Time-resolved X-ray crystal structure analysis for elucidating the hidden 'over-neutralized' phase of TTF-CA

Manabu Hoshino, Shunsuke Nozawa, Tokushi Sato, Ayana Tomita, Shin-ichi Adachi, Shin-ya Koshihara
2013 RSC Advances  
General experiment. Black suitable single crystals of TTF-CA were grown by the cosublimation of TTF and CA. The single crystal of dimensions 0.13 × 0.10 × 0.07 mm was used for the experiment. All diffraction data were collected using a marCCD X-ray detector on beamline NW14A of the Photon Factory Advanced Ring (PF-AR) at the High Energy Accelerator Research Organization (Tsukuba, Japan). Monochromatic X-ray pulses (λ= 0.56366 Å; Si(111) double crystals) were synchronized with femtosecond laser
more » ... femtosecond laser pulses. For synchronization, the frequency of the X-rays from the PF-AR was reduced to 946 Hz using an X-ray pulse selector. The sample temperature was controlled with a cold nitrogen stream from an Oxford cryogenic system. Photograph for the experimental setup is shown in Supplementary Figure S2 . Time-resolved X-ray diffraction experiments were performed at 90 K. Light-on and light-off diffraction data frames were alternately collected to minimize systematic errors. Under the present laser power conditions (< 10 mJ cm −2 ), no significant photoinduced appearance of 0k0 (k = odd) reflections was recorded at any delay time. The reflections are absent in the N phase (the suggested crystallographic space group: P2 1 /n) but are present in the I phase (Pn). The absence of the reflections indicated that the photoinduced NI transition was deactivated within a much shorter time than the pulse width of the X-ray radiation. Additionally, static X-ray diffraction experiments without laser irradiation were performed at 100 K. All crystal structures were solved by direct methods (SHELXS-97) and refined by the full-matrix least-squares method (SHELXL-97) (Ref. S1). All the non-hydrogen atoms were refined anisotropically. All the hydrogen atoms were refined isotropically using U iso = 1.2U eq of the connected carbon atom. Sample quality was checked by drawing a photo-difference Fourier map at the first delay point (Δt = 150 ps). If a poor-quality crystal was selected for measurement, the noise of the map increased and became high. One of the results is shown in Supplementary Figure S3 . Although the height and depth of the peak densities were almost the same as those described in the main manuscript (ca. ±0.25 eÅ −3 ), the noise level was almost the same as that of the peaks, indicating that the peak height and depth depended on the excitation power and data fluctuation because the sample quality was reflected in the noise density of the map. General crystallographic data. Refined formula: C 6 H 4 S 4 and C 6 O 2 Cl 4 , formula weight (M r ): 450.19, crystal system: monoclinic, Z = 2. Crystallographic data at Δt = 150 ps (On150ps) and its light-off condition Electronic Supplementary Material (ESI) for RSC Advances This journal is (Off150ps). On150ps: 11104 unique reflections merged from 46260 recorded ones (3.77 < θ < 33.56°) were used for structural analysis (R int = 0.0330). Lattice parameters, R-factor on F 2 > 2σ(F 2 ), and weighted R-factor are follows: a = 7.2140 (1) Å, b = 7.5870(1) Å, c = 14.4850(1) Å, β = 99.1310(2)°, V = 782.756(16) Å 3 , R = 0.0209, wR = 0.0593, S = 1.081. Calculated density is 1.910. Linear absorption coefficient (µ) is 0.654. Residual electron density (max/min) is 0.721/-0.413 eÅ −3 . Off150ps: 11103 unique reflections merged from 46254 recorded ones (3.77 < θ < 33.55°) were used for structural analysis (R int = 0.0326). Lattice parameters, R-factor on F 2 > 2σ(F 2 ), and weighted R-factor are follows: a = 7.2130(1) Å, b = 7.5860(1) Å, c = 14.4840(1) Å, β = 99.1270(3)°, V = 782.499(16) Å 3 , R = 0.0206, wR = 0.0587, S = 1.082. Calculated density is 1.911. Linear absorption coefficient (µ) is 0.654. Residual electron density (max/min) is 0.701/-0.487 eÅ −3 . Crystallographic data at Δt = 500 ps (On500ps) and its light-off condition (Off500ps). On500ps: 11111 unique reflections merged from 46185 recorded ones (3.77 < θ < 33.55°) were used for structural analysis (R int = 0.0342). Lattice parameters, R-factor on F 2 > 2σ(F 2 ), and weighted R-factor are follows: a = 7.2140(1) Å, b = 7.5870(1) Å, c = 14.4850(1) Å, β = 99.1300(3)°, V = 782.758(16) Å 3 , R = 0.0209, wR = 0.0595, S = 1.089. Calculated density is 1.910. Linear absorption coefficient (µ) is 0.654. Residual electron density (max/min) is 0.737/-0.443 eÅ −3 . Off500ps: 11099 unique reflections merged from 46232 recorded ones (3.77 < θ < 33.54°) were used for structural analysis (R int = 0.0329). Lattice parameters, R-factor on F 2 > 2σ(F 2 ), and weighted R-factor are follows: a = 7.2130(1) Å, b = 7.5860(1) Å, c = 14.4840(1) Å, β = 99.1250(3)°, V = 782.503(16) Å 3 , R = 0.0210, wR = 0.0603, S = 1.097. Calculated density is 1.911. Linear absorption coefficient (µ) is 0.654. Residual electron density (max/min) is 0.723/-0.457 eÅ −3 . Crystallographic data at Δt = 800 ps (On800ps) and its light-off condition (Off800ps). On800ps: 11101 unique reflections merged from 46028 recorded ones (3.77 < θ < 33.55°) were used for structural analysis (R int = 0.0326). Lattice parameters, R-factor on F 2 > 2σ(F 2 ), and weighted R-factor are follows: a = 7.2140(1) Å, b = 7.5870(1) Å, c = 14.4850(1) Å, β = 99.1300(3)°, V = 782.758(16) Å 3 , R = 0.0207, wR = 0.0583, S = 1.078. Calculated density is 1.910. Linear absorption coefficient (µ) is 0.654. Residual electron density (max/min) is 0.698/-0.513 eÅ −3 . Off800ps: 11112 unique reflections merged from 46021 recorded ones (3.77 < θ < 33.57°) were used for structural analysis (R int = 0.0323). Lattice parameters, R-factor on F 2 > 2σ(F 2 ), and Electronic Supplementary Material (ESI) for RSC Advances This journal is © The Royal Society of Chemistry 2013 4 weighted R-factor are follows: a = 7.2120(1) Å, b = 7.5860(1) Å, c = 14.4830(1) Å, β = 99.1270(3)°, V = 782.336(16) Å 3 , R = 0.0207, wR = 0.0595, S = 1.081. Calculated density is 1.911. Linear absorption coefficient (µ) is 0.655. Residual electron density (max/min) is 0.736/-0.484 eÅ −3 . Crystallographic data at Δt = 1 ns (On1ns) and its light-off condition (Off1ns). On1ns: 11106 unique reflections merged from 46004 recorded ones (3.77 < θ < 33.56°) were used for structural analysis (R int = 0.0327). Lattice parameters, R-factor on F 2 > 2σ(F 2 ), and weighted R-factor are follows: a = 7.2140 (1) Å, b = 7.5870(1) Å, c = 14.4850(1) Å, β = 99.1320(3)°, V = 782.753(16) Å 3 , R = 0.0207, wR = 0.0587, S = 1.081. Calculated density is 1.910. Linear absorption coefficient (µ) is 0.654. Residual electron density (max/min) is 0.701/-0.505 eÅ −3 . Off1ns: 11084 unique reflections merged from 46099 recorded ones (3.77 < θ < 33.55°) were used for structural analysis (R int = 0.0316). Lattice parameters, R-factor on F 2 > 2σ(F 2 ), and weighted R-factor are follows: a = 7.2120(1) Å, b = 7.5860(1) Å, c = 14.4830(1) Å, β = 99.1270(3)°, V = 782.336(16) Å 3 , R = 0.0204, wR = 0.0581, S = 1.087. Calculated density is 1.911. Linear absorption coefficient (µ) is 0.655. Residual electron density (max/min) is 0.668/-0.450 eÅ −3 . Crystallographic data at 100 K (Heated). 11130 unique reflections merged from 46042 recorded ones (3.77 < θ < 33.55°) were used for structural analysis (R int = 0.0324). Lattice parameters, R-factor on F 2 > 2σ(F 2 ), and weighted R-factor are follows: a = 7.2220(1) Å, b = 7.5910(1) Å, c = 14.4930(1) Å, β = 99.1370(2)°, V = 784.457(16) Å 3 , R = 0.0213, wR = 0.0600, S = 1.089. Calculated density is 1.906. Linear absorption coefficient (µ) is 0.653. Residual electron density (max/min) is 0.793/-0.390 eÅ −3 . Detailed method for drawing the photo-difference Fourier maps. To draw photoand thermal-difference Fourier maps, changes of the overall temperature factor (ΔB) were estimated by the Wilson-type plot.(Ref. S2) The results of the Wilson-type plots are shown in Supplementary S4. The slope of the liner fit to the Wilson-type plot corresponds to ΔB/2. The obtained ΔB values were used as a correction term to F o(light-off) and F o(90 K) in the following equation: F' o = k exp(−ΔBsin 2 θ/λ 2 ) F o where F' o is the corrected F o and k is the overall scale factor. The ΔB values are follows: 0.0254 (for F o(light-off) at Δt = 150 ps), 0.0101 (Δt = 500 ps), 0.0131 (Δt = 800 Electronic Supplementary Material (ESI) for RSC Advances This journal is © The Royal Society of Chemistry 2013 5 ps), 0.0221 (Δt = 1 ns), and 0.0714 (for F o(90 K) ). Photoinduced geometrical changes in a few percent of the species in the crystal were reflected in this map (Ref. 23 in the main manuscript and Ref. S3). All F o sets were obtained as an output of crystal structure refinement. The difference Fourier synthesis was performed using the usual Fourier synthesis equation with (light-off) as its coefficient. The lattice parameters and phases at the light-off condition were used in the synthesis. Based on those map, population of photoinduced species generated in the crystal was estimated by refinement of a disordered model. The positions and displacement parameters of the initial molecule was constrained and the positions of the disordered Cl19 and Cl25, which showed displacement change at all delay points, were refined with same displacement parameters as the initial one (using the same method as Ref. 23). From that analysis, about 1% was estimated as population of photo-converted species. The C15=O18 and C21=O24 bonds were shortened upon reaching the over-neutralized phase (at Δt = 800 ps). The shrinks were respectively 1.226(4) à 1.222(13) Å and 1.221(4) à 1.217(13) Å. These findings suggested a 4 × 10 −3 Å shortening of the C=O bond, but the standard uncertainty associated with these values was larger than their differences because of the very small population of photoinduced species (1%). If the ρ value became 0.00 by the over-neutralization, the C=O shortening would be approximately 8 × 10 −3 Å (estimated from the theoretically reported C=O lengths at ρ = 0.25 and 0.00 in Ref. S4). We therefore considered that the observed C=O shortening was qualitative in nature. All C-Cl bonds were also shortened accompanying with the over-neutralization. Two-dimensional photo-difference Fourier maps representing C=O and C-Cl shortenings were shown in Supplementary Figure S5 . The shrinks of C16-Cl19, C17-Cl20, C22-Cl25 and C23-Cl26 at Δt = 800 ps were respectively 1.721(3) à 1.711 (13) Å, 1.698(3) à 1.671(13) Å, 1.694(3) à 1.687(13) Å and 1.713(3) à 1.781(13) Å. In theoretical calculations, the 7 × 10 −3 Å shortening of C-Cl with decreasing of ρ from 0.25 to 0.00 was reported (Ref. 30). The shortenings of C16-Cl19 and C22-Cl25 were also within experimental errors, but qualitatively represent the over-neutralization of TTF-CA. The shortenings of C17-Cl20 and C23-Cl26 relatively significant compared to others. The reason of it is guessed the atomic fluctuation shown at Δt = 150 ps and 1 ns was somewhat affected to the positions of Cl20 and Cl26. The observed structural change by over-neutralization was very small and not accompanying with change of the extinction rule. Therefore, discuss about intensity change of some selected reflections would be an arbitrary way. On the other hand, photo-difference Fourier map is based on intensity change of all recorded reflections. Electronic Supplementary Material (ESI) for RSC Advances This journal is
doi:10.1039/c3ra42489h fatcat:cpc33ydhfrdyfpaei6fxgm52oe