Probing Metal−Organic Framework Design for ... - ACS Publications

Jun 26, 2018 - that are more efficient and eco-friendly is imperative for supporting the sharp ... suffer from selectivity issues and mandate expensiv...
0 downloads 0 Views 1MB Size
Subscriber access provided by TUFTS UNIV

Interfaces: Adsorption, Reactions, Films, Forces, Measurement Techniques, Charge Transfer, Electrochemistry, Electrocatalysis, Energy Production and Storage

Probing Metal-Organic Framework Design for Adsorptive Natural Gas Purification Jayraj Joshi, Guanghui Zhu, Jason Lee, Eli A. Carter, Christopher W Jones, Ryan P Lively, and Krista S. Walton Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.8b00889 • Publication Date (Web): 26 Jun 2018 Downloaded from http://pubs.acs.org on June 28, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Probing Metal-Organic Framework Design for Adsorptive Natural Gas Purification Jayraj N. Joshi, Guanghui Zhu, Jason J. Lee, Eli A. Carter, Christopher W. Jones, Ryan P. Lively, Krista S. Walton* School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive NW, Atlanta, Georgia, 30332, United States of America Abstract Parent and amine-functionalized analogues of the metal-organic frameworks (MOFs) UiO-66(Zr), MIL-125(Ti), and MIL-101(Cr) were evaluated for their hydrogen sulfide (H2S) adsorption efficacy and post-exposure acid gas stability. Adsorption experiments were conducted through fixed-bed breakthrough studies utilizing multicomponent 1% H2S/99% CH4 and 1% H2S/10% CO2/89% CH4 natural gas simulant mixtures. Instability of MIL-101(Cr) materials after H2S exposure was discovered through X-ray diffraction and porosity measurements following adsorbent pelletization, while other materials retained their characteristic properties. Linker-based amine functionalities increased H2S breakthrough times and saturation capacities from their parent MOF analogues. Competitive CO2 adsorption effects were mitigated in mesoporous MIL-101(Cr) and MIL-101-NH2(Cr), in comparison to microporous UiO-66(Zr) and MIL-125(Ti) frameworks. This result suggests the installation of H2S binding sites in large pore MOFs could potentially enhance H2S selectivity. In situ FTIR measurements in 10% CO2 and 5000 ppm H2S environments suggest framework hydroxyl and amine moieties serve as H2S physisorption sites. Results from this study elucidate design strategies and stability considerations for engineering MOFs in sour gas purification applications.

1

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Introduction Continuous development of natural gas purification strategies that are more efficient and eco-friendly is imperative for supporting the sharp rise in global natural gas demand.1–3 In natural gas reserves worldwide, hydrogen sulfide (H2S) and carbon dioxide (CO2) are particularly pervasive contaminants. These compounds are not only deleterious for fuel quality, but H2S in particular possesses strong toxicity, corrosion, and poses ecological hazards, making the removal of this compound paramount for natural gas upgrading.4–6 Gas sweetening methods commonly utilize aqueous amine solutions to capture acid gas contaminants, but these processes suffer from selectivity issues and mandate expensive absorbent regeneration schemes and process material costs.7–10 Utilizing the strong, acid gas-selective affinity of amines while selectively partitioning CO2 and H2S is therefore critical for designing next-generation sour gas filtration processes. Metal-organic frameworks (MOFs) exhibit a high degree of tunability that can be advantageous for facilitating adsorption-based processes for natural gas purification. Studies relevant to natural gas separations using MOFs have primarily employed computational methods and/or single-component isothermal adsorption experiments with spectroscopic studies.11–17 These investigations are extremely informative for identifying (1) MOFs that remain stable following acid gas exposure, and (2) understanding H2S adsorption interactions with tested frameworks. However, adsorption equilibrium measurements do not afford practical multicomponent selectivities, particularly for H2S, which exhibits non-ideal adsorbate-adsorbate interactions through hydrogen bonding and sometimes chemisorption.18 This behavior can furthermore change the adsorption behavior of other mixture constituents, as shown previously with several chemisorbing

2

ACS Paragon Plus Environment

Page 2 of 32

Page 3 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

molecules in the presence of water vapor in HKUST-1.19,20 Lack of adsorption data at low H2S partial pressures additionally complicates the comparison of MOF adsorption performance at typical concentrations of H2S

in sour gas mixtures. Consequently,

multicomponent adsorption experiments utilizing realistic H2S concentrations lend more pragmatic separation information for natural gas mixtures. Fixed-bed experiments commonly facilitate these studies. Investigations on rare-earth metal center21 and socMOFs22 evaluated in very sour (5% H2S) mixtures with CO2 and CH4 demonstrated how these less commonly studied frameworks can be designed to facilitate natural gas separations while maintaining adsorbent stability following acid gas exposure. Zou and coworkers23 also reported informative breakthrough experiments using H2S and CO2 containing mixtures to uncover low pressure H2S capture potential for more commonly studied and produced MOFs. They found MOFs containing open-metal sites exhibit the highest uptake capacities and H2S/CO2 selectivities. However, these tests were conducted in an inert carrier gas, which would produce slight selectivity differences in comparison to predominantly hydrocarbon environments. Despite a few key exceptions such as CPO27(Ni)24, MOFs with open metal sites also irreversibly bind sulfur and/or degrade through strong chemisorption reactions, limiting their cyclability.25,26 One simple strategy to enhance MOF adsorption potential while preserving structural integrity and regenerability is ligand functionalization. Modification of these nanoporous materials with a myriad of chemical groups (-NH2, -COOH, -F, -OH, etc.), using either pre-functionalized ligands or post-synthetic modification strategies, has fostered both framework stability and adsorbate selectivity in a variety of adsorption applications.27–34 Extrapolating this idea towards sour gas remediation challenges suggests the integration

3

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of acid gas selective functionalities in MOFs can generate design principles for new adsorbent materials to target H2S capture in the extremely complex multicomponent environments encountered in natural gas. This work aims to build upon the presently limited understanding of low partial pressure H2S removal from natural gas related mixtures using MOFs. Favorable interactions between primary amines and acid gas adsorbates are leveraged here to enhance H2S capture, as we explore the effects of linkerbased amine (-NH2) MOF functionalization on adsorptive sour gas purification. Three well-studied, stable MOFs previously demonstrated to allow functionalization via linkerbased amines35–38 are examined as candidate frameworks: UiO-66(Zr), MIL-101(Cr), MIL-125(Ti). Their respective amine analogues are all constructed using 2aminoterephalic acid as a framework ligand. Results from this study will serve to reveal which framework qualities are desirable for the cyclable and selective removal of H2S in MOF-based adsorbent materials. Experimental Section All chemical reagents were purchased from Sigma-Aldrich and used without further purification. MOFs were activated at 150°C for 24h under vacuum, prior to measurements. Information regarding material characterization and synthesis procedures are present in the Supplementary Information. Adsorbents were pelletized at ~8000 psig using a Carver brand powder press and held for 10s to aggregate powder particles. MOF pellets were then gently sieved through 20x40 mesh (425 µm-850 µm) to attain uniform particle sizes for fixed-bed tests. Sieved particles were added to the fixed-bed column and activated in situ under N2 flow. MOF

4

ACS Paragon Plus Environment

Page 4 of 32

Page 5 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

activation was conducted until desorbed water and entrapped solvent could no longer be observed on the outlet MS detector, with a minimum of three hours activation time. Following adsorption measurements, materials were heated to 150°C and pure nitrogen was flowed through the column until no more H2S could be detected on the outlet gas. The post-exposure MOFs were then recovered for stability analyses. Hydrogen sulfide exposure times for the materials range from 1-2.5h during the adsorption experiments. Fixed-bed system specifications and procedures are provided in more detail in the Supplementary Information. Fixed-bed adsorption experiments were performed using a custom-built breakthrough system. 100-200 mg of activated samples were tested in a fritted tube. Prior to adsorption, the challenge gas was purged to ensure equilibration of the gas mixture. 1% H2S/CH4 and 1%H2S/10%CO2/CH4 gas mixtures mentioned in the text specifically mean 1%H2S/99%CH4 and 1%H2S/10%CO2/89%CH4, respectively. These mixtures were chosen to (1) include common components of sour natural gas at realistic concentrations, (2) introduce adsorption competition typical of sour gas streams, and (3) collect saturation capacities that are comparable to existing multicomponent natural gas adsorption studies utilizing similar gas ratios.10,23 Desorption measurements were conducted by purging with pure He until no more H2S, CO2, or CH4 could be detected on the outlet. All breakthrough curves are reported here with dead-time subtracted for each adsorbate species. Dead-times were ascertained using non-adsorbing glass-beads to analogous packing heights at the same flow conditions. To ensure safe testing, the entire fixed-bed adsorption apparatus and corresponding outlet detector were housed in a wellventilated chemical hood. Additionally, these experiments were performed in an acid gas

5

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

specific testing facility at the Georgia Institute of Technology. This laboratory is housed with multiple ambient, real-time gas sensors, personal H2S sensors, and self-contained ventilated mixture housing to create a safe working environment for handling hazardous acid gases. Please refer to the Supplementary Information for specific breakthrough experiment parameters. Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) experiments were performed in situ on a FTIR spectrometer (Thermo, Nicolet iS50) equipped with a liquid nitrogen cooled MCT/A detector, a diffuse reflectance accessory (Praying Mantis, Harrick), and a high temperature reaction chamber (HVC, Harrick). The chamber used in H2S adsorption experiments were coated with SilcoNert. See SI for more details on sample activation and FTIR data collection. Results and Discussion Acid gas stability of the candidate frameworks was probed to ascertain the efficacy of these materials as reusable adsorption media. Diffraction patterns for pristine materials before and after H2S exposure are presented to elucidate possible structural degradation in Figure 1a. It should be stated that samples were exposed to the ambient air following H2S exposure, which could have affected their subsequent characterization (e.g., atmospheric water vapor exposure). There is no observable change in crystallinity for any powder samples after exposure. Figures S1-S2 and Table S2 show pelletization and subsequent H2S exposure does not affect the diffraction patterns or accessible surface areas of UiO-66(Zr), UiO-66-NH2(Zr), MIL-125(Ti), or MIL-125-NH2(Ti). However, sample pelletization did change the diffraction pattern for MIL-101(Cr). Figure 1b shows

6

ACS Paragon Plus Environment

Page 6 of 32

Page 7 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

micro-strain of crystalline features corresponding to peaks between 2θ = 2-5° for MIL101(Cr) and MIL-101-NH2(Cr) after pelletization. These low-angle features are furthermore completely absent in the pattern for MIL-101(Cr) after H2S exposure. Decreased N2 uptake at 77K after pelletization evidenced in Figures S5 and S6 support this observation. Table S2 additionally shows that BET surface areas for MIL-101(Cr) materials significantly decrease after pelletization. Qualitative changes after pelletization are also apparent in the pore size distributions for MIL-101(Cr)-type MOFs in Figures S8 and S9. A relatively larger porosity loss is experienced for these pelletized samples after acid gas exposure in Figure S5, in comparison to their pristine counterparts. Pelletized MIL-101(Cr) exhibits a 10% loss in BET surface area, greater than that seen with MIL101-NH2(Cr). These results show the pelletization process partially destroys the MIL101(Cr) framework, and in-turn compromise the characteristic acid gas stability of the material. Textural property changes in HKUST-1 reported by Peterson et al.39 within the same pelletization pressure range (1000 – 10,000 psi) as in this study (~8000 psi) demonstrate the commonality of material changes through this process on certain MOFs. Changes in structure-property relationships in other hierarchically porous structured adsorbents through similar processing have been examined as well.40,41 Walton and Sholl42 recently remarked on the importance of elucidating the effects of industrially relevant processing procedures on materials discussed in academic work, and pelletization is often required for powder adsorbents in packed-bed columns.43 Subsequently, this observation recognizes how the H2S adsorption potential of MIL101(Cr) may be deleteriously affected by probable pre-exposure processing steps. Decreased adsorption potential is most directly obvious from the measured sharp

7

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 32

decrease in accessible pore space for H2S adsorption; this reduces the adsorption capacity expected from the pristine framework. Additionally, observed structural changes may lead to further chemically-induced degradation in the presence of other possible natural gas species, such as water vapor.

(a)

(b)

Figure 1. (a) Powder diffraction patterns of pristine and 1% H2S exposed samples, and (b) powder diffraction patterns of MIL-101 (Cr) and MIL-101-NH2 (Cr) before and after pelletization and subsequent 1% H2S exposure. Exposures conducted at ambient conditions. BDC and BDC-NH2 linkers denote parent and amine functionalized MOFs, respectively

Fixed-bed adsorption experiments for the candidate frameworks were first examined using a 1% H2S/CH4 challenge gas. Desorption was performed by purging with helium at ambient temperature. Adsorption-desorption cycles for each MOF are presented in Figures S10-S15. All materials exhibit a decrease in uptake capacity following the first cycle. This is likely due to hydrogen sulfide retention between cycles. As previously discussed, surface areas and crystalline features of stable materials (UiO-66 and MIL-125) are unchanged after heated activation and vacuum. Subsequently, the initial adsorption

8

ACS Paragon Plus Environment

Page 9 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

capacity losses observed are not indicative of degradation. To further supplement this, observations reported in Table S2 and Figures S3-S4 and S6 show that degassing MOFs which are unaffected by pelletization at their activation temperatures leads to a recovery of their porosity after H2S exposure, showing their potential for cyclability when both temperature swing and vacuum desorption are utilized. Measured H2S adsorption capacities for UiO-66(Zr), UiO-66-NH2(Zr), and MIL-101(Cr) are all lower than previously reported from fixed-bed experiments at the same partial pressure of H2S by Zou and coworkers,23 albeit their experiments used balance He instead of CH4.23 Methane is largely non-adsorbing at room temperature with the selected MOFs, even at its high partial pressure relative to hydrogen sulfide.15,16,32 Considering this, a comparison of H2S uptake was performed for using methane and inert carrier gases at the same inlet H2S concentration (1%). Using UiO-66-NH2(Zr) as an example, the data in Figure S16 evidences a clear decrease in H2S adsorption when using CH4 as a carrier gas as compared to N2. These observations highlight the importance of evaluating sour gas adsorbents in the presence of even weakly adsorbing natural gas constituents to gauge competitive adsorption effects. The saturation capacity of UiO-66-NH2(Zr) here in 1% H2S/N2 (0.576 mmol H2S g-1) is still less than the 0.909 mmol H2S g-1 capacity reported by Zou and coworkers23 for a similar mixture. However, it should be noted that the previous study utilized an acid modulator in their synthesis of UiO-66-NH2(Zr), whereas a non-modulated synthesis method was utilized in this study. A comparison of the N2 adsorption isotherms at 77K reported by Zou et al. with the results reported here (Figure S4) reveals a higher accessible microporosity for the MOF in the previous report. Modulated UiO-66(Zr) materials are well known to exhibit enhanced internal pore

9

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

volumes and surface areas through induced defects.44–47 So, the lower H2S capacity for UiO-66-NH2(Zr) in the 1% H2S/N2 mixture measured here is likely a result of the lower accessible adsorption surface per mass available for H2S binding, which is characteristic of the textural properties for the non-modulated MOF utilized in this study. Acknowledging these differences across studies helps provide some metric for expected deviations in adsorption results across disparate processing conditions for MOFs.48 Integration of linker-based primary amines enhanced H2S uptake in the 1% H2S/CH4 mixture for all three parent frameworks. The degree to which H2S uptake was enhanced differed based on the textural properties of the three frameworks. Breakthrough times for the materials in Figure 2a-c reveal MIL-101-NH2(Cr) alone does not exhibit a significant enhancement in 1% H2S uptake after the addition of primary linker-based amines, at least in comparison to the other MOFs. This result is attributed to enhanced adsorption interactions of H2S with the smaller pore environments of the other two MOFs. Calculated pore sizes from previous studies suggest a ranking of MIL-125(Ti) < UiO66(Zr) < MIL-101(Cr) from smallest to largest pore size—also supported by the pore size distributions in Figure S7.16,49,50 Only MIL-101(Cr) MOFs exhibit mesoporosity in this work. Ligand-based binding sites are consequently scaffolded in smaller pores with UiO66(Zr) and MIL-125(Ti) structures. The larger pore diameters and cage sizes of MIL101(Cr) weaken physisorption interactions between framework –NH2 and –OH functionalities and adsorbed H2S. It should be restated that the previously discussed partial degradation in MIL-101-NH2(Cr) may also reduce improvement in H2S breakthrough time, as seen in a similar adsorption study by Peterson et al. for HKUST-1 and UiO-66(Zr).39

10

ACS Paragon Plus Environment

Page 10 of 32

Page 11 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Dynamic adsorption curves for amine-functionalized samples in the 1% H2S/CH4 mixture are compared in Figure 2d. MIL-125-NH2(Ti) exhibited the best H2S removal performance of the candidate MOFs, with a saturation capacity of 0.559 mmol H2S g-1 in the 99% CH4 mixture. This is also the largest increase (~170%) in H2S saturation capacity from the parent MOF observed in this study. Later discussion on the sensitivity of MIL-125-NH2(Ti) hydroxyl and primary amine sites to adsorbing H2S, investigated through in situ FTIR experiments, elaborates more on this phenomenon.

(a)

(b)

(c)

(d)

Figure 2. 1% H2S/CH4 breakthrough curves at 20°C for parent and -NH2 containing analogues of (a) MIL-125(Ti) (b) UiO-66(Zr) (c) MIL-101(Cr) (d) all –NH2 functionalized materials

11

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Carbon dioxide (CO2) is another nearly-ubiquitous natural gas constituent. Considering this, mixture complexity was increased to test the candidate MOF adsorbents using a 1% H2S/10% CO2/CH4 challenge gas. Breakthrough curves for all MOFs display a roll-up effect for CO2 upon adsorption in the new mixture—delineated in Figures S17-S22. Rollup, especially for heavier components, typically indicates the displacement of a weaker adsorbate by a stronger one in multicomponent, fixed-bed adsorption processes.51 The curves suggest H2S preferentially binds to frameworks over CO2 at the concentrations evaluated in this study. Microporous UiO-66(Zr)-NH2 and MIL-125-NH2(Ti) qualitatively exhibit more pronounced roll-up of CO2 than their parent frameworks, demonstrating stronger interactions of H2S with accessible -NH2 sites over CO2. Amine groups also facilitate greater CO2 uptake at low partial pressures. So, comparatively more adsorbed CO2 is displaced during roll-up in amine-functionalized samples. The opposite phenomenon is interestingly observed for MIL-101(Cr) and MIL-101-NH2(Cr). However, the previously discussed degradation for H2S-exposed MIL-101(Cr) materials reduces pore accessibility and partially destroys the structure—restricting adsorption interactions with–NH2 and –OH binding sites on MIL-101-NH2(Cr). Figure 3 illustrates the tendency of MIL-101-NH2(Cr) to eschew CO2 competitive adsorption effects. Unlike mesoporous MIL-101-NH2(Cr), the other two aminefunctionalized frameworks experience more significant losses in H2S capacity after introduction of CO2. This demonstrates the impact of pore environment on integrated linker-based moieties that act as adsorption sites for target molecules. The smaller pores of UiO-66(Zr) and MIL-125(Ti) facilitate CO2 adsorption at its relatively low partial pressure, enhancing interactions of the adsorbate with framework physisorption sites.

12

ACS Paragon Plus Environment

Page 12 of 32

Page 13 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Breakthrough behavior of the parent frameworks in Figure S23 supports this. The adsorption characteristics of MIL-101(Cr) are largely unchanged with added CO2, demonstrated by the relatively lower reduction in H2S breakthrough time in comparison to the microporous frameworks. Hindrance of stronger CO2 adsorption in the larger pores of MIL-101(Cr) and MIL-101-NH2(Cr) therefore mitigates the more evident adsorption competition experienced with UiO-66(Zr) and MIL-125(Ti) materials. For MOFs containing linker-based amines, the rigidity and regulated partitioning of these functionalities restricts their interactions with molecules like CO2 to weak physisorption (evidenced later in FTIR experiments), as opposed to the stronger adsorption behavior of CO2 observed in materials impregnated with flexible primary amines.52–55 Adsorption interactions are weakened further in MOFs with larger pores, as seen with MIL-101NH2(Cr) in Figure 3. These observations can importantly serve as a design heuristic for integrating gas capture selectivity in MOFs; installation of linker-based functionalities in relatively larger pore environments foster reduction in adsorption competition between target adsorbate species.

Figure 3. H2S breakthrough curves, organized left-to-right by increasing mean cage size, for -NH2 functionalized materials tested at ambient conditions under 1% H2S/CH4 (blue) and 1% H2S/10% CO2/89% CH4 (red) mixtures. Tested at 20°C.

13

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

MIL-125-NH2(Ti) was replaced by MIL-101-NH2(Cr) as the material with the greatest H2S saturation capacity when using the 1% H2S/10% CO2/CH4 mixture. The importance of adding multicomponent complexity is clearly demonstrated by this change in adsorbent ranking. For instance, this dynamic would not be easily discernable from comparing single-component isothermal uptake measurements, or previous fixed-bed adsorption studies utilizing H2S mixtures with inert balances (e.g., He, N2). Saturation capacities for all materials in both 1% H2S/CH4 and 1% H2S/10% CO2/CH4 are provided in Figure 4. The markedly lower accessible BET surface area of UiO-66-NH2(Zr) compared to MIL-125-NH2(Ti) and MIL-101-NH2(Cr), even after pelletization, likely caused this MOF to possess the lowest H2S saturation capacity of the primary amine containing adsorbents in both mixtures. H2S adsorption capacities for MIL-101(Cr) materials in the methane carrier gas here are similar to those reported by Zou and coworkers23 through 1% H2S/10% CO2/89% He breakthrough experiments. Significant methane adsorption competition at < 1 atm partial pressure is unexpected in the mesoporous adsorbent, which explains why the fixed-bed results are comparable to those acquired with an inert carrier gas. Because methane is a major constituent in natural gas, the relative insensitivity of MIL-101(Cr) to methane competition is beneficial for selective H2S removal. In contrast to MIL-101(Cr), the H2S capacities measured here for UiO-66-NH2(Zr) for 1% H2S/CH4 and 1% H2S/10% CO2/CH4 mixtures are significantly lower than the 0.909 mmol g-1 capacity reported by Zou and coworkers.23 As previously explained, this difference is mainly ascribed to differences in synthesis procedures compared with the previous study. Still, the ~50% lower H2S saturation capacity measured in 1% H2S/99% CH4 here suggests competitive binding effects from methane

14

ACS Paragon Plus Environment

Page 14 of 32

Page 15 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

significantly reduce H2S adsorption performance in small-pore UiO-66-NH2(Zr). All H2S saturation capacities exhibit order-of-magnitude agreement with previous calculations and/or equilibrium H2S adsorption experiments (previous data not available for MIL-101NH2) around P/P0 = 0.01.13,15,16,23 These comparisons to literature capacities are also provided in Table S3 in the Supplemental Information.

Figure 4. Bar graph of H2S saturation capacities for studied materials under 1% H2S / 99% CH4 (gold) and 1% H2S / 10% CO2 / 89% CH4 (blue) challenge mixtures, tested at 20°C.

In situ diffuse reflectance infrared spectroscopy (DRIFTS) experiments in 10% CO2 and 5000 ppm H2S environments were conducted to further understand adsorption behavior differences among –NH2 functionalized materials. Spectra for MOFs before and after CO2 adsorption are presented in Figure S24. Physisorbed CO2 is apparent near 2341 cm-1 after subtraction of gaseous CO2.56 Chemisorbed CO2 (1800-1300 cm-1) is not

15

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

evidenced.57 Interestingly, features in the -OH stretch region for MIL-101-NH2(Cr) are greatly varied and low in absorbance intensity following activation, unlike the other two MOFs. The evacuation of capping water molecules from Cr3+ metal sites following degassing has been previously reported.58 As a potential binding site for both CO2 and H2S, the absence of hydroxyl groups could partially explain the relatively small H2S uptake change after CO2 addition shown previously in Figure 3. Noticeable changes in the hydroxyl stretches 3800 cm-1 to 3600 cm-1 for UiO-66-NH2(Zr) and MIL-125-NH2(Ti) in Figures 5a and S31 (expanded hydroxyl and amine stretch regions presented in Figures S25-S30 in the Supplementary Information) support this hypothesis. These perturbations again help explain greater H2S capacity decreases in the presence of CO2 for UiO-66NH2(Zr) and MIL-125-NH2(Ti) than they do for MIL-101-NH2(Cr). Multiple –OH stretches in the FTIR spectrum for UiO-66-NH2(Zr) are likely indicative of either remaining water or linker defects, which are inherently present following synthesis of UiO-66(Zr) type materials.59–61 Primary amine stretches, in contrast, remain static following CO2 exposure. This may suggest that MOF linker-based amines do not largely serve as CO2 physisorption sites at the tested partial pressure. Or, linker-based –NH2 groups are weaker physisorption sites than framework hydroxyl groups at tested conditions. A previous study on UiO-66-NH2(Zr) by Bordiga and coworkers62 also noted the relative insensitivity of linker-based primary amine stretches to CO2 binding under IR, despite DFT calculations identifying electronic interactions of the adsorbate with framework –NH2 groups. FTIR results for CO2 adsorption here are consistent with this previous finding.

16

ACS Paragon Plus Environment

Page 16 of 32

Page 17 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Time resolved FTIR spectra over 45 min of H2S exposure for activated MOFs are illustrated in Figure 5b and 5c. Minor shifts in the hydroxyl and amine stretching regions are observed upon H2S exposure. In MIL-101-NH2(Cr), the relatively low absorbance intensity of hydroxyl spectral features makes it unclear if the detected changes in Figure 5b are representative of lingering hydroxyl groups or other guest molecules (e.g., H2O, solvents) interacting with adsorbing H2S. Still, changes in the hydroxyl stretch region are observed for all materials, denoting the importance of these moieties for H2S binding. Primary amine stretch changes in UiO-66-NH2(Zr) are not evidenced in Figure 5c. This may suggest H2S, at the tested partial pressure, does not significantly interact with linkerbased amines in this framework. Note the small magnitude of changes with –NH2 groups are partially due to resolution limitations in DRIFTS. However, they are also indicative of weak physisorption interactions between H2S and pendant amine moieties. Slight spectral shifts of –NH2 and –OH stretches observed for the MOFs in the presence of H2S suggest the increase in H2S affinity for amine-incorporated materials seen previously in breakthrough experiments are due to H2S interactions with these framework species. Weak H2S binding also explains the recovery of porosity in previously discussed postexposure N2 physisorption experiments after degassing under heat and vacuum. These subtle binding interactions with framework groups are expected from previous studies. For instance, weak hydrogen bonding on linker-based –NH2 groups were proposed by Vaesen et al.16 through a similar FTIR experiments and GCMC simulations on MIL-125NH2(Ti). The inherent restriction of linker-based amine interactions with H2S to weak physisorption, as observed with FTIR measurements here, is consequent of the inflexible and regularly-partitioned framework primary amines on bridging linkers, which

17

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 32

discourages stronger amine-cooperative binding interactions. This importantly lends potential for cycled, regenerable H2S adsorption, facilitated by simple MOF construction using modified ligands.

(a)

(c)

(b)

Figure 5. FTIR spectra in hydroxyl and primary amine stretching regions for all MOFs (a) before and after 10% CO2/He exposure; after 0.5 % H2S/N2 exposure in (b) hydroxyl (OH) stretching region and (c) primary amine stretching region.

Summary and Conclusions

18

ACS Paragon Plus Environment

Page 19 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

To compare the effectiveness of MOFs for hydrogen sulfide (H2S) removal from natural gas, the commonly studied frameworks UiO-66(Zr), MIL-125(Ti), MIL-101(Cr), and their amine-functionalized analogues were evaluated for their H2S stability and removal capacities using fixed-bed experiments with 1% H2S/99% CH4 and 1% H2S/10% CO2/89% CH4 test mixtures. Powder X-ray diffraction and porosity measurements indicated a partial degradation of pelletized MIL-101(Cr) type materials after H2S exposure—not observed for the pristine framework or the other samples investigated. Adsorption experiments revealed linkerbased amines enhance the H2S uptake capacity in all materials studied here. In mixtures containing CO2, MIL-101-NH2(Cr) was found to be the MOF with the greatest H2S saturation capacity and breakthrough time. FTIR measurements demonstrated that –OH groups in the microporous materials behaved as likely CO2 physisorption sites, but were absent in activated mesoporous MIL-101-NH2(Cr). Despite exhibiting acid-gas stability as a pristine material, degradation of pelletized MIL101(Cr)/MIL-101-NH2(Cr) after acid gas exposure suggests this MOF may not be a pragmatic sour gas adsorbent in large scale adsorption operations that may require similar material processing. Relationships between mechanical and chemical stabilities are understudied in current MOF literature. So, these observations demonstrate how the interplay between these two properties result in important consequences on materials used for adsorption-based separations. Adsorption data reported here provide a benchmark to evaluate selective H2S removal from natural gas using MOFs. Linker-based amine functionalization in smaller pore

19

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

MOFs significantly enhanced H2S breakthrough time as compared to their respective parent materials. As a possible design heuristic, mitigation of H2S uptake loss after addition of CO2 in mesoporous MIL-101(Cr), in comparison to the smaller pore (