Chemical Engineering and Processing 83(2014)26–32
Contents lists available at ScienceDirect
Chemical Engineering and Processing:
Process
Intensification
j o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /c e
p
Review
Catalyst retention in continuous flow with supercritical carbon dioxide
S.C.Stouten,T.Noël,Q.Wang,V.Hessel ∗
Laboratory of Chemical Reactor Engineering/Micro Flow Chemistry and Process Technology,Department of Chemical Engineering and Chemistry,Eindhoven University of Technology,PO Box 513,5600MB Eindhoven,The Netherlands
a r t i c l e
i n f o
Article history:
Received 26November 2013
Received in revised form 5March 2014Accepted 18March 2014
Available online 18July 2014
Keywords:
Supercritical carbon dioxide Continuous flow
Homogeneous catalysis Catalyst retention
a b s t r a c t
This review discusses the retention of organometallic catalysts in continuous flow processes utilizing supercritical carbon dioxide.Due to its innovative properties,supercritical carbon dioxide offers inter-esting possibilities for process intensification.As a result of safety and cost considerations,processes that use supercritical carbon dioxide are preferably done in continuous flow,as they require a pressure upwards of 74bar.Many of the reactions that benefit from the application of supercritical carbon diox-ide also involve the use of a homogeneous catalyst however,requiring efforts to recycle the catalyst when these are applied in continuous flow.Alternatively,the catalyst may be retained in the reactor by modifying the process or catalyst,such as by catalyst immobilization,membrane separation,or biphasic processing exploiting the properties of supercritical carbon dioxide.Each of these methods is discussed,including their advantages and drawbacks.Also discussed are milli-and micro-flow processes and their possibilities for integrated catalyst retention and handling supercritical carbon dioxi
de.
©2014Elsevier B.V.All rights reserved.
Contents 1.Introduction ..........................................................................................................................................262.Systems using catalyst immobilization ..............................................................................................................273.Membrane 284.Biphasic systems ......................................................................................................................................285.Supported ionic liquid 296.Continuous milli-and micro-flow systems ..........................................................................................................307.
Conclusions ...........................................................................................................................................31References ............................................................................................................................................
31
1.Introduction
Typically,the use of solvents in chemical processes is a neces-sary evil.Solvents inevitably lead to waste,yet are often required either as a medium for a chemical reaction or for downstream sep-aration [1,2].Particularly separation and purification require large amounts of solvent.In some processes,reducing or eliminating the use of solvents may be a possibility,but in most cases this is not an option.Alternatively,green solvents may be employed,such as supercritical carbon dioxide (scCO 2),which is cheap,nonflammable
∗Corresponding author.
E-mail address:v.hessel@tue.nl (V.Hessel).
and non-toxic [3–6].In addition,at depressurization scCO 2will revert to a gas,meaning no solvent residue will remain to con-taminate the final product.An example of a well-known industrial process in which these advantages are exploited is the extraction of caffeine from coffee beans [7].
Beyond this however,the use of an innovative solvent such as scCO 2also provides opportunities for pr
ocess intensification.At supercritical conditions,above 73.9bar and 304.25K,the proper-ties of carbon dioxide are somewhere in between that of a gas and a liquid.As a result,scCO 2has a high solvent power,high diffusiv-ity and can accelerate gas–liquid reactions by allowing operation under single phase conditions [8].Additionally,around the super-critical point these properties change rapidly with temperature and pressure,which provides interesting opportunities.For example,
/10.p.2014.03.017
0255-2701/©2014Elsevier B.V.All rights reserved.
S.C.Stouten et al./Chemical Engineering and Processing83(2014)26–3227
small adjustments in process conditions may result in large changes to the solubility of dissolved species to allow separation.Of course, such advantages are best exploited when process conditions can be controlled andfine-tuned to the reaction,as is the case in continu-ousflow processes.In addition,operation in a continuous process avoids the use of large pressure vessels that would be required for batch production,which are undesired due to safety issues and high costs.
Still,the use of scCO2as a solvent also comes with a price tag. While CO2’s critical pressure is fairly lo
w,compression is neverthe-less energy intensive and this energy is lost upon depressurization. Hence,scCO2should be applied in processes in which it provides a distinct benefit,as was suggested by Han and Poliakoff[9].Impor-tant reactions that have been shown to benefit from the use of scCO2,such as hydroformylation and asymmetric hydrogenation, rely on the use of expensive homogeneous catalysts.Therefore, improving the performance of such catalysts by use of scCO2is particularly worthwhile.
However,as was just argued,scCO2is best applied in a continuous process,as are the aforementioned processes.For homogeneous catalysts,which are dissolved in the reaction mix-ture,this means a method for either catalyst recycling or retention has to be found.The goal of such retention methods is to extend the use of the organometallic catalyst beyond that of a single pass through the reactor,effectively increasing the catalyst’s turnover number(TON).Several approaches to this have been developed, for instance by catalyst immobilization[10,11],membrane reactors [12,13],biphasic processes[14],or by a down-stream separation exploiting the properties of scCO2[15].The aim of this review is to provide an overview of these approaches for catalyst retention in scCO2processes.
2.Systems using catalyst immobilization
Possibly the most obvious approach to achieve catalyst reten-tion is by the immobilization of an organometallic catalyst onto a solid support,thereby creating a heterogenized version of the catalyst[10,11].Through one of its ligands,the organometallic cat-alyst is anchored to a support material.In most cases,catalytic performance suffers from this modification of the existing catalytic complex,although the mobility and performance of the immobi-lized catalyst can be improved by using a spacer between ligand and support.Catalyst leaching is an issue that may occur if the metal centre dissociates from the ligand to which it is anchored,which is why the use of multi-dentate ligands is preferred.It should also be noted that the“heterogenized”organometallic complex will be nowhere near as robust as a true heterogeneous catalyst.
Although the use of scCO2offers many advantages,many homo-geneous catalysts are poorly soluble in scCO2by themselves.To overcome this,catalysts are often modified achieve better solubil-ity[16–19].However,this does not address the catalyst retention problem for continuousflow systems.Instead,the catalyst could be immobilized on a support material,allowing both a better dis-tribution of the catalyst over the reactor and the retention of the catalyst.
Application of an immobilized cobalt complex was investigated by Lu et al.,for the conversion of CO2and ethylene oxide to eth-ylene carbonate under supercritical conditions,see Fig.1[20].In the cour
se of this reaction,the product that is formed will create a second phase which can dissolve the homogeneous catalyst,leach-ing it away from the supercritical phase and causing loss of activity
[21].This limitation was overcome by immobilizing the catalyst on
a support,improving mixing with the supercritical phase.In addi-tion,the immobilized catalyst could be used in continuousflow for 24h without loss of activity.A conversion of86%was achieved,with no formation of
by-products.
Fig.1.Continuous cycloaddition of CO2over an immobilized cobalt complex.
Redrawn from Lu et al.[20].
In the industrial hydroformylation process,catalyst separation is an important issue,particularly for reactions involving alkenes that are longer than6C atoms[22].While for shorter alkenes cat-alyst and product can be separated by distillation,the thermal sensitivity of the catalyst limits this approach to products with a relatively low boiling point.Another approach is the use of an aque-ous biphasic process;yet again this process is not feasible for longer alkenes,due to their low aqueous solubility.Meehan et al.inves-tigated the continuous hydroformylation of1-octene in scCO2by using a silica-immobilized rhodium catalyst,see Scheme1[23].At 80◦C,the observed turnover frequency(TOF)in scCO2was117h−1, which was much higher than for the batch reaction in toluene (TOF=35h−1),but lower than observed with a free homogeneous catalyst(TOF=283h−1).Performance of the catalyst was shown to be stable over30h,with no detectable rhodium leaching(<0.2%). Utilizing a controlled two-step depressurization of the CO2,about 90%of the1-octene could be removed from the product.In work by Bronger et al.,when a catalyst with a phenoxaphosphino-modified Xanthphos-type ligand was used,
see Fig.2,catalyst stability was improved in scCO2,but no activity enhancement was observed[24].
Asymmetric hydrogenation is an important tool for selective production of specific enantiomers,used in,for example,the pro-duction of pharmaceuticals such as l-DOPA[25].However,while hydrogenation is also possible over a heterogeneous catalyst,the enantioselective hydrogenation can only be achieved by use of a homogeneous catalyst.Asymmetric hydrogenation of dimethyl ita-conate with an immobilized rhodium catalyst was investigated by Stephenson et al.,revealing stable performance up to100◦C[26]. After further optimization of the ligand,an enantiomeric excess(ee) of over80%could be obtained,which even exceeded reported val-ues for hydrogenation with a free homogeneous catalyst in batch [27].
While the most common approach to immobilization is by anchoring the catalyst through a covalent bond,alternative meth-ods exist,such as by encapsulation of the catalyst.In this approach, instead of utilizing a covalent bond for immobilization,the cat-alyst is bound by trapping it inside the support structure.
Leeke Fig.2.Silica-immobilized ligand with a spacer used to anchor a hydroformylation
catalyst.
Redrawn from Bronger et al.[24].
28
S.C.Stouten et al./Chemical Engineering and Processing 83(2014)
26–32
Scheme 1.Hydroformylation of 1-octene over a rhodium catalyst.
et al.investigated the Suzuki–Miyaura cross-coupling reaction between p -tolylboronic acid and iodobenzene in toluene/methanol as well as in scCO 2,using the commercially available PdEn-Cat,see Scheme 2[28,29].In scCO 2,a conversion of 81%was obtained,compared to only 74%in organic solvent.The improved performance in scCO 2was attributed to a higher Reynolds num-ber and much lower Schmidt number for that system.This was quantified by a calculation of the mass transfer rates,which gave values of 5.82×10−4mol dm −3s −1in toluene/methanol and 1.46×10−3mol dm −3s −1in scCO 2.
Mostly,the support with the immobilized catalyst is placed in a packed bed reactor.However,due to their random structure,packed beds also have several drawbacks,such as the forma-tion of hot spots and a
large residence time distribution [30].To overcome the drawbacks of a packed bed,Burguete et al.immo-bilized a Ru-bisoxazoline catalyst on a monolithic polymer [31].Dichloromethane and scCO 2were then compared as solvents in the continuous asymmetric cyclopropanation between styrene and ethyl diazoacetate,using a monolithic milli-flow reactor.In scCO 2,a 7.7-fold increase in catalyst productivity was observed over the reaction in dichloromethane.After optimization,this productivity increase could be further increased to 16–17-fold [32].In addition,a 20%increase in enantioselectivity was reported compared to the use of a homogeneous catalyst in a batch process.Catalyst leaching was less than 1ppm.
3.Membrane reactor systems
Another seemingly straightforward method to retain a catalyst inside the reactor is to cover the exit of the reactor with a membrane [12,13].The pores of the membrane must be small enough to retain the catalyst,yet large enough to allow the product to pass through.Unfortunately,currently available membranes still have a consider-able pore size distribution,so membranes with a sharp molecular weight cut-off (MWCO)are not available.To overcome problems from a too small difference in the molecular weight between cata-lyst and product,the size of the catalyst can be increased by using enlarged ligands.Since enzymes are already very large molecules,the use of membrane reactors has
been particularly popular for enzyme catalysis.
As a consequence of the single phase-conditions that scCO 2allows,catalyst retention with a simple membrane reactor becomes possible for gas–liquid reactions.Normally,membrane separa-tion for gas–liquid systems must be done down-stream,since the gas would permeate through the membrane much more easily than liquids.For the type of membrane used with scCO 2,
ceramic
Scheme 2.Suzuki cross-coupling reaction in continuous flow using an encapsulated palladium catalyst investigated by Leeke et al.[28].
membranes are preferred over organic membranes,due to their increased stability and resistance to swelling [33].
Goetheer et al.investigated the use of a membrane reactor for the hydrogenation of 1-butene in scCO 2,see Fig.3[34].A fluorous derivative of Wilkinson’s catalyst was used,with a size of 2–4nm,much larger than the pore diameter of the silica membrane,which was 0.5–0.8nm.Continuous production of n -buta
ne was achieved at 353K and 200bar,with a TON of 1.2×105after stable operation over 32h.No detectable leaching of the catalyst was observed.A TOF of 9400was obtained in continuous flow with scCO 2,over 10times the value for organic solvent.
The influence of reaction parameters on the hydroformylation of 1-octene in scCO 2was studied by Koeken et al.,using a scCO 2-soluble homogeneous catalyst with fluorous tails [35,36].In later work,the reaction was performed in continuous flow using a mem-brane reactor [37].Unfortunately,catalytic activity was lost over time,likely attributed to incomplete retention of the catalytic com-plex by the titania membrane.
4.Biphasic systems
Another approach to catalyst retention is by using a biphasic sys-tem [38].For biphasic processes with scCO 2,the catalyst is retained in a stationary phase,while the reactants and products are trans-ported in and out of the reactor by the mobile supercritical phase.Naturally,the challenge is to find a system in which enough mass transfer occurs between the two phases,without leaching of the stationary phase and catalyst.
Ionic liquids (IL)are salts that are liquid at room temperature,which have a negligible vapour pressure a
nd whose properties are tuneable by varying the anion and cation,see Fig.4[39].The combi-nation of scCO 2with IL in particular has drawn a lot of interest since it was shown that while scCO 2dissolves well in certain IL’s,these IL’s are completely insoluble in scCO 2[40].Furthermore,although the viscosity of IL’s is typically very high,the presence of scCO 2decreases this viscosity considerably,thus improving mass trans-fer.Considering these properties,such a biphasic system shows much promise for catalyst retention.
Bösmann et al.applied the use of a scCO 2/IL biphasic system to the continuous hydrovinylation of styrene by using a
Ni-catalystreactor technology
Fig.3.Membrane reactor for the continuous hydrogenation of 1-butene as investi-gated by Goetheer et al.[34]
.
Fig.4.Structure of the common ionic liquid 1-butyl-3-methyl-imidazolium hex-afluorophosphate [BMIM]PF 6.
S.C.Stouten et al./Chemical Engineering and Processing83(2014)26–32
29
Fig.5.Continuous-flow process for CO2hydrogenation at supercritical conditions utilizing IL’s.
Redrawn from Wesselbaum et al.[48].
[41].By tuning the anion and cation of the IL,the conversion and ee values of the hydrovinylation could be increased significantly. Inflow experiments,stable catalytic activity was shown for61h, while in a comparable batch system the catalyst deactivated after only several uses.
Sellin et al.investigated the use of the scCO2/IL system for the continuous hydroformylation of1-octene[42].Stable conversion was shown for over20h,with less than1ppm leaching of the rhodium catalyst.To obtain a catalytic complex that would be sol-uble in the IL,but not in scCO2,an IL-based ligand was used.This system for continuous hydroformylation was then further opti-mized,finally achieving a TOF of500h−1with only0.012ppm rhodium leaching and with stable operation for at least72h[43,44]. The l:b ratio of2–3was still low however,yet this was improved to a ratio of40:1by use of a xantphos-derived ligand[45].Unfor-tunately,the xantphos-derived ligand proved less resistant to oxidation,resulting in a reduced performance over time.
In an alternative approach,the IL-based catalyst was dissolved solely in the mixture of reactant and product[46].This elimi-nated the use of a large amount of expensive IL and allowed the high operating pressure,necessary for extraction,to be reduced to125bar.However,this also reduced the TOF to162h−
1and increased rhodium leaching to0.1–0.5ppm.Further optimization of this approach was achieved by balancing the contents of the reac-tion mixture and increasing the pressure to140bar,increasing the TOF to180h−1and decreasing rhodium leaching to0.1ppm[47].
The hydrogenation of CO2to formic acid was investigated by Wesselbaum et al.[48],utilizing an organometallic ruthenium cat-alyst retained in an ionic liquid,see Fig.5.Due to the unfavourable thermodynamic equilibrium for this reaction,formic acid needs to be removed to achieve good conversion.While other methods rely on formation of salts,adducts or derivatives to achieve this,the integrated reaction and extraction in the IL/scCO2system allows formic acid to be obtained directly.With a TOF of314h−1,this process performs at least as good as those in conventional sol-vents.However,the rate of product extraction was still considered a limiting factor that would require further optimization.
Asymmetric hydrogenation of-keto esters in IL/scCO2was investigated by Theuerkauf et al.and was compared to a similar batch process[49].Catalyst activity remained at91%after100h on stream and69%after150h,with deactivation attributed to traces of water.Based on the production of100kg of the product methyl pro-pionylacetate,it was estimated that a batch system would require 14times as much catalyst and almost40times as much reaction solvent,with the equipment size being ten times lar
ger.Still,the continuousflow system with IL/scCO2also has a reduced enantiose-lectivity(80–82%vs.97–98%),which is a major drawback.Also,the continuousflow system requires a far higher operating pressure (250bar vs.4
bar).Fig.6.Conceptual process for continuous photo-oxidation with1O2viafluorous biphasic catalysis.
Redrawn from Hall et al.[15].
For photo-oxidation reactions with singlet oxygen(1O2),the type of solvent is of considerable importance as it affects the life-time of1O2.While CCl4has typically been the solvent of choice, its toxicity requires that a greener alternative is found.One of the few solvents that provide the necessary stability for1O2is scCO2. Furthermore,scCO2is also completely miscible with O2and is non-flammable.Continuous photo-oxidation with1O2using scCO2as a solvent was initially investigated by Bourne et al.with a dissolved photosensitiser catalyst[50,51].After promising results with the photo-oxidation of␣-terpinine,catalyst immobilization was inves-tigated by Han et al.[52].While successful,the immobilization also led to an undesired reduction of the lifetime of the catalyst due to degradation of the support.An alternative biphasic process was subsequently investigated by Hall et al.,using a catalyst dissolved influorous solvent,see Fig.6[15].In this system the catalyst and fluorous solvent are soluble in scCO2under reaction conditions, but not after depressurization,allowing downstream recovery and recycling.To create a catalyst which would preferentially dissolve in thefluorous solvent upon depressurization,the catalyst was derivatised with‘fluorous ponytails’.This approach allowed the cat-alyst to be recycled10times,although leaching of the catalyst into the product phase still occurred.In further work,application of the fluorous biphasic system was also investigated for the synthesis of antimalarial trioxanes[53].
5.Supported ionic liquid phase systems
Although the use of IL/scCO2systems looks very promising,this approach also has several limitations.As mentioned previously,one of the issues is that large volumes of IL are expensive.Another recur-ring issue is the slow extraction of product from the IL phase by scCO2,possibly as a result of diffusion limitation in the large vol-ume of IL.Both of these issues could be alleviated by an approach that requires smaller volumes of ionic liquid.This could be achieved by supporting the catalyst in a thinfilm of ionic liquid on a porous support,a so-called supported ionic liquid phase(SILP)catalyst,see Fig.7[54]
.
Fig.7.Concept of supported ionic liquid phase catalysis with scCO2,investigated by Hintermair et al.[55].
30S.C.Stouten et al./Chemical Engineering and Processing83(2014)26–32
Hydroformylation of1-octene with a SILP catalyst in scCO2was investigated by Hintermair et al.[56].With this approach a TOF of800h−1was obtained,an improvement over the previously dis-cussed system using IL/scCO2(527h−1),performing very good in comparison with processes for commercial propene hydroformyla-tion(550–700h−1)[57].In addition,the required pressure was only 100bar,in comparison with the200bar for the IL/scCO2system. Catalytic activity was stable over40h,although catalyst leaching stabilized at0.5ppm,which is relatively high.
The SILP/scCO2system was also applied in the enantioselective hydrogenation dimethyl itaconate,using a QUINAPHOS-Rh cata-lyst[58].Initially,over99%ee was obtained at full conversion,with TOF values over2000h−1and a space time yield of0.3kg L−1h−1. However,enantioselectivity suffered over time,reducing enan-tioselectivity to around70–75%ee for a run up to65h.Further investigation revealed that the drop in ee could be attributed to deactivation of the catalyst by water[55].A more stable per-formance was achieved by using a hydrophobic support surfac
e obtained byfluorination,in combination with a more hydrophobic IL.The best obtained space time yield was0.7kg L−1h−1,produc-ing100kg of product per gram rhodium.An automated,pilot scale SILP/scCO2system was investigated for its application in research as well as production,showing it can be constructed using commer-cial parts with accurate control of process conditions and integrated product separation and analysis[59].
6.Continuous milli-and micro-flow systems
Milli-and micro-flow systems offer several significant advan-tages,such as increased safety,improved heat and mass transfer and the opportunity for process intensification[60,61].Such inten-sification has an impact up to the full-process scale,enabling unique process advantages,as listed by Hessel et al.[62].Still,as the appli-cation of these systems requires an initial investment,the benefit must be sufficient to offset this,not unlike with the use of scCO2. Therefore,the application of milli-or micro-flow systems will be particularly interesting for the processes that have been discussed in the previous chapters.The improved heat and mass transfer in milli-or micro-flow systems can provide a boost to catalytic activ-ity,while the increased safety allows safe use of high pressure scCO2 without the need for extensive safety precautions[63–68].Yet,the improved control over process conditions goes further than just an optimization.Around the supercritical point,the properties of scCO2vary strongly with proce
ss conditions.To exploit this for cat-alyst retention,thefine control offered by milli-and micro-flow systems can be of great benefit.Finally,catalyst retention may be integrated more easily by utilizing milli-and micro-flow processes. To realize the benefit of these advantages however,a change in mindset towards interdisciplinarity will also be required.
The advantages of a milli-flow system can be observed in the previously mentioned biphasic system for continuous photocat-alytic oxidation with singlet O2,investigated by Poliakoff and co-workers[15,52,53].In reactors with a large diameter,illumi-nation of the catalyst is limited to the outer radius of the reactor, due to the path length of the light being limited as following from the Lambert–Beer law.In a milli-flow system on the other hand,the entire reactor volume can be illuminated.Initially,the immobiliza-tion of the photosensitiser was investigated as a retention method [52].After exploring several different options for the photosen-sitiser,tetradi(2,6)chloro-phenylporphyrin(TDCPP)was found to have both the required activity and lifetime.By bonding the TDCPP to amino-functionalised PVC beads,a supported photosensitiser with high activity was obtained,with no observed leaching.For photo-oxidation of␣-terpentine and citronellol,85%and88%yields were obtained respectively.After more than5h on stream,the
yield Fig.8.Microreactor design for the continuous hydrogenation of cyclohexene in scCO2,by Trachsel et al.[62].
Reproduced with permission from Elsevier.
for␣-terpentine photo-oxidation remained unchanged,while that for citronellol had dropped slightly,to78%.Still,the lifetime of this immobilized catalyst was limited by the degradation of the sup-port under reaction conditions.In a subsequent study,the retention method was therefore changed to a biphasic system with afluorous solvent[15].Thefluorous solvent,hydrofluoroether-7500(HFE-7500),formed a single phase with scCO2under reaction conditions only,allowing phase separation afterwards.To increase solubility in the HFE-7500/scCO2system,the photosensitiser,5,10,15,20-tetrakis-(pentafluorophenyl)porphyrin(TPFPP),was derivatised with‘fluorous ponytails’.
Another milli-flow system that has been discussed is the mono-lithic miniflow reactor investigated by Burguete et al.[31,32],to overcome problems associated with packed bed reactors[30].For a continuous asymmetric cyclopropanation reaction,theflow sys-tem provided20%higher enantioselectivity than in a homogeneous batch system,although a comparison with a packed bed reactor was unfortunately not reported.
Several more examples offlow reactions in scCO2have been demonstrated and,together with examples offlow reactions using other supercritical solvents,are summarized exclusively on that topic[68],or as part about process intensification overviews[69]. The focus of nearly all studies is on boosting the reaction speed, productivity,and space-time yield.The esterification of phthalic anhydride with methanol using scCO2as co-solvent at110bar and60◦C was shown to proceed5400-fold faster as compared to standardflow conditions[60].The hydrogenation of cyclohexene in scCO2was increased by one order of magnitude as compared to standardflow conditions[67],see Fig.8.The photo-oxidation of citronellol(within the synthesis of the commercial fragrance Rose Oxide)using scCO2as co-solvent at180bar had a space–time yield of70mmol l−1min−1as compared to0.1mmol l−1min−1for
a conventional Schlenk reactor with an immersed LED array[51].
7.Conclusions
In this review,a variety of methods have been discussed for catalyst retention in scCO2processes.In all cases,extended and
版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。
发表评论