Artículo 1 - The Future of Chemical Engineering

Chemical Engineering Science 57 (2002) 4667 4690
www.elsevier.com/locate/ces
The triplet molecular processesproductprocess engineering: the future of chemical engineering ?

Jean-Claude Charpentier
Department of Chemical Engineering=CNRS, Ecole Sup rieure de Chimie Physique Electronique de Lyon, P.O. Box 2077, e 69616 Villeurbanne Cedex, France
Abstract Today chemical engineering has to answer to the changing needs of the chemical and related process industries and to meet the market demands. Being a key to survival in globalization of trade and competition, the evolution of chemical engineering is thus necessary. Its ability to cope with the scientic and technological problems encountered will be appraised in this paper. To satisfy both the markets requirements for specic end-use properties of products and the social and environmental constraints of the industrial-scale processes, it is shown that a necessary progress is coming via a multidisciplinary and a time and length multiscale approach. This will be obtained due to breakthroughs in molecular modelling, scientic instrumentation and related signal processing and powerful computational tools. For the future of chemical engineering four main objectives are concerned: (a) to increase productivity and selectivity through intelligent operations via intensication and multiscale control of processes; (b) to design novel equipment based on scientic principles and new methods of production: process intensication; (c) to extend chemical engineering methodology to product focussed engineering, i.e. manufacturing and synthesizing end-use properties required by the customer, which needs a triplet molecular processesproductprocess engineering; (d) to implement multiscale application of computational chemical engineering modelling and simulation to real-life situations, from the molecular scale to the overall complex production scale. ? 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Future of chemical engineering; Multidisciplinary and multiscale approach; Triplet molecular processesproductprocess engineering; End-use property; Soft solids; Complex uids; Molecular modelling; Process intensication
1. Introduction The world moves forward. Industry used to be king, now the customer is. One key to survival in globalization of trade and competition, including needs and challenges, is the ability of chemical engineering to cope with the society and economic problems encountered in the chemical and related process industries. In this paper we would like to show successively some of the challenges to be taken up by chemists and consequently the waiting from chemical and process engineering. Then it will be presented the new multidisciplinary and time and length multiscale approach of chemical engineering and the necessary tools for assuring the success of this integrated approach. And nally we will propose four tracks for future researches in chemical engineering involving tailoring of materials with control structures, process intensication, product-engineering, and
multiscale simulation and modelization from the molecule scale to the overall complex product scale. 2. The world of chemical and related industries at the heart of a great number of scientic and technological challenges to be taken up by chemical engineering The world of chemistry and related industries, including process industries such as petroleum, pharmaceutical and health, agro and food, environment, textile, iron and steel, building materials, glass, surfactants, cosmetics and perfume, electronics, etc., is considerably evolving at the beginning of this new century due to unprecedent market demands and constraints stemming from public concern over environment and safety issues. The chemical knowledge is growing rapidly as the rate of discovery is increasing every day (Fig. 1). Over fourteen million di erent molecular compounds have been synthesized and about one hundred thousand can be found on the
Tel.: +33-472431702; fax: +33-472431670. E-mail address: charpentier@cpe.fr (J.-C. Charpentier).
0009-2509/02/$ - see front matter ? 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 9 - 2 5 0 9 ( 0 2 ) 0 0 2 8 7 - 7
4668
J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690
14.000.000
12.000.000
number of known compounds
10.000.000
8.000.000
6.000.000
4.000.000
2.000.000
1900
1950
2000
years
Fig. 1. Chemical knowledge is growing rapidly in consumer goods business.
market. Only a small of them are found in nature. Most of them are and will be deliberately conceived, designed, synthesized and manufactured to meet the human need or to test an idea or to satisfy his quest of knowledge (see for example the nowadays development of combinatory chemical synthesis). Already chemistry plays an essential role in mans attempt to feed the population of the planet, to tap new sources of energy, to cloth and house humankind, to improve health and eliminate sickness, to provide substitutes for rare raw materials, to design necessary materials for the new information and communication technologies and to monitor and to protect our environment. Thus to imagine reactions that will convert chemical substances we nd around us into substances or products that serve the consumers needs, such is the business of chemists and such are problems and challenges posed to and by the chemical and related industries. But the evolution of the needs has become such that the keywords associated with the modern chemistry are life sciences, information and communication sciences, and instrumentation in the 21st century. 2.1. Intellectual frontiers of chemistry especially with life sciences Up to recently Chemists have wanted to understand the atom and the structure of matter while Biologists have been more concerned about the function of matter.
Now, Chemist starts to make molecules that have function, and Biologist needs to understand structure. In the past chemical developments such as X-ray, NMR, sequencing of nucleic acids, organic synthesis, genes and recombination methods leading to discoveries in chemistry have been adopted by biology to enable biological discoveries. In the next 20 years the simulation will be reversed and chemistry may look very di erent in the post-genomic world (involving proteomics and metabolomics). Actually the Human Genome project, begun in 1985, may have led to the discovery of 25 000 40 000 new genes or at least 40 000 new chemical products which need to be investigated for both structure and function. Indeed every gene and its products, poly-ribonucleic acid (RNA) and proteins, can be classied according to their elementary functions and their systemic functions. This fascinating topic is a main eld of study in biology and biochemistry research together with bioinformatics: understanding the systemic properties of genes may revolutionize life including new cures for disease, drugs created specically for individuals (mining the genome for drugs) involving nanoscale laboratories where matter is sculptured each atom by each atom, and driving advances in technology, diagnosis and agriculture. It is a huge problem and challenge to understand and to build all the new materials and thus biology will more and more drive chemical discoveries: it will become possible to alter genes in order to alter function, to change proteins chemically to alter their function by using small ligands. Such challenges are equally posed to Chemists in biocatalysis with the design and the use of enzymes (protein molecules used as biologys catalyst to speed up the biochemical reactions in the cell) to catalyse industrial reactions to make complex products under mild conditions with little waste. The problem posed to chemists is to mimic the conditions of nature to design useful enzymes in taking into account the fact that biological systems have evolved over long periods to accomplish very specic functions and thus to adopt the natures methods to design catalysts by evolution of structure and stability. This design necessitates rapid analytical screens obtained today with the modern micro- or nanotechnology. 2.2. Natural intellectual frontiers of chemistry with information and communication sciences and processing technologies Electronics may have dominated the 20th century, but the new century will see this technology superseded by one using both electricity and light. As a result of the limitations of both electronics and all-optical processing, the hybrid technology of electro-optics, where the transmission of light encoded with information through a material is manipulated using electricity, is receiving increased attention. Examples of potential applications include turning electrical television signals into optical signals for long distance transmission by bre optics in the cable TV industry;
4669
detecting electromagnetic radiation in radar applications; voltage sensing in the electric power industry; fast switching in optical area networks; ultrafast analogue-to-digital conversion; beam steering applications including at panel displays, etc. The extent to which applications are implemented and the economic impact of electro-optic devices depends strongly on materials now being developed. For a material to be electro-optically active, it needs to have second-order optical non-linearity. An unusual requirement for second-order optical non-linearity is non-centrosymmetric material symmetry, that is, all the optical species (chromophores) must be dipolar in nature and pointing in the same direction. This feature is rare in nature and preparing such materials is a challenge for chemists and chemical engineers. For several decades, lithium niobate (LiNbO3 ) crystals have been the material of choice for making electro-optic modulator devices, despite having only modest electro-optic properties. Since LiNbO3 is crystalline, clever processing and engineering are needed to make it useable with semiconductor electronics and bre optic transmission lines. This, coupled with the di culty of growing high-quality single crystals, makes lithium niobate expensive to use and the LiNbO3 technology has probably reached maturity. Polymeric materials are now o ering intriguing new possibilities for making and using electro-optic devices and for commercialization on a broader scale. Indeed devices based on polymeric materials show bandwidths (i.e, the ability to process large amounts of information per unit time) of greater than 100 GHz compared with those around 10 GHz for inorganic materials and even smaller for purely electronic devices. Polymeric materials also o er improved processability and ease of integration into device-appropriate shapes, that is, light conning waveguides (optical wires) appropriately integrated with drive electronics (for example semiconductor VLSI) and with silica bre transmission lines. And polymeric electro-optic materials must be capable of being processed into optical quality lms, poled, hardened, processed in unit buried channel waveguide structures, and integrated with VLSI electronics and silica bres. These are challenges for both chemistry and for engineering sciences such as chemical engineering, including advances in microfabrication methods for chemical systems and in microelectromechanical systems (i.e, Wise, 1998). 2.3. Instrumentation in the 21st century To help the required multidisciplinary approach of chemists, there exists instrumentation in all its forms: laser, molecular beam, NMR, Temporal Analysis of Product, X-ray di raction, mass spectrometry, surface acoustic wave technique, spectroscopic ellipsometry, tomography, trajectography, etc. To analyse 10 000 samples a day especially in medical systems and in food production, Chemists have now the possibility of using thin-layer chromatography and its
parallel-processing capabilities. Spy satellite infrared technology that can detail limits below a microgram is becoming available in chemistry. This may be used to determine functional groups without completely resolving the substances on the thin layer plates. Mass spectrometry and ion-trap mass spectrometers with the ability to store ions in the trap will allow to do chemistry within the trap. This may develop into a gas-phase synthesis tool. Developments in holographic optical systems and new optical components and array detector technology is of great help for the development and the use of the quantitative Raman spectrometry allowing very low detection limits, in the ppm range, with the use of bulk samples with little sample preparation. Also the present day high-throughput synthesis methodologies, such as combinatorial techniques, are applied to the discovery of pharmaceuticals, catalysts, and many other new materials. So this developing instrumentation in all its forms at the disposition of chemists will be of great help either for the description of molecular complexity or for obtaining an accurate picture of chemical or enzymatic transformations. This will be necessited for the modelling and simulation in chemical and process engineering. 3. What are we waiting from chemical and process engineering? In fact there are two demands associated with the previous challenges in order to assure competitiveness and employment in such process industries: (i) How to product and with the help of which processes in order to compete in the new global economy where the keywords are globalization of business and technologies, partnership and innovation (innovation means discovery + development). This involves that the speed of product innovation is accelerating. For example in the fast moving consumer goods business to which the majority of the food business belongs, the half-times of product development has decreased from 10 years in 1970 to an estimated 23 years in the year 2000 (Fig. 2). This means that the high bonus on being rst with a product innovation of substance is getting increasingly di cult to achieve,
Half time
10
8 6 4 2 0
U Uerf Unilever
1970
1980 Year
1990
2000
Fig. 2. Acceleration of innovation time.
4670
and speeding up the product=process development cycle is therefore of paramount importance. (ii) How to answer to the evolution of market demands involving a double challenge: for developing and industrializing countries there is a low cost of manpower and less constraining local production regulations. For industrialized countries, there is a rapid development in consumer demand and constraints stemming from public and media concerns over environment and safety issues. To answer these di erent demands, chemical engineering will be in charge: (a) To research very innovative processes for the production of commodity and intermediate products where patents usually do not concern products but processes (multi-step processes for intermediate products). This leads to no longer selecting processes on the basis of economic exploitation alone but seeking compensatory gains resulting from the increased selectivity and savings linked to the process itself. Indeed the economic constraint will no longer be dened as sale price minus capital plus operating plus raw material and energy costs. But the problem becomes more complex and requires valorization of safety, health and environmental aspects, including the value of non-polluting technologies, reduction of raw material and energy losses and products and by-products recyclability as well. Indeed the customer will buy a process which is non-polluting, defect-free and perfectly safe. (b) To progress from traditional intermediate chemistry to new specialities and active material chemistry and related industries. This concerns industries involved with food products, with products for human, animal and vegetal health, along once again with the chemistry=biology interface. This concerns also upgrading and conversion of petroleum feedstocks and intermediates, conversion of coal-derived chemicals or synthesis gas into fuels, hydrocarbons or oxygenates. The aim is characterized by new market objectives, with sales and competitiveness dominated by the end-use properties of a product linked to its quality or shape and size (chemical and biological stability, degradability, chemical, biological and therapeutic activity, aptitude to dissolution, mechanical, rheological, electrical, thermal, optical magnetics characteristics for solids and solid particles together with size, shape colour, touch, handling, cohesion, friability, rugosity, tastes, succulence, esthetics, sensory properties, etc.). Control of this end-use property and expertise in the design of the process, its permanent adjustments to variety and changing demand along with speed in reacting to market conditions will be the dominant elements. Indeed for these new specialities and active materials the client will buy the product which is the most e cient and the rst on the market, thus strengthening the existing competition between the developed country producers.
Such are examples of present day problems and challenges posed by chemical and related industries to specialists of chemical and process engineering. And what is the situation concerning the participation and the implication of chemical engineering? 4. Chemical and process engineering approach in year 2001 To satisfy the previous demands involves material and energy transformations and to create again new industrial processes with zero pollution, zero defects and complete safety. Actually it needs to take into account or cope with what are called today new or emerging technologies such as biotechnology, microelectronics and microoptoelectronics, biomedical, nanotechnologies, new materials, polymers, ceramics, composites. And simultaneously, it requires to maintain competitive the classical technologies necessitated for permanent recurrent and classical problems such as renewable energies, synthetic fuels, raw material and energy savings in order to break the infernal cycle from raw material and energy to wastes in such a way that wastes constitute the raw material of the following cycle (e.g. paper, textile, sludges). Fortunately, chemical engineering is evolving to satisfy these numerous demands as the problem is not quite new for chemical and process engineering specialists. Indeed by denition, the objective of petroleum engineering, then of chemical broadened to process engineering, is the synthesis, design, scale-up or scale-down, operation, control and optimization of industrial processes that change the state, microstructure and chemical composition of material through physico (bio) chemical separations (distillation, absorption, extraction, drying, ltration, agitation, precipitation, uidization, emulsication, crystallization, agglomeration, etc.) as well as through chemical, catalytic, biochemical, electrochemical, photochemical and agrochemical reactions. Chemical engineering involves the whole of scientic and technical knowledge necessary for physicochemical and biological transformations of raw materials and energy into the targeted products required by the customer. Thus it covers areas involving a wide variety of technologies with increasing emphasis on the demand of end-use properties. However, it is important to note that today 60% of all products that a chemical company sells to its customer are crystalline, polymeric or amorphous solids. These materials need to have a clearly dened physical shape in order to meet the designed and the desired quality standards. This also applies to paste-like and emulsied products. So instead of classical basic and industrial chemicals, new developments increasingly concern highly targeted and specialized materials, active compounds and special e ect chemicals. These are much more complex in terms of molecular structure than classical chemicals.
4671
Fig. 3. Chemical supply chain (Grossmann & Westerberg, 2000).
So the purpose of basic research in chemical and process engineering is still the development of concepts, methods and techniques to better understand, conceive and design processes to change raw materials and energy into useful products. But the complexity of phenomena involved in industrial processes now increasingly forces the engineer and researcher to develop new concepts and methods sometimes encountered in other industrial activities such as defense, car, aeronautical and space or medical activities. Thus chemical and process engineering is today concerned with the understanding and development of systematic procedures for the design and optimal operation of chemical, pharmaceutical, food, cosmetics: : : process systems, ranging from nano- and microsystems to industrial-scale continuous and batch processes, as presented in Fig. 3 in using the concept of chemical supply chain (Grossmann & Westerberg, 2000). This chain starts with chemical or other products that industry must synthesize and characterize at the molecule level. Subsequent step aggregates the molecules into clusters, particles, and thin lms as single or multiphase systems that nally take the form of macroscopic mixturessolids, paste-like or emulsion products. Transitioning from chemistry or biology to engineering, one move to the design and analysis of the production units, which are integrated into a process that in turn becomes part of an industrial site with multiple processes. Finally, this site is part of the commercial enterprise driven by market considerations and demands the inclusion of the product quality. In this supply chain, it should be emphasized that product quality is determined at the micro and the nano level and that a product with a desired property must be investigated for both structure and function. This involves a thorough
Fig. 4. The length and time scales covered in the multiscale approach.
understanding of the structure=property relationship at both molecular (e.g. surface physics and chemistry) and microscopic levels. The ability to control microstructure formation to obtain the end-use properties of a uid or solid product is the key to success and will therefore help to design and control product quality and make the leap from the nano level to the process level. This requires an integrated system approach for a multiscale and multidisciplinary modelling of complex simultaneous and often coupled momentum, heat and mass transfer phenomena and processes taking place on di erent scales (Fig. 4): di erent time scales (1015 108 s) from femto- and picoseconds for the motion of atoms in a molecule during
4672
Scales and complexity levels in process engineering
Nano-scale
Molecular Processes Actives sites COMPLEXITY
between
Micro-scale
Particles Droplets Bubbles Eddies
Meso-scale
Reactors Exchangers Separators Pumps
Macro-scale
Production Units Plants
Mega-scale
Environment Atmosphere Oceans Soils
COMPLEXITY
between
MOLECULAR STRUCTURE FLUID DYNAMICS REACTION
PROCESS BUSINESS
How TO UNDERSTAND and to DESCRIBE the relationship between events at NANO and MICROscales to better convert MOLECULES onto USEFUL PRODUCTS at the PROCESS-scale
Fig. 5. Scales and complexity levels in process engineering: to understand and to describe the relationships between events at nano- and microscale to better convert molecules into useful products on the process scale.
a chemical reaction and nanoseconds for molecular vibrations up to the scale of hours for operating industrial processes and of centuries for the destruction of pollutants in the environment. di erent length scales (108 106 m) in industrial practice (Fig. 5) with approaches on the nanoscale (molecular processes, active sites), on the microscale (bubbles, droplets, particles, eddies), on the mesoscale for unit operations (reactors, exchangers, columns), on the macroscale for production units (plants, petrochemical complexes) and megascale (environment, atmosphere, oceans, soils) e.g., up to thousands of kilometers for dispersion of emissions to the atmosphere. So organizing scales and complexity levels in process engineering is necessary to understand and to describe the relationships between events at nano- and microscales to better convert molecules into useful products at the process scale. It is this approach which is required by chemical engineering today. Let us present three illustrations of this multiscale and multidisciplinary approach. (i) Transport phenomena in polyolen polymerization. This example borrowed from polymerization engineering illustrates the fact that even for processes in use on an industrial scale for quite some time, it may be necessary to re-examine the fundamental mechanisms involved at the microscopic level. This is presently the case of the polymerization of olens using highly active ZieglerNatta type catalysts and more recently supported metallocene catalysts that o er the possibility of producing tailor-made polymers in rather mild, and therefore less expensive process conditions. For a proper reactor design, optimization and control, many attempts have been made over the course of the past to model the kinetics of polymerization, the evolution of the morphology of particles formed by the catalyst and polymer
as well as the heat and mass transfer around these growing particles. To put the problem into perspective, Fig. 6 presents the di erent length scales and transfer phenomena involved in both cases of gas-phase uidized-bed reactor (FBR) (left side) and liquid phase slurry or liquid pool processes (SBR) (right side). Let us consider the gas-phase processes which are in theory particularly interesting because they use no solvents and because of the ease of separation of the nal product from the reaction medium. If the models available in the literature are used in conjunction with the classical transfer equations and correlations for transfer coe cients, they predict that experimentally observed polymerization kinetics are theoretically impossible, as they would lead to temperature gradients so high during gas-phase polymerization that the centre of the polymer particles would melt. If this happened, the pores of the growing particles would ll with molten polymer and the resulting increase in mass transfer resistance would then completely extinguish the reaction. This does not mean that signicant temperature gradients at particle levels do not exist. It means that the fundamental description currently available for such situations are inadequate. Since the reaction is very fast we do have very high levels of activity and typical rates on the order of 30 60 kg polymers=g catalyst=h and we do encounter melt-downs in industrials: the models simply do not tell us why or how. If we consider the length scales shown in the schema of Fig. 6, it is easy to see that events taking place on the nanoscale (e.g. kinetics, Fig. 5), the microscale (e.g. internal mass and energy transport), and the mesoscale (particle particle, particlewall interactions) have a signicant impact on macroscale events (e.g. global reactor behaviour), and even on megascale issues such as reactor run away, use of energy, etc. It is therefore absolutely critical that the process engineer has a fundamental understanding of events at all levels of complexity.
4673
LENGTH SCALES
Catalyst
Reactor
1-10 metres
Mixing, RTD, Homogeneity
Particle Swarm
100 m - 1cm
Local Interactions
Electrostatic/Thermal Agglomeration
Flocculation Coalescence
Single Particle
Interphase Transfer
Surface coverage
10 - 100 m
Surface temperature
Sub-particle
1 - 100 nm
Morphology, porosity Arrangement of active sites
Distribution of microdomains
Molecular Phenomena
1 - 100 A
Chemistry, kinetics, macromolecular diffusion
Active site
Fig. 6. Problems to be solved and related length scales in the heterogeneously catalysed polymerization of olens.
For example, a fundamental analysis of mass and heat transport equations reveals that convection might play a very important role during polymerization, especially during the early stages of the operation which are critical for the development of particles (Fig. 7). After the rst few seconds at most, the hydraulic forces created by the formation of solid polymer inside the particle cause the original structure to rupture or fragment. The particle retains its original shape because of the entanglement and=or crystallization of the macromolecules formed in the porous structure of the originally porous support. But values generally retained for the heat transfer coe cient around these particles growing rapidly from an original size of 10 m for fresh catalyst to a nal size on the order of 1 mm for polymer particles have to be increased by a factor of at least 10 in order of magnitude to account for experimental observations. Furthermore, it seems necessary to revise the description of the basic mechanism for heat transfer in the gas phase around the particles because the correlations used up to recently are based on the assumption that heat transfer takes place only via convection. Classical chemical engineering correlations
indeed have been developed for particles diameter 400 m whereas we are dealing with particles on the order of 20 40 m in diameter during the critical stages of low pressure olen polymerization. The use of a computational uid dynamics software package to study heat transfer from spherical particles of di erent sizes and under di erent heat transfer conditions has shown that the classical heat transfer correlations for the case where particles do not interact are not true for densely packed systems such as those encountered in reactors commonly used in olen polymerization (McKenna et al., 1999; McKenna & Soares, 2001). It was also demonstrated that convection is not the only means of removing heat from small highly active particles. Conductive heat transfer between large and small particles present in the same reactor appears to help previous alleviate problems of overheating and explain why earlier models of heat transfer in olen polymerization overpredict the temperature rise during early polymerization. (ii) Transport phenomena in metallic material elaboration. This integrated multiscale approach for the description, analysis, understanding and modelling of phenomena
4674
Fig. 7. Schematic transformation of a fresh catalyst particle into a polymer particle, and evolution of particle morphology and growth. (McKenna, Spitz, & Cokljat, 1999).
occurring at di erent scales is also met in an other important industrial case: the elaboration and processing of metallic materials (metals, alloys, materials having noticeable properties and new functions such as nanophase materials and quasicrystals). For these metallic materials the steps of cast and solidication play an important role, especially for the quality and then for the end-use property. So for the step of solidication this necessitates the homogeneization of materials at di erent scales, thus the organization of levels of complexity, from the microstructure at the microscale of the dendrite and of the grain for columnar or equiaxial structures up to the scale of the surface state of the rough-cast product, characterized by chemical macrosegregation and microsegregations. The controlled formation of these chemical segregations during the solidication requires this new approach in process system engineering in terms of coupled transport phenomena and phase changes, and also in terms of knowledge of the strong coupling existing between phenomena occurring at di erent scales such as the mesosegregation and the microsegregation. (iii) Transport phenomena in biochemical engineering. This multiscale approach is now encountered in the domains of biotechnology and bioprocesses for the knowledge and the control of biological tools (enzymes and microorganisms) to manufacture products and services. In such cases it is necessary to organize the levels of increasing complexity from the gene with known property and structure up to the productprocess couple by modelling coupled mechanisms
and processus at di erent length scales (Fig. 8): nanoscale for molecular and genomic processus and metabolic transformations, pico- and microscales for the enzymes and integrated enzymatic systems, for the populations and cellular plant, mesoscale for the biocatalyst and active aggregates, and macro- and megascales for the bioreactors, units and plants involving interactions with the biosphere. So organizing levels of complexity at di erent length scales associated with an integrated approach of phenomena and simultaneous and coupled processus underlie the new view of biochemical engineering, i.e., understanding an enzyme at the molecular level means that it may be tailored to produce a particular end product (Engasser, 1998). Also in food process engineering for the production of man-made structured foods or for (bio) converted foods there is today a signicant scope for such approaches in linking scales to model process physics, process (bio) chemistry and process microbiology from the molecular and cellular scale to the full process plant scale (Bruin, 1997). These examples underlie the new view of chemical and process engineering: organizing levels of complexity by translating molecular processes (that I call processus) into phenomenological macroscopic laws to create and control the required end-use properties and functionality of products manufactured by a continuous process. This can be dened by le G nie du Tripl processus e e produitproc d (the triplet molecular processes e e productprocess engineering) with an integrated system

BIOCHEMISTRY AND BIOCHEMICAL ENGINEERING
4675
From the GENE with known structure and function to the end-use property (or function) PRODUCT (ecoproduct)
Nano-scale
Gene
Micro-scale
Micro-organism enzyme population cellular plant
Meso-scale
Biocatalyst Environment active aggregat
Macro-scale
Bio Reactors Separators
Mega-scale
Units plants Interaction biosphere
Function
Organizing levels of complexity underlie new view of biochemical engineering

Fig. 8. Biochemistry and biochemical engineering: organizing levels of complexity underlie new view of biochemical engineering.
approach of complex phenomena occurring on di erent length and time scales. This explains why, in addition to the basic notions of unit operations, coupled transfers and classical tools of chemical reaction engineering, that is, in addition to the fundamentals of chemical and process engineering (separation engineering, chemical reaction engineering, catalysis, transport phenomena, process control), this integrated multidisciplinary and multiscale approach is a supplementary and of considerable advantage for the development and the success of this engineering science in terms of concept and paradigms. So in the future, chemical and process engineering will involve a strongest multidisciplinary collaboration among physicists, chemists, biologists, mathematicians and instrumentation specialists leading to the theoretical development of the design of products with complex structures (emulsions, paste-like products, plastics, ceramics, soft solids, etc.). Developing new concepts adapted to this idea of the product, within the framework of what could be called physicochemical (bio) engineering justies the qualication of process engineering as an extension of chemical engineering and takes on its full meaning (Charpentier & Trambouze, 1998; Bacchin et al., 1999). And improving both the design and evaluation of complex systems for the production of real products requires further research into strategies, methodologies and tools. These should be oriented toward the acquisition of basic data in thermodynamics, kinetics, rheology and transport, and toward the conception of new integrated operations allowing for coupling and uncoupling of elementary processus (transfer, reaction, separation) or combining several functions in one piece of equipment. This is clearing the way to smaller and cheaper installations requiring improved knowledge in process modelling, automation and control. But this requires mathematical models and scientic instrumentation which a ords useful basic data that can be treated using powerful computational tools. For example,
the treatment of generalized local information necessitates more and more the help of the computational uid dynamics as it is the case since a long time in combustion, car, aeronautic and spatial applications especially for the knowledge, control, stability of the ows and the characterization and the improvement of the transfer phenomena. Thus CFD, due to recent rapid advances in available software (e.g. CFDLIB, FLUENT, PHOENICS, FLOW 3 D, FIDAP, FLOW MAP, etc.) is daily becoming more important in scaling up new equipment or multifunctional unit operations by simulation of ow phenomena and processing generalized local information, i.e., for understanding the impact of complex ow geometries on mixing and reaction phenomena at the microeddy scale or for the numerical simulation of the complex hydrodynamics of multiphase catalytic gasliquidsolid reactors, or for simulating ow in complex geometries such as reactor internals (industrial distributor devices). Indeed calculations can be carried out for any geometric complexity and for single- and two-phase ow, provided that physical models are available. Nevertheless, the use of this tool becomes possible only when the calculation time is acceptable, i.e., less than few days. CFD is thus a good link between laboratory experiments, conducted at small scale with common uids (air, water, organic, hydro-carbons, etc.), and industrial operation (large scales, complex uids, severe temperature and pressure conditions) (Kuipers & Van Swaaij, 1997). 5. Necessary and indispensable tools for the success of chemical and process engineering It will be possible to understand and describe relationships between events on the nano- and microscale to convert molecules into useful products on the process and unit scales thanks to signicant simultaneous breakthroughs in
4676
three areas: molecular modelling (both theory and computer simulations), scientic instrumentation and non-invasive measurement techniques, and powerful computational tools and capabilities. It may be cited as illustration: 5.1. Modelling at di erent scales at nanoscale, the assistance of molecular modelling to better control surface states of catalysts and activators, to obtain increased selectivity and to facilitate asymmetrical syntheses (chiral technologies), or to explain the relationships structure=activity at the molecular scale in order to control the crystallization, coating and agglomeration kinetics, etc. at microscale the computational chemistry is very useful to a better knowledge of complex media such as non-Newtonian liquids, melted salts, supercritical uids, multiphase dispersions, and suspensions and more generally all systems whose properties are controlled by rheology and interfacial phenomena such as emulsions, colloids, gels, froths, foams, hydrosoluble polymers and particulate media such as powders, aerosols, charged and viscous liquids. The computational chemistry is also of great help for the knowledge of fractal structures of porous media and their in uence or mass and heat transfer, and on chemical and biological reactions. At meso and macroscales, computer uid dynamics is required for the design of new operating modes for existing equipment such as reversed ow, cyclic processes, unsteady operations, extreme conditions, i.e, high temperature, high pressure technologies, and supercritical media: : : : CFD is required for the designed, or for the design of new equipment or unit operations especially by seeking to render process step multifunctional with higher yields in coupling chemical reaction with separation or heat transfer which provides a considerable economic benet. More generally CFD is of previous assistance when it concerns the design of new equipment based on new principles of coupling or uncoupling elementary operations (transfer, reaction, separation). At the scale of production units and multiproduct plants, dynamic simulation and computer tools are more and more required and applied to analyse the operating conditions of each equipment of the production units, to predict both the material ows and states and residence times within individual pieces of equipment in order to simulate the whole production in terms of time and energetic costs. This allows for an interactive walk to predict in a few seconds the new performances (product quality and nal cost) obtained by any change due to a blocking step or a bottleneck in the supply chain. Many di erent scenarios may be tested within a short time, thus allowing the rapid identication of an optimal solu-
tion. For instance, the simulation of an entire production year takes within 10 minutes on a computer. It is clear that such computer simulations enable the design of individual steps, the structure of the whole process at the megascale and place the individual process in the overall context of production. But the previous modelization, simulation, transcription, translation and interpretation at di erent scales require also the current breakthroughs in information collecting and processing. 5.2. Breakthroughs in scientic instrumentation and non-invasive measurement techniques The development of a sophisticated instrumentation and non-invasive measurement techniques leads to notable progress in the knowledge of matterradiation interaction. In this context, the increased cooperation with physicists and physical-chemists is essential since they possess considerable expertise, especially in the application of ne, precise and instantaneous methods, which are not yet widely exploited in process engineering. For example nuclear magnetic resonance (NMR), also used in the medical world, allows one to characterize and monitor chemical and physical phenomena that occur over a wide range of length and time scales and thus provides information on structure and on structure dynamics at the molecular or Angstrom scale (speed of particle agglomeration, rate of bubbles or drop coalescence, speed of nucleation in crystallization, rate of coagulation of colloids, etc.). Magnetic resonance imaging (MRI) presents a non-invasive means to obtain specic information about structural heterogeneity of materials and porous media and concentration, temperature and velocity proles in such media. Thus with the help of performing 3D image analysis techniques such as laser scanning microscopy, one assesses local momentum, heat and mass transfer. Tomographic techniques, optical, acoustical and impedance (both resistance and capacitance) are useful local non-intrusive techniques for ow characterization and on-line control of processus. Capacitance tomography allows for the determination at microscopic scale of instantaneous local velocities, mean lengths and shape coe cients of drops and bubbles and the local fraction of each phase in multiphase ow in porous media. Also, the computed gamma-ray tomographic technique is very promising for the measurement of porosity and gasliquid ow distribution in trickle bed reactor of large diameter as well as the utilization of the computer-assisted X-ray transmission tomography for liquid imaging in trickle ow columns. We should also point out the positron emission particle tracking technique that utilizes a radioactive tracer particle to obtain the trajectories of solid or uid elements in real time either inside rotating blenders or inside agitated vessels containing
4677
Fig. 9. Computing speed acceleration.
non-Newtonian uids. Let us also add the spectroscopic and monochromatic ellipsometry to characterize the structure parameters of solid surfaces in situ and in real time at microscopic scale, i.e. porosities and thicknesses and gas sorption behaviour in thin supported membrane layers and atomic force microscopy to nely analysed surface structures. It is even possible to imagine intelligent micro- and nanotransmitter that measure every processus and model parameter at any location and any time, with the help of, e.g., optical techniques using laser beams such as laser space time resolution uorescence spectroscopy and applied particularly to real media (particulate or opaque). When will micro- or nanoelectronic transmitters be implanted directly on particles or catalytic sites to evaluate local parameters values. When will piezo-electrical polymers that generate a turbulence wake be able to continuously clean the surface of a membrane? 5.3. Breakthroughs in computational tools and capacities The considerable breakthroughs in computation technology and microelectronics must be underlined (Fig. 9). Already today informatics is of greatest importance for the engineer or the researcher in chemical and process engineering for design, control and operation of the process. But if the e ective speed of electronic hardware and software development roughly doubled every year over the past 30 years, this acceleration in computing chip power is expected to continue over the coming decade. Experts predict that around 2010 the magneto-resistive storage technology used will have reached a limit with a storage density of 10 15 gigabyte=cm2 against about 0:25 gigabyte=cm2 today. As previously mentioned as a challenge for chemists and material science specialists, holographic memory technology may substitute the magneto-resistive technology with potential storage densities of 150 gigabyte=cm2 . Manipulating individual atoms is now envisioned and considerable perspectives in molecular simulation are anticipated because the current di culty for its use is the computer calculation time which is approximately proportional to
N 2 where N is the number of atoms which limit the size of the molecules and the number of compounds of a mixture. These fantastic rapid increases in computational capabilities enable to handle more complex mathematics which permits the exhaustive solution of more and more detailed models. This will help chemical and process engineers to model process physics, process (bio) chemistry and process microbiology from the molecular and cellular scale to the full process plant scale. In addition the developments in expert systems and articial intelligence will enable more and more the process engineer to have empirical, qualitative information available virtually at his ngertips in a structured and easily accessible fashion. Fuzzy logic is of great help in the control of processes as well as the neural networks for diagnosing on-line defects, for analysing trends and for the design and the modelling of new processes. Indeed the complexity of phenomena in many cases is such that it might be too long before the obtention of the complete model or the whole of necessary experimental parameters. For example concerning these necessary experimental parameters it is interesting to utilize the advances in computers and neural networks to train a neural net model based on a huge set of available data and make predictions based on such a model. Such an approach has been applied by Larachi, Bensetiti, Grandjean, and Wild (1998) concerning the accumulation of over 30,000 data for the uid dynamic parameters in packed beds with two-phase ow. These authors have shown that if one selects randomly about 60% of the available data in concurrent up ow xed bed reactors, a neural net can be trained to achieve a remarkable t of the training set. The advantage of the approach arises when the neural net predictions are tested against the remaining 40% of the data and very good agreement is found. And nally we should emphasize that the powerful computational tools and capabilities largely contribute to the breakthroughs in signal and image processing for visualization and validation of models at di erent scales. Just to mention the case of operation in real media (particulate or opaque) and=or in complex multiphase ow conditions, the present use of sophisticated techniques such as particle image velocimetry, laser doppler anemometry, laser-induced uorescence, computed tomography, computed automated radioactive particle tracking, etc. (Chaouki, Larachi, & Dudukovic, 1997).
6. Chemical and process engineering: quo vadis ? The previous considerations on the future of chemical engineering involving an integrated multiscale approach of molecular processes, product and process engineering seem to concern four main parallel objectives for engineers and researchers.
4678
6.1. Total multiscale control of the process to increase selectivity and productivity This requires the intensication of operations and the use of precise nano- and micro technology design. Such is the case of molecular information engineering encountered for the supported organometallic catalysis or for supramolecular catalysis where instead of using porous support for heterogeneous catalyst, synthetic materials with targeted properties are now conceived and designed by chemical engineers. Indeed, central to a successful catalytic process is the development of an e ective catalyst which is a complex system in both composition and functionality. And the ability to better control its microstructure and chemistry allows for the systematic manipulation of the catalysts activity, selectivity, and stability. 6.1.1. Nanotailoring of materials with controlled structure Indeed through the control of pore opening and crystallite size and=or a proper manipulation of stoichiometry and component dispersion there exists now ability to engineer via nanostructure synthesis novel structures at the molecular and supramolecular levels, leading to the creation of nanoporous and nanocrystalline materials. These materials both possess an ultrahigh surface-to-volume ratio, which offers a greatly increased number of active sites for carrying out catalytic reactions. Nanocrystalline processing includes the tailoring of size-dependent electronic properties, homogeneous multicomponent systems, defect chemistry, and excellent phase dispersion. This provides nanocrystalline catalysts with greatly improved catalytic activity over conventional systems and multifunctionalities necessary for complex applications. We should mention that vapour phase and wet-chemical synthetic approaches have led to unpreceded control of material structures at the atomic and molecular levels, and brought about ensembles of such features in the shape of nanocrystalline systems involving crystallite-size tuning. Now complex nanocomposite systems can be built to full various roles required for the reaction mechanism and conditions and nanocomposite processing and tailoring also lends itself readily to intelligent combinatorial approaches in material design and rapid catalyst screening (Engstrom & Weinberg, 2000). Beyond catalysis, nanoparticles may be dispersed in an emulsion or a liquid for use in coating applications, food processing, cosmetic products. And organic nanoparticles can be used for drug delivery and gene therapy systems, and quantum dots for medical imaging and diagnostics and more generally for chemical, biochemical, electronic, optical, thermal and structural applications. Also through supramolecular templating, nanoporous systems can now be derived with well-dened porous size and structure, as well as compositional exibility in the form of particles and thin lms. Microporous materials
including zeolites and tailored with well-dened porous structures for excellent surface areas and product selectivity are typically derived through templating with individual molecules. The resulting zeolitic structure which consists of pore opening ( 1:5 nm) allows only small molecules to enter and react, this providing shape and selectivity in separations and catalytic reactions. There exists now the synthesis of well-dened mesoporous materials (250 nm pores) with the development of supramolecular templating which involves the use of molecular aggregates, instead of individual molecules as the framework-directing agents. Supramolecular templating processes are achieved with surfactants to guide the formation of mesostructures from solubilized silicate precursors. The anionic silicates are deposited around the posivitely charged surfactant templates to form inorganicorganic mesostructures (with hexagonal, cubic and lamellar ordering) via electrostatic interactions. For illustration, silicates with hexagonally packed cylindrical pores are obtained with ultrahigh surface areas of 1000 m2 =g. And by using surfactants with longer hydrocarbon tails and by adding polar compounds, the diameters of the mesopores can be systematically varied between 2 and 10 nm (Ying, 2000). So nanoporous structures hold many possibilities in materials applications with further development in molecular engineering such as surface functionalization of inorganic structures and extension of supramolecular templating to organic systems. And self-assembly of nanostructured building blocks (e.g, nanocrystals) combining porosities on di erent length scales will lead to interesting hierarchical structures. Indeed such systems with multiple levels of intricacies and design parameters o er the possibility to simultaneously engineer molecular, microscopic and macroscopic materials characteristics leading to the construction of such advanced systems as biomimicking medical implants or electronic=photonic devices. We could also add that in the eld of homogeneous catalysis a supramolecular ne chemistry has been recently established extending the principle of self-organization of the enzyme (catalyst) molecule to non-biological systems in using supramolecular compounds as catalysts for the shape selection of molecules. Such catalysts are formed in situ by self-organization, i.e., chemical bionics (Kreysa, 2001). So the latest advances in nanotechnology have generated materials and devices with new physical characteristics and chemical=biochemical functionalities for a wide variety of applications. And chemical engineers and researchers are uniquely positioned to play a pivotal role in this technological revolution with their broad training in chemistry, physical-chemistry, processing, systems engineering, and product design. More generally, the previous approaches concerning the tailoring of materials with controlled structure imply that chemical engineers should and will go down to the nanoscale to control events at the molecular level. Indeed at this level, we have seen that new functions such as self-organization,
4679
regulation, replication, and communication can be created by manipulating supramolecular building blocks. 6.1.2. Increase selectivity and productivity by supplying the process with a local informed ux of energy or materials At a higher microscale level, detailed local temperature and composition control through staged feed and heat supply or removal would result in higher selectivity and productivity than does the conventional approach which imposes boundary conditions and lets a system operate under spontaneous reaction and transfer processes. Finding some means to convey energy at the site (supplying the process with a local informed ux of energy) where it may be utilized in an intelligent way is therefore a challenge. Such a focused energy input may be achieved by using ultrasonic transducers, laser beams or electrochemical probes but a more fundamental approach is required to progress in this direction. Indeed the need to convey the exact amount of energy at the precise location where it has to be utilized to promote transfer or reaction requires some kind of feedback between the process and the energy source, and to drive the elementary processes within the unit is a challenge but combining microelectronics and elementary processes, e.g. tuning the selectivity by controlling catalytic reactions at the surface of electronic chips should be a track to explore for chemical engineering. 6.1.3. More clearly recognized is the necessity to increase information transfer in the reverse direction, from process to man This means developing all kinds of intelligent sensors, visualization techniques, image analysis and on-line probes giving instantaneous and local information about the process state. This opens the way to a new smart chemical and process engineering requiring close computer control, relevant models, and arrays of local sensors and actuators. Field-programmable analog arrays coupled with microreactor technology promise to change the way plants are built, as well as the methods by which their processes are designed and controlled. Rapid progress is noticeable in this area, although sensors for opaque materials and particulate solids in bulk systems are still scarce. 6.2. Design of novel equipment based on scientic principles and new operating modes and methods of production: process intensication The progress of basic research in chemical engineering has led to a better understanding of elementary phenomena and now makes it possible to imagine new operating modes of equipment or to design novel equipment based on scientic principles.
6.2.1. Process intensication using multifunctional reactors Such is the case with the multifunctional equipment that couple or uncouple elementary processes (transferreaction separation) to increase productivity, selectivity with the desired product or to facilitate the separation of undesired by-products. Indeed in recent years, extractive reaction processes involving single units that combine reaction and separation operations have received considerable attention as they o er major advantages over conventional processes: due to the interaction of reaction and mass and energy transfer, thermodynamic limitations, such as azeotrope, may be overcome and the yield of reactions increased. So the reduction in the number of equipment units leads to reduced investment costs and signicant energy recovery or savings. Furthermore improved product selectivity leads to a reduction in raw material consumption and, hence, operating costs. So globally, process intensication through use of multifunctional reactors permits signicant reductions in both investment and plant operating costs (10 20% reductions) by optimizing the process. In an era of emaciated prot margins, it allows chemical producers more leverage in competing in the global market place. There exist a great number of reactive separation processes involving unit operation hybridiztion. The concept of reactive or catalytic distillation has been successfully commercialized, both in petroleum processing, where packed bed catalytic distillation columns are used, and in manufacture of chemicals where reactive distillation is often employed. Catalytic distillation combines reaction and distillation in one vessel using structured catalysts as the enabling element. The combination results in a constant-pressure boiling system, ensuring precise temperature control in the catalyst zone. The heat of reaction directly vaporizes the reaction products for e cient energy utilization. By distilling the products from the reactants in the reactor, catalytic distillation breaks the reaction equilibrium barrier. It eliminates the need for additional fractionation and reaction stages, while increasing conversion and improving product quality. The use of reactive distillation in the production of fuel ethers or methyl acetate clearly demonstrates some of the benets. Similar advantages have been realized with the production of high-purity isobutene, for aromatics alkylation, for the reduction of benzene in gasoline and in reformate fractions, for the selective production of ethylenglycol which involves a great number of competitive reactions and for selective desulphurization of uid catalytic cracker gasoline fractions as well as for various selective hydrogenations. The next generation of commercial processes using catalytic distillation technology will be in the manufacture of oxygenates and fuel additives (Dudukovic, 1999). An alternative reaction-separation unit is the chromatographic reactor, which utilizes di erences in adsorptivity of the di erent components involved rather than di erences in their volatility. It is, especially, interesting as an alternative to reactive distillation when the species involved exhibit
4680
small volatility di erences or are either non-volatile and sensitive to temperature, as in the case, for example, in small ne chemical or pharmaceutical applications. There are several classes of reactions to which reactive chromatography is applied. The widest one is given by esterications reactions catalysed by acidic ion-exchange resins or by immobilized enzymes as the polarity di erence between the two products (ester and water) make their separation easy on many di erent adsorbents. Other applications include transesterications, alkylation, etherication, (de)hydrogenations and reactions involving sugars. Reactive chromatography has been also utilized for methane oxidation. In all these applications, special care has to be devoted towards the choice of the solid phase (sorption selectivity, sorption capacity and catalytic activity). Typical examples for the adsorbents used are activated carbon, zeolites, alumina, ion-exchange resins and immobilized enzymes (Lode, Houmard, Migliorini, Mazotti, & Morbidelli, 2001). Concerning the coupling of reaction and crystallization, there exist myriads of basic chemical, pharmaceuticals, agricultural products, ceramic powders, pigments produced by reactive crystallization based processes (i.e., processes that combine crystallization with extraction to solution mine-salts). These separation processes are synthesized by bypassing the thermodynamics barriers imposed on the system by the chemical reactions and the solubilities of the components in the mixture. By combining crystallizers with other unit operations, the stream compositions can be driven to regions within composition space where selective crystallization can occur (Berry & Ng, 1997a). The complementary nature of crystallization and distillation is also explored. Indeed the hybrids provide a route to bypass thermodynamic barriers in composition space that neither the distillation which is blocked by azeotropes and hindered by tangent-pinches in vapourliquid composition space nor the selective crystallization which is prevented by eutectics and hampered by solid solutions and temperature-insensitive solubility surfaces, can overcome when used separately (Berry & Ng, 1997b). Another advantage of such crystallizationdistillation hybrid separation processes is that they do not require the addition of solvents which may increase the process ows, create waste streams and propagate throughout a chemical plant and necessitate costly equipment to separate and recycle these solvents. Membrane technologies respond e ciently to the requirement of the so-called process intensication because they allow improvements in manufacturing and processing, in substantially decreasing the equipment-size=production-capacity ratio, the energy consumption, and=or the waste production and resulting in cheaper, sustainable technical solutions. The paper by Drioli and Romano (2001) documents the start of the art well with respect to progress and perspectives on integrated membrane operations for sustainable industrial growth. The rst studies on membrane reactors were devoted to the use of the membranes for a distributed feed of one of the reactants to a packed bed of catalyst, such
as in partial oxidation reactions in order to improve selectivity. Others have attempted selective removal of product from the reaction site in order to increase conversion of product-inhibited or thermodynamically unfavourable reactions, i.e., immobilization of biocatalysts on polymeric membranes. With such membrane bioreactors, provided that membranes of suitable molecular weight cut-o are used, chemical reaction and physical separation of biocatalysts (and=or substrates) from the products can take place in the same unit. Substrate partition at the membrane= uid interface can be used to improve the selectivity of the catalytic reaction toward the derived products with minimal side reactions. This technology can respond to the strongly increasing demand for food additives, feeds, avours, fragrances, pharmaceuticals, agrochemicals. Catalytic membrane reactors are also proposed to selective product removal to remove equilibrium limitations, i.e., catalytic permselective or non-permselective membrane reactors, packed bed (catalytic) permselective membrane reactors, uidized bed (catalytic) permselective membrane reactors. The development of such membrane reactors for high-temperature applications became realistic only in the last few years with the development of high-temperature-resistant membranes (palladium membranes) mainly to dehydrogenation reactions, where the role of the membrane is simply hydrogen removal. For more general applications material scientists must solve the problem of providing inorganic membranes of perfect integrity involving mechanical and thermal stability and membranes which will allow large uxes of desired species. Also, chemical engineers must gure out the heat transfer problem which now threatens successful scale-up. Thus it might seem reasonable to expect membrane reactors that combine oxygen transfer membranes with selective catalytic layers for partial oxidation of hydrocarbons. However, a continuous research e ort in the process dynamics of these processes and in the study of advanced control systems applied to integrated multimembrane operations is now necessary for a larger utilization of membranes in multifunctional operations combining advantageously reaction and separation in the same vessel. Conversely uncoupling the transferreactionseparation processes may benet: i.e., in a precipitator a better control of the property of the solid is obtained with the separation of the nucleationgrowthagglomeration processes and then in using an ad hoc chemical reaction, the desired product precipitates in the solution which eliminates the further utilization of a crystallization in the chain production. Finally, though multifunctional reactors are not quite new to the process industries, i.e., absorption or extraction with chemical reaction, only recently reactors incorporating several function in one reactor have been formally classied as being multifunctional and the large benets obtained in integrating progress of knowledge at di erent scale and time-lengths have been acknowledged by the process industries. This was illustrated by the rst international symposium on multifunctional reaction (Moulijn & Stankiewicz,
4681
1999) with a presentation of research and development in the main domains : reaction and heat exchange, reaction and membrane separation, reaction and sorption, reaction and power generation, reactions and distillation, reaction and catalyst regeneration and the use of non-classical structured packing. But to achieve optimal performance with multifunctional reactors, it is important to lead a scientic approach to understand where the integration of functionalities occurs, as explained by Dautzenberg and Mukherjee (2001). For example these authors have proposed a classication for the case of a catalyst particle in a reactive medium. However, we will mention more generally that the present day use of hybrid technologies encountered in a great number of multifunctional reactors is limited by the resulting problems concerning control and simulation. In fact, the interaction between simultaneous reaction and distillation introduces more complex behaviour involving the existence of multiple steady states and output multiplicities corresponding to di erent conversion and product selectivity, compared to conventional reactors and ordinary distillation columns. This will lead in the near future to interesting challenging problems in dynamic modelling, design, operation, and strong non-linear control. 6.2.2. Process intensication using new operating modes The intensication of processes may also be obtained by new modes of production which are based on scientic principles. These new operating modes are in the laboratory and=or pilot stage: reversed ow for reaction-regeneration, countercurrent ow and induced pulsing ow in trickle beds, unsteady operations, cyclic processes, extreme conditions, pultrusion, high temperature and high-pressure technologies, and supercritical media are now seriously considered for practical application. Reactors can be operated advantageously with moving thermal fronts that are created by periodic ow reversal. For example, low level contaminants or waste products such as volatile organic compounds can be e ciently removed in adiabatic xed beds with periodic reversal by taking advantage of higher outlet temperatures generated in earlier cycles to accelerate exothermic reactions. Energy and cost savings are a ected by this substitution of internal heat transfer for external exchange (Dautzenberg & Mukherjee, 2001). Some attractive options for improved catalytic reactor performance via novel modes of operation are periodic (symmetric) operation of packed beds with exothermic reaction, and coupling of an exothermic and endothermic reaction in a periodically operated (asymmetric) packed bed. The induced pulsing liquid ow in trickle beds is also proposed to improve liquidsolid contacting at low liquid mass velocities in the cocurrent down ow mode. Also when high conversions are required and the gaseous by-product of the reaction is known to inhibit the rate, such as in hydrodesulphurization, countercurrent ow mode of operation of traditional trickle beds is now preferred.
6.2.3. Process intensication using microengineering and microtechnology Current production modes are and will be more and more challenged by decentralization, modularization and miniaturization. Indeed microtechnologies recently developed, especially in Germany (i.e., IMM, Mains and Forschnungszentrum, Karlsruhe) and in USA (i.e., MIT and DuPont), lead to microreactors, micromixers, microseparators, micro heat-exchangers, and microanalysers, making possible accurate control of reaction conditions with respect to mixing, quenching, and temperature prole. Microfabrication techniques and scale-up by replication have shown spectacular advances in the electronics industry and more recently in microanalysis by biological and chemical applications referred in Section 2. Microfabricated chemical systems are now expected to have a number of advantages for chemical kinetic studies, chemical synthesis, and more generally for process development. Indeed the reduction in size and integration of multiple functions has the potential to produce structures with capabilities that exceed those of the conventional macroscopic systems and to add new functionality while potentially making possible mass production at low cost. Miniaturization of chemical analytic devices in micro-total-analysis-system ( TAS) (Berg, Van den Olthuis, & Bergveld, 2000) represents a natural extension of microfabrication technology to biology and chemistry with clear applications in combinatorial chemistry, high throughput screening, and portable analytical measurement devices. Also the merging of TAS techniques with microreaction technology promises to yield a wide range of novel devices for reaction kinetic and micromechanism studies, and on-line monitoring of production systems (Jensen, 2001). Microreaction technology is expected to have a number of advantages for chemical production as the high heat and mass transfer rates possible in micro uidic systems allow reactions to be performed under more aggressive conditions with higher yields that can be achieved with conventional reactors (Ehrfeld, Hessel, & Lowe, 2000). Also new reaction pathways considered too di cult in conventional microscopic equipment, e.g., direct uorination of aromatic compounds, could be pursued because if the microreactor fails, the small amount of chemicals released accidently could be easily contained. And the presence of integrated sensor and control units could allow the failed microreactor to be isolated and replaced while other parallel units continued production. Also these inherent safety characteristics could allow a production scale systems of multiple microreactors enabling a distributed point-of-use synthesis of chemicals with storage and shipping limitations, such as highly reactive and toxic intermediates (cyanides, peroxides, azides). Moreover scale-up to production by replication of microreactors units used in the laboratory would eliminate costly redesign and pilot plant experiments, thereby shortening the development time from laboratory to
4682
Fig. 10. The IMM micromixer: the two scanning electron microscopy images show the 2 15 interdigitated microchannels with corrugated walls (de Bellefon et al., 2000).
commercial-scale production. This approach would be particularly advantageous for pharmaceutical and ne chemicals industries where production amounts are often less than a few metric tons per year. Small size reactors are already used in testing process chemistries, e.g., catalyst testing. Indeed the small dimensions imply laminar ow, making it feasible to fully characterize heat and mass transfer and extract chemical kinetic parameters from sensor data. Also the high heat and mass transfer rates possible in micro uidic systems that are one or two order-of-magnitude greater than current heat exchangers or multiphase reactors, allow reactions to be performed under more uniform temperature conditions. For illustration of all these potentialities for process intensication using microtechnology, in the publication by Jensen (2001), chemical processing advantages from increased heat and mass transfer in small dimensions are demonstrated with model gas, liquid, and multiphase reaction systems: thin-walled microreactors (realized by MEMSmicroelectromechanical systems) for heterogeneous gas-phase reaction, membrane microreactors for hydrogenation or dehydrogenation reactions, packed bed
microreactor with high gasliquid interfacial area and high surface-to-volume ratios, and low pressure drop, and microfabricated liquid-phase reactor that integrates laminar mixing, hydrodynamic focusing, rapid heat transfer, and temperature sensing. As another illustration, recently de Bellefon, Tanchoux, Caravieilhes, Grenouillet, and Hessel (2000) proposed a new concept for high-throughput screening (HTS) experiments for rapid catalyst screening based on dynamic sequential operations with a combination of pulse injections and micromachined elements. They describe a new concept to achieve HTS of polyphasic uid reactions for two test reactions, a liquidliquid isomerization of allylic alcohols and a gasliquid asymmetric hydrogenation. The principle used for the test microreactor is a combination of pulse injections of the catalyst and the substrate, an IMM micromixer with negligible volume and residence time less than 102 s, and a tubular reactor (Fig. 10). The two scanning electron microscopy images show the micromixer with 2 15 interdigitated microchannels (40 m width) with corrugated walls. The pulses mix perfectly in the micromixer and the liquids or the gasliquid mixtures thereby emulsify and
4683
Comparison with traditional equipment (batch reactor)

80 70 60 50 40 30 20 10
4
HO
One catalyst against several substrates

HO
2
HO OH
Batch Micro
3 5
HO OH
OH OH OH OH
6 7 8
O O
9 10
0 Catalyst pulse 0.2 cm 3 of a [Rh] 0.004 kmol/m 3 solution equivalent to 80 g of Rh for each test. Catalyst : RhCl3 /TPPTS/NaOH. Rh/TPPTS = 5. Selectivity > 95%. 70 C. Flow rate: Aqueous phase 5 cm 3/min Organic phase 1 cm 3/min. [Substrate] 0.1 kmol/m 3. Residence time 100 s.
tion, microreactors must be also integrated with sensors and actuators, either on the same chip, or through hybrid integration schemes ultimately resulting in integrated chemical processors for chemical synthesis (Jensen, 2001). So the integration of multiple reactors presents signicant challenges in uid handling, local reactor monitoring, and control, thus requiring once again multidisciplinary research e orts for chemical engineers. Finally, the previous considerations on the many techniques for process intensication are very attractive. But we should underlie the disadvantage that they are generally built for a unique product. They are not suitable for products with many grades and variations. This leads to another track for the development of chemical engineering. It concerns product engineering in the current atmosphere of quick product development with its short learning and development cycles. 6.3. Formulation or product-design: Manufacturing and synthesizing end-use properties This is the answer for the nowadays ever-growing market place demand for specialty high-value-added products and sophisticated products combining several functions and properties: cosmetics, detergents, surfactants foams, lubricants, textiles, inks, paints, rubber, plastic composites, bres, pharmaceuticals, drugs, foods, agrochemicals, bituminous emulsions, high-performance composite asphalte concrete and more. In practice, the end-use property of such products is often more important than its chemical composition for the consumer. This product engineering (synthesis of properties), as also previously seen in Section 4, is the translation of molecular structure into macroscopic phenomenological laws in terms of state variables. In practice this mostly concerns complex media such as non-Newtonian liquids, gels, foams, hydrosoluble polymers, colloids, dispersions, microemulsions, suspensions for which rheology and interfacial phenomena play a major role. Also involved are the so-called soft solids, systems which have a detectable yield stress, such as ceramic pastes, foods or drilling muds. Product design concerns also particulate solids encountered in 70% of the process industries. This involves the creation and the control of the particle size distribution in operations such as crystallization, precipitation, prilling, generation of aerosols and nanoparticles as well as the control of the particle morphology and the nal shaping and presentation in operations such as agglomeration, calcination, compaction, encapsulation. Both types of operations need a better understanding as they control the end-use property and quality features, such as taste, feel, smell, colour, handling properties, sinterability or biocompatibility. Product engineering concerns also solids considered as vehicles of condensed matter from the perpective of solventless processes or non-passive intelligent solid to accomply intelligent functions such as controlled reactivity
Fig. 11. Comparison with traditional equipment (batch reactor): one catalyst against several substrates (de Bellefon et al., 2000).
behave as a reacting segment, which then travels along the tubular microreactor. Collection at the outlet of the reactor and analysis a ord the conversion and selective data. Application of this principle has been possible by using a static micromixer mounted in a dynamic microactivity test. The catalyst library was then screened. The results led to the selection of the best catalyst showing activity towards a large class of allylic alcohols. Similar results which were obtained in a microreactor and in traditional well mixed batch reactor (40 cm3 ) proves the validity of the concept (Fig. 11). In terms of catalyst and time consumption per test, the numerous tests for the liquidliquid isomerization were performed twice, to test for reproducibility, using only one or two micromoles of metal and over a total screening time of one hour. The test for the gasliquid asymmetric hydrogenation showed similar features (down to 0:2 mol of catalyst, and 35 min per test). Throughput testing frequencies of more than 500 per day are thus achievable, albeit with computer control of the apparatus. Using these microreactors for dynamic, high-throughput screening of uidliquid molecular catalysis o er considerable advantages over traditional parallel batch operations: ensuring good mass and heat transport in a small volume, reduced sample amounts (to g levels), a larger range of operating conditions (temperature, pressure) and fewer, simpler electro-mechanical moving parts. The examples shown here represent a small fraction of the many designs for microreactors being pursued or envisioned by di erent research groups. But in developing microreaction technology for process intensication, it will be essential to focus on systems where microfabrication can provide unique process advantages resulting from the small dimension, not only the high transport rates, but forces associated with high surface area-to-volume ratio. And in order to move beyond the laboratory into chemical produc-
Conversion
4684
Fig. 12. Product-oriented engineering: controlled crystallization process.
or programmed release of active components that may be obtained by multiple layer coatings. The quality and properties of emulsied or paste like and solid products is determined at the micro and the nano level. Therefore to be able to design and control the product quality and make the leap from the nano level to the process level, chemical and process engineering involved with structured materials have to face many challenges in fundamentals (structureactivity relationships on molecular level, interfacial phenomena, i.e., adhesive forces, molecular modelling, i.e., equilibria, kinetics, product characterization techniques, etc.), in product design (nucleation growth, internal structure, stabilization, additives, etc.), in process integration (simulation and design tools based on population balance) and in process control (sensors and dynamic models) (Wintermantel, 1999). For illustration, we may cite the control of the quality of microemulsions for foodstu s containing microorganism that could spoil and whose growth can be prevented in enclosing them in water-in-oil emulsion of aqueous droplet size not signicantly larger than 1 m and of a narrow size distribution which namely characterizes the product quality (Schubert, 1997). Such miniemulsions can only be generated in high-pressure homogenizers with a high-energy input and ad hoc nozzle geometry. But the droplets generated must not coalesce during emulsication which makes it necessary to nd emulsier systems which also stabilize the droplets su ciently quickly. So in modelling the emulsication process, the overall process has to be divided into two substeps: generation of droplet by mechanical energy and stabilization of the droplets before they once again coalesce and the resulting product quality is determined not only by how well the disperse phase has been broken up into small droplets but also by how well the equipment, process conditions and emulsier have been matched to one another.
And thus the kinetics of interface formation, i.e., the process at the molecular level, determines whether the desired end-product properties will be really achieved, even if the required droplet size had been achieved in the rst substep. Another illustration is the control of the shape and size of crystals in an industrial crystallization process. Wintermantel (1999) has shown that much improved process control, both in terms of crystal purity and a dened size distribution could result by detailed computer studies of the crystallization processus which can be remarkly changed by the presence of small traces of foreign substances such as unwanted by-products in the feed solution. In order to understand the mechanisms causing these changes in crystals size and shape so that it is possible to utilize them in a controlled manner, we must explain the structureactivity relationships on a molecular level. And with computer simulations, diagrams of the molecular structure of the most important crystal surfaces can be generated from X-ray crystal structure data. Also in the same computer simulation, contaminant molecules or molecules with an expected benecial e ect on crystallization processus can be placed on each crystal surface and their adsorption energy calculated. If the hypothesis is that the growth rate of surface decreases with increasing adsorption energy, a comparison of the relative adsorption energies could lead to the prediction of the expected modied crystal shape. This was illustrated with the results of crystallization from a feed ammonium sulphate solution containing the dye amaranth. It was shown that the molecule amaranth is adsorbed onto 001 surface of ammonium sulphate crystals with the highest adsorption energy in comparison with the other crystal surface. And according to the calculations, the somewhat block-shaped crystal produced in the pure system becomes a at-shaped crystal having a large 001 surface area, which was experimentally veried (Fig. 12).
4685
1
Chem Engng Specialist
2
LOCAL PROCESS PRESENTATION Velocities, temperatures and Species concentrations fields Chemical reaction rates Deposit thickness rate, growth
GLOBAL PROCESS PRESENTATION Equipment geometry Control process parameters for operating conditions
MODELISATION KINETICS
3
Material Science Specialist
?
EXPERIMENTAL ANALYSIS THEORIES
4
PROPERTIES OF THIN FILMS Shear stress, interface rugosity, optical, electrical properties End-use properties
MICROSTRUCTURE OF THIN FILMS Crystallised fraction, size and orientation of crystallites, defects.
Fig. 13. The four blocks of accessible knowledge necessited for modelling and engineering a low pressure chemical vapour deposit process.
So many quality features can only be designed in a targeted way if the molecular processes are understood at this level. And as shown by this example the analysisboth by theoretical and by experimental meansmust be carried out down to the molecular level to obtain results of real value for understanding the relationship between a certain set of product qualities and the physical product state: indeed it is well known that two pain killing tablets may have the same chemical composition, but di erent routes of production may lead to di erent crystallinity and porosity proles and therefore to totally di erent dissolution and solubility properties, namely di erent bioavailabilities (rate and extent of adsorption in the human body which are often determined by the dissolution rates of di erent crystal faces) (Villadsen, 1997). Recent models recognize the signicance of interfacial phenomena in crystal shape modelling, and lead the way for future developments, such as new simulation and=or group contribution methods for interfacial free energy production. So an important and challenging area for chemical engineering research is to link interfacial models, capable of capturing the solvent or other process conditions to process models leading to the end-product quality especially required in the specialty and ne chemical, as well as the pharmaceutical and life science industries. Another good illustration of this multilevel research effort in crystallization has been recently proposed in the area of electrical engineering: microelectronics. It concerns correlations between operating conditions and microstructure of Low Pressure Chemical Vapor Deposit (LPCVD) silicon based lms prepared from silane SiH4 (Dollet, Caussat, Couderc, & de Mauduit, 2000). A multidisciplinary working group of French and Spanish laboratories involving Motorola Semiconductors SA has been created to progress in this new and very ambitious research domain relative to LPCVD reactor modelling: mechanisms existing at substrate surface, and nally nucleation and crystal growth phenomena. The main aim of the study was the development of a
rigorous simulation model for the interpretation of the layer growth experimental data by taking into account both the mass transfer resistance at boundary layer and the solid layer growth kinetic expression. The four blocks of accessible knowledge for an LPCVD process are presented in Fig. 13: global process presentation, local process presentation, microstructure of thin lms and end-use properties of thin lms. Specialists in material sciences may relate microstructure and properties of thin lms while specialists in chemical engineering can correlate and simulate the macroscopic operating conditions of a CVD reactor and the local conditions inside the reactor. It is clear that any relationship established between blocks 2 and 3 a ords an intrinsic property of the system analysed which is independent of the equipment used and which allows to treat the complete design or optimization problems completely by theoretical methods. For example it has been possible to treat, rst, the kinetic questions such as to nd sets of conditions producing the desired thickness with the desired degree of uniformity. In a second step, one could solve the questions of material structure (amorphous, partially or totally crystallized silicon) end-use properties. The thin CVD layers were produced in the so-called hot wall tubular reactor (2 m length, 2.2 in diameter) schematically presented in Fig. 14. Numerous characterization methods have been used to study the layers microstructural evolutions as a function of their elaboration conditions. The local state of the reactor was simulated by a model CVD2 including the gaseous ow hydrodynamic parameters, the mass transfer parameters and chemical reaction parameters both in heterogeneous phase (gassolid interface) and in homogeneous phase (solid phase). Mechanism existing at surface substrate and nucleation and crystal growth phenomena together with the dynamic characteristic of the microstructure formation were modelized with simple geometric and statistic approach based on concepts of the mechanics of continuous media (instead of using molecular dynamic models requiring too much computer power).
4686
How to design theoretically the best industrial equipment and procedures (at mesoscale) (tubular reactor, length 2 metres, diameter 0,2 metre) to obtain the desired material structure and properties of use (at nanoscale)
wafers
Gas entrance
exit
Heating elements
Fig. 14. Schematic diagram of the horizontal hot wall microelectronic reactor.
The nal model that necessitates the local conditions of elaborations at surface substrate and nucleation and crystal growth laws is able to calculate the thickness, the fraction of crystallized silicon and the space distribution of silicon crystallites in the deposit layer lms from operating conditions such as gas-phase components, temperature, pressure and initial substrate nature or surface state. Fig. 15 shows transmission electronic microscopy micrographies of silicon lms from pure silane for a deposit duration of 29 and 85 nm thickness at 570 C and 0:2 Torr on di erent substrates: silica SiO2 , nitrure Si3 N4 , amorphous and polycrystalline silicon
substrate. Only silica substrate exhibits a partial crystallization. Concerning the crystal state and growth, Fig. 16 shows the in uence of deposit duration. For thin lms (short duration) silicon is crystallized, for intermediary lms, silicon is partially crystallized and for long duration, silicon becomes again totally crystallized. A comparison between these experimental data and the simulation results of the model has shown a good approximation for the crystallized fraction xc and the space distribution of crystallites (Dollet et al., 2000, Fig. 17). The results also emphasize the dynamic characteristics of lm microstructure. Similar results were obtained for complex SIPOS SiOx deposit elaborated from silane and nitrogen protoxide (Barathieu et al., 2000) and in situ boron silicon deposits from silane and boron trichloride (Caussat et al., 2001). It is interesting to note that the quality of the simulation results so obtained demonstrates the validity of the approach proposed and suggest that the way is now opened to develop a complete product engineering or engineering of materials elaboration, able to predict the kinetic and structural characteristics of LPCVD lms by numerical simulation. Finally, it must be emphasized that much progress has been realized over the past few years for the product design and the control of the process using the scientic methods of chemical engineering: thermodynamic equilibrium states are examined, transport processes and kinetics are analysed separately and these are linked again by means of models
Fig. 15. In uence of the substrate surface on the microstructure of LPCVD lms (Caussat et al., 1999).
4687
6.4. Implement multiscale and multidisciplinary computational chemical engineering modelling and simulation to real-life situations The task of chemical and process engineering has been and always will be also to design and implement the complete systems of manufacturing up to the macro- and megascales of the production site and environment. This means that global and structured optimal production of an ensemble of products having shaped mesoscopic properties needs multiscale modelization and simulation to obtain the advantages of an optimally structured complex production. Computers have opened the way for chemical and process engineering in the modelling of molecular and physical properties on the microscopic scales. For the molecular modelling, application of the principles of statistical molecular mechanics computational techniques (Monte-Carlo and molecular dynamics) and quantum mechanics constitue an area for the problem-oriented approach of chemical and process engineering. Molecular modelling starts from a consideration of microscopic structure and molecular interactions in a material system and derives thermodynamic, transport, rheological, mechanical, electrical, electronic or other properties through rigorous deductive reasoning bases on the principles of quantum and statistical mechanics. Compared to more phenomenological approaches (e.g. correlations of the group contribution type) it o ers the advantages of greater generality and reliability. Molecular modelling work can be distinguished broadly into theory and computer simulation. Theories provide closed-form descriptions of relationships between molecular constitution and macroscopic properties ; typically, mathematical approximations have to be introduced in order to reach such tractable descriptions. Simulations, on the other hand are numerical solution of quantum and statistical mechanics formulations, which may be free of the simplifying approximations invoked in theories. It is clearly impossible to cover all directions of present-day molecular modelling researches involved in a wide spectrum of problems in the chemical and material sciences. Vapourliquid equilibria, Vapourliquidliquid equilibria, liquidsolid equilibria, supercritical solution properties; amphiphiles; polymers at interfaces: adsorption on surfaces and in uence of impurities; microporous materials or ceramics structures; transport properties such as viscosity, di usivity, thermal conductivity can be calculated today by molecular modelling based on information from thermodynamic, kinetic and rheological data banks. For example, Sheehan and Sharratt (2000) reviewed the potential benets of the application of molecular modelling techniques for the prediction of solvent e ects on reaction kinetics and liquidliquid phase equilibria. They show that when the theoretical model of an equilibrium process to produce an endo- and exoproduct and the computational method used to represent this model are carefully chosen,
Fig. 16. In uence of the deposit duration on the microstructure of LPCVD lms (Caussat et al., 1999).
Fig. 17. Microstructure of silicon LPCVD lm at di erent temperatures: (a) TEM clich s, (b) simulation results, (c) schematic structure of the e lm deposit thickness (Dollet et al., 2000).
with or without the help of molecular simulation, and nally of computer tools for simulation, modelization and extrapolation at di erent scales for the whole supply chain (BASF, Unilever, Degussa, Astra, Zeneca, AKZO, Nestl , : : :). e But how can operations be scaled up now from laboratory to plant? Will the same product be obtained and will its properties be preserved? What is the role of equipment design in determining product properties? This leads to the 4th main objective for the future development of chemical and process engineering.
4688
molecular modelling can be a valuable tool for engineering design, i.e., evaluating solvent in uence on process performance. There is no doubt that molecular modelling is now playing an increasingly important role in future chemical and process engineering research and practice. And through the interplay of molecular theory, simulation, and experimental measurements evolves a better quantitative understanding of structureproperty relations, which, when coupled to macroscopic chemical engineering science, can form a basis for new materials and process design. Turning to the macroscopic scale, dynamic process modelling and process synthesis are being also increasingly developed. Indeed one must remember that the targeted products in question are generally not mass-produced products but ones which are produced in small batches and just in time for delivery to the customer whose needs are constantly changing and evolving. And to be competitive under these conditions, it is particularly important to analyse and optimize the supply chains for which we are interested in the time that individual process steps take, and these have to be simulated and evaluated also in terms of costs. But in chemical and related processes, the location of a particular component in the supply chain at a given time is not well dened. Indeed if in manufacturing industry such as car manufacturing, the component is either in a machine or in a transport device, in process industry by contrast, a batch can be found in a stirred tank, a lter, a dryer, a pump, a mill and a storage container simultaneously. New event-driven simulation tools help solve these problems by simulating both material ows and states within the individual pieces of equipment. This dynamic simulation may enable one to see in a matter of seconds whether bottlenecks may occur in the plant over the course of days, months or years. These can be eliminated by using additional pieces of equipment or by making additional resources available such as energy or manpower. The event-driven simulation also shows which alternative plant and storage strategies provide the greatest cost benet (Wintermantel, 1999). Thus the ability to use a series of coherent tools during the process life cycle is now well established: integration and opening of modelling and event-driven simulations environments are these days taking on a more and more important role, see for example the computer aided process engineering European program CAPE-OPEN which involves a great number of simulator sellers, European clients and universitary researchers in computing and simulation. The aim is to promote the adoption of a standard of communication between simulation systems at any time and length-scale level (thermodynamics, unit operations, numerical utilities for dynamic, static, batch simulations, uid dynamics, process synthesis, energetics integration, control process) in order to simulate processes allowing the customers to integrate the informations from any simulator onto another one. Indeed manufacturers in the process industries must use a collection of software (in-house, commercial, and=or academic) to perform CAPE because the market for process
simulation has been so far one of the incompatible proprietary products. Fortunately for the future, the Global CAPE-OPEN (GCO) project is expanding and developing interface specication standards to ensure interoperability of CAPE software components. And a standardization body (CAPE-OPEN Laboratories Network, CO-LaN) has been established to maintain and disseminate the software standards in the CAPE domain and to ensure that software tools used by the process industries reach a level of interoperability that will help ensure sustained growth and competitiveness (Pons et al., 2001). So we must emphasize that within the supply chain at the level of the macroscopic scale, dynamic process modelling and process synthesis is today increasingly developed. For example let us recall that full plant models may involve more then 200.000 variables, 500.000 equations and 500.000 optimization variables. Also at the scale of a single unit of production, it is possible to scale-up by a factor of 1:50.000 from the laboratory to the large scale plant by a computer simulation using mainly physical property data and computer uid dynamic or mixing. But modelling should not be confused with numerical simulation. Especially for university and industrial researchers modelling must be an activity requiring knowledge of scientic facts, experience, skills, and judgment (Villermaux, 1996). More specically the bottleneck for good models of multiphase and complex systems is the understanding of the physics, chemistry and biology of interactions, rather than the renement of numerical codes, whose sophistication is not at all concerned with real-life problems in plants and in industrial practice. Indeed chemical engineers should be looking at architects for examples on how to teach design, not at CFD packages. So in the next years attention should be more and more encouraged to systemic analytical models based on the multiscale integrated approach referred in Section 4 that considers the global behaviour of complex systems as a whole: instead of looking at more and more mathematical details, novel principles of the analytical models in chemical and process engineering should also be sought at the highest level of integration. What is needed in models is less anatomy and more physiology. 7. Conclusions: Chemical and Process Engineering: a multidisciplinary key technology serving mankind Chemical and related industries are confronted to a great number of challenges in the framework of globalization of trade and competition. These challenges have been presented and concern the contextsociety and market demands versus technological o ers (Charpentier, 2000, 2001). Actually this needs to answer to a double challenge. First, to research innovative processes for the production of commodity and intermediate products with non-polluting, defect-free and perfectly safe industrial processes. Second, to progress from
4689
traditional intermediate chemistry to new specialities and active material chemistry dominated by the synthesis of the end-use property of the product required by the customer. When used, this product should be safe and not disturb the environment too much. This needs an overall view on product design and development. To satisfy these market demands, it is shown that what the development of chemical and process engineering needs today is an integrated multidisciplinary and time and length multiscale approach from the molecular scale up to that entire production site. This approach is necessary for the understanding and the modelling of the complex simultaneous and often coupled transport phenomena and processes taking place on the di erent scales of the chemical supply chain. This was illustrated with examples concerning polyolen polymerization, metallic material elaboration and biochemical and food process engineering. It has been underlined that the previous integrated approach is today possible thanks to the considerable progress in the use of molecular modelling, scientic instrumentation and powerful computational tools and capabilities, as illustrated with the determination of physical and chemical parameters necessary for the modelling at di erent time and length scales. Finally for the future of chemical engineering, it is proposed to undertake simultaneous researches in four parallel tracks: (a) a total multiscale control of the process to increase selectivity and productivity, a good illustration is the nanostructural tailoring of required materials with controlled structure, (b) process intensication by design of novel equipment based on scientic principles and new operating methods of production, illustrations concerning multifunctional reactor, or the use of micro technology for high-throughput screening experiments for rapid catalyst screening, (c) to synthesize structured products combining several functions and properties required by the customer with special emphasis on complex uids and solids technology, illustrations concerned the quality of microemulsions for foodstu s and the control of the shape, the size and the structure of crystals in crystallization processes encountered in chemical and electronic industries, and (d) to implement multiscale and multidisciplinary computational and modelling and simulation to real-life situations, but this should concern the understanding of the physics, chemistry and biology of interactions, rather than the renement of numerical codes, whose sophistication is not at all concerned with real-life problems in industrial practice. So the future of chemical and process engineering requires the integrated approach presented as the triplet molecular processesproductprocess engineering. Moreover chemical and process engineering will be more and more involved and concerned with the application of Life Cycle Assessment to the triplet engineering, i.e, application of LCA not only to product design and its use, but also to the plant and equipment together with the associated services (Burgess & Brennan, 2001). Indeed the multiscale integrated approach adopted by chemical engineering will be of great help to
answer to the increasing environmental, society and economics requirements and to the goals of a transition to sustainability. And chemical and process engineering with its multiscale approach to understand the multidisciplinary interactions at di erent time and length scales necessary for a successful product development will more and more behave as a multisdisciplinary key technology serving mankind, regardless of the industries where chemical engineers work. References
Bacchin, P., Bergel, A., Biscans, B., Floquet, P., Wilhem, A. M., & Joulia, X. (1999). Le G nie des Proc d s: Discipline des int ractions e e e e et int ractions des disciplines Entropie, 216, 6166. e Barathieu, P., Caussat, B., Scheid, E., Schamn, S., Berjoan, R., & Couderc, J. P. (2000). Corr lations entre conditions d laboration et e e microstructure de d p t LPCVD. Partie 5: D p t de SIPOS a partir e o e o de silane et de protoxyde dazote Entropie, 227, 39. Berg, A., Van den Olthuis, W., & Bergveld, P. (2000). Micrototal analysis systems 2000. Dordrecht: Kluwer academic. Berry, D. A., & Ng, K. M. (1997a). Synthesis of reactive crystallization processes. AIChE Journal, 43(7), 17371750. Berry, D. A., & Ng, K. M. (1997b). Synthesis of crystallization in distillation hybrid separation processes. AIChE Journal, 43(7), 1751 1762. Bruin, S. (1997). Frontiers in food process engineering. 6 me Congr s e e Francais de G nie des Proc d s, Paris 1997. Entropie, 33(208), 1321. e e e Burgess, A. A., & Brennan, D. J. (2001). Application of life cycle assessment to chemical processes. Chemical Engineering Science, 56, 25892604. Caussat, B., Couderc, J. P., Vasquez, L., Figueiras, A., Van der Lee, A., Dollet, A., Berjoan, R., Scheide, E., Temple, P., Schamn, S., de Mauduit, B., Paillard, V., Villa, A., & Morante, J. R. (1999). Corr lations entre conditions d laboration et microstructure de d p ts e e e o LPCVD. Partie 3: D p t de silicium pur a partir de silane: R sultats e o e exp rimentaux Entropie, 222, 313. e Caussat, B., Scheid, E., de Mauduit, B., Vasquez, L., Cot, D., Van der lee, A., Durand, J., Berjoan, R., Paillard, V., & Couderc, J. P. (2001). Correlations entre conditions d laboration et microstructure de d p ts e e o LPCVD. Partie 6: D p t de silicium dop in situ au bore Entropie, e o e 231, 211. Chaouki, J., Larachi, F., & Dudukovic, M. (1997). Non invasive monitoring of multiphase ows. Amsterdam: Elsevier. Charpentier, J. C. (2000). It used to be industry, now the customer is king: How adapting the chemical engineering approach to the needs of process industries in year 2000: Market demands versus technological o ers Entropie, 223, 5666. Charpentier, J. C. (2001). In a context of market demands versus technological o ers is the new chemical engineering able to help and why? 3rd European congress in chemical engineering, invited plenary lecture, Nurnberg, 26 28 June 2001. Charpentier, J. C., & Trambouze, P. (1998). Process engineering and problems encountered by chemical and related industries in the near future. Revolution or continuity? Chemical Engineering and Processing, 37, 559565. Dautzenberg, F. M., & Mukherjee, M. (2001). Process intensication using multifunctional reactors. Chemical Engineering Science, 56, 251267. de Bellefon, C., Tanchoux, N., Caravieilhes, S., Grenouillet, P., & Hessel, V. (2000). Microreactors for dynamic, high troughput screening of uid=liquid molecular catalysis. Angewandte Chemie, International Edition, 39(19), 34423445. Dollet, A., Caussat, B., Couderc, J. P., & de Mauduit, B. (2000). Corr lations entre conditions d laboration et microstructure de d p ts e e e o LPCVD. Partie 4: D p ts de silicium pur a partir de silane: Pr vision e o e
4690
J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690 proposals for future developments Chemical Engineering Science, 56, 39313949. McKenna, T. F., Spitz, R., & Cokljat, D. (1999). Heat transfer from catalysts with computational uid dynamics. AIChE Journal, 45(11), 22402392. Moulijn, J. A., & Stankiewicz, A. I. (1999). 1st International symposium on multifunctional reactors, special issue. Chemical Engineering Science, 54(10), 1297. Pons, M., Braunschweig, B., Irons, K., Koller, J., Kuckelberg, A., & Roux, P. (2001). CAPE-OPEN (CO) standards: Implementation and maintenance. 3rd European congress of chemical engineering, computational engineering session 7, Poster 188, Nurnberg 26 28 June 2001. Schubert, H. (1997). Advances in the mechanical production of food emulsions. In: R. Jawitt (Ed.), Engineering and food at ICEF 7, Part 1 (pp. AA 8287). She eld: She eld Academic Press. Sheehan, M., & Sharratt, P. N. (2000). Predicting solvent e ects on reactions and liquidliquid equilibrium by computer simulation. Transactions of the Institution of Chemical Engineers, 78, 10771083. Villadsen, J. (1997). Putting structure into chemical engineering. Chemical Engineering Science, 52(17), 28572864. Villermaux, J. (1996). New horizons in chemical engineering. Plenary lecture at the 5th World congress of chemical engineering, San Diego. Wintermantel, K. (1999). Process and product engineering achievements, present and future challenges. Chemical Engineering Science, 54(11), 16011620. Wise, K. D. (1998). Special issue on integrated sensors, microactuaters, and microsystems (MEMS). Proceedings of the IEEE, 86, 1531. Ying, J. (2000). Nanostructure tailoring: Opportunities for molecular engineering in catalysis AIChE Journal, 46(10), 19021906.
de la microstructure des lms par simulation num rique Entropie, 226, e 311. Drioli, E., & Romano, M. (2001). Progress and new perspectives on integrated membrane operations for sustainable industrial growth. Industrial and Engineering Chemistry Research, 40, 12771300. Dudukovic, M. P. (1999). Trends in catalytic reaction engineering. Catalysis Today, 48, 5. Ehrfeld, W., Hessel, V., & Lowe, H. (2000). Microreactors: New technology for modern chemistry. Weihgheim: Wiley-VCH. Engasser, J. M. (1998). La cr ation de cellules-usines: Un nouvel enjeu e pour le G nie Biochimique. Entropie, 212=213, 117122. e Engstrom, J. R., & Weinberg, W. H. (2000). Combinatorial material science: Paradigm shift in material discovery and optimization AIChE Journal, 46, 2. Grossmann, I. E., & Westerberg, A. E. (2000). Research challenges in Process Systems Engineering. AIChE Journal, 46(9), 17001703. Jensen, K. F. (2001). Microreaction engineeringis small better? Chemical Engineering Science, 56, 293303. Kreysa, G. (2001). Future trends in chemical engineering. 6th World congress in chemical engineering, Melbourne, 2327 September. Kuipers, J. A. M., & Van Swaaij, W. P. M. (1997). Application of computational uid dynamics to chemical reaction engineering. Review on Chemical Engineering, 3, 1. Larachi, F., Bensetiti, Z., Grandjean, B. P. A., & Wild, G. (1998). Two-phase frictional pressure drop in ooded-bed reactors: A state-of-the-art correlation Chemical Engineering Technology, 21, 887. Lode, F., Houmard, M., Migliorini, C., Mazotti, M., & Morbidelli, M. (2001). Continuous reactive chromatography. Chemical Engineering Science, 56(2), 269. McKenna, T. F., & Soares, J. B. P. (2001). Single particle modelling for olen polymerization on supported catalyst: A review and

Artículo 1 - The Future of Chemical Engineering

Caricato da

Informazioni sul documento

Descrizione originale:

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Artículo 1 - The Future of Chemical Engineering

Caricato da

Copyright:

Formati disponibili

Chemical Engineering Science 57 (2002) 4667 4690

The triplet molecular processesproductprocess engineering: the future of chemical engineering ?

Tel.: +33-472431702; fax: +33-472431670. E-mail address: charpentier@cpe.fr (J.-C. Charpentier).

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

number of known compounds

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

Fig. 2. Acceleration of innovation time.

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

Fig. 3. Chemical supply chain (Grossmann & Westerberg, 2000).

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

Scales and complexity levels in process engineering

MOLECULAR STRUCTURE FLUID DYNAMICS REACTION

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

Mixing, RTD, Homogeneity

Morphology, porosity Arrangement of active sites

Chemistry, kinetics, macromolecular diffusion

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

Organizing levels of complexity underlie new view of biochemical engineering

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

Fig. 9. Computing speed acceleration.

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

Comparison with traditional equipment (batch reactor)

One catalyst against several substrates

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

Fig. 12. Product-oriented engineering: controlled crystallization process.

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

J.-C. Charpentier / Chemical Engineering Science 57 (2002) 46674690

Potrebbero piacerti anche