Reduced thermal expansion by surface-mounted nanoparticles in a pillared-layered metal-organic framework

Control of thermal expansion (TE) is important to improve material longevity in applications with repeated temperature changes or fluctuations. The TE behavior of metal-organic frameworks (MOFs) is increasingly well understood, while the impact of surface-mounted nanoparticles (NPs) on the TE properties of MOFs remains unexplored despite large promises of NP@MOF composites in catalysis and adsorbate diffusion control. Here we study the influence of surface-mounted platinum nanoparticles on the TE properties of Pt@MOF (Pt@Zn2(DP-bdc)2dabco; DP-bdc2-=2,5-dipropoxy-1,4-benzenedicarboxylate, dabco=1,4-diazabicyclo[2.2.2]octane). We show that TE is largely retained at low platinum loadings, while high loading results in significantly reduced TE at higher temperatures compared to the pure MOF. These findings support the chemical intuition that surface-mounted particles restrict deformation of the MOF support and suggest that composite materials exhibit superior TE properties thereby excluding thermal stress as limiting factor for their potential application in temperature swing processes or catalysis.


S1.1 General remarks
All chemicals were purchased from commercial suppliers and used without further purification. MOF synthesis was conducted with pure N,N-dimethylformamide (DMF) (99.8%). Fresh DMF was used for solvent exchange. Ethanol for solvent exchange was purchased as technical grade and redistilled prior use. All experiments and procedures were carried out on air unless stated otherwise.
Liquid state NMR spectra were recorded on a Bruker Ultrashield DRX400 spectrometer ( 1 H: 400.13 MHz) at ambient temperature (298 K). The 1 H NMR spectroscopic chemical shifts δ are reported in ppm relative to tetramethylsilane. 1 H NMR spectra are referenced against the residual proton resonances of the respective deuterated solvent as an internal standard (DMSO-d6: δ (1H) = 2.50 ppm). MOF samples were digested and subsequently measured in 0.5 ml DMSO-d6 with 0.05 ml DCl (7.6 N), all other substances were dissolved and measured in DMSO-d6.

S1.2 Linker synthesis
Organic linkers were synthesized via Williamson Etherification according to literature known procedures, 1 albeit slightly altered/optimized.

Washing and activation procedure
Washing and solvent exchange steps are conducted by vigorously shaking solid and solvent in a capped centrifuge tube, then centrifugation and decanting of the supernatant.
The obtained powder is washed immediately with DMF (3x 10 ml) until the supernatant is clear to yield the MOF material in its as-synthesized (as) state. While wet, this state is verified by X-ray powder diffraction (see chapter S2).
To remove DMF the as material is first washed with ethanol (10 ml), then twice soaked in fresh ethanol (2x 10 ml) overnight, and lastly soaked once in dichloromethane (10 ml) overnight. The wet powder is then pre-dried on air at ambient conditions before proper activation in vacuo at r.t. for 16 h and subsequently at 70 °C for 4 h. The resulting dry sample is stored under argon for further use to rule out possibility of gradual hydrolysis during prolonged exposure to ambient moisture. Linker and pillar incorporation ratio was verified via 1 H NMR (see below for specifications), CHNS contents were determined via combustion analysis. Zn and Pt contents were determined by atom adsorption spectroscopy or photometry for samples with more than 50 mg available with uncertainty of +-0.1 w%. Elemental analysis was provided and conducted by the TUM CRC Microanalytical laboratory.

S1.4 MOF structure and flexibility
Zn2(DP-bdc)2dabco is a pillared-layered MOF consisting of two-dimensional square-lattice layers spanned by paddlewheel units and DP-bdc 2linkers which are congruently stacked and connected via the dabco pillars in the third dimension. Many derivatives of the Zn2(fu-bdc)2dabco family (fu = 2,5functionalization) are flexible MOFs which can transition between a contracted np phase and an expanded lp phase (in a first-order phase transition). 2 The in this work studied MOF only undergoes this transition in response to polar solvents (like DMF, EtOH) and CO2. Other functionalizations, however, unlock this flexibility in response to temperature or mechanical pressure as well. 3,4 These MOFs are synthesized in their lp state (due to the reaction medium DMF) and transitioned to their np phase during careful solvent exchange and removal. This state is metastable until exposure to abovementioned polar adsorbates which switch the MOF to its lp phase. This is then again reversible by i.e. solvent evaporation, drying, gas removal etc. The interested reader is directed to this review article on flexible MOFs. 5 Supplementary Figure 1: From left to right structure of the terephthalic acid derivative linker, dabco pillar, and Zn paddlewheel building block. Only coordinating nitrogen atoms shown in the paddlewheel unit for clarity. Right hand structures visualize the np to lp phase transition as is present and triggered in Zn2(DP-bdc)2dabco by polar solvents (like DMF, EtOH) and CO2. Blue nodes = paddlewheel unit, grey struts = DP-bdc 2-, green struts = dabco.

S2 Powder X-ray diffraction
PXRDs of the as-synthesized (as) and resolvated samples were measured using Bragg-Bentano geometry with a silicon wafer plate on a Rigaku Benchtop MiniFlex 600-C. X-ray Cu Kα radiation (λ = 1.5406 Å) with a voltage of 40 kV and current of 15 mA was used.
Activated (dry) samples and samples after the CO2 and N2 physisorption cycles were filled in glass capillaries in a glovebox under argon atmosphere and PXRDs were measured in Debeye-Scherrer geometry on a PANalytical Empyrean diffractometer. X-ray Cu Kα radiation (λ = 1.5406 Å) with a voltage of 45 kV and current of 40 mA was used.

Supplementary Figure 2:
PXRD patterns in the range of 2θ = 6-12° and 6-50°. Black: as-synthesized (as) with infiltrated DMF; red: activated (dry); grey: re-infiltrated with DMF after activation; blue: material state after all conducted gas adsorption measurements presented in this work (N2 and CO2). Transition from lp state in as to np state in dry is observable by shift of the 110 reflection from 2θ = 8.2-8.4° to 9.7° and intensity increase of the 001 reflection at 2θ = 9.2°. 6 Due to the very low NP loading and significant peak broadening coming from diffraction domains in the nano regime, no peaks corresponding to Pt nanoparticles can be observed at the expected angles around 40° (I(111)) and 47° (I(200)) 2θ. 7 Free space of the sample tube was determined after measuring each adsorption isotherm using helium (>99.999 vol%). A liquid nitrogen bath was used for measurements at 77 K and a dry ice -acetone bath was used for measurements at 195 K.

Supplementary Figure 3:
Isotherms of N2 at 77 K (green) and CO2 at 195 K (blue/purple) on all materials with adsorption (filled symbols) and desorption (empty symbols), respectively. All materials with platinum nanoparticles (D1-3) show slightly decreased total absolute uptake compared to the reference material (D0), which we attribute to a small loss in accessible porosity due to nanoparticles blocking pores or pore access, and a widened np to lp transition pressure range. This more gradual opening is attributed to surface stress exhibited by the NPs which slightly rigidify the underlying framework. Layers around the NPs do as a result require a higher CO2 partial pressure to undergo the opening breathing motion. This could explain both the widening of the opening pressure range, as well as the almost identical onset of the step. The widening is correlated to Pt nanoparticle presence, but not linearly to Pt content. The reverse lp to np transition during desorption occurs in line with D0. After desorption framework integrity and completed np phase transition of all materials was confirmed by PXRD (see section S2).

S8
S4 Thermogravimetric analysis and differential scanning calorimetry Thermogravimetric analysis coupled with differential scanning calorimetry (TGA-DSC) was conducted on a Netzsch TG-DSC STA 449 F5 in a temperature range from 25 °C to 800 °C with a heating rate of 10 K min -1 under argon flow (flow rate: 20 mL min -1 ). It should be noted that the sample is briefly (few seconds) exposed to air before the measurement when the aluminium oxide pan is transferred from the argon filled transport vial to the sample holder stage. Darker purple corresponds to higher reflection intensity. Each set of 28 patterns is normalized to its highest intensity signal. Prominent gradual shift of peaks composites to lower angles is attributed to the literature known thermal expansion of Zn2(DP-bdc)2dabco. S11

S6 Pawley fitting and cell parameter
Raw VTXRD pattern (see chapter S5) were fitted directly.

S8 Electron microscopy
Scanning transmission electron microscopy (STEM) micrographs with energy dispersive X-ray spectroscopy (EDS) elemental mappings were recorded with a JEM-ARM200F "NEOARM" microscope from JEOL (Germany) GmbH with a cold FEG electron source operated at 200 kV. Samples were prepared by depositing a drop of the solid dispersed in ethanol onto carbon-coated copper grids (200 mesh) and dried in air. The electron tomography was carried out in a 2 or 3° step using the TEMography™ software for both recording and 3D-image reconstruction.