The assembly of water at the interface of materials has become one of the central topics in biology, chemistry, and materials sciences . It can lead to catalytic reactions at oxide surfaces [2,3,4] and has also attracted considerable interest due to promising applications in photocatalysis, electrochemistry, and sensors [5,6,7,8]. Although understanding the mechanism of water dissociation at surfaces is of paramount interest, its study on flat oxide surfaces is unfortunately highly non-trivial—mostly due to experimental challenges in structural characterization and high pressure required (on the order of 20 atm) for in-situ studies on flat surfaces [1,7,9,10,11,12]. In this respect, metal organic frameworks (MOFs) are attractive systems to study reactions since they are well-controlled crystalline environments with well-defined metal oxide centers [3,4,13,14], in which high gas densities can be achieved at relatively low external pressures. MOFs are networks of metal ions linked by organic ligands, forming a well-characterized cross-linked structure. Due to their porous nature, gas molecules can penetrate deeply into the MOF network  and experience adsorption forces that allow for significantly longer residence times than at surfaces under similar pressures and temperatures, thus fostering an environment much more conducive to trigger, observe, and study desired reactions.
MOFs are already well studied for their high surface areas and nano-porous structure as they provide an ideal environment for many applications ranging from gas storage and separation to sensors and catalysis [16,17,18,19,20]. Amongst the vast amount of existing MOFs, MOF-74 is of great interest since it contains a high density of coordinately unsaturated metal centers (also called open metal sites) in metal-oxide pyramid clusters, which act as active adsorption sites for many small molecules such as H2, CO, CO2, NO, CH4, and H2O [21,22,23,24,25,26,27,28,29]. In particular, Zn-MOF-74 exhibits a strong affinity to water, making it ideal for examining various water reaction mechanisms . We have already studied the reaction of water molecules alone in Zn-MOF-74 [2,3,14,31], finding direct evidence for water dissociation at only 150 °C—this was achieved through the use of D2O instead of H2O, observing a clear fingerprint peak at 970 cm−1 that corresponds to a O–D bending vibration of OD groups formed upon deuteration of the organic linker . The concerted action between open metal sites and linker phenolate group plays an essential role in breaking up water molecules since water molecule establish strong coordinative bond with metal center via its oxygen and hydrogen bonding interaction with nearby –C–O– moiety from organic linker. With an increase in temperature the water molecule dissociates into D and OD . The OD binds to the open metal sites, while the D atom is transferred to the oxygen of phenolate group. The reaction with water (H2O) itself follows of course the same dissociation pathway, but the spectroscopic signature is impossible to detect with H2O as hydrogen blue shifts the peak outside the phonon gap of the MOF, where it strongly couples to many other modes. It was also demonstrated that the formation of water networks within the MOF pores was crucial to further lower the corresponding reaction barrier via a proton-exchange mechanism and that physical obstruction of this network with inert molecules (He, Ar) hinders the reaction .
In the present work, we explore the role of co-adsorbed alcohol molecules and present a kinetic analysis of the water dissociation reaction inside Zn-MOF-74 in the presence of additional guest molecules. We find that alcohol molecules affect the dissociation rate, either by enhancing or blocking the reaction depending on the type of interaction resulting from such co-adsorption. Specifically, methanol enhances the reaction due to H-bonding interactions, while isopropyl alcohol (IPA) hinders it because steric interactions dominate in that case. The knowledge derived from these combined studies is directly relevant to fields such as catalysis and sensors by providing a means to control water-related reactions, as well as to applications such as gas storage and separation by suppressing water dissociation and therefore extending the MOF lifetime.
2.1. Sample Preparation and In-Situ Infrared Spectroscopy
Zn-MOF-74 powder (1.5 mg) was gently pressed onto a KBr pellet and placed inside a high-pressure cell (Specac Ltd., Orpington, UK; product number P/N 5850c). This high-pressure cell was located in the sample compartment of a Nicolet 6700 FTIR spectrometer (Thermo Scientific Inc., Mountain View, CA, USA) with the sample at the focal point of the beam. All the infrared (IR) spectroscopic data were collected with a liquid N2-cooled mercury cadmium telluride (MCT-B or MCT-A from Thermo Scientific Inc., Mountain View, CA, USA) detector. The cell was connected to different gas lines for exposure and a vacuum line for evacuation. All spectra were recorded in transmission mode from 400 cm−1 (MCT-B) or 650 cm−1 (MCT-A) to 4000 cm−1 (4 cm−1 spectral resolution). Regular water (H2O) and heavy water (D2O) were used, with most of the work performed with D2O to avoid a MOF phonon overlap in the spectral range of the O–H bending vibration [2,3].
2.2. Experimental Measurement Conditions
Once inside the high-pressure cell, the MOF powder was activated by annealing at 180 °C for at least 4 h in vacuum (<50 mTorr) to remove the solvent and ambient humidity from the inside of MOF’s pores. The sample was then cooled down to room temperature. In a mixing gas chamber connected to the high-pressure cell (where the sample is located). 8 Torr of alcohol (methanol or isopropyl alcohol) was prepared with 8 Torr of deuterated water (D2O). The alcohol/D2O mixture was then introduced into the main chamber and allowed to stabilize for 10 min, after which the temperature was raised to 180 °C. Spectra were recorded as a function of time until stabilization of the 970 cm−1 peak was reached (~8 h). For completeness, the results are compared to data previously reported by our group using He and pure D2O . Measurements with pure MeOH and its deuterated analog (MeOD) are also presented as reference in the Supplementary Information (See Figure S2).
2.3. Computational Details
Ab initio modeling was performed at the density functional theory (DFT) level in Vienna Ab initio simulation package (VASP) [32,33]. To include van der Waals interactions of guest molecules within the MOF, the exchange-correlation functional vdW-DF was used [34,35,36,37]. The energy cutoff was set to 600 eV and only the Γ point was considered due to the size of the system. The 54 atom rhombohedral primitive cell of Zn-MOF-74 was first relaxed until the forces on the atoms were below 1 meV/Å. Binding energies were then calculated for MeOH, H2O, OH, and H (and deuterated equivalents), with the hydrogen or deuterium at the metal center and on the oxygen atom of the MOF organic linker. After the guest molecules were placed into the MOF, each structure was again relaxed until the forces on all atoms were below 1 meV/Å. Reaction barriers were calculated with a standard transition-state search algorithm, i.e., the nudged-elastic band method [38,39], as implemented in VASP.
3. Results and Discussion
3.1. Experimental Quantification of the Water Dissociation Reaction Rate
To determine the impact of guest molecules on the water dissociation rate through chemical interactions, we studied this reaction in the presence of alcohol molecules inside the MOF as a function of temperature and time. After MOF activation, the alcohol/D2O mixture was introduced into the cell at room temperature and IR absorption spectra were subsequently recorded with the sample at 180 °C, monitoring the above-mentioned O–D bending vibration at ~970 cm−1, which is the product of the D2O dissociation . The IR spectra were recorded until the intensity of the 970 cm−1 peak stabilized. In this manner, the kinetics of methanol/D2O and isopropanol/D2O mixtures could be compared for different alcohol pressures with previously reported data with either pure D2O or inert gas/D2O mixtures , using the integrated area of the 970 cm−1 peak as a quantification of D2O dissociation. Results with pure methanol and pure deuterated methanol are also presented for comparison in the supplementary information (see Figure S2).
Figure 1a shows the representative water dissociation fingerprint at 970 cm−1 after 8 h of reaction at 180 °C for the different mixtures analyzed inside the MOF, which is sufficient time for the reaction to stabilize in all environments. All spectra are normalized to the quantity (weight) of the MOF powder. The feature at 1003 cm−1, only observed for the mixture with 8 Torr MeOH, will be discussed shortly. Figure 1b shows the kinetic evolution of the tested reactions. It is clear that water dissociation rates strongly depend on the environment: the D2O 2D + O reaction is faster in the presence of methanol and slower in the presence of isopropanol or inert gases such as He. Experimentally, we find that the water reaction rates inside MOF-74 are fastest to slowest as follows: 8 Torr MeOH > 4 Torr MeOH > Pure D2O > 8 Torr IPA > 950 Torr He. Table 1 reports the percent of dissociated water molecules, as compared with pure water (100%), as function of guest molecules. These data can be used to examine the dependence on the alcohol pressure, for instance in the case of methanol. The dependence is clearly not linear, but can be well fitted with a quadratic fit, , where y is the percent dissociation and x the pressure, pointing to the importance of the local environment. Indeed, water dissociation is faster when there are more MeOH molecules inside the MOF pore, i.e., when a network can be formed.
Methanol is interesting, namely because of its small size that minimizes steric hindrance within the MOF pore. Different pressures of methanol were tested while keeping the D2O pressure at 8 Torr for all experiments. The infrared spectra measured after introduction of a MeOH/D2O mixture as a function of temperature and time are presented in Figure 2. After introducing a methanol/D2O mixture into Zn-MOF-74 at room temperature, the C–O stretch mode of methanol gas phase was observed at 1033 cm−1 (see Figure S3). As the temperature increases, this band red-shifts to 1010 cm−1 at 170 °C and to 1003 cm−1 at 180 °C, which suggests a strong interaction of methanol with the MOF metal center .
Water dissociation only proceeds at 180 °C, as evidenced by the appearance of the 970 cm−1 peak. At that point, the peak at 1003 cm−1 weakens, indicating that C–O dissociation occurs. The synergistic dissociation of methanol and D2O results in a faster D2O dissociation reaction. In fact, the increased dissociation with MeOH pressure described earlier suggests that the presence of methanol fosters the reaction, mediated by the metal site, almost doubling the reaction yield in 8 Torr MeOH. This observation can be rationalized with first-principles calculations as detailed next.
3.2. Theoretical Analysis of the Water Dissociation Reaction Rate
Calculations summarized in Table 2 show that, when both water and methanol are present in the system, methanol associates to the metal center and water binds to the organic linker. Water alone can associate with both the metal linker and organic linker almost equally, but the MeOH has a stronger binding (association) to the metal center than water. In fact, MeOH does not interact with the oxygen atom of the linker, in contrast to water. Consequently, the OH group of MeOH binds to the metal center upon cooperative adsorption of MeOH and D2O, while the D atom from water binds to the oxygen atom of the linker since the metal site is occupied, and the OD part of D2O binds to the CH3 group of MeOH, as illustrated in Figure 3a. The dissociation reaction then proceeds as follows: First, there is a lengthening and thus weakening of the C–O bond in MeOH, which is in agreement with the observed redshift of ν(C–O) from 1033 cm−1 to 1003 cm−1 and facilitates subsequent OH exchange. Indeed, at the transition state, the OD group from the neighboring D2O molecule approaches the methyl group, while the other D atom bonds to the oxygen of the ligand. As OH (from methanol) bonds to the Zn center very strongly (see Table 2) and D (from D2O) binds strongly to O (see Table 2), the OD group is transferred to the methyl group to complete the hydroxyl exchange on the methyl group. While such an exchange is not common, the strong energetic gain drives this reaction. As MeOD is formed, a single D remains at the oxygen of the linker, characterized by the well-studied 970 cm−1 IR fingerprint . Note that in this cooperative reaction the energy barrier for dissociation is also reduced by a clustering effect . The products of the dissociation reaction (OH on the Zn2+ site or D on oxygen site) bind much more strongly than the initial reactants (H
«Совершенный квадрат»? - переспросил Джабба. - Ну и что с. Спустя несколько секунд Соши преобразовала на экране, казалось бы, произвольно набранные буквы.
Теперь они выстроились в восемь рядов по восемь в каждом. Джабба посмотрел на экран и в отчаянии всплеснул руками.