5. This model is based on a three-phase clustered system with intercomecting channels within the polymer. The three regions consist of (A) a fluorocarbon backbone, some of which is microcrystalline, (B) an interfacial region of relatively large fiactional void volume containing some pendant side chahs, some water, and those sulfate ~ O U P S
and counter ions which are not in clusters, and (C) the clustered regions where the

majority of the ionic exchange sites, counter ions, and sorbed water exists (Yeager and Eisenberg, 1982).

From experimental means, such as small-angle-X-ray scattering (SAXS), it has been detemined that the phase-separated morphology is on the order of 30-50A Bragg spacing (Brookrnan and Nicholson, 1986). Howeve- upon hydration, Nafion@,with its ability to sorb relatively large amounts of water, cm increase its dry weight by as much as 50% or more, depending on equivdent weight, counter ion, and temperature. Upon hydration, cluster diarneter and the nurnber of exchange sites are thought to increase (Brookrnan and Nicholson, 1986).

Fig. 4.5 The Yeager 3 Phase Mode1 of N&on@ Clusters (Mauritz - ~ a f i o n @ Research)

78

The Nafion@NR50 resin begins to lose the sulfonate groups at about 280C (Harmer et
ai., 1996). This temperature stability is much higher than hydrocarbon-based sulfonate

t~ ion-exchange resins, such as ~ m b e r l y s 15, which is stable only up to 130C.

NR-50 resin is very low, typically 0.02 m' g-' or less, and The surface area of the ~ a f i o n @
most of the active sites are buried within the polymer beads. Under many types of reaction conditions, these sites are inaccessible or poorly accessible, and as a result the observed activity for rnany reactions is very low (Harmer et al., 1996).

4.2.1.2 High Surface Area Piafion@Resin

In order to compensate for the shortcoming outlined in the end of the previous section, a
newly developed matenal (Harmer et al., 1996) was tested. It is a nanocomposite of nanometer shed ~ a f i o n @ particles entrapped withui a highly porous silica network. resin

It is commercially available and it is known as ~afion@' SAC-13 resin (for a content of 13

wt % of ~ a f i o n @ the Nfo @ in ai n '

residsilica material).

With the new material, the

accessibility of the acid groups is supposed to significantly improve. The inherent

thermal stability of the Nafion@SAC-13 resin is about the same as Nafion' M - 5 0 resin.

4.3 Procedures

4.3.1 Reactor Loading and Set-up

4.3.1.1 NMR Tube Reaction
1. Weigh a clean, empty NMR tube.

2. Weigh 0.04 g of catalyst in the NMR tube (Note: The desired quantity of catalyst was

measured directiy in the tube to avoid transfer losses.).

3. Prepare a mixture of acetone and phenol with a molar ratio of 1:2 acetone:phenol.

4. Take 40 pl of the mumire w i t h an Eppendorf pipette and put them into the tube.

5. Add 700 pl of deuterated chloroform to the tube. 6. Cover the NMR tube and seal carefully with parafilm.
7. Insert the tube in the appropriate spinner and adjust the depth using the sample depth gauge.

8. T m on the air and position the tube in the magnet.

9. Lower the tube into the magnet by tuming the air off.

10. Start the spinning air.

11. When the spinning rate reaches its set value, adjust lock power and lock gain.

12. Center and then lock the signal.

13. Start heating the magnet by tuming on the heater, setting the temperature at 343 K,
and increase the air flow through the magnet.
14. Once the temperature is reached, shim the field, and start the routine for data

acquisition.

4.3.1.2 Batch Reactor

1. Put on protective clothing, lab coat, goggles, and gloves.

Ail the steps were

perfomed in the fume hood. 2. Weigh the desired quantity of catalyst (10 g or 20 g) in a clean measuring dish.

3. Transfer the catalyst into the flask.

4. Weigh the desired quantities of phenol and acetone (75.55 g or 89 g of phenol and

24.12 g or 11 g of acetone). 5. Add the measured quantities of reagents to the reactor.

6. Turn on cooIing water to condenser.
7. Turn on magnetic stirrer.

8. Turn on the heat and adjust the thermostat for the desired temperature.

4.3.1.3 Flow Reactor

Put on protective clothing, [ab coat, goggles, and gloves. All the steps were perfomed
in a ventilated explosion proof bunker.

Place the desired catalyst in the reaction zone. Assemble cleaned reactor. Set temperature to desired value to dry catalyst. Open nitrogen tank and pressurize the reactor to 50 - 55 psi. Adjust back-pressure valve to maintain this pressure. Check for leaks.

8. Set flowmeter (NI 12-02) to maintain nitrogen pressure at 50 - 55 psi. 9. Let catalyst dry ovemight.
10. Set temperature to desired value for m.
11. Vent reactor to remove water that came off the catalyst.

12. Let the pressure stabilize. It shodd corne back to where it was set.
13. Weigh the desired quantities of phenol and acetone.
14. Load reactants into the pump.

15. Open valve to reactor.

16. Close bunker door.

17. Set pump control to get desired flow rate and turn the power on to the pump.

4.3.2 Reactor Sampling Procedure and Sample Preparation

4.3.2.1 NMR Tube Reaction

The routine was set to sample the reaction every 20 minutes. No samples were removed fiom the tube, because the tube itself, that is the reactor, was the sample. The Free
Induction Decay (FID) for each sample was saved in a file. At the end of the reaction the

l files with d the FID's were saved on disk and M e r analyzed with appropnate
software.

4.3.2.2 Batch Reactor

Turn the magnetic stirrer off.
Collect 40 pl of reaction mixture with the Eppendodpipette and put them in an NMR tube. Dilute the sample with 700 p l of deuterated chloroform. Coliect 200 pl of reaction mixture with the Eppendorf pipette an put them in a 2 ml sample vial. Dilute the sample with 1ml of acetone.

Tum the magnetic stirrer back on.

The reaction was sampled every 24 hours.

4.3.2.3 Flow Reactor

After 24 hours of reaction, the liquid in the receiver was collected in a clean flask for

sampling. Two samples were prepared the same way as described for the batch reactor,
one in the NMR tube and the second in a sampling vial.

4.3.3 Reactor Shut-Down and Clean-Up Procedure 4.3.3.1 NC R Tube Reaction lP

1. Remove the tube fiom the magnet.

2. Turn off the heater, and set the temperature back to the room temperature value.
3. When the magnet cooled down, reduce the air flow through the magnet.

4.3.3.2 Batch Reactor

Put on protective clothing, lab coat, goggles, and gloves.

Turn off the heater.

T m off the magnetic stirrer.

Tum off the cooling water to the condenser.

When reactor is at roorn temperature, remove the reactor fkom the c-clamp. Separate the solid catalyst fiom the reaction mixture on a Bchner funnel.

Wash the catalyst with acetone on the filter.

Allow catalyst to dry on the filter, then dispose in a special container. Dispose of the separated liquid, reaction mixture and wash acetone into the solvent cm.
10. Wash the Bask, the stopper, and the themorneter with acetone, then with water and

soap, then with distilled water, and allow them to dry. 11. Rime thoroughly the sampling syringe with acetone to remove any residual BPA or phenol.

4.3.3.3 Flow Reactor

1. Put on protective clothing, lab coat, goggles, and gloves.

2. Turnpump off.

3. Vent the receiver into a clean flask for sampling.

4. Set temperature to about 50C and let the reactor cool down so that it can be handled.

When the reactor cooled down, tum off the nitrogen tank and vent the reactor by opening the back-pressure valve. Open the bunker door. Disassemble the reactor. Clean the reaction zone.

Clean the ceramic saddles with sulfi.uk acid, then dry them in a clean oven.
10. Drain the pump of remaining reactants and rime with acetone to remove any left over

phenol which could solidfi in the pump when it is not operational. Place solvents in solvent can for disposal.

4.4 Sample Analysis

:deally, the sampling and the analysis of the reactor content should have been done on line, on a High Pressure Liquid Chromatograph (HPLC) equipped with a split injector. Unfominately the necessary equipment to do so was not available, and alternate options

had to be considered, that is a Gas Chromatography/Mass Spectrometer (GCMS), and an NMR Spectrometer.
The calibration of the GC/MS was not possible due to the fact that most of the byproducts produced in the reaction were not available as standards. Also a split injector

was not available. Because the samples were concentrated, and an adequate solvent for
the reaction mixture was not determineci, acetone was used to dilute the samples for the

GC/MS. As a result the analysis on the GCMS was used to monitor the selectivity of the
process, and the NMR analysis was used to determine the quantity of BPA produced.

4.4.1 Gas Chrornatography/Mass Spectrometry Analysis

A Varian Satum II Gas Chrornatography/Mass Spectrometer with an ion trap detector

was used for product identification and to determine the seiectivity of the process of BPA

formation under heterogeneous catalysis. A fused silica capillary column SPBTM-20, 30

m long, 0.32 mm diameter, and 0.25 pn film thickness, supplied by Supelco Inc." was
employed. The apparatus employs a high performance Varian Mode1 3400 Gas

Chromatograph (GC) with a 1093 Septum equipped Programmable Injector (SPI) and a
1077 Splitless Capillary Injector.

The GC is also equipped with thermostated

pneumatics, for improved retention time stability. The Satum uses an ultratrace ion trap

mass spectrometer.

Cornplete separation of the components is necessary in order to get an accurate rneasure

of the quantities of each within the sample. Complete separation was not possible for the
chromanes, and they were considered together. The temperature profile of the GC

method is displayed in Figure 4.6. Table 4.2 outlines the acquisition method of the MS,

and the boiling points and retention times of the products fiom this reaction are presented
in Table 4.3. A representative chromatogram of the products from the condensation of
acetone with phenol is presented in Figure 4.7.

Fig. 4.6 Temperature Profile of Method Used o n GC

Time [min]

Table 4.2: Acquisitior method on the Mass Spectrometer 1 29 to 350 Mass range ( m u ) Seconddscan (4Scans) 1 .O00 Acquire time (minutes) 12 ~il/Ml delav (seconds) 1110 Il Peak threshoid (counts) Mass defect (mmu/lOOamu) 1 85 Backrrround mass (arnu) 1 28 Ioniz. mode EI ON Auto ion control Cal gas ON

Table 4.3: Peak tabIe with retention times and boiling points of the products Peak Retention Time (min) Molar Weight (ghofe) Boiling Point (C)lmm Hg Mesityl oxide 152 98.14 1 301760 Mesitylene 216 120.19 164.71760 Cla 174 323

470 O-pisomer 546 Chromanes (601 ;607) BPA 632 " 4-@-hydroxypheny1)-2-rnethyl-1,3-pentadiene 2-@-hydroxyphenyl)-4-methylpent-3-en-2-oi

0-0 isomer

228.29 228.29 268.34 228.29

17010.25 165-17010.25 17010.3

Fig. 4.7 Chrornatogram of the Products Obtained fiom the Condensation Process
C a m m e n t : EXERIHENT 4 W I O H FE39 WTER 96 HOURS Scan No: 1 RetentionTima' 8!01 RIC: 8 Range: 1 ta 728 Plotted: 1 to 728

Chramaabgram P l o t

-C:~ATLIRWUILICIP(AWXP1\FEBSs4ts W W 9 8 a s
Range: 8

16:B1!38
8

ieq

IB&A = 351694116

TOT

The calibration of the G C M S method was not possible because most of the by-products
observed in the reaction were not availabie as standards. NI the sarnples contained

relatively the same components in comparable amounts, the sarnpling procedure. The preparation of the sample and the analysis conditions were identical. Therefore it was decided to consider the area of the peaks proportional with the mass of the corresponding compound. The GC/MS analysis was used to detelmine the selectivity of the process.

4.4.2 NMR AnaIysis

A Bruker ACF-200 NMR Spectrometer with a

MHz

frequency was used to

deterrnine the amount of BPA produced with heterogeneous catalysis. acquisition parameters are summarized in Table 4.4. commercial NMR data processmg 1995).

The data
a

N M R data are anaIyzed using

as NUTS (Acom NMR,

Table 4.4: Data acquisition parameters

Parameter 1 Value 16 Number of Acquisitions 5.O Pulse Width (usec) 3 .O Recycle DeIay (sec) 200.132339 Frequency (MHz) 4032.3 Sweep Width (Hz) Dwell Titne (usec) 248.O 2.032 Acquisition Time (sec) 1 104.9 Offset Freauencv 1 8 192 Number of Points Time Domain TPPI Acquisition Type

4.4.2.1 General Introduction to the NMR Procedure Used in

this Study
The procedure for calculating the yield in BPA in this study is based on the fact that

acetone and BPA have peaks that do not overlap and can be integrated and compared.

The procedure for calculating the error associated with the NMR analysis is also

described. The NMR spectra of acetone, phenol, and bisphenol A are presented in Fi-ures 4.8,4.9, and 4.10, respectively.

Fig. 4.8 NMR Spectmn for Acetone (CDCI,)

Fig. 4.9 NMR Spectnim for Phenol

Fig. 4.10 NMR Specaum for Bisphenol A (CDCI,)

4.4.2.2 Calculation of the Error Associated with the NMR

Analysis
In the NMR spectrum of the initial mixture acetone:phenol (Fig. 4.1 l), the integrai of the
acetone peak was attributed the value of 100. The portion of the spectnim where the two methyl groups of the BPA should appear was also integrated. Since there is no BPA initially present in the system, the value of this integral should be zero, and if it is nof
then the error is attributed to the noise level within the integrated interval of the spectrum.

Fig. 4.11 NMR Spectrum of the Initial Mixture of Reaction (fiom 0.4 ppm to 3.0 ppm)

Since the processing error seems to be under 0.01 %, it was necessary to attribute a value
of 10,000 to the integral of the acetone peak in order to calculate the processing error.

Then the integrated noise showed a non zero value (Fig. 4.12).

The signal to noise (sh) is calculated as a ratio of two ratios. The numerator is the area
of a peak divided by the frequency domain over which the peak was integrated. The denominator is the area integrated over a region in the spectrum where no peak is supposed to appear divided by the frequency domain over which it was integrated. The eequency domain can be expressed both in Hertz or in ppm; ppm represents the ratio of
the resonant frequency (in Hz) and the fiequency of the magnet (in Hz).

In the

caiculation presented for signal to noise the fiequency domains were expressed, by choice, in Hertz. The ratio signal to noise (sh) is:

The error associated with NMR analysis is:

Fig. 4.12 NMR Specum of the Initial Mixture of Reaction (fiom 1.O ppm to 3.0 ppm)

4.4.2.3 Procedure for Calculating the Yield in BPA

Another sample, taken at the end of the reaction, was analyzed on the NMR, and
processed with NUTS (Acom NMR, 1993,1994,1995),using the same normalization

constant, as the sample shown in Figure 4.1 1. The specend of the reaction is presented in Figure 4.13.

of the sample taken at the

When processing spectra with the same normalization constant, by setting the value of an integral at 100, in al1 subsequent spectra the values of the integrals considered are going to be percentages of the set integral. It is known that both integrals considered in this application (the acetone peak and the peak of the methyl groups in the BPA) are accounted for six protons each, and it is also known that the integrals are proportional to the number of moles in the mixture corrected with the a m b e r of protons (a correction not necessary since the number of protons is the same). This means that if the initial composition of the mktwe is known, the number of moles of BPA and of acetone in a subsequent spectnim c m be calculated.

Fig. 4.13 NMR Spectrum of the Final Mixture of Reaction (fiom 0.4 pprn to 3.0 pprn)

The initial mixture contains: 24.12 g of acetone and 75.55 g of phenol. The molecular
weights for acetone and phenol are: 58.08 g/mole and 94.1 1 g/mole respectively. Therefore the initial rnixhue consists of 0.41 mole of acetone and 0.8 mole of phenol, or, in mole percent, 34.1 mole % acetone and 65.9 mole % phenol.

The yield in BPA is caiculated with the following formula:

where

7 is the yieId in BPA;

q p ~ , f i n the number of moles of BPA present in the finai reaction mixture, and it is is

caiculated as the product of the initiai number of moles of acetone and the value of

the BPA methyl peak integrai in the NMR spectnun;

~ is A molecuiar weight of the BPA; the

m i , is the weight of hc reaction mixture.

4.5 Summary
This chapter presented the matenals and the analytical methods used for this research, and the apparatus employed by the experimentd part of this study. The NMR tube reaction was used to identifi new catalysts. The batch reactor was used to perform the experiments with ~ a f i o n @ , order to assess the effects of the selected process parameters in on the synthesis of BPA. The plug flow reactor was employed for the reactions with acidified activated alumina, since this catalyst was not mechanically robust enough to undertake the mixing in the batch reactor. A significant amount of time was required to ensure that al1 safety concerns were satisfied.

in the next chapter the experimental results are presented and discussed. Several sets of

experiments were performed to evaluate the reactivity of the system of interes&to veri& the experimental reproducibility, and to narrow down the experimental region which will be investigated usng an experimental design.

Chapter 5

Experimental Results and Discussion

Eight sets of experiments were performed to examine the synthesis of bisphenol A under various reaction conditions. The first set used homogeneous and heterogeneous catalysis
at roorn temperature. The second set used ~afTon@ various temperatures in a batch at

reactor. The third set used AmberlystB 1 5 as heterogeneous catalyst, with the purpose of evaluating the experimental reproducibility. The fourth set consisted of one reaction with heterogeneous catalyst, and had the purpose of validating the simulation prediction that
the reaction goes to depletion of acetone. The nfth set consisted of reactions performed
with heterogeneous catalyst in an NMR tube. The sixth set used heterogeneous catalysis

in a batch reactor with the purpose of cornparhg the performance of ~ a f o n @ NR-50

versus ~mberlyst@ Finally, the seventh and the eighth sets employed heterogeneous 15. catalysis at various ternperatures, catalyst concentrations, and molar ratios acetone:phenol in a batch reactor. Ln these experiments a two factorial design was perfomed to examine
the effects of catalyst type, catalyst concentration, temperature, and molar ratio acetone to

phenol. AU the experiments presented in this chapter are summarized in Table 5.1.

Table 5.1 Summary of the experiments 1 Catalyst Conc. Temp. Acetone:Phenol Reactor Type Catalyst Exp. # 1 Time 1 Molar Ratio (Oc) (wt %) (h) Type Batch 1:2 25 HC1 10 288 1.1 1:2 Batch 25 ~mberlyst"15 10 2 16 1.2 1:2 Batch 25 72 No Catalyst 1.3 Batch 1:2 63 10 96 ~afion@NR-50 II. 1 Batch 1:2 72 10 Nafionam-50 11.2 96 Batch 1:2 83 10 96 ~ a i o nNR-50 @ 11.3 Batch 92 1:2 ~ a f i o n NR-50 @ 10 96 11.4 Batch 102 1:2 10 Nafion@ NR-50 96 11.5 Batch 72 12 10 5 ~rnberlyst~ 15 IL?.1 72 1:2 Batch 10 5 ~mberIyst@ 15 111.2 Batch 72 1:2 10 5 ~ m b e r l y s 15 t m.3 Batch 72 1:2 10 240 Amberiyst@15 IV. 1 NMR Tube 1:2 70 6 ~ m b e r l y s 15 t~ 10 V. I NMR Tube 1:2 70 10 3 Nafion@ NR-50 V.2 NMR Tube 1:2 70 10 3 ~ a f i o n NR-50 @ V.3 V.4 1 6 1 AA 300/HC1 1 10 1 70 1 1 :2 1 NMR Tube NMR Tube 1:2 70 AA 300/HCl 6 10 V.5 Batch 1:5 92 20 27 ~ m b e r l y s 15 t~ VI. 1 Batch 92 15 VI.2 20 27 ~ a f i o n NR-50 @ Batch 102 1:2 10 24 VII. 1 Nafion@ NR-50 102 Batch 1:5 10 24 VIL2 Nafion@ NR-50 82 Batch 1:2 10 24 NafionaNR-50 VII.3 Batch 82 1:2 24 Nafion@ NR-50 30 VI1-4 Batch 82 1:5 24 10 ~~on@NR-50 VIL5 102 1 :2 Batch 20 24 VII.6 Nafiona NR-50 Batch 15 102 20 24 VII.7 Nafionam-50 Batch 1:5 82 20 24 ~ @ VII.8 - a f i o n NR-50 Batch 1:2 92 10 24 Nafion@NR-50 VII.9 92 Batch 1:s 10 24 VXI.10 NafionBNR-50 Batch 1:2 92 10 VII.1 1 24 1 NafionQNR-50 Batch 1:2 92 24 10 Vn.12 Nafion@ NR-50 Batch 15 102 24 10 VIII. 1 NafionQNR-50 Batch 15 102 24 20 ~afion@NR-50 VIII.2 102 Batch 1:2 20 %iT24TizzmmT 1:2 Batch 10 VIII.4 24 82 Nafion@ NR-50 Batch 1:5 82 24 VIII.5 10 ~afion@NR-50 102 1:2 Batch 24 10 Nafion@NR-50 VIII.6 82 Batch 15 24 20 Nafion@ NR-50 VIII.7 Batch 1:2 82 20 ~afion@NR-50 VIIT.8 24

5.1 Preliminary Investigation

The purpose of the preliminary investigation was to evaluate system reactivity, blank reactions, experimental region, scheme of reaction, and experimental reproducibility. Homogeneous and heterogeneous catalysts were used at this stage, at various temperatures.

5.1.1 Evaluation of System Reactivity and Blank Reactions

For evaiuating the system reactivity, three experiments were conducted (experiments 1.1, 1.2 and 1.3 presented in Table 5.1). Based on the results of the simulation, discussed in Chapter 3, a 1:2 mixture acetone:phenol was used in al1 three experiments. Two of the experiments were conducted under catalytic conditions: for the first experiment concentrated hydrochloric acid (HCL) was used, and for the second experiment ~ r n b e r l y s t15 was used. In the diird experiment no catalyst was used. ~

Ali three

experiments were conducted at room temperature. The experiments with hornogeneous

catalyst and the reaction with no catalyst were conducted without stirring. The reaction

with heterogeneous catalyst was under continuous stirring (the magnetic stirrer was set on
position 5).

The homogeneous catalysis was sampled and analyzed four times, once a day, every three days. The heterogeneous catalysis was sampled and analyzed three times, once a day, every three days. consecutively.

The non-catalytic process was analyzed once a day, three days

It was reported (McKetta and Cunningham, 1976) that a 1:2 mixture of acetone and phenol treated with concentrated hydrochloric acid which is allowed to stand for some hours at room temperature deposits a mass of crude bisphenol A crystals. Another statement found in the literature is that the activity of the catalysts for the formation of bisphenol A, at temperatures appropriate for the said process, decreases as follows

(Singh, 1992):
HC1> ~mberlyst' 15 > Acidic zeolites with large pore openings (> 7.0 A)

The intention was to validate these two statements, and to veriQ how fastklow the acetone and the phenol react at room temperature when no catalyst is added. This last verification is needed because some of the samples taken fiom the reaction mixture were to be analyzed several hours afier sampling and it is important to make sure that there is no reaction between sampling and analysis.

The reaction mixture treated with hydrochloric acid started to darken (brown) the next day and it continued to do so until the last day of analysis (the 12' day). No crystals were observed on the bottom of the flask. The reaction mixtures from the experiment with heterogeneous catalyst and fkom the experiment with no catalyst showed no change i n color and no crystals were observed on the bottom of the flask either.

Figures 5.1, 5.2, 5.3, and 5.4 represent the NMR spectra of the samples taken h m the reaction mixture with homogeneous catalyst (HCI). The high peak at approximately 2.2 ppm is the acetone (Pouchert and Behnke, 1993). The split peaks in the region 6.5 ppm -

7.5 ppm belong to the aromatic protons fiom phenol, bisphenol A,

and

hydroxyisopropylidene phenol (the product of the fust step of condensation) (Pouchert and Behnke, 1993). The solvent peak appears in the overlapped region, at 7.24 ppm

(Pouchert and Behnke, 1993). The peak at approximately 1.59 ppm is given by the
methyl protons in bisphenol (Pouchert and Behnke, 1993). The peak at approximately

1.27 ppm is believed to be given by the methyl protons fiom the hydroxyisopropylidene phenol.

Fig. 5.1 Analysis of the Reaction with Homogeneous Catalyst (after h e e days)

i " "

i " '

'

Fig. 5.2 Analysis of rhe Reaction with Homogeneous Catalyst (after six days)

Fig. 5 3 Analysis of the Reaction with Homogeneous Catalyst (after nine days) .

Fig. 5.4 Analysis of the Reaction with Homogeneous Catalyst (after twelve days)

These specea, although not quantitative, show that bisphenol A is formed when reacting acetone with phenol in the presence of hydrochloric acid at room temperature. The reaction mixture used in this experiment was left in the fume hood. Three months later crystals of BPA were observed on the bottom of the flask. They were needle shaped, about 1 cm in length, growing in d l directions. These crystals continued to grow, meaning that the process of bisphenol A formation continued. Figure 5.5 illustrates the crystals after several rnonths.

Fig. 5.5 Crystais of BPA

Figure 5.6 represents the NMR spectrum of the sarnple taken from the reaction mixture
with heterogeneous catdyst (AmberlystB 15), after nine days. The description of the

peaks is the same as the one given for Figures 5.1 to 5.4.

Fig. 5.6 Analysis of the Reaction with Heterogeneous Catalyst (afier nine days)

"

"

'

"

"m

'

'I r " "

mm

The spectra of the samples taken fiom the reaction with heterogeneous catdyst show that bisphenol A is not obtained when reacting acetone with phenol in the presence of ~ m b e r l y s t ~ under continuous stin-ing, at room temperature. This demonstrates that in 15, the same range of temperatures, the hydrochloric acid is more active with respect to the formation of bisphenol A than the ~mberlyst@ 15.

Figure 5.7 represents the NMR spectrum of the sample taken fiom the reaction mixture
with no cataiyst, after three days. The description of the peaks is the same as the one

given for Figures 5.1 to 5-4.

Fig. 5.7 Analysis of the Reaction with No Catalyst (afler three days)

The spectra of the samples taken fiom the reaction with no catalyst show that bisphenol A is not obtained when reacting acetone with phenol at room temperature. The results of the experiment with heterogeneous catalyst and of the experiment with no cataiyst leads to the conclusion that there is no risk of the reaction advancing during the period between the sarnpling and the NMR analysis, since there was no detectable conversion at ny time.

5.1.2 Evaluation of Experimental Region

In order to evaluate the temperature range and the duration that should be considered for

the reaction, five reactions were performed in a batch reactor, for a duration of 96 hours,
using Nafion@catalyst, and a molar ratio acetone:phenol of 1:2 (experiments II. 1, I . , I2 11.3, Ii.4 and 1 . presented in Table 5.1). The reactions were sampled every 24 hours ind 15

the sarnples were analyzed on the GC

- MS.

One sample was prepared and analyzed on

the NMR at the end of each reaction as well.

The results are presented in Table 5.2. Figures 5.8 to 5.14 present the variation of BPA, O-p isomer, and chromanes selectivity in tirne, variation of BPA, O-p isomer, and chromanes selectivity with the temperature, and variation of yield in BPA with the temperature.

Table 5.2: Results of the second set of experiments Exp.

#

Temp. ("Cl
63

Catalyst Reaction Time Conc.

(wtOh)
@)

Yield

Product distribution (wt%) II II[I+NOthers

II. 1

24 48
72

96
11.2
72

24

48
72 96

11.3

83

24 48 72 96

11.4

92

11.5

102

1 is bisphenol

idenediphenol;

; trimethyl chroman 1 IV is 4'-hydroxyphenyl-2,4,4-trimethyi chroma. II; Others are cornpounds found only in trace quantities. OnIy mesityl oxide was detected. The mesityl oxide was not detected.

Fig. 5.8 Variation o f BPA Selectivity in Time

1-

Time [hl

Fig. 5.9 Variation of Selectivity of O-p lsomer in Time

Time
102C

F]

Fig. 5.10 Variation of Chromanes Selectivity in Time

Time Eh]

Fig. 5.1 1 Variation of BPA Selectivity with Temperature

Temperature [Cl

Fig. 5.12 Variation of O-p Isorner Selectivity with Temperature

1-

Temperature [Cl

Fig. 5.13 Variation of Chromanes Selectivity with Temperature

Temperature [ l C

Fig. 5.14 Variation of BPA Yield with Temperature

Temperature [ l C

Figure 5.8, variation of BPA selectivity in time, shows that the data for the selectivity of

BPA formation converge, which indicates that different temperatures result in a

difference in the rate of BPA formation, but not in a difference in mechanism. The negative influence of the temperature on the selectivity of the BPA formation can also be observed in Figure 5.1 1. At 83OC the rate of BPA formation is lower than at higher temperahires, 92OC and 102"C, and as the reaction advances, the weight percent of BPA in the m a s of reaction decreases until it reaches a plateau, which is considered to be the equilibrium composition. The explanation for this is that although the number of moles

of by-products formed is smaller, most of them have a rnolecular weight greater than

BPA. At higher temperatures the weight percent of BPA in the rnass of reaction stays
fairly constant (a small decrease of the weight percent is observed at 92OC and a small
increase at 1 0 2 O C ) because in this case the rate of formation of BPA is higher and the

higher amount of BPA formed compensates for the difference in molecular weight between BPA and the heavier products.

The graph presented in Figure 5.9 shows that the selectivity of the O-p isomer formation increases in tirne for the reaction at 83 OC, is constant for the reaction at 92 OC, and dramatically decreases in Ume for the reaction performed at 102 OC. Figure 5.12 shows that, depending on the tirne, the temperature has a different effect on the selectivity of O-p isomer formation. That is: for a 24 h o u reaction, the selectivity remains constant, also for a 48 hour reaction the selectivity of O-p isomer formation is slightly icreasing, and for a 72 hour reaction the selectivity of O-p isomer formation is slightly decreasing, but the effect of temperature is sri11 negligible. For a 96 hour reaction the effect of the temperature is significant, the selectivity of the O-p isomer decreases with increasing temperature.

Figure 5.10 shows that the formation of the chromanes increases in tirne, also the temperature has a positive effect on the formation of chromanes. This can be dso seen in Figure 5-13. As expected, the yield in BPA increases with the temperature and it tends to level off. This can be seen in Figure 5.14.

It is not appropriate to compare the data obtained in this set of experirnents with the data

in Table 2.1. Recall that the data in Table 2.1 are equilibnum data. There is no evidence

that equilibrium has been achieved, for the data shown in Table 5.2. We know that there is still some acetone present i the mixture. n

Based on the above mentioned conclusions, it was decided to run the subsequent reactions for only 24 hours. It was also decided that the temperature region to be investigated will be between 82 O and 102 OC. C

5.1.3 Scheme of Reaction

Based on the findings in the previous section, a scheme of reaction was proposed. The most favored reaction is by far the formation of BPA. Therefore the process starts with the condensation of one molecuie of acetone with a molecule of phenol. Then the hydroxyisopropylidene phenol (ortho and para isomers) formed reacts very fast with a second molecule of phenol, leading to the formation of BPA or its O-p isomer. The
0-0

isomer also forms, but at a much lower rate, and the condensation of acetone with itself or the dehydration of acetone and some other by-products formation is actually favored.

Phenol

Acetone

p-2 -hydroxg-

=OH
Phenol

+ O=C,

/ CH3

CH3 Acetone

isopropylidenephenol

CH, I =?-OH CH ,

O-2-hikoxgisopropylidenephenol

p - 2-hydroxyPhenol i opropylidenephenol s

Phenol

isopropylidene phenol
HF,

C=O H~/ C
Acetone

+ O=C,

/ CH3

CH3
Acetone

OH 2,4 r-isopropyiidenediphenol

YC, H,C

'

C-HC-C-CH,+H,O

O M e e l Oxide

II

Acetone

Mesitylene
+

OH

H? CH,

CH 3

, a OH- H3C, C ' + H,O H,C/ 0

b~
2,2*-isopropylidenediphenol

O-2-hydroxyPhenoi isopropylidene phenol

The mesityl oxide can react with a rnolecule of phenol and form a product of
condensation named 2-@-hydroxyphenyl)-4-methylpent-3-en-2-01. This product can get dehydrated in the presence of acids and form 4-@-hydroxypheny1)-2-methyl-1,3pentadiene.

MesqlOxide

Phenol

2-methyi, 4-p-phenyl, 4-01 2-p entanone

4-@-hy&oxpphenyl)-2-methyl-

1,3-p entadiene

The mesityl oxide can react with two molecules of phenol and fomi chromanes.
H3C,
2 C=HC-C-CH, II H,C O

'

OH -ZH;O @,C ,!H

Phenol

O4CH,

Mesityl Oxide

H3C
4'hydroxy phenyl4'-hydroxyphenyl2,2,4-trimethyl chroman 1 2,4,4-tnmethyl chroman II

The triphenol formation indicated by the simulation was not confirmed, while chromanes were formed, as expected based on the data found in the literature. Some products which were not indicated in the literature, and therefore not considered in the simulation either, were obtained. These were the result of the reaction of acetone with itself (mesityl oxide and mesitylene), and the result of the reaction between phenol and mesityl oxide (2-(p-

hydroxypheny1)-4-methylpent-3-en-2-01), product which undergoes dehydration and

forms another by-product observed in the final reaction mixture (4-@-hydroxypheny1)-2methyl- l,3-pentadiene).

This mechanism was based on the observation of the appearance and disappearance of the
product peaks in the chromatograms of the mass of reaction. The
O-

and p-Zhydroxy-

isopropylidene phenol were not observed in the chromatograms at any t h e . These two cornpounds are the products of the first step of condensation in the process of formation of BPA, 0-0, and O-p isomers. They are definitely produced, and the reason for not being

seen in the chromatograms is that the second step of condensation is very fast, therefore
they are consumed as soon as they are formed.

In most of the experiments, after 24 hours the only products observed were mesityl oxide,
bisphenol A and the O-p isomer. Also, the amount of the bisphenol A (the p-p isomer) was obviously greater than the amount of the O-p isorner produced at al1 tirnes during the reaction.

The ~esi?ylene,the

0-0

isomer and the chromanes peaks appear in the chromatograms

almost at the same tirne, and their production stays at low quantities. The areas of the corresponding peaks indicate that the amount of chromanes in the reaction mixture increases faster than the amounts of mesitylene and 0-0isomer.

In some experirnents (11.4 and 11.5) the mesityl oxide was observed afier 24 hours and
after 48 h o u , but after 72 hours the corresponding peak disappeared and another product appeared in the chromatogram.

This last compound was identified

as 2-@-

hydroxypheny1)-4-methylpent-3-en-2-01,the product of condensation between mesivl

oxide and phenol. Almost at the sarne tirne another peak appeared in the chromatogram

of the reaction mixture, which was determined to be 4-@-hydroxypheny1)-2-methyl- 1,3 pentadiene, the product of dehydration of 2-@-hydroxyphenyl)-4-methylpent-3-en-2-o1. To summarize then, the experimental data strongly support the scheme of reaction proposed in this section.

5.1.4 Experimental Reproducibility

The experimentai reproducibility was evaluated both qualitative1y and quantitatively during the preliminary nuis. A quantitative rneasure of the experimentai reproducibility

was also estimated from the replicate runs in the experimentid design. The experiments
' were performed in the batch reactor for five hours, with ~ m b e r l y d15 as heterogeneous catalyst, at 7ZC. The results are presented in Table 5.3. The standard deviation was calculated as a rneasure of reproducibility and it was found to be 0.104 wi??, and the average of the yield in BPA was 0.37 wt%.

Table 5.3: Results of the experiments performed with ~ r n b e r l y 15 in the batch reactor t~
Exp.
#

Temp.

cc)
72 72

Catalyst TWe

Catalyst Conc.
(wt%)

Reaction Time (h)

5

Acetone:Phenol Molar Ratio

1:2 1:2 1:2

Yield in %PA
(wt%)

III. 1 111.2
111.3

72

10 ~ m b e r l y s 15 t~ ~mberlyst@ 15 10 ~ r n b e r l ~ s t @ 1 5 10

O -42
0.25
0.44

5 5

5.1.5 Validity of Simulation Prediction for Depletion of Acetone

One experiment was performed in the batch reactor, with ~ m b e r l y f l 15 as heterogeneous catalyst, at 72OC. nie intent of this experiment was to examine whether the reaction depletes acetone as predicted by the simulation's results. It was observed that, after 10 days, there was no acetone lefi. The results are presented in Figure 5.15. The disappearance of acetone indicates that the reaction is zero order in acetone, while the literature indicates that the reaction is first order in acetone. The difference is most

likely due to different reaction conditions, including catalyst type and initial reaction stoichiometry .

Fig. 5.15 Disappearance of Acetone

5.2 Investigation of Suitability of New Catalysts

Two new catalysts were tried, one was a solid super acid catalyst, ~ a f i o n @ , other one the
was activated alumina acidified with hydrochlonc acid (AA 3 00/HC1). These catalysts were selected since it was reported (Singh, 1992) that with higher acidity catalysts the
yield and the selectivity of the BPA formation are increasing.

One experiment was performed with each of the two catalysts. A third expeninent was
carried out using Amberlyst@' 15. The reproducibility of the processes was also verified.

The reaction with Amberlyst@15 served as basis of cornparison for the other two, since

~ m b e r l y f is the solid catalyst widely used in industry. Since the amount of Nafion@ 15
available was small and there was little chance of getting more, this set of reactions was

performed in N M R tubes to rninimize the quantities of catalysts and reagents used. The molar ratio acetone:phenol was 1:2, and the temperature was 70C.

The purpose of these nins was only to identie the formation of the desired product, BPA.
Therefore the values obtained for the yield and selectivity served only the purpose of cornparison of the processes catalyzed by the new catalysts with the process catalyzed by Amberlystm 15. The intent was not to obtain a quantitative result. The reaction mixtures were also analyzed on the GC 5.27, and 5.18.

- MS.

The chromatograms are shown in Figures 5.16,

Fig. 5.16 Chromatogram of the Products for the Process Catalyzed by ~ m b e r l y s 15(6 h) t~

Fig. 5.17 Chromatogram of the Products for the Process Catalyzed by Nafion@(3 h)

Fig. 5.18 Chromatogram of the Products for the Process Catalyzed by AA 300/HC1(6h)

CosaPerrt:

Scan Ho: 628 Retention Tine: Plcrttud: 1 to 988

The resdts showed that both the ~ a f i o n @ the AA 300/HCl can be used as catalysts and for the process of BPA formation. Both the yield in BPA and the selectivity of BPA formation are better for the processes that used new catalysts than for the process that used ~ r n b e r l y s t 15. ~ Comparing the two new catalysts, the yields in BPA are

comparable, but the selectivity of the process is clearly higher in the process catalyzed by
AA 300/HC1.

5.3 Performance Cornparison of Nafione and ~ m b e r l y s 15 t~

In order to compare the performance of Nafion" with ~ m b e r l y s t ~ two expenments 15,
were performed, one with Nafion@,one with ~mberlyst@ in the conditions described 15, by Singh (Singh, 1992) and compared with the results claimed by this author. A mumire of 1 5 acetone:phenol was reacted for 27 hours, at 92OC, under heterogeneous catalysis

(20 % catalyst). The results are presented in Table 5.4.

Table 5.4: Results of performance cornparison between Nafion@and Amberlystm15

Catalyst Reaction Yield tirne (h) i BPA n (wtY0) Product distribution (wt%)' 1 II m+N v
89.31 71.15 80.27 4.87 16.29 15.01

Product ratios II/I III+~VA 0.05 0.23 0.18

27 ~ m b e r l y s 15" t~ 27 AmberlystB 15 27 Nafion@ " Singh, 1992 t See notation in Table 5.2

3.54
12.56

18.08 25.21

0.00

2.28 0.00 4.72

0.04
0.18

0.00

The results published by Singh could not be reproduced. The two experiments conducted gives better results under the sarne conditions. for cornparison show that Nafion@

5.4 Experimental Design

An experimental design is a disciplined plan for collecting data (McLeIlaq 1998). It is

an integral component of quality improvement, and supports improvement in product design, process design, and process operation. Designed experiments are fiequentiy used in industry in order to detexmine what factors influence the process of interest. These experiments corne in many different forms since the purposes of the experiment may differ fiom situation to situation. In many cases, the expenment is simply used to

distuiguish the significant factors fiom the insignificant factors. Once these significant factors have been identified, other designs c m be used in order to determine the optimum setting for the process.

Designed experiments require that a set of controlled process variables, also called factors, be identified dong with the appropriate response variables. The factors are controllable variables thought to have influence on the responses. The responses are rneasurable outcornes of interest. It is important to realize that a factor is considered controllable by the experimenter if the values of the factor, known as levels, c m be determined prior to the beginning of the test program and can be maintained at the values required by the experimental design.

Continuhg with the terminology, the test run, the noise variables (covariates), the extraneous variation, and the effect needs to be defined. The test run is a set of factor level combinations for one experimental nui. The noise variables are variables affecthg

the process or product performance which cannot or are not controlled. The extraneous variation is the variation in rneasured response values in an experiment attributable to sources other than deliberate changes in the level of the factors. And finally, the effect of factors on response is measured by change in average response under two or more factor level combinations.

5.4.1 Factors Chosen and Responses

The purpose of the factorial design in this snidy is to investigate the effects of three
factors on the yield and selectivity of the production of bisphenol A, and on the selectivity of O-p isomer and chromanes formation. The three factors are: catalyst concentration (C), temperature (T), and acetone:phenol molar ratio (R).

The two levels for catalyst concentration were chosen tu be 10% and 20% by weight of the mass of reaction. Based on previous experimental expenence and simulation results, al1 correlated with data f?om literature, the levels for temperature were chosen to be 82C and 102OC, and for acetone:phenol rnoiar ratio, 0.2 and 0.5.

It is desired to conduct a minimum nurnber of runs at each level. The magnitude of the
effects to be detected was arbitrarily chosen as 2.5 the size of the inherent noise standard deviation, with a fdse detection probability of 0.05 and a failure to detect probability of 0.1. The minimum number of runs was calculated with the formula 5.1 (McLellan, 1998):

where z

'-7

,is the criticai value of the standard normal random variable with upper tail

probability d 2 , and z,-, is the critical value of the standard normal random variable with upper tail probability P. Typical values for a and Pare 0.95 and 0.9 respectively. Since the proposed design is a two-Ievel factorial design, the minimum number of required is 8. The detection of effects was considered A = 2 . 5 ~ .
nins

5.4.2 Evaluation of Results from Experimental Design

The values of each of the factors of interest were converted to their corresponding coded values. The hi& value, low value, midpouit, range and half range for each of the factors are summarized in Table 5.5 and the formulae used for coding were (McLellan, 1998):

where [Cl, [Tl, and CR] are the uncoded catalyst concentration, temperature and molar ratio acetone:phenol respectively.

Table 5.5: High value, low value, midpoint, range and half range for each factor

High Level Low Level Midpoint Range % Range

C (Y%$%) 20 T(OC) 1102 R(moVrn01) 1 0.5
4-1

+1 +1

10 1 -1 8 2 1 -1 0.2 ( -1

15 92 0.35

10 20

0.3

10 O. 15

The nuis that were performed in the experimental design are listed in Table 5.6 in the
randomized order they were performed. Replicate nuis were also executed in order to
v e a the reproducibility of the results.

The main effect of a factor is the average influence of a change in level of the single

factor on the response. This is calculated as the difference between the average of responses at high Ievel of the factor and the average of responses at low Ievel of the factor (McLellan, 1998).

Ma W f f e c t =

&PC,Or=,

-Y

,ktor=-,

(5-3)

Table 5.6: Experimental runs w d to investigate the effect of catalyst concentration (C),

temperature (T) and molar ratio of acetone and phenol (R). Nurnber W1.I VII.2
C -1 -1

T
1

R
I

1 -1

The interaction is the extent to whch the influence of one factor on response depends on

the level of another factor, or combination of other factors. The interaction effects,
considering n factors (x,. .

.a, caiculated as the difference between the average of the are

responses at high level and the average of responses at Iow level (McLeIlan, 1998). The difference in calculating interaction effects versus calcuiating main effects is that the high and low level in calcuiating interaction effects actually mean levels at which the product

x, *x,. ..x, equais +1 and - 1, respectively.

InteractionEfect = YX, *=.,
Il

=,

-y=,

**? Am =-,

5.4.3 Precision of Calculated Effects

The significance of the calculated effects is assessed by estimating the precision of the calculated effect with respect to the factor of interest. Confidence intervals were

constnicted. If the confidence interval contains zero, the effect is plausibly zero, and therefore it is not statistically significant. Otherwise it is significant, and it is taken into consideration. Standard t-tables and the standard deviation of the calculated effects are required for estimating the precision of each eEect (McLellan, 1998).
Precision = f ty 4/2 * seffeCt
(=)

where v is the number of degrees of fkeedom of the inherent noise variance estimate, a is the desired confidence level (a = O.OS), and seffecfis an estimate of the standard deviation of the calcdated effect. v = n - 1, where n is the number of runs in the replicate set.

The standard deviation of the calculated effect is calculated as the square root of the variance of the calculated efEect. The variance of the calculated effect is e s h a t e d fiom the number of runs in the designed experiment (2'3 and the inherent noise variance (sih ), which can be estirnated fiom replicate nins (McLellan, 1998).

5.4.4 Effects Analysis

Table 5.7 contains the data for ail four responses for the corresponding run. The

experiment number (Exp. #) corresponds to the same designation numbers outlined in Table 5.1 and Table 5.6 (randomized order). The main effects, two-factor interaction and three-factor interaction effects were calculated for each of these responses and are summarized in Table 5.8. Table 5.9. presents the precision of the calculated effects.

Table 5.7: Responses for the experirnents performed in the 2 experimental design '
Selectivity of O-pisomer formation; yz
(wt%)

Exp. X

vn.i
WI.2 VIL3 VIL4 VII. 5 VU.6 VIL7 VII.8 VII.9 VII.10 VII. 11 VIL 12

Selectivity of BPA formation; y, (WtYo) 70.07 68.47 84.55 78.68 77.23 64.90 63.O2 73-95 73-97 80.27 71.83 72.76

Selectivity of chromanes , formation; y

(wt%)

Yield in BPA; Y4 (wtO/o)

19-32 30.00 10.74 13-47 22.77 16.59 27.17 26.05 13-54 15.01 21.28 2 1-27

0.00 11.53 0.00 0.00 0.00 11.29 9.8 1 0.00 2.9 0.0 1 0.00 0.00

2-35 10.5 1 9.76 2.44 2.42 7.32 7.56 2.50 2.35 3 .69 1.95 2.32

Table 5.8 : Calculated effects

Table 5.9 : Precision of calculated effects

Response YI(m%) 14.52 Inherent Noise Variance 7.26 Variance 2.69 Standard Deviation 8-561 t 6.33b Precision a 95 % conndence; 90 % confidence
I

--

Y? (m%)
1

16.69 8.35 2.89 9.19' 1 6.80~

Y4 (m%) 0.58 0.29 1.O3 0.584 3.26' 1 f 2.41b t 1.72a 1 + 1.37~ I

Y3

(m%) 2.10 1.O5

'

5.4.4.1 Selectivity of BPA Formation

At the 95% confidence level, the temperature was the only factor that significantly af3ected the selectivity of BPA formation. The catalyst concentration, acetone:phenol molar ratio and al1 the associated interaction effects were insignificant, at the specified level of confidence. The significance of the effects was also analyzed for 90%

confidence. The temperature is still the only significant factor for the selectivity of BPA formation. The effects and their significance at the wo levels of confidence are presented in Figure 5.19.

Fig. 5.19 Effects o f Considercd Factors on Srlectivity o f BPA Formaticn and their Siynificanct:

-15

4
O 1

1 _ _ _ _ 90%

---95% confidence
confidence

Note: 1-C;

2-T;

3+R; 4+CT; 5+CR; 6+TR; 7+CTR

However, for the conditions chosen for these experiments, the temperature was the single factor to significantly affect the selectivity of the BPA formation. According to these results, increasing the temperature, the selectivity of the BPA formation will decrease.

5.4.4.2 Selectivity of O-pIsomer Formation

At 95% confidence Level, none of the factors had a significant effect on the formation of
the O-p isomer. The effect of the ratio of the reagents is very close to being significant,

and it actually becornes significant at the 90% confidence level. The effects and their significance at the two levels of confidence are presented in Figure 5.20.

Fig. 5.20 Efcrs of Considered Factors on Sclectivity of O-p isomer Formation and their Sinificance

Note: 1+C; 2+T; 3+R; 4+CT; 5+CR; 6+TR; 7+CTR

According to these results, as the acetone:phenol molar ratio is increased, the selectivity

of the O-pisomer formation will decrease. This could be explained by the fact that the
lower the concentration of phenol in the mass of reaction, the lower the probability that

the formation of 0.p-isomer will occur, since the formation of the p-p isomer is favored in
comparison with the formation of the O-p isomer.

OH
Fig. 5.21 Effects Present in the Molecule of Phenol and the Nucleophiiic Attack

As shown in the above figure (Fig. 5.21), there is a higher density of electrons in the para position, therefore the molecules of phenol will be more likely to react with the etectrophilic substrate in the para position, fonnuig the p-p isomer (BPA).

5.4.4.3 Selectivity of Chromanes Formation

At the 95% confidence level, the temperature was the only factor that significantly afEected the selectivity of chromanes formation. The hvo-factor interaction associated
widi the catalyst concentration and the molar ratio (CR) and the three-factor interaction

( C m ) were very close to being significant. The catalyst concentration, acetone:phenol

molar ratio and al1 the other associated interaction effects were insignificant, at the specified level of confidence.

According to these results, the selectivity of the chromanes formation will increase by increasing the temperature. This is explained by the fact that an increase in temperature will result in an increase of the energy of the molecules. Therefore, molecules which did

13 1

not have the energy to react, will now be accelerated and will be more likely to react.

The reaction of mesityl oxide with phenol, also the cyclization of the dimer formed by
dimerization of isopropenyl phenol, both leading to chromanes, have higher activation energies than the formation of BPA or its o-p isomer, and the higher the temperature, the higher the probability of reaction occurrence.

The signiscance of the effects was also analyzed for 90% confidence levei. In this case,
the effects which were very close to being significant became significant (CR and CTR).

Aiso the main effect of the molar ratio (R) and the two-factor interaction effect associated
with the temperature and the molar ratio (TR) became significant. The effects and their

significance at the two levels of confidence are presented in Figure 5.22.

Fig. 5.22 Effects o f Considcred Factors on Selcctivity o f Chromanes Formation and their Significance

95% conficience 90% confidence

Note: 1-C;

2+T; 3 4 2 ; 4+CT; 5+CR; 6+TR; 7+CTR

According to these results, the selectivity of the chromanes formation will increase by decreasing the molar ratio acetone:phenol. This is explained by the fact that the more

phenol available, the higher the possibility of less thermodynamically favored reactions to occur, such as the formation of chromanes.

5.4.4.4 Yield in BPA

At the 95% confidence level, temperature was the oniy main factor that significantly affected the yield in BPA.
As well, the two-factor interactions associated with the

catalyst concentration and the temperature (CT), and the temperature and the molar ratio

(TR), also the three-factor interaction (CTR) were significant. The catalyst concentration,
acetone:phenol molar ratio and al1 the other associated interaction effects were insignificant, at the specified level of confidence.

Accordhg to these results, the yield in BPA will increase by increasing the temperature. The catalyst concentration and the molar ratio of the reagents have significant effects only in conjunction with the temperature. The three-factor interaction effect is also significant.

The analyses performed for 90% confidence level added the significance of the catalyst concentration. The yield in BPA increases by decreasing the catalyst concentration. This does not mean that the reaction with no catalyst gives 100% yield, the observation is valid only for the studied range of catalyst concentration, 10 to 20% of the mass of reaction. This can be explained by the fact that the process has two steps. The higher the concentration of catalyst, the higher the concentration of active sights, and therefore more molecules can participate in the first step of condensation. In this marner more phenol is

consumed in the fint step of condensation, and Iess phenol will be available for the second step of condensation. Also the high concentration of active sites encourages other side reactions, like the condensation of acetone with itself and others. The effects and

their sipificance at the two levels of confidence are presented in Figure 5.23.

Fig. 5.23 Effects o f Considcred Factors on Yield in B P A and their Significance

95% confidence

90% confidence
Note: l+C; 3-T;

3-R;

4+CT; 5+CR; 6+TR; 7-iCTR

5.4.5 Regression Analysis

The regression analysis was performed based on the previous results of the experimental design. Models linear in parameters are fitted to the data for the selectivity of BPA, O-p isomer, and chromanes formation, and for the yield in BPA. The least squares estimates

for the parameters were detemiined for each model. The adequacy of the models was
assessed using F ratio tests.

The sum of squared residuals (SSR), the surn of squared errors (SSE), the error vector, and the total sum of squares (TSS) were calculated w"iththe formulas (McLellan, 1998):

TSS = SSR + SSE

(5.11)

The least square method h d s the parameter values that rninimize the s u . of squares of
the residuals over the data set. The assumptions for the least squares estimation are

(McLellan, 1998):
1. The values of the explanatory variables are known exactly;

2. The form of the equation provides an adequate representation for the data;
3. The variance of the random error is constant over the range of data collected;
4. The random fluctuations i each measurement are statistically independent fiom those n

of other measurements.

In the assessrnent of mode1 adequacy one most often makes the assumption that the
random fluctuations are normally distnbuted.

5.4.5.1 Mode1 for Selectivity of BPA Formation

The analysis of the effects of the diEerent factors on the selectivity of BPA formation lead to the conclusion that only the temperature has a significant effect on this response. Therefore the suggested mode1 is:

The T and the Y, vectors are:

In order to calculate the least squares estimates:

The least squares parameter estirnates are obtained as:

B,= (TT)-'~ E ; T

The parameters for this model are:

The results of the regression are presented in Table 5.10. The significance of this model
was assessed using the Mean Square Regression Ratio, and the Residual Variance Ratio. The Mean Square Regression Ratio is compared against F,, Residual Variance Ratio is compared against F ,

, ,

5.9874, and the

,-, = 8.9406. , ,

This analysis confims

the statisticd significance of the temperature effect.

Table 5.10: Results of the Regression Analysis for the Selectivity of BPA Formation

df

SS

MS

1 287.4003 287.4003 6 90.15157 15.02526

7 377.5519
Parameter

Coefficients Standard Emr

t Stat

Lower 95%

Upper 95%

72.60875 1.370459 52.98134 -5.99375 1.370459 4.37353

It is also straightforward to calculate individual confidence intervals for the parameters.

The results of the regression analysis are summarized in Table 5.10.

To summarize, the proposed model is:

y, = 72.6 1- 5.99 T
(*3.35) (+ 3.35)

(+ value) represents the 95% confidence intervals for the individual parameters.

5.4.5.2 Mode1 for the Selectivity of O-p Isomer Formation

The analysis of the effects of the different factors on the seIectivty of O-p isorner

formation shows that only the molar ratio acetone:phenol has a significant effect on this response. Therefore the suggested model is:

j, =

a., w . , ~
A A

The least square parameter estimates for this model are:

The results of the regression are presented in Table 5.11. The significance of this model
was assessed using the Mean Square Regression Ratio, and the Residual Variance Ratio.

The Mean Square Regression Ratio is compared against FI+ ,

Residual Variance Ratio is compared against F,

,, = 5.9874, ,

and the

,,, = 8.9406. ,

This anaiysis confirms

the statistical significance of the temperature effect. The Mean Square Regression Ratio

ht test shows t a the proposed model accounts for significant trend, and the Residual
Variance R t o indicates that the model is adequate. ai

Table 5.11: Results of the Regression Analysis for the Selectivity

Formation

Isomer

Regession Statistics Multiple R 0.828565 R Square 0.686519 Adjusted R 0.634273 Square Standard 3.498563 Erro r Observations ( 8
F SS MS SignifTcance F 1 160.8321 160.8321 13.13994 0.01 1032 6 73.43968 12.23995 7 234.27181 I 1 Coeficients Standard t Stat P-value Lower 95% Upper 95% Error 19.51375 1-236929 15.77597 4.1 1E-06 16.48709 22.54041 -4.48375 1.236929 -3.62491 0.01 1032 -7.51041 -1 -45709

df

Regression Residual Total Parameter lntercept R

To summarize, the proposed mode1 is:

Y2
A

= 1931-4.48R (+ 3.03) (k 3.03)

(it value) represents the 95% confidence intervals for the individual parameters.

5.4.5.3 Mode1 for the Selectivity of Chromanes Formation

The analysis of the effects of the different factors on the selectivity of chromanes formation shows that the factors that have a significant effect on this response are the temperature (T), the molar ratio acetone:phenol (R), both two-factor interaction effects associated with the molar ratio acetone:phenol (CR and TR), and the three-factor interaction effect (CTR). Therefore the suggested mode1 is:

The least square parameter estimates for this mode1 are:

The results of the regression are presented in Table 5.12. The significance of this mode1
was assessed using the Mean Square Regression Ratio, and the Residual Variance Ratio.

The Mean Square Regression Ratio is compared against F,, Residual Variance Ratio is compared against F,

,,

5.9874, and the

,, ,

8.9406. The Mean Square

Regression Ratio test shows that the proposed model does not account for signficant trend, and the Residual Variance Ratio indicates that the model is adequate.

To summarize, the calculated model is:

j3 = 4.08 + 4.O8T - 1.26R + 1.63CR - 126TR + 1.63CTR (k3-49) ( 3-49) ( 3 k 5-49) (+3 -49) (G.49) (k3.49) (+ value) represents the 95% confidence intervals for the individual parameters.
At the 95% confidence level, ody the intercept and the temperature coefficient are statistically significant. At the 90% confidence levef, al1 the parameters, except the intercept and the temperature coefficient, are on the verge of statistical signincance. This

analysis is in agreement with the calculated effects analysis presented in section 5.4.4.3
and Tables 5-8 and 5.9.

The intercept and the parameter associated with the temperature are significant, d l the
other parameters are not. It should be considered dropping the molar ratio term, al1 the
two and three factor interaction terrns. The new mode1 is:

Table 5.12: Results of the Regression AnaIysis for the Selectivity of Chromanes
Formation
Regression ~tatistjcs 0.947407 Multiple R R Square 0.89758 Adjusted R 0.641529 Square Standard 3.383506

Error
Observations
df
Regression Residual Total Parameter Iritercept T R CR TR CTR

SS

MS

Significance F

5 200.6553 4 . 3105 3.505473 01 2 22.89623 1 1.4481 1 7 223.5515 I P-value Coefficients Standard t Sa tf

Error

0.236722

Lower 95%

Upper 95%

4.07875 4.07875 -1 .25625 1.62625 -1.25625 1 -62625

1.19625 3.409613 1.19625 3.409613 1.1 9625 -1 .O501 6 1 -19625 1.359457 6 1.19625 -1 -0501 1.19625 1.359457

0.076303 0.076303 0.403823 0.306988 0.403823 0.306988

0.5857 0.5857 -4.7493 -1 -8668 -4.7493 -1 -8668

7.571785 7.571785 2.236785 51 19285 . 2.236785 5 1i 9285 .

5.4.5.4 Model for the Yield in BPA

The analysis of the effects of the different factors on the yield in BPA shows that the
factors that have a significant effect on this response are the catalyst concentration (C), the temperature (T), both two-factor interaction effects associated with the temperature

(CT and TR), and the three-factor interaction effect (CTR). Therefore the suggested
model is:
y4

=4 . 0
A

+b4.1c+&.2~fh.3c~+k.4~~+d,5c~~

The least square parameter estimates for this mode1 are:

The results of the regression are presented in Table 5.13. The significance of this model
was assessed using the Mean Square Regression Ratio, and the Residual Variance Ratio.
The Mean Square Regression Ratio is compared against F,, Residual Variance Ratio is compared against F,

,, ,

5.9874, and the

,. ,,,

= 8.9406.

The Mean Square

Regression Ratio test shows that the proposed model accounts for significant trend, and

the Residuai Variance Ratio indicates that the model is adequate.

Table 5.13: Results of the Regression Analysis for the Yield in BPA

1 Regression

lResidual

Ilntercept

SS 5 88.2931 5 2 0.1906 7 88.48375 Coeficients Standard Emr 5.6075 0.109144 -0.6525 0.1 09144 1-3275 0.1 09144 1.1 575 0.109144 -1 -96 0.1 09144 1.915 0.109144
df

F [ Significance F MS 0.005376 17.65863 185.2952 0.09531

t Stat

P-value 0.000379 0.026858 0.006692 0.008774 0.003087 0.003233

Lower 95%

Upper 95%

51.37689 -5.97832 12.16279 10.60522 -17.9579 17.54556

5.137889 6.0771 11 -1-122111 -0.18289 0.8578891 1.7971 11 0.6878891 1.6271 11 -2.42961 -1 -49039 1.4453891 2.38461 1

The calculated mode1 is:

j, = 5.61 - 0.65C+ l.33T

+ l38CT - 1.96TR + 1.92CTR

(20.47) (k0.47)

(3.47) (k0.47) (?0.47)(&0.47)

(k value) represents the 95% confidence intervals for the individual parameters.

5.4.6 Summary of z3 Experimental Design

The results obtained in this experimental design indicate that a 1:2 molar ratio of acetone:phenol in the initial mixture should be used. Moderate temperatures are

desirable. This conclusion agrees with the data obtained fiom both the literature and the

simulation. A 10% catalyst concentration is preferred.

Higher catalyst concentrations favor the production of chromanes and O-p isomer, in other words, the by-products formation. Lower molar ratios favor the by-products

formation. High temperatures influence positively both the formation of BPA and byproducts, but this factor is not significant for the formation of O-p isomer, and the chromanes are produced in much lower quantities than BPA. it is also concluded that more experimental points would be necessary to better estimate the parameters for the models proposed for the selectivity of the chromanes formation and the yield in BPA.

5 5 Additional Runs .
Eight additional nuis were conducted at the sarne levels as for the first set of experiments (see Table 5.1). Table 5.14 lists the experiments in the randomized order they were performed. The particle size of the catalyst used i this second set of experiments was n smailer, therefore some changes in the significance of the effects of the investigated parameters might appear. The main effects, two-factor interaction effects and three-factor interaction effects were calculated and are summarized in Table 5.15.

TabIe 5.14: Additional runs Number VILI.1 VIII.2 VIII.3 VIII.4 W.5
C -1

VLII.6
W.7 VIII.8

+1 +l -1 -1 -1
+1

T +1 +1 +1 -1
-1

R -1 -1
+i

y,

Y2

+1

+1 -1 -1

+1 -1 +1 -1 +1

80.25 75.86 71.67 76.98 100.00 81.57 83.67 f 72.49

17.86 15.53 14.53 12.46 0.00 15.27 16.33 11-61

1 ( 1

Y3

Y 4

1.89 3.00 4.62 0.00 0.00 0.00 0.00 9.09

0.50
2.02

6.57 9.09 1.62 1.70 1.75 4.39

Table 5.15: Calcdated effects for the additional nuis

Effect Y1 Y2 Y3
Y4

C T -8.78 -5.95 3.10 5 . 6 9 0.11 3.71 0.46 1 1 - 5 2

R
-9.27 1.04 2.21 3.97

1 1

CT
1.63 -4.64 -0.84 2.74

CR
1.58 -3.89 3.15 -0.37

TR
7.83 -2.83 -2.34 -1.09

CTR -4.34
4.69 -1.39 2.05

As expected, there are sorne changes in the significance of the calculated effects for the second set of experiments, compared with the fxst set of experiments. The cornparison

n between the two sets of experiments is presented i tabulated form in Table 5.16 (blank
space means that the effect is not significant, - means that the ef5ect is significant and negative, and + means that the effect is significant and positive).

Table 5.16: Cornparison between the calculated effects in the first set and the second set
of experiments

The results require explanation: By increasing the catalyst concentration y), the selectivity of the BPA formation decreases, due to the fact that the catalyst used had a bigger specific active surface, and increasing its quantity, it means that there are more active sites available to cataiyze the formation of the by-products (b).

By increasing the molar ratio acetone:phenol (3, the selectivity of the BPA formation
decreases, which is explained by the fact that the more acetone in the system, the higher the probability that the acetone will react with itself, or maybe with already formed products, to yield by-products. The fact that excess phenol favors the BPA

formation was mentioned in the literature, and contradicted by the simulations and the
first set of experiments. between the two catalysts.

The explanation might be in the daerence in acidity

It is important to ernphasize the fact that the only main factor that appeared significant for the yield of BPA formation in both 23 experimental designs was the temperature. Another noticeable fact is the significance of the two factor interaction effect associated with the catalyst concentration and the temperature, and the significance of the three factor interaction effect.

5.6 z4 Experirnental Design

Due to the changes in significance of the effects of the investigated paraneters, the
second set of experiments cannot be adopted as a replicate set. Therefore , a new qualitative factor was introduced, to account for the two different batches of catalyst, and
a new, Z4 factorial design was considered. The coding for the experhnents which used

cataiyst with bigger particle size was +1 (first set), and the coding for the experiments

which used catalyst with smaller particle size was -1 (second set). The coding for al1

other factors remains unchanged. The data for the z4experimental design are presented in

Table 5.17.

Table 5.17 Data for 2' expenmentai design

VII.12 1 -1 1 O 1 +1 1 +L 1 72.76 * Effect associated with the difference between the twc i batches of catalyst

Main effects, two-factor interaction effects, three-factor interaction effects, and fourfactor interaction effects were calculated and their statistical significance was assessed. It was expected that the four-factor interaction effects are not significant, and indeed, they were not. The calculated effects are presented in Table 5.18, and their significance is presented in tabulated form in Table 5.19 @la& space means that the effect is not

significant, - means that the effect is significant and negative, and + means that the effect is significant and positive).

Table 5.18: Calculated effects for the 2' design

Table 5.19: Significant effects

According to these results, increasing the catalyst concentration, the selectivity of BPA formation will decrease, and the selectivity of chromanes formation will increase. This is due to the fact that higher catalyst concentrations translate into more active sites capable of catalyzing the process, increasing the probability of already formed products to M e r react, and, in this case, to form more chromanes.

The effect of temperature is significant for the selectivity of BPA and chromanes formation, and for the yield in BPA. The higher the temperature, the lower the selectivity of BPA formation, the higher the selectivity of chromanes formation, and the higher the yield in BPA. This is explained by the fact that a higher temperature increases the reactivity of the system. More BPA is formed, but also more by-products are formed.

Increasing the particle size of the catalyst the selectivity of BPA formation decreases, and

hs the yield in BPA increases. T i effect is similar to the catalyt concentration eEect,
since smdler paaicle size means that the active sites inside of the catalyst bead are not as accessible as in the case of the bigger catalyst beads.

if the temperature and the particle size increase simultaneously, the result is an increase in the reactivity of the system, and the combined effect is an increase in the selectivity of the chromanes formation and in the yield in BPA. If the particle size and the molar ratio acetone:phenol are both decreased, the combined effect is an increase in the selectivity of BPA formation and a decrease in the yield in BPA.

5.7 Regression Analysis for the z4Experimental Design

5.7.1 Model for Selectivity of BPA Formation
Based on the significance of the calculated effects, the proposed mode1 for the selectivity of BPA formation (y,) is:

The parameters for this model are:

The results of the regression are presented in Table 5.20. The significance of this model
was assessed using the Mean Square Regression Ratio, and the Residual Variance Ratio.

The Mean Square Regression Ratio is compared against F1.,4,0B5 4.6001 and the Residual Variance Ratio is compared against F,

,- ,, = 3.1 122. ,

The calculated model is:

9, = 76.46 - 3.43C - 4.49T - 335N + 329RN

(52.69) (f2.69)( S . 6 9 ) (52.69) (k2.69) (+ value) represents the 95% confidence intervals for the individual parameters.

It can be noticed that the effect of the temperature on the selectivity of BPA formation is slightly smaller in the case of the combined design in cornparison with the 23 design (-

8.97 compared with -1 1.99). In the combined design other factors become significant
(cataiyst concentration-C, particle size-N, and interaction effect associated with the molar ratio of the reagents and the particle size-RN).

Table 5.20: Results of the Regression Analysis for the Selectivity of BPA Formation

IR Square
Square

Z I
Regression Statistics

) 0.778371

Observations 1
I

Total

261.9971 1182.136 Standard Emr 1.22009 122009 1.22009 1.22009 1.22009

23.81791 t Sfat 62.66751 -2.81 127 -3.67493 -3.1 5653 2.694473

P-value Lower 95% Upper 95% Lower 90.0% Upper 90.0%
2.1 2E-15 0.01 693 0.003658 0.0091 35 0.020862

73.7746 79.1454 -0.7446 -6.1 154 -7.1 6915 -1 -79835 -6.53665, -1-16585 0.602099) 5.972901

74.26886 -5,621 14 -6.67489 -6.042391 1.096361

78.651 14 -1 -23886 -2.29261 -1 -66011 5.47864

5.7.2 Model for Selectivity of Chromanes Formation

Based on the significance of the calcuiated effects, the proposed model for the selectivity
of chromanes formation (y,) is:

The parameters for this model are:

The resdts of the regression are presented in Table 5.21. The significance of this mode1
was assessed using the Mean Square Regression Ratio, and the Residual Variance Ratio.

The Mean Square Regression Ratio is compared against F,.,,,o,

Residual Variance Ratio is cornpared against F,.

4.6001 and the

,. , = 3.1 122. ,

Table 5.21: Results of the Regression Analysis for Selectiviq of Chromanes Formation

Adjusted R Square Standard Observations

0.639385 2.726238

MS SS 5 234.8299 46.96599 1 O 74.32371 7.432371 Residual 1 5 309.1536 1 Total tStat Parameter Coefficients Standard Enor I 4.697866 lntercept 3.201875 0.681 559 1-524375 0.6815591 2.236599 C 3.030734 2.065625 0.681559 T 2.348475 0.681559 CR 1.600625 -1-77993 TR -1.21313 0.681 559 1 2.013125 0.681559 2.953704 TN

F Significance F' 6-349112 0.006741

I
1

P-value

1 1

0.000844 0.049288 0.012661 0.040746 O. IO544 0.0 144-43

Lower 95% 1.683266 0.005766 0.547016 O.082Ol6 -2.731 73 0.49451 6

Upper 95% 4.720484 3.042984 3.584234 3.1 19234 0.305484 3.531734

Lower

90.0%
1.966575

Upper 90.0% 4.4371 75

0.289075
0.830325 0.365325 -2.44843 0.777825

2.759675
3.300925 2,835925 0.022175 3.248425

The calculated mode1 is:

It can be noticed that the effect of the temperature on the seIectivity of chromanes formation is two times smaller in the case of the combined design in cornparison with the
3 design (4.13 compared with 8.16).

The effect of the molar ratio (R) becomes

insignificant in the combined design. The significance of the two factor interaction eEects (CR and TR) is almost the same for both 2) and designs. The three factor

interaction effect (CTR), which appeared in the 2 design, loses its significance in the Z4 design. In the combined design the two factor interaction effect associated with the temperature and the particle size (TN) becomes significant.

The parameter for the two factor interaction associated with temperature and molar ratio
acetone:phenoI is not significant and its deletion is considered. The new model is:

5 7 3 Mode1 for Yield in BPA ..

Based on the significance of the calcdated effects, the proposed model for the yield in
BPA (y,) is:

The parameters for this model are:

The resuits of the regression are presented in Table 5.22. The significance of this model
was assessed using the Mean Square Regression Ratio, and the Residual Variance Ratio. The Mean Square Regression Ratio is compared against F,.,,,,,
= 4.6001 and the

Residual Variance Ratio is compared against F,

,,-,,, = 3.1 122.

TabIe 5.22: Results of the Regression halysis for the Yield in BPA

Regression Stafistk 1 I Multiple R 10.953169 R Square 0.90853 Adjusted R 0.803994 Square 1-488213 Standard

Error Observations
df

16

Regression

Residual
Total Parameter

lntercept T R N
CT TR TN RN CTR

F SS MS SignKcance F 8 153.9895 19.24869 8.691 023 0.004978 7 15.50345 2.214779 15 169.493 Coeficients Standard t Stat P-value Lower Lower Upper Emr 90.0% 95% 95% 4.531 25 0.372053 12.17903 5.76E-06 3.651484 5.41 1016 3.826366 0.285 0.372053 0.766019 0.468703 -0.59477 1.164766 -0.41988 0.92125 0.372053 2.476124 0,04245 0.041484 1.801016 0.216366 1.O7625 0.372053 2.892731 0.023226 0.7 96484 1.956016 0.371 366 1.26375 0.372053 3.396691 0.01 149410.383984 2.1435161 0.558866 -1.2525 0.372053 -3.36645 0.01 19781 -2.1 3227 -0.37273 -1-95738 1.0425 0.372053 2.802018 0.026447 0,162734 1.922266 0.337616 -1 .O6125 0.372053 -2.85241 0.024603 -1-94102 -0.18148 -1.7661 3 1-46875 0.372053 3.947687 0.005548 0.588984 2.348516 0.763866

Upper 90.0% 5.2361 34 0.989884 1.626134 1.781 134 1.968634 -0.54762 1.747384 -0.35637
2.173634

The proposed mode1 is:

The effect of the catalyst concentration ( C ) becomes insignificant in the combined design. The effect of the temperature on the yield in BPA is just slightly srnaller in the case of the combined design in comparison with the 2' design (2.09 compared with 2.66). The significance of the two factor interaction effect associated with the catalyst concentration
' and the temperature (CT) is almost the same for both z3 and 2 designs. The two factor

interaction eEect associated with the temperature and the molar ratio of the reagents (TR) is smaller in the case of the combined design i comparison with the 2' design (-2.51 n compared with -3.92). The three factor interaction effect (CTR) is significant in both z3

and z4 designs. In the combined design some factors become significant: the molar ratio

(R), the particle size 0, factor interaction effect associated with the temperature the two
and the particle size (RI), and the two factor interaction effect associated with the molar ratio of the reagents and the particle size (RN).

This chapter presented the results obtained for the experiments performed in the NMR tube and in the batch reactor. These experiments investigated and evaluated the reactivity of the system, blank reactions, and experimental reproducibility. A scherne of reaction

was set up, based on the results obtained. New catalysts were tested and found suitable
for producing BPA.

The effects of temperature, catalyst concentration, and molar ratio acetone:phenol in the

initiai reaction mixture were examined in depth, also the variation of BPA and by-

products formation and the variation of the yieId of BPA with respect to t h e were analyzed. The results were compared to data obtained fiom Literature and simulation.

The examination of the experimental design provides a better understanding of the operating conditions and the effects the chosen factors have on the system under investigation.

The purpose is to maximize the amount of BPA produced while

minimizing the number and the amount of by-products produced.

The analysis of the results obtained in the 2" experimental design indicate that a moderate temperature is desirable, a 10 wt% catalyst concentration and a 1:2 molar ratio acetone:phenol. At temperatures close to the upper limit of the experimental range, the

n yield i BPA is higher, but so is the formation of chromanes. This fact c o b s the
fndings in the literature and the simulation results. The initial molar ratio of acetone and phenol is significant only for the yield in BPA, and a stoichiometric ratio is defnitely preferred, which confirms the simulation results and contradicts the data in the literature.

The particle size of the catalyst beads also influences the production of BPA. Larger
amounts of BPA were obtained with the cataiyst with bigger particle size, and better selectivities of the BPA formation w t the catalyst with smaller particle size. This can ih

be explained by the fact that the occurrence of swelling of the smalier particles of catalyst

was insuffkient and the access of the reagents to the acidic sites inside the catalyst
particle w s reduced. Harmer et al., 1996 indicate that the accessibility of the active sites a inside the catalyst can be improved by using as catalyst a new rnaterial instead of the
156

basic stmctural polyrner Nafion@,that is Nafion" SAC-13, which is essentially silica irnpregnated with the basic smictural polymer ~ a f i o n ? T i new material seems to hs fomuiately combine the benefits of the porous structure of the silica and the super acid capabifities of ~afion".

The next chapter presents the results of the experiments performed in the plug flow reactor. The intent of these experirnents is to take M e r the investigation of the process

of BPA formation and to identiw Iines of future work.

Chapter 6

Reactions in the Plug Flow Reactor

Preliminary results obtained in the experiments performed in the plug flow reactor (PFR)
are presented and discussed in this chapter (see PFR diagram in Figure 4.1).

The

catalysts used for these

nins

were: activated alumina acidified with hydrochloric acid

(AA 300/HCl), Nafion@NR-50, and bJafonmSAC- 13. Al1 the experiments presented in

this chapter are summarized in Table 6.1.

Table 6.1: Summary of the experiments

Exp. #

Tirne (h)
24 24 24 24 24

Catalyst Type

Flow Rate Temperature Acetone:Pheool Molar Ratio Wh) (Oc)

4.0 4.8 4.8 4.8 4.0

IX.2 X. 1 X.2 X.3 XI. 1

AA 300/HC1 Nafion@ NR-50 Nafion@ NR-50 NafionmNR-50 Nafion@ SAC-13

92 102 102 102 92

1:2 1:s 15 15 1:2

6.1 Reaction with Acidic Activated Alumina

Several experiments were tried with AA 300/HC1 in the batch reactor, but only after half hour the reaction mixture became cloudy and the reactions were stopped. The activated
alumina was not robust enough to withstand mixing and it crushed. This is the reason

why a plug flow reactor was necessary to Uivestigate this system. Two reactions were performed using as catalyst activated alumina acidified with hydrochlonc acid (AA 300MC1)-

For the first reaction, a 8 1 mesh AA 300 was acidified for two hours with a 2:l x4 solution of hydrochloric acid and water (volurneiric proportion) at room temperature. The catalyst was dried in the reactor with hot nitrogen (105"C),overnight. The reactor
was fed with a 1:5 initial mixture of acetone and phenol (molar ratio) using a syringe

pump previously calibrated, at a rate of 4.8 cch. The temperature was maintained constant during the reaction, at 102OC, by means of a thennocouple located in the catalyst bed, attached to a PID (proportional-integral-derivative) controller, that regulated the power of the heater. The reaction was stopped after 24 hours.

The product was cloudy. M e r separation, a sarnple was anaiyzed on the GC-MS. Although a BPA peak was observed, the quantity produced was small. No other products were observed.

Taking into account the resdts of the first reaction, a second experiment was performed.
With the intent of increasing the yield in BPA, some modifications were considered. 20 grarns of a 14x18 mesh AA 300 were acidified for two hours with concentrated hydrochloric acid at room temperature. The catalyst was dried ovemight in the reactor
with hot nitrogen (105C). The reactor was fed with a 1:2 initial mixture of acetone and

phenol (molar ratio) using a syringe pump previously calibrated, at a rate of 4.0 cch. The temperature was maintained constant during the reaction, at 92"C, by means of a thermocouple located in the catalyst bed, attached to a P D controller, that regulated the power of the heater. The changes in process conditions were intended to increase the activity of the cataiyst and the retention time in the reactor. Both variables were changed to increase the contact time between the catalyst and the reagents. A 1:2 molar ratio of

rm the initial reagents was preferred, based on the conclusions f o the previous
investigations. The reaction was stopped after 24 hours. The o d y product observed in significant amount was bisphenol A. The results are presented in Table 6.2.

Table 6.2: Results of the experiments with AA 3001 HCl

Exp. #

Product Distribution' (wtO/o) Temp. Yield Reaction Molar Flow Time (h) Ratio cc/h O C 1 n III+N V wt% 102 0.00 0.00 IX. 1 24 4.8 100.00 0.00 1:5 c 0.5 0.00 92 0.00 0.00 IX.2 4.0 1.79 100.00 24 1:2 ' I is bisphenol A; II is 2,4'-isopropylidenediphenol; III is 4'-hydroxyphenyl-2,2,4; trimethyl chroman 1 IV is 4'-hydroxyphenyl-2,4,4-trimethylchroman II; V are other byproducts.

The yields in BPA obtained in these two experiments are comparable to the yields

obtained for some of the experiments presented in the previous chapter (experiments

WI. 11, VIII. 1, VIII.2, W . 5 , VIII.6, VIII.7).

The selectivities obtained in the

experiments performed in the PFR using AA 300/HC1 as catalyst are 100%, higher than the selectivity obtained in any previous experiment These prelirninary results prove that acidified activated alumina is a suitable catalyst for the production of bisphenol A.

6.2 Reactions with Piafion@ NR-50

The reactions with Nafion@NR-50 (the basic polymer) were performed at 102OC. The
catalyst was dned prior to the reaction with Nz at 105OC. The reactor was fed with a 1:5
initial mixture of acetone and phenol (molar ratio), using a syringe purnp previously

calibrated, at a rate of 4.8 cch. The temperature was maintained constant during the reaction, at 102"C, by means of a thermocouple located in the catalyst bed, attached to a

PID controller, that regulated the power of the heater. The reaction was stopped after 24
hours. Between the two reactions the catalyst was regenerated with a 15% solution of niic acid, at 50C, which was allowed to flow through the catalyst bed at a rate of 20

cc& for six hours.

Two samples fiom the fist reaction were prepared and analyzed, one on the GC-MS, the other one on the NMR. The product fiom the second reaction separated hto two layers, a light one and a dark one. Both layers were analyzed on the GC-MS.

nie reason for the separation is believed to be the fact that the catdyst was not washed
and dried well enough after regeneration, therefore an aqueous and an organic layer

existed in the system. Also the presence of the nitric acid made possible the formation of some nitro compounds, which were observed in the light colored fiaction. No BPA or any of its isomers were detected in the light colored bction. Some BPA, 0-0 isomer and o-p isorner were observed in the dark colored fraction. distribution and yield for both reactions are presented in Table 6.3. The product

Table 63: Results of the experiments with ~ a f i o n NR-50 @

Exp. #

Product Distribution (wt%) wt% I II III+TV V VIc X. 1 4.1 1 82.47 17.53 0.00 0.00 X.2 " 1O 0 x.2 31.17 7.97 0.00 0.00 60.86 79.63d 20.37' a Light colored fraction of the product in the second reaction Dark colored fiaction of the product in the second reaction Products unidentified, and obtained oniy in the reaction performed with regenerated catalyst, believed to be nitro derivatives Selectivity of BPA formation, if the unidenified products are not taken into account Selectivity of o-p isomer formation, if the unidentified products are not taken into

Reaction 1 Molar Flow Time Ratio c c 5 24 4.8 15 24 i:5 4.8 24 1 5 4.8

Temp. O C 102 102 102

Yield

A third experiment was tried with the sarne catalyst, but it was stopped because the

pressure was building up rapidly. The reason appeared to be the fact that the polymer particles swelled and expanded to the extent that they almost formed a block inside the reactor, without leaving any space for the reactants or the nitrogen to flow through. This situation might be solved by mWng the Nafion@NR-50 particles with g l a s beads, to ensure the flow space in the reactor. Another possible and interesting experiment to examine is to mix the Nafion@ NR-50 with acidified activated alumina. Besides solving

the flow problem inside the reactor. one couid also snidy the combined effect of the two cataiysts on the process of BPA formation.

6.3 Reaction with ~ a f i o n @ SAC-13

~afion@' SAC-13 is essentially silica impregnated with the basic polymer Nafion@ NR-50.

This new materid combines the benefits of the porous structure of the silica with the
super acid properties of ~ a f i o n ? The purpose of this reaction is to evaluate 60m the qualitative point of view the suitability of this new material to catalyze the production of bisphenol A.

The reaction with Nafion@ SAC-13 was performed at 92OC. The catalyst was dried prior
to the reaction with N,at

Mac. The drying temperature, 155"C, was arbitrarily chosen

between 105C and 160C, since the drying process can occur ordy at temperatures above 100aC, and on the instructions that came with the catalyst drying temperatures under

160C were indicated. The reactor was fed with a 1:2 initiai mixture of acetone and
phenol (molar ratio), using a syringe pump previously calibrated, at a rate of 4.0 cch.

The temperature was maintained constant during the reaction, at 9SaC, by means of a
thermocouple located in the catalyst bed, attached to a P D controller, that regulated the power of the heater. The reaction was stopped after 24 hours.

The phenol used in this reaction was not the liquid form as for al1 the other experiments.

The crystals form was used instead. The reason for this was that a1 of the liquid phenol l

was consurned, and Fisher Scientific stopped distributhg this product, because they could

not stabilize their b ~ c h e s . Other suppliers were contacted, but they do not fiunish the phenol in liquid fom, with 9% water as impurity. As a result, the water content in this system was lower and a decrease i the yield was expected, since it was found in the n Iiterature that about 10% water in the initial reaction mixture increases the rate of the reaction (Scheibel, 1974).

The only products observed were mesityl oxide,

0-0

isomer, and BPA, in very s m d

quantities. Two samples were prepared and analyzed, one on the GC-MS, the other one on the NMR. The product distribution and the yield for the reaction are presented in Table 6.4.

Table 6.4: Results of the experiment with ~ a f i o n @ - 13 SAC

1 Exp. 1 Reacion 1

Molar ( Flow

Temp.

Yield

Product Distribution (wtY0)

This reaction proves that bisphenol A can be obtained in a process that uses Nafion@ SAC

13 as catalyst. The yield obtained in this experiment is comparable with the yield

obtained in experiment VIII.1 and the selectivity obtained is much lower than the selectivities obtained in any of the previous experiments. Further investigation is

required to see if the reactor configuration and the catalyst are feasible for commercial production. The fact that only mesityl oxide and 0-0 isomer were obtained, besides BPA,

indicates that an excess phenol and a lower ternpenture might be more appropriate for
this systern.

6.4 Summary
This chapter presented the results obtained for the experiments performed in the plug flow reactor. These experiments investigated the siiitability for producing bisphenol A of

~ r new catalyst, that could not be investigated in the batch reactor. These experiments o

NR-50 in the plug flow reactor. also studied the behaviour of the Nafion@

A three new catalysts: acidified activated alumina, Nafion@NR-50,and Nafion@SACU

13, were found suitable to catalyze the production of bisphenol A in a flow system, using
phenol and acetone as starting materials. The temperature, the molar ratio of the initial reagents in the feed, the retention tirne, and the type of cataiyst seem to be factors that significantly influence the process of BPA formation.

The next chapter presents the final conclusions of the present study and identifies
directions for future investigation. Keeping in mind that the motivation for this work is
the production of bisphenol A via catalytic distillation, the next step is to optimize the

reaction configuration in the plug flow reactor.

Chapter 7

Conclusions and Recommendations

7.1 Conclusions
The synthesis of bisphenol A @PA) with heterogeneous catalysts was investigated in a batch system and in a plug flow reactor. Experiments were conducted with ~rnberlyst@'

15, Nafion@NR-50, ~afon@' SAC-1 3, and activated alumina acidified with concentrated
hydrochlonc acid (AA300/HCl). Gibbs reactor simulations were also conducted. The resdts of this investigation are surnmarized below.

Both the simulation and the experirnental work indicated that the species present in the system are remarkably reactive in the provided conditions. It was also shown that the stoichiometric ratio acetone to phenol (1:2) represents a better choice for operation instead of excess phenol as mentioned in the literature. The results indicate that a moderate temperature and a 10% catalyst concentration are desirable. A t temperatures close to 100C the yield in BPA is higher, but so is the formation of chromanes. This fact confirms the findings in the literature and the simulation results.

The initiai molar ratio of acetone and phenol is signifiant o d y for the yield in BPA;

the closer to the stoichiometric ratio, the higher the yield in BPA.

The particle size of the catalyst beads in the case of Nafion@NR-50 also influences
the production of BPA. Larger amounts of BPA were obtained with the catalyst with bigger particle size, and better selectivities of the BPA formation with the catalyst with smaller particle size.

Ml three new catalysts: AA300/HC1, Nafiof NR-50, and ~ a f i o n @ SAC-13, were

found suitable to catalyze the production of bisphenol A in a flow system, using phenol and acetone as starting matenals. The temperature, the molar ratio of the initial reagents in the feed, the retention t h e , and the type of catalyst are factors that significantly (in a statistical sense) infiuence the process of BPA formation.

z4 experimental design was

performed. The experiments were done in a stirred

batch reactor under continuous reflux, using Nafion@NR-50 as catalyst and acetone

and phenol as starting matenals.

detailed below:

The results of the Z4 experirnental design are

1. The selectivity of BPA formation decreases by increasing the catalyst concentration. 2. The selectivity of chromanes formation increases by increasing the catdyst concentration.

3. The higher the temperature the lower the selectivity of BPA formation.
4. The selectivity of chromanes formation increases with increasing temperature.
5. The yield of the process increases with the temperature.

6 . The selectivity of the BPA formation decreases by increasing the particle size
of the catalyst bead.
7. The yield of the process increases with increasing the size of the catalyst bead.

This effect is similar with the catalyst conceniration eEect, since smaller
paaicle size means that the active sites inside the catalyst bead are not as accessible as for the larger catalyst beads.

8. If the temperature and the particle size increase simultaneously, the result is an
increase in the reactivity of the system, and the combined effect is an increase in the selectivity of the chromanes formation and in the yield in BPA. If the particle size and the molar ratio acetone:phenol are both decreased, the combined effect is an increase in the selectivity of BPA formation and a decrease in the yield in BPA.

9 . The significance of the three factor interaction involving the catalyst

concentration, the temperature and the initial molar ratio of acetone and phenol indicates the possibility that an optimum might exist in the operation of the system.

The available results indicate that the operating conditions that rnaximized the

quantities of bisphenol A and muiimized the quantities of by-products are: a

temperature of 82"C, a molar ratio of 1 2 acetone:phenol in the initial reaction

mixture, and a 10% catalyst concentration, for both the catalyst w t smaller particle ih
size and the catalyst with bigger particle size. The quantities of BPA obtained in these two cases were 9.09g BPA/100g crude, and 9.79g BPN100g cmde respectively. No chromanes were obtained in any of these two experiments. The experiment performed with catalyst with larger particle size had a higher selectivity of BPA formation (84.55%, compared to 76.98% for the experiment performed with catalyst
with catdyst with smaller particle size).

The significance of this work is that the yields and the selectivities obtained for the
processes conducted with the newly identified catalysts are better than for the process
15. catalyzed by ~ r n b e r l y s t ~ While the importance of a better yield is obvious, higher

selectivities for this process is critical, since the demand of high purity bisphenol A on the market is constantly increasing, and the separation and purification processes for obtaining higher punty bisphenol A are complicated and costly.

7.2 Recommendations
Future experiments to investigate the synthesis of BPA should be continued in the plug flow reactor. The choice of the catalyst type effect wodd be of great

importance. Future experirnents should be conducted to investigate the effect on yield and selectivity of compounds containing mercapto groups. Also the effect of the water content in the initiai reaction mixture is a factor requirhg investigation.

The resdts indicate the possible presence of an optimum within the ivestigated

experimental region. The search for the o p h u m imply an elaborate experimental

program. Further experiments should be performed to verie the presence of the optimum and to identiQ the optimum. in case it exists.

One of the most hstrating parts of this work was to find a method of analysis. Two
reasons contributed to this hstration:
1. Most of the by-products obtained in the process were not available as

standards;
2. Complete separation was not possible for the chromanes.

In order to overcome these difficulties, the use of high pressure liquid

chromatography is advisable, as is the use of standards for calibration. These standards could be obtained fiom a big producer of BPA, Shell Ltd. for example, by getting them invo lved in the research project.

Appendix A: Health and Safety Considerations

The ex~erimental aspect of this research entailed several safety concerns which had to be de& with. The batch reactor was installed in the fiune hood, and the PFR was installed
within a c o n t h e n t bunker to reduce the possibility of incidents. The fume hood

shielded the experimenter fiom splashes and protected against toxic vapors. The bunker shielded the experimenter fiom debris in case the pressure would have increased and run out of control, and it provided a barrier against toxic vapors. Al1 the components used or produced in the reaction could have been hamiful in sufficient quantities, except the solid catalysts. Standard laboratory procedures required the mandatory use of lab coats, safety glasses and appropriate gloves whenever handling these materials. Table A.l lists the components used or produced in the reaction, hazards associated with these components and the suggested safety requirements. The information in table A.l was obtained Erom the MSDS for the respective chemicais.

Table A.l: Chemicals used in experiments, associated hazards and safety requirements
Safety Requirements Hazards Skin and eye irritant, narcotic, Latex gloves/Eye protectiodLab coat toxic, flammable Neoprene gloves/Eye Phenol Severe eye and skin irritant, protectiodab coat narcotic, toxic, carcinogenic Mild eye and skin irritant, Neoprene gloves1Eye Bisphenol A protectiodLab coat toxic Latex g1ovesEye protection Asphyxiant Nitrogen Latex glovesEye protection Mild eye and skin irritant ~afion@' Latex gloves/Eve * protection Not hazardous Activated alumina Hydrochloric acid 1 Severe eye and skin irritant, 1 Neoprene gloves/Eye 1 toxic and corrosive 1 ~rotectiodLab coat * Latex glovesEye protection Skin and severe eye irritant Amberlyst 15 Polyvinyl alcohol gloves/Eye Severe eye and skin irritant, Deuterated chloroform protectiofiab coat toxic and carcinogenic Neoprene gloves/Eye Severe eye and skin irritant, Nitric acid 1 toxic and corrosive 1 protectiodLab coat

Chemical Acetone

Appendix B: PRO/II@Input File

The simulations were run using a windows based version of PROLI'.

Acetone:phenol

ratio and temperature were the oniy factors changed at each nui.

The flow sheet for the simuiation is:

-FIT
Table B. 1 contains the input file used for the simulations.

Table B.l:

PRODI" Keyword Input File

Generated by PRO/iI Keyword Generation System <version 2.71 - 02-14-95> E Generated on: Thu Oct 09 15:39:38 1997

rITLE DIMENSION SI, STDTEMP=273.15, STDPRES= 101.325 SEQUENCE SIMSCE CALCULATION RVPBASIS=APIN, TVP=3 10.93 COMPONENT DATA LIBID 1,H20/2,ACETONE/3 ,PHENOL/4,BSPHNOLA FILL=SIMSCI NONLIB 5,Isopropylidene/6,Chroman/7,Triphenol, STRUCTURE 5,111 (2),lOO(8), 1201(2),9OO(2),9O3(1)/ & 6,li 1(1), 1200(8),1225(2),900(3),90 1(1),655(1)/ & 7,111(3), 1ZOO(l l), 1201(4),9OO(4),9O3(2) NBP 4,493 SOLUPARA 5,9.6034 N M P 5,383.15 CNUM 5,15 ZNUM 5,-14 FORMATION(V,M) 5,-3.6928E5 T'HERMODYNMC DATA METHOD SYSTEM=NRTL, ENTROPY(L)=SRK, SET-NRTLOI, & DEFAULT STREAM DATA PROPERTY STREAM=S 1, TEMPERATURE=343.15, PHASE=M, & RATE(M)=45 -3593, COMPOSITION(M)=2,25/3,75, & NORMALIZE, SET=DEFAULT UNIT OPERATIONS GIBBS UID=Rl FEED S l PRODUCT M=S2 OPERATION PHASE=M, I S O T H E W PARAMETER PHYSPROP=1 ELEMENTS REACTANTS= 1/2/3/4/5/6/7, COMPONENTS=1,2, 1,O/2,6,1,3/ & 3,6,1,6/4,16,2,15/5,16,2,15/6,20,2,18/ & 7,26,3,24, NAMES=H,O,C END
-

Appendix C: Summary of Simulation Results

The following 5 tables show the moi percent content of bisphenol A, og-isomer, and triphenol in the final product Stream as a function of temperature and molar ratio over the considered range of temperature, at 1 atm. The final table shows the variation of

selectivity of bisphenol A (I), ogisomer ( I , I and triphenol (III) with the temperature at ) various acetone:phenol molar ratios. The initial reaction mixture consists of acetone and phenol only.

Table C.1: Variation of bisphenol A, o,p-isomer, and triphenol formation with the acetone:phenol molar ratio at 323.15 K. The results are presented in mol %.
Temperature

Molar Ratio

Bisphenol A (

og-Isomer

Triphenol

Table C.2: Variation of bisphenol A, o,p-isomer, and triphmol formation with the acetone:phenol rnolar ratio at 333.15 K. The resdts are presented in mol %.
Temperature

Molar Ratio

Bisphenol A

Table C.3: Variation of bisphenol A, o,p-isomer, and triphen01 formation with the acetone:phenol molar ratio at 343.15 K. The results are presented in mol %.
Temperature

Molar Ratio

Bisphenol A

* The equilibrium was not achieved.

the 50 iterations.

The numbers r lect the result obtained at the end of c

Table C.4: Variation of bisphenol A, o,p-isomer, and triphend formation with the acetone:phenol rnolar ratio at 353.15 K. The results are presented in mol %.
Temperature

Molar Ratio

BisphenoiA

op-Isomer

Triphenol

Table CS: Variation of bisphenoi A, o,p-isomer, and aiphenol formation with the acetone:phenol molar ratio at 3 63.1 5 K. The results are presented in mol %.
- -

Temperature

Molar Ratio

Triphenol
0.0006 0.0028 0.0074 0.0 156 0.0305 0.060 1 0,1336 0.2967

Table C.6:

Variation of selectivity of bisphenol A ( ) I , o,p-isomer (I, and triphenol I) (III) with the temperature at various acetone:phenol molar ratios.
Molar Ratio Ac:Ph 0.05
Temperature (K)
323.15

Selectivity (%)
1 69.77

I I
30.23

III 0.00

Appendix D: The NMR Phenomenon

Nuclear Magnetic Resonance (NMR) spectroscopy is a method of great interest and importance for the study of chemical substances. The use of puised Fourier transform methods with spectnim accumulation made it possible to obtain high resolution spectra (Sanders and Hunter, 1993).

Do1 Magnetic Energy Levels and Transitions

When the spin quantum number 1 of a nucleus is nonzero, the nucleus possesses a magnetic moment. This condition is met if the mass number and atomic number are not even. The proton ('H) has a spin 1 of X. When placed in a magnetic field of strength Bo, nuclei with nonzero 1 occupy quantified magnetic energy levels, cailed Zeeman levels, the number of which is equal to 21 + 1 (Sanders and Hunter, 1993). The relative population of the Zeeman levels are normally given by a Boltzmann distribution. Transitions between energy levels can be made to occur by means of a resonant radio r) fiequency (f field B, of fiequency vo. A way of picturing the resonance phenornenon e s e s fiom the fact that when placed in a magnetic field, a nucleus undergoes Larmor precession about the field direction at a rate given by v, or a, ( a o is the resonant fiequency, o, = 2nv0, rad/s). Transitions between energy levels occur when the fiequency of the rf field equals the Larmor precession frequency (Sanders and Hunter, 1993).

D.2 The Chernical Shift

Resonaxce occurs at slightly different fiequencies for each type of proton, depending on its chemical binding and position in a molecule. This variation is caused by the cloud of electrons about each nucleus, which shields the nucleus against the magnetic field, thus requiring a slightly lower value of v, to achieve resonance than for a bare proton (Sanders and Hunter, 1993). Protons attached to or near electronegative groups such as OH, OR, OCOR, COOR, and halogens expenence a lower density of shielding electrons and resonate at higher v,. Protons farther removed fiom such groups, as in hydrocarbon chains, resonate at lower v,. These variations are calied chemical shifts and are commonly expressed in relation to the resonance of the tetramethylsilane (TMS) as the zero of reference. The total range of proton chernical shifts in organic compounds is on

the order of 10 ppm, e.g. ca 1 kHz in a magnetic field of 2.34 T (Sanders and Hunter, 1993). For any nucleus, the separation of chemically shifted resonmces, expressed in H z , are proportional to Bo. When expressed in ppm, as common, the chemical shifts are independent of Bo. The eiectronic screening of nuclei is actually anisotropic so that the chemical shift is a directional quantity and depends on the orientation of the molecule with respect to the direction of the magnetic field. In solution, the motional averaging produces an isotropie value of the chemical shift.

D.3 Nuclear Coupling

Nuclei sficiently removed fiom each other do not feel the effects of the magnetic fields of the other nuclei. In this case, the locai magnetic field at each nucleus is essentially equal to Bo. If Bo can be made very homogeneous over the sample, the width of the resonance lines may be very small.

D.3.1Direct Dipole - Dipole Coupling

In most substances, protons contribute to local fields and are sufficiently numerous to atoms also conhbute to the local fields, but their naturai have a marked effect. The 13c abundance is very small, therefore they do not have a visible effect.

D3.2Indirect Nuclear Coupling

Magnetic nuclei may transmit information to each other concerning their spin states not only directly through space, but also through the intervening covalent bonds. This is indirect or scalar nuclear coupling, aiso known as J coupling. Rapid tumbling of the molecde does not reduce this interaction to zero. If the nucleus has n suffIcient1y close, equivalently coupled spin -% neighbors, its resonance will be split into n + 1 spin states of the neighboring group of spins. Thus one neighboring spin splits the observed resonance to a doublet, two produce a 1:2: 1 triplet, three a 1:3 :3: 1 quartet, and so on. The strength of the coupling is denoted by a coupling constant J and is expressed in Hz.

Referen
Acorn NMR, www.acornnrnr.com (1 993, 1994, 1995) Anderson, W.M., G.B. Carter and A.J. Landua, Anal. Chem., 31, 1214-1217 (1959). Arnold-Stanton, R. and D.M. Lemal, The Journal of Orgunic Chemistry, 56, 146-151
(1991).

Asaoka, S., T. Maejima, K. Sakashita, N. Yoneda, M. Yasui, N- Onda, T. Watanabe, N. Moriya, A. Shindo, H. Nishijima, A. Kukidome, R. inaba, T. Imazeki, K. Shimogawara, U.S. Patent 5,324,867 (1994). Asaoka, S., T. Maejima, K. Sakashita, N.Yoneda M. Yasui, N. Onda, T. Watanabe, N. Moriya, A. Shindo, H. Nishijima, A. Kukidome, R. Inaba, T. Imazeki, K. Shimogawara, US.Patent 5,382,7 12 (1995). Berg, K. and H.J. Buysch, US Patent 5,302,774 (1994). Berg, K., G. Malamet, A. Eitel, C. Wulff, U.S. Patent 5,395,857 (1995). Brookman, P.J. and J.W. Nicholson, Developments in Ionic Polymers, 2,269-283 (1986). Butler, G.B., K.F. O'Driscoll, G.L. Wilkes, JMS-Rev. Mucromol. Chern. Phys., C34(3), 325-373 (1994).
Catana, E., 1 Mateescu, V. A. Giurcaneanu, Pinte* Revista de Chimie, 44(5), 475-481 . (1993).

Challa, G. and P.H. Hermans, Anal- Chern., 32,778-779 (1960).

Chernical Market Reporter, 254(3) (1998).
Cipullo, M.J., US Patent 5,146,007 (1992). Cuais, FLF., J. Chem. Soc., 415-41 8 (1962). Dachs, N. W., J. W. Ginter, J.A. Pawlak, U.S. Patent 4,359,590 (1982). Dethmers, F.H.D. and E. de Jong, Red. T m Chim. Pavs-Bas, 84,460-464 ( 1965). Faler, G.R. and G.R. Loucks, US Patent 4,294,995 (198 1).

Faler, G.R. and G.R. Loucks, US Patent 4,346,247 (1982). Fder, G.R. and G.R. Loucks, US Patent 4,396,728 (1983). Faler, G.R. and G.R. Loucks, US Patent 4,424,283 (1984). Faler, G.R. and G R . Loucks, US Patent 4,455,409 (1984). Flippen, J.L., J. Karle, and LL. Karle, J. Am. Chem. Soc-, 92,3749-3755 ( 1 970). Faler, G.R., M.J. Cipullo, U.S. Patent 4,766,254 (1988).

Harmer, M.A., W.E. Farneth, Q. Sun, ,LA. C.S., 118(33), 7708-77 15 (1996).
Hileman, B., Chernical and Engineering News, 75(12) (1997). Hill, D.J.T., D.A. Lewis, and J.H. 09Donnell, J-MacromoZ-Sci-Chem., A29(1), 11-18 (1992).
Iimuro, S., T. Kitamura, Y. Morimoto, US Patent 4,918,245 (1990).

Jackson, B.W., Chee 470-Compurer Aided Plant Design, Course Notes, Queen's University, Departmant of Chemicai Engineering (1997). Kato, T., Nippon Kagaku Zasshi, 84,458-46 1; Chem. Abstr., 59, 13781 (1963).
Kiedik, M., J. Kolt, J. Marszycki, E. Zajac, T. Bek 2. Swiderski, A. Rzodeczko, J. Mroz, J. Olkowska, U S . Patent 5,198,59 1 (1993).

Kishore, K, K.S. Annakutty, I.M. Mallick, PoIymer 29(4), 762-764 (1988). Korenman, Ya.1. and A.A. Goronkhov, Zh. Khirn., 47, 2058-2059 (1973); Russ. J. Phys. Chem., 47, 1157-1 158 (1973); Zh-Prikl. Khim., 46, 2597-2599 (1973); J. Appl. Chern. U.S.S.R., 46,275 1-2753 (1 973).

Li, S.M., US Patent 4,822,923 (1989).

Mahajan, S.S., B.B. Idage, N.N. Chavan, S. Sivaram, Journal o Applied Pot'yrner f Science, 6l(l3), 2297-2304 (1996). Mauritz - ~af5on" Research, www.psrc.usm.edu/mauritz/nafion.htm.

McKetta, J.J. and W.A. Cunningham, (eds.), Encyclopedia o Chernical Processing and f Design, 4,406-425 (1976).

McLellan, P.J., Chee 80 I -Strategies for Process Investigation, Course Notes, Queen's University, Departmant of Chernicd E n g i n e e ~ g 998). (1 Nenitescu, C.D., Chimie Organica, 2,47,691 (1980).
Ning, X. and H. Ishida, Journal o Polymer Science, Par[ A: Polymer Chernisny, 3 ( ) f 26, i 121-1129 (1994).

NTP Chernical Repository Bisphenol A, http://ntp-server.niehs.nih.gov (199 1).

Podrebarac, G.G., F.T.T. Ng, G.L. RempeI, C h t e c h . , 27,37-45(1997). Pouchert, C.J. and I. Behnke, The Aldrich Librmy o 1 3 and I H FT NMR Spectra f ~ (1993). Powell, J.B. and C.W. Uzelmeier, US Patent 5,008,470 (199 1).

Prahl, W.H., S.J. Lederman and E.I. Lichtblau, U.S. Patent 3,219,549 (1965).
Prahl, W.H., S.J. Lederman and E.I. Lichtblau, U.S. Patent 3,290,390 (1 966).

Pressman, E.J. and P.R. Willey, US Patent 4,584,4 16 (1 986). Reference Manual, Simulation Sciences Inc. O (1994 - 1997). Rowton, R.L. and R.R. Russell, J. Org. Chem., 23, 1057-1O58 (1958). Rudolph, U.,N. Bachem, C. WdR, US Patent 5,284,98 1 (1994). Sakashita, K., S. Asaoka, A. Shindo, N. Moriya, Chemeca, 21(2), 543-548 (1993). Sanders, J.K. and B.K. Hunter, Modern NMR Specrroscopy (1993). Scheibel, E. G. German Offen. 2,353,940; Chem. Abstr., 81,64271 (1974).

Shaikh, A.G. and S. Sivaram, Polymer, 36(16), 3223-3228 (1 995).

Shaikh, A.G., S. Sivaram, C. Puglisi, F. Samperi, G. Montaudo, Polqmer Bulletin (Berlin), 32(4), 427-432 (1994).

Shinohara, Y. Japanese patent 70,39251 (1970); Chem. Absn:, 74,142598 (1971). Singh, A.P. Catalysis Letrers, 16,43 1-435 (1992).

Smith, W.R and R.W. Missen, Chernical Reuction Equilibrium Analysis: Theory und Algorithm ( 1 99 1).

Takenaka, J, T. Yamamoto, Y . Hirano, E. Nishi, S . Kodaka and T. Yoshida, Kogvo . 69, Kagaku Zasshi, 70,2081-2088 (1967);Chem. Abst~., 18759 (1968).
Unno, M., T . Kotaki, Y.Mikuriya, T. Dojyo, US.Patent 5,604,072 ( 1 997). Van Ness, H .C . Classical thermodynamics of nonelectrolite solutions: with applications to phase equilibria ( 1982). Verkhovskaya, Z.N. et al. Neftpererab. Nefekhim. (Moscow), 5, 34-35 (1973); Chem. Abstr., 79,66836 (1973). Wasilewska, M. Spec~ochimica Acta Port B: Atomic Specnoscopy, S2(9- 1 O, 1321-1 326 ) (1997).

Yeager, H.J. and A. Eisenberg, Perfluorinated Ionomer Membranes, 1-6,41-63 ( 1982).