Porous metal–organic alloys based on soluble coordination cages†

Diverse strategies for the preparation of mixed-metal three-dimensional porous solids abound, although many of them lend themselves only moderate levels of tunability. Herein, we report the design and synthesis of surface functionalized permanently microporous coordination cages and their use in the isolation of mixed metal solids. Judicious alkoxide-based ligand functionalization was utilized to tune the solubility of starting copper(ii)-based cages and their resulting compatibility with the mixed-cage approach described here. We further prepared a family of isostructural molybdenum(ii) cages for a subset of the ligands. The preparation of mixed-metal cage solids proceeds under facile conditions where solutions of parent cages are mixed and product phases isolated. A suite of spectroscopic and characterization tools confirm the starting cages are intact in the amorphous product. Finally, we show that utilization of precise ligand functional groups can be used to prepare mixed cage solids that can be easily and cleanly separated into their constituent components through simple solvent washing or solvent extraction techniques.


Synthesis of 5-nonoxy-1,3-benzenedicarboxylic acid (5-nonoxy-H2bdc):
Dimethyl 5-hydroxyisophthalate (4 g, 19 mmol), potassium carbonate (2.9 g, 21 mmol), and 1iodononane (4.14 mL, 21 mmol) were added to a 250 mL RBF with 80 mL of acetone and stirred at 40 ˚C for 12 hours. The colorless solid was removed via vacuum filtration. The filtrate was added back to the RBF along with potassium hydroxide (2.46 g, 43.8 mmol) in 80 mL of DI H2O. The solution was stirred for 12 hours at 50 ˚C. The solution was concentrated to remove the acetone, then acidified to a pH of 1. The white precipitate was isolated via vacuum filtration then dried overnight at 60 ˚C to yield 64 % final product. 1 Figure S4. 1 H-NMR spectrum of 5-nonoxy-1,3-benzenedicarboxylic c in DMSO-d6.

Synthesis of 5-decoxy-1,3-benzenedicarboxylic acid (5-decoxy-H2bdc):
Dimethyl 5-hydroxyisophthalate (4 g, 19 mmol), potassium carbonate (2.9 g, 21 mmol), and 1iododecane (4.48 mL, 21 mmol) were added to a 250 mL RBF with 80 mL of acetone and stirred at 50 ˚C for 12 hours. The colorless solid was removed via vacuum filtration. The filtrate was added back to the RBF along with potassium hydroxide (2.46 g, 43.8 mmol) in 80 mL of DI H2O. The solution was stirred for 12 hours at 50 ˚C. The solution was concentrated to remove the acetone, then acidified to a pH of 1. The white precipitate was isolated via vacuum filtration then dried overnight at 60 ˚C to yield 76 % final product. 1 Figure S5. 1 H-NMR spectrum of 5-decoxy-1,3-benzenedicarboxylic acid in DMSO-d6.

Synthesis of 5-undecoxy-1,3-benzenedicarboxylic acid (5-undecoxy-H2bdc):
Dimethyl 5-hydroxyisophthalate (4 g, 19 mmol), potassium carbonate (2.9 g, 21 mmol), and 1iodoundecane (4.86 mL, 21 mmol) were added to a 250 mL RBF with 80 mL of acetone and stirred at 50 ˚C for 12 hours. The colorless solid was removed via vacuum filtration. The filtrate was added back to the RBF along with potassium hydroxide (2.46 g, 43.8 mmol) in 80 mL of DI H2O. The solution was stirred for 12 hours at 50 ˚C. The solution was concentrated to remove the acetone, then acidified to a pH of 1. The white precipitate was isolated via vacuum filtration then dried overnight at 60 ˚C to yield 91% final product. 1 Figure S6. 1 H-NMR spectrum of 5-undecoxy-1,3-benzenedicarboxylic acid in DMSO-d6.

Synthesis of 5-dodecoxy-1,3-benzenedicarboxylic acid (5-dodecoxy-H2bdc):
Dimethyl 5-hydroxyisophthalate (4 g, 19 mmol), potassium carbonate (2.9 g, 21 mmol), and 1iodododecane (5.18 mL, 21 mmol) were added to a 250 mL RBF with 80 mL of acetone and stirred at 50 ˚C for 12 hours. The colorless solid was removed via vacuum filtration.  Cage Synthesis: Synthesis of Cu24(5-hexoxy-bdc)24. Cu(NO3)2•2.5H2O (0.754 g, 3.24 mmol) and 5-hexoxy-1,3benzenedicarboxylic acid (0.862 g, 3.24 mmol) were dissolved in DMF (72 mL) and ethanol (228 mL) in a 500 mL jar and sealed with a Teflon-lined cap. The resulting solution was then heated at 80 ˚C for 3 days resulting in blue, rhombohedral shaped crystals and blue powder. The residual DMF/ethanol solution was decanted, and the crystals were washed with methanol. The methanol was decanted and fresh methanol was added and decanted 3 additional times over the course of 3 days to fully exchange and remove any DMF. The sample was then dried under vacuum to remove solvent and activated for gas adsorption measurements.
Synthesis of Cu24(5-heptoxy-bdc)24. Cu(NO3)2•2.5H2O (0.754 g, 3.24 mmol) and 5-heptoxy-1,3benzenedicarboxylic acid (0.908 g, 3.24 mmol) were dissolved in DMF (288 mL) and methanol (72 mL). The solution (15 mL) was dispersed in 20 mL scintillation vials (x20) and sealed with a Teflon-lined cap. The vials were then heated at 60 ˚C for 5 days resulting in blue, plate shaped crystals. The residual DMF/methanol solution was decanted, and the crystals were combined and then washed with methanol. The methanol was decanted and fresh methanol was added and decanted 3 additional times over the course of 3 days to fully exchange and remove any DMF. The sample was then dried under vacuum to remove solvent and activated for gas adsorption measurements.
Synthesis of Cu24(5-octoxy-bdc)24. Cu(NO3)2•2.5H2O (0.116 g, 0.5 mmol) and 5-octoxy-1,3benzenedicarboxylic acid (0.147 g, 0.5 mmol) were dissolved in DMF (2 mL) and ethanol (13 mL) in a 20 mL scintillation vial sealed with a Teflon-lined cap. The vial was then heated at 60 ˚C for S10 3 days resulting in a blue powder. The residual DMF/ethanol solution was decanted, and the material was washed with methanol. The methanol was decanted and fresh methanol was added and decanted 3 additional times over the course of 3 days to fully exchange and remove any DMF. The sample was then dried under vacuum to remove solvent and activated for gas adsorption measurements.
Synthesis of Cu24(5-nonoxy-bdc)24 bulk material. Cu(NO3)2•2.5H2O (0.344 g, 1.48 mmol) and 5-nonoxy-1,3-benzenedicarboxylic acid (0.456 g, 1.48 mmol) were dissolved in DMF (87.6 mL) and methanol (49.3 mL) in a 250 mL jar and sealed with a Teflon-lined cap. The resulting solution was then heated at 60 ˚C for 3 days resulting in a mixture of blue, rhombohedral shaped crystals and a blue powder. The residual DMF/methanol solution was decanted, and the material was washed with methanol. The methanol was decanted and fresh methanol was added and decanted 3 additional times over the course of 3 days to fully exchange and remove any DMF. The sample was then dried under vacuum to remove solvent and activated for gas adsorption measurements.
Synthesis of Cu24(5-decoxy-bdc)24 single crystals. Cu(NO3)2•2.5H2O (0.0063 g, 0.027 mmol) and 5-decoxy-1,3-benzenedicarboxylic acid (0.0087 g, 0.027 mmol) were dissolved in DMF (2.4 mL) and ethanol (0.6 mL) in a 4 mL vial and sealed with a Teflon-lined cap. The resulting solution was then heated at 80 ˚C for 3 days resulting in blue, rhombohedral shaped crystals. The residual DMF/ethanol solution was decanted, and the crystals were washed with methanol. The methanol was decanted and fresh methanol was added and decanted 3 additional times over the course of 3 days to fully exchange and remove any DMF. The sample was then dried under vacuum to remove solvent and activated for gas adsorption measurements.
Synthesis of Cu24(5-decoxy-bdc)24 bulk material. Cu(NO3)2•2.5H2O (0.126 g, 0.54 mmol) and 5-decoxy-1,3-benzenedicarboxylic acid (0.174 g, 0.54 mmol) were dissolved in DMF (12 mL) and ethanol (3 mL) in a 20 mL scintillation vial and sealed with a Teflon-lined cap. The resulting solution was then heated at 80 ˚C for 3 days resulting in a blue powder. The residual DMF/ethanol solution was decanted, and the material was washed with methanol. The methanol was decanted and fresh methanol was added and decanted 3 additional times over the course of 3 days to fully exchange and remove any DMF. The sample was then dried under vacuum to remove solvent and activated for gas adsorption measurements.
Synthesis of Cu24(5-undecoxy-bdc)24. Cu(NO3)2•2.5H2O (0.058 g, 0.25 mmol) and 5-undecoxy-1,3-benzenedicarboxylic acid (0.084 g, 0.25 mmol) were dissolved in DMF (12 mL) and ethanol (3 mL) in a 20 mL scintillation vial and sealed with a Teflon-lined cap. The resulting solution was then heated at 80 ˚C for 3 days resulting in a blue powder. The residual DMF/ethanol solution was decanted, and the material was washed with methanol. The methanol was decanted and fresh methanol was added and decanted 3 additional times over the course of 3 days to fully exchange and remove any DMF. The sample was then dried under vacuum to remove solvent and activated for gas adsorption measurements.

Synthesis of Cu24(5-dodecoxy-bdc)24.
Procedure was modified from a previously published synthesis. 1 Accordingly, Cu(OAc)2•H2O (0.082 g, 0.41 mmol) and 5-dodecoxy-3,5benzenedicarboxylic acid (0.147 g, 0.42 mmol) were dissolved separately in DMF (10 mL). The two solutions were subsequently combined. Methanol (15 mL) was added to the solution to S11 precipitate out the cage as a blue powder. The supernatant was decanted and fresh methanol was added and decanted 3 additional times over the course of 3 days to fully exchange and remove any DMF. The sample was then dried under vacuum to remove solvent and activated for gas adsorption measurements.
Synthesis of Crystalline Mo24(5-hexoxy-bdc)24. Mo2(OAc)4 (0.1849 g, 0.432 mmol) and 5hexoxy-1,3-benzenedicarboxylic acid (0.1150 g, 0.432 mmol) were added to a 20 mL scintillation vial and dissolved in 9.6 mL DMA and 2.4 mL MeOH. Upon heating at 100 °C in an N2 glovebox for 2 days, crystalline powder was formed. The powder was decanted of excess mother liquior and washed with anhydrous methanol three times, exchanging the solvent every 8 hours, and dried under vacuum to yield 0.1199g of an orange powder.
Synthesis of Mo24(5-heptoxy-bdc)24. Mo2(OAc)4 (0.0462 g, 0.108 mmol) and 5-heptoxy-1,3benzenedicarboxylic acid (0.0303 g, 0.108 mmol) were added to a 20 mL scintillation vial and dissolved in 3 mL of DMF upon heating at 100 °C in a dry bath in an N2 glovebox. About 15 mL of anhydrous methanol was used to precipitate out orange powder. The vial was centrifuged down and the powder was washed with anhydrous methanol three times, exchanging the solvent every 8 hours. It was dried under vacuum to yield an orange powder.

Synthesis of Mo24(5-octoxy-bdc)24.
Mo2(OAc)4 (0.0462 g, 0.108 mmol) and 5-octoxy-1,3benzenedicarboxylic acid (0.0318 g, 0.108 mmol) were added to a 20 mL scintillation vial and dissolved in 3 mL of DMF upon heating at 100 °C in a dry bath in an N2 glovebox. About 15 mL of anhydrous methanol was used to precipitate out orange powder. The vial was centrifuged down and the powder was washed with anhydrous methanol three times, exchanging the solvent every 8 hours, and dried under vacuum to yield an orange powder.
Synthesis of Crystalline Mo24(5-nonoxy-bdc)24. Mo2(OAc)4 (0.1849 g, 0.432 mmol) and 5nonoxy-1,3-benzenedicarboxylic acid (0.1332 g, 0.432 mmol) were added to a 20 mL scintillation vial and dissolved in 10.8 mL DMF and 1.2 mL EtOH. Upon heating at 100 °C in an N2 glovebox for 2 days, crystalline powder was formed. The powder was decanted of excess mother liquior and washed with anhydrous methanol three times, exchanging the solvent every 8 hours. It was dried under vacuum to yield 0.0914g of an orange powder.

Synthesis of Mo24(5-decoxy-bdc)24.
Mo2(OAc)4 (0.0462 g, 0.108 mmol) and 5-decoxy-1,3benzenedicarboxylic acid (0.0348 g, 0.108 mmol) were added to a 20 mL scintillation vial and dissolved in 3 mL of DMF upon heating at 100 °C in a dry bath in an N2 glovebox. About 15 mL of anhydrous methanol was used to precipitate out orange powder. The vial was centrifuged down and the powder was washed with anhydrous methanol three times, exchanging the solvent every 8 hours, and dried under vacuum to yield an orange powder.

Synthesis of Mo24(5-undecoxy-bdc)24.
Mo2(OAc)4 (0.0462 g, 0.108 mmol) and 5-undecoxy-1,3benzenedicarboxylic acid (0.0363 g, 0.108 mmol) were added to a 20 mL scintillation vial and dissolved in 3 mL of DMF upon heating at 100 °C in a dry bath in an N2 glovebox. About 15 mL of anhydrous methanol was used to precipitate out orange powder. The vial was centrifuged down and the powder was washed with anhydrous methanol three times, exchanging the solvent every 8 hours, and dried under vacuum to yield an orange powder.

Synthesis of Crystalline Mo24(5-dodecoxy-bdc)24.
Mo2(OAc)4 (0.1849 g, 0.432 mmol) and 5dodecoxy-1,3-benzenedicarboxylic acid (0.1514 g, 0.432 mmol) were added to a 20 mL scintillation vial and dissolved in 9.6 mL DMF and 2.4 mL EtOH. Upon heating at 100 °C in an N2 glovebox for 2 days, crystalline powder was formed. The powder was decanted of excess mother liquor and washed with anhydrous methanol three times, exchanging the solvent every 8 hours. It was dried under vacuum to yield 0.1182g of an orange powder.

Synthesis of Cu/Mo Cage Alloys:
Synthesis of Cu/Mo(5-hexoxy-bdc). Mo24(5-hexoxy-bdc)24 (0.0360 g, 0.1 mmol) and Cu24(5hexoxy-bdc)24 (0.0327 g, 0.1 mmol) were added to a 20 mL scintillation vial and dissolved in ~5mL of DMF. About 15 mL of anhydrous methanol was used to precipitate out a brown powder. The vial was centrifuged down and the powder was washed with anhydrous methanol three times, exchanging the solvent every 8 hours, and dried under vacuum to yield 0.0614g of a brown powder.

Synthesis of Cu/Mo(5-nonoxy-bdc).
Mo24(5-nonoxy-bdc)24 (0.0402 g, 0.1 mmol) and Cu24(5nonoxy-bdc)24 (0.0369 g, 0.1 mmol) were added to a 20 mL scintillation vial and dissolved in ~5mL of THF. About 15 mL of anhydrous methanol was used to precipitate out green powder. The vial was centrifuged down and the powder was washed with anhydrous methanol three times, exchanging the solvent every 8 hours, and dried under vacuum to yield 0.065g of a green powder.

Synthesis of Cu/Mo(5-dodecoxy-bdc).
Mo24(5-dodecoxy-bdc)24 (0.0444 g, 0.1 mmol) and Cu24(5-dodecoxy-bdc)24 (0.0411 g, 0.1 mmol) were added to a 20 mL scintillation vial and dissolved in ~5mL of benzene. The green solution was pipetted into a gas adsorption tube and freeze dried using liquid N2. A fluffy green powder (0.0639 g) was formed after the frozen cage was thawed slowly by immersing the tube in an ice bath under dynamic vacuum.

Copper(II) Cage NMR Spectra:
In order to obtain NMR, all copper(II) cages were digested via DCl and dissolved in DMSO-d6 after activation.

Molybdenum(II) Cage NMR Spectra:
In order to obtain the most soluble cages for NMR, all molybdenum(II) cages were crashed out from DMF and washed three times with MeOH or EtOH at room temperature.

NMR Spectra of Mo(II) Ligand Exchange Controls
In these experiments, a Mo(II) alkoxy cage was combined with a different length alkoxy ligand in a 1:1 ratio and heated at 100 °C in DMF to produce a homogenous solution. For instance, 0.1 mmol of Mo24(5-hexoxy-bdc)24 was combined with 0.1 mmol of 5-nonoxy-1,3-benzenedicarboxylic acid in DMF. Upon addition of MeOH, a powder crashed out and was washed with MeOH three times. The samples were then dried under dynamic vacuum and dissolved in NMR solvent in order to determine if ligand exchange occurred.

Characterization of Metal and Ligand Exchange Controls
In this experiment, 0.1 mmol of Cu(5-hexoxy-bdc)24 was combined with 0.1 mmol Mo(5-nonoxybdc)24 in DMF and heated to 100 °C to produce a homogenous solution. The Cu/Mo alloy powder was crashed out of solution using MeOH and washed three times with MeOH to remove residual DMF. After the powder was dried under dynamic vacuum, the Mo(5-nonoxy-bdc)24 was extracted via THF and separated into another vial. The alloy powder was washed thrice with THF to remove residual Mo(5-nonoxy-bdc)24 cage and the washed were combined. Methanol was used to crash out the Mo(5-nonoxy-bdc)24 cage from the THF solution. The powder was washed three times with MeOH, dried under vacuum, and dissolved in NMR solvent in order to determine if ligand or metal exchange occurred. The isolated Cu(5-hexoxy-bdc)24 was dried under dynamic vacuum for use in UV-Vis experiments to determine if ligand or metal exchange occurred. Figure S25. 1 H-NMR spectrum of Mo24(5-nonoxy-bdc)24 separated from Cu(5-hexoxy-bdc)24 after isolation from DMF solution and THF extraction. The NMR was taken in THF-d8.

Phase Transfer of CuMo Alloy for Cage Separation
A 1:1 ratio of Mo24(5-hexoxy-bdc)24 or Mo24(5-nonoxy-bdc)24 to Cu24(5-dodecoxy-bdc)24 and Cu24(5-hexoxy-bdc)24 or Cu24(5-nonoxy-bdc)24 to Mo24(5-dodecoxy-bdc)24 was added to a 20 mL scintillation vial and dissolved in DMF. Then hexane, heptane or octane was layered on top of the DMF. Upon shaking, the Mo24(5-dodecoxy-bdc)24 or Cu24(5-dodecoxy-bdc)24 was extracted from the DMF to the alkane layer as seen by the color change of the alkane solvent. For the hexane phase transfer, this layer was collected via pipette after extracting three times to remove residual cage. The hexane layer was then pumped down to dryness and the DMF layer was crashed out with MeOH, washed three times and dried via vacuum. For the octane phase transfer, as MeOH is not miscible with octane, the DMF layer was simply crashed out with MeOH, washed three times and dried via vacuum.

X-Ray Crystallography:
X-ray structural analysis for Cu24(5-hexoxy-bdc)24, Cu24(5-nonoxy-bdc)24, and Cu24(5-decoxybdc)24. Crystal data and refinement details are shown in Table S1. Crystals were mounted using viscous oil onto a plastic mesh and cooled to the data collection temperature. Data were collected on a Bruker-AXS APEX II DUO CCD diffractometer with Cu-Kα radiation (λ = 1.54178 Å) focused with Goebel mirrors. Unit cell parameters were obtained from 48 data frames, 0.5º ω, from different sections of the Ewald sphere. The unit-cell dimensions, equivalent reflections and systematic absences in the diffraction data are uniquely consistent with C2/m, C2 and Cm. Refinement in the centrosymmetric space group options, C2/m, yielded chemically reasonable and computationally stable results of refinement. The data were treated with multi-scan absorption corrections. 2 Structures were solved using intrinsic phasing methods 3 and refined with full-matrix, least-squares procedures on F2. 4 Residual electron density, solvent molecules and atoms that cannot be assigned a reasonable model, were treated as diffused electron density using Squeeze. 5 The ligand alkyl chains were confirmed by digestion followed by NMR spectroscopy and thus the electron density that was deducted from the diffused alkyl chain ends could be assigned. Although the assignments of void contents were based on the best spatial and electron count fit of the reaction solvent DMF and water, we cannot discount the possibility of alternative void contents that might consist of thermal decomposition products of amide solvents. The compound molecules were located at an intersection of a mirror plane and a two-fold rotation axis. A coordinated DMF molecule was found disordered in two positions with a 53/47 refined site occupancy ratio.
These compounds consistently deposit as weakly diffracting crystals and the results herein represent the best of several trials. In order to compensate for the low-resolution data, various constraints and restraints were applied. All non-hydrogen atoms were refined with anisotropic displacement parameters. All H-atoms were treated as idealized contributions with geometrically calculated positions and with Uiso equal to 1.2-1.5 Ueq of the attached carbon atom. Atomic scattering factors are contained in the SHELXTL program library. 4 The structures have been deposited at the Cambridge Structural Database under the following CCDC deposition numbers: 2018425-2018427. *These data-sets were processed with Squeeze to account for diffused electron density in areas of the difference map that cannot be modelled as chemically reasonable, computationally convergent, moieties. The moieties in square brackets are based on the best fit of the known ligand alkyl chain, majority solvent (DMF) and water.                Figure S68. Thermogravimetric analysis of methanol washed Cu24(5-heptoxy-bdc)24 cage from 25 °C to 600 °C at a ramp rate of 2 degrees per minute. Figure S69. Thermogravimetric analysis of methanol washed Cu24(5-octoxy-bdc)24 cage from 25 °C to 600 °C at a ramp rate of 2 degrees per minute. Figure S70. Thermogravimetric analysis of methanol washed Cu24(5-nonoxy-bdc)24 cage from 25 °C to 600 °C at a ramp rate of 2 degrees per minute. Figure S71. Thermogravimetric analysis of methanol washed Cu24(5-decoxy-bdc)24 cage from 25 °C to 600 °C at a ramp rate of 2 degrees per minute. Figure S72. Thermogravimetric analysis of methanol washed Cu24(5-undecoxy-bdc)24 cage from 25 °C to 600 °C at a ramp rate of 2 degrees per minute. Figure S73. Thermogravimetric analysis of methanol washed Cu24(5-dodecoxy-bdc)24 from 25 °C to 600 °C at a ramp rate of 2 degrees per minute. Figure S74. Thermogravimetric analysis of methanol washed Mo24(5-hexoxy-bdc)24 (orange), CuMo(5-hexoxy-bdc) alloy (green) and Cu24(5-hexoxy-bdc)24 (blue) from 25 °C to 600 °C at a ramp rate of 2 degrees per minute. Figure S75. Thermogravimetric analysis of methanol washed Mo24(5-nonoxy-bdc)24 (orange), CuMo(5-nonoxy-bdc) alloy (green) and Cu24(5-nonoxy-bdc)24 (blue) from 25 °C to 600 °C at a ramp rate of 2 degrees per minute.

Gas Adsorption Measurements:
Dinitrogen and Carbon dioxide were used for adsorption measurements. To measure the isotherms, a Micromeritics 3Flex gas adsorption analyzer was used. Samples were loaded into weighed analysis tubes under N2 atmosphere to prevent oxidation or water absorption into samples. The tubes were capped with Transeals and removed from the glovebox. They were activated under vacuum at various temperatures on a Smartvac degas system until the static outgas rate was less than 2 µbar/min. After degassing, the tube was removed from the Smartvac under vacuum and weighed to determine the mass of the sample in the tube. For cryogenic N2 measurements, an isothermal jacket was fitted on the tube. Degas surveys using CO2 at 195 K or N2 at 77 K were performed on the samples after they were heated in increments of 25 °C to determine the optimal activation temperature. Langmuir Surface Areas were calculated via the Micromeritics software. BET calculations were calculated via the first and second consistency check. 6 Both N2 and CO2 sorption measurements were performed on some materials due to the lack of porosity to N2.