Molecular Design of a Metal-Nitrosyl Ferroelectric with Reversible Photoisomerization

The development of photo-responsive ferroelectrics whose polarization may be remotely controlled by optical means is of fundamental importance for basic research and technological applications. Herein, we report the design and synthesis of a new metal-nitrosyl ferroelectric crystal (DMA)(PIP)[Fe(CN)5(NO)] (1) (DMA = dimethylammonium, PIP = piperidinium) with potential phototunable polarization via a dual-organic-cation molecular design strategy. Compared to the parent non-ferroelectric (MA)2[Fe(CN)5(NO)] (MA = methylammonium) material with a phase transition at 207 K, the introduction of larger dual organic cations both lowers the crystal symmetry affording robust ferroelectricity and increases the energy barrier of molecular motions, endowing 1 with a large polarization of up to 7.6 μC cm–2 and a high Curie temperature (Tc) of 316 K. Infrared spectroscopy shows that the reversible photoisomerization of the nitrosyl ligand is accomplished by light irradiation. Specifically, the ground state with the N-bound nitrosyl ligand conformation can be reversibly switched to both the metastable state I (MSI) with isonitrosyl conformation and the metastable state II (MSII) with side-on nitrosyl conformation. Quantum chemistry calculations suggest that the photoisomerization significantly changes the dipole moment of the [Fe(CN)5(NO)]2– anion, thus leading to three ferroelectric states with different values of macroscopic polarization. Such optical accessibility and controllability of different ferroelectric states via photoinduced nitrosyl linkage isomerization open up a new and attractive route to optically controllable macroscopic polarization.


Experimental Synthesis
All chemicals were commercially available and used without further purification. Synthesis of Ag 2 [Fe(CN) 5 NO] was reported by E. Reguera et al. [1] Compound 1 was prepared by a reaction of Ag 2 [Fe(CN) 5 NO] (10 mmol) and dimethylamine hydrochloride (10 mmol) and piperidine hydrochloride (10 mmol) in 8 mL of deionized water. After the AgCl precipitate was filtrated, reddish-brown flaky crystals were obtained after the solution evaporated for several days. Yield: 87% based on Ag 2 [Fe(CN) 5

X-ray Crystallographic Analysis
The in-situ variable-temperature single-crystal diffraction intensities data were collected on a Bruker Smart APEX diffractometer equipped with Mo Kα sealed tube ( = 0.71073 Å). The APEX3 software package was used for data collection, cell refinement, and data reduction. Using Olex 2 program, [2] the structures were solved by using Intrinsic Phasing with the SHELXT structure solution program and using full-matrix least-squares method with the SHELXL refinement program. [3] Non-hydrogen atoms were refined anisotropically and the positions of the hydrogen atoms were generated geometrically. The crystal data and structure refinement results for 1 are listed in Table S1. Powder X-ray diffraction (PXRD) patterns (Cu-K  ,  = 1.54184 Å) were collected on Panalytical Empyrean with Cu-Kα X-ray radiation (40 kV, 45 mA).

Elemental analysis
Elemental analyses for C, N, and H were performed with a Truspec Micro CHNS 630-200-200 elemental analyzer.

Thermal Analysis
Differential Thermal Analysis (DTA) and thermogravimetric analysis (TGA) were carried out on a Hitachi NEXTA STA300 with a heating rate of 10 K min -1 from 298 to 673 K under a nitrogen atmosphere. Differential scanning calorimetry (DSC) was carried out on a TA DSC Q2000 instrument under a nitrogen atmosphere in aluminum crucibles with heating and cooling rates of 10 K min -1 from 195 to 405 K.

Dielectric and P−E hysteresis loop measurements.
The dielectric measurements were carried out on a Keysight E4990A impedance analyzer at 16 frequencies from 500 Hz to 2 MHz, with an applied voltage of 1.0 V and a temperature sweeping rate of 3 K min -1 approximately in the range of 80-450 K in a Mercury iTC cryogenic environment controller of Oxford Instrument. The powder sample of 1 was ground and pressed into tablets under a pressure of around 5 GPa. The pressed-powder pellets were deposited with a magnetic sheet used as an electrode. P-E hysteresis loops measurements of 1 were done on ~2.0 x 1.5 x 1.5 mm 3 sized single crystal sample with silver paste electrodes by using a TF analyzer (TFA-1000).

SHG Measurement.
Variable-temperature SHG experiment was executed by Kurtz-Perry powder SHG test using an Nd:YAG laser (1064 nm) with an input pulse of 570 V under a programmable cryogenic cooling system.

Infrared (IR) spectroscopy.
KBr pellets of of compound 1 were mounted in a liquid nitrogen-cooled cryostat to allow laser irradiation without changing the optical geometry. The samples were irradiated with 405 nm light from a 25-mW continuous diode laser (BH, New York during 60 min. The spectra of the samples were obtained at increasing temperatures, from 77 K to 280 K, on a Bruker V70 FTIR spectrometer with an MCT wide band detector. In the investigation of photoisomerization reversibility, the spectra were recorded after alternating irradiation of the sample with 405 nm for 50 minutes followed by irradiation at 800 nm for 5 min (800-mW power Ti: sapphire laser). The spectra were recorded between 4000 cm -1 and 1200 cm -1 , with 4 cm -1 resolution (cut-off of the cryostat windows at 1000 cm -1 ).

PFM measurements.
Imaging of the topography, domain configurations with high spatial resolution and ferroelectric

Theoretical calculation.
Quantum chemical calculations were performed in the HyperChem 7.01 package using a semi-empirical PM3 method including the restricted Hartree-Fock approximation. Convergence limit was set 10 -8 , iteration limit 32767. The calculations were performed for unit cell of 1 ( Figure S1a,b) and individual anions and cations ( Figure S1c-e). The atomic positions corresponding to the unit cell of 1 in the ground state were experimentally determined at 100 K by single crystals X-ray diffraction. The unit cell was first generated using CCDC Mercury software [4] and then transferred to HyperChem for calculations ( Figure   S1a,b). The metastable states were obtained by manual modification of nitroprusside anions in the same unit cell. The bond lengths and angles of nitroprusside anion in metastable phases were taken from reference. [5] Only the configuration of nitroprusside anions was changed; the positions of other molecules were constrained. No additional geometry optimization of the modified unit cell was done.
The calculations provided the dipole moments of the unit cell. The unit cell polarization, P [C/m 2 ], was determined using equation: P = 3.33556255×D/V, where D is the total dipole moment in Debye, and V is the unit cell volume in Å 3 determined from the X-ray analysis (Table S1). The numerical coefficient represents the combination of conversion factors to SI units. The obtained dipole moments and polarizations are presented in Table S1.   Scheme S1. The structural formula of 1. Figure S2. The powder XRD patterns confirmed the phase purity of the as-synthesized sample 1.   K for 5 minutes, that is, above the temperature of the transition to HTP. The image reveals a system of alternating ridges and furrows evidencing sample twinning upon transition from orthorhombic to monoclinic crystal structures. Corrugation of the surface appears due to opposite signs of the spontaneous strain (monoclinic distortions of the higher-symmetry orthorhombic structure) in adjacent ferroelastic domains on cooling from the annealing temperature. The monoclinic distortion at the HTP-to-ITP transition is significantly larger than at the ITP-to-LTP transition. As a result, the clapping angle, that is, the angle needed to turn one of the adjacent ferroelastic domains towards the other to preserve the material integrity is much larger at the HTP-to-ITP transition than at the ITP-to-LTP transition. [6] As can be determined from the profile of the surface corrugation in the image, the clapping angle between adjacent ferroelastic domains is between about 10° and 14°, while the values calculated with the use of the data of Table S1 yield 12.7° for the complete HTP-to-LTP transition. In turn, for the ITP-to-LTP transition, the clapping angle is 3.8°. This value would result in a topography with a corrugation amplitude comparable or below the surface feature height seen in the topography images in Figure 5 of the main text, which makes it difficult to reveal twinning in the topography PFM images displayed in the main text. PFM images after application of +60 V. Image size 60x60 µm 2 . S12 Figure S11. FTIR spectrum of 1 before and after 405 nm light irradiation at the specified temperatures.