from Tertiary Amyl Methyl Ether using a
Polymeric Membrane
Percy van der Gryp (B.Ing, Chemical)
Dissertation submitted in fulfilment of the requirements for the degree of Master of Engineering in the School of Chemical and Minerals Engineering of the
Potchefstroomse Universiteit vir Christelike Hoer Onderwys
Supervisor: Prof. R.C. Everson
Co-Supervisor: Dr. S. Marx
Prof. H.W.J.P Neomagus
May 2003
In this study, the sorption and pervaporation characteristics of the commercial PERVAP2256® membrane with methanol and tert-amy\ methyl ether (TAME) mixtures were investigated. The feed concentration and the feed temperature were varied and quantities such as the degree of swelling, flux and selectivities were measured. The main objective of this study was to investigate the selective removal of methanol from mixtures of methanol and TAME (with an azeotropic point) by means of pervaporation. The experiments were conducted over the entire methanol concentration range at different temperatures ranging from 25 °C to 45 °C.
The membrane was found to be highly methanol selective, with the permeate selectivity 9 1
varying between 5 and 53, and fluxes as high as 12 kg m" h" . It was shown that methanol preferentially absorbed in the membrane and preferentially permeated through the membrane. Methanol can thus be successfully separated by the separation process of pervaporation from mixtures of methanol and TAME that include an azeotropic point.
The permeation of the components through the membrane was modelled, based on the solution-diffusion model. The Flory-Huggins theory was used to describe the sorption step while Fick's first law of diffusion was used to describe the transport step. The predicted values showed good agreement with the experimental measurements in specific regions.
Keywords:
Pervaporation; sorption; methanol-TAME separation; PVA-membranes; solution-diffusion model
In hierdie studie is die sorpsie- en pervaporasiekarakteristieke eienskappe van die kommersiele PERVAP2256® -membraan met metanol en tersiere amiel metiel eter (TAME) mengsels ondersoek. Die voerkonsentrasie en die voertemperatuur is gevarieer en eienskappe soos die swellingsgraad, vloed en selektiwiteit is gemeet. Die hoof doelstelling van hierdie studie was om die selektiewe verwydering van metanol vanuit mengsels van metanol en TAME (met 'n aseotropiese punt) deur middel van pervaporasie te ondersoek. Die eksperimente is uitgevoer oor die hele metanolkonsentrasie gebied by verskillende temperature vanaf 25°C tot 45°C.
Die membraan is gevind om hoogs metanolselektief te wees, met 'n 0 1 permeasieselektiwiteit wat wissel tussen 5 en 53 en die vloed so hoog as 12 kg.m" .hr" . Daar is aangetoon dat metanol by voorkeur in die membraan geabsorbeer en by voorkeur deur die membraan gedring het. Metanol kan dus suksesvol van mengsel van metanol en TAME, wat 'n aseotropiese punt bevat, deur middel van die skeidingstegniek pervaporasie geskei word.
Die permeasie van die komponente deur die membraan is gemodelleer, gebaseer op die oplossingsdiffusiemodel. Die Flory-Huggins-teorie is gebruik om die sorpsiestap te beskryf, terwyl Fick se eerste diffusiewet gebruik is om die oordragstap te beskryf. Die gemodelleerde waardes het goeie ooreenkomste met die eksperimentele waardes getoon in sekere gebiede.
Sleutelwoorde:
Pervaporasie; sorpsie; metanol-TAME-skeiding; PVA-membrane; oplossings diffusiemodel
I, Percy van der Gryp, the undersigned, hereby declare that this dissertation, Separation by Pervaporation of Methanol from Tertiary Amyl Methyl Ether using a Polymeric Membrane, is my own work.
Percy van der Gryp POTCHEFSTROOM
Where do toe Begin with such an essential-part of once work...
"There have been names that I have loved to hear, There have been names that I have loved to mention,
'But never has there been a name so dear To this heart of mine, as the 9{ame divine,
The precious, precious 9{ame of
Jesus Christ."
adapted from Lela B. Long
-Isaiah 42 vers 8 (Soli Deo Gloria)
and then the humans enter the arena of praise, starting first with Sanette, my dear friend, who I wish to acknowledge with gratitude for her constructive comments, suggestions and help in completing this project.
To my Wise father-li/(e supervisor Trof Tjverson who Beyonddoubt plantedthe great seedofresearch in my heart. Ooo... yes andttein (or must I say Trof. 9{eomagus ©), you truly watered my seed for now I can almost see a small flower flourishing. Than^a Iotfor all the help.
To my parents for always listening and sharing in both the joy and the frustration of doing research.
To my friend, Leon, for your understanding and programming sfqfls in modelling of this project.
To Antionette and all my dear friends who prayed a Iotfor me, it made the days seem shorter and the big mountain smaller... Qod bless you all.
"When we become servants of Jesus Christ we quickly recognise the fact that we are NOTHING but sinners saved by GRACE. We have NO POWER of
our own, neither do we control the power of our MASTER. " -Rebecca
Brown-TITLE PAGE i ABSTRACT ii UITTREKSEL iii DECLARATION iv ACKNOWLEDGEMENTS v TABLE OF CONTENTS vi NOMENCLATURE xii LIST OF FIGURES xiv LIST OF TABLES xviii
Chapter 1
GENERAL INTRODUCTION
Overview 1 1.1 Background and motivation 2
1.2 Objectives 6 1.3 Scope of Investigation 6
Chapter 2
BACKGROUND AND LITERATURE SURVEY
Overview 10 2.1 Introduction 11 2.2 General Membrane terminology 12
2.2.1 Definition of membrane 12 2.2.2 Membrane effectiveness parameters 13
2.2.3 Fouling 14 2.2.4 Concentration polarization 14
2.2.5 Swelling of the membrane 15
2.3 Types of Membranes 16 2.3.1 Polymeric membranes 16
2.3.2 Ceramic membranes (zeolites) 17 2.3.3 Advantages and disadvantages 18 2.4 Membrane separation processes 19
2.5 Pervaporation 20 2.5.1 Background 20 2.5.2 Process Description 20
2.5.3 Industrialisation and future industrial development 21
2.5.4 Suppliers of pervaporation equipment 23
2.5.5 Industrial applications 24 2.5.5.1 Hydrophilic pervaporation 25
2.5.5.2 Hydrophobic pervaporation 25 2.5.5.3 Organophilic pervaporation 26 2.5.6 Advantages and Disadvantages of Pervaporation 27
2.6.2 Temperature 28 2.6.3 Permeate pressure 29 2.6.4 Membrane properties 29
2.6.4.1 Thickness of the membrane 29 2.6.4.2 Membrane pretreatment 30 2.7 Separation of alcohol/ether mixtures by polymeric membranes 31
Chapter 3
PERVAPORATIVE MODELLING OF A POLYMERIC MEMBRANE
Overview 36 3.1 Introduction 37 3.2 Theory of mass transport through a membrane 38
3.3 Theory of the Solution-Diffusion Model 40
3.3.1 Model Assumptions 41 3.3.2 Sorption Equilibria 41
3.3.2.1 Interaction parameters 45 3.3.3 Mass transport in pervaporation 47
3.3.3.1 Diffusion coefficients 48 3.3.3.2 Chemical potential 50 3.3 Activation Energy of Pervaporation Transport 52
Chapter 4 EXPERIMENTAL Overview 54 4.1 Membrane used 55 4.2 Chemicals used 56 4.3 Sorption experiments 56
4.3.1 Apparatus and methodology 56
4.4 Pervaporation experiments 58 4.4.1 Apparatus and methodology 58
4.4.2 Experimental equipment 59
4.4.3 Membrane cell 60 4.4.4 Analytical equipment 60 4.5 Experimental planning and design 61
4.5.1 Assumptions during planning 61 4.5.2 Dependent and independent variables 61
4.5.3 Experimental design 61 4.5.3.1 Sorption experiments 62
4.5.3.2 Pervaporation experiments 62 4.6 Reproducibility and experimental error 63
Overview 64 5.1 Sorption characteristics of the membrane 65
5.1.1 Influence of Feed Composition and temperature 65
5.2 Pervaporation characteristics of the membrane 68
5.2.1 Influence of Feed Composition 68 5.2.1.1 Total flux and selectivity 68 5.2.1.2 Partial fluxes of MeOH and TAME 71
5.2.2 Influence of operating temperature 74 5.3 Separation capabilities of pervaporation 78
5.4 Concluding remarks 80 Chapter 6
MODELLING PERVAPORATION PROCESS
Overview 81 6.1 Introduction 82 6.2 Sorption 82
6.2.1 Interaction parameters 83 6.2.2 Modelling of the preferential sorption 85
6.3 Pervaporation 88 6.3.1 Diffusion coefficients 88
6.3.2 Modelling of the partial fluxes 89 6.3.2.1 Concentration-independent 89 6.3.2.2 Concentration-dependent 91 6.3.2.1 Chemical potential 94 6.3 Critical assesment 102 6.3 Concluding remarks 103 Chapter 7
CONCLUSIONS AND RECOMMENDATIONS
Overview 104 7.1 Main Objective 105 7.2 Sorption characteristics 105 7.3 Pervaporation characteristics 105 7.4 Modelling 106 7.5 Recommendations 107 REFERENCES 108
Overview 122 A. 1 Membrane processes 123
A.l.l Microfiltration (MF). 123 A.1.2 Ultrafiltration (UF) 123 A.1.3 Nanofiltration (NF) 124 A. 1.4 Reverse Osmosis (RO) 124
A. 1.5 Gas separation 125 A.1.6 Dialysis 125 A. 1.7 Electrodialysis 126 A. 1.8 Emerging processes 126 A.2 Suppliers of pervaporation equipment 127
Appendix B
SORPTION EXPERIMENTS
Overview 128 B.l Measured results 129
B.2 Sample calculations 130 B.2.1 Sample calculation of swelling ratio (sorption factor) 130
B.2.2 Sample calculation of sorption selectivity 130
B.3 Calculated results 131 B.4 Graphical representation of results 132
Appendix C
PERVAPORATION EXPERIMENTS
Overview 134 C. 1 Measured results 135
C.2 Sample calculations 138 C.2.1 Sample calculation of the fluxes 138
C.2.2 Sample calculation of the selectivity 139 C.2.3 Sample calculation of steady state fluxes and selectivities 139
C.2.4 Sample calculation of the deviation coefficients 140
C.3 Calculated results 141 C.3.1 Total flux 141 C.3.2 Methanol partial flux 144
C.3.3 TAME partial flux 147
C.3.4 Selectivity 150 C.3.5 Fluxes and selectivities at steady state 153
C.3.6 Deviation coefficients 157 C.4 Graphical representation of results 159
C.4.1 Total flux 159 C.3.2 Methanol partial flux 160
C.3.3 TAME partial flux 161
Overview 163 D.l Calibration Curve of MeOH-TAME 164
D.l.l Sample calculation for calibration curve 167 D.2 Determination of composition from calibration curve 167
Appendix E
STATISTICAL INFERENCE OF EXPERIMENTS AND RESULTS
Overview 169 E.l The uncertainties and confidence intervals in measurements 170
E.2 Calculation of GC-measuring error 172 E.3 Calculation of sorption experimental error 173
E.3.1 Swelling ratio 173 E.3.2 Sorption selectivity 173 E.4 Calculation of pervaporation experimental error 174
E.4.1 Pervaporation flux 176
E.4.2 Selectivity 176 Appendix F
MODELLING OF SORPTION-STEP
Overview 177 F. 1 Interaction parameter gi2 178
F.2 Interaction parameters gi3 and g23 181 F.3 Modelling sorption according to the solution-diffusion model 183
F.3.1 Calculated results 183 F.3.2 Accuracy of model 185 F.3.3 Graphical representation 185 Appendix G MODELLING OF PERVAPORATION Overview 187 G.l Diffusion coefficients 188
G. 1.1 Sample calculation of transport parameters 188
G.1.2 Calculated results 189 G.2 Modelling pervaporation according to the solution-diffusion
model 193 G.3 Calculated results 194
G.3.1 Methanol flux (entire concentration region) 194 G.3.2 TAME flux (entire concentration region) 198 G.3.3 Methanol flux (region above azeotropic) 200 G.3.3 TAME flux (region above azeotropic) 202
G.4 Graphical representation 204 G.4.1 Greenlaw et al. 's model for MeOH flux (entire concentration
region) 204 G.4.2 Long's model for MeOH flux (entire concentration
region) 205 G.4.3 Suzuki and Onozato model for MeOH flux (entire
G.4.5 Suzuki and Onozato model for TAME flux (entire
concentration region) 208 G.4.6 Greenlaw et a/.'s model for TAME flux (region above
azeotropic point) 209 G.4.7 Suzuki and Onozato model for TAME-flux (region above
azeotropic point) 210 G.4.8 Yang eta.'s model for MeOH flux 210
Sxinhnl ])(sriipiion I nil Di diffusion coefficient of component i m2/s
DS degree of swelling g/g
G* excess free energy of mixing J/mol.K
gij binary interaction parameter
-GM gibbs free energy of mixing J/mol
HM the heat of mixing J/mol
J Flux kg/m2-hr
M swelling ratio g/g
n; mole fraction of component i
-P permeability m3m/m2s.Pa.m
R universal gas constant J/mol.K
S solubility coefficient mVnr'.Pa
SM total entropy change on mixing J/mol
T temperature K o r ° C
Ui relative volume fraction of component i of a
binary mixture in the polymeric phase
Vi volume fraction in the liquid phase (feed) of
component i
Vi molar volume of component i. m3/mol
W mass of the swollen or unswollen membrane g
X mass fraction of component in feed g/g
a selectivity (reparation factor)
-P
enrichment factor-4> volume fraction in the ternary (membrane)
phase
8 preferential sorption
Mi chemical potential of component I J/mol
n
viscosity N.s/m2n
osmotic pressure Pa% interaction parameter
-Subscripts Description
i orj component for example methanol
MeOH methanol
TAME tert-amyl methyl ether
y mass fraction of component in permeat
X mass fraction of component in feed
oo equilibrium or finale
Oor o initial
1 component 1 (methanol)
2 component 2 (TAME)
CTA Cellulose triacetate
D Diffusivity
ETBE ethyl tertiary butyl ether
GS gas separation
MeOH methanol
MF microfiltration
MTBE methyl tert-butyl ether
NO nanofiltration
PV pervaporation
RO reverse osmosis
TAME tert-amyl methyl ether
UF ultrafiltration
Figure 1.1: Figure 1.2: Figure 2.1: Figure 2.2: Figure 2.3: Figure 2.4: Figure 2.5: Figure 3.1: Figure 4.1: Figure 4.2: Figure 4.3: Figure 4.4: Figure 4.5: Figure 5.1: Figure 5.2: Figure 5.3: Figure 5.4: Figure 5.5: Figure 5.6: Figure 5.7:
The reaction network in the formation and splitting of TAME 3 Schematically representation of the scope of this investigation 7 Representation of a two-phase system separated by a membrane 12 Schematic drawing of the pervaporation process, (a) Vacuum
pervaporation, (b) Inert purge pervaporation. 21 Number of patents on pervaporation between 1983 and 1999. 22
Number of scientific articles on pervaporation. 23 Areas of pervaporation: applications and membranes. 25 Schematic representation of the solution-diffusion model 40
SEM image of the PERVAP2256® membrane. 55 Schematic diagram of the apparatus used to determine the
composition of the liquid inside the membrane. 57
Photo of the pervaporation apparatus. 58 Schematic diagram of the pervaporation apparatus. 59
Photo of the flat sheet membrane cell. 60 Swelling ratio ( ■ ) and sorption selectivity (O) at 25°C. 65
Polymer-liquid mixture equilibrium curve at 25°C ( ♦ ) , 35°C 67 ( ) a n d 4 5 ° C ( 0 ) .
Influence of feed composition on total flux at a constant
temperature of 25°C (♦), 35°C (□) and 45°C (•) 69 Influence of feed composition on selectivity at a constant
temperature of 25°C (♦), 35°C (□) and 45°C (•) 69 Influence of feed composition on partial fluxes at 35°C. 72
MeOH-flux ( ■ ) and TAME -flux (O)
Deviation coefficients of MeOH ( ■ ) and TAME (O) at 35°C. 73 Influence of operating temperature on total flux at constant feed
compositions (Feed mass fraction methanol: • 15%, ■ 40% and
Figure 5.9: Figure 5.10: Figure 5.11: Figure 6.1: Figure 6.2: Figure 6.3: Figure 6.4: Figure 6.5: Figure 6.6: Figure 6.7: Figure 6.8: 77 79 84 86 40%, and + 90%)
Arrhenius plot of methanol flux vs. reciprocal temperature.
(Feed mass fraction methanol: ♦ 15%, ■ 30%, A 40%, x 50%, * 76 60%, • 75% and + 90%)
Activation energy for MeOH ( ♦ ) and TAME ( □ ) flux (Ej) vs. feed methanol concentration.
Separation diagram for methanol at 35°C where (—) represents VLE, (•) pervaporation and (□) sorption.
Fourth-order polynomial relation between gn and v2 at 35°C
Comparison of experimental values (•) for the methanol sorption and calculated values using both concentration independent (—) and concentration dependent (—) interaction parameters at 25 °C.
Comparison of experimental values (•) for the methanol sorption and calculated values using both concentration independent (—) and concentration dependent (—) interaction parameters at 45 °C
Comparison of experimental MeOH-flux (•) with the
concentration-independent diffusion coefficients (—) at 25°C. Comparison of experimental TAME-flux (•) with the
concentration-independent diffusion coefficients (—) at 25°C. Comparison of experimental MeOH-flux at 25°C with the models. (♦) Experimental flux and (—) Predicted flux Comparison of experimental TAME-flux at 25°C with the models. (♦) Experimental flux, (—) Greenlaw and (—) Suzuki. Comparison of experimental MeOH flux at 30°C for the region above the azeotropic point with the models. (♦) Experimental flux and (—) Predicted flux
86 89 90 91 91 92
above the azeotropic point with the models. (♦) Experimental
flux and (—) Predicted flux 93 Figure 6.10: Comparison of experimental MeOH flux with Yang et al.'s
model. (♦) Experimental flux and (—) Predicted flux. 96 Figure 6.11: Comparison of experimental TAME flux with Yang et al.' s
model. (♦) Experimental flux and (—) Predicted flux. 97 Figure B. 1: Swelling ratio ( ■ ) and sorption selectivities (O) at 25°C 132 Figure B.2: Swelling ratio ( ■ ) and sorption selectivities (O) at 35°C 132 Figure B.3: Swelling ratio ( ■ ) and sorption selectivities (O) at 45°C 133 Figure B.4: Polymer-liquid mixture equilibrium curve at
25°C ( ♦ ) , 35°C ( ) a n d 4 5 ° C ( 0 ) 133 Figure C. 1: Influence of feed composition on total flux at different
temperatures 159 Figure C.2: Influence of feed composition on methanol flux at different
temperatures 160 Figure C.3: Influence of feed composition on TAME flux at different
temperatures 161 Figure C.4: Influence of feed composition on selectivity at different
temperatures 162 Figure D. 1: Calibration curve for MeOH-TAME mixtures (A = methanol, B
= TAME) 166 Figure D.2: Calibration curve for MeOH-TAME mixtures (A = MeOH, B =
TAME) 166 Figure D.3: Areas of peaks for MeOH and TAME from GC 167
Figure E. 1: Reproducibility curve of flux experiments 175 Figure E.2: Reproducibility curve of selectivity experiments 175
Figure F.l: Curves of gi2(v2) vs. V2 179 Figure F.2: Curves of gni^i) vs. U2 180 Figure F.3: Computation scheme for calculation of g13 and g23 182
Figure F.5: Figure F.6: Figure G.l: Figure G.2: Figure G.3: Figure G.4: Figure G.5: Figure G.6: Figure G.7: Figure G.8: Figure G.9: Figure G.IO:
and concentration dependent (—) interaction parameters at 25
°C. 1 8 5
Comparison of experimental values (•) for the methanol sorption and calculated values using both concentration independent (—) and concentration dependent (—) interaction parameters at 25
°C. 1 8 6
Comparison of experimental values (•) for the methanol sorption and calculated values using both concentration independent (—) and concentration dependent (—) interaction parameters at 25
°C. 1 8 6
Comparison between calculated and experimental values. 189 Comparison of experimental MeOH flux with Greenlaws et al.'s
model. 204 Comparison of experimental MeOH flux with Long's model 205
Comparison of experimental MeOH flux with Suzuki and
Onozato model 206 Comparison of experimental TAME flux with Greenlaws et ai's
model. 207 Comparison of experimental TAME flux with Suzuki and
Onozato model 208 Comparison of experimental TAME flux with Greenlaws et al.'s
model. 208 Comparison of experimental TAME flux with Suzuki and
Onozato model 209 Comparison of experimental MeOH flux with Yang et al.'s
model 211 Comparison of experimental TAME flux with Yang et al.'s
Table 1.1: Table 2.1: Table 2.2: Table 2.3: Table 4.1: Table 4.2: Table 4.3: Table 5.1: Table 5.2: Table 6.1: Table 6.2: Table 6.3: Table 6.4: Table 6.5: Table 6.6: Table 6.7: Table 6.8: Table 6.9: Table 6.10 Table 6.11 Table 6.12
Results of studies on the separation of alcohol/tertiary ether 5 mixtures by polymeric pervaporation membranes
Some advantages and disadvantages of polymeric and zeolite 18 membranes
Different membrane processes 19 Application of polymeric membranes to the separation of
alcohol from alcohol/ether mixtures 29 Specification sheet for the PERVAP2256® polymeric membrane 56
Specifications of experimental equipment 59
Experimental planning 63 Interaction parameters of methanol and TAME 67
Activation energy of pervaporation for MeOH and tertiary ethers
through different membranes. 78 Concentration independent interaction parameters 83
Concentration-dependent interaction parameters (gn and g23) 83
Coefficients of the fourth-order polynomial of gi2(v2) 84 Coefficients of the fourth-order polynomial of gi2(u2) 85 Standard deviation and R values for the sorption models 87 Expressions for the diffusion coefficients of MeOH and TAME 88
The accuracy (R -values) of the partial fluxes 90 Accuracy of partial MeOH fluxes for the region above the
azeotropic point (Revalues) with the different models 93 Accuracy of partial TAME fluxes for the region above the
azeotropic point (R -values) with the different models 94 Accuracy of partial MeOH and TAME fluxes with Yang et al.'s
approach 97 The limiting diffusion coefficients and plasticization coefficient
for MeOH and TAME 98 Limiting diffusion coefficients and plasticization coefficient for
Table C.l: Measured pervaporation experimental data (mass permeated)
over time at 25°C 135 Table C.2: Measured pervaporation experimental data (mass permeated)
over time at 30°C. 135 Table C.3: Measured pervaporation experimental data (mass permeated)
over time at 35°C. 136 Table C.4: Measured pervaporation experimental data (mass permeated)
over time at 40°C 137 Table C.5: Measured pervaporation experimental data (mass permeated)
over time at 45 °C 137 Table C.6: Pervaporation sampling data at 35°C and 50 wt.% MeOH 138
Table C.7: Calculated total flux at 25°C. 141 Table C.8: Calculated total flux at 30°C. 141 Table C.9: Calculated total flux at 35°C. 142 Table C.10: Calculated total flux at 40°C. 143 Table C.l 1: Calculated total flux at 45°C. 143 Table C.12: Calculated methanol flux at 25°C. 144 Table C.13: Calculated methanol flux at 30°C. 144 Table C.14: Calculated methanol flux at 35°C. 145 Table C. 15: Calculated methanol flux at 40°C. 146 Table C. 16: Calculated methanol flux at 45°C. 146 Table C. 17: Calculated TAME flux at 25°C. 147 Table C.l8: Calculated TAME flux at 30°C. 147 Table C. 19: Calculated TAME flux at 35°C. 148 Table C.20: Calculated TAME flux at 40°C. 149 Table C.21: Calculated TAME flux at 45°C. 149 Table C.22: Calculated Selectivity at 25 °C. 150 Table C.23: Calculated Selectivity at 30°C. 150 Table C.24: Calculated Selectivity at 35°C. 151 Table C.25: Calculated Selectivity at 40°C. 151 Table C.26: Calculated Selectivity at 45°C. 152
Table C.28: Table C.29: Table C.30: Table C.31: Table C.32: Table D.l: Table D.2: Table E.l: Table E.2: Table E.3: Table E.4: Table E.5: Table F.l: Table F.2: Table F.3: Table F.4: Table F.5: Table F.6: Table F.7: TableG.l: Table G.2: Table G.3: Table G.4: Table G.5: Table G.6:
Calculated methanol flux at steady state. Calculated TAME flux at steady state. Calculated selectivities at steady state. Calculated methanol Deviation coefficient. Calculated TAME Deviation coefficient.
154 155 156 157 158 Calculated and measured values for MeOH-TAME compositions 164 Area ratios for MeOH-TAME mixtures from
gaschromatographic analysis 165 Z's for use in two-sided large-n intervals for |X 171
GC-experiments for calculating error 172 Sorption experimental reproducibility data 173 Pervaporation experimental reproducibility data 174
Steady-state experimental results. 176
Coefficients of gi2 180 Calculated values of Xi3 and X23 181
Calculated constant of gi3 and g23 182 Calculated results for sorption-step in solution-diffusion model
at25°C 183 Calculated results for sorption-step in solution-diffusion model
at35°C 184 Calculated results for sorption-step in solution-diffusion model
at45°C 184 Standard deviation and R values for the sorption models 185
Experimental sorption fraction and MeOH-flux at 25°C 188 Calculated values of the transport parameters for MeOH at 25°C. 189
The concentration-independent diffusion coefficients 189 The limiting diffusion coefficients (Di°) and plasticization
coefficients (p\j and pa) for MeOH and TAME 190
Different models used for predicting the values of the fluxes 193 The experimental MeOH fluxes and predicted values for the
Table G.8: Experimental MeOH fluxes and predicted values for the
different models 200 Table G.9: Experimental TAME fluxes and predicted values for the
CHAPTER 1
GENERAL INTRODUCTION
"The Beginning is the most important part of the worl^ "
Plato, The Republic
'"Why, a four-year-old child could understand this. Someone get me a jour-year-old child. "
Groucho Marx
'7 have yet to see any problem, however complicated, that looked
at in the right way, did not become stillmore complicated."
Anderson's Law:
Overview
This Chapter is subdivided into three sections, starting in section 1.1 (background and motivation) with a general background about the legislation concerning the phase-out of leaded petrol that led to the motivation of this project. The main objectives of the project are given in section 1.2 (objective) and the outline of the dissertation with the scope of investigation is given in section 1.3 (Scope of investigation).
1.1 Background and motivation
For many years tetraethyl lead or tetramethyl lead was added to petrol to increase the octane rating and improve fuel efficiency. In the USA it was found, however, that the extremely high levels of airborne lead in urban areas were directly traceable to the combustion of leaded petrol (Corbett, 1991). In 1985 the U.S Environmental Protection Agency ruled that petrol lead content be reduced by over 90 percent, and by the early
1990s unleaded petrol had become the standard throughout the United States.
Unleaded petrol is usually introduced, for example, in the European Union, USA, Canada and Japan for environmental reasons. In South Africa, however, unleaded petrol is primarily introduced to ensure technological compatibility and therefore more favourable economies of scale (Botha, 1996).
The South African government has set 2006 as target date for a complete phase-out of leaded petrol in South Africa. The lead content of petrol sold in SA was reduced from 0.84 gram per litre in 1983 to 0.4 gram per litre in 1989 and early in 1996 unleaded petrol was introduced (Botha, 1996). The result has been a dramatic drop in airborne lead emissions and has fostered great interest in the use of alternative octane enhancers. Various oxygenated compounds (those containing oxygen) like alcohols and ethers have proved to be suitable for octane enhancement and carbon monoxide (CO) emission reduction (Oost & Hoffmann: 1995), but some oxygenates possess better blending characteristics than others (Brockwell etal., 1991).
In general, ethers like methyl tert-buty\ ether (MTBE), tert-butyl ethyl ether (ETBE), tert-amyl methyl ether (TAME) and tert-amyl ethyl ether (TAEE) are preferred over alcohols because of their low blending vapour pressure (Jensen & Datta, 1995).
The predominant oxygenate used, is MTBE. This is due to the relatively lower cost of the methanol (MeOH) used and the ready availability of isobutylene from catalytic cracking and steam cracking fractions. TAME is also coming into vogue due to the pressure to reduce the olefin content in the petrol. Using both TAME and MTBE can give refiners increased blending flexibility for volatility control (Koskinen etal, 1996).
SASOL, on the other hand, uses TAME as a petrol additive instead of MTBE because TAME has better blending properties for the fuel produced by SASOL and 2-methyl-2-butene (2M2B) is a side-product from one of SASOL's product streams (Botha, 1996).
Commercially, TAME is typically produced by the reaction of methanol with isoamylenes (2-methyl-l-butene and 2-methyl-2-butene) over an acidic catalyst. The reaction proceeds through protonation to give tertiary butyl cation as a common intermediate, followed by the nucleophilic addition of methanol to produce TAME
(Oost etal., 1995): CH, C H3— C H2— C = C H2 2MB1 + C H3O H MeOH
A
.CH. C H3— C H = C 2MB2 \ CH. + C H3O H MeOH CH O I , — CH,— C — O — CH,I
CH, TAMEFigure 1.1 - The reaction network in the formation and splitting of TAME.
This reaction is limited by the reaction equilibrium and therefore an excess of methanol is required to reach a high conversion of isoamylenes relative to the ether. The net product stream of the reactor consists of unconverted methanol, isoamylenes, TAME and the inert of other butenes. The butenes are separated by distillation. The excessive methanol must be removed from the final product, but methanol forms an azeotrope with TAME at a composition of 40 wt% methanol (Oost et al, 1995). In the industrial processes, the reaction mixtures are washed with a large amount of water and distilled to methanol separation (Yang et ah, 1994). This conventional process is very cost- and energy-intensive.
An increasingly popular route for producing tertiary ethers is reactive distillation with a solid catalyst contained within a distillation column. In this manner both chemical reaction and fractionation of products can proceed simultaneously. This combination offers potential advantages over the conventional process mentioned above, including the following (Nijhuis et al, 1991):
• Use of the heat of reaction for product separation.
• A relative easily controllable temperature profile in the catalytic section. • Lower operating costs due to higher obtainable conversions.
• Lower capital cost due to less equipment.
Although reactive distillation is less expensive than the conventional process, it is still an expensive operation since the entire feed stream needs to be heated to the reaction temperature and the liquid still needs to be heated to the mixture boiling point for separation. An alternative separation method with low energy and cost consumption is therefore needed. One possible method is the membrane separation process, pervaporation. Pervaporation is attractive for removing minor components from liquid mixtures and for purification or recovery purposes (Feng & Huang, 1997). The process is even more attractive for separating azeotropic mixtures, since the separation is not based on the relative volatility of the components in the mixture, but only depends on the relative affinity of one of the components for the membrane (Mulder, 1996).
A wide selection of papers on the separation of ether mixtures with polymeric membranes has been published recently as shown in Table 1.1 (Cao et al, 1999; Hung
et al, 1998; Luo et al., 1997). A summary of the most important results from these
studies is shown in Table 1.1. It is evident from these studies that polar hydrophilic polymeric materials led to the methanol and ethanol permselective membranes. It is well kown that the cross-linked poly-vinyl alcohol (PVA) membranes have been used for producing absolute alcohol on an industrial scale. One of the most suitable PVA membranes for this purpose is the commercially available, dense PERVAP2256® membrane from Sulzer Chemtech (Germany). The PERVAP2256® membrane had been previously used for the separation of methanol-MTBE (Gonzalez et al, 2001) and ethanol-ETBE (Ortiz et al, 2002) that showed high alcohol fluxes as well as high permeate selectivity.
Table 1.1- Results of studies on the separation of alcohol/tertiary ether mixtures by polymeric pervaporation membranes
Maximum MciiihriiiiL'
Mi'iiilmiui' Mux Suk'Clml} Conditions TIlk'kHL'SS Ki'l'i'i'i'iicc
(jj.in'Mir"1) (11111)
MeOH from TAME Polyion complex (PIC) hollow fiber prepared
from polyacrylonitrile 1000 50
20 wt% methanol in feed, 30°C
0.53 mbar 150 Hung etal., 1998
MeOH from MTBE Dense Cellulose
triacetate (CTA) 550 1150 38 wt% methanol in feed, 50°C,
2 mbar 25 Cao etal., 1999
Poly(phenylene oxide) 500 7.7 21 wt% methanol in feed, 22°C,
1.3 mbar 40 Doghieri et al., 1994
Polymer blend of poly (acrylic acid) and
poly(vinyl alcohol) 600 50
20 wt% methanol in feed, 50°C
2.7 mbar 20 Park etal., 1995
Cellulose acetate (CA)
and CTA 450 200 30 wt% methanol in feed, 40°C
Downstream pressure not reported 20 Yang et al, 1998 EtOH from ETBE
Polymer blend of Cellulose acetate butyrate (CAB) and
Cellulose Acetate 3000 34
30 wt% ethanol in feed, 60°C,
1 mbar 50 Luo *?/«/., 1997
1.2 Objectives
The objective of this study was to investigate the performance of a polymeric membrane for the separation of methanol from tertiary amyl methyl ether by means of pervaporation, based on specific criteria such as magnitude of flux and selectivity. The commercial membrane was further investigated to obtain information about the phenomena govering the separation. For this purpose an extensive experimental programme was completed and a suitable mathematical model developed.
1.3 Scope of Investigation
The basic scope of this investigation is summarised in Figure 1.2. The dissertation is subdivided into seven chapters that consist of the following, in order to achieve the above-mentioned objectives:
(i.) An overview of the terminology, theory and current literature on the subject of pervaporation technology is evaluated in Chapter 2 (Background and Literature Survey). The focus here was to acquire knowledge on the pervaporation process and to review previous work that had been done on the separation of alcohol and ether mixtures by polymeric membranes by pervaporation.
(ii.) An overview of the theory and principle approach to describe mass transport through a membrane using the simplified solution-diffusion model is discussed in Chapter 3 (Pervaporative modelling of a polymeric membrane). The focus here was to acquire knowledge on a simplified model that could be used to describe the pervaporation separation of MeOH and TAME, by using both the sorption and pervaporation data obtained with the PERVAP2256® membrane.
Scope of Investigation
Sorption characteristics
Here the focus was on the swelling properties of the PERVAP2256® membrane and the
sorption of methanol in the membrane. The effect of both the feed composition and
temperature was investigated
Influence of feed composition
The solution composition was varied between 0 and 100 wt. % MeOH and the respons of the
total amount absorped in the membrane and the MeOH sorption was observed.
Influence of temperature
The solution temperature was varied between 25°C and 45°C with a resolution of three and the respons of the total amount absorped in the membrane and MeOH sorption was observed.
Pervaporation characteristics
Here the focus was on the separation of methanol from the binary mixtures of MeOH and TAME.
The effect of both the feed composition and temperature was investigated.
Influence of feed composition
The feed composition was varied between 0 and 100 wt. % MeOH with a resolution of seven and the respons of the total flux
and selectivity was observed
Influence of temperature
The feed temperature was varied between 25°C and 45°C with a resolution of five
and the respons of of the total flux and selectivity was observed
Modelling the process
Here the focus was to examine whether a simplified model, based on the solution-diffusion
mechanism, can be developed by using both the sorption and pervaporation data obtained with the
PERVAP2256® membrane.
(iii.) The experimental apparatus and methodology that were used in this investigation are discussed in Chapter 4 (Experimental). The two aspects that are addressed in this dissertation are, (i) the sorption characteristics of the membrane and (ii) the pervaporation characteristics of the membrane.
Sorption characteristics:
The focus was here to acquire knowledge of the influence of the solution composition and operation temperature on the swelling properties of the membrane and the preferential sorption capabilities of the membrane. From these experiments, relevant parameters for solving the solution-diffusion model could be obtained.
Pervaporation characteristics:
The focus was here on the selective removal of methanol from the mixtures of MeOH and TAME by means of pervaporation with the PERVAP2256® membrane.
The effect of both the concentration of the feed mixture and the temperature on the pervaporation process of the binary mixtures was investigated. A range of 25°C to 45°C for the feed temperatures with approximately 5°C intervals was chosen, while a range between 0 and 100 wt.% MeOH for the feed composition was chosen for all the experiments.
(iv.) The results obtained for both the sorption and pervaporation experiments are discussed in Chapter 5 (Experimental results and discussion). The main focus here was to verify, from the experimental results, that the pervaporation process can be used to separate an azeotropic mixture and that the PERVAP2256® membrane is MeOH selective.
(v.) The modelling procedure followed to model the sorption and pervaporation data is discussed in Chapter 6 (Modelling). The main focus here was to describe the simplified model that was developed for both the sorption and pervaporation data obtained with the PERVAP2256® membrane. The modelling results are also compared with the experimental results.
(vi.) Finally, in Chapter 7 (Conclusions and recommendation) a detailed discussion of the main conclusions that can be drawn from this investigation is given with some recommendation for further study.
CHAPTER 2
BACKGROUND AND
LITERATURE SURVEY
"In literature quotation is good only when the writer whom I follow goes my way, and, being better mounted, than I,gives me a cast."
Emerson
"Quotations are best brought in to confirm some opinion controverted."
Swift
Overview
The appearance of a variety of membrane processes to treat liquid mixtures has given rise to an extensive, if not altogether consistent, terminology and application. It is therefore the aim of this Chapter to introduce an overview of the terms, concepts and applications used in membrane technology, particularly focusing on organic-organic separation by the pervaporation process with polymeric membranes.
The Chapter is subdivided into six sections, in order to achieve the above-mentioned aim, starting with a general overview of the terms and concepts (section 2.2) used in membrane technology. The different types of membranes (section 2.3) and separation processes (section 2.4) that can be found in membrane technology are discussed in sections 2.3 and 2.4. The main focus of this survey was on the pervaporation process (section 2.5) that includes a description of the process and process variables (section 2.6), characteristics of the process, as well as the membranes used for pervaporation. Finally a brief overview of previous work done on the separation of alcohol/ether mixtures by polymeric membranes is discussed in section 2.7.
2.1 Introduction
During the past decade, industrial membranes have established themselves as indispensable components in chemical processing industries. Membrane-based technology is currently regarded as a new frontier of chemical engineering and has been widely used for purification, concentration and fractionation of fluid mixtures. (Feng & Huang, 1997)
Pervaporation as a membrane process is an attractive separation technique that has been the object of numerous experimental and theoretical investigations (see Table 1.1) Pervaporation can be considered a basic unit operation with significant potential for the solution of various environmental and energy processes. The basis of this process is that a liquid (mixture) is in contact with a membrane. At the permeate side a partial pressure difference is generated by means of a vacuum pump, or by means of an inert gas flow. The components of the liquid that move through the membrane are vaporised by the low pressure, removed and condensed.
The process is even more attractive for separating azeotropic mixtures, since the separation is not based on the relative volatility of the components in the mixture, but depends essentially on the relative affinity of one of the components for the membrane. Recommended review articles on the pervaporation process are Jonquieres et al. (2002), Feng et al. (1997), Dutta et al. (1997), Mulder (1996), and Zhang et al. (1995).
2.2 General Membrane terminology
2.2.1 Definition of membrane
The membrane is at the heart of every membrane process and can be defined as a permselective barrier or interphase between two phases (Mulder, 1996). Figure 2.1 gives a schematic representation of a membrane separation process.
Figure 2.1 - Representation of a two-phase system separated by a membrane.
Phase 1 is usually considered as the feed or upstream side while phase 2 is considered the permeate or downstream side. Separation is achieved because the membrane, under a certain driving force, has the ability to transport one component from the feed mixture more readily than any other component or components. This may occur through various mechanisms. The extent of this force is determined by the potential gradient, or approximately by the difference in potential, across the membrane divided by the membrane thickness.
The main driving force can be the partial pressure difference, the chemical potential difference or the concentration difference. In fact, all three could contribute to the actual driving force for transport. Other possible forces, such as magnetical fields, centrifugal fields and gravity will not be considered.
The barrier is most often a thin, nonporous polymeric film, but may also be a porous polymer, ceramic, or metal materials, or even a liquid or gas.
2.2.2 Membrane effectiveness parameters
Membranes are rated by the rate at which they produce permeate, and the ability to discriminate between compounds they retain and compounds they pass. Therefore, the performance or effectiveness of a membrane in separating a liquid mixture is characterised by two parameters - flux and selectivity (separation factor) (Dutta et al.,
1997). While flux or permeation rate is expressed as the amount of permeate collected per unit time per unit membrane area (kg/m -hr or kmol/m -hr), selectivity of the permeation process for a particular solute is expressed in terms of either the selectivity (a), or the enrichment factor (P).
The selectivity (a) is patterned after the relative volatility of the components of binary liquid mixtures as (Dutta et al, 1997):
a =
* / 0 z * )
(2.i)
*,./(!-*,.)
with i the preferentially permeating species and y and x the mass fraction of each component in the permeate and the feed, respectively. As a ratio of ratios, the selectivity is independent of the concentration units used.
The enrichment factor (P) is simply the ratio of concentrations of the preferentially pervaporating species in permeate and feed (Dutta et al., 1997):
P = ^ - (2-2)
and is related to solute rejection in reverse osmosis (P=l-R). The relation between the selectivity and enrichment factor is readily obtained as:
aJ-^^ (2-3)
The parameter a is more commonly used because it fits the general definition of the separation factor.
Membranes are available in several different configurations - tubular, hollow-fiber, plate-and-frame, and spiral-wound. Some of these designs may work better than others for a particular application, depending on such factors as viscosity, concentration of suspended solids, particle size, and temperature.
2.2.3 Fouling
Fouling is one of the most frequently encountered problems with membranes, especially with membrane processes involving liquid feeds. Fouling can be seen as the coating of the membrane surface or blocking of the surface with a solid or gelatinous material, which creates a barrier through which the permeating species must pass (Humphrey et al., 1997). The net effect of this blockage is a decrease in the flux passing through the membrane. The blockage may also take the form of a second but nonselective resistance and thus decrease the overall selectivity of the membrane.
2.2.4 Concentration polarization
Concentration polarization is similar to fouling in that a layer is formed near the surface of the membrane, which is different in concentration from the bulk fluid. The restraining layer consists of a build-up in the concentrations of nonpermeating or slowly permeating components in the feed as the more permeable components pass through the membrane (Humphrey et al., 1997).
The net effect of concentration polarization is a decrease in the permeation rate of the more rapidly permeating components and the perrmeation rates, on the other hand, for the components that permeate relatively slowly can increase. The major disadvantage of concentration polarization is a decrease in membrane selectivity.
2.2.5 Swelling of the membrane
Swelling of the membrane occurs when polymeric membranes are used. The liquid, in contact with the membrane, dissolves into it and causes membrane swelling. Swelling tends to alter the membrane properties and generally leads to higher permeability and lower selectivity. The swelling state of the membrane is the unique feature of the pervaporation process, ranging from fully swollen at the liquid feed side to virtually dry at the permeate side which is under constant vacuum. There is thus a swelling gradient within the thickness of the membrane (Boddeker, 1990).
The swelling ratio (M„) and degree of swelling (DS) are used interchangeably to describe the swelling of the membrane. The degree of swelling is defined as the amount of equilibration solution sorbed in the membrane (WTO - Wo) compared to the mass of
the unswollen membrane (W0) (Richau et al., 1996; Nam et al, 1999), i.e. the degree of
swelling is the mass % uptake at equilibrium:
DS = W^-Wp xlOO (2-4)
The swelling ratio ( M „ ) relates the amount of equilibration solution sorbed (WTO - Wo)
to the total mass of the ternary system (membrane and equilibration solution) (WTO)
(Yang et al., 1998):
M = \ +<p2 (2-5)
2.3 Types of membranes
The choice of membrane materials is dictated by the application, environments, the separation mechanisms by which they operate and economic considerations. Finding a suitable membrane is the most important hurdle to pass when devising any membrane system. A large number of membranes have been developed and their efficiency demonstrated, at least in the laboratory, in separating liquid mixtures. These include (i) polymeric membranes; (ii) organic/inorganic composite membranes, e.g., zeolite-filled polymeric membranes; (iii) inorganic membranes, e.g., zeolite membranes; and (iv) liquid membranes.
2.3.1 Polymeric membranes
The industry has so far been dominated by polymeric membranes which, after many years of research and development and marketing by many companies, start to enjoy applications ranging from desalination of sea and brackish waters, food and beverage processing, gas separations, hemodialysis to controlled release.
The first widespread use of polymeric membranes for separation applications dates back to the 1960-70s when cellulose acetate was cast for desalination of sea and brackish water. Since then many new polymeric membranes came to the market for applications extended to ultrafiltration, microfiltration, dialysis, gas separations and pervaporation. So far, ultrafiltration has been used in more diverse applications than any other membrane process.
For pervaporation and gas separation, dense membranes (usually polymeric membranes) are required, preferably with an anisotropic morphology, an asymmetric structure with a dense top layer and an open porous sublayer, as found in asymmetric and composite membranes (Mulder, 1998). The requirements for the substructure are in fact the same as for gas separation membranes i.e.:
• An open substructure to minimize resistance to vapor transport and to avoid capillary condensation.
A number of investigations have been reported which were mainly applied to the pervaporation of mixtures of aliphatic alcohols with hydrocarbons. From these investigations it was observed that generally polar molecules permeated faster through polar membranes than non-polar molecules, and vice versa (Park et ah, 1995). Polar polymers such as polyvinylalchol, polyacrylicacid, Nafion, and cellulose acetate were found to be selectively permeable for alcohols over hydrocarbons.
2.3.2 Ceramic membranes (zeolites)
Zeolites are three dimensional, microporous, crystalline solids with well-defined structures that contain aluminium, silicon, and oxygen in their regular framework. The silicon and aluminium atoms are tetrahedrally connected to each other through shared oxygen atoms. Thus the framework of every zeolite is constructed from tetrahedral building blocks, TO4, where T is a tetrahedrally coordinated. (Tavolaro et ah, 1999)
Zeolites can be classified according to their framework symmetry with an identification code of three letters used by the International Zeolite Association (IZA). Therefore, MFI designates the MFI-type that includes synthetic species with different chemical composition (silicalite and ZSM-5).
When the zeolite is composed solely of Si4+02"4 tetraheda, the framework is neutral,
since an oxygen atom bridges two T atoms. In that case the zeolite is hydrophobic. When aluminium is incorporated, the framework becomes negatively charged since the valency of aluminium is +3. This negative charge is then compensated by a positive ion, e.g. Na+, K+ or Ca + or a proton, H+. The presence of aluminium in the zeolite
framework has several effects: the zeolite becomes hydrophilic and acidic, it has ion-exchange capacity and the presence of the counter-ion might obstruct the pores thus reducing the pore size of the zeolite. The presence of the counter-ion can render the zeolite catalytically active. Decreasing the Si/Al ratio increases the hydrophilicity of the zeolite and the number of cations that are needed to balance the charge. (Van de Graaf,
2.3.3 Advantages and disadvantages
Zeolite membranes have considerable advantages over other types of membranes like polymer membranes, in that they are highly stable under thermal cycling, high temperatures, and harsh physical and chemical environments which other membranes cannot withstand. The chemistry of the zeolites can be modified to provide catalytic properties, to change them between hydrophobic and hydrophilic surfaces, to change the pore size and structure (creating different types of zeolites with potential molecular
sieving action), etc., which make them useful for many different applications. A summary of the advantages and disadvantages of polymeric and zeolite membranes are given in Table 2.1
Table 2.1. Some advantages and disadvantages of polymeric and zeolite membranes.
Polymeric membranes Zei)lile membrane
Advantages Advantages
Low capital cost Long-term stability at high temperatures Well-settled market Resistance to harsh environments Achieve high selectivities Resistance to high pressure drops
Good fluxes Easy catalytic activation
Disadvantages Easy cleanability after fouling Unstable at high temperatures Molecular sieving
Unstable to harsh environments Disadvantages Difficult to clean after fouling High capital cost
Low membrane surface per module volume Difficulty in achieving high selectivities in large scale microporous membranes
2.4 Membrane separation processes
The membrane processes discussed in this dissertation are classified according to the driving force used in the process. The various membrane processes along with the driving force and environmental applications are described in more detail in Appendix A and listed in Table 2.2.
All definitions, applications and summaries of Table 2.1 were obtained from the following references: Lipnizki et al. (1999), Mulder (1998) and Perry & Green (1996).
Table 2.2 - Different membrane processes. MiMiil>r;inr prmvsMs I.MimpU' <>t';i|)|>lif;i1i<ms Pressure driven
Microfiltration (MF) Separation of bacteria and cells from solutions; removal of colloids from waste streams;
removal of dust particles from air Ultrafiltration (UF) Separation of proteins and viruses,
concentration of oil-in-water emulsions Nanofiltration (NF) Separation of dye and sugar, water softening Reverse osmosis (OR) Desalination of sea and brackish water, process
water purification
Driven by activity gradient (solution diffusion mechanism applies)
Gas separation (GS) Hydrogen recovery from process gas streams, dehydration and separation of air; removal of CO2 and H2S from landfill gas
Vapour permeation (VP) removal of condensable solvents from air Pervaporation (PV) Dehydration of ethanol and organic solvents Other driving forces
Electro-dialysis Removal of metals from wastewater; acid (electrical potential) recycle from "pickling" baths; desalination of
brackish water
Liquid membranes Recovery of plating chemicals removal of acid (gradient in chemical gases from air
potential based on solubility)
Membrane distillation Water purification and Desalination of brine (temperature difference
2.5 Pervaporation
2.5.1 Background
Though pervaporation is one of the most popular areas of current membrane research, the concept of pervaporation separation is not new. The phenomenon of pervaporation was first reported by Kober (Kober, 1917), who coined the term "pervaporation" by combining the words "permeation" and "evaporation" in a publication. He reported the selective permeation of water from aqueous solutions of albumin and toluene through cellulose nitrate films.
The usefulness, however, of pervaporation for separation and concentration was recognised in 1935 by Faber (Feng et al., 1997). Heisler (Heisler et al, 1956), on the other hand published the first known quantitative work on pervaporation for the separation of water-ethanol mixtures using a cellulose membrane. It was the work of Binning (Binning et al., 1961) who established the principles and highlighted the potential of pervaporation technology.
Recent advances in pervaporation technology have made it one of the latest commercially developed membrane processes. In 1982 the first pilot plant was commissioned in Brazil. The plant removed water from ethanol. The first commercial scale plant started operating in 1988 in Betheniville, France. Once again the application was the dehydration of ethanol (Neel, 1991).
2.5.2 Process Description
Pervaporation is a relatively new membrane separation process in which a liquid mixture (feed) is in contact with a permselective membrane (Feng & Huang, 1997). One component is transported through the membrane preferentially. It evaporates on the permeate side of the membrane leaving as a low-pressure vapour. The permeate vapor can be condensed and collected or released as desired.
The chemical potential gradient (difference of partial pressure or activity) across the membrane is the driving force for the mass transport (Mulder, 1996). The driving force
can be created by applying either a vacuum pump or an inert purge (normally N2, He or Ar) on the permeate side to maintain the permeate vapor pressure lower than the partial pressure of the feed liquid. A schematic drawing of this process is shown in Figure 2.2.
feed relentate feet 1 1 etentate
. « - . » . . . - . . _ relentate im» I. — Wm . « - . » . . . - . . _
1
i H . , . / * ^ ^ . I. — conden! >er1
i H . , . / * ^ condenser " &1
i H . , . / * ^ ^m . ,i,x j — vacuum pump 00i
CH*^
permeate (»> vacuum pump carrier gas 00 per neateFigure 2.2 - Schematic drawing of the pervaporation process. (a) Vacuum pervaporation, (b) Inert purge pervaporation.
Vacuum pervaporation, which is customarily referred to as the standard pervaporation, is the most widely-utilized mode of operation, while inert purge pervaporation (also known as carrier gas pervaporation) is normally of interest if the permeate can be discharged without condensation. An advantage of the inert purge pervaporation is that no vacuum equipment is necessary, which lowers the cost substantially (Feng & Huang,
1997).
2.5.3 Industrialisation and future industrial development
An indication of the industrialisation or future industrial development of a given technology can be see by the number of patents issued. The number of European and USA patents issued for pervaporation during 1983 and 1999, according to Jonquieres et
50 i O 25 V n | 20--15 10
5--H
jz\ f~\
B 83 84 85 86 87 88 89 90 91 92 93 94 95 Year|BEuropean patent applications BEuropean patents PUS patents]
Figure 2.3 - Number of patents on pervaporation between 1983 and 1999.
It can be seen from Figure 2.3 that a few European patent applications were issued yearly with a maximum of eight demands in 1994 for a corresponding total of 57. The number of European patents issued during the same period of time gradually increases up to a maximum of 6 in 1995 and then decreases to reach a mean number of 2, corresponding to the level of 1989, for a total number of 37. The number of USA patents on the other hand issued between 1983 and 1999 were 263 which is seven times more then the European patents issued.
Concurrent with very promising beginnings in Europe, one notes also a great interest in pervaporation in the USA, characterised by a very sharp increase in the number of USA patents issued between 1985 and 1992 as shown in Figure 2.3. However, during the last 7 years, there has been an important decrease in the number of USA patents on pervaporation with a comeback to a level similar to that in 1989 just before the sharp peak of the years 1990-1992. Subsequent to very promising beginnings in the 1980s, pervaporation now appears to be a technology in a dormant stage.
The interest in research on pervaporation as a separation technique on the other hand can also be seen from Figure 2.4 (Lipnizki et al., 1999). Figure 2.4 shows the numbers of papers published on pervaporation between 19981 and April 1998 obtained from Bath Information Data Service (BIDS). It is also forecast that the market share of pervaporation will grow at rates even above the market rate for membrane processes.
Figure 2.4 - Number of scientific articles on pervaporation.
2.5.4 Suppliers of pervaporation equipment
Industrial suppliers of pervaporation equipment are extremely scarce worldwide. The European market is clearly dominated by the German company Sulzer Chemtech, formerly known as GFT and GKSS. They were also the pioneers in the first industrial development of pervaporation (Jonquieres et al, 2002). Other European countries seem to be well behind regarding the industrialisation of pervaporation. The Swiss company Kuhni commercialises separation systems based on modules developed by other industrial partners likes CM-Celfa Membrantrenntechnik. The Smart Chemical Company (Great Britain) started in 1993 to commercialise PV systems for the concentration or drying of organic solvents by tubular zeolite membranes. In the Netherlands, the same type of approach is also claimed by Pervatech BV, formerly known as Velterop Ceramic Membrane Technology, that provides separation systems equipped with tubular ceramic membranes for drying gases or organic solvents.
The leading USA companies for supplying pervaporation equipment are: Membrane Technology and Research, Bend Research and Isotronics. In 1996, Texaco gave up its projects in pervaporation owing to the important difficulty of finding industrial clients. Note that this date also corresponds very well with the very sharp interruption of Texaco US Patents on PV, despite a very strong research activity and development of industrial PV modules in the 1990s. Another US company, Exxon, did not proceed further with real industrialisation owing to a too low return on investment. Appendix A.2 gives more details and a summary of the main suppliers of pervaporation equipment.
2.5.5 Industrial applications
The applications of pervaporation can be classified into three main categories as shown in Figure 2.5, namely:
• hydrophilic pervaporation (dehydration of organic solvents),
• hydrophobic pervaporation (removal of organic compounds from aqueous solutions), and
PERVAPORATION
4
4
4
Hydrophilic Hydrophobic Organophilic
The target compound water is The target organic compounds are The target organic compound is separated from an aqueous-organic separated from an aqueous-organic separated from an organic-organic mixture by being preferentially mixture by being preferentially mixture by being preferentially permeated through the membrane permeated through the membrane permeated through the membrane
Examples ol Membrane Materials: Examples of Membrane Materials: Examples of Membrane Materials:
Polyvinylalcohol (PVA) Polydimethylsiloxane (PDMS) Polydimethylsiloxane (PDMS) Polyacrylonitrile (PAN) Polyether-Block-Polymide (PEBA) Polyether-Block-Polymide (PEBA) Polyetherimide (PEI) Polytetrafluoro-ethylene (PTFE) Polyvinylalcohol and Polyacrylonitrile Caesium Polyacrylate Polybutadiene (PB) (PVWPAN)
Some Applications: Polypropylene (PP) Polyetherimide (PEI) Some Applications: Some Applications: Breaking of azeotropes of
binary mixtures
dehydration of multi-component mixtures
batchwise dehydration in
Waste water treatment Removal of organic traces from ground and drinking water Removal of alcohol from beer and wine
Separation of ethanol from ETBE Separation of methanol from MTBE Separation of methanol from TAME Separation of benzene and cyclohexane
discontinuous processes Recovery of aromatic compounds in
Separation of ethanol from ETBE Separation of methanol from MTBE Separation of methanol from TAME Separation of benzene and cyclohexane
food technology
Separation of compounds from fermentation broth in biotechnology
Figure 2.5 - Areas of pervaporation: applications and membranes.
2.5.5.1 Hydrophilic pervaporation
Pervaporation has been commercially exploited for dehydrating organics, especially ethanol, isopropanol and ethylene glycol. In the mid-1970's Sulzer Chemtech (then know as GFT) commercialised an economical pervaporation process to produce high purity ethanol from 5-7 wt.% ethanol in fermentation broths. Two plants were built in Brazil and Philippines respectively. This was followed by the installation of several other integrated distillation-pervaporation plants in Europe and Asia. The world's largest pervaporation plant is operating in Bethenville, France, to produce 150,000 litters per day of ethanol using 2 400 m2 membrane area. About seventy commercial plants for
alcohol dehydration are now in operation world-wide. (Dutta et al., 1997)
2.5.5.2 Hydrophobic pervaporation
Most pervaporation applications can be found in the chemical process industry, but there are also other areas, such as food and pharmaceutical industries to concentrate heat-sensitive products or remove aroma compounds, for environmental problems to
remove volatile organic contaminants (VOC's) from waste water or in analytical applications to enrich a given component for quantitative detection. (Mulder, 1998)
Treating wastewaters before discharge by methods such as steam stripping, adsorption, or bioremediation, is expensive and pervaporation may offer an attractive alternative. By using hydrophobic or organophilic membranes, VOC's can be concentrated by orders of magnitude at the permeate side, compared to the feed, and recovered by condensation. The VOC concentrate can either be disposed of by combustion, or be recycled for reuse (Dutta et al., 1997).
Experimental and theoretical work on the removal of VOC's from dilute aqueous solutions have focused on the following three areas (Dutta et al., 1997):
• The selection of hydrophobic and/or organophilic membranes for different VOC contaminants and the study of the relations between membrane materials and their selectivities and permeabilities with different VOC systems.
• The study of mass transfer of organic solutes through membranes, i.e., sorption and diffusion processes within membranes, as well as organic transport in the liquid and vapour phases.
• System design and optimisation to determine the best module configuration and operating conditions.
2.5.5.3 Organophilic pervaporation
The third application of pervaporation lies in the separation of an organic compount from its mixture with another organic compound. This separation is presently the least developed application of pervaporation because of the problems normally associated with membrane stability under relatively harsh conditions, such as high temperatures and pH, but it represents the largest opportunity for energy and cost savings (Feng et al.,
1997). Among these organic-organic separations most of the alcohols form azeotropes with other organics, like aromatics, alkanes, ethers, ester, etc.
A commercial pervaporation plant for the methanol/MTBE separation using a cellulose acetate membrane has been developed by Air Products. These membranes worked well at methanol concentration up to 6%. However, above this concentration, membrane
selectivity was lost due to membrane plasticization. The operating temperature was limited to less than 50°C since the membrane was not stable in the methanol/MTBE mixtures at temperatures above 50°C (Dutta et al., 1997).
2.5.6 Advantages and Disadvantages of Pervaporation
The most obvious advantage of pervaporation is the minimized energy consumption and the separation of close-boiling mixtures or azeotropic mixtures. Furthermore, the required equipment is small, hence it takes less space, and the maintenance cost is low with simple operation and control.
The major drawback against pervaporation in the industry is the difficulty in implementation of the system. The mains reason for that is a lack of knowledge about the capabilities and also a general mistrust of the technology, even though it has been validated on the pilot scale. There is, however, some other drawbacks that should be mentioned (Mulder, 1998):
• The right membrane that would provide high flux and satisfactory separation factor is not always available.
• Low selectivity or flux. • Low membrane lifetime.
2.6 Process Variables
The process factors that influence pervaporation flux and selectivities are: (1) feed composition, (2) feed temperature, (3) permeate pressure (vacuum) and (4) membrane properties. It is essential to understand the effects of these factors so as to select proper operating conditions for the separation of a particular mixture.
2.6.1 Feed composition
The feed composition is the single most important factor in determining pervaporation flux and selectivity. It affects liquid sorption, membrane swelling and diffusion, i.e. the permeation rate of a particular component is affected by all other components present.
