FEATURE ARTICLE www.rsc.org/materials | Journal of Materials Chemistry Hydrogen storage in metal–organic frameworks{ David J. Collins and Hong-Cai Zhou* Received 26th February 2007, Accepted 21st March 2007 First published as an Advance Article on the web 11th April 2007 DOI: 10.1039/b702858j For any potential hydrogen-storage system, raw uptake capacity must be balanced with the kinetics and thermodynamics of uptake and release. Metal–organic frameworks (MOFs) provide unique systems with large overall pore volumes and surface areas, adjustable pore sizes, and tunable framework–adsorbate interaction by ligand functionalization and metal choice. These remarkable materials can potentially fill the niche between other physisorbents such as activated carbon, which have similar uptake at low temperatures but low affinity for hydrogen at ambient temperature, and chemical sorbents such as hydrides, which have high hydrogen uptakes but undesirable release kinetics and thermodynamics. Introduction Interest is high worldwide in replacing the burning of fossil fuels with generation of energy through hydrogen fuel cells. Electricity generation from hydrogen and oxygen, especially for transportation applications, has the potential to result in a significant reduction of CO2 emissions and air pollution in general. Of course, any fuel, especially one for use in transportation applications, would be useless without an effective storage technology. The United States Department of Energy has set a number of short-term goals for on-board hydrogen storage systems: 6.0 wt% and 45 g L21 by the year 2010, and 9.0 wt% and 81 g L21 by the 2015.1 Additionally, these goals should be met at near-ambient temperatures and applicable pressures (less than 100 bar), and any storage technology must Department of Chemistry and Biochemistry, Miami University, Oxford, OH, USA. E-mail: zhouh@muohio.edu; Fax: +1 513 529-0452; Tel: +1 513 529-8091 { This paper is part of a Journal of Materials Chemistry theme issue on New Energy Materials. Guest editor: M. Saiful Islam. David J. Collins David Collins received his B.S. (Materials Science & Engineering) from Ohio State University in 1995. After graduation, he taught high school physics and chemistry in Houston, Texas, and Cincinnati, Ohio, before returning to school to pursue a Ph.D. in chemistry in 2003. At Miami University, working with Professor Hong-Cai Zhou, David has investigated the synthesis of heterometallic paddlewheel clusters and the synthesis and properties of novel MOF materials, while also pursuing a project to bring inorganic chemistry and crystallography concepts to local high school classes. 3154 | J. Mater. Chem., 2007, 17, 3154–3160 minimize weight and volume contributions to the potential fuel-cell powered vehicle. To maintain current transportation ranges, the average light-duty vehicle would require 5–15 kg of onboard hydrogen storage.2 While liquid fuels such as gasoline and diesel fuel can be easily stored in simple tanks at ambient pressures, a gaseous fuel such as hydrogen poses a real challenge. One alternative is liquefaction; however, this requires cryogenic temperatures, extremely high pressures, or both; compressed H2 technologies have addressed storage pressures of 700 bar at room temperature. In addition, the density of liquid H2 is 70.8 g L21; the 2015 gravimetric goal (81 g L21) precludes any purely physical storage of molecular dihydrogen as a liquid or compressed gas. Aside from improvements in current tank technology, the storage methods presently under the most consideration are physisorbents and chemical and metal hydrides. In addition to total storage capacity, the kinetics and thermodynamics of release and recharging must be considered. Undesirable kinetics and thermodynamics can not only limit the release and recharging rate, but may also add unnecessary Hong-Cai Zhou Hong-Cai ‘‘Joe’’ Zhou obtained his B.S. from Beijing Normal University, and his Ph.D. from Texas A&M University, under the direction of Professor F. A. Cotton, followed by a postdoctoral stint in the laboratory of Professor R. H. Holm at Harvard University. He is currently an associate professor in the Department of Chemistry and Biochemistry at Miami University. His research is in the fields of biomimetic materials chemistry and bio-inspired coordination chemistry. This journal is ß The Royal Society of Chemistry 2007 weight and volume to the vehicle because of the need for extra heat exchangers—this leads to consideration of the heat of formation of the hydrogen-carrying species or heat of hydrogen adsorption of sorbents, DHf or DHads, when evaluating a potential storage method. Chemisorption techniques, involving the formation of hydrides, suffer from binding hydrogen too tightly, i.e., DHf too large—ranging from 50 to over 200 kJ mol21. For metal hydrides, such as LaNi5H6.5, the storage system must operate above ambient temperature (often .400 K) in order to release the hydrogen fuel, while refueling liberates large amounts of energy.2 The thermodynamics of chemical hydrides (such as LiBH4) are such that there are questions regarding the reversibility of the hydrogen uptake and release, while the chemical reaction which releases hydrogen only occurs at elevated temperatures; additionally, the energy input to create the hydride reduces the overall efficiency of the system.2 As a contrast, most physisorption techniques, including carbon nanotubes and other porous materials (activated carbon, zeolites, etc.), suffer from the opposite problem: DHads is typically small, less than 10 kJ mol21, and appreciable adsorption can only be achieved at very low temperatures (typically ,100 K). MOFs as physisorbent materials In the last 10 years, metal–organic frameworks (MOFs, also known as coordination networks or coordination polymers) have become a burgeoning field of research and a promising candidate for hydrogen storage materials due to their exceptionally high porosity, uniform but tunable pore size and well-defined hydrogen occupation sites. Generally speaking, these materials are constructed by coordinate bonds between multidentate ligands and metal atoms or small metalcontaining clusters (referred to as the secondary building unit or SBU, a term borrowed from the description of zeolites). Most have three-dimensional structures incorporating uniform pores and a network of channels. These pores and channels are often filled with guest species, usually solvent from synthesis. Removal of these guests often leads to framework collapse; however, in many cases, framework integrity is preserved, and these voids remain and other guest molecules can then be adsorbed onto this porous structure.3 In May 2003, Yaghi, et al. reported what is believed to be the first measurements of hydrogen adsorption on a MOF: a remarkable 4.5 wt% at 77 K and pressures less than 1 atm, and 1.0 wt% at room temperature at 20 bar on the material Zn4O(bdc) (bdc = 1,4-benzenedicarboxylate) (also referred to as MOF-5 and IRMOF-1).43 These values were later adjusted downward based on follow-up studies,12 but the idea remained: for hydrogen storage, these porous MOFs were a competitive alternative to other physisorption-based materials such as zeolites or activated carbon. Since 2003, at least 60 unique MOFs have been evaluated for their ability to store hydrogen (summarized in Table 1). Coupled with measurements of porosity and surface area based on nitrogen adsorption, some understanding of the many factors that determine the hydrogen uptake by a porous MOF has been developed. Recent computational studies, involving both electronicstructure methods (ab initio and DFT) and molecular This journal is ß The Royal Society of Chemistry 2007 mechanics (Grand Canonical Monte Carlo methods), in addition to increasingly detailed structural characterization of the hydrogen-adsorbed species (including neutron scattering and synchrotron X-ray diffraction) have added insight to these remarkable materials and the mechanisms of hydrogen adsorption. Rather than present merely a summary of all studies that have been performed to date, this Feature Article will focus on those characteristics of MOFs which must be achieved and/or improved in order to reach the promise of hydrogen storage at near-ambient temperatures and reasonable pressures. These factors include achieving sufficient surface area and pore volume, the inclusion of appropriately sized pores, and the formation of high-energy hydrogen binding sites (on either the metal cluster, ligand, or both). The first factors, surface area and pore volume, scale with the overall hydrogen saturation uptake; existing studies (vide infra) show that the surface areas and porosity achieved currently by MOFs can nearly reach the 2010 US DOE gravimetric adsorption goals, albeit at 77 K. The remaining factors influence the interaction between the framework and the hydrogen molecule, DHads; their effects are observable primarily at low pressures and temperatures. In this regime, the pores remain mostly unfilled; hydrogen–framework interactions dominate the shape of the adsorption isotherm. The most facile method to determine the heat of adsorption, DHads, is application of the Clausius–Clapeyron equation to adsorption data collected at two temperatures (typically 77 K and 87 K).44 Other methods include microcalorimetry and estimation via simulation (GCMC, etc.). As can be seen in Table 1, only about one-third of published hydrogen adsorption studies include DHads data; increasing this fraction is necessary to further elucidate the relationships between various structural and chemical factors and this key variable. Since many studies do not directly measure DHads, the hydrogen uptake at 77 K and 1 atm can be used as a comparative proxy for the heat of adsorption. Maximizing this hydrogen affinity will increase the temperatures at which MOFs can adsorb large amounts of hydrogen; this is necessary in order to develop a storage system that meets the above targets at near-ambient temperatures. Pore volume and surface area In a study of zeolites and activated carbons, it was proposed that higher gas adsorption could be achieved on materials with a large volume of micropores with an appropriate diameter (although the specific ‘‘appropriate diameter’’ was left undetermined).45 Much early work was directed toward meeting the first part of this goal—namely, the synthesis of highly porous frameworks, which could then be filled with hydrogen gas. Some examples include several of the ‘‘MIL’’ series of MOFs reported by Férey et al., with pore sizes greater than 25 Å,11 and the isostructural ‘‘IRMOF’’ series with progressively larger pores reported by Yaghi et al.,7,12 all based on carboxylate ligands and possessing noninterpenetrated networks. These materials possess pore volumes greater than 1.5 cm3 g21, and in turn adsorb a sizable amount of hydrogen at high pressures: 6.01 wt% for J. Mater. Chem., 2007, 17, 3154–3160 | 3155 MIL-10111 and 6.7 wt% for IRMOF-20 at 77 K.8 As much as 80% of the volume of these materials is empty space. In view of Aristotle’s observation that ‘‘nature abhors a vacuum,’’ it is unlikely that a material with a significantly greater permanent porosity fraction can be created. Nonetheless, as shown in Fig. 1, it is obvious that large pore volumes and surface areas are necessary for high hydrogen saturation uptakes at 77 K; the same will be true for hydrogen storage at ambient temperature. Pore size and interpenetration Various studies have reported the perhaps counterintuitive finding that smaller pores actually take up hydrogen more effectively than very large ones.16,37 The ideal pore size seems to be 4.5–5 Å, or approximately 2.8–3.3 Å when the van der Waals radii of the atoms composing the pore walls are excluded; this is comparable to the y2.8 Å kinetic diameter of H2. Pores of this size allow the dihydrogen molecule to interact with multiple portions of the framework rather than just one SBU or organic linker, increasing the interaction energy between the framework and H2. Some have also proposed that the curvature of pores or nonlinearity of channels, in addition to the size or chemical functionality, plays a role, although this is difficult to quantify experimentally.37,46 One problem encountered when attempting to quantify the exact relationship between pore size and hydrogen uptake is the variety of methods used to calculate and Surface area, porosity, and hydrogen adsorption data for selected MOFs Table 1 a Material Mn(HCO2)2 Mg3(HCO2)6 MOF-5, Zn4O(bdc)3, IRMOF-1 Sc2(bdc)3 MIL-53(Al), Al(OH)(bdc) MIL-53(Cr), Cr(OH)(bdc) MIL-101, Cr3OF(bdc) IRMOF-2, Zn4O(bbdc)3 IRMOF-3, Zn4O(abdc)3 IRMOF-6, Zn4O(cbbdc)3 IRMOF-8, Zn4O(ndc)3 IRMOF-9, Zn4O(bpdc)3 IRMOF-11, Zn4O(hpdc)3 IRMOF-13, Zn4O(pydc)4 IRMOF-18, Zn4O(tmbdc)3 IRMOF-20, Zn4O(ttdc)3 Mg3(ndc)3 Mn(ndc) Zn4O(L1)3 Zn4O(L2)3 Er2(pdc)3 Y2(pdc)3 MAMS-1, Ni8(tbbdc)6 MOF-74, Zn3O3(dhbdc) HKUST-1, Cu3(btc)2 MIL-96, Al3O(btc)3 MIL-100, Cr3OF(btc) Dy(btc) TUDMOF-1, Mo3(btc)2 PCN-6, Cu3(tatb)2 PCN-9, Co4(tatb)8/3 MOF-177, Zn4O(btb) MIL-102, Cr3OF(ntc)3/2 MOF-505, Cu2(bptc) Cu2(tptc) Cu2(qptc) Zn3(bdt)3 Mn3(bdt)3 Mn3(bdt)8Cl2 Cu(bdt) Mn(btt) Pd(pymo)2 Cu(pymo)2 Zn7O2(pda)5 ZIF-8, Zn(mim)2 ZIF-11, Zn(pim)2 Zn(ndc)(bpe)K Zn3(bpdc)3(bipy) Co3(bpdc)3(bipy) Ni(cyclam)(bpydc) Zn2(bdc)(dabco) b 2 SA /m g 21 Pore volume/cm3 g21 d 240 150d 4170 721d 1590 1500 5500 2544 3062 3263 1818 2613 2340 2100 1501 4590 520 191 502d 396d 427d 676d 0.043 0.332 0.59 0.56 1.9 0.88 1.07 1.14 0.90 0.73 1.53 0.068 0.20 0.13 0.186 0.294 1132 2175 0.39 0.75 2800 655d 2010 3800 1355 5640 42.1 1830 2247d 2932d 640d 290d 530d 200f 2100d 600d 350d 1.0 1810 1947g 303 792e 922e 817 2090 0.67 1.453 0.51 0.12 0.680 0.886 1.138 0.795 0.17 0.663 0.582 0.2 0.33 0.38 0.37 0.75 3156 | J. Mater. Chem., 2007, 17, 3154–3160 H2 uptake at 77 K, 1 atm (wt%) 0.9 0.60 1.5 1.5 2.1 1.8 2.5 1.21 1.42 1.48 1.50 1.17 1.9 1.73 0.89 1.35 0.78 0.57 Maximum H2 uptake (wt%) 77 K 298 K 5.2, 48 bar 0.45, 60 bar 3.8, 16 bar 3.1, 15 bar 6.1, 40 bar 0.43, 80 bar 4.9, 32 bar 3.6, 10 bar 0.4, 30 bar 3.5, 34 bar DHadsc/ kJ mol21 4.8 10i 6.1 9.0 6.7, 70 bar 9.5 1.12, 48 bar 0.98, 48 bar 0.675 0.760 0.6 1.77 2.54 1.6 1.0 1.32 1.75 1.9 1.53 1.25 0.65 2.59 2.52 2.24 1.4 0.9 0.8 0.66 2.1 1.2 0.8 2.3, 26 bar 3.6, 10 bar 1.96, 3 bar 3.28, 26.5 bar 0.35, 65 bar 8.3 6.8 0.15, 73.3 bar 6.3i 10.1 7.5, 70 bar 0.9, 10 bar 4.2h 6.7h 7.01h 0.05, 35 bar 5.99i 6.5j 8.7 8.4 8.8 6.9, 90 bar 1.4, 90 bar 10.1 1.01, 71.43 bar 1.29 1.37 0.8 1.74 1.98 1.1 2.01 3.1, 55 bar 2.0, 40 bar 0.3, 65 barh 7.1 6.8 Reference 4 5 6–8 9 10, 11 10, 11 11 7 7 7, 8 12, 13 7 7, 8 7 12 7, 8 6, 14 15 16 16 17 17 18 7, 8 7, 19 20 11 21 22 23, 24 25 8, 12 26 27–29 29 29 30 30 30 30 31 32 32 33 34 34 35 36, 37 36, 37 38 37 This journal is ß The Royal Society of Chemistry 2007 Fig. 1 Correlation between surface area (Langmuir method, N2) and saturation hydrogen uptake at 77 K. report the size of pores. One set of methods, the application of Dubnin–Astakhov analysis or the Horvath–Kawazoe model to gas sorption data, provides an estimation of pore size, but is limited by the quality of the adsorption data and the gas used. Given that MOFs typically have a highly ordered crystalline structure, it is generally not difficult to determine a highquality single-crystal X-ray structure of these materials. Many researchers apply a variety of software tools to estimate accessible pore volume and pore sizes based on these structures; however, this method relies both on estimations of the van der Waals radii of the atoms along the pore or channel walls, and the particular algorithms of the software packages used. We have found perhaps the most useful information to be simple atom-to-atom distance measurements across the pores or channels, as measured directly from the crystal structure; this allows the reader or reviewer to make ‘‘apples-to-apples’’ comparisons from one paper to the next, and to apply any additional assumptions or estimations as necessary. Interpenetration, the phenomenon in which two or more networks in a structure are physically entangled (as links in a chain), is a common motif in MOF structures.47 The typical effect of interpenetration on porosity is to subdivide large single pores, each bounded by the entire organic linker, into several smaller ones, each bounded by smaller portions of the organic linker. One extreme example of a highly interpenetrated structure with small pores is Zn4O(L1)3 (L1 = 6,69-dichloro-2,29-diethoxy-1,19-binaphthyl-4,49-dibenzoate),16 with a fourfold interpenetrating structure, open channels of less than 5 Å, and BET surface area of only 502 m2 g21; this material adsorbs 1.12 wt% of hydrogen at room temperature and 48 bar, as contrasted to MIL-101, with a surface area of 5500 m2 g21 and hydrogen uptake of only 0.43 wt% under the same conditions. We have recently reported a system which can be synthesized in either an interpenetrated or a noninterpenetrated form, allowing evaluation of interpenetration as an independent criterion resolved from other factors.23 This MOF, Cu3(tatb)2 (also designated PCN-6), contains the copper-carboxylate paddlewheel SBU linked by a trigonal triazine-based ligand, and the catenation isomerism is controlled by the presence or absence of oxalic acid (apparently in a templating role, as it does not appear in the final structure). Not surprisingly, the noninterpenetrated form (PCN-69) has a Table 1 Surface area, porosity, and hydrogen adsorption data for selected MOFs (Continued ) a Material b 2 21 SA /m g Pore volume/cm3 g21 H2 uptake at 77 K, 1 atm (wt%) Maximum H2 uptake (wt%) 77 K 298 K DHadsc/ kJ mol21 Reference Ni2(dhtp) 1083 0.41 0.7 1.8, 60 bar 0.3, 65 bar 39 700d 0.94 10.4 40 NaNi3(OH)(sip) 0.181 0.8 41 Ni2(bipy)3(NO3)4 0.63 0.7 2.5, 15 bar 0.15, 15 bar 41 Ni3(btc)2(pic)6(pd) 1670 0.59 2.08 37 Zn2(bdc)(tmbdc)(dabco) 1400 0.50 1.85 37 Zn2(tmbdc)2(dabco) 1450 0.52 1.70 37 Zn2(ndc)2(dabco) 1610 0.57 1.78 37 Zn2(tfdbc)2(dabco) 1740 0.62 1.68 37 Zn2(tmbdc)2(bipy) IRMOF-8 + Pt/AC 3.5, 100 bar 24.8 42 a Abbreviations: bdc = 1,4-benzenedicarboxylate, bbdc = 2-bromo-1,4-benzenedicarboxylate, abdc = 2-amino-1,4-benzenedicarboxylate, cbbdc = 1,2-cyclobutane-3,6-benzenedicarboxylate, ndc = 2,6-naphthalenedicarboxylate, bpdc = 4,49-biphenyldicarboxylate, hpdc = 4,5,9,10tetrahydropyrene-2,7-dicarboxylate, pydc = pyrene-2,7-dicarboxylate, tmbdc = 2,3,5,6-tetramethylbenzene-1,4,-dicarboxylate, ttdc = thieno[3,2b]thiophene-2,5-dicarboxylate, L1 = 6,69-dichloro-2,29-diethoxy-1,19-binaphthyl-4,49-dibenzoate, L2 = 6,69-dichloro-2,29-benzyloxy1,19-binaphthyl-4,49-dibenzoate, pdc = pyridine-3,5-dicarboxylate, tbbdc = 5-tert-butyl-1,3-benzenedicarboxylate, dhbdc = 2,5-dihydroxy-1,4benzenedicarboxylate, btc = 1,3,5-benzenetricarboxylate, tatb = triazine-4,49,40-s-triazine-2,4,6-triyltribenzoate, btb = 1,3,5-benzenetribenzoate, ntc = naphthalene-1,4,5,8-tetracarboxylate, bptc = biphenyl-3,39,5,59-tetracarboxylate, tptc = terphenyl-3,30,5,50-tetracarboxylate, qptc = quaterphenyl-3,3-9,5,5--tetracarboxylate, bdt = 1,4-benzeneditetrazolate, btt = 1,3,5-benzenetristetrazolate, pymo = 2-pyrimidinolate, pda = 1,4phenylenediacrylate, mim = methylimidizolate, pim = phenylimidizolate, bpe = trans-1,2-bis(4-pyridyl)ethene, bipy = 2,29-bipyridine, cyclam = 1,4,8,11-tetraazacyclotetradecane, bpydc = 2,29-bipyridyl-5,59-dicarboxylate, dabco = 1,4-diazabicyclo[2.2.2]octane, dhtp = 2,5dihydroxyterephthalate, sip = 5-sulfoisophthalate, pic = 3-picoline, pd = 1,2-propanediol, tmbdc = 2,3,5,6-tetramethylbenzene-1,4,-dicarboxylate, tfbdc = 2,3,5,6-tetrafluoro-1,4-benzenedicarboxylate. b Surface area determined using Langmuir method, N2 adsorption, 77 K, unless otherwise noted. c At low coverage, calculated from adsorption isotherms at two or more temperatures, unless otherwise noted. d BET method, N2 adsorption, 77 K. e BET method, Ar adsorption, 87 K. f BET method, O2 adsorption, 77 K. g Calculated from crystal structure. h Estimated saturation limit from Langmuir plot. i Determined directly by microcalorimetry. j Calculated from Grand Canonical Monte Carlo simulation. This journal is ß The Royal Society of Chemistry 2007 J. Mater. Chem., 2007, 17, 3154–3160 | 3157 higher overall porosity, based on the solvent-accessible volume calculated from the single-crystal X-ray structure. However, the interpenetrated form of this MOF exhibits a 41% increase in surface area, 133% increase in volumetric hydrogen uptake, and 29% increase in gravimetric hydrogen uptake when compared to the noninterpenetrated form.23 Two Grand Canonical Monte Carlo (GCMC) studies on interpenetrated MOFs in the IRMOF series demonstrated that in these structures, the smaller pore size and multiplicity of networks allows the dihydrogen molecule to interact with the central (phenyl-containing) portion of multiple ligands, thus increasing the relevance of the non-coordinating portion of the ligand to hydrogen storage.28,48 Both studies agree that the importance of a high heat of adsorption is greatest at low loadings (low H2 pressures), and that the overall pore volume becomes more important at higher loadings. The net result of interpenetration is an increase in the interaction energy, DHads, and is reflected in an increase in the H2 uptake at 77 K and 1 atm. Thermal activation and metal-based H2 binding sites In addition to solvent molecules trapped within the pores of the material, as-synthesized MOFs may also have solvent molecules attached as ligands to the metal centers or incorporated as part of the SBU. Removal of these solvent ligands is referred to as thermal activation, and is often a necessary step to access the full gas-adsorption potential of a material. In some cases, these coordinated solvent molecules may merely protrude into windows or channels, blocking access of dihydrogen molecules into the larger spaces within the framework. In other cases, removal of these ligands (often aqua ligands) leaves the metal cation exposed on the interior surfaces and open to direct approach by the dihydrogen molecule. Chen and coworkers have shown that the removal of axial aqua ligands from dicopper paddlewheel SBUs via thermal activation exposes the copper binding sites in MOF-505;27 Long et al. have demonstrated the same phenomenon in a Mn-containing MOF with tetrazolate ligands,30 as have Bordiga et al. in HKUST-1.49 This generates a so-called unsaturated metal center (UMC), which then interacts strongly with the dihydrogen molecule. These UMCs can be seen as analogous to entatic metal centers in bioinorganic chemistry, in which metal ions (such as the iron in hemoglobin) are forced into an unusual coordination geometry (see Fig. 2)— such a concept has been advanced in our lab by the study of PCN-9, a MOF containing a coordinatively unsaturated Co4(m4-O)(CO2)8 SBU.25 This study also demonstrated via IR spectroscopy that the cobalt site can bind probe molecules CN2 and CO;25 similar results were found in CO binding to the copper center of HKUST-1.49 Exposing these sites greatly increases the ability of the material to adsorb hydrogen; the incorporation of accessible UMCs has been shown to be a viable strategy to increase the hydrogen uptake by MOFs. Single-crystal neutron diffraction of MOF-5 revealed two hydrogen-binding sites, one higher-energy site over the center of the Zn4(m4-O)(CO2)8 SBU, and a second site over the face of 3158 | J. Mater. Chem., 2007, 17, 3154–3160 Fig. 2 (a) Schematic drawing of the active center of hemoglobin; the gold sphere represents an iron atom. (b) Schematic drawing of the tetranuclear M4(m4-O)(CO2)8 SBU containing coordinatively unsaturated metal atoms. (c) The M2(CO2)4 paddlewheel SBU. a ZnO4 tetrahedron.50 Neutron powder diffraction reveals two additional sites in MOF-5 at increased loading: one associated with the zinc-carboxylate moiety Zn(CO2), and one over the phenyl ring of the ligand.51 Generally these agree with an inelastic neutron scattering experiment performed on the same material, differing only in the preferred order of site occupation. MP2 ab initio calculations of simplified MOF models predict the metal-carboxylate linker to be a highenergy binding site, with additional lower-energy sites located around the phenyl ring of the ligand.52 Additional studies probing H2 binding sites have been performed on a number of other MOFs as well. Long et al. have employed neutron powder diffraction to confirm that H2 is closely associated with the UMCs found in a Mn-tetrazole MOF.31 A combination of temperature-programmed desorption and inelastic neutron scattering of H2-loaded NaNi(sip)2 (sip = 5-sulfoisophthalate) has revealed a number of discrete H2 binding sites, the strongest of which can be associated with an unsaturated Ni site.40 A combined DFT and GCMC dynamical study of the dicopper-paddlewheel-containing MOF-505 shows that, as expected based on the earlier thermal-activation study (vide supra), the binding energy of H2 on MOF-505 is highest at the copper UMC sites exposed by thermal activation.27 Most recently, neutron powder diffraction of D2-loaded HKUST-1 identified six distinct D2 sites, shown in Fig. 3. The first, highest-energy site is associated with the copper UMCs (on the axes of the paddlewheel SBUs), near enough to indicate significant interaction with the d9 Cu(II) center. The remaining sites fill competitively from the smallest to largest pores, with these sites located near the benzene ring and carboxylate moieties of the ligand.53 Ligand structure and functionalization Although the metal sites and/or the SBU are the preferential adsorption sites for hydrogen, the organic linker can play an important secondary role in increasing adsorption further, as shown above. In the IRMOF series developed by Yaghi et al., the basic structural motif of Zn4(m4-O)(CO2)8 SBUs connected by aromatic phenyl-containing linkers is repeated to generate a series of isostructural materials, which differ only in the central portion of the ligand.7,12 Increasing the aromaticity of this central portion, from a simple phenyl ring (MOF-5/IRMOF-1) to cyclobutylbenzene (IRMOF-6) to naphthalene (IRMOF-8) This journal is ß The Royal Society of Chemistry 2007 Dissociative adsorption of hydrogen Fig. 3 D2 Sites in HKUST-1, identified via neutron powder diffraction, numbered in order of occupation with increased loading. Top: shown along [001] (left) and [111] (right). Bottom: axial Cu(II) paddlewheel UMC site (left), along [111] in the 5 Å small pore with 3.5 Å side windows (middle), and along [100] showing the 9 Å pore. Reprinted with permission from ref. 53. Copyright 2006, American Chemical Society. One potential method for enhancing the hydrogen uptake involves the concept of ‘‘hydrogen spillover’’. Long known on supported metal catalysts, spillover occurs by the dissociative chemisorption of hydrogen onto the metal surface, followed by the migration of monoatomic hydrogen onto the supporting material.56 Intimate contact between this supporting material and a third material allows ‘‘secondary spillover’’; this technique has been applied by Li et al. in preliminary studies, enhancing the hydrogen adsorption of MOF-5 and IRMOF-8 as much as eightfold.42 To achieve this remarkable result, 5% Pt on activated carbon was mixed with the MOF and sucrose, then ground to make an intimate mixture; melting and subsequent carbonization of the sucrose provided a carbon ‘‘bridge’’ permeating the mixture, allowing atomic hydrogen to migrate from the metal surface across the carbon to the MOF. The smaller size of the hydrogen atom should allow the adsorbate to penetrate smaller pores which are inaccessible to the dihydrogen molecule; in other words, the potential exists to bypass the gravimetric storage limits imposed by the density and incompressibility of liquid dihydrogen. Dissociative adsorption should also increase the affinity of the framework for hydrogen, as the adsorbed species in this case is the hydrogen atom rather than the dihydrogen molecule. Conclusions increases the hydrogen uptake dramatically, from 0.5 wt% to 1.0 wt% and 1.5 wt%, respectively.12 One further way to enhance the affinity of the ligand for the dihydrogen molecule is by chemical functionalization: the introduction of an electron-donating group (or groups) to the central portion of the ligand. Again, the IRMOF series provides illustration: replacing one hydrogen atom on the central benzene ring of the linker in MOF-5 with –Br or –NH2 affords IRMOF-2 and -3, respectively; replacing all four hydrogens with methyl groups affords IRMOF-18.7 A fifth member of the series, IRMOF-20, replaces the phenyl ring of bdc with a thieno-[3,2-b]thiophene moiety; this was expected to enhance hydrogen adsorption significantly due to the increased polarizability of the heteropolycyclic ligand.7 MP2 computational studies suggest that these electrondonating substituents should increase the affinity of the aromatic ring for H2;54 however, little enhancement was found in these functionalized IRMOFs in practice.7 A similar lack of enhancement was reported by Kim et al. for pillared MOFs constructed in part by ligands with all phenyl H atoms replaced with either –F or –CH3.37 This may be due to partial restriction of the pore size or blocking of the high-affinity metal-based binding sites by the larger ligand, effectively canceling out any benefit derived from electronic enhancement of the ligand. It has also been proposed that cyanide and N-heterocyclic ligands in general may have a higher affinity for dihydrogen than purely graphitic ligands, in part based on hydrogen adsorption studies with carbon, carbon nitride, and boron nitride nanotube structures;55 MOFs that exhibit this functionality include systems using triazine- and tetrazolatebased ligands.24,30 This journal is ß The Royal Society of Chemistry 2007 In order to reach hydrogen storage goals, a method must be found which has both sufficient capacity and acceptable uptake/release kinetics and thermodynamics at ambient temperatures and reasonable operating pressures. The density of molecular dihydrogen imposes strict limits on the potential of tank storage technologies to meet the 2015 gravimetric goals; hydride species uptake sufficient amounts of atomic hydrogen, but the thermodynamics of these systems raises daunting challenges. Current MOFs and other porous physisorbents present moderate uptake volumes but insufficient interaction energy to retain hydrogen at ambient temperature. However, MOFs offer a path to the middle ground—the volume advantages of tanks and porous adsorbents and the potential of moderate affinity to hydrogen, allowing uptake and release at reasonable temperatures. Although the heat of adsorption of dihydrogen on current MOFs is still quite low, this property appears ‘‘tunable’’ in MOFs by a variety of methods, potentially allowing these materials to reach a ‘‘quasi-chemisorptive’’ regime in which a higher-energy interaction between dihydrogen and the framework exists but remains short of true chemical bonding. Catalyzed dissociation allows the material to interact with atomic hydrogen rather than molecular dihydrogen, increasing further both the interaction energy and the volume fraction which can be penetrated by the adsorbed hydrogen. The next several years are sure to see both an increased understanding of the tunability of these interactions and the further development of novel MOFs with ever-greater uptake capacities—all in the hopes of a more efficient, cleaner energy future. J. Mater. Chem., 2007, 17, 3154–3160 | 3159 References 1 Hydrogen, Fuel Cells & Infrastructure Technologies Program: Multi-year Research, Development, and Demonstration Plan, U.S. Dept. of Energy, 2005, http://www.eere.energy.gov/hydrogenandfuelcells/mypp/. 2 L. Schlapbach and A. Züttel, Nature, 2001, 414, 353–358. 3 A. J. Fletcher, K. M. Thomas and M. J. Rosseinsky, J. Solid State Chem., 2005, 178, 2491. 4 D. N. Dybtsev, H. Chun, S. H. Yoon, D. Kim and K. Kim, J. Am. Chem. Soc., 2004, 126, 32. 5 J. A. Rood, B. C. Noll and K. W. Henderson, Inorg. Chem., 2006, 45, 5521. 6 M. Dincă and J. R. Long, J. Am. Chem. Soc., 2005, 127, 9376. 7 J. L. C. Rowsell and O. M. Yaghi, J. Am. Chem. Soc., 2006, 128, 1304. 8 A. G. Wong-Foy, A. J. Matzger and O. M. Yaghi, J. Am. Chem. Soc., 2006, 128, 3494. 9 J. Perles, M. Iglesias, M. A. Martin-Luengo, M. A. Monge, C. Ruiz-Valero and N. Snejko, Chem. Mater., 2005, 17, 5837. 10 G. Férey, M. Latroche, C. Serre, F. Millange, T. Loiseau and A. Percheron-Guégan, Chem. Commun., 2003, 2976. 11 M. Latroche, S. Surblé, C. Serre, C. Mellot-Draznieks, P. L. Llewellyn, J.-H. Lee, J.-S. Chang, S. H. Jhung and G. Férey, Angew. Chem., Int. Ed., 2006, 45, 8227. 12 J. L. C. Rowsell, A. R. Millward, K. S. Park and O. M. Yaghi, J. Am. Chem. Soc., 2004, 126, 5666. 13 A. Dailly, J. J. Vajo and C. C. Ahn, J. Phys. Chem. B, 2006, 110, 1099. 14 I. Senkovska and S. Kaskel, Eur. J. Inorg. Chem., 2006, 4564. 15 H. R. Moon, N. Kobayashi and M. P. Suh, Inorg. Chem., 2006, 45, 8672. 16 B. Kesanli, Y. Cui, M. R. Smith, E. W. Bittner, B. C. Bockrath and W. Lin, Angew. Chem., Int. Ed., 2005, 44, 72. 17 J. Jia, X. Lin, A. J. Blake, N. R. Champness, P. Hubberstey, L. Shao, G. Walker, C. Wilson and M. Schröder, Inorg. Chem., 2006, 45, 8838. 18 S. Ma, D. Sun, X.-S. Wang and H.-C. Zhou, Angew. Chem., Int. Ed., 2007, 46, 2458. 19 B. Panella, M. Hirscher, H. Pütter and U. Müller, Adv. Funct. Mater., 2006, 16, 520. 20 T. Loiseau, L. Lecroq, C. Volkringer, J. Marrot, G. Férey, M. Haouas, F. Taulelle, S. Bourrelly, P. L. Llewellyn and M. Latroche, J. Am. Chem. Soc., 2006, 128, 10223. 21 X. Guo, G. Zhu, Z. Li, F. Sun, Z. Yang and S. Qiu, Chem. Commun., 2006, 3172. 22 M. Kramer, U. Schwarz and S. Kaskel, J. Mater. Chem., 2006, 16, 2245. 23 S. Ma, D. Sun, M. W. Ambrogio, J. A. Fillinger, S. Parkin and H.-C. Zhou, J. Am. Chem. Soc., 2007, 129, 1858. 24 D. Sun, S. Ma, Y. Ke, D. J. Collins and H.-C. Zhou, J. Am. Chem. Soc., 2006, 128, 3896. 25 S. Ma and H.-C. Zhou, J. Am. Chem. Soc., 2006, 128, 11734. 26 S. Surblé, F. Millange, C. Serre, T. Düren, M. Latroche, S. Bourrelly, P. L. Llewellyn and G. Férey, J. Am. Chem. Soc., 2006, 128, 14889. 27 B. Chen, N. W. Ockwig, A. R. Millward, D. S. Contreras and O. M. Yaghi, Angew. Chem., Int. Ed., 2005, 44, 4745. 28 Q. Yang and C. Zhong, J. Phys. Chem. B, 2006, 110, 655. 29 X. Lin, J. Jia, X. Zhao, K. M. Thomas, A. J. Blake, G. S. Walker, N. R. Champness, P. Hubberstey and M. Schröder, Angew. Chem., Int. Ed., 2006, 45, 7358. 3160 | J. Mater. Chem., 2007, 17, 3154–3160 30 M. Dincă, A. F. Yu and J. R. Long, J. Am. Chem. Soc., 2006, 128, 8904. 31 M. Dincă, A. Dailly, Y. Liu, C. M. Brown, D. A. Neumann and J. R. Long, J. Am. Chem. Soc., 2006, 128, 16876. 32 J. A. R. Navarro, E. Barea, J. M. Salas, N. Masciocchi, S. Galli, A. Sironi, C. O. Ania and J. B. Parra, Inorg. Chem., 2006, 45, 2397. 33 Q.-R. Fang, G.-S. Zhu, M. Xue, Q.-L. Zhang, J.-Y. Sun, X.-D. Guo, S.-L. Qiu, S.-T. Xu, P. Wang, D.-J. Wang and Y. Wei, Chem.–Eur. J., 2006, 12, 3754. 34 K. S. Park, Z. Ni, A. P. Côté, J. Y. Choi, R. Huang, F. J. UribeRomo, H. K. Chae, M. O’Keeffe and O. M. Yaghi, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 10186. 35 B. Chen, S. Ma, F. Zapata, E. B. Lobkovsky and J. Yang, Inorg. Chem., 2006, 45, 5718. 36 J. Y. Lee, L. Pan, S. P. Kelly, J. Jagiello, T. J. Emge and J. Li, Adv. Mater., 2005, 17, 2703. 37 H. Chun, D. N. Dybtsev, H. Kim and K. Kim, Chem.–Eur. J., 2005, 11, 3521. 38 E. Y. Lee and M. P. Suh, Angew. Chem., Int. Ed., 2004, 43, 2798. 39 P. D. C. Dietze, B. Panell, M. Hirsche, R. Blo and H. Fjellvåg, Chem. Commun., 2006, 959. 40 P. M. Forster, J. Eckert, B. D. Heiken, J. B. Parise, J. W. Yoon, S. H. Jhung, J. S. Chang and A. K. Cheetham, J. Am. Chem. Soc., 2006, 128, 16846. 41 X. Zhao, B. Xiao, A. J. Fletcher, K. M. Thomas, D. Bradshaw and M. J. Rosseinsky, Science, 2004, 306, 1012. 42 Y. Li, F. H. Yang and R. T. Yang, J. Phys. Chem. C, 2007, 111, 3405; Y. Li and R. T. Yang, J. Am. Chem. Soc., 2006, 128, 726; Y. Li and R. T. Yang, J. Am. Chem. Soc., 2006, 128, 8136. 43 N. L. Rosi, J. Eckert, M. Eddaoudi, D. T. Vodak, J. Kim, M. O’Keefe and O. M. Yaghi, Science, 2003, 300, 1127. 44 J. Rouquerol, F. Rouquerol and K. Sing, Adsorption By Powders and Porous Solids, Academic Press, London, 1998. 45 M. G. Nijkamp, J. E. M. J. Raaymakers, A. J. van Dillen and K. P. de Jong, Appl. Phys. A: Mater. Sci. Process., 2001, 72, 619. 46 K. W. Chapman, P. J. Chupas, E. R. Maxey and J. W. Richardson, Chem. Commun., 2006, 4013. 47 S. T. Batten and R. Robson, Angew. Chem., Int. Ed., 1998, 37, 1460. 48 H. Frost, T. Duren and R. Q. Snurr, J. Phys. Chem. B, 2006, 110, 9565. 49 C. Prestipino, L. Regli, J. G. Vitillo, F. Bonino, A. Damin, C. Lamberti, A. Zecchina, P. L. Solari, K. O. Kongshaug and S. Bordiga, Chem. Mater., 2006, 18, 1337. 50 E. C. Spencer, J. A. K. Howard, G. J. McIntyre, J. L. C. Rowsell and O. M. Yaghi, Chem. Commun., 2006, 278. 51 T. Yildirim and M. R. Hartman, Phys. Rev. Lett., 2005, 95, 215504. 52 T. Sagara, J. Klassen and E. Ganz, J. Chem. Phys., 2004, 121, 12543. 53 V. K. Peterson, Y. Liu, C. M. Brown and C. J. Kepert, J. Am. Chem. Soc., 2006, 128, 15578. 54 T. Sagara, J. Klassen, J. Ortony and E. Ganz, J. Chem. Phys., 2005, 123, 014701. 55 X. D. Bai, D. Zhong, G. Y. Zhang, X. C. Ma, S. Liu and E. G. Wang, Appl. Phys. Lett., 2001, 79, 1552; S. H. Jhi and Y. K. Kwon, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 69, 245407. 56 A. J. Robell, E. V. Ballou and M. Boudart, J. Phys. Chem., 1964, 68, 2748; S. T. Srinivas and P. K. Rao, J. Catal., 1994, 148, 470; W. C. Conner and J. L. Falconer, Chem. Rev., 1995, 95, 759. This journal is ß The Royal Society of Chemistry 2007