Skip to main content
NCBI home page
ACS Omega logoLink to ACS Omega
. 2024 Oct 15;9(43):43469–43476. doi: 10.1021/acsomega.4c04872

Hydrolysis of Dimethyl Phosphite by Zr- and Hf-UiO-66

Brandon T Yost †, Addison Wilson †, Bradley Gibbons ‡, Muhammad Kasule †, Yue Wu †, Amanda J Morris ‡, L E McNeil †,*
PMCID: PMC11525503  PMID: 39494028

Abstract

graphic file with name ao4c04872_0007.jpg

Metal–organic frameworks have been utilized as heterogeneous catalysts for the degradation of chemical warfare agents, typically organophosphorous nerve agents. Vibrational spectroscopy techniques coupled with nuclear magnetic resonance (NMR) were utilized to study the adsorption and degradation of dimethyl phosphite (DMP), a simulant molecule of the organophosphorus nerve agent Soman (GD), by Zr- and Hf-UiO-66 as a function of particle size, defect type, and defect density. Defective Zr- and Hf-UiO-66 have been synthesized via a modulated synthesis protocol to engineer missing linker and missing cluster defects into the crystal structure. The adsorption of DMP to UiO-66 was observed to be surface-limited, suggesting maximal DMP adsorption occurs with maximized external surface area. In addition, Hf-UiO-66 samples engineered with large quantities of missing cluster defects are observed to more efficiently hydrolyze DMP into phosphonic acid when compared to less-defective samples. The increase in reactivity is attributed to the greater accessibility of the internal particle volume and, thus, access to a higher number of Lewis-acidic open metal sites, facilitated by missing cluster defects. Taken together, these two observations indicate that to create maximally efficient MOF catalysts for chemical warfare agent degradation one must obtain frameworks with large surface area to maximize adsorption of the simulant as well as a large accessible free volume obtained through the introduction of missing cluster defects to maximize the degradation of the simulant at the Lewis acid sites found in the interior of the framework.

Introduction

Metal–organic frameworks (MOFs) are a class of porous coordination networks that have recently been studied for applications in gas storage1,2 and sensing devices,3 drug delivery systems,4,5 wastewater treatment,6−8 and catalysis.9−13 This is due primarily to their high porosity, internal surface area, stability, and the tunability of their crystal structures.14,15 Recently, more interest has emerged in MOF-based catalysis, particularly for the degradation of organophosophorous chemical warfare agents (CWAs) and their simulants. Specifically, members of the ZrIV-family of MOFs (UiO-66, UiO-67, MOF-808, and NU-1000) have been studied for application in CWA degradation.16−19 Understanding the mechanism of adsorption and degradation of CWAs will allow MOF design to be tailored specifically to maximize the catalytic capacity of a MOF to degrade CWAs.

MOF-based hydrolysis of CWAs and their simulants is driven by the Lewis-acidic open metal sites within a MOF.9,20 The high porosity, large surface area, and large internal free volume of MOFs allow a chemical warfare agent to penetrate the framework and interact with the Lewis acid sites in the interior. However, decomposition of CWAs by UiO-66 is notoriously surface-limited due to the small (0.6 nm) pores created by the short 1,4-benzenedicarboxylic acid (BDC) linker molecules. Despite this, defect engineering has been utilized to increase the pore apertures through the removal of linker molecules and/or metal secondary building units (SBUs). Rather than increased defect density driving enhanced catalytic capabilities, recent work by Gibbons et al.9 suggests a complex interplay between defect density and particle size together determines the catalytic capacity in UiO-66.

In this project, pristine and defective Hf- and Zr-UiO-66 samples were synthesized to have varying quantities of missing linker and missing cluster defects and were exposed to dimethyl phosphite (DMP), a simulant for the CWA Soman or GD, in order to obtain vibrational spectroscopic signatures of the hydrolysis reaction. The various samples and their defect states are discussed in our previous work.21 Based on our vibrational spectroscopy and nuclear magnetic resonance (NMR) measurements, we conclude that DMP is hydrolyzed into phosphonic acid by the Lewis acid sites inside the framework and is subsequently left strongly bound to the framework. The reaction pathway is shown in eqs 1 and 2. Spectroscopically, we see that the adsorption of DMP to the framework is surface limited: decreasing particle size (and therefore increasing surface area) increases the amount of DMP adsorbed to the framework. However, as the hydrolysis of DMP into phosphonic occurs at the Lewis acidic metal sites, which lie primarily in the interior of the framework, increasing the density of missing cluster defects increases the measured quantity of phosphonic acid, suggesting an increased catalytic activity. Together, a combination of high surface area coupled with a large quantity of missing cluster defects leads to the most catalytically active framework.

graphic file with name ao4c04872_m001.jpg 1
graphic file with name ao4c04872_m002.jpg 2

Experimental Section

All Zr- and Hf-UiO-66 sample preparation is discussed in our previous work21 and follows modified synthetic protocols developed by Shearer et al. To summarize, pristine Zr-UiO-66 was synthesized following the procedure described by Shearer et al.22 Defective Zr-UiO-66 was synthesized using a modulated synthesis protocol utilizing formic acid or acetic acid as a modulator with various molar equivalents of modulator relative to Zr. For clarity, all samples will be referred to herein by their modulator and molar equivalents: FA100, FA200, FA300, and AA300. Here, FA refers to formic acid modulator, AA refers to acetic acid modulator, and X00 refers to the molar equivalent of the respective modulator acid used in synthesis relative to Zr. The defect density of each sample was determined by thermogravimetric analysis (TGA) and the defect densities and particle sizes for all samples studied herein have been collated from ref (9). and are shown in Table 1 for convenience. Similarly, pristine Hf-UiO-66 was synthesized following the same procedure as that of pristine Zr-UiO-66 with ZrCl4 replaced by HfCl4. Defective Hf-UiO-66 was synthesized following a similar modulated synthesis approach as for defective Zr-UiO-66, however formic acid was the only modulator acid utilized. As discussed in our previous work,21 defective Hf-UiO-66 tends to preferentially form missing cluster defect sites evidenced by the emergence of two low-angle reflections in powder X-ray diffraction patterns, indicating the presence of (100) and (110) Bragg plane reflections of the reo topology created by missing cluster defect sites.23 Herein, all defective Hf-UiO-66 samples will be referred to by their approximate crystal volume in the reo topology based on the relative intensity of the low-angle reflections in their PXRD patterns: 5% reo, 10% reo, 17% reo, and 32% reo. Importantly, all samples studied herein were from the same synthesis batch of samples discussed in,9 so crystallographic data will not be duplicated herein as they are discussed in detail in the previous work.

Table 1. Defect Density and Particle Size Data for Zr-UiO-66 Samples Studieda.

Sample Name Defect Density (%) Approximate Particle Size (nm)
AA300 14 600
FA100 22 100
FA200 21 600
FA300b 24 650
Open in a new tab
a

These data have been collated from Ref (9).

b

Dominated by missing cluster defects.

Regarding defectivity of the UiO-66 frameworks, missing cluster defects are paired with missing linker defects. However, they are classed separately (herein and elsewhere) due to the ordering of those defects on a larger scale than the random distribution of missing linker defects. That is, powder X-ray diffraction (PXRD) of the samples with large quantities of missing cluster defects (independent of the quantity of linkers removed alongside the removed clusters) show reflections related to the formation of nanoregions of altered topology relative to the standard face-centered cubic topology of UiO-66 in its pristine form. However, defects are also typically quantified by thermogravimetric analysis (TGA) whereby an assumption baked into the mathematical framework of the technique is that there are no missing metal centers: the defect density is calculated assuming the final product after exposure to high temperatures consists on the same number of metal centers present at the outset. When comparing samples such as FA100 and FA300, it is often more concise to refer to FA300 as “more” defective than FA100 based on the propensity for the difference in their defectivity to yield increased internal surface areas in FA300. This is not simply relating particle size to surface area or defect level, but rather utilizing results previously published by our collaborators. In their previous work studying these samples (from the same synthesis batch) it is shown that both FA200 and FA300 contain missing cluster defects, but FA300 is approximately 3% more defective than FA200 by TGA and 5% more defective than FA200 when studied with 1H NMR. Similarly, synthesis with 200 mol equiv of a formic acid modulator (FA200) was shown to produce an average particle size within error of that produced by 300 mol equiv of formic acid modulator (FA300) over several batches. It is also shown in their previous work (see Figure 5 of ref (9).) that FA300 has slightly larger average pore sizes measured via N2 isotherms. Therefore, we conclude that the sample with the higher modulator concentration (FA300) has more surface area than FA200 and contains more missing cluster defects than FA200.9

To study the uptake and hydrolysis of a chemical warfare agent simulant, each sample was exposed to dimethyl phosphite (CH3O)2POH (DMP), a simulant for the chemical warfare agent known as GD or Soman. For each Zr-UiO-66 sample, 20 mg of powder was placed into a glass vial and heated under vacuum (approximately 30 mTorr) at 150 °C for 24 h to drive off any excess moisture present within the framework. After 24 h, the oven was flooded with nitrogen gas. Liquid DMP was placed into a reservoir at the bottom of a plastic vial and similarly held under vacuum for 30 min at room temperature to remove any air and moisture mixed in with the as-purchased DMP. After each of the powder samples was dried, the sample vial was placed inside the DMP vial and sealed. The two vials were held in a fume hood for 72 h at room temperature to ensure saturation loading of DMP into Zr-UiO-66. After 72 h of exposure, each powder sample was placed onto a glass slide and Raman spectra were acquired immediately to determine the vibrational signature of DMP adsorbed on Zr-UiO-66 with minimal hydrolysis due to the lack of physisorbed water. Rehydrated Zr-UiO-66 samples were created by leaving the powder on the glass slide exposed to ambient air for 1 week (relative humidity ∼30–40%), permitted by the high stability of UiO-66 in ambient conditions. This procedure was repeated for Hf-UiO-66, but no heating was involved in the drying process prior to the DMP exposure and rehydration took place over the span of 2 weeks. This is not to say that the Hf-UiO-66 samples were not activated: activation took place immediately upon fabrication, the Hf-UiO-66 samples were simply not reheated after transportation in a desiccant-loaded vial to a new laboratory.

All Raman spectra were collected on a Dilor Triple XY Raman system coupled to an optical microscope. A 473 nm excitation laser source at 50 mW power was used to illuminate the powder samples on a glass slide open to air on the microscope stage. Scattered light is collected by a 100× microscope objective and dispersed across a liquid-nitrogen-cooled CCD detector after encountering a series of three gratings. Counts on the CCD are plotted against the Raman shift (difference in frequency between the incident and scattered light). The frequency axis is calibrated utilizing the well-known Raman-active mode of Si at 520.5 cm–1. The intensity of each spectrum is then internally normalized to the aromatic C–H vibrational mode found at 865 cm–1.

Results and Discussion

Raman Spectroscopy

Raman spectra were first obtained for all samples prior to exposure to DMP to provide a baseline. A detailed description of the Raman signatures for both Zr- and Hf-UiO-66 is given in our previously published work21 and Zr-UiO-66 spectra are shown in Figure 1. To summarize, the most intense Raman-active modes are dominated by motion confined to the linker molecule and unsurprisingly remain unchanged between samples due the consistent benzene dicarboxylic acid linker present in all samples. However, spectroscopic signatures of defects were found in the low-wavenumber “fingerprint” region (200 cm–1 - 400 cm–1), namely a slight red-shift and concurrent quench of the stretching mode of the Zr–O coordination bond (similarly for the Hf–O coordination bond in Hf-UiO-66 samples). The observed red-shift was attributed to defect-induced rigidity changes of the metal–oxygen coordination bond.21 In Figure 1, the slightly negative dip near 1100 cm–1 is an artifact of the background subtraction algorithm caused by the tail of a peak centered near 1150 cm–1 that is cut off by the 2048-data-point-wide window. However, this has no undesirable effects on the rest of the spectra.

Figure 1.

Figure 1

Open in a new tab

Zr-UiO-66 Raman spectra prior to exposure to DMP. Here, near-perfect agreement among the spectra of the different samples is seen regardless of defect densities. Slight differences in the low-wavenumber region are seen when comparing AA300 to the rest of the samples, but the reason for the discrepancy is likely due to defects as discussed in ref (21).

To investigate the adsorption of DMP to the MOF samples, samples were exposed to the DMP and transferred to a glass slide one at a time and Raman spectra were immediately acquired to minimize the amount of time samples spent in air following DMP exposure. As DMP hydrolysis requires water, care was taken to minimize water exposure and isolate the spectroscopic fingerprint of DMP adsorption in the absence of reactivity. While the samples are exposed to humid air during sample transfer and scanning (duration approximately 20 min), it is not expected that water adsorption would be significant over this period. In each sample, a new vibrational mode was seen to emerge at approximately 765 cm–1 while all modes attributed to the framework itself remain unchanged. These data are highlighted in Figure 2.

Figure 2.

Figure 2

Open in a new tab

(a) Raman spectra of DMP-exposed Zr-UiO-66 pristine and defective samples. The vertical dashed line marks the central frequency of the P–O stretches (symmetric and antisymmetric) of DMP. (b) P–O stretching modes of DMP centered at 765 cm–1 consisting of two closely spaced vibrational modes (symmetric and antisymmetric stretch).

Gas-phase density functional theory (DFT) calculations were performed in the Gaussian software package24 in order to assign the measured vibrational modes to specific atomic motion. The structure for DMP was optimized while simultaneously computing vibration modes at the hybrid DFT level utilizing the B3LYP functional and the split-valence triple-ζ basis set including extra d-type polarization orbitals for heavy (non-hydrogen) atoms as well as p-type polarization orbitals for hydrogen.25,26 The B3LYP/6-311G(d,p) level of theory for DMP suggests that the observed broad feature centered at 765 cm–1 is in fact two unresolved vibrational modes: one centered at approximately 763 cm–1 with a broad shoulder centered at approximately 774 cm–1. These modes represent symmetric and antisymmetric P–O stretches of the DMP molecule, respectively. Since each sample was exposed for the same length of time, it is likely that the intensities of these vibrational modes correlate with the quantity of adsorbed DMP. Given that each sample was exposed to DMP for 3 days and believed to be fully saturated, the intensities of the DMP vibrational modes are hypothesized to correlate with the quantity of DMP adsorbed. Thus, the relative intensities can provide insight into the role of particle size and defect density on adsorption capacity. The results support the conclusion that particle size plays a major role in the adsorption of DMP. The trend in peak intensity and peak area for the 765 cm–1 peak is generally consistent with increased external surface area, i.e., the FA100 sample with an average particle size and total surface area per particle of 100 nm and 2168 cm2, respectively, displays larger peak intensity and peak area when compared to the FA300 with an average particle size and external surface area of 600 nm and approximately 514 cm2, respectively. Importantly, the total surface area for FA300 is not given in ref (9)., but is given for FA200. However, it is stated in ref (9). that this calculated total surface area is inversely proportional to particle size and FA 200 and FA 300 have nearly identical particle size. Therefore, the particle size for FA200 will be used as an estimated value for that of FA300 here. This is contrasted against defect density, which must play a minor role, considering FA300 is more defective and dominated by missing cluster defects when compared to the missing linker defect level of FA100.9

To explore the reactivity of DMP, the saturated samples were rehydrated in ambient air for 1 week and again characterized with Raman spectroscopy. The Raman spectra show a clear quenching of the 765 cm–1 P - O stretching vibrational modes of DMP. Along with this quenching, a new Raman-active mode emerges centered at 830 cm–1 (Figure 3). As this mode only appears in the DMP-exposed and hydrated samples and at the expense of the P - O stretch of DMP, we hypothesize that this peak is a signature of phosphonic acid from hydrolyzed DMP rather than the DMP chemically bonding anywhere within the framework. In pristine Zr-UiO-66, a framework known to have catalytic activity limited to the surface, the P - O stretch of DMP continues to be observed upon rehydration, though it is very weak relative to its intensity immediately after exposure. This is likely due to slow evaporation of physisorbed DMP that does not coordinate any Lewis-acidic defect sites on the surface of the MOF. The P - O single bond stretch of the phosphonic acid is very weak in pristine Zr-UiO-66 as well and is seen as a broadening of the framework mode centered at 816 cm–1 along with a shoulder forming near 830 cm–1. This indicates either the presence of limited Lewis acid sites on the surface of the crystal, or that over time some DMP is able to penetrate into the pores; it is likely that both are true. The two possibilities cannot be disentangled by these measurements as the incident laser light probes both the surface and interior of the powder samples. In contrast, the defective samples show a nearly complete quenching of the P - O stretching mode of DMP in concert with the emergence and growth of the P - O stretch of phosphonic acid (Figure 3 (b)). This suggests that the defective samples will more readily hydrolyze DMP into phosphonic acid and the complete quenching of the P - O stretch of DMP implies that nearly all of the adsorbed DMP has been hydrolyzed.

Figure 3.

Figure 3

Open in a new tab

(a) Raman spectra of DMP-exposed and rehydrated Zr-UiO-66 defect series. The asterisk (*) represents the P–O stretch of DMP that is quenched with rehydration in defective samples and still weakly present in the pristine Zr-UiO-66. (b) Detail of the P–O stretch of phosphonic acid centered at 830 cm–1. The peak between the solid lines represents the coalescence of two peaks: symmetric and antisymmetric P–O stretches of DMP. The broad peak between the dotted lines represents the P–O stretch of phosphonic acid. The frequency shift observed between FA200 and FA300 as compared to FA100 and AA300 is likely due to the nature and location of the DMP adsorption: increased access to internal Lewis acid sites in FA200 and FA300 likely allows for a more strongly bound product.

In contrast to DMP adsorption, DMP reactivity appears to correlate to the level of missing cluster defects. While FA100 with a higher external surface area adsorbed more DMP, the intensity and areas of the peaks associated with the phosphonic acid hydrolysis product were lower in comparison to the larger-particle FA300. Given that FA300 contained more missing clusters defects than FA100,21 it is therefore likely that the large accessible internal volume of FA300 relative to FA100 as a consequence of missing cluster defects allows for increased access to Lewis acid sites on the interior of the defective framework.

More evidence of hydroylzed DMP is seen at higher frequencies as is highlighted in Figure 4. A new feature centered around 1125 cm–1 appears after rehydration and is a function of defect density and particle size. The higher-frequency vibrational mode in Figure 4 remains the same across all samples and is a MOF-specific vibrational mode that is independent of DMP exposure. As shown in eqs 1 and 2, methanol is a secondary product of the breakdown of DMP in the presence of water as the decomposition occurs through the removal of two methyl groups. Anhydrous methanol is known to have a Raman-active mode at 1120 cm–1 attributed to the rocking of the CH3.27 However, the peaks observed slightly above 1120 cm–1 suggest that this mode is not free methanol, but rather methanol hydrogen-bound to water. When methanol and water are mixed in a roughly 1:1 ratio by volume, the CH3 rocking mode shifts to 1125 cm–1 due to the formation of strong hydrogen bonds.27 It is important to note that in Figure 4, the presence of the new vibrational band is all that can be ascertained fairly. While FA200 appears more-active than FA300 based on the methanol–water feature at 1125 cm–1, these data were taken multiple weeks apart and exposed to ambient moisture in the meantime. The reason for this is 2-fold: first, focus was placed on the low-wavenumber fingerprint region to compare the uptake and hydrolysis as the hydrolysis was hypothesized to take place at undercoordinated metal centers. The hardware utilized for the Raman measurements allows narrow focus on a small 2048-pixel-wide window, so immediate measurements did not reveal the methanol–water feature at 1125 cm–1 as this frequency was just outside of the measured window. The measurement window was expanded only after appropriate data were acquired all on Zr- and Hf-UiO-66 samples, and measurements took place in ambient air. Therefore, subsequent measurements of the methanol vibrational mode were on samples of slightly varied hydration, so no attempt is made here at quantifying methanol production. The presence of the vibrational signature of this methanol–water pair supports the claim that the frameworks are catalytically active and that the vibrational mode at 830 cm–1 is properly assigned to the P - O stretch of phosphonic acid. Attempts to remove phosphonic acid from the framework following the hydrolysis reaction were unsuccessful (Figure S1).

Figure 4.

Figure 4

Open in a new tab

Emergence of a new feature upon rehydration centered around 1125 cm–1 attributed to methanol hydrogen-bonded to water. The decomposition of DMP into phosphonic acid occurs through the removal of methyl groups in the presence of water, so this new feature corroborates the catalytic activity of the framework.

Similar results are obtained for Hf-UiO-66. Hf-UiO-66 Raman data prior to exposure to DMP are shown in Figure 5(a). Slight differences in the spectra are caused by missing cluster defects present in the Hf-UiO-66 samples and are discussed in our previous work.21 Following DMP exposure and rehydration (Figure 5(b)), the vibrational mode that was assigned to phosphonic acid in the Zr-UiO-66 samples appears at 830 cm–1. It is important to note that the DMP-exposed sample does indicate some reaction with MOF-adsorbed water, as the samples were not as rigorously dried as the Zr-UiO-66 samples. Nuclear magnetic resonance (NMR) data also suggest DMP uptake has a strong dependence on defect density. These data are discussed further in the SI. In Zr-UiO-66, the intensity of the phosphonic acid P - O vibrational mode was significantly lower than 1.0 on the normalized scale (relative to the 865 cm–1 linker vibrational mode). However, in Hf-UiO-66, the more defective samples (10% reo and 30% reo following the naming convention discussed in the Experimental section) show the phosphonic acid mode with intensities greater than 1.0. This suggests that the Hf-UiO-66 samples are able to hydrolyze more DMP than are the Zr-UiO-66 samples. This is unsurprising as the defective Hf-UiO-66 samples are dominated by missing cluster defects whereas Zr-UiO-66 samples are generally comprise missing linker defects, and therefore the Hf-UiO-66 samples offer a much larger accessible free volume for DMP to adsorb (also seen via nuclear magnetic resonance in Figure S2) and more open metal sites to catalyze the hydrolysis reaction in Hf-UiO-66. Similarly, the emergence of the 1125 cm–1 methanol–water vibrational mode is seen in the Hf-UiO-66 samples and is more intense in the more-defective samples.

Figure 5.

Figure 5

Open in a new tab

(a) Raman spectra of Hf-UiO-66 samples prior to exposure to DMP. (b) Raman spectra of Hf-UiO-66 samples after being exposed to DMP and rehydrated. Each curve is internally normalized to the peak centered at 865 cm–1.

At higher frequencies, more signatures of the hydrolysis reaction emerge in the Raman spectra and support conclusions drawn above. In Figure 6 the spectrum of pristine Hf-UiO-66 prior to exposure is compared to those of pristine and defective Hf-UiO-66 after exposure and rehydration. Similar to the lower-frequency region (Figure 5), the framework modes remain intact and several new modes emerge. As in Figure 5, the features at 830 and 1125 cm–1 in the defective samples represent P–O stretching of the phosphonic acid (based on DFT calculations) and vibrations of the methanol–water bound pairs27 respectively. In addition, new features appear at approximately 1120 cm–1, 1180 and 1300 cm–1. Based on DFT calculations performed on DMP, methyl hydrogen phosphonate, and phosphonic acid, it is hypothesized that each of these are vibrational modes of phosphonic acid: symmetric and antisymmetric O–H/P–H wagging and a P=O stretching, respectively. For free DMP (single molecule gas-phase) there are several weakly active Raman modes calculated between 1100 and 1500 cm–1, but these new features in the measured spectra are unlikely to be due to excess DMP. The calculated Raman activities of the vibrational modes between 1100 and 1500 cm–1 are approximately 20 times weaker than that of the 760 cm–1 mode, which is completely quenched in the defective samples. Therefore, it is unlikely that these strong modes at 1180 and 1300 cm–1 represent DMP. This is also true for the vibrational modes calculated within this frequency range for methyl hydrogen phosphonate. Coupled with its apparent instability and weak Raman activity, it is unlikely that these modes represent motion of methyl hydrogen phosphonate. However, precisely three modes are calculated within this region for phosphonic acid, at approximately 1020 cm–1, 1080 and 1180 cm–1. While there is not perfect agreement, one must first consider that these calculations are done on a single molecule of phosphonic acid in the gas phase whereas within the UiO-66 framework the phosphonic acid is hypothesized to be tightly bound to the Lewis acid sites. For this reason, the frequencies of the measured vibrational modes are not expected to match those of the calculated modes precisely, but rather are expected to be red- or blue-shifted from the calculated frequencies. Therefore, more-detailed calculations would need to be performed to better represent the binding phenomenon. Phosphonic acid is the only molecule introduced to the UiO-66 system with Raman-active modes in this frequency region. Infrared absorption spectroscopy, including diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) techniques has been shown to be a powerful tool in the characterization of guest molecule introduction into the UiO-66 framework. For example, Grissom and co-workers utilized time-resolved DRIFTS to determine CO2 binding processes within the pores of UiO-66. During CO2 exposure, a new DRIFTS peak was observed at 3641 cm–1 that was attributed to CO2 hydrogen bonding to the μ3–OH groups of UiO-66.28 Similarly, Yousuf et al. utilized infrared spectroscopy and DRIFTS to characterize various Cu@UiO-66 frameworks through the use of a CO probe molecule, determining that both the Cu(I) and Cu(II) oxidation states participated in the oxidation reaction.29 However, in the case of DMP hydrolysis catalyzed by UiO-66, the molecules produced in each stage of the hydrolysis reaction (that is, the methyl hydrogen phosphonate and phosphonic acid) are all strongly IR-active and so have correspondingly intense features in the infrared absorption spectra. Therefore, the spectroscopic signature of a mixture of all three of these species are visible in the FTIR spectra, even after rehydration (Figure S3).

Figure 6.

Figure 6

Open in a new tab

Raman signatures of the high-frequency region for Hf-UiO-66 samples. A vertical shift of the samples exposed to DMP is applied for clarity and does not signify true intensity changes. Each curve was background subtracted prior to addition of the vertical shift. The framework-specific vibrational modes remain intact, as is seen by comparing each DMP-exposed sample’s spectra to the pristine Hf-UiO-66 spectrum (shown in magenta). However, new peaks emerge that are hypothesized to be O–H and P–H wagging of phosphonic acid (1120 and 1180 cm–1 respectively) as well as a P=O stretching mode at 1300 cm–1.

Conclusion

In conclusion, Raman spectroscopy is seen to be extremely sensitive to the uptake and hydrolysis of DMP by UiO-66. In Zr-UiO-66 there is a complex interplay between defect density and particle size in their effects on DMP adsorption and hydrolysis. While samples with a small average particle size (FA100) are able to adsorb a large quantity of DMP relative to samples with larger average particle size (e.g., FA300), it appears that this adsorption is limited primarily to the surfaces of the crystals based on the relatively small quantity of the product of the catalysis, phosphonic acid, seen by Raman spectroscopy. Similarly, Hf-UiO-66 samples that are dominated by missing cluster defects show that the quantity of phosphonic acid produced by catalysis in such structures is larger than that of comparable Zr-UiO-66 samples, likely due to the ability of DMP to penetrate into the larger mesopores generated by the removal of metal SBUs. This is corroborated by NMR data that suggest that the more-defective Hf-UiO-66 samples are able to adsorb more DMP as well as generate more bound phosphonic acid. Together, these findings help to determine a set of design rules for the optimization of DMP hydrolysis by UiO-66. Namely, samples with small particle sizes and large quantities of defects (without sacrificing structural stability) allow for significant DMP uptake on the surface coupled with enhanced access to the Lewis acid sites that catalyze the hydrolysis reaction.

Acknowledgments

This work was performed with the support of the Department of Defense Threat Reduction Agency under grants HDTRA11910008 (UNC-CH) and W911NF-19-2-0156 (Virginia Tech).

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.4c04872.

  • Attempted removal of DMP as studied by Raman spectroscopy, nuclear magnetic resonance (NMR), and Fourier transform infrared absorption spectroscopy (PDF)

The authors declare no competing financial interest.

Supplementary Material

ao4c04872_si_001.pdf (1.2MB, pdf)

References

  1. Li B.; Wen H.-M.; Zhou W.; Chen B. Porous Metal - Organic Frameworks for Gas Storage and Separation: What, How, and Why?. J. Phys. Chem. Lett. 2014, 5, 3468–3479. 10.1021/jz501586e. [DOI] [PubMed] [Google Scholar]
  2. Li Y.; Yang R. T. Gas adsorption and storage in metal-organic framework MOF-177. Langmuir 2007, 23 (26), 12937–12944. 10.1021/la702466d. [DOI] [PubMed] [Google Scholar]
  3. Li Y.; Xiao A.-S.; Zou B.; Zhang H.-X.; Yan K.-L.; Lin Y. Advances of metal–organic frameworks for gas sensing. Polyhedron 2018, 154, 83–97. 10.1016/j.poly.2018.07.028. [DOI] [Google Scholar]
  4. Beg S.; Rahman M.; Jain A.; Saini S.; Midoux P.; Pichon C.; Ahmad F. J.; Akhter S. Nanoporous metal organic frameworks as hybrid polymer–metal composites for drug delivery and biomedical applications. Drug Discovery Today 2017, 22 (4), 625–637. 10.1016/j.drudis.2016.10.001. [DOI] [PubMed] [Google Scholar]
  5. Nasrabadi M.; Ghasemzadeh M. A.; Zand Monfared M. R. The preparation and characterization of UiO-66 metal-organic frameworks for the delivery of the drug ciprofloxacin and an evaluation of their antibacterial activities. New J. Chem. 2019, 43 (40), 16033–16040. 10.1039/C9NJ03216A. [DOI] [Google Scholar]
  6. Azhar M. R.; Abid H. R.; Periasamy V.; Sun H.; Tade M. O.; Wang S. Adsorptive removal of antibiotic sulfonamide by UiO-66 and ZIF-67 for wastewater treatment. J. Colloid Interface Sci. 2017, 500, 88–95. 10.1016/j.jcis.2017.04.001. [DOI] [PubMed] [Google Scholar]
  7. Fan J.; Chen D.; Li N.; Xu Q.; Li H.; He J.; Lu J. Adsorption and biodegradation of dye in wastewater with Fe3O4@MIL-100 (Fe) core–shell bio-nanocomposites. Chemosphere 2018, 191, 315–323. 10.1016/j.chemosphere.2017.10.042. [DOI] [PubMed] [Google Scholar]
  8. Zhan X.-Q.; Tsai F.-C.; Xie L.; Zhang K.-D.; Liu H.-L.; Ma N.; Shi D.; Jiang T. Ligands-coordinated Zr-based MOF for wastewater treatment. Nanomaterials 2018, 8 (9), 655. 10.3390/nano8090655. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Gibbons B.; Bartlett E. C.; Cai M.; Yang X.; Johnson E. M.; Morris A. J. Defect Level and Particle Size Effects on the Hydrolysis of a Chemical Warfare Agent Simulant by UiO-66. Inorg. Chem. 2021, 60 (21), 16378–16387. 10.1021/acs.inorgchem.1c02224. [DOI] [PubMed] [Google Scholar]
  10. Kalaj M.; Momeni M. R.; Bentz K. C.; Barcus K. S.; Palomba J. M.; Paesani F.; Cohen S. M. Halogen bonding in UiO-66 frameworks promotes superior chemical warfare agent simulant degradation. Chem. Commun. 2019, 55 (24), 3481–3484. 10.1039/C9CC00642G. [DOI] [PubMed] [Google Scholar]
  11. Wang G.; Sharp C.; Plonka A. M.; Wang Q.; Frenkel A. I.; Guo W.; Hill C.; Smith C.; Kollar J.; Troya D.; Morris J. R. Mechanism and Kinetics for Reaction of the Chemical Warfare Agent Simulant, DMMP(g), with Zirconium(IV) MOFs: An Ultrahigh-Vacuum and DFT Study. J. Phys. Chem. C 2017, 121 (21), 11261–11272. 10.1021/acs.jpcc.7b00070. [DOI] [Google Scholar]
  12. Vellingiri K.; Philip L.; Kim K.-H. Metal–organic frameworks as media for the catalytic degradation of chemical warfare agents. Coord. Chem. Rev. 2017, 353, 159–179. 10.1016/j.ccr.2017.10.010. [DOI] [Google Scholar]
  13. Kalaj M.; Palomba J. M.; Bentz K. C.; Cohen S. M. Multiple functional groups in UiO-66 improve chemical warfare agent simulant degradation. Chem. Commun. 2019, 55 (37), 5367–5370. 10.1039/C9CC02252J. [DOI] [PubMed] [Google Scholar]
  14. Winarta J.; Shan B.; Mcintyre S. M.; Ye L.; Wang C.; Liu J.; Mu B. A Decade of UiO-66 Research: A Historic Review of Dynamic Structure, Synthesis Mechanisms, and Characterization Techniques of an Archetypal Metal-Organic Framework. Cryst. Growth Des. 2020, 20 (2), 1347–1362. 10.1021/acs.cgd.9b00955. [DOI] [Google Scholar]
  15. Gentile F. S.; Pannico M.; Causa M.; Mensitieri G.; Di Palma G.; Scherillo G.; Musto P. Metal defects in HKUST-1 MOF revealed by vibrational spectroscopy: a combined quantum mechanical and experimental study. Journal of Materials Chemistry A 2020, 8 (21), 10796–10812. 10.1039/D0TA01760D. [DOI] [Google Scholar]
  16. Harvey J. A.; McEntee M. L.; Garibay S. J.; Durke E. M.; DeCoste J. B.; Greathouse J. A.; Sava Gallis D. F. Spectroscopically Resolved Binding Sites for the Adsorption of Sarin Gas in a Metal-Organic Framework: Insights beyond Lewis Acidity. J. Phys. Chem. Lett. 2019, 10 (17), 5142–5147. 10.1021/acs.jpclett.9b01867. [DOI] [PubMed] [Google Scholar]
  17. Feng X.; Jena H. S.; Krishnaraj C.; Leus K.; Wang G.; Chen H.; Jia C.; Van Der Voort P. Generating Catalytic Sites in UiO-66 through Defect Engineering. ACS Appl. Mater. Interfaces 2021, 13 (51), 60715–60735. 10.1021/acsami.1c13525. [DOI] [PubMed] [Google Scholar]
  18. Kim M.-K.; Kim S. H.; Park M.; Ryu S. G.; Jung H. Degradation of chemical warfare agents over cotton fabric functionalized with UiO-66-NH 2. RSC Adv. 2018, 8 (72), 41633–41638. 10.1039/C8RA06805D. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Plonka A. M.; Grissom T. G.; Musaev D. G.; Balboa A.; Gordon W. O.; Collins-Wildman D. L.; Ghose S. K.; Tian Y.; Ebrahim A. M.; Mitchell M. B.; Hill C. L.; Morris J. R.; Frenkel A. I. Effect of Carbon Dioxide on the Degradation of Chemical Warfare Agent Simulant in the Presence of Zr Metal Organic Framework MOF-808. Chem. Mater. 2019, 31 (23), 9904–9914. 10.1021/acs.chemmater.9b04565. [DOI] [Google Scholar]
  20. Mondloch J. E.; Katz M. J.; Isley III W. C.; Ghosh P.; Liao P.; Bury W.; Wagner G. W.; Hall M. G.; DeCoste J. B.; Peterson G. W.; Snurr R. Q.; Cramer C. J.; Hupp J. T.; Farha O. K. Destruction of chemical warfare agents using metal-organic frameworks. Nat. Mater. 2015, 14 (5), 512–516. 10.1038/nmat4238. [DOI] [PubMed] [Google Scholar]
  21. Yost B. T.; Gibbons B.; Wilson A.; Morris A. J.; McNeil L. E. Vibrational spectroscopy investigation of defects in Zr- and Hf-UiO-66. RSC Adv. 2022, 12, 22440–22447. 10.1039/D2RA03131K. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Shearer G. C.; Chavan S.; Bordiga S.; Svelle S.; Olsbye U.; Lillerud K. P. Defect Engineering: Tuning the Porosity and Composition of the Metal-Organic Framework UiO-66 via Modulated Synthesis. Chem. Mater. 2016, 28 (11), 3749–3761. 10.1021/acs.chemmater.6b00602. [DOI] [Google Scholar]
  23. Cliffe M. J.; Wan W.; Zou X.; Chater P. A.; Kleppe A. K.; Tucker M. G.; Wilhelm H.; Funnell N. P.; Coudert F.-X.; Goodwin A. L Correlated defect nanoregions in a metal-organic framework. Nat. Commun. 2014, 5 (1), 4176. 10.1038/ncomms5176. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Frisch M. J., Trucks G. W., Schlegel H. B., Scuseria G. E., Robb M. A., Cheeseman J. R., Scalmani G., Barone V., Petersson G. A., Nakatsuji H., Li X., Caricato M., Marenich A. V., Bloino J., Janesko B. G., Gomperts R., Mennucci B., Hratchian H. P., Ortiz J. V., Izmaylov A. F., Sonnenberg J. L., Williams-Young D., Ding F., Lipparini F., Egidi F., Goings J., Peng B., Petrone A., Henderson T., Ranasinghe D., Zakrzewski V. G., Gao J., Rega N., Zheng G., Liang W., Hada M., Ehara M., Toyota K., Fukuda R., Hasegawa J., Ishida M., Nakajima T., Honda Y., Kitao O., Nakai H., Vreven T., Throssell K., Montgomery J. A. Jr., Peralta J. E., Ogliaro F., Bearpark M. J., Heyd J. J., Brothers E. N., Kudin K. N., Staroverov V. N., Keith T. A., Kobayashi R., Normand J., Raghavachari K., Rendell A. P., Burant J. C., Iyengar S. S., Tomasi J., Cossi M., Millam J. M., Klene M., Adamo C., Cammi R., Ochterski J. W., Martin R. L., Morokuma K., Farkas O., Foresman J. B., Fox D. J.. Gaussian16 Revision C.01; Gaussian Inc.: 2016.
  25. Francl M. M.; Pietro W. J.; Hehre W. J.; Binkley J. S.; Gordon M. S.; DeFrees D. J.; Pople J. A. Self-consistent molecular orbital methods. xxiii. a polarization-type basis set for second-row elements. J. Chem. Phys. 1982, 77, 3654–3665. 10.1063/1.444267. [DOI] [Google Scholar]
  26. McLean A. D.; Chandler G. S. Contracted gaussian basis sets for molecular calculations. i. second row atoms, z = 11–18. J. Chem. Phys. 1980, 72, 5639–5648. 10.1063/1.438980. [DOI] [Google Scholar]
  27. Hester R. E.; Plane R. A. Raman spectra of methanol solutions-I+ Saturated solutions of some electrolytes. Spectrochimica Acta Part A: Molecular Spectroscopy 1967, 23 (8), 2289–2296. 10.1016/0584-8539(67)80121-1. [DOI] [Google Scholar]
  28. Grissom T. G.; Driscoll D. M.; Troya D.; Sapienza N. S.; Usov P. M.; Morris A. J.; Morris J. R. Molecular-Level Insight into CO2 Adsorption on the Zirconium-Based Metal-Organic Framework, UiO-66: A Combined Spectroscopic and Computational Approach. J. Phys. Chem. C 2019, 123 (22), 13731–13738. 10.1021/acs.jpcc.9b02513. [DOI] [Google Scholar]
  29. Yousuf M. R.; Johnson E. M.; Maynes A. J.; Johnston C. R.; Karim A. M.; Morris A. J.; Morris J. R.; Troya D. Catalytic CO Oxidation by Cu Single Atoms on the UiO-66 Metal-Organic Framework: The Role of the Oxidation State. J. Phys. Chem. C 2022, 126 (30), 12507–12518. 10.1021/acs.jpcc.2c03463. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ao4c04872_si_001.pdf (1.2MB, pdf)

Articles from ACS Omega are provided here courtesy of American Chemical Society

RESOURCES