Ligand Engineering for Metal-Organic Frameworks with Hydrogen Storage Capabilities
L. Ciemnolonski
Drexel University, Dept. of Chemistry, Philadelphia, PA
December 5, 2009
Abstract
The dihydrogen (H2) physisorption of a metal-organic framework (MOF) can be optimized by tailoring the organic ligand used in the MOF synthesis. The functional groups, body and steric interactions are all important considerations when engineering the ligand. Other techniques such as catenation, ligand doping, the use of secondary functional groups, ligand distortion and impregnation are also strategies to enhance H2 physisorption.
Introduction
Bridging the worlds of organic and inorganic research, exists a fairly new class of materials known as metal-organic frameworks (MOF's). MOF's are microporous crystalline materials created by the coordination of metal ions or clusters, commonly called secondary building units (SBU's), with organic ligands or linkers. MOF's are also known as polymer coordination networks or hybrid frameworks. MOF's are of great interest, especially within the last 10 years, for a variety of reasons. Most importantly, MOF's have the potential to be functional in a wide variety of applications, such as gas storage and/or separation, catalysis, ion exchange and magnetism. However, as any chemist knows, function is determined by chemical structure. Thus, MOF's are even more so intriguing because of the degree of control that can be executed over the final structure, rather than the shake-and-bake technique. This element of control separates MOF's from other porous solids, such as activated carbons. At the core of MOF engineering is ligand engineering, in which you can create an organic ligand specifically tailored to meet your needs, depending on which application it will be used for.
As we all know, green energy and more specifically, hydrogen as a fuel source, has been of interest since the late 1970's due to economic and environmental concerns. However, one of the biggest challenges we face using hydrogen as a fuel source is the storing of hydrogen itself. MOF's have been shown to be a promising hydrogen storage system through physical adsoprtion, or physisorption, due to their immense surface area and other unique properties [1]. This paper will focus on the engineering of porous MOF's constructed from metal ions and one organic ligand for hydrogen storage applications. Specific emphasis will be placed on the ligand engineering.
Background
A great deal of research is being done on MOF's and their applications. A major player in this field is Dr. Omar M. Yaghi of UCLA. Yaghi and his colleagues created an entirely new branch of chemistry called "reticular chemistry" [2] in the 1990's. Reticular chemistry is the study of "molecular building blocks of synthetic and biological origin into predetermined structure[s] using strong bonds" [2]. Other pioneers in structure engineering using strong bonds are Hoskins and Robson [3], whose work dates back to 1990. Today, four MOF's are being commercially manufactured by BASF Chemical Company [4] and sold by Sigma Aldrich. These MOF's, named Basolite, have extremely high surface areas and exhibit adsorption and desorption properties due to their highly porous frameworks.
Coordination Chemistry
A metal-ligand complex, or coordination compound [5], is created by the interaction of a Lewis acid (metal atom or ion) with Lewis bases (ligands). The ligands donate electrons to form a bond with the Lewis acid. The number of ligands attached to the Lewis acid is called the coordination number. The coordination number [5] depends on the size of the central ion and the steric interactions between the ligands. In terms of simple 3-dimensional MOF's with cube-like pores, it is of interest to investigate complexes with square-planar or octahedral geometry, both of which are based on 6-coordination [5]. Six coordination is typical for metal ions with d0 to d9 electronic configurations, such as divalent cationic copper (d9). Beyond the 3d and 4d transition metals, as the atomic radii increases, higher coordination is possible. For the purposes of this discussion, we will focus on six-coordination.
To determine what ligands will best participate in the formation of a 6-coodination complex, one can study the spectrochemical series [5] created by R. Tsuchida. From Crystal-field theory [5] based on octahedral geometry, it was found that the ligand-field splitting parameter is greatly influenced by the ligand. A large slitting parameter denotes a large energy gap between the eg and t2g orbitals and thus a stable system. It was found that CO and nitrogen-containing ligands, like bipyridine, have large splitting parameters and so they are considered strong-field ligands. In reality, most ligands used in MOF synthesis do in fact usually contain carbonyl and nitrogen functional groups.
Dihydrogen Chemistry
Dihydrogen (H2) has a high specific enthalpy of 142 kJ/g, which makes it an obvious candidate as a fuel source [5]. H2 can act as both a weak Lewis base or a weak Lewis acid [6]. As a weak Lewis base, H2 can bind to electron-rich metal atoms or ions. Through the phenomenon known as backdonation [6], transition metals can use a filled d-orbital to form a sigma-bond with an empty antibonding orbital of H2. Backdonation and thus strong interaction with H2 is not seen with M+1 alkali metal ions due to their closed-shell electronic configuration [7]. The H2 affinity to 3d M2+ metal ions has also been shown to harmonize with the Irving-Williams series [8].
Backdonation (6)
H2 has also been shown to have an affinity to carbon-based surfaces, such as carbon nanotubes and aromatic rings, through Van der Waals interactions with the sp2 carbons [9]. Another means for H2 physisorption to a carbon surface has been said to be due to the hydrogen "spillover" effect [10]. The spillover effect involves the use of a catalyst such as platinum and a bridge to help move H2 from the catalyst to the surface. While this technology is currently being explored for use in MOF's, it is not desirable due to the use of a catalyst.
Spillover of dihydrogen from catalyst to the carbon-based surface (10)
Specifically concerning MOF's, the Clausius-Clapeyron equation, shown below, is often used as a measure of H2 affinity to the MOF [11].
P=pressure, n= amount of gas adsorbed, T=temperature, R=universal gas constant, C=constant (11)
MOF Synthesis
Among the many MOF synthesis methods, such as sol-gel and diffusion, hydrothermal or solvothermal synthesis seems to be the most prevalent. In these syntheses methods [12], the metal salt, organic ligand(s), solvent and any other necessary ingredients, are all placed into an teflon-lined autoclave "bomb" under autogenous pressure and heated to at least 100°C for a specified time frame, usually a few hours to days. The synthesis conditions allow diffusion and crystallization to occur. MOF synthesis is based upon self-assembly of the starting materials to create crystalline solid structures [12]. High quality single-crystals are desired so that they may be analyzed by X-ray crystallography.
Self-assembly of metal ions and ligands to create 1D, 2D or 3D MOF's (12)
Requirements for Hydrogen Storage Systems
The United Stated Department of Energy [13] requires MOF's to hold about 6% H2 by weight at ambient temperatures by 2010 and 9% by 2015 to be considered as a potential storage system. Because of this weight percent requirement, the use of lightweight ligands and/or metals would be ideal to lower the overall effective density of the resulting MOF. Aside from the H2 uptake targets, there are many other requirements [6] for a chosen hydrogen storage system. First of all, the raw materials needed to generate the storage system must be cheap and in high supply. The materials used to create MOF's are indeed cheap and readily available. Furthermore, the physisorption of H2 must be reversible, as to easily allow use of the stored hydrogen. Also, the physisorption of H2 cannot require massive amounts of heat or pressure. On the same note, the desorption of H2 from the system cannot generate massive amounts of heat or pressure. Ideally, H2 could be physisorbed quickly without a catalyst, to make for quick and easy re-fueling. Because of this, our discussion will not cover H2 physisorption by the spillover effect, which does require a catalyst.
MOF Engineering
Now that we understand the chemistry, necessary requirements and synthesis procedure, we can begin to design our MOF. While most argue that open metal coordination sites are the main route to H2 physisorption in a MOF, it is important to remember that MOF's are not made entirely of metals. Thus, the ligand needs to be thoroughly explored a major factor in H2 physiorption as well.
Ligand Functional Groups
Ligands usually contain O-, N- and S- donor atoms at their "appendages" so that they can strongly coordinate with the metal ion [12]. Functional groups commonly seen in MOF ligands are carboxylates, carbonyls, amines, amides, thiols and cyano groups. The ligands usually contain two or more functional groups and are termed bi- and multi-dentate ligands [12]. Ligands containing only one functional group are called terminal ligands, because they discourage high dimensionality of the resulting MOF.
Ligand Body
The space between the functional groups, or the "body" [12] of the ligand, is also very important. The body creates the pore walls and is a key factor in MOF stability. Therefore, it is of no surprise that aromatic rings, such as benzene rings, are commonly incorporated into the ligand body. The sp2 carbon atoms of benzene create a planar and rigid ring system, compared to the sp3 carbon atoms of an alkane, which have free bond rotation. Take for example 1,4-benzenedicarboxylic acid and malonic acid, shown below. Both ligands contain the same two functional groups, but they differ in their body.
Montney et al. [14] have actually used malonic acid in conjunction with another ligand to create new MOF's. However, even they admit that the rigid nature of 1,4-benzenedicarboxylic acid is more beneficial for porous MOF's and that the flexible nature of malonic acid can lead to many possible MOF structures. Another danger of using non-rigid ligands is the possibly of decomposition of the final structure. Thus, for our purposes of designing a particular porous MOF, rigid ligands are preferred over aliphatic ligands.
Benzene rings are not the only ring systems used in ligands. Polycyclic aromatic hydrocarbons (PAH's) such as triphenylene and pyrene are also used. Again, the rigid properties of the PAH's are key for the MOF's structural and stability properties, however they also prove very useful in the physisorption properties of the MOF as well. For example, Roswell et al. [15] used pyrene in the body of their ligand. The resulting MOF showed increased H2 physisorption which they attributed to the favorable pyrene-H2 interaction. Furthermore, Scanlon et al. [16] evaluated many PAH's for their affect on H2 physisorption, some of which were planar, curved, neutral and charged. It was found that anionic charged systems were doubly effective due to the electrostatic "charge-induced dipole interactions" with H2, in addition to the usual Van der Waals interactions. The PAH surface geometry was also found to be a key factor in the H2 physisorption for neutral PAH systems as well. For instance, Okamoto [16] reported that graphene can adsorb three times as much H2 if the surface is not planar. Scanlon et al. [16] continue to say that concave surfaces have twice the calculated H2 binding energy than convex surfaces. Another study by Li et al. [11] is also in agreement that nonlinear, or curved, MOF pores help increase the interaction with and physisorption of H2.
The body of the ligand can also be distorted to affect the pore shape and overall dimensionality of the MOF. Furukawa et al. [17] did such experiments starting with the 1,4-benzenedicarboxylic acid ligand linking together two SBU's. When non-distorted, the ligand is linear and the two carboxylate groups lie in the same plane as the benzene ring ("A" in the figure below). However, bending the ligand out-of-plane ("C" in the figure below) creates a polyhedral molecule with a spherical-like pore. The curved nature of this pore may help enhance H2 physisorption.
Distorted ligands and the affect on pore shape and MOF dimensionality (17)
These kind of polyhedral molecules have divided into a branch of their own, called metal-organic polyhedra (MOP's), and are also of great interest now [17].
Although they are not as common as ring systems, double bonds are also incorporated into ligand bodies, although usually in conjunction with ring systems. One study reported by Wang et al. [18] shows that a nitrogen-nitrogen double bond in the body of the ligand is more favorable than a carbon-carbon double bond for H2 physisorption.
The body of the ligand also has an obvious affect on the pore size. Clearly, the use of 4,4'-biphenyldicarboxylic acid versus 1,4-benzenedicarboxylic acid, shown below, will increase the pore dimensions due to the presence of another benzene ring. Thus, pore dimension seem fairly easy to control.
It has been reported by several sources that the size of the pores is critical for H2 physisorption. Surprisingly, Latroche et al. [19] state that big pores are actually unfavorable if H2 is to be physisorbed as a monolayer onto the pore walls. Rather, there should be many small pores. Small pores will allow H2 to interact with two or more of the pore walls and will better help the physisorption of the H2. Thus, pore dimensions should be slightly larger than that of H2, which has a bond length of 0.74 angstroms and a Van der Waals diameter of 2.4 angstroms [11]. Chun et al. [20] did such experiments and created a MOF with pore channels 4-6 angstroms wide, which is unusual in MOF research. As a result of the uniquely small channel dimensions, the MOF showed good H2 physisorption onto the pore walls, 2.45 weight percent at 77K and 0.98 bars pressure. Furthermore, Chun et al. [20] state that this MOF is not even saturated at this point and that further pore engineering could lead to even higher H2 uptake. Still, some [15] argue that large pore volume is necessary for H2 physisorption by pore filling, versus pore wall coverage.
Catenation
Along the same lines of creating small pores to ensure sufficient interaction between the pore walls and H2, catenation has shown to be an efficient technique to encourage H2 physisorption. Catenation [15] is basically the interpenetration or interweaving of two or more MOF's. As to allow catenation to occur in-situ, the use of longer ligands is found to be the most useful. While long ligands would usually results in large pores, catenation reduces pore volume dramatically, as shown below.
Roswell et al. [15] created a catenated MOF (IRMOF-13) with pore volume reduced about 50% compared to the non-catenated MOF's. Nonetheless, IRMOF-13 was measured to have about a 10% increase in H2 adsoprtion versus a comparable, non-catenated MOF.
Unsaturated Metal Sites
Because open metal sites do have a have strong affinity for H2 and allow for more efficient packing of H2 inside the pores [7], it is necessary to explore how ligands can be used to encourage the presence of open metal coordination sites. The most common means of creating open metal sites is to use small terminal ligands, such as solvents like water or DMF, which can be removed from the structure with heat [7]. Of course, the danger in this technique is that your MOF structure could decompose after the removal of those molecules. Such tests can be done using thermal gravimetric analysis (TGA).
Another possible technique to create open metal sites is to use bulky ligands. As discussed previously, steric interactions between the ligands determine the coordination number about the metal ion. Thus, if the ligands are bulky and will not allow 6-coordination, an open metal site is likely to be created. However, this may also decrease the overall MOF dimensionality.
Ligand Doping & Modification
The use of metal ions within the ligands themselves is a fairly new and very hot topic of research right now in the MOF world. One way to introduce metal ions into the ligands is chelation. Ligands containing chelated metals have been reported by Dinca et al. [7]. Examples of such ligands they described are shown below.
Metal-containing ligands. Left to right: A hypothetical ligand, the salen-Mn3+ complex, and a half-sandwich unit (7)
Metal ions incorporated into the ligand can also be modified post-synthesis, such as the removal of the CO ligands from the half-sandwich unit [7]. Right now, this proves to be challenging. However, once unsaturated, the metal ions should be a very useful way to increase H2 physisorption. Another new way to increase H2 physisorption is called impregnation [7]. This involves the use of multi-anionic ligands to increase the presence of metal ions through normal coordination. The large volume of metal ions can then be used for ion-exchange purposes to balance the framework and bring in even more metals to the framework.
Side groups, or secondary functional groups, on a ligand can affect the electronic character of the ligand and the overall MOF structure. For instance, Eddaoudi et al. [21] tested several variations of 1,4-benzenedicarboxylic acid ligand, shown below:
Variations of 1,4-Benzenedicarboxylic acid (R1-BDC) ligand (21)
Interestingly, in the final MOF, all the side groups were found to be directed into the pores, thus reducing the overall free volume in the pore, compared to the MOF made with R1-BDC. This could serve as a novel way to increase H2 physisorption, if a suitable side group was used. In fact, the MOF made using R6-BDC was found to have good physisorption capabilities for methane gas due to the hydrophobic character of it's side group [21]. In another study, Roswell et al. [15] report that the use of an electron-donating side group like methyl onto a benzene ring could increase the H2 affinity to the ligand by about 15%. On the other hand, they report that electron-withdrawing side groups could decrease the H2 affinity to the ligand.
The use of secondary functional groups on a ligand can also serve as a way to insert a metal ion onto the ligand. Such research was done by Zhou et al. [22]. They found that simple side groups such as amines can be converted to open vanadium cation sites, which can be further converted into other metal ions. Conversion to transition metal ions, such as cobalt, would be even more ideal since Zhou and his colleagues also found that open cobalt ions are more useful for H2 physisorption, compared to iron and manganese [22].
Conversion of a secondary functional group to an open metal site (22)
Conclusion
The organic ligands used in MOF's have a significant impact on the structure's H2 physisorption ability. First and foremost, a stable structure is necessary and can be controlled by the presence of strongly-coordinating functional groups at the ligand's "appendages" and ring systems in the ligand body. Preferentially, the ring systems should be curved, not linear. Curved pores can also be achieved by distortion of the ligand's backbone. The length of the ligand is critical. Either short ligands should be used or catenation should be employed to create many small pores rather than a few large pores. Small pores greatly enhance H2-pore interaction and thus physisorption. Ligands can also be tailored to create open metal sites with high H2 affinity. Such techniques include the use of volatile solvent molecules as terminal ligands, bulky ligands, ligands containing chelated metals, strategic use of secondary functional groups in the body of the ligand, and impregnation.
References
(1) Saha, D.; Wei, Z.; Deng, S. Int J Hydrogen Energy2008, 33, 7479-7488. DOI
(2) Center for Reticular Chemistry. Link (accessed 10/2009).
(3) Paz, F. A. A.; Rocha, J.; Klinowski, J.; Trindade, T.; Shi, F.; Mafra, L. Progress in Solid State Chemistry2005, 33, 113-125.DOI
(4) BASF The Chemical Company. Basolite Metal-Organic Frameworks. Link (accessed 10/2009).
(5) Shriver, D. F.; Atkins, P. W. Inorganic Chemistry 3rd ed. W.H. Freeman and Company: New York, NY, 2003, 211-228, 256.
(6) Kubas, G. J. Chem. Rev. 2007, 107, 4152-4205. DOI
(7) Dincă; M.; Long, J. R. Angewandte Chemie International Edition2008, 47, 6766-6779. DOI
(8) Ma, S.; Yuan, D.; Chang, J.; Zhou, H. Inorg. Chem.2009, 48, 5398-5402. DOI
(9) Zhou, L.; Zhou, Y.; Sun, Y. Int J Hydrogen Energy2006, 31, 259-264. DOI
(10) Chen, L.; Cooper, A. C.; Pez, G. P.; Cheng, H. The Journal of Physical Chemistry2007, 111, 18995-19000. DOI
(11) Li, Y.; Xie, L.; Liu, Y.; Yang, R.; Li, X. Inorg. Chem.2008, 47, 10372-10377. DOI
(12) Robin, A. Y.; Fromm, K. M. Coord. Chem. Rev.2006, 250, 2127-2157. DOI
(13) U. S. Department of Energy. Energy Efficiency & Renewable Energy. Fuel Cell Technologies Program. Hydrogen Storage. Link to siteLink to pdf (accessed 11/2009)
(14) Montney, M. R.; Supkowski, R. M.; LaDuca, R. L. Polyhedron2008, 27, 2997-3003. DOI
(15) Rowsell, J. L. C.; Yaghi, O. M. J. Am. Chem. Soc.2006, 128, 1304-1315. DOI
(16) Scanlon, L. G.; Feld, W. A.; Balbuena, P. B.; Sandi, G.; Duan, X.; Underwood, K. A.; Hunter, N.; Mack, J.; Rottmayer, M. A.; Tsao, M. The Journal of Physical Chemistry B2009, 113, 4708-4717. DOI
(17) Furukawa, H.; Kim, J.; Ockwig, N. W.; O'Keeffe, M.; Yaghi, O. M. J. Am. Chem. Soc.2008, 130, 11650-11661. DOI
(18) Wang, X.; Ma, S.; Rauch, K.; Simmons, J. M.; Yuan, D.; Wang, X.; Yildirim, T.; Cole, W. C.; López, J. J.; Meijere, A. d.; Zhou, H. Chemistry of Materials2008, 20, 3145-3152. DOI
(19) Latroche, M.; Surblé, S.; Serre, C.; Mellot-Draznieks, C.; Llewellyn, P. L.; Lee, J.; Chang, J.; Jhung, S. H.; Férey, G. Angewandte Chemie International Edition2006, 45, 8227-8231. DOI
(20) Chun, H.; Jung, H.; Koo, G.; Jeong, H.; Kim, D. Inorg. Chem.2008, 47, 5355-5359. DOI
(21) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; O'Keeffe, M.; Yaghi, O. M. Science2002, 295, 469-472. DOI
(22) Zhou, H. A Biomimetic Approach to Metal-Organic Frameworks with High H2 Uptake. 2009Link (accessed 11/2009).
Chemical Information Retrival (Chem 767)
lciemnolonski log
Ligand Engineering for Metal-Organic Frameworks with Hydrogen Storage Capabilities
L. Ciemnolonski
Drexel University, Dept. of Chemistry, Philadelphia, PA
December 5, 2009
Abstract
The dihydrogen (H2) physisorption of a metal-organic framework (MOF) can be optimized by tailoring the organic ligand used in the MOF synthesis. The functional groups, body and steric interactions are all important considerations when engineering the ligand. Other techniques such as catenation, ligand doping, the use of secondary functional groups, ligand distortion and impregnation are also strategies to enhance H2 physisorption.
Introduction
Bridging the worlds of organic and inorganic research, exists a fairly new class of materials known as metal-organic frameworks (MOF's). MOF's are microporous crystalline materials created by the coordination of metal ions or clusters, commonly called secondary building units (SBU's), with organic ligands or linkers. MOF's are also known as polymer coordination networks or hybrid frameworks. MOF's are of great interest, especially within the last 10 years, for a variety of reasons. Most importantly, MOF's have the potential to be functional in a wide variety of applications, such as gas storage and/or separation, catalysis, ion exchange and magnetism. However, as any chemist knows, function is determined by chemical structure. Thus, MOF's are even more so intriguing because of the degree of control that can be executed over the final structure, rather than the shake-and-bake technique. This element of control separates MOF's from other porous solids, such as activated carbons. At the core of MOF engineering is ligand engineering, in which you can create an organic ligand specifically tailored to meet your needs, depending on which application it will be used for.
As we all know, green energy and more specifically, hydrogen as a fuel source, has been of interest since the late 1970's due to economic and environmental concerns. However, one of the biggest challenges we face using hydrogen as a fuel source is the storing of hydrogen itself. MOF's have been shown to be a promising hydrogen storage system through physical adsoprtion, or physisorption, due to their immense surface area and other unique properties [1]. This paper will focus on the engineering of porous MOF's constructed from metal ions and one organic ligand for hydrogen storage applications. Specific emphasis will be placed on the ligand engineering.
Background
A great deal of research is being done on MOF's and their applications. A major player in this field is Dr. Omar M. Yaghi of UCLA. Yaghi and his colleagues created an entirely new branch of chemistry called "reticular chemistry" [2] in the 1990's. Reticular chemistry is the study of "molecular building blocks of synthetic and biological origin into predetermined structure[s] using strong bonds" [2]. Other pioneers in structure engineering using strong bonds are Hoskins and Robson [3], whose work dates back to 1990. Today, four MOF's are being commercially manufactured by BASF Chemical Company [4] and sold by Sigma Aldrich. These MOF's, named Basolite, have extremely high surface areas and exhibit adsorption and desorption properties due to their highly porous frameworks.
Coordination Chemistry
A metal-ligand complex, or coordination compound [5], is created by the interaction of a Lewis acid (metal atom or ion) with Lewis bases (ligands). The ligands donate electrons to form a bond with the Lewis acid. The number of ligands attached to the Lewis acid is called the coordination number. The coordination number [5] depends on the size of the central ion and the steric interactions between the ligands. In terms of simple 3-dimensional MOF's with cube-like pores, it is of interest to investigate complexes with square-planar or octahedral geometry, both of which are based on 6-coordination [5]. Six coordination is typical for metal ions with d0 to d9 electronic configurations, such as divalent cationic copper (d9). Beyond the 3d and 4d transition metals, as the atomic radii increases, higher coordination is possible. For the purposes of this discussion, we will focus on six-coordination.
To determine what ligands will best participate in the formation of a 6-coodination complex, one can study the spectrochemical series [5] created by R. Tsuchida. From Crystal-field theory [5] based on octahedral geometry, it was found that the ligand-field splitting parameter is greatly influenced by the ligand. A large slitting parameter denotes a large energy gap between the eg and t2g orbitals and thus a stable system. It was found that CO and nitrogen-containing ligands, like bipyridine, have large splitting parameters and so they are considered strong-field ligands. In reality, most ligands used in MOF synthesis do in fact usually contain carbonyl and nitrogen functional groups.
Dihydrogen Chemistry
Dihydrogen (H2) has a high specific enthalpy of 142 kJ/g, which makes it an obvious candidate as a fuel source [5]. H2 can act as both a weak Lewis base or a weak Lewis acid [6]. As a weak Lewis base, H2 can bind to electron-rich metal atoms or ions. Through the phenomenon known as backdonation [6], transition metals can use a filled d-orbital to form a sigma-bond with an empty antibonding orbital of H2. Backdonation and thus strong interaction with H2 is not seen with M+1 alkali metal ions due to their closed-shell electronic configuration [7]. The H2 affinity to 3d M2+ metal ions has also been shown to harmonize with the Irving-Williams series [8].
H2 has also been shown to have an affinity to carbon-based surfaces, such as carbon nanotubes and aromatic rings, through Van der Waals interactions with the sp2 carbons [9]. Another means for H2 physisorption to a carbon surface has been said to be due to the hydrogen "spillover" effect [10]. The spillover effect involves the use of a catalyst such as platinum and a bridge to help move H2 from the catalyst to the surface. While this technology is currently being explored for use in MOF's, it is not desirable due to the use of a catalyst.
Specifically concerning MOF's, the Clausius-Clapeyron equation, shown below, is often used as a measure of H2 affinity to the MOF [11].
MOF Synthesis
Among the many MOF synthesis methods, such as sol-gel and diffusion, hydrothermal or solvothermal synthesis seems to be the most prevalent. In these syntheses methods [12], the metal salt, organic ligand(s), solvent and any other necessary ingredients, are all placed into an teflon-lined autoclave "bomb" under autogenous pressure and heated to at least 100°C for a specified time frame, usually a few hours to days. The synthesis conditions allow diffusion and crystallization to occur. MOF synthesis is based upon self-assembly of the starting materials to create crystalline solid structures [12]. High quality single-crystals are desired so that they may be analyzed by X-ray crystallography.
Requirements for Hydrogen Storage Systems
The United Stated Department of Energy [13] requires MOF's to hold about 6% H2 by weight at ambient temperatures by 2010 and 9% by 2015 to be considered as a potential storage system. Because of this weight percent requirement, the use of lightweight ligands and/or metals would be ideal to lower the overall effective density of the resulting MOF. Aside from the H2 uptake targets, there are many other requirements [6] for a chosen hydrogen storage system. First of all, the raw materials needed to generate the storage system must be cheap and in high supply. The materials used to create MOF's are indeed cheap and readily available. Furthermore, the physisorption of H2 must be reversible, as to easily allow use of the stored hydrogen. Also, the physisorption of H2 cannot require massive amounts of heat or pressure. On the same note, the desorption of H2 from the system cannot generate massive amounts of heat or pressure. Ideally, H2 could be physisorbed quickly without a catalyst, to make for quick and easy re-fueling. Because of this, our discussion will not cover H2 physisorption by the spillover effect, which does require a catalyst.
MOF Engineering
Now that we understand the chemistry, necessary requirements and synthesis procedure, we can begin to design our MOF. While most argue that open metal coordination sites are the main route to H2 physisorption in a MOF, it is important to remember that MOF's are not made entirely of metals. Thus, the ligand needs to be thoroughly explored a major factor in H2 physiorption as well.
Ligand Functional Groups
Ligands usually contain O-, N- and S- donor atoms at their "appendages" so that they can strongly coordinate with the metal ion [12]. Functional groups commonly seen in MOF ligands are carboxylates, carbonyls, amines, amides, thiols and cyano groups. The ligands usually contain two or more functional groups and are termed bi- and multi-dentate ligands [12]. Ligands containing only one functional group are called terminal ligands, because they discourage high dimensionality of the resulting MOF.
Ligand Body
The space between the functional groups, or the "body" [12] of the ligand, is also very important. The body creates the pore walls and is a key factor in MOF stability. Therefore, it is of no surprise that aromatic rings, such as benzene rings, are commonly incorporated into the ligand body. The sp2 carbon atoms of benzene create a planar and rigid ring system, compared to the sp3 carbon atoms of an alkane, which have free bond rotation. Take for example 1,4-benzenedicarboxylic acid and malonic acid, shown below. Both ligands contain the same two functional groups, but they differ in their body.
Montney et al. [14] have actually used malonic acid in conjunction with another ligand to create new MOF's. However, even they admit that the rigid nature of 1,4-benzenedicarboxylic acid is more beneficial for porous MOF's and that the flexible nature of malonic acid can lead to many possible MOF structures. Another danger of using non-rigid ligands is the possibly of decomposition of the final structure. Thus, for our purposes of designing a particular porous MOF, rigid ligands are preferred over aliphatic ligands.
Benzene rings are not the only ring systems used in ligands. Polycyclic aromatic hydrocarbons (PAH's) such as triphenylene and pyrene are also used. Again, the rigid properties of the PAH's are key for the MOF's structural and stability properties, however they also prove very useful in the physisorption properties of the MOF as well. For example, Roswell et al. [15] used pyrene in the body of their ligand. The resulting MOF showed increased H2 physisorption which they attributed to the favorable pyrene-H2 interaction. Furthermore, Scanlon et al. [16] evaluated many PAH's for their affect on H2 physisorption, some of which were planar, curved, neutral and charged. It was found that anionic charged systems were doubly effective due to the electrostatic "charge-induced dipole interactions" with H2, in addition to the usual Van der Waals interactions. The PAH surface geometry was also found to be a key factor in the H2 physisorption for neutral PAH systems as well. For instance, Okamoto [16] reported that graphene can adsorb three times as much H2 if the surface is not planar. Scanlon et al. [16] continue to say that concave surfaces have twice the calculated H2 binding energy than convex surfaces. Another study by Li et al. [11] is also in agreement that nonlinear, or curved, MOF pores help increase the interaction with and physisorption of H2.
The body of the ligand can also be distorted to affect the pore shape and overall dimensionality of the MOF. Furukawa et al. [17] did such experiments starting with the 1,4-benzenedicarboxylic acid ligand linking together two SBU's. When non-distorted, the ligand is linear and the two carboxylate groups lie in the same plane as the benzene ring ("A" in the figure below). However, bending the ligand out-of-plane ("C" in the figure below) creates a polyhedral molecule with a spherical-like pore. The curved nature of this pore may help enhance H2 physisorption.
These kind of polyhedral molecules have divided into a branch of their own, called metal-organic polyhedra (MOP's), and are also of great interest now [17].
Although they are not as common as ring systems, double bonds are also incorporated into ligand bodies, although usually in conjunction with ring systems. One study reported by Wang et al. [18] shows that a nitrogen-nitrogen double bond in the body of the ligand is more favorable than a carbon-carbon double bond for H2 physisorption.
The body of the ligand also has an obvious affect on the pore size. Clearly, the use of 4,4'-biphenyldicarboxylic acid versus 1,4-benzenedicarboxylic acid, shown below, will increase the pore dimensions due to the presence of another benzene ring. Thus, pore dimension seem fairly easy to control.
It has been reported by several sources that the size of the pores is critical for H2 physisorption. Surprisingly, Latroche et al. [19] state that big pores are actually unfavorable if H2 is to be physisorbed as a monolayer onto the pore walls. Rather, there should be many small pores. Small pores will allow H2 to interact with two or more of the pore walls and will better help the physisorption of the H2. Thus, pore dimensions should be slightly larger than that of H2, which has a bond length of 0.74 angstroms and a Van der Waals diameter of 2.4 angstroms [11]. Chun et al. [20] did such experiments and created a MOF with pore channels 4-6 angstroms wide, which is unusual in MOF research. As a result of the uniquely small channel dimensions, the MOF showed good H2 physisorption onto the pore walls, 2.45 weight percent at 77K and 0.98 bars pressure. Furthermore, Chun et al. [20] state that this MOF is not even saturated at this point and that further pore engineering could lead to even higher H2 uptake. Still, some [15] argue that large pore volume is necessary for H2 physisorption by pore filling, versus pore wall coverage.
Catenation
Along the same lines of creating small pores to ensure sufficient interaction between the pore walls and H2, catenation has shown to be an efficient technique to encourage H2 physisorption. Catenation [15] is basically the interpenetration or interweaving of two or more MOF's. As to allow catenation to occur in-situ, the use of longer ligands is found to be the most useful. While long ligands would usually results in large pores, catenation reduces pore volume dramatically, as shown below.
Roswell et al. [15] created a catenated MOF (IRMOF-13) with pore volume reduced about 50% compared to the non-catenated MOF's. Nonetheless, IRMOF-13 was measured to have about a 10% increase in H2 adsoprtion versus a comparable, non-catenated MOF.
Unsaturated Metal Sites
Because open metal sites do have a have strong affinity for H2 and allow for more efficient packing of H2 inside the pores [7], it is necessary to explore how ligands can be used to encourage the presence of open metal coordination sites. The most common means of creating open metal sites is to use small terminal ligands, such as solvents like water or DMF, which can be removed from the structure with heat [7]. Of course, the danger in this technique is that your MOF structure could decompose after the removal of those molecules. Such tests can be done using thermal gravimetric analysis (TGA).
Another possible technique to create open metal sites is to use bulky ligands. As discussed previously, steric interactions between the ligands determine the coordination number about the metal ion. Thus, if the ligands are bulky and will not allow 6-coordination, an open metal site is likely to be created. However, this may also decrease the overall MOF dimensionality.
Ligand Doping & Modification
The use of metal ions within the ligands themselves is a fairly new and very hot topic of research right now in the MOF world. One way to introduce metal ions into the ligands is chelation. Ligands containing chelated metals have been reported by Dinca et al. [7]. Examples of such ligands they described are shown below.
Metal ions incorporated into the ligand can also be modified post-synthesis, such as the removal of the CO ligands from the half-sandwich unit [7]. Right now, this proves to be challenging. However, once unsaturated, the metal ions should be a very useful way to increase H2 physisorption. Another new way to increase H2 physisorption is called impregnation [7]. This involves the use of multi-anionic ligands to increase the presence of metal ions through normal coordination. The large volume of metal ions can then be used for ion-exchange purposes to balance the framework and bring in even more metals to the framework.
Side groups, or secondary functional groups, on a ligand can affect the electronic character of the ligand and the overall MOF structure. For instance, Eddaoudi et al. [21] tested several variations of 1,4-benzenedicarboxylic acid ligand, shown below:
Interestingly, in the final MOF, all the side groups were found to be directed into the pores, thus reducing the overall free volume in the pore, compared to the MOF made with R1-BDC. This could serve as a novel way to increase H2 physisorption, if a suitable side group was used. In fact, the MOF made using R6-BDC was found to have good physisorption capabilities for methane gas due to the hydrophobic character of it's side group [21]. In another study, Roswell et al. [15] report that the use of an electron-donating side group like methyl onto a benzene ring could increase the H2 affinity to the ligand by about 15%. On the other hand, they report that electron-withdrawing side groups could decrease the H2 affinity to the ligand.
The use of secondary functional groups on a ligand can also serve as a way to insert a metal ion onto the ligand. Such research was done by Zhou et al. [22]. They found that simple side groups such as amines can be converted to open vanadium cation sites, which can be further converted into other metal ions. Conversion to transition metal ions, such as cobalt, would be even more ideal since Zhou and his colleagues also found that open cobalt ions are more useful for H2 physisorption, compared to iron and manganese [22].
Conclusion
The organic ligands used in MOF's have a significant impact on the structure's H2 physisorption ability. First and foremost, a stable structure is necessary and can be controlled by the presence of strongly-coordinating functional groups at the ligand's "appendages" and ring systems in the ligand body. Preferentially, the ring systems should be curved, not linear. Curved pores can also be achieved by distortion of the ligand's backbone. The length of the ligand is critical. Either short ligands should be used or catenation should be employed to create many small pores rather than a few large pores. Small pores greatly enhance H2-pore interaction and thus physisorption. Ligands can also be tailored to create open metal sites with high H2 affinity. Such techniques include the use of volatile solvent molecules as terminal ligands, bulky ligands, ligands containing chelated metals, strategic use of secondary functional groups in the body of the ligand, and impregnation.
References
(1) Saha, D.; Wei, Z.; Deng, S. Int J Hydrogen Energy 2008, 33, 7479-7488. DOI
(2) Center for Reticular Chemistry. Link (accessed 10/2009).
(3) Paz, F. A. A.; Rocha, J.; Klinowski, J.; Trindade, T.; Shi, F.; Mafra, L. Progress in Solid State Chemistry 2005, 33, 113-125.DOI
(4) BASF The Chemical Company. Basolite Metal-Organic Frameworks. Link (accessed 10/2009).
(5) Shriver, D. F.; Atkins, P. W. Inorganic Chemistry 3rd ed. W.H. Freeman and Company: New York, NY, 2003, 211-228, 256.
(6) Kubas, G. J. Chem. Rev. 2007, 107, 4152-4205. DOI
(7) Dincă; M.; Long, J. R. Angewandte Chemie International Edition 2008, 47, 6766-6779. DOI
(8) Ma, S.; Yuan, D.; Chang, J.; Zhou, H. Inorg. Chem. 2009, 48, 5398-5402. DOI
(9) Zhou, L.; Zhou, Y.; Sun, Y. Int J Hydrogen Energy 2006, 31, 259-264. DOI
(10) Chen, L.; Cooper, A. C.; Pez, G. P.; Cheng, H. The Journal of Physical Chemistry 2007, 111, 18995-19000. DOI
(11) Li, Y.; Xie, L.; Liu, Y.; Yang, R.; Li, X. Inorg. Chem. 2008, 47, 10372-10377. DOI
(12) Robin, A. Y.; Fromm, K. M. Coord. Chem. Rev. 2006, 250, 2127-2157. DOI
(13) U. S. Department of Energy. Energy Efficiency & Renewable Energy. Fuel Cell Technologies Program. Hydrogen Storage. Link to site Link to pdf (accessed 11/2009)
(14) Montney, M. R.; Supkowski, R. M.; LaDuca, R. L. Polyhedron 2008, 27, 2997-3003. DOI
(15) Rowsell, J. L. C.; Yaghi, O. M. J. Am. Chem. Soc. 2006, 128, 1304-1315. DOI
(16) Scanlon, L. G.; Feld, W. A.; Balbuena, P. B.; Sandi, G.; Duan, X.; Underwood, K. A.; Hunter, N.; Mack, J.; Rottmayer, M. A.; Tsao, M. The Journal of Physical Chemistry B 2009, 113, 4708-4717. DOI
(17) Furukawa, H.; Kim, J.; Ockwig, N. W.; O'Keeffe, M.; Yaghi, O. M. J. Am. Chem. Soc. 2008, 130, 11650-11661. DOI
(18) Wang, X.; Ma, S.; Rauch, K.; Simmons, J. M.; Yuan, D.; Wang, X.; Yildirim, T.; Cole, W. C.; López, J. J.; Meijere, A. d.; Zhou, H. Chemistry of Materials 2008, 20, 3145-3152. DOI
(19) Latroche, M.; Surblé, S.; Serre, C.; Mellot-Draznieks, C.; Llewellyn, P. L.; Lee, J.; Chang, J.; Jhung, S. H.; Férey, G. Angewandte Chemie International Edition 2006, 45, 8227-8231. DOI
(20) Chun, H.; Jung, H.; Koo, G.; Jeong, H.; Kim, D. Inorg. Chem. 2008, 47, 5355-5359. DOI
(21) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; O'Keeffe, M.; Yaghi, O. M. Science 2002, 295, 469-472. DOI
(22) Zhou, H. A Biomimetic Approach to Metal-Organic Frameworks with High H2 Uptake. 2009 Link (accessed 11/2009).