Structural characterization, thermal properties, and molecular motions near the phase transition in hybrid perovskite [(CH2)3(NH3)2]CuCl4 crystals: 1H, 13C, and 14N nuclear magnetic resonance

The structural characterization of the [(CH2]3(NH3)2]+ cation in the perovskite [(CH2)3(NH3)2]CuCl4 crystal was performed by solid-state 1H nuclear magnetic resonance (NMR) spectroscopy. The 1H NMR chemical shifts for NH3 changed more significantly with temperature than those for CH2. This change in cationic motion is enhanced at the N-end of the organic cation, which is fixed to the inorganic layer by N–H···Cl hydrogen bonds. The 13C chemical shifts for CH2-1 increase slowly without any anomalous change, while those for CH2-2 move abruptly compared to CH2-1 with increasing temperature. The four peaks of two groups in the 14N NMR spectra, indicating the presence of a ferroelastic multidomain, were reduced to two peaks of one group near TC2 (= 333 K); the 14N NMR data clearly indicated changes in atomic configuration at this temperature. In addition, 1H and 13C spin–lattice have shorter relaxation times (T1ρ), in the order of milliseconds because T1ρ is inversely proportional to the square of the magnetic moment of paramagnetic ions. The T1ρ values for CH2 and NH3 protons were almost independent of temperature, but the CH2 moiety located in the middle of the N–C–C–C–N bond undergoes tumbling motion according to the Bloembergen–Purcell–Pound theory. Ferroelasticity is the main cause for the phase transition near TC2.


Scientific Reports
| (2020) 10:20853 | https://doi.org/10.1038/s41598-020-77931-0 www.nature.com/scientificreports/ The organic chains are extended along the a direction. The organic chains NH 3 -CH 2 -CH 2 -CH 2 -NH 3 are almost identical, and the skeleton N-C-C-C-N is planar. Above 434 K, the symmetry is monoclinic with space group B2/m and lattice constants a = 7.309 Å, b = 8.866 Å, c = 7.614 Å, α = 95.365°, and Z = 2 22 . The lattice constants a and c in the monoclinic structure are comparable with those in the room temperature structure, whereas the b parameter in the monoclinic structure is half of that in the room temperature structure. According to the previously reported, the Phelps et al. 21 and Czupinski et al. 23 determined the structural phase transition for (CH 2 ) 3 (NH 3 ) 2 CuCl 4 . And, the structural, dielectric, and conductive properties were discussed by Mostafa et al. 20 . In addition, the structural phase transition was analysed by x-ray and optical studies 22 , where ferroelastic multidomain walls were observed in the orthorhombic phase. Iqbal et al. 1 reported Raman scattering results at various temperatures above and below the respective magnetic ordering temperature (149 K) and in a magnetic field up to 10 kg. The crystal structure, magnetic and optical properties have been studied by only a few researchers. In addition, the thermal properties, the structural and molecular dynamics of the [(CH 2 ) 3 (NH 3 ) 2 ] CuCl 4 crystal have not been studied in detail.
Here, differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA) experiments were performed to provide a better understanding of the phase transition temperatures and thermal properties of [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 . In addition, the structural characterizations of the [(CH 2 ) 3 (NH 3 ) 2 ] cation were studied in detail by magic angle spinning (MAS) nuclear magnetic resonance (NMR) and static NMR methods. The temperature dependences of the chemical shifts and spin-lattice relaxation times T 1ρ were measured by 1 H MAS NMR and 13 C cross-polarization (CP)/MAS NMR to highlight the role of the cation in [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 . In addition, 14 N static NMR spectra of [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 single crystals were acquired. Based on these results, the structural characterizations for NH 3 -CH 2 -CH 2 -CH 2 -NH 3 are discussed as a function of temperature. In particular, the hydrogen bonding of the N-H···Cl between the Cu-Cl layer and the alkylammonium chain within the [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 is expected to give important information regarding the fundamental mechanisms that enable various potential applications.
The structure of the [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 crystal at 300 K was analysed using an X-ray diffraction system equipped with a Cu-Kα radiation source at the KBSI, Seoul Western Center. DSC (TA Instruments, DSC 25) was conducted at a heating rate of 10 °C/min from 190 to 600 K under nitrogen gas. TGA was performed using a thermogravimetric analyser (TA Instruments) from 300 to 680 K at the same heating rate. The sample weights used for DSC and TGA experiments were 6.23 and 7.53 mg, respectively. Optical observations were performed using an optical polarized microscope in the temperature range of 300-600 K, where the as-grown crystals were placed on a Linkam THM-600 heating stage.
NMR spectra of [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 crystals were obtained using a 400 MHz Avance II + Bruker solid-state NMR spectrometer, equipped with 4 mm CP/MAS probes (at the KBSI, Seoul Western Center). The Larmor frequencies to 1 H MAS NMR and 13 C CP/MAS NMR experiments were at ω 0 /2π = 400.13 and 100.61 MHz, respectively. A MAS rate of 10 kHz was used to minimize the spinning sideband. The NMR chemical shifts were recorded using tetramethylsilane (TMS) as the standard. The T 1ρ values were measured using a π/2 − t sequence by changing the spin-locking pulses, and the width of the π/2 pulse was 3.3 μs. The spin-lock power on the 1 H and 13 C channel was 75.76 kHz. The 13 C T 1ρ values were obtained by changing the duration of the 13 C spin-locking  . NMR data could not be obtained because the NMR spectrometer could not operate at temperatures above 430 K. The true temperature at spinning condition of 10 kHz was adjusted based on the sample temperature, suggested by Guan and Stark 24 . The temperature change was maintained within the error range of ± 0.5 K while adjusting nitrogen gas flow and heater current.

Experimental results
The X-ray powder diffraction pattern of the [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 crystal at room temperature is displayed in Fig. 2, and this result was consistent with that reported by Czapla et al. 23 The results of the DSC analysis of [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 under a nitrogen atmosphere are shown in Fig. 3. An endothermic peak at 434 K and a exothermic peak at 539 K where observed. However, the peak around 334 K reported previously 20 was not observed. To confirm that the DSC peaks at 434 K and 539 K were consistent with the structural phase transition, TGA was performed. The measured TGA curves are also shown in Fig. 3. Good thermal stability was observed up to around 480 K; above this temperature, the first signs of weight loss were observed, indicating the onset of partial thermal decomposition. The crystalline structure of the compound [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 (M = 280.49 mg) breaks down at high temperatures. Considering the TGA results and possible chemical reactions, the solid residue amounts were calculated. The weight loss of 13% at around 539 K (see Fig. 3) was likely due to the decomposition of the HCl moieties, which is consistent with the exothermic peak at the same temperature in the DSC curve. The weight sharply decreased between 500 and 600 K, with a corresponding weight loss of 65% around 650 K. This result is consistent with previous TGA data 23 . Further, optical polarizing microscopy was used to understand the crystal's phase transition, thermal decomposition, and melting mechanism. The color of the crystal was dark  www.nature.com/scientificreports/ brown at room temperature, as illustrated in the inset of Fig. 3. While there were no changes observed from room temperature to 523 K, it began to melt slightly at approximately 539 K. Above 600 K, the crystal emitted an odour, and its surface and edges melted considerably (see Suppplementary Information 1). The chemical shifts of the 1 H NMR spectrum of [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 crystals were obtained with increasing temperature, as shown in Fig. 4. Two peaks in the NMR spectra are indicated in the figure; the spinning sidebands for CH 2 are represented with crosses, and those for NH 3 are marked with open circles. At 300 K, the 1 H NMR chemical shift for CH 2 was observed at δ = 2.76 ppm, whereas that for NH 3 was at δ = 11.48 ppm. Below 300 K, the signal for 1 H of CH 2 had very low intensity and could not be easily identified. The 1 H peak for CH 2 did not significantly change with increasing temperature, while for NH 3 , the change in the chemical shift was dependent on temperature (see the Supplementary Information 2).
The 1 H NMR spectra were also measured with several delay times, and the intensity of NMR spectra as a function of delay time followed a single exponential function. The rate of decay of the spin-locked proton magnetization is characterized by T 1ρ 25-27 : where I(t) and I(0) are the signal intensity at time t and t = 0, respectively. The 1 H NMR signals of CH 2 and NH 3 measured at 300 K were plotted as a function of delay time over the range of 0.2-80 ms, as shown in the inset of Fig. 5. It can be seen that the 1 H NMR signal intensities varied with the delay time. From the slope of the intensity vs. delay time curve, 1 H T 1ρ values for [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 were obtained from the CH 2 and NH 3 peaks as a function of inverse temperature. Changes in T 1ρ values above T C1 were not observed outside this temperature because of the limitation of the NMR spectrometer. The 1 H T 1ρ values for CH 2 and NH 3 were of the order of 10 ms, and their values were almost independent of temperature (see Fig. 5). The 13 C NMR chemical shifts for CH 2 in [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 were measured as a function of temperature, as shown in Fig. 6. At all temperatures, the 13 C MAS NMR spectra showed two resonance signals. The 13 C MAS NMR spectrum for TMS was observed at 38.3 ppm at 300 K, which was used to calibrate the device to 0 ppm for determining the chemical shift in 13 C 28 . Here, CH 2 between CH 2 and CH 2 is named CH 2 -1, and CH 2 close to NH 3 is named CH 2 -2. At 300 K, the two resonance signals were recorded at chemical shifts of δ = 28.78 and δ = 124.97 ppm for CH 2 -1 and CH 2 -2, respectively. The 13 C chemical shifts for CH 2 were different CH 2 -1 far away from NH 3 and CH 2 -2 close to NH 3 . The small 13 C resonance peaks indicated by arrow at 420 K and 430 K were attributed to a splitting of the CH 2 -2. The 13 C chemical shift for CH 2 -1 increased slowly and monotonously without an anomalous change with increasing temperature, whereas those for CH 2 -2 moved abruptly to the lower side with increasing temperature compared to CH 2 -1, as shown in the inset in Fig. 6.
The 13 C full-width at half-maximum (FWHM) values of the NMR peaks for CH 2 -1 and CH 2 -2 decreased with increasing temperature. Broader line widths are observed for more rigid lattices, where motional narrowing is quenched, as shown by the increase in line widths at lower temperatures. The line widths of 13 C for CH 2 -1 and CH 2 -2 were the same within experimental uncertainty, where the line width narrowed from 30 to 10 ppm with increasing temperature from 180 to 430 K, respectively (see the Supplementary Information 3).
The integration change of the 13 C NMR spectra obtained by increasing the delay time was measured. All decay curves for CH 2 -1 and CH 2 -2 were described by a single exponential function, as shown by Eq. (1). 13 C T 1ρ values were measured by the spin-locking pulse sequence with a locking pulse of 75.76 kHz. From the slope of their recovery traces, the 13 C T 1ρ values were obtained for the CH 2 -1 and CH 2 -2 as a function of 1000/temperature, as shown in Fig. 7. Although no change in T 1ρ values was observed near T C2 , T 1ρ values measured for 180-430 K    where , and f e = τ C /[1 + ω e 2 τ C 2 ]. Here, C is a coefficient, γ e is the gyromagnetic ratio of the electron, S is the spin number of the paramagnetic ion, r is the distance between the paramagnetic ion and the carbon, ω e is the Larmor frequency of electron, and ω 1 is the spin-lock field. When ω C τ C = 1, T 1ρ is at its minimum, so a relationship between T 1ρ and ω 1 was applied to obtain the coefficient in Eq. (2). Using this coefficient, we calculated τ C as a function of temperature. According to BPP theory, the local field fluctuation is governed by the thermal motion of CH 2 -1 and CH 2 -2, which is activated by thermal energy. In this case, τ C is described by Arrhenius behaviour: τ C = τ o exp(-E a / k B T), where τ o , E a , and k B are the pre-correlation time, activation energy of the motions, and Boltzmann constant, respectively 27 . As the magnitude of E a depends on the molecular dynamics, we plotted τ C vs. 1000/T on a logarithmic scale (inset of Fig. 7), which gave E a values for CH 2 -1 and CH 2 -2 of 8.93 ± 0.54 and 6.85 ± 0.48 kJ/ mol, respectively. 14 N NMR investigations were performed using a [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 single crystal over the temperature range of 180-430 K. The 14 N spectra were obtained using the solid-state echo method by static NMR at a Larmor frequency of 28.90 MHz. Two 14 N NMR signals were derived from the quadrupole interactions due to the spin number I = 1. Near 333 K (= T C2 ), the number of resonance lines and resonance frequency of the NMR spectrum showed abruptly changes, as shown in Fig. 8. Above T C2 , the spectrum showed one pair of lines, whereas below T C2 it showed two pairs. The lines with the same colour below T C2 indicate the same pairs for 14 N. The changes in the 14 N resonance frequency as a function of temperature were attributed to variations in the structural geometry, corresponding to changes in the quadrupole coupling constant 30,31 . The resonance frequency of the 14 N signals below T C2 changed almost continuously, and those of the 14 N signal above this temperature also varied abruptly. Near T C2 , the electric field gradient tensors at N sites varied, reflecting changes in the atomic configuration around the nitrogen atom. Although the phase transition temperature at T C2 reported previously 20 was not observed in our DSC experimental results, the 14 N NMR spectrum showed changes near T C2 . The phase transition at T C2 exists, and 14 N in the NH 3 groups plays an import role in this phase transition. In contrast, the two different 14 N spectra below T C2 are thought to have two inequivalent N sites or be due to twin domains. However, according to the previouslyreported X-ray results 21 , there have been no reports of two different N sites, and twin domains have been reported 22 . Czapla et al. 22 suggested that the ferroelastic domains observed in the orthorhombic phase could be connected to a prototype tetragonal phase. Here, the [(CH 2 ) 3 (NH 3 ) 2 ]CuCl 4 crystal existed in three crystallographic phases: monoclinic (2/m) above 434 K, tetragonal (4/mmm) between 334 and 434 K, and orthorhombic (mmm) below 334 K. For the transition from the 4/mmm of the tetragonal phase to the mmm of the orthorhombic phase, the domain wall orientations were expressed as x = 0 and y = 0. According to Aizu 32 and Sapriel 33 , the equations of the twin domain walls reflected the ferroelasticity of the 4/mmmFmmm. Hence, our results are thought to support the mechanism of ferroelastic twin domains. As a result, the separation of two 14 N NMR lines into four 14 N NMR lines under T C2 was due to the ferroelastic twin domain structure.
(2)   13 C CP/MAS NMR, and 14 N static NMR as a function of temperature. The changes in chemical shifts in the 1 H and 13 C NMR spectra indicated changes in crystallographic symmetry. The NMR chemical shifts were related to the local field at the location of the resonating nucleus in the crystals. The 1 H NMR chemical shift for NH 3 changed more significantly with temperature than that of CH 2 because being H-bonded, the 1 H NMR chemical shift of the NH 3 moiety is much more sensitive to temperature fluctuations, and varies significantly due to the variation in H-bond length with temperature. The 13 C NMR chemical shift for CH 2 -1 increased slowly with increasing temperature, without any anomalous change. However, the shift for CH 2 -2, moved significantly to lower values with increasing temperature compared to CH 2 -1. The 13 C NMR chemical shifts of CH 2 -2 closer to the N-H···Cl bonds were higher those of CH 2 -1. In addition, the abrupt change in the resonance frequency of the 14 N nuclei observed near T C2 was attributed to a ferroelastic phase transition. The previously reported phase transition at T C2 20 was not observed in DSC, but the 14 N NMR data clearly indicated changes in atomic configuration at this temperature. The NH 3 groups are coordinated by N-H···Cl bonds; thus, atomic displacements with temperature in the environment of the 14 N nuclei are correlated with CuCl 4 . 1 H and 13 C T 1ρ have lower values in the order of milliseconds because T 1ρ is inversely proportional to the square of the magnetic moment of paramagnetic ions. The T 1ρ values for CH 2 and NH 3 protons were almost independent of temperature, but the CH 2 moiety located in the middle of the N-C-C-C-N bond undergoes tumbling motion according to the BPP theory. The increase in 13 C T 1ρ at high temperatures may be simply due to the change in distance rather than the change in correlation time. More importantly, the total correlation time τ C is dominated by the electric relaxation correlation time, rather than the rotational correlation time of the paramagnetic.