Coordination Polymer Flexibility Leads to Polymorphism and Enables a Crystalline Solid–Vapour Reaction: A Multi-technique Mechanistic Study

Despite an absence of conventional porosity, the 1D coordination polymer [Ag4(O2C(CF2)2CF3)4(TMP)3] (1; TMP=tetramethylpyrazine) can absorb small alcohols from the vapour phase, which insert into Ag–O bonds to yield coordination polymers [Ag4(O2C(CF2)2CF3)4(TMP)3(ROH)2] (1-ROH; R=Me, Et, iPr). The reactions are reversible single-crystal-to-single-crystal transformations. Vapour-solid equilibria have been examined by gas-phase IR spectroscopy (K=5.68(9)×10−5 (MeOH), 9.5(3)×10−6 (EtOH), 6.14(5)×10−5 (iPrOH) at 295 K, 1 bar). Thermal analyses (TGA, DSC) have enabled quantitative comparison of two-step reactions 1-ROH→1→2, in which 2 is the 2D coordination polymer [Ag4(O2C(CF2)2CF3)4(TMP)2] formed by loss of TMP ligands exclusively from singly-bridging sites. Four polymorphic forms of 1 (1-ALT, 1-AHT, 1-BLT and 1-BHT; HT=high temperature, LT=low temperature) have been identified crystallographically. In situ powder X-ray diffraction (PXRD) studies of the 1-ROH→1→2 transformations indicate the role of the HT polymorphs in these reactions. The structural relationship between polymorphs, involving changes in conformation of perfluoroalkyl chains and a change in orientation of entire polymers (A versus B forms), suggests a mechanism for the observed reactions and a pathway for guest transport within the fluorous layers. Consistent with this pathway, optical microscopy and AFM studies on single crystals of 1-MeOH/1-AHT show that cracks parallel to the layers of interdigitated perfluoroalkyl chains develop during the MeOH release/uptake process.


Crystal Structure of 3
Compound 3 has been synthesized as a colourless crystalline solid by diffusion between two acetonitrile solutions of silver(I) heptafluorobutanoate and TMP, respectively (see Experimental section). The coordination polymer forms a 1D zig-zag tape, in which silver atoms have a distorted tetrahedral geometry. Each Ag(I) centre is coordinated to one chelating heptafluorobutanoate and two TMP ligands ( Figure S1). Thus, single TMP ligands link the Ag(O 2 C(CF 2 ) 2 CF 3 ) units, forming a coordination polymer which extends in the [001] direction. The 1D zig-zag geometry of the S4 coordination polymer arises from the distorted tetrahedral coordination geometry of the Ag(I) ion, with an NAgN angle of 133.48(7) °. The perfluoroalkyl chains are interdigitated in the (100) plane.  Figure S1.

Powder X-ray Diffraction
Phase purity of 1-MeOH. A polycrystalline sample of 1-MeOH was lightly ground in an agate mortar and pestle and loaded into a 0.7 mm borosilicate glass capillary prior to being mounted and aligned on a Bruker-AXS D8 Advance powder diffractometer operating with Ge-monochromated Cu-K α1 radiation (λ = 1.54056 Å). Powder patterns were measured at a scan rate no faster than 1 º/min in the range 4 ≤ 2θ ≤ 40 º. The powder pattern was indexed using the program TOPAS. S3 A unit cell was found corresponding to crystal structure of 1-MeOH already established from single crystal diffraction. A Pawley refinement, S4 conducted using TOPAS, was implemented, confirming the phase purity of 1-MeOH. Pawley refinement converged to R wp of 0.0518 , R wp ' = 0.1334 (R wp ' is the background subtracted R wp ); see Figure S3. Phase purity of 1-EtOH. White microcrystalline 1-EtOH, product of the solution phase synthesis, was loaded into a 0.7 mm borosilicate capillary and X-ray diffraction data were collected (λ= 0.799993(8) Å) at station ID31 S5 at the European Synchrotron Radiation Source (ESRF) using a 9channel multi-analyser crystal (MAC) detector. All data were collected at room temperature. The powder pattern was indexed using the program TOPAS. A unit cell was found corresponding to crystal structure of 1-EtOH already established from single crystal diffraction. The starting model used for Rietveld refinement, conducted using TOPAS, was the single crystal structure of 1-EtOH. The model for the structure was refined with one global isotropic thermal parameter. A 6 th order spherical harmonic correction of the intensities for preferred orientation was applied in the final stage of refinement. Rietveld refinement converged to R wp of 0.14630 , R wp ' = 0.23936 (R wp ' is the background subtracted R wp ); see Figure S4. Phase purity of 1-iPrOH. White microcrystalline 1-iPrOH, product of the solid-vapour synthesis from 1 and iPrOH, was loaded into a 0.7mm borosilicate capillary and X-ray diffraction data were collected (λ= 0.826741(1) Å) at beamline I11 at Diamond Light Source, S6 equipped with a wide angle (90 °) PSD detector comprising 18 Mythen-2 modules. S7 A series of 14 pairs of scans were S6 conducted at room temperature, each pair related by a 0.25 ° detector offset to account for gaps between detector modules. The resulting 28 patterns were summed to give the final pattern for structural analysis. The powder pattern was indexed using the program TOPAS. A unit cell was found corresponding to the known crystal structure of 1-iPrOH. The starting model used for Rietveld refinement, conducted using TOPAS, was the single crystal structure of 1-iPrOH. This model was refined with one global isotropic thermal parameter. A 6 th order spherical harmonic correction of the intensities for preferred orientation was applied in the final stage of refinement. Rietveld refinement converged to R wp of 0.10456, R wp ' = 0.18195 (R wp ' is the background subtracted R wp ); see Figure S5.  Figure S6. In situ X-ray powder diffraction study of heating of 1-MeOH, adapted from reference S2. The original report identified the two polymorphs of 1 as 1 and 1 HT , rather than 1-A HT and 1-B HT used here.

Solid-state reaction transforming 1-EtOH → 1-A HT + 1-B HT
The gaps in the patterns that are not included in the fits result from the data collection method using the PSD detector on beamline I11. When this detector was first used by us at beamline I11, the data acquisition software did not provide an automated means to collect offset patterns and sum these to eliminate the gaps between modules. However, current data collection strategy and software can avoid this, as seen for the study of 1-iPrOH.
Compound 3, present as an impurity from the synthesis of 1-EtOH, appears unchanged over the course of the loss of EtOH by 1-EtOH. The gradual, but small, increase in percentage composition of 3 is attributed to some loss in crystallinity during the heating experiment. The contribution of amorphous material to the overall composition has not been determined quantitatively. After 20 minutes at 340K   After 60 minutes at 340K After 80 minutes at 340K

Gas-phase Fourier-transform infra-red spectroscopy (FTIR)
Fourier transform infrared (FTIR) spectroscopic experiments were conducted using a doublewalled 10 cm glass IR absorption cell fitted with either KCl or KBr windows and a suspended sample container (Figure S19) as previously described. S8 To obtain the calibration line for methanol, ethanol and isopropanol vapours, gas-phase IR spectra in the region of 400-4000 cm −1 were acquired using a FTIR spectrometer (Perkin-Elmer Paragon 1000, resolution 1 cm −1 , no apodization). The spectrometer was operated in the single-beam mode, that is, sample and background (empty cell) spectra were recorded separately. The calibration for the FTIR experiment was performed by introducing a known amount of dry ROH (ROH = MeOH, EtOH and i PrOH), which was previously de-gassed by freeze-pump-thaw methods, into an empty gas cell (evacuated under vacuum) and acquiring IR spectra at 22 ºC (range of pressures used for calibration: 5-30 Torr). The absorbance is given as lg(I 0 /I). The area under the methanol C-O stretching absorption band was integrated from 950 to 1100 cm −1 after background subtraction and baseline correction to determine the partial pressure of methanol ( Figures S20 and S21). For ethanol and isopropanol the areas under the C-O stretching and C−H bending absorption bands, which overlap, was integrated from 950 to 1175 cm −1 and 900 to 1000 cm −1 , respectively, after background subtraction and baseline correction (Figures S22-S25). S12 Figure S19. Double-walled 10 cm glass IR absorption cell fitted with KCl windows.

Gas phase methanol FTIR
Dry methanol at pressures of 5, 10, 15, 20 and 30 Torr was used to determine the calibration line. The integrated absorbance was measured by integrating the area of the peaks in the region of 950-1100 cm -1 (C−O stretching band, Figure S20) Figure S20. Methanol C−O stretching band in the gas phase including P and R branches of the rotational fine structure at a partial pressure of 30 Torr.
One hundred spectra were measured for each pressure. The final calibration line has R 2 = 0.9999 corresponding to Integrated Absorbance (cm -1 ) = 4.1826  Pressure (bar) ( Figure S21).

Gas phase ethanol FTIR
Dry ethanol at pressures at 5, 10, 15, 20 and 30 Torr was used to determine the calibration line. The integrated absorbance was measured by integrating the area of the peaks in the region of 950-1175 cm -1 (C-O stretching and C-H bending bands, Figure S22 One hundred spectra were measured at each pressure. The final calibration line has R 2 = 0.9998 corresponding to Integrate Absorbance (cm -1 ) = 4.8512  Pressure (Torr) ( Figure S23).

Gas phase Isopropanol FTIR
Dry isopropanol at pressures of 5, 10, 15, 20 and 25 Torr was used to determine the calibration line. The integrated absorbance was measured by integrating the area of the peaks in the region of 900-1000 cm -1 (C−O stretching and C-H bending bands, Figure S24). One hundred spectra were measured at each pressure. The final calibration line has R 2 = 0.9972 corresponding to Integrated Absorbance (cm -1 ) = 0.643  Pressure (Torr) ( Figure S25).
-S16 Figure S25. Isopropanol pressure calibration line The equilibria linking the crystalline solids 1 and 1-ROH and the vapour ROH are: (1-ROH) In order to calculate the equilibrium constant a series of equations were used. Equation (1) defines the equilibrium constant, where a 1 is the activity of compound 1; f ROH is the fugacity of the alcohol released and, a 1-ROH is the activity of the alcohol-containing coordination polymer 1-ROH. Equation (2) defines the equilibrium constant in pressure terms, where p ROH is the equilibrium pressure of the alcohol and φ ROH is the fugacity coefficient, which is defined in equation (3). As the final alcohol pressures were very low (see Table 2), it can be supposed that the alcohols behave as perfect gases, so the fugacity coefficient is unity. In order to study the activity of the coordination polymers, an independent methanol release experiment was done, doubling the amount of initial 1-MeOH coordination polymer. The equilibrium pressure (p MeOH = 0.00753(6) bar) is similar to the experiment done with only 50 mg of starting 1-MeOH coordination polymer. It can be concluded that the equilibrium pressure is independent of the amount of starting material. As a consequence, the activity coefficients of the two solid components of the reaction can be assumed to be unity. K eq = a 1 x (f ROH /p ө ) 2 / (a 1-ROH ) K p = (f ROH /p ө ) 2 = (φ ROH x p ROH /p ө ) 2 Isopropanol calibration line S17 The equilibrium constant and the Gibbs free energy for each of the three equilibria are provided in Table 4.

TGA and DSC traces
a) S18 b) c) Figure S26. TGA-DSC analysis for (a) 1-MeOH, (b) 1-EtOH and (c) 1-iPrOH. See Table 1 for numerical data and associated text for discussion of these experimental data.