User:Eddie McCabe

1. Introduction Carbogenic molecular sieves (CMS) are nanoporous membranes mainly composed of carbon with a few impurities of some other elements. These materials have molecular sieving properties due to their pore size and internal structure. Basis of adsorption in this kind of membranes is size and shape of molecules. As Foley [1] indicated “molecular sieving is an operational or a phenomenological description and not one based inherently on the structure of the material under examination”. One of similar membranes commonly confused with CMS membranes are activated carbons. Although both are mainly composed of carbon, they adopt different mechanisms in separation. Activated carbons separate different molecules based on their different adsorption equilibrium constants. On the other hand, CMS membranes perform separation based on the rate of adsorption. Another category of molecular membranes are zeolitic membranes. Unlike activated carbons zeolites have different constituting elements. However, they expose the same behavior in separating molecules. To reemphasize the importance of “operational definition” of molecular sieving, we should note that zeolites and CMS materials are both molecular sieves. However, the first group is generally crystalline while the second group is globally amorphous. 2. fabrication of carbogenic molecular sieves To fabricate a CMS membrane we pyrolyze a natural or synthetic precursor under controlled conditions. The whole fabrication process might possibly include some pretreatment and/or post-treatment stage so as to adjust the quality of product. For the pyrolysis process to be efficient it should have at least 25-50% carbon yield on the basis of precursor mass [1]. Generally, the fabrication of carbogenic membranes is comprised of six consecutive steps, as shown in Figure 1. The heart of this process is the pyrolysis process. It is when nanoporous structure of carbon membrane rises which is the basis for its interesting separation properties. 2.1. Precursor selection Carbogenic membranes are generally produced from a number of different carbon containing natural or synthetic materials. This includes a variety of synthetic and natural polymers, thermosetting resins, graphite, coal, pitch and plants, under inert atmosphere or vacuum conditions during carbonization [1, 2]. Decomposition of synthetic polymeric precursors produces some light gases while leaving behind a carbogenic porous material. For example carbonization of widely used synthetic precursor polyfurfuryl alcohol (PFA) produces carbon dioxide, water, and methane in various compositions depending on pyrolysis conditions leaving behind pyrolyzed PFA as the main product [1, 2].

Figure 1. Carbon membrane fabrication process [2]. Fabrication of membrane using these synthetic precursors results in a highly pure carbon membrane. However the cost of producing such molecular sieves is an important barrier to their commercialization. At the moment commercial CMS membranes are all prepared from naturally occurring precursors like anthracite coal. These have found attractive application in industrial separation of nitrogen and oxygen. Thermosetting polymers are often resistant to very high temperatures which make them favorable for CMS membrane fabrication. This thermal resistant is of paramount importance due to high temperature requirement of different stages during the fabrication process specially carbonization. This has directed most of recent work toward the application of such precursors [2]. 2.2. Polymeric membrane preparation From here on we focus more on synthetic precursors since they have more favorable characteristics as we mentioned above. First by polymeric membrane we mean uncarbonized precursor which undergoes pyrolysis during the next stages of fabrication. These precursors are cast in two major configurations, unsupported membranes and supported membranes. Unsupported membranes themselves might be flat (film), hollow fiber, or capillary. On the other hand, supported membranes might be flat or tube regarding their configuration. Supported polymeric membranes are used most of the time, because of the poor mechanical stability of unsupported carbon membranes [2].

Figure 2. Precursor pretreatment methods.

2.3. Pretreatment The goal of pretreatment is increasing the stability of the prepared polymeric membrane which further goes under harsh condition of pyrolysis. There are a number of pretreatment schemes used so far among which the most popular one has been chemical oxidation pretreatment. Generally pretreatment methods fall into two category of chemical and physical methods. Physical ones are mostly used for hollow fiber membranes and include stretching or drawing. On the other hand chemical pretreatments consist of introducing a chemical reagent to the prepared polymer membrane. Figure 2 summarizes different pretreatment methods [2]. 2.4. Pyrolysis Pyrolysis or carbonization is the process of heating polymeric precursor in a controlled atmosphere (vacuum or inert) up to the pyrolysis temperature. Heating rate, thermal soak time and final temperature might be controlled for specific design considerations [2]. As we mentioned before this is the most important stage of fabrication process in which micropores with dimensions comparable to molecular size are produced. These micropores are the basis of molecular sieving behavior in such membranes. During carbonization, release of significant amounts of light gases result in appreciable weight loss of precursor. This weight loss plus micropore generation leads to high surface area product. Typical volatile byproducts frequently observed in pyrolysis of different precursors include ammonia, hydrogen cyanide, hydrogen, nitrogen, carbon monoxide, carbon dioxide and methane, which might or might not release depending on precursor nature [1, 2]. An important reaction coupled to CMS preparation procedure is the cross-linking which is either conducted during precursor preparation or pyrolysis. It is the cross-linking that is responsible for globally amorphous nanoporous structure of CMS. This prevents formation of crystalline graphitic carbon during carbonization, otherwise produced due to its favorable state at high temperatures of pyrolysis. Figure 3 shows the disordered structure that comes about due to cross-linking [2, 3]. During the pyrolysis a pore structure composed of two different kinds of porosity with different sizes is formed. Smaller pores which are narrowly distributed in size are connected via some larger pores that work as conduits for gas molecule permeation. In this regard we have some “wide openings” plus a few “constrictions” as shown in Figure 4. Small di denotes those “ultramicropores” that are responsible for molecular sieving properties (< 10 Å). The larger pores denoted by Di (6–20 Å) allow gas diffusion through the porous structure. Special attention should be paid during the preparation of a microporous structure so that it would have both kinds of the above pores and therefore it can behave as a molecular sieve while providing relatively high gas diffusion through its structure [2].

Figure 3. Cross-linking and evolution of disordered nanoporous structures (non-graphiting carbons) [1].

Figure 4. Idealized structure of a pore in a carbon material [2]. The underpinning discussion that governs CMS evolution is as follows [3]; when we heat the precursor amorphous sp2 carbon gradually transforms into aromatic domains that become larger with time and temperature (Figure 5). As this evolution progresses, we will have more ordered structures; also it tends to form open domains with high free volume. Carbon precursor ultimately converts to graphite if sufficient time and energy are provided. CMS materials from polymer pyrolysis are kinetically-frozen and metastable. By virtue of this evolutionary trajectory we can control pore size distribution of the CMS materials.

Figure 5. Evolution of CMS structure [3]. 2.5. Post-treatment This is step is in fact a means towards adjusting pore size dimension and distribution in pyrolyzed carbon. This stage might include a thermochemical treatment like post-oxidation, chemical vapor deposition (CVD), post-pyrolysis and coating. In addition to previously mentioned effect this process might be beneficial in repairing minor defects in porous structure [2]. 2.6. Module Construction Module is the device in which the membrane is installed. Important considerations for the preparation of membrane modules are summarized in Table 1 [2]. Table 1. Important considerations for the preparation of a membrane module [2].

3. Potential catalytic applications of CMS Due to molecular sieving properties that CMS materials exhibit, they can be employed for shape selective catalysis. This includes their application as support as well as catalyst itself. The first application consists of embedding metal catalyst particles or metal oxides inside the porous structure (which is generally done during pyrolysis).The second potential application has been proven by observing their active sites capable of performing dehydrogenation catalysis. The ultramicroporous structure of CMS materials gives rise to both reactant and product shape selectivity (Figure 6).

Figure 6. Schematics of shape selectivity In Figure 7, a number of catalytic applications for CMS materials are shown. Among these, we will discuss hydrogenation and Fischer-Tropsch synthesis briefly as examples of reactant and product shape selectivity respectively [1]. 3.1. Hydrogenation, Example of Reactant Shape Selectivity The application of CMS materials for reactant shape selectivity goes back to 1970s when Trimm and Cooper [4] studied Platinum containing CMS (Pt-CMS) fabricated by reduction of chloroplatinic acid in PFA during pyrolysis of the later. Their studies show that these catalysts exhibit reactant shape selectivity in hydrogenation of olefins. In other words, they showed that Pt-CMS catalyst hydrogenates small linear olefins more efficiently than small branched olefins. Some other researchers have demonstrated reactant shape selectivity with composite catalysts. These catalysts were made of Pt supported on activated carbon or Fe supported on silica mixed with PFA and hydrolyzed. These catalytic structures have been able of selectively hydrogenating propene against isobutene [1, 3]. For some novel applications of CMS containing catalytic materials refer to reference [5]. 3.2. Fischer–Tropsch Synthesis, Example of Product Shape Selectivity Fischer–Tropsch synthesis refers to a complex network of chemical reactions that convert a mixture of carbon monoxide and hydrogen into liquid hydrocarbons. Generally reactions producing alkanes are useful and take place as follows: (2n + 1) H2 + n CO → CnH(2n+2) + n H2O Where n is a positive integer. Methane is generally an undesirable product as it is gaseous at room temperature. Alkanes produced are mostly unbranched, appropriate for diesel motor application. In addition to alkanes, some other reactions produce small amounts of alkenes, alcohols and some other oxygenated hydrocarbons. Generally the distribution of produced hydrocarbons in the Fischer–Tropsch synthesis obeys Anderson–Schulz–Flory probability distribution described by: W_n/n = 〖(1 - α)〗^2 α^(n-1) (1) Where Wn is the weight fraction of hydrocarbon with n carbon atoms and parameter α is chain growth probability. The value of α is essentially dictated by the catalyst nature as well as process conditions. The above distribution implies that methane is the most abundant product for α < 0.5; however, as α approaches 1, the total amount of methane formed lessens with respect to overall products. Increasing α enhances the formation of long-chained hydrocarbons. However, very long-chained hydrocarbons (waxes) are solid at room temperature, hence not desirable. Therefore, for production of liquid fuels it is necessary to crack some of the Fischer–Tropsch products. In order to avoid this (which surely increases cost significantly and is an energy consuming process), some researchers have proposed using molecular sieves like zeolites and CMS with fixed sized pores that can restrict the formation of hydrocarbons longer than some characteristic size (usually n < 10). This way we can minimize methane formation without producing lots of long-chained hydrocarbons. Although this offers a potential application for CMS materials it needs more study to improve available technologies. A recent paper on the applications of molecular sieves in Fischer-Tropsch process is presented by Sun and coworkers, 2012 [6].

Figure 7. Catalytic applications of CMS materials 4. research Paper Review The rest of our report is devoted to a discussion on the results of the following research paper: Maryam Peer, Ali Qajar, Ramakrishnan Rajagopalan, Henry C. Foley, “On the effects of emulsion polymerization of furfuryl alcohol on the formation of carbon spheres and other structures derived by pyrolysis of polyfurfuryl alcohol”, Carbon 51 (2013) 85 – 93. In this paper researchers have investigated fabrication of carbon spheres using emulsion polymerization as the precursor preparation method and subsequent pyrolysis of polymeric membrane. Furfuryl alcohol has been used as the monomer that was later polymerized to polyfurfuryl alcohol (PFA). During the emulsion polymerization process we need a surfactant and in this study they have used Pluronic F-127 as the amphiphilic agent. This also has the effect of structure-directing which led to fabrication of spherical polymeric membranes that were converted to CMS membranes after carbonization. These spherical membranes had an average size of 50 nm to a few microns depending upon the conditions of process. An interesting observation in their results is in the case of surface area per unit mass of membrane. As-synthesized carbon spheres have a surface area of about 480 m2/g with an average mean pore size of 0.5 nm. This high surface area (note that this value is before any post-treatment or activation) might be increased using an activation process (e.g. using CO2) to more than 1500 m2/g. In the above paper, a nice diagram of different compositional regions in the form of a pseudo-ternary phase diagram of surfactant/monomer/solvent is presented in order to determine the effects of changes in the emulsion polymerization variables on the kinds of carbon morphologies that might emerge from PFA after pyrolysis. It is observed that morphology and size of the carbon spheres is sensitive to monomer and surfactant concentrations, acid molarity and solvent composition. The effect of these different parameters is summarized in Table 2. In the following sections we shall discuss the details of these results along with underpinnings that justify the results. Table 2. Summary of the effect of different parameters on CMS fabrication process.

4.1. Experimental Aspects Emulsion polymerization is a kind of radical polymerization that is conducted in an emulsion consisting a solvent (most of the time water), monomer, and surfactant. Peer and coworkers, have used an amphiphilic triblock copolymer (EO106PO70EO106) Pluronic F-127 as surfactant and FA as monomer. Their typical synthesis scheme is shown in Figure 8.

Figure 8. Scheme used by Peer and coworkers to synthesize CMS spheres.

Figure 9. Pseudo-ternary phase diagram of solvent/surfactant/furfuryl alcohol system showing morphology evolution of carbon, A: spherical particles, B: interconnected structure, C: flaky structure, D: interconnected structure and E: carbon chunks, Black points are showing the samples synthesized [7]. 4.2. Results and discussion 4.2.1. Pseudo-terary phase digram Five different regions labeled A–E are determined on the phase diagram as shown in Figure 9. In region A we observe the following: Spherical particles Retained morphology even after pyrolysis Uniform and transparent micellar solution formed with the surfactant/acid/solvent mixtures Depending on the composition of surfactant, monomer and solvent, well separated carbon spheres with different diameters (from 50 nm to a few micro meters) are formed narrow pore size distribution (mean = 0.5 nm) (measured using methyl chloride gas adsorption) surface area = 480 m2/g before activation Investigators, conducted physical activation of produced CMS spheres under flowing CO2 with a flow rate of 900 ml/min and a temperature of 900˚C for 3 h. This process did not change the spherical shape of carbon particles. Activation led to 45% decrease in the initial mass which in turn brought about an increase in surface area to 1200 m2/g with a mean micropore size of 0.63 nm and narrow pore size distribution. They reported that the extent of activation might be controlled by changing the CO2 flow rate or the stream temperature. This also leads to control over pore size and total porosity (Figure 10). Carbon materials within regions B and D, show interconnected porous reticulated morphology. The mixture with composition in region B is monomer (FA) deficient (<10 wt. %) while that in region D is solvent deficient. Polymeric mixtures in both of these regions have been very viscous and this in turn significantly hinders the formation of homogeneous morphological features. In region C we have surfactant-rich phase and this results in a highly viscous gel. They observed that final carbon structure in this region is a dense, flaky and highly irregular structure. Finally, structures formed from mixtures with compositions in the FA rich region of E show dense, flaky morphology.

Figure 10. Characterization of carbon spheres synthesized by emulsion polymerization, as-synthesized and 45% activated, (a) methyl chloride pore size distribution, (b) cumulative pore volume, (c and d) TEM images of pyrolyzed carbon spheres [7]. 4.2.2. EFFECT OF MICELLAR SOLUTION COMPOSITION Monomer Concentration: The first parameter that has been studied by investigators is the monomer (FA) concentration. Figure 11 shows that there is a strong linear dependence between the size of the final carbon particles after pyrolysis and the FA concentration. The concentration of surfactant and the mass ratio of ethanol to total solvent (ethanol + water) were kept constant at 20 wt. % and 0.7, respectively for all the samples. The spheres have been mostly but not perfectly monodispersed. It seems that it shall be a suitable composition region for many applications considering observed monodispersity since handling similar particles is easier most of the time. The description of this behavior requires considering the kinetics of polymerization. The rate of polymerization for this process is given by R_p=(k_p n ̅N_p 〖[M]〗_p)/N_A (2) where kp is the polymerization rate constant, n ̅ is the number of cationic species per micelle, Np is the number of micelles and [M]p is the concentration of monomer in micelles and NA is Avogadro’s number. Equation (2) implies that the rate of polymerization is directly proportional to the monomer concentration in each micelle. The studies show that carbon spheres grow linearly with an increase in the monomer concentration (as shown in Fig. 10) as we would expect from the rate expression. Surfactant Concentration: The second studied variable is surfactant concentration. Unlike the monomer concentration, any changes in surfactant concentration have a significant effect on the morphology of carbon structure. At very low concentrations of pluronic F-127 (<2 wt. %), a large and highly distorted spherical morphology with huge dispersion (as shown by the error bars) is observed (Figure 12). On the other hand, when the surfactant concentration is very high (>35 wt. %), indistinct, agglomerated and highly irregular small spherical particles are observed. When the surfactant concentration ranges between 5 and 35 wt. %, we see regular spherical morphology and a monotonic decrease in size with increase in surfactant concentration (Figure 12).

Figure 11. Effect of monomer (furfuryl alcohol) amount on average carbon particles size, all samples synthesize with 20 wt. % pluronic F-127 [7].

Figure 12. Effect of surfactant (Pluronic F-127) on average carbon particles size, all samples made with 9 wt. % furfuryl alcohol [7]. The rationale for these results might be presented as follows; when surfactant concentration is between 5 and 35 wt. %. An increase in surfactant concentration leads to a reduction in interfacial surface tension and an increase in the number of micelles which in turn leads to an increase in the available surface area for polymerization. Finally this leads to a lot of small micelles that bring about the formation of small spheres with relatively narrow size distribution. At very low concentrations of surfactant the number of micelles formed is less and polymerization occurs both within the micelles and in the solvent-rich phase simultaneously, leading to the formation of irregular and larger particles. Finally at the other extreme (high surfactant concentration) a large number of very small micelles are formed that leads to an increased tendency to aggregate and to produce less orderly morphology.

Figure 13. Effect of initiator (HCl) molarity on average carbon particles size and morphology, Ethanol/ (Ethanol + water) mass ratio: 0.7, red: (10.5 wt. % FA, 10.5 wt. % pluronic F-127), blue: (40 wt. % pluronic F-127, 12 wt. % FA) [7].

Figure 14. (a) SEM and (b) TEM images of carbon sample synthesized at HCl molarity of 7, 40 wt.% pluronic F-127, 12 wt.% FA, and 48 wt.% solvent [7]. Initiator Concentration: The size of the carbon particles increases acid concentration, but only up to a certain limit, beyond which, the final particle size decreases significantly (Figure 13). The limiting acid concentration at which this reversal (maximum size) occurs is impressed upon by surfactant concentration, as it is shifted to lower molarities at lower surfactant concentrations. Also, it is obvious that polydispersity increases in the vicinity of maximum point and we observe a bimodal distribution in this region. At very high acid concentrations (>6 M), interconnected, possibly hollow, rod-like structures are formed (Figure 14). To describe this rather complex behavior we should first know the mechanism that governs the polymerization reaction. The underlying mechanism is shown in Figure 15. On the one hand, at higher acid concentrations we have higher rate of initiation and larger number of available carbocations in each micelle. Therefore we would experience higher rate of polymerization and larger carbon particles. However, on the other hand, as shown in the second step of polymerization mechanism, higher acid concentrations increases the rate of crosslinking hence increases the rate of termination. Therefore, beyond a certain critical acid concentration, due to simultaneous opposite effects of initiation-termination, a reversal and a significant decrease in the particle size is observed. This complex behavior obligates one to adopt suitable initiator concentration in accordance with surfactant concentration for specific design objectives. Solvent Composition: The last parameter is composition of ethanol/water mixture. In general, the average size of carbon particles increases as ethanol fraction grows in the solvent (Figure 16). When the ethanol fraction is less than 0.5, an irregular dense morphology is emerges. As the ethanol fraction increases, spherical carbon particles are formed. When the ethanol fraction is between 0.5 and 0.8, monodispersed carbon spheres with diameters in the range of 150 nm to 1 lm are formed. At very high ethanol fractions, a bimodal dispersion is what we see with particles ranging in diameter from 150 to 300 nm and from 1.2 to 1.4 μm (Figure 17).

Figure 15. Acid catalyzed polymerization scheme of furfuryl alcohol [7].

Figure 16. Effect of ethanol and water composition in the solvent on average particles size, samples synthesized with 10.5 wt.% furfuryl alcohol and 10.5 wt.% pluronic F-127 and 79 wt.% solvent [7].

Figure 17. Effect of ethanol and water composition on the size distribution of carbon particles, samples synthesized with 10.5 wt.% furfuryl alcohol and 10.5 wt.% pluronic F-127, (a and c): ethanol fraction of 0.7, (b and d): ethanol fraction of 0.9 [7]. Application of ethanol as cosolvent decreases the interfacial tension between the hydrophobic tail of surfactant and the solvent. This reduction in interfacial tension decreases the micelle core and core + corona radii. Also it lowers the micelle association number which is the number of surfactant species included in micelle. In fact addition of ethanol improves solubility of surfactant in solvent. It also lessens hard sphere interaction diameter. The sum of these effects leads to higher number density of micelles in micellar mixture. These micelles become smaller by increasing ethanol fraction. Therefore at high concentrations of ethanol (VEtOH/ V(H2O + EtOH) > 0.8) we will have rather tiny micelles with a high number density. So two distinct mechanisms of micellar growth take place, first nucleation and growth of individual micelles, and second the Ostwald ripening which is coagulation of smaller micelles leading to larger particles (note that the later mechanism is driven by small hard sphere interaction diameter). As a result, larger spheres are synthesized. At lower ethanol fractions the first mechanism dominates and determines the particle size. Finally at very low ethanol fractions (VEtOH/ V(H2O + EtOH) < 0.5), the critical micellar concentration (defined as a range of concentrations separating the limit below which virtually no micelles are detected and above that all additional surfactant molecules form micelles) is less leading to increased aggregation of micelles and an irregular carbon morphology. 4.3. critique In the paper reviewed the exact mechanism of termination and its precise relation to HCl concentration is not described well. It has been mentioned that increase of acid initiator would lead to more termination. However it is not clear how this effect takes place (we can make some predictions but yet not enough from a scientific point of view). Also the third step of mechanism in part (a) should be described more to elucidate the underpinnings of relation between acid concentration, particle size and morphology. Another thing not discussed is the pore size distribution. There is only a short indication at the end of the paper that says there is no appreciable effect on micropore structure by varying studied variables. However it seems that more investigation is needed for such a claim. Another problem is regarding Figure 10. In parts (a) it is not clear what variable does the vertical axis refers to. Finally, the last thing that needs more elaboration is the effect of ethanol fraction in the limit of very low EtOH fractions. For future studies it is advised to select wisely a number of other combinations of surfactants and initiators as well as other kinds of solvent mixtures and conduct the same experiments to find out a reasonable relation between characteristics of each agent (monomer, initiator, surfactant, and solvent) and design parameters of CMS materials. Refrences

[1] Foley, H.C., “Carbogenic Molecular Sieves: Synthesis, Properties and Applications”, Microporous Materials 4 (1995) 407-433. [2] Saufi, S.M., Ismail, A.F., “Fabrication of Carbon Membranes for Gas Separation––A Review”, Carbon 42 (2004) 241–259. [3] Foley, H.C., Kane, M.S., Goellner, J.F., “New Directions With Carbogenic Molecular Sieve Materials”, Access in Nanoporous Materials, Edited by T. J. Pinnavaia and M. F. Thorpe, Kluwer Academic Publishers, 2002. [4] Trimm, D.L., Cooper, B.J., “The Preparation of Selective Carbon Molecular Sieve Catalysts”, Journal of the Chemical Society, 477 (1970). [5] Holbrook, B.M., Rajagopalan, R., Dronvajjala, K., Choudhary, Y.K., Foley, H.C., “Molecular sieving carbon catalysts for liquid phase reactions: Study of alkene hydrogenation using platinum embedded nanoporous carbon”, Journal of Molecular Catalysis A: Chemical, xxx (2012) xxx– xxx. [6] Sun, B., Qiao, M., Fan, K., Ulrich, J., Tao, F., “Fischer–Tropsch Synthesis over Molecular Sieve Supported Catalysts”, ChemCatChem 2011, 3, 542 – 550. [7] Peer, M., Qajar, A., Rajagopalan, R., Foley, H.C., “On the effects of emulsion polymerization of furfuryl alcohol on the formation of carbon spheres and other structures derived by pyrolysis of polyfurfuryl alcohol”, Carbon 51 (2013) 85 – 93. [8] Alexandridis, P., Yang, L., "SANS Investigation of Polyether Block Copolymer Micelle Structure in Mixed Solvents of Water and Formamide, Ethanol, or Glycerol", Macromolecules 2000, 33, 5574-5587.

Caption1
Caption2