Accepted Manuscript

Title: Optimal design of intensified processes for DME

synthesis

Author: <ce:author id="aut0005"

author-id="S0098135417300042-
d5d5310a81bd6c30e975ca40a17754dd"> Costin Sorin
Bı̂ldea<ce:author id="aut0010"
author-id="S0098135417300042-
c228a91b1610d584b3e3898077dd12bc"> Romuald
Győrgy<ce:author id="aut0015"
author-id="S0098135417300042-
7a8846e23d24b848cde7be4cbe642bc2"> Cristian C.
Brunchi<ce:author id="aut0020"
author-id="S0098135417300042-
2797caab2b75845e48ac8d889432d060"> Anton A.
Kiss

PII: S0098-1354(17)30004-2
DOI: http://dx.doi.org/doi:10.1016/j.compchemeng.2017.01.004
Reference: CACE 5660

To appear in: Computers and Chemical Engineering

Received date: 26-8-2016

Revised date: 28-12-2016
Accepted date: 3-1-2017

Please cite this article as: Bı̂ldea, Costin Sorin., Győrgy, Romuald.,
Brunchi, Cristian C., & Kiss, Anton A., Optimal design of intensified
processes for DME synthesis.Computers and Chemical Engineering
http://dx.doi.org/10.1016/j.compchemeng.2017.01.004

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Optimal design of intensified processes for DME synthesis

Costin Sorin Bîldea,1 Romuald Győrgy,2 Cristian C. Brunchi,3 Anton A. Kiss 4,5

1
University “Politehnica” of Bucharest, Polizu 1-7, 011061 Bucharest, Romania
2
Aristotle University of Thessaloniki, Department of Chemical Engineering, Greece
3
Delft University of Technology, Process & Energy Department, Leeghwaterstraat 34, 2628
CA Delft, The Netherlands
4
AkzoNobel Research, Development & Innovation, Process Technology SRG, Zutphenseweg
10, 7418 AJ Deventer, The Netherlands. E-mail: [email protected]
5
Sustainable Process Technology Group, Faculty of Science and Technology, University of
Twente, PO Box 217, 7500 AE Enschede, The Netherlands

*
Corresponding authors’ e-mail addresses: [email protected], [email protected]

Highlights
 Rigorous design and simulation of gas-phase and reactive distillation DME processes
 Reactive distillation is suitable for new DME plants due to lowest CapEx and OpEx
 Existing plants can be revamped to gas-phase reactor coupled with reactive distillation

Abstract
Dimethyl ether (DME) is widely used as green aerosol propellant, precursor to other organic
compounds, or as a clean fuel for diesel engines or in combustion cells. The classic method
for producing DME is by dehydration of methanol in a catalytic gas-phase reactor, and
purification in a direct sequence of two distillation columns. Reactive distillation (RD) is a
much better alternative for DME synthesis, based on process intensification principles.
This paper presents the optimal design of novel DME processes based on reactive distillation,
and makes a fair comparison with the classic reactor-separation-recycle process (for a plant
capacity of 100 ktpy DME). The new RD processes were optimized in terms of minimizing
the total annual costs, leading to savings of 30% in CapEx and 6% in energy requirements for
the RD process. The results indicate that a RD column is recommended for new DME plants,
while a combination of gas-phase reactor and RD is suitable for revamping existing plants.
Optimal design of intensified processes for DME synthesis

Keywords
Dimethyl ether, gas-phase reactor, reactive distillation, process design, process optimization

1. Introduction
Dimethyl ether (DME) has remarkable properties (e.g. non-toxic colorless gas, non-corrosive,
non-carcinogenic, environmentally friendly) that make it usable as green aerosol propellant,
as precursor to organic compounds, and as clean fuel for diesel engines or combustion cells
(Muller and Hubsch, 2005). It is worth noting that DME market is estimated to worth 9.7
billion USD by 2020 (www.marketsandmarkets.com).
Many of the current industrial DME plants are based on the methanol dehydration route – in a
conventional reactor-separation-recycle process (Kiss et al., 2007; Dimian et al., 2014) –
whereas methanol can be produced from (bio-based) syngas or by CO2 hydrogenation (Kiss et
al., 2016). The chemical recycling of carbon dioxide to methanol and dimethyl ether has been
advocated by Olah et al. (2009) and it plays a key role in a methanol economy.
Unexpectedly, there are only few references in literature related to the design and economic
evaluation of an entire process for DME synthesis. Lei et al. (2011) discuss the economics of
a process consisting of a fixed-bed reactor and two ordinary distillation columns.
Nevertheless, the reactor is modeled as a simple Gibbs reactor where equilibrium is attained.
The recycle of unreacted methanol was not considered, neither the issues related to heating of
the reactant, cooling the reactor effluent, or heat-integration. Bai et al (2013) presented the
simulation of a process for DME synthesis by vapor phase methanol dehydration in a fixed-
bed adiabatic reactor, with 1000 ktpy production rate. The simulation model is also compared
to data acquired from a set of industrial production equipment with a capacity of 200 ktpy. An
industrial solid acidic catalyst was used, and the kinetic model was developed by the same
authors (Li et al, 2009). A follow-up paper (Bai et al., 2014) mentions a five-bed industrial
reactor, operated adiabatically and provided with heat-exchangers for inter-stage cooling.
They also present detailed simulation results for a three-bed adiabatic reactor where inter-
stage cooling is achieved by injecting (a fraction of) the fresh, reactor cold feed. Farsi et al.
(2010) used a genetic algorithm to find the temperature distribution in a multi-tubular reactor
which maximizes the DME production rate. Compared to the conventional adiabatic reactor,
the methanol conversion increases from 81.9% to 85.75% which results in 4.3% increase of
DME production rate. However, this is achieved at the cost of a more complex reactor. The
authors also remark that the optimal temperature profile is difficult to be obtained in practice.
None of the above-mentioned papers (Bai et al., 2014, Farsi et al., 2010) discuss the positive

2
Optimal design of intensified processes for DME synthesis

impact that the more complex reactor setup has on the plant profitability. Han et al (2009)
presented the simulation of a one-step process for DME production in a slurry reactor, from
coal-based syngas (66.25 ktpy production rate). A simple CSTR model was used, after
verification on a set of experimental data obtained on a laboratory-scale reactor fed with 0.93
mol/h syngas.
To increase the capacity of DME production, the chemical industry needs novel eco-efficient
processes that can meet the growing market demand (Arcoumanis et al., 2008; Azizi et al.,
2014). The suggested improvements of the methanol dehydration process followed a number
of technologies reported in literature: fully thermally-coupled distillation columns (Petlyuk)
or dividing-wall column replacing the direct distillation sequence (Kiss, 2013; Kiss and Ignat,
2013); self-heat recuperation (Kansha et al., 2015); simultaneous synthesis and separation in a
catalytic distillation process (An et al., 2004; Lei et al., 2011); catalytic cyclic distillation
(Patrut et al., 2014); or reactive dividing-wall column (Kiss and Suszwalak, 2012). Additional
DME technologies (e.g. coupled and dual type reactors, micro-reactors, membrane reactors,
and spherical reactors), including the direct route from syngas, are also well described in the
review paper of Azizi et al. (2014).
Among these technologies, reactive distillation (RD) seems to be the most promising (Azizi et
al, 2014), being a proven process intensification method that effectively combines the reaction
and separation into a single unit (Agreda et al., 1990). RD can considerably improve the
performances of an equilibrium limited process, by pulling the conversion to completion (thus
avoiding recycles), increasing selectivity and productivity, reducing the energy use, and
overall reducing the CapEx and OpEx (Huss et al., 1999, Malone et al., 2003; Schoenmakers
and Bessling, 2003; Sundmacher and Kienle, 2003; Harmsen, 2010; Kiss and Bildea, 2012).
This study is the first to provide the optimal design of novel DME processes based on reactive
distillation, one process (single RD column) being most suitable for new DME plants, and
another one (coupling a gas-phase reactor and a RD column) that is more appropriate for
revamping existing DME plants based on gas-phase methanol dehydration. These novel
process alternatives are optimized in term of minimum total annual costs (TAC) and
compared to the classic reactor-separation-recycle process (at a capacity of 100 ktpy DME,
over 99.99% purity), in terms of key performance indicators (e.g. CapEx, OpEx, and specific
energy usage). Rigorous design and simulations were performed in Aspen Plus, including
accurate LHHW kinetics for both the gas-phase and liquid-phase reactions as well as rigorous
phase equilibrium calculations.

3
Optimal design of intensified processes for DME synthesis

2. Problem statement
Due to the ever increasing demand in cheap chemicals and alternative fuels, the industry
strives to deliver DME at even lower costs using the current plants (optimized and revamped)
or new plants based on process intensification technologies. Most of the DME technologies
described in literature (Azizi et al., 2014) are hindered by the incomplete conversion of
methanol which requires costly downstream processing, recovery and large recycle of
methanol (up to 40% of the fresh methanol feed). To solve this problem, this study proposes
DME process alternatives based on reactive distillation, which allows the complete
conversion of methanol (and avoids recycle streams) at competitive costs.

3. Simulation approach
Rigorous simulations were performed in Aspen Plus for each of the process alternatives
investigated in this study, considering a plant capacity of 100 ktpy DME. This section
provides details on the thermodynamics and kinetic parameters, as well as the methodology
used for the process optimization (i.e. minimization of the total annual costs) and economic
evaluation (e.g. estimation of capital and operating expenditures).

3.1 Physical properties

The non-ideality of the liquid phase was modeled using the UNIQUAC liquid activity model.
The methanol-water and methanol-DME binary interaction parameters are available in the
Aspen Plus database, while the water-DME binary interaction parameters were taken from the
NIST database. The non-ideality of the vapor phase was modeled by the Redlich-Kwong
equation of state. Redlich-Kwong EOS requires critical properties of the pure components,
which are also available in the Aspen Plus database. The property model was also validated
against experimental data reported in literature (Teodorescu and Rasmussen, 2001; Ihmels
and Lemmon, 2007; Wu et al, 2011). Figure 1 shows the residue curves map (RCM) and the
ternary map of the DME-methanol-water mixture (Kiss et al., 2012). Fortunately, there are no
azeotropes present in this system, but there is a small region where phase splitting is possible.

3.2 Reaction kinetics

DME is formed in the reversible chemical reaction of methanol dehydration, either in gas-
phase (catalyzed by -alumina) or in liquid-phase (catalyzed by ion exchange resins).

 CH3OCH3  H 2O
2CH3OH 
 (1)

4
Optimal design of intensified processes for DME synthesis

Gas-phase reaction. The reaction rate of the gas phase methanol dehydration, on a
commercial -Al2O3 catalyst, was described by the following kinetics. The apparent kinetic
parameters provided by Berčič and Levec (1993) were used (the original apparent rate
constant ks was increased by a factor 100, in order to match the effectiveness factor presented
in the same paper). Other similar kinetics are reported in literature (Mollavali et al., 2008).
 y y 
ks K M2  yM2  W D 
 K eq
r   (2)
1  2 
4
K M yM  KW yW

With the following constants:

ks = 3.30×109 exp(-10800/T), kmol/(kg∙h)
KM = 0.72×10-2 exp(830/T), dimensionless
KW = 0.45×10-2 exp(1130/T), dimensionless
The equilibrium constant Keq was calculated (by Aspen Plus) from Gibbs free energies.

Liquid-phase reaction. The reaction rate of the liquid phase methanol dehydration on
Amberlyst-35 was described by the following rate equation – using the kinetic parameters
from Hosseninejad et al. (2012):

ks  1 cD cW 
r 2 
1  (3)
  2
KW cW   K eq cM 
1  
 K M cM 
With the following constants:
ks = 6.12×109exp(-11793/T), kmol / (kg∙s)
KW/KM = exp(-6.46+2964.0/T)
Keq = exp(-2.6305+2787/T), regressed from equilibrium constant (Keq) values calculated by
Aspen Plus from Gibbs free energies.

3.3 Optimization and economics

Optimizing a chemical process is usually a mixed-integer nonlinear problem that is non-
convex and likely to have multiple locally optimal solutions. Hence a guarantee of
convergence to the globally optimal solution is not possible for non-convex problems. More
details about the decision variables, constrains and solution method are provided later, within
the appropriate context. In this work, each process design is optimized using the minimization

5
Optimal design of intensified processes for DME synthesis

of the total annual cost (TAC) as objective function:

CapEx
TAC  OpEx  (4)
payback period
A payback period of 3 years was used, and a running time of 8000 hours/year was considered.
The following heating and cooling costs were taken into account: high-pressure (HP) steam
(42 bar, 254 °C, $9.88/GJ), medium-pressure (MP) steam (11 bar, 184 °C, $8.22/GJ), low-
pressure (LP) steam (6 bar, 160 °C, $7.78/GJ), and cooling water ($0.72/GJ). These costs of
utilities are typical for a US plant (Luyben, 2011), but they might differ for other locations.
The total investment costs (CapEx) include the reactor, all heat exchangers, and distillation
columns. The cost of the heat exchangers (reboiler, condenser, FEHE, steam generator,
cooler) are given by (Turton et al., 2009):


CHEX (US $)   M & S / 280    474.7  A0.65  2.29  Fm  Fd  Fp   (5)

C furnace (US $)   M & S / 280   15668  Q0.85  1.27  Fc  (6)

where M&S is the Marshall & Swift equipment cost index (M&S=1536.5 in 2012), A is the
area (m2), Fm = 1 (carbon steel), Fd = 0.8 (fixed-tube), Fp = 0 (less than 20 bar). A heat
transfer coefficient U=0.5 kW/m2/K was assumed to calculate the heat transfer area. For the
reboilers, the design factor was taken as Fd = 1.35. For the furnace, Q is the duty (in MW),
while the correction factor was taken as Fc = 1 (process heater, carbon steel, design pressure
less than 40 bar).
The distillation columns diameter (D) were obtained by the tray sizing utility from Aspen
Plus, while the height was evaluated from the number of trays (NT), as: H = 0.6∙(NT - 1) + 2
(m). Afterwards, the cost of the columns shell was calculated as:
Cshell (US $)   M & S / 280    957.9  D1.066  H 0.82    2.18  Fc  (7)

The cost of the trays was given by (Turton et al., 2009):

Ctrays (US $)  NT   M & S / 280   97.2  D1.55   Ft  Fm  (8)

with Ft = 0 (sieve trays) and Fm = 1 (carbon steel)

where Fc = Fm∙Fp, Fm = 1 (carbon steel) and Fp  1  0.0074   P  3.48  0.00023   P  3.48

2

The price of the structured packing (KATAPAK-SP 11) was taken as 10,000 $/m3. The
relationship (7) was also used to estimate the cost of the tubular reactor. For the solid acid
catalyst (Amberlyst-35 ion exchange resin, with a bulk density of 560 kg/m3) and the gas-

6
Optimal design of intensified processes for DME synthesis

phase catalyst (-alumina, bulk density 882 kg/m3, particle size 3 mm), a purchased cost of 10
$/kg was considered. Clearly, the price of catalyst differs per country and manufacturer.

4. Results and discussion

This section provides the simulation results for the three DME process options considered:
reactor-separation-recycle (R-S-R) process, single step process using a reactive distillation
column, and combined gas-phase reactor and reactive distillation process.

4.1 Reactor-separator-recycle process

In the conventional R-S-R process, methanol is dehydrated over a γ-alumina catalyst (or
HZSM-5, silica-alumina, phosphorous-alumina and fluorinated-alumina) to produce DME.
The fixed-bed gas-phase reactor is operated at 250-400C and pressure up to 20 bars. Due to
incomplete conversion, a direct sequence of two distillation columns is required to deliver
high-purity (min. 99.99 %wt) DME (Muller and Hubsch, 2005; Kiss and Ignat, 2013).
Figure 2 shows the process flowsheet, mass balance and the key design parameters of the
reactor-separation-recycle process for DME production. The fresh methanol feed is mixed
with the recycled methanol stream and fed via a feed-effluent heat exchanger (FEHE) to a
furnace, which vaporizes the whole stream. Methanol vapor is converted in the reactor to
DME and water, but the conversion is incomplete. The hot reactor outlet stream is used in
FEHE to pre-heat the feed to the reactor, and then the remaining heat of the hot stream is used
for steam generation (which can be used for the reboiler of methanol recovery column). The
outlet of the steam generator is fed directly, as vapor, to the DME distillation column (COL-1).
Note that further cooling to liquid at boiling point (about 64 ˚C) increases the TAC, as an
additional heat exchanger is needed and transferring the heat to cooling water increases the
duty required for DME separation. In COL-1, the DME is separated as top distillation, while
the bottom product (mixture of methanol and water) is fed to the methanol recovery column
(COL-2), which separates methanol as distillate (a stream that is recycled in the process) and
water by-product as bottom stream.
For the conventional process, the choice of the reactor is a major design decision, as it
determines the achievable conversion. In general, approaching the chemical equilibrium
condition has a beneficial effect on the TAC, as the reduction of separation section CapEx and
OpEx overcome the expense of using more catalyst.
When a single-bed adiabatic reactor is used, the design decisions concern the reactor size
(diameter and length) and the reactor-inlet temperature. The diameter was calculated by

7
Optimal design of intensified processes for DME synthesis

imposing a fluid velocity u = 0.25 m/s, which is in line with the reported value of an industrial
DME reactor (Bai et al, 2013). Concerning the reactor length, a longer reactor (more catalyst)
leads to higher conversion and thus reduces the separation costs, but pressure drop increases
and the reactor is more difficult to build. Higher feed temperature increases the reaction rate
and therefore reduces the amount of catalyst required to achieve a certain conversion. On the
other hand, it requires more heat and reduces the achievable conversion due to the exothermal
nature of the chemical reaction. Moreover, it increases the temperature along the reactor bed,
with negative effect on catalyst stability. Figure 3 presents the TAC versus inlet temperature,
for reactors with different lengths. For each reactor, the optimal operating point (minimum
plant TAC) is marked, together with the values of the total annual cost, TCA*, total
investment cost TIC*, total operating cost (TOC*), pressure drop along the reactor, P, and
maximum temperature along the reactor bed, Tout. It can be observed that the reactor leading
to minimum TAC while fulfilling the constraint of outlet temperature below 400 C (when the
catalyst starts to deactivate) has the 12 m length, being fed at 275 °C. This will be considered
for further comparison with other design alternatives.
Figure 4 presents lines of constant reaction rate in the temperature – conversion plane. The
line of maximum reaction rate (at a given conversion) is represented by the dashed line. Also,
the figure shows the reaction trajectories corresponding to reactors of length 8 m, 12 m and 16
m, fed at the optimum temperature. It can be noticed that all the optimal reactors approach the
equilibrium, therefore the reduction of separation costs exceeds the cost of using more
catalysts.
While evaluating the TAC shown in Figure 3, the performance of the distillation columns was
specified in terms of distillate purity (99.99 %wt DME, COL-1) and bottoms purity (>99.95
%wt water, COL-2). The usual design condition of R = 1.2 Rmin (Luyben, 2011) was used,
while the location of the feed tray was chosen such that the minimum reboiler duty is
achieved. The specification for the methanol recycle purity (in the range 0.95 %wt to 0.999
%wt) had little influence on the optimal reactor design and minimum TAC value. The DME
column (COL-1) was operated at 10 bar, which allows use of cooling water in the condenser.
The FEHE increases the temperature of the reactor feed stream up to 140 ˚C, by using the
reactor effluent as heat source. Note that the amount of heat that can be recovered in the
FEHE is limited by the temperature crossover which occurs when higher outlet temperature of
cold stream is specified (Figure 5).
Table 1 and Table 2 list the simulation results of the R-S-R process for DME production,
while Table 3 provides a summary of the economic results. Note that the furnace contributes

8
Optimal design of intensified processes for DME synthesis

by 29% to CapEx and 61% to OpEx, so heat-recovery (using FEHE, steam generation and
vapour feed to the DME column) plays an important role in the process. The specific energy
requirements are 2.58 MJ/kg (714 kWh/ton) DME.

4.2 Reactive distillation process

RD conveniently combines reaction and separation in a single unit that can drive conversion
to completion by continuously removing the products from the system. In such processes,
designing a RD column at the maximum driving force results in an optimal design in terms of
controllability and operability, which is less sensitive to disturbances in the feed and has the
inherent ability to reject disturbances (Mansouri et al., 2015).
As shown in Figure 6, methanol is fed at the top of the reactive zone in the RD column (on
stage 9). The reaction takes place in the reactive zone where the solid acid catalyst is placed
(stages 9-42). DME is the lightest component in the system hence removed as top distillate,
while water is the heaviest component hence removed as bottoms. Unlike the gas-phase
reaction, the dehydration of methanol in liquid phase is catalyzed by thermally stable resins,
such as Amberlyst-35 which has high activity and selectivity at temperatures up to 150 °C.
The design of RD processes can be performed using single and multi-objective optimization
approaches (Segovia-Hernandez et al., 2015). The optimal design of the RD process (and the
next one using reactive distillation) is a mixed-integer nonlinear problem (MINLP). Such
problems are intrinsically very difficult to solve, and the solution time increases rapidly with
the number of variables and constraints.
The decision variables used in the optimization are the following:
 Discrete (integers): number of stages (NT), feed stage (NF), reactive stages (NR1 – NR2)
 Continuous (real numbers): reflux ratio (RR), pressure (P), amount of catalyst / tray (mcat)
The optimization considers the following constraints:
 Distillate purity: 99.99 %wt DME
 Bottoms purity: exceeding 99.95 %wt
 Temperature on the reactive stages: below 150 °C (to avoid catalyst deactivation)
 Catalyst must fit into the available space (no more than 20% of the packing volume)
Note that the optimal steady state operating point is often defined by the intersection of active
constraints (Kookos and Perkins, 2016). This well-established feature of chemical processes
is neatly exploited to simplify the solution of the optimization problem. Thus, the RD column
for DME synthesis has the following features, confirmed by several sensitivity analysis runs:

9
Optimal design of intensified processes for DME synthesis

 When the distillate rate is set to half of the feed (based on reaction stoichiometry), due to
the relative volatilities of the species involved in the process, methanol is equally
distributed between the distillate and the product streams. Therefore, achieving 99.99 %wt
distillate purity (DME) by manipulating the reflux ratio also ensures that the purity of the
bottoms stream (water) exceeds the 99.95 %wt requirement.
 The operating costs (OpEx) represent more than 75% of the TAC. The main contributions
to the capital costs (CapEx) are the condenser, reboiler, column shell and structured
packing, with the catalyst representing less 3%. For this reason, increasing the amount of
catalyst per theoretical stage invariably results in a lower value of the TAC, as the
reduction of the other costs clearly outweighs the cost of catalyst. However, the amount of
catalyst is limited to 20% of the available volume (Götze et al., 2001).
 Increasing the operating pressure leads to higher temperature along the RD column. This
results in faster reaction rates and, in turn, to lower TAC due to reduced effort necessary
to achieve the separation.
Based on these observations, the following optimization strategy is employed. In Aspen Plus,
set the discrete decision variables: number of stages, feed stage, and reactive stages. The
distillate molar rate is set to half of the feed, according to reaction stoichiometry. By means of
Design Specification blocks, the continuous decision variables (amount of catalyst on each
tray, pressure, and reflux ratio) are adjusted such that the constraints are fulfilled (distillate
purity 99.99 %wt DME; temperature on reactive stages < 150 °C; catalyst fits into the
available space of 20% of the packing volume). Due to mass balance and high purity of the
distillate, the purity of the bottoms stream always exceeds the 99.95 %wt requirement. The
simulation is run and the value of the TAC is obtained. Afterwards the discrete variables are
adjusted, and new runs are performed, until the minimum TAC is obtained. In this work, the
optimal values of the discrete variables are found using the genetic algorithm implemented in
Matlab by means of the ga function. The user-provided Matlab objective function uses the
COM interface to communicate with Aspen Plus: sends to Aspen Plus the values of the
discrete decision variables, requests the simulation run and checks the convergence. After a
successful run, the objective function (TAC, calculated in Aspen Plus by a FORTRAN
Calculator block) and the continuous decision variables (obtained by the Design Specification
blocks) are retrieved and saved in an Excel file together with the discrete decision variables.
Then, the value of the objective function is passed to the ga function, which performs a new
iteration. If Aspen Plus simulation does not converge, the simulation is re-initialized and a
new run is attempted. If this also fails, a large value is returned to ga as the value of the

10
Optimal design of intensified processes for DME synthesis

objective function. During an initial run, rather large ranges of the discrete variables were
assumed and many Aspen Plus simulations were unsuccessful. After a better estimation of the
optimum, the ranges were restricted as follows: feed tray: 7…12; first reactive stage: 7…12;
last reactive stage 30…50; total number of stages: 45 … 60. In this case, more than 95% of
the simulations were successful. The progress of the optimization run is shown in Figure 7.
Figure 6 also shows the process flowsheet, mass balance and the key operating parameters,
while Table 4 provides the optimal design parameters of the RD column for DME production.
The operating pressure (11.36 bar) leads to reactive stages temperature in the range 130-150
°C. The amount of catalyst (193.2 kg/stage) occupies 20% of the stage volume, this being in
line with the recommendations for KATAPAK packing (Götze et al., 2001). The reflux ratio,
(6.174 kmol/kmol) ensures the required purity. Along the reactive stages, the liquid-phase
methanol mole fraction exceeds 80%, which leads to a large reaction rate. For convenience,
Figure 8 plots the temperature and reaction rate profiles while Figure 9 shows the liquid
composition profiles along the RD column. Remarkable, for the RD process, the CapEx is
only 2395 k$, the OpEx is 2604 k$/year, and the specific energy requirements are 2.43 MJ/kg
(672.5 kWh/t) DME.

4.3 Combined gas-phase reactor and reactive distillation

The main idea of the combined process is to use only the reaction section of existing R-S-R
processes for DME production, and include a reactive distillation column in the downstream
processing, as shown in Figure 10. This allows complete reactant conversion and avoids the
need of recycling. The reactor diameter decreases, but the effect on the TAC is small.
However, it should be remarked that the gas-phase reactor of existing DME plants can process
more fresh methanol (for example, the R-S-R process illustrated in the previous section could
process Figure 2748 kmol/hr instead of 546 kmol/hr). The methanol conversion is eventually
driven to completion by the RD column. As the feed to the downstream processing section is
not pure methanol (as in the single step RD process), but a ternary mixture (DME, methanol
and water from the gas-phase reactor), an additional unit is necessary. Because water has the
highest boiling point (and therefore in the RDC column will remain in the liquid phase) and
has a detrimental effect on the reaction rate (see equation 3), this component is removed in the
first distillation column. The water column was designed for 99% water recovery, at 99.97
%wt purity. The methanol – DME mixture is fed above the reactive section of the RDC.
In the RD column, the amounts of catalyst (67 kg/stage), is the maximum occupying 20% of
the available space. Despite the higher methanol liquid phase concentration, the reaction rate

11
Optimal design of intensified processes for DME synthesis

is lower compared to the single RD Column process due to less catalyst. Along the reactive
stages, the temperature is in the range 134-147 °C. Figure 11 and Figure 12 present
temperature, reaction rate and composition profiles along the RD column. Table 5 and Table 6
list the simulation results of the combined process for DME production, while Table 7
provides a summary of the economic results. For the same production capacity (100 ktpy) the
cost of the reaction section is lower as compared to the classic R-S-R process due to the
absence of the methanol recycle. However, the downstream processing section is somewhat
more expensive than the classic direct distillation sequence, as the reaction must also be
carried out together with the separation. Moreover, the reactor outlet is not fed as vapor to the
RD column (due to the reaction taking place in liquid phase) but as saturated liquid
(condensed in the prefractionator). As a consequence, the key economic indicators are slightly
increased: CapEx is 3437 k$, OpEx is 3152 k$/year and the specific energy requirements are
2.905 MJ/kg (807 kWh/ton) DME.

5. Process comparison
The economic results are summarized for all cases in Table 3 (R-S-R process alternatives),
Table 4 (single step process in RD column) and Table 7 (combined process). Based on these
figures, it is clear that for a new DME plant the single step process using a reactive distillation
column is the best option with the lowest CapEx (about 70% of the classic R-S-R process)
and similar OpEx, with a reduction of 6% in the specific energy requirements.
However, in case of increasing the capacity of an existing DME plant, it would be cheaper to
just revamp the plant by keeping the whole reaction section and reuse the columns of the
direct distillation sequence to separate water (first column) and convert the remaining MeOH
to DME (second column, reactive distillation). By doing so, the additional investment is
minimized and the TAC reduced in spite of the slight increase in OpEx. Another key
advantage of the combined process is possibility to significantly increase the plant capacity
due to the absence of the methanol recycle.

6. Conclusions
The rigorous process simulations showed that the novel process alternatives based on reactive
distillation can significantly improve the classic reactor-separator-recycle process for DME
production. Based on the results presented here, the following conclusions can be drawn:
 The classic reactor-separator-recycle DME process could be improved by feeding the
reactor outlet as vapor to the direct distillation sequence (about 8% savings in TAC).

12
Optimal design of intensified processes for DME synthesis

 The single step process using a reactive distillation column is the most promising for
new DME plants due to the lowest CapEx (2395 k$) and OpEx (2604 k$/year) - for
100 ktpy plant capacity - as well as specific energy requirements (2.43 MJ/kg DME).
 The combined process (gas-phase reactor + reactive distillation) is not recommended
for new plants (due to higher costs than the single step RD process), but it is suitable
for revamping existing DME plants. With only a minor additional investment, the
plant capacity can be significantly increased by having complete methanol conversion
and no recycles, thus allowing a higher processing capacity of the fresh methanol feed.

Acknowledgement
Financial support of the European Commission through the European Regional Development
Fund and of the Romanian state budget, under the grant agreement 155/25.11.2016 (Project
POC P-37-449, acronym ASPiRE) is gratefully acknowledged.

References
1. Agreda V. H., Partin L. R., Heise W. H., High-purity methyl acetate via reactive
distillation, Chemical Engineering and Processing, 86 (1990), 40-46.
2. An W., Chuang K., Sanger A., Dehydration of methanol to dimethyl ether by catalytic
distillation, Canadian Journal of Chemical Engineering, 82 (2004), 948-955.
3. Arcoumanis C., Bae C., Crookes R., Kinoshita E., The potential of di-methyl ether (DME)
as an alternative fuel for compression-ignition engines: A review, Fuel 87 (2008), 1014-
1030.
4. Azizi Z., Rezaeimanesh M., Tohidian T., Rahimpour M. R., Dimethyl ether: A review of
technologies and production challenges, Chemical Engineering and Processing: Process
Intensification, 82 (2014), 150-172.
5. Bai Z., Ma H., Zhang H., Ying W., Fang D., Process simulation of dimethyl ether
synthesis via vapor phase dehydration. Polish Journal of Chemical Technology, 15 (2013),
122-127.
6. Bai Z., Ma H., Zhang H., Ying W., Fang D., Simulation of a multi-stage adiabatic reactor
with inter-stage quenching for dimethyl ether synthesis. Chemical Industry & Chemical
Engineering Quarterly, 20 (2014), 481-490.
7. Berčič G., Levec J., Catalytic dehydration of methanol to dimethyl ether. Kinetic
investigation and reactor simulation, Industrial & Engineering Chemistry Research, 32
(1993), 2478-2484.

13
Optimal design of intensified processes for DME synthesis

8. Dimian A.C., Bildea C.S., Kiss A.A., Integrated design and simulation of chemical
processes, 2nd edition, Elsevier, Amsterdam, 2014.
9. Farsi M., Jahanmiri A., Eslamloueyan R., Modeling and optimization of MeOH to DME
in isothermal fixed-bed reactor, International Journal of Chemical Reactor Engineering, 8
(2010), article A79.
10. Götze L., Bailer O., Moritz P., von Scala C., Reactive distillation with KATAPAK®,
Catalysis Today, 69 (2001), 201-208.
11. Han Y., Zhang H., Ying W., Fang D., Modeling and simulation of production process on
dimethyl ether synthesized from coal-based syngas by one-step method. Chinese Journal
of Chemical Engineering, 17 (2009), 108-112.
12. Harmsen J., Process intensification in the petrochemicals industry: Drivers and hurdles for
commercial implementation, Chemical Engineering and Processing, 49 (2010), 70-73.
13. Hossenininejad S., Afacan A., Hayes R. E., Catalytic and kinetic study of methanol
dehydration to dimethyl ether. Chemical Engineering Research and Design, 90 (2012),
825-833.
14. Huss R.S., Chen F., Malone M.F., Doherty M.F., Computer-aided tools for the design of
reactive distillation systems, Computers & Chemical Engineering, 23 (1999), S955-S962.
15. Ihmels E. C., Lemmon E. W., Experimental densities, vapor pressures, and critical point,
and a fundamental equation of state for dimethyl ether, Fluid Phase Equilibria, 260
(2007), 36-48.
16. Kansha Y., Ishizuka M., Song C., Tsutsumi A., Process intensification for dimethyl ether
production by self-heat recuperation, Energy, 90 (2015), 122-127.
17. Kiss A.A., Advanced distillation technologies - Design, control and applications. Wiley,
Chichester, UK, 2013.
18. Kiss A. A., Bildea C. S., Dimian A. C., Design and control of recycle systems by non-
linear analysis, Computers & Chemical Engineering, 31 (2007), 601-611.
19. Kiss A. A., Bildea C. S., A review on biodiesel production by integrated reactive
separation technologies, Journal of Chemical Technology and Biotechnology, 87 (2012),
861-879.
20. Kiss A. A., Ignat R. M., Revamping dimethyl ether separation to a single-step process,
Chemical Engineering & Technology, 36 (2013), 1261-1267.
21. Kiss A. A., Pragt J. J., Vos H. J., Bargeman G., de Groot M. T., Novel efficient process
for methanol synthesis by CO2 hydrogenation, Chemical Engineering Journal, 284
(2016), 260-269.

14
Optimal design of intensified processes for DME synthesis

22. Kiss A. A., Suszwalak D. J-P. C., Innovative dimethyl ether synthesis in a reactive
dividing-wall column, Computers & Chemical Engineering, 38 (2012), 74-81.
23. Kookos I.K, Perkins J.D., Control structure selection based on economics: Generalization
of the back-off methodology, AIChE Journal, 62(2016), 3056-3064.
24. Lei Z., Zou Z., Dai C., Li Q., Chen B., Synthesis of dimethyl ether (DME) by catalytic
distillation, Chemical Engineering Science, 66 (2011), 3195-3203.
25. Li R., Zhang H. T., Ying W. Y., Fang D. Y., Research on the intrinsic kinetics of
methanol dehydration to dimethyl ether. Natural Gas Chemical Industry, 39 (2009), 9-12.
26. Luyben W. L., Principles and case studies of simultaneous design, AIChE Wiley,
Hoboken, 2011.
27. Malone M., Huss R., Doherty M., Green chemical engineering aspects of reactive
distillation, Environmental Science & Technology, 37 (2003), 5325-5329.
28. Mansouri S. S., Sales-Cruz M., Huusom J. K., Woodley J. M., Gani R., Integrated process
design and control of reactive distillation processes, IFAC-PapersOnLine, 48 (2015),
1120-1125.
29. Mollavali M., Yaripour F., Atashi H., Sahebdelfar S., Intrinsic kinetics study of dimethyl
ether synthesis from methanol on γ-Al2O3 catalysts, Industrial & Engineering Chemistry
Research, 47 (2008), 3265-3273.
30. Muller M., Hubsch U., Dimethyl ether, in Ullmann's Encyclopedia of Industrial
Chemistry, 7th Edition, Wiley-VCH, Weinheim, 2005.
31. Olah G. A., Goeppert A., Surya Prakash G. K., Chemical recycling of carbon dioxide to
methanol and dimethyl ether: From greenhouse gas to renewable, environmentally carbon
neutral fuels and synthetic hydrocarbons, The Journal of Organic Chemistry, 74 (2009),
487-498.
32. Patrut C., Bildea C. S., Kiss A. A., Catalytic cyclic distillation - A novel process
intensification approach in reactive separations, Chemical Engineering and Processing:
Process Intensification, 81 (2014), 1-12.
33. Schoenmakers H. G., Bessling B., Reactive and catalytic distillation from an industrial
perspective, Chemical Engineering and Processing, 42 (2003), 145-155.
34. Segovia-Hernandez J. G., Hernandez S., Bonilla-Petriciolet A., Reactive distillation: A
review of optimal design using deterministic and stochastic techniques, Chemical
Engineering and Processing: Process Intensification, 97 (2015), 134-143.
35. Sundmacher K., Kienle A. (Eds.), Reactive distillation: Status and future directions,
Wiley-VCH, Weinheim, 2003.

15
Optimal design of intensified processes for DME synthesis

36. Teodorescu M., Rasmussen P., High-pressure vapor-liquid equilibria in the systems
nitrogen plus dimethyl ether, methanol plus dimethyl ether, carbon dioxide plus dimethyl
ether plus methanol, and nitrogen plus dimethyl ether plus methanol, Journal of Chemical
and Engineering Data, 46 (2001), 640-646.
37. Turton R., Bailie R.C., Whiting W.B., Shaeiwitz J.A, Analysis, synthesis and design of
chemical processes, 3rd edition, Prentice Hall, USA, Appendix A, 2009.
38. Wu J. T., Zhou Y., Lemmon E. W., An equation of state for the thermodynamic properties
of dimethyl ether, Journal of Physical and Chemical Reference Data, 40 (2011), Article
Number: 023104.

16
 Optimal design of intensified processes for DME synthesis

Residue curve for DME/METHANOL/WATER

Mo
METHANOLTernary
(136.8 1Map
C) (Mole Basis)

le f
ER

0. 1

0. 9

rac
T

0. 2
WA

0. 8

M ET

0.8
0. 3
rac

0. 7

HA
le f

0. 4

0. 6

NO
Mo

0.6
L
0. 5

0. 5
0. 6

0. 4

0.4
0. 7

0. 3
0. 8

0.2
0. 2
0. 9

0. 1 WATER DME
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 (1 79.98 C) 0.10.20.30.40.50.60.70.80.9 (4 4.40 C)
Molefrac DME

Figure 1. Residue curve map and ternary diagram of the mixture DME-MeOH-H2O at 10 bar

Methanol recycle
F = 121.3 kmol/h Qc = 6.453 MW
xM = 0.999
64.2 °C, 1 bar
STEAM F = 273 kmol/h
xDME = 0.9999
GENERATOR 44.4 °C, 10 bar DME
2
219.2 °C 170 °C RR = 3.94
COL-1
Qc = 2.82 MW
D = 1.68 m
Q = 0.576 MW
10

REACTOR 2
FURNACE RR = 1.37
Methanol FEHE L = 12 m
275 °C D = 1.89 m 395.4 °C COL-2
F = 546 kmol/h 11.5 bar 17
33 °C 140 °C 12 bar
25 °C, 1 bar
D = 1.16 m
21
396 °C F = 667.3 kmol/h
Q=6.89 MW xM = 0.9998 F = 667.3 kmol/h 155.8 C 1.25 bar
QR = 0.625 MW 10.17 bar Water
xM = 0.182
xDME = 0.409
xW = 0.409 F = 273 kmol/h
29
xW = 0.9999
106.9 °C, 1.29 bar

QR = 2.04 MW

Figure 2. Reactor-separation-recycle process for DME production

17
Optimal design of intensified processes for DME synthesis

4400 L / [m] = 8

10
4200
TAC / [103 US$]

12

4000
16

3800 TAC* / [106 US$/yr] = 3.813 3.791 3.782 3.784

TIC* / [106 US$] = 3.594 3.417 3.334 3.252
TOC* / [106 US$/yr] = 2.615 2.652 2.671 2.700
P / [bar] = 0.54 0.46 0.38 0.30
Tout / [˚C] = 391.4 395 403.8 412
3600
250 260 270 280 290 300
T / [ C]
Figure 3. TAC vs reactor-inlet temperature, for reactors with various lengths. Total annual
cost, pressure drop and reactor-outlet temperature are indicated, for each optimal point.

1 u = 0.25 m/s u = 0.25 m/s u = 0.25 m/s

Equilibrium line L = 16 m L = 12 m L =8 m
D = 1.87 D = 1.89 m D = 1.93 m
TAC = 3.8 M$/yr TAC = 3.79 M$/yr TAC = 3.78 M$/yr
0.05
0.1
0.8 0.2
0.4
0.8 Line of maximum
0.6 1.6 reaction rate

3.2
X

0.4 6.4

0.2
parameter:
reaction rate / [10-3 kmol/(kg∙h)]
0
250 300 350 400 450
T / [ C]
Figure 4. Methanol conversion versus temperature, at fixed reaction rate.

18
Optimal design of intensified processes for DME synthesis

400 400
350 350
300 300
250 250
200 200

T / [˚C]
T / [˚C]

Tcold,out = 140 ˚C Tcold,out = 150 ˚C

150 150
100 100
50 50
0 0
0 500 1000 1500 2000 2500 0 2000 4000 6000 8000
Q / [kW] Q / [kW]

Figure 5. T–Q curve of the feed-effluent heat exchanger (FEHE), for two different values of
the cold stream outlet temperature.

Water
F = 273 kmol/h
1 P = 11.36 bar xW = 0.9997 %wt
35 °C, 1 bar

QC= 9.15 MW
AC = 755 m2
Methanol 2
RR = 6.177 DME
F = 546 kmol/h
F = 273 kmol/h
25 °C, 12 bar 85.4 °C
xDME = 0.9999 %wt
9
50.6 °C, 11.36 bar
QFEHE= 0.943 MW
AFEHE = 132.2 m2
RDC

D = 2.1 m

42
QR= 8.49 MW
AR = 248.8 m2
53
11.42 bar 1 bar
54 185 °C 99.6 °C

Figure 6. Single step process using a reactive distillation column for DME production

19
Optimal design of intensified processes for DME synthesis

8000
7000
Average
6000
TAC / [k$/year]

5000
4000
Best
3000
2000
1000
0
0 50 100 150 200
Iteration
Figure 7. Progress of the RDC optimization

200 10
Feed stage
r T
Reaction rate / [kmol/h]
8
Temperature / [°C]

150

6
100
4

50
2

0 0
0 5 10 15 20 25 30 35 40 45 50
Stage
Figure 8. Temperature and reaction rate profile along the RD column

20
Optimal design of intensified processes for DME synthesis

1
Methanol
0.8
Mole fraction

0.6

0.4
Water
DME
0.2

0
0 10 20 30 40 50
Stage
Figure 9. Liquid (continuous line) and vapor (dashed line) composition profiles along the RD
column

Water

F = 273 kmol/h
xw = 0.9997

Qc = 7.12 MW
Qc = 3.82 MW
1 DME + MEOH 1

STEAM 5 bar 11 bar 10 bar

GENERATOR DME
208 °C 170 °C RR=1.867 D1 = 1.24 m F = 273 kmol/h
RR=1.937 xDME = 0.9999
D = 1.38m 8
F = 324.8 kmol/h
COL-1 xM = 0.305
Q = 0.36 MW xDME = 0.688
9
RDC

FURNACE REACTOR 25
FEHE
275 °C 395°C F = kmol/h
Methanol bar xW = 0.9997
546 kmol/h
25 °C 140 °C 12 bar QR = 3.16 MW
18 1 bar

395 °C L = 12 m 35
Q= 1.98 MW D = 1.71 m Water
QR = 1.727 MW

F = kmol/h
xM = 0.18
xDME = 0.41
xW = 0.41

Figure 10. Combined gas-phase reactor and reactive distillation process for DME production

21
Optimal design of intensified processes for DME synthesis

200 10
180

Reaction rate / [kmol/h]

160 8
Temperature / [°C]

140
120 6
T
100
80 r 4
60
40 2
20
0 0
0 5 10 15 20 25 30 35
Stage
Figure 11. Temperature and reaction rate profiles along the RD column, in the combined gas-
phase reactor and RDC process

1
Water
DME Methanol
0.8
Mole fraction / [°C]

0.6

0.4

0.2

0
0 5 10 15 20 25 30 35
Stage
Figure 12. Liquid (continuous line) and vapor (dashed line) composition profiles along the
RD column, in the combined gas-phase reactor and RDC process

22
Optimal design of intensified processes for DME synthesis

Tables

Table 1. Simulation results for the reactor-separator-recycle process for DME production
(100 ktpy): reaction system (without downstream processing)

Parameter / unit Value

Reactor
Length / [m] 12
Diameter / [m] 1.89
CapEx [k$] 754.48

FEHE
Area / [m] 12.32
Duty / [MW]Tab 2.29
CapEx [k$] 41.18

Furnace
Duty [MW] 6.88
CapEx / [k$] 1006
OpEx / [k$/year] 1983.7

Steam generator
Duty / [MW] 0.576
Area / [m2] 33.3
CapEx / [k$] 104.1
OpEx / [k$/year] -129.21

23
Optimal design of intensified processes for DME synthesis

Table 2. Simulation results for the reactor-separator-recycle process for DME production
(100 ktpy): downstream processing (direct distillation sequence)

DME Methanol
Column Column
Feed condition
Feed rate / [kmol/h] 667.25 394.25
Feed temperature / [°C] 170.0 86.5

Design data
Number of stages, NT 18 30
Feed stage, NF 10 21
Diameter / [m] 1.68 1.16
Height / [m] 11.6 18.8
Reboiler area / [m2] 44.5 77.0
Condenser area / [m2] 681.1 144.7

Operating data
Distillate rate / [kmol/h] 273 121.27
Bottoms rate / [kmol/h] 394.3 273
Reflux ratio 3.94 1.37
Reboiler duty / [MW] 0.625 2.044
Condenser duty / [MW] 6.45 2.82
Condenser pressure / [bar] 10 1
Distillate purity / [%wt] 99.99 99.95
Bottoms purity / [%wt] - 99.98

Economic data
CapEx / [k$] 911.4 601.0
Condenser / [k$] 558.9 204.2
Reboiler / [k$] 111.7 159.2
Column shell / [k$] 221.6 218.2
Trays / [k$] 19.2 18.9

OpEx / [k$/year] 282.2 516.6

Cooling / [k$/year] 133.9 58.5
Heating (HP steam] / [k$/year] 148.2 458.1
TAC / [k$/year] 586.0 716.9

24
Optimal design of intensified processes for DME synthesis

Table 3. Summary of economic results for the R-S-R process for DME production (100 ktpy)
CapEx / [k$] OpEx / [k$/year] TAC / [k$/year]
Reactor 754.5 - 251.5
FEHE 41.2 - 13.7
Furnace 1006.4 1983.7 2319.2
Steam generator 104.1 -129.2 -94.5
DME column 910.8 281.9 585.5
Methanol column 600.6 516 716.2
TOTAL 3417.6 2652.4 3791.6

25
Optimal design of intensified processes for DME synthesis

Table 4. Optimal design parameters of a RD column for DME production (100 ktpy)
Parameter / unit Value
Design data
Number of stages, NT 54
Feed stage, NF 9
Reactive stages, NR1 – NR2 9 - 42
HETP / [m] 0.5
Amount of catalyst / [kg/stage] 193.2
Catalyst volumetric fraction 0.2
Diameter / [m] 2.1
Height / [m] 28.0
Reboiler area / [m2] 248.8
Condenser area / [m2] 755.0
Feed preheater / [m2] 132.3

Operating data
Feed rate / [kmol/h] 546
Feed temperature / [°C] 85.4
Distillate rate / [kmol/h] 273
Bottoms rate / [kmol/h] 273
Reflux ratio 6.177
Reboiler duty / [MW] 8.485
Condenser duty / [MW] 9.148
Feed preheater / [MW] 0.943
Condenser pressure / [bar] 11.36
Distillate purity / [%wt DME] 99.99
Bottoms purity / [%wt water] 99.97

Economic data
CapEx / [k$] 2394.9
Condenser / [k$] 597.6
Reboiler / [k$] 342.2
Column shell / [k$] 578.4
Packing / [k$] 586.5
Catalyst / [k$] 65.7
Trays / [k$] 31.9
Feed pre-heater 192.6

OpEx / [k$/year] 2604.2

Cooling / [k$/year] 189.7
Heating (HP steam] / [k$/year] 2414.5
TAC / [k$/year] 3402.5

26
Optimal design of intensified processes for DME synthesis

Table 5. Simulation results for the combined gas-phase reactor and reactive distillation
process for DME production (100 ktpy): gas-phase reactor section
Parameter / unit Value
Reactor
Length / [m] 12
Diameter / [m] 1.71
CapEx [k$] 231.8

FEHE
Area / [m] 11.8
Duty / [MW] 1.981
CapEx [k$] 7.65

Furnace
Duty [MW] 5.63
CapEx / [k$] 848.66
OpEx / [k$/year] 1671.55

Steam generator
Duty / [MW] 0.366
Area / [m2] 25.1
CapEx [k$] 86.54
OpEx [k$/year] -82.08

27
Optimal design of intensified processes for DME synthesis

Table 6. Simulation results for the combined gas-phase reactor and reactive distillation
process for DME production (100 ktpy): downstream processing section
Parameter / unit Water-column RD column
Feed condition
Feed rate / [kmol/h] 546 324.8
Feed temperature / [°C] 164.2 66.6

Design data
Number of stages, NT 18 35
Feed stage, NF 9 8
Reactive stages - 8 - 25
Diameter / [m] 1.25
Height / [m] 1.38 21.8
Reboiler area / [m2] 112 90.8
Condenser area / [m2] 1066 386.4
Catalyst / [kg/stage] - 67
Catalyst volumetric fraction [%] - 0.2

Operating data
Distillate rate / [kmol/h] 324.8 273
Bottoms rate / [kmol/h] 221.2 51.8
Reflux ratio 1.867 1.937
Reboiler duty / [MW] 1.727 3.159
Condenser duty / [MW] 7.118 3.822
Condenser pressure / [bar] 5 10.2
Distillate purity / [%wt] 0.003 (Water), 0.9999 (DME)
Bottoms purity / [%wt] 0.9997 (Water) 0.9997 (Water)

Economic data
CapEx / [k$] 1143.8 956.7
Condenser / [k$] 747.9 386.8
Reboiler / [k$] 203.6 177.8
Column shell / [k$] 177.5 270.2
Trays / [k$] 14.8 13.3
Packing / [k$] - 108.6

OpEx / [k$/year] 556.6 1006.7

Cooling / [k$/year] 147.6 79.3
Heating (HP steam] / [k$/year] 409.0 927.4
TAC / [k$/year] 936.8 1325.6

28
Optimal design of intensified processes for DME synthesis

Table 7. Summary of economic results for the combined gas-phase reactor and reactive
distillation process for DME production (100 ktpy)
CAPEX / [k$] OPEX / [k$/year] TAC / [k$/year]
Reactor 231.8 - 77.3
FEHE 7.65 - 2.55
Furnace 848.67 1671.55 1954.44
Steam generator 86.54 -82.08 -53.23
Water column 936.8 556.6 868.86
RD Column 1325.6 1006.7 1448.56
TOTAL 3437.06 3152.77 4298.4

29