Development of reaction calorimetry applied to supercritical CO2 and methanol-CO2 critical mixture

The combination of reaction calorimetry and supercritical fluids as reactants and/or solvent holds great potential for the optimization of chemical processes using high pressure near-critical and/or supercritical mixtures. The very specific properties of supercritical fluids and fluid mixtures have been under investigation already from the nineteenth century and some are quite well understood. But there is still a lack of knowledge in macroscopic properties and raw data for pure fluids and fluid mixtures. This cumulates with the difficulty to apply simple equations of state for the prediction of phase behavior and the calculation of the thermodynamic and the derived transport properties. As a whole, the relatively poor engineering information about supercritical fluids is one of the major reasons that explains the slow industrial development of supercritical processing. The work presented here concerns the complete development of the new "supercritical reaction calorimetry field", which is in fact the application of reaction calorimetry to supercritical fluids media. As far as we know we are pioneers in this domain as only relatively small scale calorimeters have been successfully used with SCFs and only scarcely. The fact is that the use of reaction calorimetry provides essential information for scale-up purposes. Moreover it allows additional in-line equipment to be used, in order to get complementary information. This renders the reaction calorimeter much lesser specialized and potentially attractive as a tool for industrial application developments. In order to correctly position this project in the scale of laboratory and industrial researches, the first chapter is dedicated to the review of SCFs and SCF mixtures properties, their potential as volatile organic solvent relievers and their actual industrial applications. As a result there is still a great need in engineering data for pure SCFs and their mixtures, although their unique properties make them suitable for several processes, mostly driven by environmental considerations. Classical reaction calorimetry is also shortly discussed with the determination of the technical challenge that arose from the use of a supercritical fluid, as it basically occupies all the available space. Thus we point out that both cover and flange should be separately controlled and adjusted to the inner temperature in order to avoid side heat transfer from the reaction media to those elements. Moreover, the regulation parameters for the jacket and the additional cover and flange should be optimized. This has proved to be difficult as these parameters have been shown to depend strongly on the supercritical carbon dioxide pairs, density and temperature. Nevertheless satisfactory pairs of P, I parameters for all three parts of the reactor have been found using the Ziegler-Nichols approach for the jacket and a pure trial-and-error approach for the two others. The heat transfer with SCFs should be carefully taken into account in order to proceed with correct chemical reactions' evaluation. Heat transfer in scCO2 and CO2-methanol mixtures can be described in stirred tank reactors using the well known Wilson plot methods. The behavior of the internal film heat transfer coefficient with the temperature is drastically different from the one observed for classical liquid systems. In contrast to classic liquids, in supercritical CO2, the lower the temperature (above the critical point) the higher the internal film heat transfer coefficient. It shows a clear divergence when approaching the critical point. This tendency could be explained by the evolution of the thermodynamical and transport properties of scCO2 around the critical point. This confirms some observations in the literature with some continuous flow heat transfer apparatuses working with SCFs. The heat transfer coefficients for the cover and the flange have been evaluated and show the same general tendency as the one of the jacket in regard to the temperature. The hydrodynamics of scCO2 in a stirred tank reactor have been investigated using Laser Doppler Anemometry measurements and indicating that the stirred scCO2 has a similar behavior with liquid water. However, the turbulent kinetic energy and the mean velocities are 20% lower for scCO2, which can be explained by its lower density. Information about the phase behavior during a reaction is essential in order to proceed homogeneously. Under this condition, the system allows for higher reaction rates and for easier process control. This has motivated the complementary development of a video imaging surveillance system for the inside of the reactor, as well as the development of a high pressure variable-volume view cell for the screening of mixtures' phase behavior. As the "top-view" observation is not ideal for phase change observation (no meniscus visualisation), the further development of an in-situ high pressure ultrasonic probe is also provided in this work. Sound speed measurements have been proven to be not only useful as a "phase-change" detector, but also allowed determining the critical point of pure fluids and mixtures. This work also pointed out the real difficulty to combine a homogeneous supercritical reaction media to a sufficiently high heat produced signal. To access the limit of detection of the calorimetric system and to find out the characteristic time constant of the reactor and its variation with temperature and density, we created an amplifier system based on a Labview® interface program. Using some defined functions for the calibration heater released heat, it has been possible to simulate some typical reaction curves and thus validate the heat flow measurements with the supercritical reaction calorimeter.

Optimization of reaction calorimetry for monitoring chemical reactions in supercritical fluids

Charalampos Mantelis

The PhD thesis deals with the optimization of the reaction calorimetry technique to monitor chemical reactions in supercritical fluids. The aim is to develop this thermal analysis technique to monitor the heat released by a chemical reaction in a high pressure reactor and consequently discuss upon the reaction evolution. The experimental equipment used was previously developed in the group of chemical and physical safety at the EPFL. Additionally, a mass flow meter was installed during this thesis to measure the quantity of supercritical fluid inserted in the reactor. Initially, the core of the technique, meaning the heat flow equation, was examined in a term-by-term analysis to adapt and optimize each term for supercritical reaction systems. For this analysis a model reaction was chosen, namely the free-radical dispersion polymerization of methyl methacrylate in supercritical carbon dioxide, and a set of conditions was set as a reference. The particularities linked to the supercritical nature of the solvent were taken into consideration and more precisely the fact that the solvent occupies the entire available reactor volume. As a result it was found that every reactor part has to be thermally controlled and that special caution has to be paid on the estimation of the overall heat transfer coefficient between the reaction mixture and the temperature regulating fluid running in the reactor jacket and on the estimation of the specific heat capacity of the reaction mixture. Furthermore, the injection phase of the additional reactants, according to the reaction protocol, was found to be subject to considerable calorimetric errors; therefore an optimized injection pattern was designed to minimize the undesired temperature oscillations during this phase. The second step of the thesis consists of the discussion on the reaction evolution based on the optimized results obtained. First a direct comparison is presented between the calorimetric results before and after the analysis to highlight the points were significant improvements were introduced, mainly in terms of accuracy and reproducibility of the results. Then the discussion focuses on the nucleation phase of the polymer particles, where the data show that the reaction takes place almost exclusively in the continuous phase. This conclusion is also found to be in very good agreement with the results of other experimental techniques. Further, the role of the pressure on the reaction evolution was examined through a series of experiments, based on a small reaction deceleration observed, and was found to have a drastic effect on the creation of stable dispersion conditions. A key parameter in this investigation was the partitioning of the solvent, the monomer and the produced polymer in the two reaction phases. The latter also helped in the formulation of an explanation for the measured pressure profiles. Finally, the reaction heat rate data of the previous tests were used to discuss some safety aspects of the reaction, primarily through a cooling system failure scenario. Once the results on the model reaction were sufficiently confident, different reaction systems were tested to demonstrate the robustness and to explore the limits of the equipment and of the technique. It is shown that the calorimeter succeeds in detecting very small amounts of heat; therefore can monitor much less exothermic reactions, like the dispersion polymerization of styrene in supercritical carbon dioxide. On the other hand, the applicability of the technique is limited for very exothermic chemical reactions, like the precipitation polymerization of acrylic acid in supercritical carbon dioxide. Additionally, non-polymerization reaction systems can also be monitored and this is demonstrated with the example of the esterification of acetic anhydride with methanol in supercritical carbon dioxide. Finally, some preliminary tests were carried out to investigate the possibility of working with the high pressure reactor in a continuous mode but the results show that the developed set-up is not suited for such an application. The last part of the thesis deals with the use of the reaction calorimeter to produce polymeric foams using supercritical carbon dioxide as the blowing agent. The capability to control the reactor temperature with high accuracy and measure precisely the temperature and the pressure were exploited in this case. The equipment was initially calibrated in terms of the depressurization profiles that can be achieved and consequently polymer foams were produced. The effects on the final foam morphology of the polymer type, the temperature, the pressure, the depressurization rate and of the calorimeter operation were studied and various trends were identified.

EPFL2008

Reaction calorimetry in supercritical fluids

Sophie Fortini

This thesis is devoted to the study of the dispersion polymerization of methyl methacrylate in supercritical carbon dioxide (scCO2), using a poly(dimethylsiloxane) macromonomer (PDMS macromonomer) as stabilizer. Supercritical fluids (SCF) and SCF mixtures are characterized by a temperature and a pressure above their critical point(s), which is the last point on the vaporization line of a pure component. This means that these fluids operate from moderate to high pressure. They can be used in various processes ranging from extractions, nanoparticle formation for controlled drug release, chemical reactions and polymer processing. Nowadays, the best candidate for SCF processing is carbon dioxide. The fundamental motivation of using scCO2 as a solvent is based on its potential to replace harmful chemical organic volatile compounds (VOCs) in order to develop more sustainable and environmentally friendly chemical processes. At this point, the crucial role of CO2 in the development of the so-called "green chemistry" comes on the stage. CO2 is a natural abundant compound with low toxicity exhibiting no inflammability. This last property is very advantageous considering the cost investments spent by the chemical industry to control the safety of the chemical processes using highly flammable compounds like VOC solvents. As expected, environmental arguments are not sufficient to motivate the development of new chemical process routes. Therefore, additional arguments to use SCFs have to be found, and they do exist. As supercritical fluids are compressible fluids they can exhibit liquid-like and gas-like properties, which can be tuned easily by varying the operating conditions, like pressure and temperature. This fundamental behavior of SCF is their main asset and demonstrates their superiority to develop more flexible processes. The polymer industry is one of the industries that uses the largest volumes of organic solvents and sometimes halogenated ones, well known to destroy the ozone layer. The use of scCO2 gives to chemists and engineers the opportunity to develop more sustainable polymer processes, considering the numerous chemical and physical advantages of carbon dioxide. The processing of scCO2 for polymer production is no more than fifteen years-old. This means that a certain quantity of knowledge has been acquired but still a lot of unknowns hinder their promotion at industrial level. This work is inserted in this context and finds there its main motivations. This thesis is composed of two different but intrinsically connected approaches of the dispersion polymerization of the methyl methacrylate (MMA) in scCO2. A part of this thesis is devoted to the development of techniques allowing the on-line monitoring of polymerizations in scCO2 at "larger" scale, conducted from an engineering approach. The intrinsically connected part is devoted to the understanding of the fundamental phenomena that govern the dispersion polymerization of MMA in scCO2, its kinetic features and the product characteristics of the polymer produced, being conducted from a chemical approach of the subject. Up to now, most of the studies dealing with polymerization reactions in scCO2 are realized in small autoclaves between 2 and 60 ml allowing pertinent fundamental analysis but with poor similarities with an industrial reactor. A keystone of this work is the development of a supercritical reaction calorimeter composed of a 1.3 liters high pressure reactor allowing the kinetic study of the dispersion polymerization, which in turn gives a direct insight into the parameters that control the dispersion polymerization stability, efficiency and mechanism. Based on an adapted heat balance, the calorimeter can give the profile of monomer conversion (thermal conversion). Furthermore, the volume of the reactor allows inserting on-line sensors inside the reactor. This possibility led to the development of an ultrasonic sensor to monitor the polymerization process. The combination between the calorimetric information and the sensor signal shows the potential of these sensors to monitor polymerization reactions in scCO2. By measuring the speed of sound evolution during the course of the polymerization, it is possible to calculate the composition of the medium and thus evaluate the monomer conversion. The analysis of the effects of temperature, stirring speed and impeller types demonstrate that the dispersion polymerization of MMA can be effective under a wide range of operating conditions using the PDMS macromonomer as stabilizer. The experiments point out that the stability of the dispersion polymerization and in turn the rate of polymerization, the polymerization loci1 and the polymer quality depend strongly on the stabilizer concentration but more fundamentally on its degree of solubility in the carbon dioxide. Phase behavior measurements demonstrate that the 5'000 g/mol PDMS macromonomer exhibits a relative good solubility in carbon dioxide and that this solubility can be greatly improved by the presence of the monomer in the mixture. In fact, the monomer acts as a cosolvent for the PDMS macromonomer improving the solvency of the scCO2 (polarity, density). This study demonstrates that the concentration of the monomer in the reacting medium is a key parameter to control the stability of the dispersion throughout the polymerization process. A dispersion polymerization is characterized by polymer-rich particles dispersed in a continuous phase, i.e. the CO2-rich continuous phase. The results give evidence of the fact that the polymerization can take place in both phases depending on the concentration of stabilizer and its solubility in the reacting medium. In this case, bimodal molecular weight distributions and intermediate rate of polymerization are observed. When only one reaction locus is active, monomodal molecular weight distributions are obtained. Furthermore, the locus of the polymerization influences directly the rate at which the polymer is produced and the degree of polymerization. In order to complete the study, a model has been developed demonstrating that diffusion limitations are operative in the dispersion polymerization of MMA in scCO2. A gel effect is present in the polymer-rich phase that leads to the increase of the molecular weight of the polymer produced, as observed also experimentally. Moreover, the presence of this gel effect occurring inside the polymer-rich particles explains the auto-acceleration of the polymerization rate as the conversion increases. The model demonstrates that in the case of an effective dispersion polymerization of the methyl methacrylate in scCO2 the main reaction loci are the polymer-rich particles. ______________________________ 1 A dispersion polymerization is composed of polymer-rich particles dispersed in the CO2-rich continuous phase. The polymerization locus is referred to the phase into which the polymerization can take place.

EPFL2006

Photocatalytic and Electrocatalytic Reduction of Carbon Dioxide in Pressurized Systems

Patrick Voyame

The depletion of carbon-based fossil fuels and the rise in atmospheric carbon dioxide concentration will force an inevitable change in the future global energy landscape. CO2 reduction presents the advantages of decreasing its atmospheric concentration and storing energy in chemical form in CO2 reduction products. With a predicted conversion to renewable energy such as solar or wind energy, energy storage will become a key process in the near future for buffering the fluctuating energy production. The objective of the present work was to study the efficiency towards CO2 reduction of different molecular catalysts. In particular, conducting experiments in high-pressure and supercritical conditions and observing the effect of CO2 pressure or concentration on the efficiency and selectivity of the reaction. As supercritical CO2 (scCO2) has poor solubilisation capabilities, experiments were conducted in biphasic systems or with addition of an organic co-solvent. To drive the reduction reaction of CO2, a catalyst is needed to overcome the kinetic limitations of the reaction, but an energy input is also necessary. Three different forms for this energy input were used in this work. In a first time a sacrificial product, decamethylferrocene (DMFc), was used to transfer electrons to CO2 in biphasic water/scCO2 systems. Complete oxidation of the DMFc was observed in presence of anion capable of transporting protons from water to the DMFc present in the supercritical phase. A photosensitization cycle was used to supply a water soluble catalyst, NI(II)Cyclam, in electron at the required potential to drive the reduction of CO2 into carbon monoxide in water/scCO2 system. The creation of the interface in the system appeared highly favourable to the efficiency of the catalyst. A second catalyst, a ruthenium polypyridyl carbonyl complex, was used for the photocatalytic reduction of CO2. Pressure had an important impact on the production of one of the two reduction product, carbon monoxide, while the production of formate was unaffected by CO2 pressure. As limitations in productivity in photocatalytic experiments were coming principally from the photosensitizer cycle or the sacrificial electron donor, photosensitizer was replaced by an electrode to provide the catalyst in electrons. Voltammetry in CO2-expanded liquids was described and determination of important parameters such as catalyst concentration and diffusion coefficient as a function of pressure was performed. Electrocatalytic reduction of CO2 by ruthenium and rhenium polypyridyl carbonyl complexes was studied in CO2-expanded liquids. The catalytic mechanisms were observed to be highly influenced by CO2 concentration.

EPFL2016