Understanding Volumetric and Gravimetric Hydrogen Adsorption

Forum Article www.acsami.org

Understanding Volumetric and Gravimetric Hydrogen Adsorption Trade-off in Metal−Organic Frameworks Diego A. Gómez-Gualdrón,*,†,‡ Timothy C. Wang,†,§ Paula García-Holley,§ Ruth M. Sawelewa,‡ Edwin Argueta,⊥ Randall Q. Snurr,⊥ Joseph T. Hupp,§ Taner Yildirim,*,∥ and Omar K. Farha*,§,# ‡

Department of Chemical and Biological Engineering, Colorado School of Mines, Golden, Colorado 80401, United States Department of Chemistry and ⊥Department of Chemical and Biological Engineering, Northwestern University, Evanston, Illinois 60208, United States ∥ NIST center for Neutron Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United States # Department of Chemistry, King Abdulaziz University, Jeddah 22254, Saudi Arabia §

S Supporting Information *

ABSTRACT: Metal−organic frameworks (MOFs) are porous crystalline materials that are promising for adsorption-based, on-board storage of hydrogen in fuel-cell vehicles. Volumetric and gravimetric hydrogen capacities are the key factors that determine the size and weight of the MOF-filled tank required to store a certain amount of hydrogen for reasonable driving range. Therefore, they must be optimized so the tank is neither too large nor too heavy. Because the goals of maximizing MOF volumetric and gravimetric hydrogen adsorption loadings individually are incompatible, an in-depth understanding of the trade-off between MOF volumetric and gravimetric loadings is necessary to achieve the best compromise between these properties. Here we study, both experimentally and computationally, the trade-off between volumetric and gravimetric cryo-adsorbed hydrogen deliverable capacity by taking an isoreticular series of highly stable zirconium MOFs, NU-1101, NU-1102, and NU-1103 as a case study. These MOFs were studied under recently proposed operating conditions: 77 K/100 bar →160 K/5 bar. We found the difference between highest and lowest measured deliverable capacity in the MOF series to be ca. 40% gravimetrically, but only ca. 10% volumetrically. From our molecular simulation results, we found hydrogen “monolayer” adsorption to be proportional to the surface area, whereas hydrogen “pore filling” adsorption is proportional to the pore volume. Thus, we found that the higher variability in gravimetric deliverable capacity in contrast to the volumetric capacity, occurs due to the proportional relation between gravimetric surface area and pore volume in the NU-110x series in contrast to the inverse relation between volumetric surface area and void fraction. Additionally, we find better correlations with geometric surface areas than with BET areas. NU-1101 presents the highest measured volumetric performance with 46.6 g/L (9.1 wt %), whereas NU-1103 presents the highest gravimetric one with 12.6 wt % (43.2 g/L). KEYWORDS: nnoporous materials, zirconium MOFs, molecular modeling, cryoadsorption, energy storage

■

INTRODUCTION Hydrogen gas can be used on-board to power fuel cell vehicles (FCVs) while producing zero carbon emissions.1 Widespread use of FCVs, however, depends on safe, economical and practical strategies to store and release on-board enough hydrogen for commercially attractive driving ranges.2 Current FCVs commercially distributed by Toyota3 and Honda4 rely on storage at ambient temperature with compression of hydrogen gas to 700 bar (CHG). Still, the high storage pressure creates concerns regarding compression cost and on-board storage safety, which imposes significant design constraints on the tank, especially for passenger vehicles.5 To reduce the storage pressure, some prototype FCVs by BMW use hydrogen storage at cryogenic temperature with compression to 350 bar.6 At © XXXX American Chemical Society

cryogenic temperature, the required storage pressure could be further reduced by filling the tank with a suitable porous adsorbent material such as a metal−organic framework (MOF)which is constituted by inorganic nodes connected by organic linkers7,8and exploiting hydrogen-adsorbent interactions to enhance hydrogen gas densification.9 Design of cryo-adsorption tanks for hydrogen storage in FCVs has been explored by GM subsidiaries.10 Recently Special Issue: Hupp 60th Birthday Forum Received: January 23, 2017 Accepted: March 30, 2017

A

DOI: 10.1021/acsami.7b01190 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Forum Article

ACS Applied Materials & Interfaces

not been directly studied for NU-1101 and NU-1102, but given a similar coordination chemistry and even smaller pores than NU-1103, the two former MOFs are expected to be at least as stable as the NU-1103.

proposed cryo-adsorption operating conditions correspond to cycling between 100 bar at 77 K and 5 bar at 160 K.11 The lowest desorption pressure is constrained by engine settings, and the moderate temperature increase at this pressure enhances the adsorbent deliverable capacity, which is the difference in the amount of hydrogen adsorbed at the storage pressure and the amount retained at the lowest desorption pressure. In recent work,8 we used molecular simulation to predict hydrogen deliverable capacities at the proposed cryoadsorption conditions in more than 13,000 MOFs constructed in the computer. We identified some MOFs with predicted volumetric deliverable capacities as high as 57 g/L, and a large number of MOFs with a predicted deliverable capacity higher than that of both the adsorbent-free tank under the same conditions (30 g/L) and CHG (37 g/L). Some of the identified MOFs had already been synthesized (e.g., MOF-5,12 NU125,13 NOTT-101,14 NU-110315) but not yet experimentally tested for hydrogen storage at the 100 bar/77 L → 5 bar/160 K operating conditions. Indeed, MOFs have been studied for hydrogen storage for over a decade motivated by their highly tunable modular structure, but the focus has been placed on adsorption at ambient temperature.9,16−18 To design a suitable on-board hydrogen storage system, both volumetric adsorption and gravimetric hydrogen adsorption are important. High values for the latter are desired to ensure the on-board tank is not prohibitively heavy. On the other hand, volumetric hydrogen adsorption is crucial because it directly relates to the required volume of the on-board tank. However, hydrogen adsorption studies have tended to focus primarily on gravimetric adsorption.9,16,18 In our recent large-scale computational study on hydrogen cryo-adsorption on MOFs,8 we found that for different MOF isoreticular series (of different topologies), gravimetric and volumetric deliverable capacities initially increase together until the latter reaches a maximum. Beyond this point, the volumetric deliverable capacity decreases as the gravimetric one increases. However, the decrease is surprisingly slow, which allows one to find MOFs possessing similarly high volumetric deliverable capacities and markedly different gravimetric ones. To elucidate how the trade-off between these two critical performance metrics occurs around the volumetric adsorption maximum, we chose to study experimentally and computationally hydrogen cryo-adsorption in three members of the zirconium-based isoreticular NU-110x series15 of ftw topology, namely NU-1101, NU-1102, and NU-1103. Our choice is motivated by their similarly high stabilities and their similar pore structures, yet systematically varying pore sizes and surface areas. NU-1103 has been reported15 to have the highest surface area for any synthesized zirconium-based MOFa subclass of MOFs notable for their exceptional hydrothermal stability.19−25 Indeed, as part of our recent hydrogen cryo-adsorption study, we measured the hydrogen deliverable capacity of NU-1103 (43.2 g/L and 12.6 wt %)8 on a sample synthesized and activated two-years before adsorption tests. The studied NU1103 sample did not evince degradation during the two-year “wait” period. This remarkable display of stability motivated us to investigate the isostructural MOFs, NU-1101 and NU-1102. Note that the water-stability of NU-1103 has been thoroughly studied in ref 15, where an activated NU-1103 sample was shown to fully retain its original porosity and structure (confirmed by nitrogen adsorption and X-ray diffraction) after being soaked in water for 18h and reactivated via solventexchange and supercritical CO2 activation. Water-stability has

■

METHODS

Experimental Section. MOF samples NU-1101, NU-1102, and NU-1103 were synthesized following previously reported procedures.15 Briefly, in a glass vial, a zirconium source (ZrCl4 or ZrOCl2) and corresponding organic linkers were dissolved in DMF (or DEF), along with optimized amount of benzoic acid as a modulator. The reaction vial was heated at elevated temperature in a gravity convection oven. After the synthesis, the MOF sample was washed and soaked with fresh DMF for three times over the course of 1 day, and then solvent-exchanged with acetone. After the solvent was fully exchanged to acetone, the sample was thermally activated under dynamic vacuum at 120 °C for 12 h. For NU-1103, the sample was activated using the supercritical CO2 drying technique reported in ref 26 and subsequently heated at 80 °C under dynamic vacuum for 12 h. Low and high pressure gas sorption measurements were performed at the NIST Center for Neutron Research, using a carefully calibrated, high accuracy, computer-controlled Sieverts apparatus. A detailed description of the experimental setup, calibration and the isotherm has been previously published.27,28 High-purity grade (99.999%) gases were used in adsorption measurements. NU-1103 originally underwent activation two years ago, after which a nitrogen isotherm was collected. The activated sample was then kept in a He-glovebox for two-years until a new nitrogen isotherm was measured for this study; notably the stored sample was not reactivated. The pair of nitrogen isotherms proved nearly identical (consistent with nearly identical values for pore volume and BET area; see Figure S2) indicating negligible sample degradation, even if the initial quality of the sample was slightly lower than for the previously reported preparation of NU1103.15 NU-1101 and NU-1102 were synthesized two years ago and stored in a jar (without activation). At the start of the current study we activated the stored samples as described above and then recorded N2 isotherms. The pore volumes and BET areas obtained from the isotherms are close to those of the published for separately prepared samples that had not been subjected to extended storage.15 Taken together, the sorption results for NU-1101, NU-1102, and NU-1103 highlight the resistance to degradation of these high-porosity materials, regardless whether they are stored in activated (with solvent absent) or nonactivated (i.e., as-synthesized with solvent present) form. Computational. NU-1101, NU-1102, and NU-1103 structures were constructed using the topologically based crystal constructor (ToBaCCo) code we used in previous work,8 by using the appropriate building blocks and the ftw blueprint. The constructed structures were optimized in Materials Studio29 with the unit cell parameters constrained to match experiments.15 Note that crystallographic structures have been determined experimentally for these three MOFs, but as for many MOFs, they are not suitable for simulation because of crystallographic disorder.30 Adsorption simulations were performed using grand canonical Monte Carlo (GCMC) simulations using the RASPA code.31 Each isotherm point was calculated using 2500 equilibration cycles followed by 2500 cycles for data collection. Each cycle corresponds to N Monte Carlo moves, where N is the maximum between 20 and the number of adsorbate molecules in the unit cell. Adsorbates underwent equal probability translation, rotation, insertion and deletion moves. MOF atoms remained immobile. Nonbonded interactions were described by a Lennard-Jones (LJ) plus Coulomb potential. Hydrogen LJ parameters and charges were assigned according to the DarkimLevesque model with Feyman-Hibbs corrections.32 Nitrogen LJ parameters and charges were assigned according to the TraPPE model.33 MOF LJ parameters were assigned according to the Universal Force Field.34 Lorentz−Berthelot mixing rules were used to derive cross-term LJ parameters. MOF electronic charges were obtained via density functional theory (DFT) calculated on appropriately “capped” versions of the B


Forum Article


Figure 1. MOF structures and absolute adsorption nitrogen isotherms. (a−c) MOF structures with largest pore in each structure represented by a yellow sphere for (a) NU-1101, (b) NU-1102, and (c) NU-1103. (d−f) Comparison of measured (line with points) and simulated (smooth line) nitrogen isotherms for (d) NU-1101, (e) NU-1102, and (f) NU-1103. MOF samples were synthesized 2 years ago.

Figure 2. Absolute volumetric hydrogen adsorption isotherms in studied MOFs. (a) NU-1101, (b) NU-1102, and (c) NU-1103. Simulated (dashed lines) and experimentally measured (lines with points) at the same temperature appear in the same color.

■

constituent building blocks (see Figure S1 for details) using Gaussian 09.35 The charges were fitted to reproduce the electrostatic potential surrounding the building blocks according to the CHELPG method.36 The building blocks carry the charge information when assembled into the corresponding MOFs, with only a typically small residual charge being redistributed to ensure MOF neutrality. Hydrogen molecules contacting MOF pore walls in simulation snapshots were counted using the same routine we used for counting nitrogen molecules contacting MOF pore walls in earlier work.15,37 Geometric nitrogen-accessible surface areas (NASAs) and BET areas (from simulated and measured nitrogen isotherms) were calculated according to the procedures explained in detail in earlier work.37 Pore size was calculated according to the method of Gelbs and Gubbins,38 which calculates the largest sphere that can enclose randomly selected points inside the MOF pores.

RESULTS AND DISCUSSION Hydrogen Adsorption Measurements. Figure 1 illustrates the structures of the MOFs investigated in this work and compares the measured nitrogen isotherms in NU-1101, NU1102, and NU-1103 with the simulated isotherms in the perfect crystals. The differences between measured and simulated nitrogen saturation loading indicate that the quality of the activated samples is high but, especially in the case of NU-1102, somewhat lower than the quality of the activated samples we reported in ref 15. Recall, however, that nitrogen adsorption measurements two-years apart for NU-1103 indicate that the starting quality of the specific samples is retained. The measured pore volumes of the NU-1101, NU-1102, and NU1103 samples studied here are 99, 81, and 91%, respectively, of the pore volumes of the corresponding perfect crystals. Similarly, the BET areas calculated from measured nitrogen

C


Forum Article


Table 1. Measured and Simulated Properties for NU-110x Series, Including Absolute H2 Adsorption Loadings and H2 Deliverable Capacity (ΔH2) NU-1101

NU-1102

NU-1103

property

exp.

sim.

exp.

sim.

exp.

sim.

pore volume (cc/g) gravimetric BET area (m2/g) gravimetric NASA (m2/g) volumetric BET area (m2/cm3) volumetric NASA (m2/cm3) smallest pore diameter (Å) largest pore diameter (Å) heat of adsorption (kJ/mol) H2@100 bar/77 K (g/L) ΔH2@100 bar/77 K → 5 bar/77 K (g/L) ΔH2@100 bar/77 K →5 bar/160 K (g/L) H2@100 bar/77 K (wt %) ΔH2@100 bar/77 K →5 bar/77 K (wt %) ΔH2@100 bar/77 K →5 bar/160 K (wt %)

1.72 4340

1.72 4440 4420 2030 2025 9.5 17.2 5.5 52.9 33.3 50.7 10.2 5.9 9.8

1.65 3720

2.05 4710 4710 1900 1900 11.1 20.5 4.9 54.2 34.1 52.4 11.6 7.6 11.3

2.72 6245

2.97 6885 5645 2050 1682 13.5 24.2 4.5 52.6 40.0 51.1 15.1 11.9 14.4

1990

5.5 48.7 29.9 46.6 9.5 6.1 9.1

1500

4.5 45.3 30.5 43.7 9.9 6.9 9.6

1860

3.8 44.9 33.3 43.2 13.0 10.1 12.6

presents a deliverable capacity of 31 g/L between 100 bar/77K and 5 bar/160 K). Remarkably, the highest (46.6 g/L for NU-1101) and lowest (43.2 g/L for NU-1103) measured nonisothermal volumetric deliverable capacities among the three MOFs differ by only about 7% despite appreciable differences in pore size, pore volume, and surface area. In contrast, the highest (12.6 wt % for NU-1103) and the lowest (9.1 wt % for NU-1101) measured nonisothermal gravimetric deliverable capacities differ by ca. 40%. Such behavior is consistent with previous observations from computational screening of large numbers of candidate MOFs, namely, nonisothermal, volumetric vs gravimetric deliverable capacity plots presenting a relatively flat maximum in volumetric deliverable capacity, after which this quantity slowly decreases as the corresponding gravimetric deliverable capacity decreases. Hence NU-1101, NU-1102, and NU-1103 should be excellent candidates to analyze in order to better understand this trade-off behavior. Since the same behavior is evident in the data obtained via simulations with idealized versions of NU-1101, NU-1102, and NU-1103 and since simulated isotherms agree reasonably well with measured ones (as do other measured and simulated properties in Table 1), we now perform our analysis of volumetric and gravimetric adsorption based on simulation data. This approach eliminates varying and potentially confounding contributions from structural imperfections, residual solvent, and similar factors. Analysis of Gravimetric Hydrogen Adsorption. Because the nonisothermal deliverable capacities for NU-1101, NU1102, and NU-1103 correlate well with loadings at 100 bar, we focused our analysis on understanding the latter. We considered the relation of these loadings with the MOF surface area since adsorption loadings have often been correlated to this property.39−41 For instance, Chahine’s rule39 states that the maximum gravimetric hydrogen excess loading, Nexcess, in a carbon-based adsorbent at cryogenic temperature relates to the gravimetric surface area, GSA, by

isotherms in NU-1101, NU-1102, and NU-1103 samples are 98, 79, and 91%, respectively, of the BET areas calculated from the corresponding simulated isotherms. Thus, sample quality according to pore volumes and BET areas are consistent. The measured hydrogen isotherms for the MOF samples are presented in Figure 2 and compared with corresponding simulated ones on the perfect crystal. We measured and simulated hydrogen adsorption at 77 K, 160 and 295 K up to 100 bar. At 100 bar, measured uptakes at 77 and 160 K are nearly six and three times, respectively, higher than at 295 K. Measured and simulated hydrogen isotherms essentially overlap at 295 and 160 K up to 100 bar, and at 77 K up to 5 bar. At higher pressure, simulated hydrogen loadings at 77 K somewhat overestimate measured loadings partly due to somewhat lower pore volumes and BET areas in the samples relative to the values for the corresponding perfect crystals. Indeed, measured hydrogen loadings at 100 bar and 77 K for NU-1101, NU1102, and NU-1103 are 92, 83, and 86%, respectively, of the corresponding simulated loadings. Table 1 lists the various measured and simulated properties for NU-1101, NU-1102, and NU-1103. Conversion between gravimetric and volumetric quantities is done using the crystallographic density of the MOFs. The listed properties include hydrogen loadings at 100 bar/77 K and isothermal and nonisothermal hydrogen deliverable capacities. Hydrogen isotherms at 77 K (Figure 2) show that at 5 bar NU-1101, NU-1102, and NU-1103 retain 39, 33, and 26%, respectively, of the hydrogen adsorbed at 100 bar, significantly lowering the isothermal deliverable capacity. In fact, although NU-1103 adsorbs somewhat less hydrogen at 100 bar on a volumetric basis than NU- 1101 and NU-1102, it presents a slightly higher isothermal deliverable capacity because of lower hydrogen retention at low pressure. In any case, isothermal volumetric deliverable capacities at cryogenic conditions for these MOFs (30−33 g/L) appear a little lower than the deliverable capacity of CHG (37 g/L) and only a little higher than an adsorbentfree, isothermal cryo-tank at identical operating conditions (30 g/L between 100 and 5 bar). On the other hand, by moderately raising the desorption temperature to 160 Kas recently proposed11only 4% of the hydrogen adsorbed at 100 bar remains at 5 bar, resulting in attractive values for the nonisothermal deliverable capacity (note that the adsorbentfree, nonisothermal cryo-tank at identical operating conditions

Nexcess[wt%] = 2.28 × 10−3GSA[m 2/g]

(1) 2

which roughly corresponds to 1 wt % for each 500 m /g of surface area. This rule, which was derived based on hydrogen adsorption on ideal graphene, overestimates the measured excess hydrogen maxima when instead applied to NU-1101, NU-1102, and NU-1103 samples (see Figures S3−S6). We still D


Forum Article

ACS Applied Materials & Interfaces Ntotal[wt%] = 4.04 × 10−3GSA[m 2/g] − 8.4

do observe, however, that excess hydrogen increases with gravimetric surface area. Because for vehicular applications total hydrogen loading, Ntotal, is more relevant, let us recall that Ntotal = Nexcess + ρbulk Vp

The gravimetric loadings at 100 bar in NU-1101, NU-1102 and NU-1103 predicted by eq 6 are plotted in Figure 3b along with corresponding simulated loadings vs NASA. Equation 6 seems to work well to explain the gravimetric performance trends for the NU-110x MOFs. A sizable improvement of 48% in the gravimetric loading of NU-1103 at 100 bar relative to that of NU-1101 is primarily the result of a 28% increase in the gravimetric surface area of NU-1103 relative to that of NU1101. Analysis of Volumetric Hydrogen Adsorption. Since increasing the MOF gravimetric surface area increases total gravimetric hydrogen loading at 100 bar, the question that follows then is whether something similar occurs with the MOF volumetric surface area and the volumetric hydrogen loading at 100 bar. We can rewrite eq 2 in volumetric terms as

(2)

where ρbulk is the hydrogen density at the relevant bulk conditions (thus independent of the MOF) and Vp is the MOF pore volume. Assuming that Nexcess is proportional to the gravimetric surface area, we can write for total hydrogen at 100 bar: Ntotal = a1GSA + a 2Vp

(3)

where a1 and a2 are constants (and a2 is evidently ρbulk). Data for 18 MOFs listed in previous work42 shows that pore volume increases with gravimetric surface area (see Figure S9) roughly as Vp = a3GSA + a4

(6)

(4)

Ntotal = Nexcess + ρbulk Vf

where a3 and a4 are constants. Thus, it is apparent that simply focusing on increasing either property will result in higher gravimetric hydrogen loadings. Based on simulated properties for NU-1101, NU-1102, and NU-1103, a3 and a4 are equal to 0.0010 and −2.7, respectively, according to Figure 3a. In this

(7)

where Vf is the MOF void fraction. If we assume, analogous to eq 3, that Nexcess is proportional to the volumetric surface area, we have Ntotal = b1 VSA + b2 Vf

(8)

where b1 and b2 are constants (where analogous to eq 3, b2 is ρbulk). The challenge here is that contrary to the relation between gravimetric surface area and pore volume, the volumetric surface area is known to initially increase with void fraction until it reaches a maximum, beyond which the volumetric surface area decreases as void fraction continues to increase.24 The latter is the case for NU-1101, NU-1102, and NU-1103, which appear to present an inverse linear correlation between void fraction and volumetric surface area in the form of Vf = b3 VSA + b4

(9)

where, from Figure 4a, we can see that b3 is −0.0003 and b4 is 1.358. Thus, it is evident that there are two opposing effects in eq 8. Again, we used NASA because it correlated better (R2 = 0.998) than BET area (R2 = 0.021) with void fraction. Substituting eq 9 into eq 8, we get to Ntotal = D1 VSA + D2

(10)

where D2 is equal to b2b4 and D1 is equal b1 + b2b3. Using the volumetric loading at 100 bar (50.1 g/L) and volumetric surface area (2025 m2/cm3) of NU-1101 we solve for b1 to be 0.014634 to make eq 10 into Ntotal = 5.33 × 10−3VSA + 42.1

Figure 3. Correlation between MOF gravimetric properties for NU110x series. (a) MOF pore volume vs NASA, (b) simulated (with GCMC: solid circles) and calculated (with eq 6: empty circles) MOF gravimetric hydrogen loading vs NASA.

The volumetric loadings at 100 bar in NU-1101, NU-1102, and NU-1103 predicted by eq 11 are plotted in Figure 4b along with the corresponding simulated loadings vs NASA. Although the predictions with eq 11 miss the subtle maximum at NU1102, they capture the most salient feature of the volumetric loadings, which is that the values are closely similar for the three sorbents. NU-1101 only presents 3.5% higher volumetric loading at 100 bar than NU-1103 despite the 20% higher volumetric surface area of NU-1101 relative to NU-1103. This occurs due to lower porosity of NU-1101 as inferred from the discussion, around eq 8 (vide supra), on opposing effects. Now we proceed to understand these opposing effects by analyzing simulation snapshots up to 100 bar.

figure we used NASA because it correlated better (R2 = 0.998) with pore volume than BET area did (R2 = 0.984). Substituting eq 4 into eq 3 results in the linear relationship Ntotal = C1GSA + C2

(11)

(5)

where C2 is equal to a2a4 and C1 is equal a1 + a2a3. To this point only a1 is unknown, which determines the assumed proportionality between gravimetric excess hydrogen at 100 bar and gravimetric surface area. Using the simulated 9.5 wt % Ntotal of NU-1101, a1 would be 0.00094. Thus, eq 5 would become E


Forum Article


to 40 bar, the hydrogen density at the center of both large and small pores increases moderately, but still higher hydrogen density is observed at locations where hydrogen contacts framework atoms. The high hydrogen density (darker in Figure 5) region now extends from the unit cell corners and central pyrenes to linker arms. As the pressure increases from 40 to 100 bar, it becomes more difficult to discern changes in the density maps, perhaps only the increase in the density at the center of the pores being notable. To better understand our observations from the density maps in Figure 5, we breakdown the absolute hydrogen adsorption isotherm at 77 K for NU-1101, NU-1102, and NU1103 on the basis of hydrogen siting. Figure 6 illustrates the hydrogen adsorption breakdown for these simulated isotherms. The breakdown quantifies what amount of adsorbed hydrogen is in contact or not with the pore walls as the pressure increases up to 100 bar. To use an analogous terminology to that we used to breakdown nitrogen adsorption in MOFs in previous work,37 we denote hydrogen contacting the pore walls as contributing to the formation of an adsorption monolayer and hydrogen not contacting the walls as contributing to pore filling. For the three MOFs, up to 5 bar there is a steep increase in total hydrogen adsorption (Figure 2), which mostly corresponds to hydrogen contributing to monolayer formation (Figure 6). The monolayer formation is essentially complete around 40 bar, which is close to the pressure at which the excess maximum occurs for these materials. Note also that the changes in relative contribution to total loading by monolayer formation and pore filling present clear trends across the MOFs. Higher MOF volumetric surface area (NU-1101 > NU-1102 > NU-1103) results in higher hydrogen loading corresponding to monolayer formation. On the other hand, higher MOF void fraction or pore volume (NU-1103 > NU-1102 > NU-1101) results in higher hydrogen loading corresponding to pore filling. These opposing trends, as observed in Figure 7 for adsorption at 100 bar in the three studied zirconium MOFs, tend to compensate with each other, which explains the relatively unchanged total volumetric loading despite the different properties of the MOFs. As Figure 8 shows, the pore filling loading for the NU-110x series correlates linearly with void fraction, while the monolayer loading correlates with volumetric surface area. Considering that nonisothermal deliverable capacities for the 100 bar/77K → 5 bar/160 K operating conditions and volumetric loading at 100 bar correlate well, it is clear that, to maximize hydrogen

Figure 4. Correlation between MOF volumetric properties for NU110x series. (a) MOF void fraction vs NASA, (b) simulated (with GCMC: solid circles) and calculated (with eq 11: empty circles) MOF volumetric hydrogen loading vs NASA.

Hydrogen Adsorption Siting. Adsorbate density maps for NU-1103 as the pressure increases and the MOF fills with hydrogen are shown in Figure 5. Density maps for NU-1101 and NU-1102 (Figures S10−S15) show that, at least qualitatively, these structures fill up with hydrogen in a similar way to NU-1103 in terms of where hydrogen is primarily located at each pressure. At pressures below 5 bar, the density maps show that hydrogen is concentrated at locations where it contacts framework atoms. Specifically, hydrogen density appears primarily at the corners of the unit cell, which correspond to strong interaction “pockets” created by the convergence of linker arms around the zirconium nodes, and secondarily at the surface of the central pyrenes. At these low pressures, the hydrogen density at the center of the large (centered at [1/2, 1/2, 1/2]) and small pores (centered at [1/ 2, 0, 0] and symmetrically related positions) is extremely low (Figure S16 illustrates pore locations). As the pressure increases

Figure 5. Hydrogen average density maps on NU-1103 obtained from simulations at different pressures. NU-1103 framework in blue, hydrogen density increases as map color changes from white (minimum density at the given pressure) to black (maximum density at the given pressure). Complete set of density maps for all samples presented as Figures S10−15. F


Forum Article


Figure 6. Hydrogen siting breakdown for NU-110x series. Simulated absolute volumetric adsorption loadings are divided between loading corresponding to hydrogen molecules contacting (black line) and not contacting (purple line) the pore walls. (a) NU-1101, (b) NU-1102, (c) NU1103.

storage performance, the MOF design must maximize volumetric surface area while maintaining pore volume as high as possible and vice versa. On the basis of the fittings in Figure 8, a hypothetical ftw Zr-MOF with a volumetric surface area and void fraction combination to the right of the 70 g/L dashed line in Figure 9 could potentially have an adsorption

Figure 7. Hydrogen siting breakdown for absolute hydrogen adsorption loading at 100 bar for absolute NU-1101, NU-1102, and NU-1103. Hydrogen from monolayer shown in black (top curve) and hydrogen from pore filling shown in purple (bottom curve).

Figure 9. Combination of NASA and void fraction for 2133 Zr-MOFs. Points colored according to MOF topology. The dashed lines are obtained based on the fitted equations in Figure 8. The region to the right of each line provides a combination of NASA and void fraction that could potentially meet the hydrogen loading indicated in each line at 100 bar/77 K. NU-110x MOFs (star points) are roughly over the 50 g/L line.

loading of 70 g/L. However, no ftw Zr-MOF presented surface area-void fraction combinations in that region. The exact volumetric surface area and void fraction combination required to meet a given hydrogen loading combination will likely somewhat change with MOF chemistry and topology, so the fittings in Figure 8 may not exactly apply to other MOFs. However, to provide some perspective on what property combinations can be obtained with chemically similar MOFs (i.e., Zr-MOFs, absence of polar organic functional groups) to the NU-110x series, we have added combinations of surface area and void fraction computationally calculated for 2133 zirconium-based MOFs of 17 different framework topologies (these structures were obtained as a subset of the 13,512 MOFs we studied in ref 8) Note that none of these MOFs falls to the right of the 70 g/L dashed line either.

Figure 8. Fitting of hydrogen monolayer and pore filling loadings to MOF properties. (a) hydrogen pore filling loading vs void fraction. (b) Hydrogen monolayer loading vs NASA. G


Forum Article

ACS Applied Materials & Interfaces Note that our interest in Zr-MOFs stems from their typically high stability, but there are other examples of stable MOFs based on other metals. Two important, highly stable MOFs based on a similar strategy of strong coordination bonds are the Fe(III)-, Al(III)-, and Cr(III)based MIL-100 and MIL-101 MOFs developed by Ferey and co-workers.43 For instance, MIL-101 was found to remain stable after months of exposure to ambient conditions. From crystallographic structures, we estimate the void fractions and volumetric surface areas for perfect MIL-100 and MIL-101 crystals to be 0.75 and 0.81 and 1890 m2/cm3 and 1480 m2/cm3, respectively. It is also noteworthy that, although generally less stable, researchers have intensely studied MOFs with under-coordinated metal sites for hydrogen storage. These MOFs often correspond to MOFs with Cu-paddlewheels.8,16,44,45 The rationale is that under-coordination boosts hydrogen−metal interactions that in turn may boost hydrogen loadings. In a literature survey, we identified NOTT-11214 to have the highest volumetric loading among previously experimentally tested MOFs at 77 K and high pressure (ca. 50 g/L at 80 bar), and a reported heat of adsorption of 5.6 kJ/mol. Although NOTT112 presents Cu-paddlewheels, its hydrogen loading and heat of adsorption is only slightly different than NU-1101 (ca. 48 g/ L and 5.5 kJ/mol). This suggests textural properties to play a more prominent role than most chemistries found in MOFs. This partly explain our focus on void fraction and surface area in the present work. Note, however, that chemistries such as metal alkoxides have been predicted46 to bind hydrogen with binding energies in the 20−80 kJ/mol range.

■

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. ORCID

Timothy C. Wang: 0000-0002-2736-2488 Randall Q. Snurr: 0000-0003-2925-9246 Joseph T. Hupp: 0000-0003-3982-9812 Omar K. Farha: 0000-0002-9904-9845 Author Contributions †

D.A.G.-G. and T.C.W. contributed equally

Funding

Colorado School of Mines and National Science Foundation (DMR-1308799). Notes

The authors declare the following competing financial interest(s): O.K.F., J.T.H., and R.Q.S. have a financial interest in NuMat Technologies, a start-up company that is seeking to commercialize MOFs.

■

ACKNOWLEDGMENTS D.A.G-.G. is grateful for start-up funds from Colorado School of Mines. R.Q.S. thanks the National Science Foundation (DMR-1308799) for financial support. Calculations were possible thanks to the BlueM supercomputer cluster at Colorado School of Mines.

CONCLUSIONS

■

Combined experimental and computational studies of an isoreticular series of high-porosity zirconium-based MOFs have proven useful for understanding and gaining insight into the structural and textural trade-offs entailed in simultaneously obtaining high volumetric and high gravimetric uptake (and release) of molecular hydrogen under cryogenic conditions. The measured nitrogen isotherms indicated little degradation for NU-1101, NU-1102, and NU-1103 during a two-year period. Good performance for these MOFs, close to theoretical predictions, was observed despite the lengthy time between synthesis and adsorption tests. Measured nonisothermal deliverable capacities of NU-1101, NU-1102, and NU-1103 are 46.6 g/L (9.2 wt %), 43.7 g/L (9.6 wt %), and 43.2 g/L (12.6 wt %), respectively. (Simulations indicate that ideal versions of these materials (i.e., defect-free and perfectly activated versions) may yield nonisothermal deliverable capacities as high as 14.4 wt % (NU-1003) and ∼51 to 52 g/ L (all three materials). As informed by molecular simulation, the pronounced differences in gravimetric performance are due to the concerted increase of gravimetric surface area with pore volume, whereas the relatively similar volumetric performance for the three MOFs is due to the conflicting trends for volumetric surface area and void fraction.

■

experimental heat of adsorption calculations, complete density maps for simulated hydrogen adsorption (PDF)

REFERENCES

(1) Thomas, C. E.; James, B. D.; Lomax, F. D., Jr; Kuhn, I. F., Jr Fuel Options for the Fuel Cell Vehicle: Hydrogen, Methanol or Gasoline? Int. J. Hydrogen Energy 2000, 25, 551−567. (2) Felderhoff, M.; Weidenthaler, C.; von Helmolt, R.; Eberle, U. Hydrogen Storage: The Remaining Scientific and Technological Challenges. Phys. Chem. Chem. Phys. 2007, 9, 2643−2653. (3) The 2017 Toyota Mirai: The Future Available Now; http://www. toyota.com/mirai/fcv.html (accesed Dec 2016). (4) Clarity Fuel Cell: A Clear Path to the Future; http://automobiles. honda.com/clarity, (accessed Dec 2016). (5) Schlapbach, L.; Zuttel, A. Hydrogen-Storage Materials for Mobile Applications. Nature 2001, 414, 353−358. (6) First Hydrogen Station with Two Types of Refueling Technology Opening in Munich; http://www.bmwblog.com/2015/07/17/firsthydrogen-station-with-two-types-of-refuelling-technology-opening-inmunich/ (accessed Dec 2016). (7) Furukawa, H.; Cordova, K. E.; O’Keeffe, M.; Yaghi, O. M. The Chemistry and Applications of Metal-Organic Frameworks. Science 2013, 341, 1230444. (8) Gomez-Gualdron, D. A.; Colon, Y. J.; Zhang, X.; Wang, T. C.; Chen, Y.-S.; Hupp, J. T.; Yildirim, T.; Farha, O. K.; Zhang, J.; Snurr, R. Q. Evaluating Topologically Diverse Metal-Organic Frameworks for Cryo-Adsorbed Hydrogen Storage. Energy Environ. Sci. 2016, 9, 3279− 3289. (9) Sculley, J.; Yuan, D.; Zhou, H.-C. The Current Status of Hydrogen Storage in Metal-Organic Frameworks-Updated. Energy Environ. Sci. 2011, 4, 2721−2735. (10) Arnold, G.; Eberle, U.; Hasenauer, D. U.S. Patent US 7517396 B2, April 14, 2009. (11) Engineering an Adsorbent-Based Hydrogen Storage System: What Have We Learned? energy.gov/sites/prod/files/2015/02/f19/fcto_h2_ storage_summit_siegel.pdf (accessed Sept 2015).

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.7b01190. Additional details on computational and experimental methods, excess adsorption isotherms and details on H


Forum Article

ACS Applied Materials & Interfaces (12) Wong-Foy, A. G.; Matzger, A. J.; Yaghi, O. M. Exceptional H2 Saturation Uptake in Microporous Metal−Organic Frameworks. J. Am. Chem. Soc. 2006, 128, 3494−3495. (13) Wilmer, C. E.; Farha, O. K.; Yildirim, T.; Eryazici, I.; Krungleviciute, V.; Sarjeant, A. A.; Snurr, R. Q.; Hupp, J. T. GramScale, High-Yield Synthesis of a Robust Metal-Organic Framework for Storing Methane and Other Gases Energy. Energy Environ. Sci. 2013, 6, 1158−1163. (14) Yan, Y.; Lin, X.; Yang, S.; Blake, A. J.; Dailly, A.; Champness, N. R.; Hubberstey, P.; Schroder, M. Exceptionally High H2 storage by a Metal-Organic Polyhedral Framework. Chem. Commun. 2009, 1025− 1027. (15) Wang, T. C.; Bury, W.; Gomez-Gualdron, D. A.; Vermeulen, N. A.; Mondloch, J. E.; Deria, P.; Zhang, K.; Moghadam, P. Z.; Sarjeant, A. A.; Snurr, R. Q.; Stoddart, J. F.; Hupp, J. T.; Farha, O. K. Ultrahigh Surface Area Zirconium MOFs and Insights into the Applicability of the BET Theory. J. Am. Chem. Soc. 2015, 137, 3585−3591. (16) Zhao, D.; Yuan, D.; Zhou, H.-C. The Current Status of Hydrogen Storage in Metal-Organic Frameworks. Energy Environ. Sci. 2008, 1, 222−235. (17) Colón, Y. J.; Fairen-Jimenez, D.; Wilmer, C. E.; Snurr, R. Q. High-Throughput Screening of Porous Crystalline Materials for Hydrogen Storage Capacity near Room Temperature. J. Phys. Chem. C 2014, 118, 5383−5389. (18) Basdogan, Y.; Keskin, S. Simulation and Modelling of MOFs for Hydrogen Storage. CrystEngComm 2015, 17, 261−275. (19) Cavka, J. H.; Jakobsen, S.; Olsbye, U.; Guillou, N.; Lamberti, C.; Bordiga, S.; Lillerud, K. P. A New Zirconium Inorganic Building Brick Forming Metal Organic Frameworks with Exceptional Stability. J. Am. Chem. Soc. 2008, 130, 13850−13851. (20) Morris, W.; Volosskiy, B.; Demir, S.; Gándara, F.; McGrier, P. L.; Furukawa, H.; Cascio, D.; Stoddart, J. F.; Yaghi, O. M. Synthesis, Structure, and Metalation of Two New Highly Porous Zirconium Metal−Organic Frameworks. Inorg. Chem. 2012, 51, 6443−6445. (21) Feng, D.; Gu, Z.-Y.; Li, J.-R.; Jiang, H.-L.; Wei, Z.; Zhou, H.-C. Zirconium-Metalloporphyrin PCN-222: Mesoporous Metal−Organic Frameworks with Ultrahigh Stability as Biomimetic Catalysts. Angew. Chem., Int. Ed. 2012, 51, 10307−10310. (22) Mondloch, J. E.; Bury, W.; Fairen-Jimenez, D.; Kwon, S.; DeMarco, E. J.; Weston, M. H.; Sarjeant, A. A.; Nguyen, S. T.; Stair, P. C.; Snurr, R. Q.; Farha, O. K.; Hupp, J. T. Vapor-Phase Metalation by Atomic Layer Deposition in a Metal−Organic Framework. J. Am. Chem. Soc. 2013, 135, 10294−10297. (23) Feng, D.; Chung, W.-C.; Wei, Z.; Gu, Z.-Y.; Jiang, H.-L.; Chen, Y.-P.; Darensbourg, D. J.; Zhou, H.-C. Construction of Ultrastable Porphyrin Zr Metal−Organic Frameworks through Linker Elimination. J. Am. Chem. Soc. 2013, 135, 17105−17110. (24) Gomez-Gualdron, D. A.; Gutov, O. V.; Krungleviciute, V.; Borah, B.; Mondloch, J. E.; Hupp, J. T.; Yildirim, T.; Farha, O. K.; Snurr, R. Q. Computational Design of Metal−Organic Frameworks Based on Stable Zirconium Building Units for Storage and Delivery of Methane. Chem. Mater. 2014, 26, 5632−5639. (25) Lin, Q.; Bu, X.; Kong, A.; Mao, C.; Zhao, X.; Bu, F.; Feng, P. New Heterometallic Zirconium Metalloporphyrin Frameworks and Their Heteroatom-Activated High-Surface-Area Carbon Derivatives. J. Am. Chem. Soc. 2015, 137, 2235−2238. (26) Nelson, A. P.; Farha, O. K.; Mulfort, K. L.; Hupp, J. T. Supercritical Processing as a Route to High Internal Surface Areas and Permanent Microporosity in Metal−Organic Framework Materials. J. Am. Chem. Soc. 2009, 131, 458−460. (27) Peng, Y.; Krungleviciute, V.; Eryazici, I.; Hupp, J. T.; Farha, O. K.; Yildirim, T. Methane Storage in Metal−Organic Frameworks: Current Records, Surprise Findings, and Challenges. J. Am. Chem. Soc. 2013, 135, 11887−11894. (28) Zhou, W.; Wu, H.; Hartman, M. R.; Yildirim, T. Hydrogen and Methane Adsorption in Metal−Organic Frameworks: A High-Pressure Volumetric Study. J. Phys. Chem. C 2007, 111, 16131−16137. (29) Material Studio; Accelrys Software: San Diego, CA, 2001−2011.

(30) Chung, Y. G.; Camp, J.; Haranczyk, M.; Sikora, B. J.; Bury, W.; Krungleviciute, V.; Yildirim, T.; Farha, O. K.; Sholl, D. S.; Snurr, R. Q. Computation-Ready, Experimental Metal−Organic Frameworks: A Tool to Enable High-Throughput Screening of Nanoporous Crystals. Chem. Mater. 2014, 26, 6185−6192. (31) Dubbeldam, D.; Calero, S.; Ellis, D. E.; Snurr, R. Q. RASPA: Molecular Simulation Software for Adsorption and Diffusion in Flexible Nanoporous Materials. Mol. Simul. 2016, 42, 81−101. (32) Levesque, D.; Gicquel, A.; Darkrim, F. L.; Kayiran, S. B. Monte Carlo Simulations of Hydrogen Storage in Carbon Nanotubes. J. Phys.: Condens. Matter 2002, 14, 9285. (33) Potoff, J. J.; Siepmann, J. I. Vapor−Liquid Equilibria of Mixtures Containing Alkanes, Carbon Dioxide, and Nitrogen. AIChE J. 2001, 47, 1676−1682. (34) Rappe, A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A.; Skiff, W. M. Uff, a Full Periodic Table Force Field for Molecular Mechanics and Molecular Dynamics Simulations. J. Am. Chem. Soc. 1992, 114, 10024−10035. (35) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F. o.; Bearpark, M. J.; Heyd, J.; Brothers, E. N.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J.; Gaussian; Gaussian, Inc.: Wallingford, CT, 2009. (36) Breneman, C. M.; Wiberg, K. B. Determining Atom-Centered Monopoles from Molecular Electrostatic Potentials. The Need for High Sampling Density in Formamide Conformational Analysis. J. Comput. Chem. 1990, 11, 361−373. (37) Gómez-Gualdrón, D. A.; Moghadam, P. Z.; Hupp, J. T.; Farha, O. K.; Snurr, R. Q. Application of Consistency Criteria to Calculate BET Areas of Micro- and Mesoporous Metal−Organic Frameworks. J. Am. Chem. Soc. 2016, 138, 215−224. (38) Gelb, L. D.; Gubbins, K. E. Pore Size Distributions in Porous Glasses: A Computer Simulation Study. Langmuir 1999, 15, 305−308. (39) Panella, B.; Hirscher, M.; Roth, S. Hydrogen Adsorption in Different Carbon Nanostructures. Carbon 2005, 43, 2209−2214. (40) Frost, H.; Snurr, R. Q. Design Requirements for Metal-Organic Frameworks as Hydrogen Storage Materials. J. Phys. Chem. C 2007, 111, 18794−18803. (41) Goldsmith, J.; Wong-Foy, A. G.; Cafarella, M. J.; Siegel, D. J. Theoretical Limits of Hydrogen Storage in Metal−Organic Frameworks: Opportunities and Trade-Offs. Chem. Mater. 2013, 25, 3373− 3382. (42) Farha, O. K.; Eryazici, I.; Jeong, N. C.; Hauser, B. G.; Wilmer, C. E.; Sarjeant, A. A.; Snurr, R. Q.; Nguyen, S. T.; Yazaydın, A. Ö .; Hupp, J. T. Metal−Organic Framework Materials with Ultrahigh Surface Areas: Is the Sky the Limit? J. Am. Chem. Soc. 2012, 134, 15016− 15021. (43) Ferey, G.; Mellot-Draznieks, C.; Serre, C.; Millange, F.; Dutour, J.; Surble, S.; Margiolaki, I. A Chromium Terephtalate-Based Solid with Unusually Large Pore Volumes and Surface Area. Science 2005, 309, 2040. (44) Yuan, D.; Zhao, D.; Sun, D.; Zhou, H.-C. An Isoreticular Series of Metal−Organic Frameworks with Dendritic Hexacarboxylate Ligands and Exceptionally High Gas-Uptake Capacity. Angew. Chem. 2010, 122, 5485. (45) Ma, S.; Simmons, J. M.; Sun, D.; Yuan, D.; Zhou, H.-C. Porous Metal_organic Frameworks Based on an Anthracene Derivative: I


Forum Article

ACS Applied Materials & Interfaces Syntheses, Structure Analysis, and Hydrogen Sorption Studies. Inorg. Chem. 2009, 48, 5263. (46) Getman, R. B.; Miller, J.H.; Wang, K.; Snurr, R. Q. Metal Alkoxide Functionalization in Metal-Organic Frameworks for Enhanced Ambient-Temperature Hydrogen Storage. J. Phys. Chem. C 2011, 115, 2066.

J


Understanding Volumetric and Gravimetric Hydrogen Adsorption

Recommend Documents