Research topics include (but are not limited to):
~ Eco-efficient intensified fluid separations for pre-treatment and downstream processing
~ Systematic identification and integration of process intensification alternatives
~ Optimal process integration of technologies for e-Refinery
~ Design, control and optimization of intensified distillation technologies
~ Eco-friendly entrainers for extractive distillation processes
~ Modeling and optimization of circular chemical processes and systems

We developed a novel heat-integrated reactive absorption process that eliminates all conventional catalyst related operations, efficiently uses the raw materials and equipment, and considerably reduces the energy requirements for biodiesel production - 85% lower as compared to the base case (equivalent to only 21.6 kW.h / ton FAME). Rigorous simulations based on experimental results were carried out using Aspen Plus and Dynamics. Despite the high degree of integration, the process is very well controllable using an efficient plantwide control structure proposed in this work. The main results are given for a plant producing 10 ktpy fatty acid methyl esters from methanol and waste vegetable oil with high free fatty acids content, using sulfated zirconia as solid acid catalyst.

Several successful examples of integrated processes can be found among reactive separations that combine reaction and separation steps in a single unit (e.g. reactive distillation). Note that such an integration requires a match of the reaction and separation conditions. Compared to traditional reactor distillation sequences, the integrated reactive-distillation design brings several advantages such as: 1. increased conversion due to overcoming equilibrium limitations, 2. increased selectivity via suppression of secondary reactions, 3. reduced energy consumption via in-situ heat integration, 4. avoidance of hot spots, 5. ability to separate close boiling components.

Along with reactive separations, there is also the possibility to integrate different separation units together. The direct or indirect sequence of two distillation columns evolved via the Petlyuk column to the concept of dividing-wall column (DWC). This is a very attractive design alternative as it saves the cost of building two columns and cuts operating costs by using a single condenser and reboiler. Compared to the conventional distillation, DWC offers the following advantages:
1. reduced number of equipment units, 2. lower energy consumption compared to (in-)direct separation sequences, 3. high thermodynamic efficiency due to reduced remixing effects, 4. high purity for all three product streams reached in only one column, 5. better economics: DWC can save up to 30% in capital invested and operating costs.

Both reactive distillation and dividing-wall columns are developments of a conventional distillation column. However, at the same time they are two different ways of integration. The advantages of both integrated units could be further enhanced if they are combined via an additional integration step. The resulting unit called reactive dividing-wall column (RDWC) has a highly integrated configuration that consists of one condenser, one reboiler, the reactive zones, the pre-fractionator and the main column together in a single-shell column. RDWC offers an alternative to conventional reactive distillation towers or multicolumn arrangements, with potential significant cost savings.

Biodiesel consists of fatty acid methyl esters (FAME), currently manufactured by either trans-esterification using liquid Na/KOH catalyst, or batch esterification of free fatty acids using H2SO4 as catalyst. These catalysts are not only corrosive and toxic, but they require neutralization and an expensive separation, thus making biodiesel an attractive but costly alternative fuel. The complete catalyst removal is imperative due to the EU restrictions on sulfur content in diesel fuels (< 15 ppm as of 2006).

To solve these problems, we propose the replacement of the homogeneous acid catalyst with solid acids and develop a sustainable esterification process based on catalytic reactive distillation (RD). Solid acids can be easily separated from the biodiesel product; they need less equipment maintenance and form no polluting by-products. Finding catalysts that are active, selective, water-tolerant and stable under the process conditions is the main challenge for a successful design. We have screened a large number of zeolites, heteropoly-compounds, metal oxides, ion-exchange resins, and carbon-based solid acids.

The catalyst development was integrated in the process design at an early stage, by data mining and embedding of reaction kinetics in the process simulation. The integrated reactive-separation design is able to shift the chemical equilibrium to completion and preserve the catalyst activity by continuously removing the products. This process dramatically improves the biodiesel synthesis and reduce the number of downstream steps. The key benefits are: efficient use of the reactor volume leading to high unit productivity; stoichiometric reactants ratio hence no external recycles of excess alcohol; no neutralization step since the solid acids do not mix with the products hence no salt waste streams; reduced capital and operating cost due to the integration of the reaction and separation into one unit and the elimination of additional separation steps; and sulfur-free fuel as solid acids do not leach into the biodiesel product.

This research aims to replace the conventional batch synthesis of fatty esters by a novel continuous sustainable technology based on catalytic Reactive Distillation (RD) enhanced by entrainers. The key features are: 1. Continuous process that makes use of a stoichiometric feed of reactants, it has no external recycles and delivers high-purity product. 2. Involves minimum equipment, instrumentation, energetic and manpower costs since additional steps for separation and purification are absent. 3. Makes use of solid catalyst. No auxiliary polluting by-products are formed. 4. Multi-product plant can be designed to produce a variety of esters in the same equipment.

The key novelty element in this project is the intensification of the RD process by the use of a Mass Separation Agent (MSA), known also as entrainer. The presence of a MSA is necessary to modify the physical properties of the reactive mixture, namely the phase equilibrium, in order to reconcile optimally the conditions for separation and reaction. The entrainer should be non-toxic and acceptable as impurity for the target products.

Because of equilibrium limitations of esterification, high conversions can be obtained only by using a large excess of alcohol. Batch technologies are highly inefficient involving costly separations, large energy requirements and polluting by-products. A continuous process based on solid catalyst would bring significant advantages. In this respect the use of Reactive Distillation seems appropriate. The use of a solid catalyst is crucial in this project. Some critical issues to address are: 1. High activity in view of the low residence-time achievable in RD columns, below 30 minutes. 2. Prevent the deactivation by water due to the hydrophilic nature of acidic sites. Free-water is obviously prohibited, but low-content water of the organic phase is also necessary in order to guarantee long-life operation. 3. Because the alcohol/acid molar ratio inside the RD column may vary over five orders of magnitude, this may lead to undesired dehydration of alcohols to ethers.

This research consists of screening suitable acid solid catalysts for continuous reactive distillation. Robotic means are used for the fast and reliable synthesis. A new micro-reactor device is built for activity characterization and systematic ranking of candidates. The micro RD device is built in such a way to reproduce the key physical aspects. Next, a reduced number of catalysts with several substrates are investigated in detail. Accurate kinetic measurements and data mining complement previous tasks. The results are imbedded directly in the simulation the RD distillation column.

Another important topic is the intricate relation between plantwide control structures and reactor design, an aspect largely ignored by now both by designers and control engineers. It is demonstrated that the so-called snowball effect (high sensitivity) is merely a matter of reactor design than control. A clear distinction is made between self-regulation and controlled-regulation structures in relation with the stoichiometry of the reaction network and the make-up policy of reactants. In general, using the recycle flow rate to manipulate the production rate is a good design practice, better than fixing the fresh feeds of reactants. However, for some complex reactions, the simple self-regulation feed policy is workable. In addition it may ensure a desired selectivity pattern.

This study proposes a design methodology that allows the designer to screen by a simple and quick procedure feasible design(s) and select promising alternatives that integrates plant wide controllability at an early conceptual stage. The approach is illustrated by realistic case studies. Some of the main conclusions of this project are summarized below:

1. Design of a chemical reactor involved in recycle system should be viewed from a systemic perspective. Reactor volume should exceed a critical value for feasible operation. The dynamic behaviour and control of the Reactor-Separator-Recycle system depends on the reactor size and the control structure for the make-up of reactants.

2. Snowball effect is a steady state pheno­menon that may upset the operation of units involved in recycles, namely the separators. The snowball effect is rather a problem of design than of process control. From control viewpoint larger reactors behave better than smaller ones.

3. Design of chemical reactors should take into account the state multiplicity when more than one reactant is involved, even with isothermal reactors. When only mass recycle is concerned, the reactor type is not important for the type of non-linear behaviour, the results being similar for both CSTR's and PFR's. The superposition of feedback of energy on material feedback gives a more complex pattern of state multiplicity. Operating point should be always selected on the stable branch, far from the fold.

4. Multiplicity pattern depends also on the control strategy. Simplified realistic models can be developed and bifurcation diagrams can be traced by means of standard tools or even flowsheeting simulators.

5. Make-up of reactants should be analysed in connection with reactor design. A clear distinction can be made between self-regulation (fixed fresh feeds) and controlled regulation (make-up in recycle). Self-regulation may be applied if the number of reactants equals the number of independent reactions.

6. The issue of selectivity in recycle systems looks differently as from stand-alone viewpoint. For consecutive/parallel reactions the ratio between desired product and sub-product can be simply set from the initial ratio of reactants, independently from the reactor type or design.

7. Results of this study have been translated in simple design guidelines, such as:
a. The recycle should be preferably of high purity, but can tolerate a limited amount of product.
b. Recycle flow rate should be set at the highest value allowed by the economical trade-off.
c. The ratio of kinetic constants and ratio of fresh feeds can be used to manipulate the selectivity.

This study presents results of dynamic modelling, simulation and optimisation of a sulphuric acid process. In a typical plant (as for example the Monsanto process), conversion of SO2 to SO3 takes place in an adiabatically operated catalytic reactor. Because the reaction is reversible and exothermal, several conversion steps, addition of fresh air and inter-stage cooling are necessary. Further, SO2 conversion is improved and tail gas emissions are reduced through an intermediate SO3 absorption step.

In this study we show that operational problems may occur when the process is disturbed, for example by changing production rate, catalyst deactivation, or varying mass flow rate of air fed into the sulphur burner due to day-night temperature differences and constant volumetric flow rate operation. For certain disturbances, the dynamic response of the plant is nonlinear and leads to sustained oscillations. This behaviour is explained by the presence of an inverse response for the temperature through the reactor beds, combined with the positive energy feedback caused by the two feed-effluent heat exchangers. This study demonstrates two approaches to stabilize the plant, namely changing the catalyst distribution among the conversion stages and employing feedback temperature controllers.

In addition, key operating parameters were determined by sensitivity analyses. This study proves that further improvements are possible through optimisation. The challenge is to maximize the amount and quality of steam generated and, simultaneously, to minimize the environmental impact of the process. For example, the amount of SOx released in the atmosphere can be reduced by 40% if the amount of air fed to the sulphur burner, the flow rates of air fed to converter passes R3 and R4, and split fractions of cold streams entering the gas-gas heat exchangers FEHE1 and FEHE2 are set to optimal values.

The study is based on a dynamic model developed in gPROMS, which includes a graphical user interface build in MS Excel. Therefore, the model is helpful also for operator training, besides controllability, operability and optimisation studies. An excellent agreement exists between the real plant data and the results of the simulations, the relative error being tipically below 1%. For example, the temperature differences between simulation results and real plant data do not exceed 5-10 K.

Further usage of this dynamic robust model includes: controllability analysis (determine the controllability of the plant and possible improvements of the control structure, as well as optimal control policy), other dynamic simulations (detect potential sensitivity problems and which model parameters have greater importance), operator training (due to the Excel interface that plays the role of a control panel that simulates the real plant behaviour for any change of model parameters), other optimization studies (e.g. maximize the amount of energy produced by the plant or optimize the air feed policy accounting for day/night temperature variations).

Characterization of foam and its stability is considered an essential issue. The experiments performed at lab-scale allowed gathering of information about foam and bulk liquid rheology. Along with real plant data, this information was used for modeling the system. The modeling of the aerated stirred tank reactor integrates foam formation as well as most important design parameters. Sensitivity analyses of key operating and design parameters are performed. The results of the model were validated against experimental data. Errors of maximum ± 20% are expected at factory-scale.

Additionally, the use of foam control agents and mechanical foam-breakers was investigated. Antifoaming lab-scale experiments were carried out at and the choice of an antifoaming agent and its practical applicability was optimized. Various designs of mechanical foam-breakers and their practical applications were examined. Operating parameters, such as air sparging regime or mixing conditions, were optimized. Consequently, feasible foam control recommendations are proposed. These alternatives are reviewed and a critical analysis is performed.

Full model results were validated against lab-scale and factory-scale data. Subsequently model adjustments were performed when required. After processing the simulation results, long-term and short-term foam-control solutions are proposed for practical implementation. These solutions have been demonstrated at pilot- and factory-scale.

Monitoring the environmental pollution needs the prediction of toxicity of chemicals in air, waste waters and sole. QSARs (Quantitative Structure-Property Relationships) can be used to predict the toxicity accurately, without using more expensive experimental methods. Drug research and production is also related to the QSAR techniques.

In this work new correlating results by using Cluj and Szeged topological indices are reported, with the aim to demonstrate the capability of our indices to model the molecular properties of organic compounds. The above mentioned indices are calculated on the ground of the Cluj and Szeged matrices, respectively.