Chemical Engineering Science 54 (1999) 2839 }2847

Multiple reactions in catalytic distillation processes for the production

of the fuel oxygenates MTBE and TAME: Analysis by rigorous model

and experimental validation

Kai Sundmacher

, Gerd Uhde, Ulrich Ho

!mann*

Institut fuKr Chemische Verfahrenstechnik, Technische UniversitaKt Clausthal, Leibnizstrasse 17, D-38678 Clausthal-Zellerfeld, Germany

Abstract

The combination of a chemical reaction and a distillative separation in one apparatus shows several advantages compared to the

separately performed processes. The present contribution presents a comparative study of several possible models of di!erent
complexity for this reactive distillation process. These models consider multiple chemical main and side reactions which are always
present in the industrial production of the fuel ethers MTBE and TAME. Due to the strong nonideality of the reaction mixtures,
liquid-phase activities are used for the formulation of the reaction kinetics. The simulated results were experimentally validated in two
packed laboratory-scale columns. It can be shown that the consideration of side reactions and the modelling of internal catalyst
phenomena play an important role for the interpretation of the experimental results. The comparison of a rate-based model and
a Murphree equilibrium stage model shows that the latter one with a lower complexity yields equivalent model predictions for the
MTBE-system.

 1999 Elsevier Science Ltd. All rights reserved.

Keywords: Reactive distillation; Packed column; Fuel ethers; Multiple reactions; Modelling; Experimental validation

1. Introduction

Reactive distillation is a multifunctional reactor con-

cept which combines a distillative separation with a
chemical reaction, preferably heterogeneously catalysed.
The expected bene"ts from such a synergetic interaction
of two unit operations in one column are an enhanced
conversion in excess of chemical equilibrium, an in-
creased selectivity, the avoiding of hot spots by using
heat of reaction for distillation and the overcoming of
azeotropic limitations (e.g. Sundmacher and Ho!mann,
1996).

Fig. 1 shows the #ow scheme of a common industrial

process (Obenaus and Droste, 1980) for the production of
the fuel ethers tert-amyl-methylether (TAME) or methyl-
tert-butylether (MTBE) in comparison to a possible pro-
cess in which a reactive distillation is involved. In the "rst

*Corresponding author.
Present address: Max-Planck-Institute for Dynamics of Complex

Technical Systems, Leipziger Strasse 44, ZENIT-Building, D-39120
Magdeburg, Germany.

step of both processes impurities are removed from the
mixture of the reactive ole"ns and inert hydrocarbons in
a feed wash. In the "rst "xed-bed reactor the main part of
the overall conversion of reactive ole"ns is attained.
This reactor is "lled with an acid ion exchange resin
as catalyst. For the further increase of the ole"n conver-
sion a second reactor is necessary in the common indus-
trial process. In a fourth process step the product
TAME is isolated via distillation from hydrocarbons
and methanol. This unreacted methanol is removed from
unreacted and inert hydrocarbons via an extraction
step and separated from water via distillation. Then,
methanol can be recycled to the "xed-bed reactor. The
process which involves a reactive distillation combines
three process steps of the common process and, by this,
signi"cant savings of energy and investment costs are
achieved.

In the last two years, quite a number of papers were

published on design principles (e.g. Bessling et al., 1997;
Espinosa et al., 1996; Nisoli et al., 1997; Sneesby et al.,
1997), modelling and operation (e.g. Alejski and Duprat,
1996; Bock et al., 1997; Hauan et al., 1997; Isla and
Irazoqui, 1996; Sundmacher and Ho!mann, 1996) of

0009-2509/99/$ } see front matter

 1999 Elsevier Science Ltd. All rights reserved.

PII: S 0 0 0 9 - 2 5 0 9 ( 9 8 ) 0 0 5 2 0 - X

Fig. 1. (a) Flow scheme of a conventional process for the production of fuel ethers. (b) Flow scheme of a process including a reactive distillation
column for the production of fuel ethers.

reactive distillation columns. Most of these consider only
a single desired reaction taking place. In contrast to that,
in the present contribution multiple main and side reac-
tions in catalytic distillation columns are theoretically
analysed and experimentally validated. This is of remark-
able relevance for the application of reactive distillation
columns under industrial conditions where, in most
cases, at least one side reaction is involved.

2. Considered reaction systems

For our analysis the syntheses of the fuel ethers MTBE

and TAME were chosen as two examples of high indus-
trial importance. The reaction schemes of these two
reaction systems are depicted in Fig. 2. In case of TAME-

synthesis, the fuel ether is formed from methanol and the
two isomers 2-methyl-1-butene (2MB1) and 2-methyl-2-
butene (2MB2). The latter two components simulta-
neously isomerise, forming a reaction triangle with the
two TAME synthesis reactions (Oost and Ho!mann,
1996). The production of MTBE from methanol (MeOH)
and isobutene (IB) is accompanied by the undesired par-
allel formation of isobutene-dimers (DIB) which deacti-
vate the catalytic distillation packing.

In both reaction systems additional side reactions can

occur. One of these reactions is the etheri"cation reaction
of methanol to dimethylether (DME):

2MeOH

& DME#HO.

(1)

Water can react in a consecutive reaction with an ole"n
to the appropriate alcohol. In case of the MTBE-system

2840

K. Sundmacher et al./Chemical Engineering Science 54 (1999) 2839}2847

Fig. 2. Schemes of the considered reaction systems.

Table 1
Parameters for vapor pressure correlations

Component

AG

BG

CG

DG

¹

AG [K]

p

AG [bar] Eq. no. Ref.

MeOH

!

8.5480

0.7698

!

3.1085

1.5448

512.6

80.9

(I)

Reid et al. (1987)

IB

!

6.9554

1.3567

!

2.4522

!

1.4611

417.9

40.0

(I)

Reid et al. (1987)

MTBE

!

7.8252

2.9549

!

6.9408

12.174

497.2

34.8

(I)

Reid et al. (1987)

1-Butene

!

6.8820

1.2705

!

2.2628

!

2.6163

419.6

40.2

(I)

Reid et al. (1987)

DIB

9.1168

2932.2

!

52.535

*

*

*

(II)

Mohl et al. (1996)

2MB1

!

6.8299

0.7266

!

2.1536

!

3.6223

465.0

34.5

(I)

Reid et al. (1987)

2MB2

!

7.7144

1.9595

!

3.1571

!

2.2252

470.0

34.5

(I)

Reid et al. (1987)

TAME

9.1556

2782.4

!

55.243

*

*

*

(II)

Cervenkova and
Boublik, (1984)

nP

!

7.2894

1.5368

!

3.0837

!

1.0246

469.7

33.7

(I)

Reid et al. (1987)

(I) ln

p

QG

p

AG



"

(1!x)

\ [AGx#BGx #CGx#DGx] with x"1!

¹

¹

AG

.

(II) ln (p

QG)"AG!BG/[¹#CG] Remark: pQG in (bar); ¹ in (K).

DIB"2, 4, 4-trimethyl-1-pentene.

tert-butanol (TBA) results from the reaction of isobutene
and water:

IB#HO &TBA.

(2)

3. Chemical and phase equilibria

For the formulation of the kinetics of the reversible

TAME-reactions three coupled chemical equilibria have
to be considered. Rihko et al. (1994) proposed the follow-

ing expressions to calculate the equilibrium constants of
the two TAME-formation reactions:

K?2+#"1.057;10\exp



4273.5 K

¹



,

(3)

K?2+#"1.629;10\exp



3374.4 K

¹



.

(4)

To obtain a kinetic model of the reversible MTBE-
reaction the chemical equilibrium has to be considered,
too. For the simulations presented here we used the
correlation published by Reh"nger and Ho!mann (1990)
for the temperature dependence of the equilibrium con-
stant K?+2 #.

The vapor}liquid-phase equilibria are calculated as-

suming an ideal vapor phase. Thus, the phase equilibria
can be calculated with the liquid-phase activity coe$-
cients

cG and the vapor pressures of the pure compo-

nents p

QG:

x

*C

G

)

cG(x*C, ¹))pQG(¹)"x4G)p.

(5)

The equations and coe$cients used for the calculation of
the pure component vapor pressures p

QG are summarized

in Table 1.

For the estimation of the activity coe$cients in the

TAME-system the Wilson model was applied. This
model predicts a homogeneous mixture of methanol with
the C5 hydrocarbons for the whole concentration range
at boiling temperature. This is in agreement with the
experimental data. In contrast to that the UNIQUAC-
model predicts phase splitting, and therefore it is not an
adequate model for the TAME-system. The activity coef-

"cients

of the MTBE-system are calculated from the

UNIQUAC-model. The parameters of both activity coef-

"cient models, Wilson and UNIQUAC, are summarised

in Tables 2 and 3.

K. Sundmacher et al./Chemical Engineering Science 54 (1999) 2839}2847

2841

Table 2
Wilson interaction parameters for activity coe$cient calculation in the TAME reaction system from Mohl et al. (1996)

Component i

Component j

MeOH

2MB1

2MB2

TAME

n!P

(

kGH!kGG) (kJ/mol)

MeOH

0

9.7723

10.147

4.8263

11.749

2MB1

1.3765

0

0.47880

!

0.61175

0.32674

2MB2

0.96881

!

0.47794

0

!

0.38604

0.36228

TAME

!

0.17700

0.95133

0.71233

0

1.1439

nP

1.9467

!

0.19418

!

0.26549

!

0.44784

0

Table 3
UNIQUAC parameters for activity coe$cient calculation in the MTBE reaction system from Reh"nger and Ho!mann (1990)

Component i

qG

rG

Component j

MeOH

IB, 1-Butene

MTBE

TMP1

(

jGH!jGG) (kJ/mol)

MeOH

1.432

1.431

0

!

0.2941

!

0.7320

!

0.05964

IB, 1-butene

2.684

2.920

5.873

0

0.4340

0.2121

MTBE

3.632

4.068

3.897

!

0.2047

0

!

0.1424

DIB

4.920

5.616

5.873

!

0.1637

0.3804

0

Mohl et al. (1996).

DIB - 2, 4, 4-trimethyl-1-pentene.

4. Reaction kinetics

4.1. TAME-kinetics

The microkinetics of the heterogeneously catalyzed

liquid-phase TAME-formation are formulated according
to Oost and Ho!mann (1996). These authors derived
a Langmuir}Hinshelwood rate expression in terms of
the liquid-phase activities. The two isomeric ole"ns
2M1B and 2M2B show ideal mixture behavior in the
liquid phase. Therefore they can be lumped together
to one isoamylene (IA) fraction. Then, based on the
assumptions that the TAME-formation reactions of the
absorbed species are rate-determining and that nearly all
active sites are occupied by methanol molecules, the
following rate equation for the net turn-over-number
results:

r2+#"k2+#(¹))

a'

a+-&

!

1

K2+#(¹)

#

1

K2+#(¹)



a2+#

a

+-&



.

(6)

The reaction rate constant k2+# is given by the fol-

lowing Arrhenius equation:

k2+#(¹)"2.576

mmol

s eq

exp



!

89.5 kJ/mol

R

;

1

¹

!

1

333 K



.

(7)

Eq. (6) implies that the isomerization of the C5-ole"ns
runs much faster than the two TAME-formation reac-
tions and therefore the chemical equilibrium of the
isomerization reaction is established. It should be men-
tioned that this is in contrast to the work of Rihko et al.
(1997), who found that the isomerization reaction is
much slower than expected by Oost and Ho!mann
(1996). Therefore, additional kinetic experiments are ne-
cessary to clarify the importance of the isomerization
reaction kinetics.

4.2. MTBE-kinetics

For the microkinetic description of the heterogeneous-

ly catalyzed liquid-phase MTBE-formation, the well-es-
tablished rate expression of Reh"nger and Ho!mann
(1990) is used. These authors introduced activity-based
Langmuir isotherms to formulate the sorption equilibria
of the components between the liquid phase and the
catalyst gel phase. If the reaction of sorbed species is
the r.d.s. and if methanol is sorbed highly selective in the

2842

K. Sundmacher et al./Chemical Engineering Science 54 (1999) 2839}2847

Fig. 3. Simulated reaction rates and component activity pro"les within
a catalyst particle for the MTBE system.

Fig. 4. Model family for heterogeneously catalysed reactive distillation
process.

ion-exchange resin, the following microkinetics result:

r+2 #"k+2 #(¹)

a'

a+-&

!

1

K?+2 #(¹)

a+2 #

a

+-&



.

(8)

The temperature dependence of the rate constant is given
by the Arrhenius equation

k+2 #(¹)"15.5

mmol

s eq

exp



!

92.4 kJ/mol

R

1

¹

!

1

333 K



.

(9)

The dimerization of isobutene forms the two isomers 2, 4,
4-trimethyl-1-pentene and 2, 4, 4-trimethyl-2-pentene.
They were lumped to one pseudo-component di-
isobutene (DIB). According to Haag (1967) the formation
of DIB follows a Langmuir}Rideal mechanism:

Adsorption:

IB#S

8IB ) S.

(10)

Surface reaction:

IB ) S#IB &&

I

"'

DIB ) S.

(11)

Desorption:

DIB ) S

8 DIB#S.

(12)

The second reaction step, Eq. (11), is assumed to be rate
determining and irreversible. At low methanol concen-
trations, where this reaction takes place, only methanol
and isobutene molecules are absorbed in the ion-
exchange resin. From these assumptions the following
rate equation is obtained:

r"' "k"' (¹))

a

'

Ka+-&#a'

where K,

K+-&

K'

.

(13a,b)

As can be seen from Eq. (13b), the parameter K is de"ned
as the ratio of the sorption constants of methanol and
isobutene. By "tting of published experimental data this
parameter was estimated to K"500. The rate constant
k"' was determined from the experimental rate data of

Haag (1967):

k"' (¹)"1270

mmol

s eq

exp



!

66.7 kJ/mol

R

1

¹

!

1

333 K



(14)

A detailed analysis of the simultaneous mass transport
and reaction phenomena inside the porous catalyst body
(Uhde et al., 1998) revealed that, for low methanol bulk
concentrations, the dimerization reaction of isobutene
will occur in the inner core of the catalyst where the
MTBE-formation rate vanishes. This is illustrated in
Fig. 3 for the case of a spherical catalyst body.

This knowledge was used to calculate the e!ective rate

of the side reaction with the help of a modi"ed catalyst
model of Sundmacher and Ho!mann (1994) which ac-
counts for the e!ect of internal mass transport resistances
on the etheri"cation macrokinetics.

5. Modelling of packed sections in a reactive distillation
column

A detailed comparative study of several possible react-

ive distillation models of di!erent complexity was
performed. The model family consists of four di!erent
models which are illustrated in Fig. 4.

Models (1) and (3) assume that vapor}liquid equilib-

rium is established between the bulk phases. Models
(2) and (4) are rate-based, i.e. the multicomponent
mass transport between the vapor phase and liquid phase
is taken into account. The interfacial transport is
modelled on the basis of a matrix solution of the general-
ized Maxwell}Stefan equations (Taylor and Krishna,
1993).

Furthermore, the models can be distinguished with

respect to the consideration of internal catalyst mass
transport phenomena. In models (1) and (2) the solid
catalyst is treated quasihomogeneously, i.e. at each cata-
lytically active site the liquid bulk phase composition is

K. Sundmacher et al./Chemical Engineering Science 54 (1999) 2839}2847

2843

Table 4
Column con"guration and catalyst properties

TAME

MTBE

Column

H  (m)

0.547

0.510

H  (m)

0.547

0.510

dA (m)

0.076

0.054

Catalyst

e1

0.49

c* [eq(H



)l

\]

1.2

Shape

Raschig rings

d ;h ;t  (mm)

9

;9;2

present. Models (3) and (4) include possible mass trans-
port resistances inside the catalyst. These resistances are
treated via the catalyst e!ectiveness factor concept based
on the model of Sundmacher and Ho!mann (1994).

Models (1) and (3) consist of

E the mass balances of the components for a packed

column section,

E the above formulated chemical microkinetics,

E the overall energy balance for a packed column sec-

tion,

E the vapor

} liquid phase equilibria; any deviation from

phase equilibria is taken into account by a Murphree-
e$ciency stage approach.

In addition to this, models (2) and (4) contain

E the mass balances of the components in the vapor

phase,

E the mass balances of the components at the va-

por}liquid interface.

The corresponding formulation of these balance equa-
tions can be found in detail elsewhere (Sundmacher and
Ho!mann, 1996). The steady-state solution of the model
equations is obtained by introducing accumulation terms
to the component mass balances (&&False Transient
Method''). The other balance equations are treated alge-
braically. This leads to a set of di!erential and algebraic
equations. For their numerical integration the extrapola-
tion integrator LIMEX (Deu#hard et al., 1987) was
applied.

6. Simulated results and experimental validation

Based on the model approach described above,

detailed simulations were carried out to investigate
the in#uence of selected operating parameters (re#ux
ratio r, reboiler heat input Q
, pressure p) on the

process performance (conversion, yield, selectivity) for
the TAME reaction triangle and the MTBE parallel
reactions.

The reactive distillation experiments are performed in

two di!erent stainless-steel laboratory columns with un-
structured Raschig ring packings. Both columns can be
operated at elevated pressures up to p"1 MPa and
temperatures up to ¹"2003C. The MTBE reaction sys-
tem was investigated in a column with 54 mm inner
diameter, and the experiments for the TAME reaction
system were performed in a larger column with 76 mm
inner diameter. The catalytically active packing consists
of 9

;9 mm porous glass rings with a strong acidic ion

exchange resin inside the pores. This catalyst is an in-
house development of our institute (Kunz and Ho!mann,
1995). The ion exchange capacity of this catalyst per
column volume is about 0.6 eq(H



)/l.

The noncatalytic packing consists of the same porous

glass rings without the ion-exchange polymer. In both
columns the packing is distributed in an upper catalytic
part and a lower noncatalytic part. The feed point of the
premixed reactants and the inert component was in the
middle between the two packings. Table 4 gives an over-
view of the column geometry and the catalyst properties.

For all simulations presented below we assumed a heat

loss coe$cient of 2 W/(m

 K) through the column wall.

This value is based on a comparison of simulated results
with various experimental data from our reactive distilla-
tion laboratory columns (Gravekarstens, 1998). All simu-
lations were carried out with a number of 10 packed
column sections. The simulations with model (1) are
carried out with a Murphree-e$ciency of 0.8. This
value coincides with the experimental "ndings of Bessling
et al. (1998). These authors investigated the separation
e$ciency of the catalytic Raschig rings which were used
for the discussed reactive distillation experiments. They
found that the number of theoretical separation stages
per meter is about 8 for the vapor load range established
in

our

columns

(vapor

F-factor:

F"0.3 Pa

....

0.9 Pa

). Since the total height of the column packing is

about 1 m the overall Murphree-e$ciency stage factor of
the 10 simulated stages is 0.8.

First, simulated and experimental results of the

TAME-reaction system are presented. Due to the fact
that the TAME-formation rate is more than ten times
lower than the MTBE-formation rate, mass transfer re-
sistances inside the catalyst bodies are negligible, and
therefore a quasihomogeneous approach should be su$-
cient to describe the TAME-reactive distillation column.

In Fig. 5a calculated and measured temperature pro-

"les along the column are compared, and in Fig. 5b the

corresponding liquid-phase compositions are depicted.
The two reactive isoamylene isomers 2M1B and 2M2B
and the inert solvent n-pentane are lumped together to
one single C5-fraction. The simulated results were ob-
tained with the vapor}liquid nonequilibrium model (2).
Both "gures show a good agreement between simulated
and experimental data. The bottom product contains no
methanol and the composition of the distillate is domin-
ated by the existence of the binary azeotrope between

2844

K. Sundmacher et al./Chemical Engineering Science 54 (1999) 2839}2847

Fig. 5. (a) Experimental and simulated temperature pro"les for TAME
reactive distillation column. (b) Experimental and simulated composi-
tion path for TAME reactive distillation column; experimental conver-
sion of isoamylenes: X'"37%; experimental selectivity w.r.t. TAME:

S2+#+100%.

Fig. 6. Experimental and simulated composition path for TAME react-
ive distillation column with high isoamylene content in feed mixture;
experimental conversion of isoamylenes: X'"39%; experimental sel-

ectivity w.r.t. TAME: S2+#"88%.

methanol and the C5-components. The concentration of
the desired product TAME in the bottom is not very
high. This is caused by a relatively low reaction rate and
a high re#ux ratio which forces considerable amounts of
C5-components to leave the column at the bottom.

According to the TAME-reaction kinetics, Eq. (6),

higher etheri"cation rates in the catalytic column section
should be achievable at a higher content of isoamylenes
and a lower content of methanol in the feed mixture.

In correspondence to this, Fig. 6 shows the liquid-

phase composition path for an experiment with
a changed feed composition and lower re#ux ratio. In
fact, the TAME content in the bottom is much higher
than in Fig. 5b. The conversion of isoamylenes slightly
increased from X' "0.37 (exp. in Fig. 5b) to X'"0.39

(exp. in Fig. 6) which is mainly due to the high content of
isoamylenes in the feed. However, by the undesired
formation of dimers from isoamylenes, the selectivity
w.r.t.

TAME

decreased

from

S2+#+100% to

S2+#"88%. Note that the dimers (12 mol% in reboiler

liquid) are lumped together with TAME in the concen-
tration triangle of Fig. 6.

Fig. 7a depicts experimental and simulated temper-

ature pro"les of the MTBE-system. Two models are used
to simulate the experimental results. Model (1) is
a quasihomogeneous approach which accounts for the
main reaction only. The parallel occurrence of the dimer-
ization of isobutene can not be described by this
approach since this requires a catalyst model which ac-
counts for mass transport resistances inside the catalytic
rings. Therefore, model (1) underestimates the temper-
ature in the reboiler where the high boiling by-product
DIB is mainly located. The more detailed model (3)
includes the side reaction and predicts a higher temper-
ature in the reboiler. Fig. 7b shows the corresponding
simulated liquid-phase composition path and some ana-
lyzed liquid samples. The component isobutene and the
inert component 1-butene are lumped together to one
C4-fraction. The analyzed experimental samples clearly
revealed that a considerable amount of dimers is formed
under the given process conditions. Therefore in the
lower part of the column, signi"cant deviations between
the experimental concentrations and those, which are
predicted by model (1), exist.

Model (3) predicts a composition of the reboiler liquid

which is in good agreement with the experimental data.
The compositions in the upper part of the column (com-
position trajectory near the C4 corner of the tetrahedron)
are predicted quite well by both models. In a previous
work, Sundmacher et al. (1997) had already shown that
the most complicated model (4) yields nearly the same
results as the much simpler equilibrium stage model (3).
As a consequence, an equilibrium stage approach is su$-
cient for the adequate description of the MTBE process.

K. Sundmacher et al./Chemical Engineering Science 54 (1999) 2839}2847

2845

Fig. 7. (a) Experimental and simulated temperature pro"les for MTBE
reactive distillation column. (b) Experimental and simulated composi-
tion path for MTBE reactive distillation column.

7. Conclusions

For the simulation of a heterogeneously catalyzed re-

active distillation process several models with di!erent
complexity were tested. The simulated results were ex-
perimentally validated for the TAME- and the MTBE-
reaction system in two laboratory-scale columns. In case
of TAME-column a fully rate-based nonequilibrium
model was applied successfully. For the MTBE-system it
was shown that a Murphree-equilibrium stage model
which includes the calculation of simultaneous mass
transport and reaction phenomena inside the catalytic
distillation packing is well applicable. The formation of
high boiling dimers from isobutene and their accumula-
tion in the liquid phase lead to a considerable increase of
the temperatures in the bottom section of the MTBE-
column. Obviously, it is necessary to account for the side
reaction in designing a catalytic distillation column for
this reaction system.

As a work for the future, the column operating condi-

tions have to be adjusted to optimize the conversion of
ole"ns in the reactive distillation process. This process
optimization should be based on the validated models
which are presented in this contribution.

Acknowledgements

The authors wish to thank the Volkswagen-Stiftung in

Germany for "nancial support of this research work
within the project &&Modellierung der Integration von Stof-
ftrennprozessen und komplexen Reaktionsnetzwerken''.

Notation

aG

liquid-phase activity of component i

c*

ion-exchange capacity related to catalyst vol-
ume, eq(H



)/l

dA

inner column diameter, m

d 

outer diameter of ring shaped catalyst, mm

FK

feed mass #ow rate, kg/h

h 

height of ring shaped catalyst, mm

H 

height of purely distillative column section, m

H 

height of catalytic column section, m

kH

reaction rate constant, mol/(eq(H



) s)

K

ratio of sorption constants, Eq. (18)

KQG

sorption constant of component i

K?H

activity-based chemical equilibrium constant
of reaction j

p

total operating pressure, MPa

pA

critical pressure, bar

p

QG

saturated vapor pressure of component i, bar

Q

reboiler heat input, W

r

re#ux ratio

rH

rate of reaction j related to catalytically active
sites, mol/(eq(H



)s)

R

universal gas constant, 8.314 J/(mol K)

S2+#

selectivity w.r.t. TAME

t 

wall thickness of ring-shaped catalyst, mm

¹

temperature, K

¹A

critical temperature, K

xG

mole fraction of component i

Greek letters

e1

volume fraction of solid catalyst in catalytic
packing section

Subscripts

DIB

DIB reaction (see Fig. 2)

ISO

isomerization reaction of 2M1B and 2M2B
(see Fig. 2)

MTBE

MTBE formation (see Fig. 2)

TAME1

TAME formation from 2M1B (see Fig. 2)

TAME2

TAME formation from 2M2B (see Fig. 2)

Superscripts

e

equilibrium

F

related to the feed

2846

K. Sundmacher et al./Chemical Engineering Science 54 (1999) 2839}2847

<

related to the vapor phase

¸

related to the liquid phase

S

related to the solid catalyst

Abbreviations

C4

C4-fraction ("isobutene#1-butene)

C5

C5-fraction ("nP#2M1B#2M2B)

DIB

diisobutene (2,4,4-trimethyl-1-pentene#
2,4,4-trimethyl-2-pentene)

IA

isoamylene (2M1B#2M2B)

IB

isobutene

MeOH

methanol

MTBE

methyl-tert-butylether

nP

n-pentane

S

active site of catalyst

TAME

tert-amyl-methylether

TBA

tert-butanol

2M1B

2-methyl-1-butene

2M2B

2-methyl-2-butene

References

Alejski, K., & Duprat, F. (1996). Dynamic simulation of the multicom-

ponent reactive distillation. Chem. Engng Sci., 51, 4237}4252.

Bessling, B., LoKning, J.M., OhligschlaKger, A., Schembecker G., & Sun-

dmacher, K. (1998). Investigations on the synthesis of methyl acetate
in a heterogeneous reactive distillation process. Chem. Engng ¹ech-
nol., 21, 393}400.

Bessling, B., Schembecker, G., & Simmrock, K.H. (1997). Design of

processes with reactive distillation line diagrams. Ind. Engng Chem.
Res., 36, 3032}3042.

Bock, H., Jimoh, M., & Wozny, G. (1997). Analysis of reactive distilla-

tion using the esteri"cation of acetic acid as an example. Chem.
Engng ¹echnol., 20, 182}191.

Cervenkova, I., & Boublik, T. (1984). Vapor pressure, refractive indices

and densities at 203C and vapor}liquid equilibria at 101.325 kPa in
the TAME}methanol system. J. Chem. Engng Data, 29, 425}427.

Deu#hard, P., Hairer, E., & Zugck, J. (1987). One step and extrapola-

tion methods for di!erential-algebraic systems. Numer. Math., 51,
501}516.

Espinosa, J., Aguirre, P., & PeHrez, G. (1996). Some aspects in the design

of multicomponent reactive distillation columns with a reacting
core: Mixtures containing inerts. Ind. Engng Chem. Res., 35,
4537}4549.

Gravekarstens, M. (1998). Ein-u} des Detaillierungsgrades auf die

Modellierung von heterogen katalysierten Reaktivdestillationsprozes-
sen am Beispiel der M¹BE-Synthese. Ph.D. thesis, TU Clausthal.

Haag, W.O. (1967). Oligomerization of isobutene on cation exchange

resins. Chem. Engng Prog. Symp. Ser., 63, 140}147.

Hauan, S., Hertzberg, T., & Lien, K.M. (1997). Multiplicity in reactive

distillation of MTBE. Comput. Chem. Engng, 21, 1117}1124.

Isla, M.A., & Irazoqui, H.A. (1996). Modelling, analysis and simulation

of a methyl tert-butyl ether reactive distillation column. Ind. Engng
Chem. Res., 35, 2696}2708.

Kunz, U., & Ho!mann, U. (1995). Preparation of catalytic poly-

mer/ceramic ion exchange packings for reactive distillation col-
umns. Preparation Catalysts <I, 91, 299}309.

Mohl, K.D., Kienle, A., & Gilles, E.D. (1996). Personal communication,

Institut fuKr Systemdynamik und Regelungstechnik, UniversitaKt
Stuttgart.

Nisoli, A., Malone, M.F., & Doherty, M.F. (1997). Attainable regions

for reaction with separation. A.I.Ch.E. J., 43, 374}387.

Obenaus, F., & Droste, W. (1980). HuKls process: methyl tertiary

butylether. ErdoKl Kohle-Erdgas-Petrochem. Brennsto+chem., 33,
271}275.

Oost, C., & Ho!mann, U. (1996). The synthesis of tertiary amyl methyl

ether (TAME): microkinetics of the reactions. Chem. Engng Sci., 51,
329}340.

Reh"nger, A., & Ho!mann, U. (1990). Kinetics of methyl tertiary butyl

ether liquid phase synthesis catalysed by ion exchange resin. Chem.
Engng Sci., 45, 1605}1626.

Reid, C.R., Prausnitz, J.M., & Poling, B.E. (1987). ¹he properties of

gases and liquids (pp. 656}732). New York: McGraw-Hill.

Rihko, L.K., Kiviranta-PaKaKkkoKnen, P.K., & Krause, A.O. (1997). Kin-

etic model for the etheri"cation of isoamylenes with methanol. Ind.
Engng Chem. Res., 36, 614}621.

Rihko, L.K., & Krause, A.O. (1995). Kinetics of heterogeneously

catalyzed tert-amyl methyl ether reactions in the liquid phase. Ind.
Eng. Chem. Res., 34, 1172}1180.

Rihko, L.K., Linnekoski, J.A., & Krause, A.O. (1994). Reaction

equilibria in the synthesis of 2-methoxy-2-methylbutane and 2-
ethoxy-2-methylbutane in the liquid phase. J. Chem. Engng Data, 39,
700}704.

Sneesby, M.G., TadeH, M.O., Datta, R., & Smith, T.N. (1997). ETBE

synthesis via reactive distillation. 1. Steady-state simulation and
design aspects. Ind. Engng Chem. Res., 36, 1855}1869.

Sundmacher, K., Gravekarstens, M., Rapmund, P., Thiel, C., & Hof-

fmann, U. (1997). Intensi"ed production of fuel ethers in reactive
distillation columns } systematic model-based process analysis.
AIDIC Conference Series (Vol. 2, pp. 215}221).

Sundmacher, K., & Ho!mann, U. (1994). Macrokinetic analysis of

MTBE-synthesis in chemical potentials. Chem. Engng Sci., 49,
3077}3089.

Sundmacher, K., & Ho!mann, U. (1996). Development of a new cata-

lytic distillation process for fuel ethers via a detailed nonequilibrium
model. Chem. Engng Sci., 51, 2359}2368.

Taylor, R., & Krishna, R. (1993). Multicomponent mass transfer. New

York: Wiley.

Uhde, G., Sundmacher, K., & Ho!mann, U. (1998). AktivitaKt und

SelektivitaKt makroporoKser Ionenaustauscher-Katalysatoren fuKr die
Veretherung von Ole"nen. Chem. Ing. ¹ech., 70, 886}890.

K. Sundmacher et al./Chemical Engineering Science 54 (1999) 2839}2847

2847