I NTRODUCTION

Opnmi.zanon of Design & Opemllon rn Reactn>e Distd/atton 1

CHAPTER I

INTRODUCllON

Th~ opti1nizatiun of design & opr.rnt;on. of reactive

distillation processes is in vestigated in th is Ph.D.

thesis. Thro technologicul lilld scientific scenarios used

in this resear c h a r e dcs<.;ribed in this c h apter. First,

t he drivers for c hange tn the current dynam ie

c nvironmc-:nt of chemical processing industry arr:

iden t ified . Then the reactive d ist illation processing is

introduced. The gt:mer alities of this process togeth er

with its technical challtmges in desi~n and operation

are ~ddresscd with the: d evelopment of a reactive

distillation process for the production of methyl aceta t ~

are presented The scientific setting of conreptuw

design & open:tliun in process systPms engineering,

with an emphi.\sis on the key cht:1.1lcnges in the:: <..les1gn

of r eactive distillation is addressed. The c h apte1· is

concluded with a thesis outline tlnd a concise..:

description of the scientific novc;;lty of this rc::scarch.

Optimizotinn of Design & Operation in Rco ctiv<: Di.still<AtWn 2

1. 1 A CHANGING ENVIRONMENT FOR THE C HEMICAL P ROCESS I NDUSTRY

ThP. chemical process indu~try is subject to a rapidly changing

environment, characterized by slim profit margins and fierce

competitiveness. Rapid changes are not exclusively found in the demands of

society for new, high quality, safe, clean and environmentally benign

products, they can be found in the dynamics of business operations, which

include global operations, competition and strategic alliances mapping,

among others.

Being able to operate at reduced costs with increasingly shorter time­

to-market times is the common denominator of successful companies;

however, attaining this performance level is not a straight forward or trivial

issue. Success is dependant on coping effectively with dynamic

environments and short process development and design times. Taking

into account life span considerations of products and processes is becoming

essential for development and production activities. Special attention

needs to be pa.id the potential losses of resources over the process lite span.

Since these resources differ in nature, for example they can be capital, raw

materials, labor, and energy. Implementing this life-span aspect is a

challenge for the chemical industry.

Moreover, manufacturing excellence practice needs to be pursued,

with a stress on the paramount importance of stretching profit margins,

while maintaining safety procedures. In addition, society is increasingly

demanding sustainable processes and products. It is no longer innovative

to say that the chemical industry needs to take into account biospheres

sustainability. Closely related to sustainable development, risk

minimization, another process aspect, must also be taken inlo

consideration. In today's world, processes and products must be safe for

their complete life span. Major incidents such as F'lixborough (1974) with

28 casualties and Bhopal (1984) with 4000+ casualties may irreversibly

affect society's perception of the chemical industry and should be a thing

of the past.

Opbmizallon o//Nsign & Opera!Jon 111 Rmaiw Dl:>'tl1/a1Jon 3

Addressing all these process aspects, given the underlying aim of

coping effectively with the dynamic environment of short process

development and design llmes, has resulted in a wide set of techniraJ

responses. Examples of these responses includl" advanced process control

strategies and real-time optimization. Special attention is paid to t ht­

synthesis of novel uml operations that can integrate several functions and

units to give substantial increases in process and plont efficiency.

These operations nrc conventionally re-fcrred to as hybrid and

mtrn!<ifie-d units, respectively and arc characterized by reduced costs and

process complexity. Reactive Distillation (RD) is an example of such an

operation.

1 .2 REACTIVE DISTILLATION P OTENTIAL

1 .2. l M AIN FEATURES & SUCCESS STORIES

Reactive Dist illation is n promising operation wherein reaction and

separation takes place within a smgle distillation column. The synergislif'

effcct of this combination has the potential to increase conversion,

improve seltictivity and facilitate separation Lask. The performance of

rc-action with separation in one piece of equipment offers distinct

advantages over the conventional, sequential approach. Especially for

equilibrium limited reactions such as esteri fication and ester hydrolysis

reactions, conversion can be increased far beyond chemical <>quilibrium

conversion due to the contmuous removal of reaction products from the

reactive zone. This may lead to an enormous reduction of capital and

investment costs anc! mRy be important for sustainable devtilopmcnt due

lo a lower consumption of resources. Reactive distillation has many

advantages over the sequential processtng, such as affixed bed reactor

followed by a fractionating column, in which the disti llate or bottom

fraction is recycled back lo the reactor inlet. The most important

adv1tnt.1ge m the use of Rcu.ctive distillntion is for the equihhrium limit~d

reactions is the elimination of conversions limitations by the continuous

Optimization of Design & Operation in Reactive Distillation 4

removal of the product. Apart from the increased conversions following

benefits can be obtained.

o The reduction in capital inves tment due to process integration, wh ich

reduces cost on piping, pumps and instrumentation.

o Savings in energy cost in case of exothermic reaction as the h eat of

reaction can be u sed for vaporizing the liquid.

o The h ot spot formation on the cata lyst is limited as the maximum

operating temperature in the reaction zone is limited to the boiling point

of th e liquid. Longer catalyst lifetimes as compared to conventional

systems.

Before applying Reactive distillation technology on industrial scale

three constraints are to be fulfilled.

• The temperature window of the vapor liquid equilibrium must be

equivalent to the reaction temperature . However the th ermal stability

of the catalyst can limit th e upper operating tempera ture of th e

distillation column.

The necessity of the wet catalyst pellet makes it compulsory to

conduct the reaction in liquid phase only.

Catalysts with long lifetimes are recommended.

1.2.2 The Basics

Reactive distillation is used with reversible, liquid phase reactions. Suppose a reversible reaction h ad the following chemical equation:

A+B~C+D ... ( 1. 1)

For many reversible reactions the equilibrium point lies far to the left and little product is formed :

A+B -.:=- C+D ... (1 .2)

However , if one or more of th e products a re removed more of the product will be formed because of Le Chatlier 's Principle:

A+B~C+~ ... (1.3)

Removing one or more of th e produ cts is one of the principles behind

Optimization of Design & Operation in Reactiue Distillation 5

reactive distillation. The reaction mixture is heated and the product(s) arc

boiled off. However, caution must be taken that the reactants won'L boil

off before the products.

For example, Reactive Distillation can be used in removing acetic acid

from water. Acetic acid is the byproduct of several reactions and is very

useful in its own right. Derivatives of acetic acid are used in foods,

pharmaceuticals, explosives, medicinal and solvents. It is also found in

many homes in the form of vinegar. However, it is considered a pollutant

in waste water from a reaction and must be removed.

CH3COOH + CH30 H ~ CH3COOCH3 + H20 Acetic Acid :\letha111JI J\1eth)'l .llcelflle 1Yt1ter

.. . (1 .4}

The first patents for this processing route appeared in the 1920s,

Backhaus (1921a,b,c), but little was done with it before the 1980s Malone

and Doherty (2000); Agreda and Partin (1984) when reactive distillation

gained increasing attention as an alternative process that could be used

instead of the conventional sequence chemical reaction-distillation.

Using this example one can qualitatively assess the inherent value of

this processing strategy. The process costs are substantially reduced

(- 80%) by the elimination of units and the possibility of heat integration.

Using task integration-based synthesis the conventional process,

consisting of 11 different steps and involving 28 major pieces of

equipment, is effectively replaced by a highly task-integrated RD unit. The

last decades have seen a significant increase in the number of

experimentally research sludies dealing with RD applications. For example,

Doherty and Malone (2001) state more than 60 RD systems have been

studied, with the synthesis of methyl t - butyl ether (MTBE) and ethyl

t-butyl ether (ETBE) gaining considerable attention. Recently, in the frame

of fine chemicals technology Omota et al. (2001, 2003) propose an innovative

RD process for the esterification reaction of fatty acids. The feasibility of this

process is firstly suggested using a smart combination of thermodynamic

analysis & computer simulation (Omota et al.,2003).Secondly, the proposed

Optimization of Dc•ign & Operation in Heaciiue Disrillarion 6

M.o:_H __ -(>

'------1-<>H....-

Conventional Column Figure 1. 1.

Methyl aoei.te

water

RD C<>lumn

Schematic representation of the conventional process for the synthesis of methyl acetate (left) and the highly task-integrated RD unit (right). Legend: RO I : reactor; SO I : splitter; $02: extractive distillation; $03: solvent reeove.ry; S04: MeOH recovery; SOS: extractor; S06: azeotropic column; S07, S09: Gash columns; $08: color column; VOl: decanter

design methodology is successfully applied to a representative eslerification

reaction in the kinetic regime (Omota el aL, 2001). Process development,

design and operation of RD processes arc highly complex tasks. The

potential benefits of this intensified process come with significant

complexity in process development and design. The nonlinear coupling of

reactions, transpo11: phcmomena and phase equilibria can give rise to highly

system-dependent features, possibly leading to the presence of reactive

azeotropes and/or the occurrence of steady-state multiplicities.

Furthermore, the number of design decision variables for such an

integrated unit is much higher than the overall design degrees of freedom

of separate reaction and separation units. As industrial relevance

requires that design issues are not separated from the context of process

development and plant operations.

Optim~ohl>ll of D~~'(Jn & 0pr.mt1on m #it?Octiv« l)i:i;tillation 7

Taking an industrial perspective Stankiewicz (2003) lists the following

processes as potential candidates for RD technology:

Type Example

EsteriTICation - - - - - - - -Methyl acetate fro~ethanol and acetic acid Methyl acetate from methanol and acetic acid Ethyl acetate from ethanol and acetic acid Butyl acetate from butanol and acetic acid

Transestcrification Ethyl acetate from ethanol and butyl acetate Diethyl carbonate from ethanol and dimethyl carbonate

1 lydrolysis Acetic acid and methanol from methyl acetate and water ETBE from isobutene and ethanol TAME from isoamvlene and methanol

Alkylation Cumene from propylene and benzene

Condensation Diacetone alcohol from acetone Bisohenol-A from ohenol and acetone

Hydration Mono ethylene glycol from ethylene oxide and water -Nitration 4 -Nitrochlorobenzene from chlorobcmzene and nitric acid

1.2 .3 Technical Challenges In The Process Design & Operation

To design a Reactive distillation process it makes sense to start

with the feasibility studies. A prerequisite for installing a Reactive

distillation column it necessary to check the boundary condition s . The

process synthesis tools must be utilized whether the process can be

carried out as reactive distillation .If one or more process va riants arc

idcntLf1ed than on can !':tart with the ::itudy of chemical kinetics and the

determination of proper thermodynamic data. Applying these data to

calculate steady state conditions the results then provide the basis for

laboratory experiments. An iterative procedure leads to an optimization of

modeling and experiments.

In Reactive distillation as in conventional distillation the knowledge

of the vapor - liquid equilibrium is vita l. Whe n this equilibrium is

superimposed on the chemical reaction the concentration curve in the

distillation column is appreciably influence. The change in the

Optimization of Design & Operation in Reactive Distillation 8

composition of th e reaction mixture due to distillation becomes apparent

with the aid of distillation curves. By superposition of the two processes

Reactive distillation curves can be determined to clarify the concentration

profile in the reaction column. In reactive distillation to represent our

four component graphically the concept of transformed Reactive

distillation proposed by Buza & Doherty 1995: Espinosa, Agurrie & Perez

1995 are used. This dimensional reduction will not only simplify the

graphic representation but also enable the Reactive distillation to be

described with the same equations as conventional distillation.

Depending on the depth of models, various combinations of models

including mass transfer and reaction can be used to calculate th e

Reactive distillation. In the majority of the cases the equilibrium stage

model is used for simulation of distillation without chemical reaction.

The most important issue for scale up for the design of industrial

columns can be resolved by the three design aids simulation,

experimentations and reference can be used in close linkages with each

other. Simulation is the decisive basis when all major data are available.

A generalized applicability of RD technology is a key challenge for the

process-oriented community. Operational applicability is seen as strategic

goal coupled with the development of (conceptual) design methodologies

that can be used to support the RD decision making process. Thus, the

process systems engineering community is expected to provide tools and

supporting methods that can be used to faster develop and better

operation of RD processes. Designing chemical process involves the joint

consideration of process unit development and design programs and these

are key challenges in RD process design.

A topic that is emerging as a challenge in the RD arena, is that, due to its

system-dependency, RD processing is strongly limited by its reduced

operation window (P, T). The key challenges for the RD community include:

The introduction of novel and more selective catalysts;

The design of more effective and functional packing structures; and

Finding new applications.

Optimization of Desi.gn & Operation in Reactive Distillation 9

The first two challenges are strongly driven by the need to expand the RD

operational window beyond t h e current bounds for a given application.

1.3 SIGNIFICANCE OF CONC EPTUAL DESIGN IN PROCESS SYSTEMS ENGG

1.3.1 SCIENTIFIC SETTING OF CONCEPTUAL DES I GN

Since its introduction, process systems engineering (PSE) h as been

used effectively by chemical engineers to assist the d evelopment of chemical

engineering. In tying science to engineering PSE provides engineers with the

systematic design and operation methods, tools that they require to

successfully face the challenges of today's chemical-oriented industry.

At the highest level of aggregation and regardless of le ngth scale (i.e.

from micro-scale to industrial-scale) the field of PSE discipline relies

strongly on engineers being a ble to identify produc tion systems. For the

particular case of chemical engineering, a production system is defined as a

purposeful sequence of phy !'<ir.R l, r.h emical and biological transformations

used to implement a certain function (Marquart.It, 2004). A production

s ystem is character ized by its function, deliberate delimita tion of its

boundaries within the environment, its internal n e twork structure and its

physical behavior and performa n ce. These production systems are used to

transform raw materia ls into product materials characterized by diffe rent

c h emical identities, compositions, morph ologies and shapes. From a PSE

perspective the most remarkable feature of a system is its ability to b e

decomposed or aggregated in a goal-oriented manner to generate smaller

or larger s ystems (Frass, 2005). Evidently, t h e level of scrutiny is very

much linked to the trade-off between complexity a nd transparency .

The conceptual process design (CPD) is made at the following level

of r edu ced aggregation. ln the remarnder of this section particular attention

is given to CPD in the c ontext of PSE. Thus, CPD is d efined as the task of

frnding the best process flowsheet, in terms of selecting the process units

and interconnec tions among these units and estimating the optimum

design condition s. The best process is regarded as the one that allows for

an economical, safe and e n vironmenta l responsible conversion of specific

Optimization of Dei.ign & Operation in Reactive Distilk.tion 1 O

feed stream(s) into specific product (s).

Although this CDP definition might suggest a straigh t-forward and

viable activity, the art of process design is complicated by the nontrivial

tasks of

• identil)'ing and sequencing the physical and chemical tasks;

• selecting feasible types of unit operations to perform these tasks;

• finding ranges of operating conditions per unit operation;

• establishing connectivity between units with respect to mass & energy

streams;

• selecting suitable equipment opt ions and dimensioning; and

• Control of process operations.

Moreover, the design activity increases in complexity due to the

combinatorial explosion of options. Therefore, systematic methods, based

on process knowledge, expertise and creativity, are requ ired to determine

which will be the best design given a pool of thousands of alternatives.

1 . 3 .2 KEY CHALLENGES JN T HE D ESIGN OF R D

Due to its highly complex nature, the RD design task is sti ll a

challenge for the PSE community. The following intellectually challenging

problems need to be considered by the PSE commu nity (Grossman and

Westerberg, 2000i:

design methodologies for s ustainable and environmentally benign

processes;

design methodologies for intensified processes;

t ighter integration between design and the control of processes; (iu)

synthesizing plant wide control systems;

optimal planning and scheduling for new product discovery;

planning of prncess networks;

flexible modeling environments;

life-cycle modeling;

advanced large-scale solving methods; and

Availability of industrial nonsensitive data.

OprimtZOll<>n of D<!sign & Operanon on Reactwe Di$tillutitm 11

An additional specific challenge is to define n widely accepted set of

effecuveness criteria that can be used to assess process performance. These

criteria i.hnulrl take into account the economic, susto.inauility und

responsiveness/ controllability performances of the design alternatives.

1.4 I NTRODUCTION or THE PRESENT WORK

1.4. 1. S CIENTIFIC NOVELTY & RELEVANCE

The scientific novelty of this work is embedded in several areas.

A detailed analysis to check the role of external and intraparucle

mass transfer 1s necessary before proceeding to evaluate kinetics. The

quantitative criteria involves evaluation of factors, a2 and qlecp

(experimental Thiele parameter). which are defined as the ratios of

observed rates of esterification to the maximum rates of liquid-solid and

intra particle mass transfer respectively.

The temperature dependence of the rate parameters k:i and k1 arc

determined to evaluate the activation energy for the forward and the

reverse heterogeneous reaction. The kinetics and equilibrium of auto

catalyzed and ion exchange resin (Amberlyst-15, Dowex SOW) catalyzed

estenficat1on of acetic acid with methanol are studied in a temperature

range of 303 - 343 K. For the heterogeneous catalytic esterificatton

reaction, the effect of temperature, catalyst loading, and catalyst

particle size and reactant concentrations on the Initial rate of reaction

are studied.

The residue curve mapping technique is extended to the RD case

and systematically applied to reactive mixtures outside conventioi1al

composition ranges. This Lcchnique is found to be particularly useful for

the sequencing of (non-) reactive separation trains. Secondly, the models

of the process synthesis building blocks arc refined leading to th<­

following sub-improvements: (i) 1:1 refmed modular representation of the

building blocks; (ii) changes/improvements in the models of the building blocks.

The analysis of initial rate data is useful in understanding the

OptimlZ<loon of Des>gn & Operooon in RttJdi.,,. Drm11all0n 12

dependence of the reaction rate on individual reaction paramrters and

also in the evaluation of mass transfer effects. Dynamic reactive distillation expenments at pilot plant scale were

perfonned using the Amberlyst 15 and Dowex. 50 catalysts. The

perfonnanoes of these two catalystR in application to the Rd for syntho::sis of

mrthyl acetate are undertaken. These RD studies include varintion of

temperature, catalyst loadmg, reflux ratios. flow rates of the reaclants,

location of the feed trays etc.

The effect of a side reaction on the sdectivity and product punty,

potential side reaction of methanol dehydrating to form dimethyl ether

(DME) and water was studied.

Rate-based and equilibrium model are developed which contains

detailed reaction kinetics and which describes the process behavior

accurately. The benefits of usmg dynamic models of different complexity

and size for process design, optimaJ operation and control of catalytic

distillation (CD) processes an: discussed for the case study of the

heterogeneously catalyzcu reactive d1stillat1on (RD) of methyl acetate. The

Rd system for synthesis of methyl acetate is modeled and simulated using

ChrmCad 5.0.

1.4.1. O U TLINE o r T H& T H El!IIS

This dissertation is divided into several chapters, covering the design

and operation of grassroots RD proCl"sscs. The chnpter I reviews tht"

literature work and research undertaken in reactive distillation to dnte.

The fundamentals, weaknesses and opportunities of RD processing arc

addressed in chapter II. A brief description of the current design

methodologies in RD forms the subject of chapter Ill. Special attention is

paid to the identification of the methodologies' strengths and missing

opportunities and a combination of mt:thodology capabilities is used to

derive a new design approach, which is presented in detail in chapter iV.

The el!lsential elcmenls or this design approach are then addressed using

the synthesis of methyl acetate as tutoriaJ example. Chapter V deals with

Optimizotion of De$ign & Operation in Reactive Disn1latwn 13

the feasibility analysis of RD processing based on an improved residue

curve mapping technique and column sequencing. In chapter Vi, a

steady-state equilibrium model and a rate-based model are developed

and compared for packed reactive distillation columns for the

production of methyl acetate. ChemCad 5.0 version is used in design

and simulations. The material setup and the methodology adopted

during these studies are described in chapter vii. The methods and

calculation techniques that support the design of reactive distillation

columns supported by experimental work and discussion are covered

in chapter viii.The thesis is concluded with a summary of the findings and

accomplishment of the studies followed by remarks nnd recommendntions

for further research in the RD field.