Research Data Supporting “Phase Behavior of Light-Responsive Lyotropic Liquid Crystals for Molecular Solar Thermal Energy Storage”

SAXS heating scans with in-situ pseudo-DSC measurements were performed at beamline BM26, at the European Synchrotron Research Facility (Grenoble, France). The X-ray beam energy was 12 keV and the camera length was 3 m. Samples (~5 mg) were loaded into aluminium DSC pans and hermetically sealed before loading into a Linkam DSC600 for measurement. The temperature was controlled using a T96 controller and LINK software, with cooling using liquid nitrogen. For the charging studies, samples were irradiated from the top in the DSC pan for a known time before sealing and measuring. X-ray exposure time was 20 s for every sample, with frames taken every 30 – 60 s during heating cycles. Heating cycles varied slightly depending on the sample, to maximize the beamtime available, whilst still measuring all phase transitions. For 100 wt% samples, cycle 1 was used; for 50 – 90 wt% LLC samples, cycle 2 was used; for UV irradiated samples, cycle 3 was used (Table 1). Table 1. DSC parameters used during in-situ SAXS and DSC studies, where rate is the heating or cooling rate, Tstart and Tend are the start and end temperatures of that ramp and the dwell time is the wait time between reaching the final temperature and ending measurement. Cycle Step Rate / °C min-1 Tstart/ °C Tend/ °C Dwell time / min 1 1 5 -20 80 0 2 2 80 150 0 3 75 150 25 20 2 1 5 -20 150 0 2 75 150 25 0 3 1 5 -20 150 0 2 75 150 25 15 SAXS images were reduced to 1D curves using the processing functionality in DAWN including calibration relative to a silver behenate standard, normalization to measured beam intensity on a beamstop mounted photodiode, definition of Poisson errors and solid angle correction before azimuthal integration. For phase analysis, peak positions were found by subtracting the background and using the Find Peaks tool in Origin(Pro), Version 2021b, OriginLab Corporation. Interlamellar spacings, d, were calculated using the position of the primary SAXS peak, Q0, using the relationship: d = 2π / Q0. All other SAXS measurements were performed at the high-throughput SAXS beamline B21, Diamond Light Source (Oxfordshire, UK).Click or tap here to enter text. The X-ray beam energy was 13.1 keV and detector distance set to 3.7 m, giving a Q range of 0.0045 – 0.34 Å-1. Samples were loaded into polyimide capillaries (low viscosity) or Kapton tape. Background of water, or ethylene glycol, in polyimide capillary or Kapton tape alone were subtracted using ScÅtter and 20 frames of 1 s exposure time were averaged to give the final data pattern. Unless otherwise stated, samples were measured at 25°C. For temperature ramp samples, the time needed to equilibrate the sample to the given temperature was measured using a thermocouple. The sample was then held at the temperature for this equilibration time before taking a SAXS measurement. Peak fitting and d spacing calculations were carried out as above. POM images were taken using a Leica EC3 camera fitted to an Olympus BHM microscope. Samples were placed between 2 glass slides and observed at room temperature under crossed polars. For heating scans, slides were placed on a Linkam PE120 Peltier heat stage controlled by a T96 LinkPad controller and cooled using a water circulation pump. DSC scans were taken using a Mettler Toledo DSC1 with a liquid nitrogen cooling system and Argon purge gas. Analysis was conducted in the STARe software. Due to the variation in peak positions between samples, peak integration limits were determined by eye and have been included in the results tables as lower and upper limits (°C). For DSC scans of the E isomer in the solid-state, samples (~1 mg) were loaded into aluminum DSC pans and sealed with a pierced lid. Samples were equilibrated at -20°C for 2 minutes, heated from -20 – 150°C at 5 °C min-1 and subsequently cooled from 150 – -20°C at 5 °C min-1. One out of three repeats for each chemically distinct sample was heated a second time. To determine the isomerization enthalpies of the AzoPS, AzoPS were first dissolved in CDCl3 (~10 mg mL-1) and irradiated at 365 nm in a glass vial, whilst stirring, for 2 – 24 h. An NMR scan was taken to confirm isomerization. The sample was then dried using compressed air, whilst still irradiating, for ~4 h. The sample was then transferred to an aluminum DSC pan (~1 mg), with a hole-punched lid for measurement. The remaining dry sample was re-dissolved in CDCl3 and an NMR was taken to calculate the isomerization degree after the drying process. DSC scan consisted of heating from 0 – 150°C at 6 °C min-1, cooling to 0°C at 40 °C min-1 and a second heat from 0 – 150°C at 6 °C min-1. After DSC measurement, the pan was broken open and sample dissolved in CDCl3. An NMR spectrum was taken to determine the change in isomerization degree before and after the DSC measurement. To get the theoretical isomerization enthalpy, the measured enthalpy change on isomerization was divided by the percentage change in isomerization degree before and after the DSC scan, to obtain the theoretical enthalpy change on 100% conversion. TGA measurements were performed on a TA Instruments Discovery SDT650. C8AzoC4E4 at 50 wt% water was measured, with an initial mass of 9.2 mg. The sample was measured from 0 – 700 ºC at a heating rate of 10 ºC min-1, using nitrogen as a purge gas. To calculate the isomerization degree, spectra were taken using a Bruker 400 MHz spectrometer, over 16 scans. Spectra were analyzed using Topspin 4.1.4. For solution studies, AzoPS were dissolved in DMSO-d6 (10 mM). The sample (1 mL) was loaded into a quartz NMR tube and irradiated with UV (365 nm) light for 60 min, with NMR spectra taken at periodic time stamps (1, 2, 3, 4, 5, 10, 15, 30, 45 and 60 minutes). After 60 minutes, there was negligible further change to the NMR spectra on irradiation. All spectra were calibrated to DMSO-d6 (δ = 2.50 ppm). The positions of the azobenzene peaks characteristic to the E and Z isomers were determined from the 0 and 60 min UV irradiation spectra. For the E these are: δ = 7.85 (doublet, 2H), 7.76 (doublet, 2H), 7.38 (doublet, 2H) and 7.11 (doublet, 2H) ppm. For the Z these are: δ = 7.14 (doublet, 2H), 6.83 (singlet, 4H), and 6.74 (doublet, 2H) ppm. The peaks were integrated and the doublet at 7.85 ppm assigned an integral of 2. The singlet at 6.83 ppm was divided by 2 and set as 2 separate values for the error calculations. The doublets at 7.11 ppm (E) and 7.14 ppm (Z) were excluded from analysis as the overlap means they cannot be deconvoluted. The percentage isomerization was calculated by taking a ratio of the average of the Z integrals to the sum of the average of E and Z integrals and multiplying by 100%. Errors were determined using the standard deviations of the integrals of different peaks within the E and Z assignments. To determine the percentage isomerization for AzoPS in LCs (50, 70, 90 and 100 wt% in D2O), samples were loaded into DSC pans (~7 mg) and top-irradiated using UV light for time between 1 – 24 h. After the irradiation period, the pan was transferred to a vial and DMSO-d6 (500 μL) added. The sample was shaken to dissolve, transferred to an NMR tube and measured immediately. Percentage isomerization was calculated as above. UV-Vis absorbance spectra were measured using a Perkin Elmer Lambda 750 spectrometer with a slit width of 1 nm and a scan speed of 266.75 nm min-1. Measurements were taken at 1 nm intervals from 700 – 200 nm, using quartz cuvettes with a 10 mm path length. Temperature control was achieved using Perkin Elmer Peltier temperature controller 201. For cycling experiments, a sample of C8AzoC4E4 (2 mL, 50 μM in water) was irradiated for 20 minutes with either UV (365 nm) or blue (455 nm) before measurement. This was repeated for 19 complete UV-blue cycles. To determine the thermal half-life of the Z isomer, solutions of AzoPS in water (C8AzoC4E4: 100 μM; C6AzoC4E4 and C10AzoC4E4: 50 μM) were made up and loaded (2 mL) into cuvettes. A reference sample of water was used for all spectra. The UV-Vis absorbance spectrum in the native, E state was first taken at 25°C. The sample was irradiated for 20 minutes under UV light, and the formation of the Z PSS was confirmed by taking another spectrum at 25°C. The sample was irradiated for a further 10 minutes whilst the sample cell was heated to the measurement temperature (45 – 65°C). Spectra were then taken every 3 minutes for 80 measurements; except for C8AzoC4E4 at 45°C, where spectra were taken every 5 minutes for 72 measurements. Subtracted data were plotted and the position of the absorption maximum for the E π-π* transition (λmax) was found using the using the Find Peaks tool in Origin(Pro), Version 2021b, OriginLab Corporation. The percentage isomerized (%Z) was calculated using: %Z= (A_t- A_0)/(A_PSS- A_0 ) (1) where A0 = absorbance at λmax before UV irradiation, APSS = absorbance at λmax in the photostationary state and At = absorbance at λmax at time t. Plots of ln(%Z) against time were then fitted to a straight-line to give the first-order rate parameter (k) according to: k t = ln(%Z) (2) Eyring plots of ln(k/T) vs. 1/T gave the enthalpy change (H) and entropy change (S) for isomerization according to: k= (k_B T)/h exp⁡((-∆H)/RT) exp⁡〖(∆S/R)〗 (3) where kB = Boltzmann’s constant, T = absolute temperature in Kelvin, h = Planck’s constant, and R = molar gas constant. The thermal half-life at 20°C was determined by extrapolating back from the straight-line plot. For irradiation stability measurements, C8AzoC4E4 (2 mL, 50 μM in water) in water was placed in a quartz cuvette with 10 mm path length. The sample was irradiated with UV light (wavelength = 365 nm, irradiance = 6.00 mW cm-2) constantly, whilst taking UV-Vis absorbance spectra every 30 minutes for 24 hours using an Ocean Optics FX Miniature spectrometer with an integration time of 5 ms. The spectra were plotted using Origin(Pro) (\Version 2021b, OriginLab Corporation) and absorbance maxima were found using the Find Peaks tool. The absorbance maximum for the Z-rich PSS (315 nm) was plotted against time to obtain a gauge for damage to the photoswitch on irradiation, which would result in a gradual decrease in this value over time. A conformer search was performed for both the E and Z isomers of each molecule using the GFN2-xTB methods by Bannwarth and co-workers and the CREST code. GFN2-xTB incorrectly predicts the Z isomer to be lowest in energy, unconstrained conformer searches for the E isomer would yield Z isomers instead. Hence, in the case of the conformer search for the E isomer, the torsion angle around the -N-N- bond was kept fixed. The lowest conformer for each isomer from the conformer search were subsequently re-optimized in DFT using B97-3c counterpoise corrected composite method by Brandenburg and co-workers,10 both in vacuum and in water. In the latter case, the solvation in water was described using the COSMO implicit solvation model (r 80). All DFT calculations were performed using the Turbomole code.