Reversible quantitative guest sensing via spin crossover of an iron(ii) triazole

Discrete mononuclear [FeII(tolpzph)2(NCS)2]·THF (1·THF), shows highly sensitive, robust and reversible solvent-dependent spin crossover, enabling it to act as a quantitative small molecule sensor.


Complex synthesis General remarks regarding complex synthesis
Dry THF and dry Et 2 O were obtained from a Pure Solv MD-6 Solvent Purification System and degassed by bubbling with dry Ar (g) for 20 min before use. Other chemicals were bought commercially and used as received.

(NCS) 2 ]·THF (1·THF)
Method A (microcrystals). Under standard Schlenk conditions, a yellow solution of anhydrous FeCl 2 (16.2 mg, 0.127 mmol) in dry THF (10 mL) was added via cannula to solid NaNCS (21.0 mg, 0.255 mmol) and the mixture stirred for 20 minutes causing the solution to turn deep red and a white precipitate to form. The suspension was filtered and the deep red filtrate transferred via cannula to a colourless solution of tolpzph (80 mg, 0.255 mmol) in dry THF (5 mL), resulting in a violet solution. The solution was stirred for 30 minutes, then reduced to half the volume by bubbling with N 2(g) . On standing for 48 hours at room temperature dark green microcrystals precipitated. These were filtered and surface dried by repeated brief vacuum and refill (dry N 2(g) ) cycles (x5) on the Schlenk line (82 mg, 94.3 μmol, 74%). These green microcrystals were stored under Ar (g)

Method B (single crystals).
Under standard Schlenk conditions, a yellow solution of anhydrous FeCl 2 (24.3 mg, 0.192 mmol) in dry THF (5 mL) was added via cannula to a colourless solution of NaNCS (31.2 mg, 0.384 mmol) in dry THF (5 mL) and the mixture stirred for 20 minutes causing the solution to turn deep red and a white precipitate to form. The suspension was filtered and the deep red filtrate transferred via cannula to a colourless solution of tolpzph (120 mg, 0.384 mmol) in dry THF (15 mL), resulting in a violet solution. The solution was stirred for 12 hours under an Ar (g) balloon, then transferred via cannula to an H-tube and subjected to vapour diffusion of dry, degassed Et 2 O (40 mL). After 1 week at room temperature, dark green, rod shaped, single crystals of 1·THF suitable for X-ray crystallography formed. These were filtered off and dried by repeated brief vacuum and refill (dry N 2(g) ) cycles (x5) on the Schlenk line (51 mg, 58.6 μmol, 31%). These crystals were stored under Ar (g)

S6
Additional information regarding solvent content of samples obtained as described above. When drying the compound using repeated brief vacuum and refill (dry N 2(g) ) cycles it is possible to remove some of the solvent of crystallisation. This can be avoided if sufficient care is taken. For three batches of 1·THF: TGA ∆mass (at 403 K, after 30 min): 8.60 %, 9.31 % and 8.27 % calc. for ·THF (78.1 g mol -1 ): 8.96%.
However, when a sample was instead kept under vacuum for approximately 10 min, TGA gave a ∆mass (at 403 K, after 30 min): 6.63 % calc. for 0·75THF (78.1 g mol -1 ): 6.72%. Please note that if a sample is excessively dried like this, then exposure to THF vapour for 2 hours, followed by 30 minute drying in air, can be used to re-solvate it and obtain 1·THF.

(NCS) 2 ]·CHCl 3 (1·CHCl 3 )
A sample of 1·THF (30 mg) was placed in a small sample tube (~5 cm 3 ). This sample tube was then placed in a large sample tube (~30 cm 3 ) containing CHCl 3 (5 mL) and the large vial capped. After 24 hours of vapour diffusion the THF of solvation had been replaced by CHCl 3 . Samples were then left exposed to air for 30 min before use in fractional solvent content experiments as described on page S7, or microanalysis and TGA, in order to allow any excess surface adsorbed solvent to evaporate. As there was only minimal surface solvent (for this vapour method, as compared to the samples isolated from solution), this air-drying was sufficient to ensure the sample dried without water damage as a result of the solvent becoming hydrated. Elemental analysis calc. for 1·CHCl 3

Details of instrumentation and techniques
Elemental analyses were measured at the Campbell Micro-analytical Laboratory, University of Otago. IR spectra were recorded on a Bruker ATR-IR spectrometer with an Alpha-P module. 1 H and 13 C NMR spectra were recorded on a 400 MHz Varian 400MR spectrometer at 298 K. Chemical shifts are referenced to residual solvent peaks (CDCl 3 : 1 H 7.26, 13 C 77.16 ppm). ESI mass spectra were recorded on a Bruker micrOTOF Q mass spectrometer in MeOH.
X-Ray crystallographic data for were collected using graphite monochromated Cu Kα radiation on an Oxford Diffraction SuperNova diffractometer equipped with a Cryostream N 2 open-flow cooling device, 3 at 100 and 273 K. Series of scans were performed in such a way as to collect a complete set of unique reflections to a maximum resolution of 0.80 Å. Raw frame data (including data reduction, inter-frame scaling, unit cell refinement and absorption corrections 4 ) for all structures were processed using CrysAlis Pro. 5 Structures were solved using SHELXS-2014 6 and refined using full-matrix least-squares on F 2 within the X-Seed graphical user interface. 7 All nonhydrogen atoms were refined with anisotropic displacement parameters. Hydrogen atoms were restrained in calculated positions using a riding model. Crystallographic data for the structures have been deposited with the Cambridge Crystallographic Data Centre, CCDC 1434103-1434104.
TGA data were collected on a TA Instruments TGA Q 50 . In each case the temperature was ramped at 1 K min -1 between 298 K and 403 K and then held constant until a total elapsed time of 150 min after which the mass was essentially constant (Δmass < 1% per hour).
Solid state visible reflectance data were recorded on a Perkin Elmer Lambda 950 UV/Vis/NIR spectrometer either in the range 400-1500 nm or in the range 400-800 nm, as indicated. Samples were attached to a solid support of Labsphere reflectance standard using double-sided tape (Sellotape® Double-Sided). Strong absorbance of this tape precluded measurement below 400 nm. See Figure . Variable temperature magnetic susceptibility measurements were carried out in house on a Quantum Design Versalab, Cryogen-free PPMS susceptometer equipped with a vibrating sample mount (VSM) in the temperature range 50-400 K. Data were corrected for the diamagnetism of the sample and a background correction for the sample holder was applied. The data were collected as indicated for each experiment, in either (a) 'settle mode' (5 K min -1 between data points which were recorded 1 min after the temperature was within the smaller value of 0.5 K or 0.5% of the target value) or (b) 'sweep mode' (scan rate of 1 K min -1 with data collected continuously). These data were also corrected for temperature lag between the sample and instrument thermometer. This was done by measuring the temperature difference between the heating and cooling runs (outside of the range of the SCO event) that gave the same magnetic susceptibility, and assigning the mean temperature to this susceptibility whilst also noting the temperature correction. Doing this at several different places resulted in a constant temperature correction so this was applied across the 50-300 K range.
Full methodology for determining the relationship between fractional solvent content and spin crossover behaviour: Samples weighing between 14 mg and 20 mg (2 samples for CHCl 3 , 6 samples for THF) were loaded into a PPMS sample holder and χ M T measured in the Versalab PPMS over the range 280→200→280 K at a scan rate of 1 K min -1 . The average of the temperature at which χ M T = 1.5 cm 3 mol -1 K in the cooling and heating directions was recorded and then a small sample (2-4 mg) was removed from the PPMS holder for TGA to determine the corresponding mass loss on heating it (see above) and therefore solvent content. The remaining sample in the PPMS holder was then returned to the Versalab PPMS and heated in the range 280→400→280 K at scan rates ranging from 1-5 K min -1 and with isothermal periods at 400 K in the range 0-120 min. The variation in time spent at 400 K was used to remove differing amounts of solvent from the sample in order to provide a wide range of data points. After this heating cycle χ M T was again measured in the range 280→200→280 K at a scan rate of 1 K min -1 , then another small sample (2-4 mg) taken for TGA. This process was then repeated for a total of 4 or 5 cycles per 14-20 mg sample (until there was no sample left). The molecular weight used in the data analysis was always that for 1·THF (870.8 g mol -1 ).
Refinement details for 1·THF: A crystal was mounted and a dataset collected at 273 K (mostly HS), then the temperature was dropped to 100 K (mostly LS) and a second dataset collected on the same crystal. At both temperatures there is one half of the complex and one half THF in the asymmetric unit (and two complexes and two THF molecules in the unit cell). The THF is disordered over two overlapping quarter occupancy sites, with all four carbon atoms common to S8 both positions (hence half occupancy). PART -1 contains all four half occupancy C atoms and one quarter occupancy O atom (O35), whilst PART -2 contains the other one quarter occupancy O atom (O45). H atoms were not included on this disordered THF. DFIX was used to restrain the disordered THF geometries. Unsurprisngly, the U values of this THF were far better at 100 K than at 273 K.

S16
Temperature dependence of the unit cell parameters for 1•THF Figure S14. Temperature dependence of the unit cell a-axis. The error bars shown represent a 95% confidence interval.

S18
Electronic absorbtion spectra Figure S18.Solid state reflectance UV-vis spectra of forest green 1·THF (as prepared) and violet 1 (after drying in a N 2(g) stream at 403K for 1 h) in the range 400-1500 nm. Note: the data is presented as collected; no correction for the tape has been made. Figure S19. Solid   The sample on the right was initially treated in the same way, then exposed to THF vapour for 4 hours to regenerate 1·THF (shown) -this is the sample that was used to give the blue, pink and green lines in Figure S19. Figure S29. The same data as shown in Figure 3, but with all calculations done using MW(1·THF).