The importance of the perovskite-type structure for many practically useful substances, such as perovskite-sensitized solar cells, is well known . Their characteristic properties are associated with phase transitions in the crystal structures. It is therefore important to investigate the dynamical behavior of the cations under the restriction of the anion sub-structures of each phase. We have already reported the so-called layered perovskite-type structure of the compound 3, where the inorganic and organic layers are stacked alternatingly . The crystal undergoes successive phase transitions at T=208, 165, 135, and 121 K, as was demonstrated by 81Br NQR (nuclear quadrupole resonance) as well as differential scanning calorimetry and differential thermal analysis (DTA) measurements. Moreover, the phase transitions of the chloride analog n-(CH3CH2CH2NH3)2CdCl4 have been precisely investigated. It shows a similar layered perovskite-type structure with a different anion layering. It has been clarified that the successive phase transitions of the crystal at T=179, 157, and 105 K are associated with its normal (N), incommensurate (IC), commensurate (C), and another commensurate (C′) structures , , . One of us also reported on the dynamics and physical properties of many perovskite-type compounds of the general formula [Cl(CH2)nNH3]2MX4 (n=2–3) with M=Cd, Hg, Cu, Mn, Pb, Sn, and X=Cl, Br, I . In the present study, we investigated the structures of 1 and 2 by single-crystal X-ray diffraction. 81Br NQR and DTA methods have also been applied to these compounds to investigate the chemical bonding as well as the phase transitions. The cations of these compounds can be regarded as derived from the n-propylammonium cation CH3CH2CH2NH3+ by replacing a methyl group by a Br atom, or by replacing a hydrogen atom of a methyl group by a Br atom, respectively.
2 Results and discussion
2.1 Crystal structures
The crystallographic data and the experimental conditions of the structure determinations carried out at low temperature are given in Table 1. The bond lengths and angles for the anions are listed in Table 2. By reference to the results of the DTA measurements as described below, the present structures of 1 and 2 correspond to phases VI and IV, respectively. Both crystal structures are isostructural with sequences of alternating inorganic and organic layers parallel to the crystallographic ab plane. Compound 1: orthorhombic, Pbcn, a=843.8(4), b=775.4(4), c=2339.6(11) pm, T=113 K; compound 2: orthorhombic, Pbcn, a=858.7(3), b=783.6(2), c=2519.4(7) pm, T=113 K. The inorganic layer is a sheet of [CdBr6]4− octahedra connected through the Br atoms at the equatorial corners. The organic layer is made up of the double sub-layers of BrCH2CH2NH3+ or BrCH2CH2CH2NH3+ cations, respectively, which direct their aliphatic portions in head-to-head fashion and their polar ammonium substituents toward the inorganic sheets. The sub-layers are connected via van der Waals forces. In the inorganic layers, the planes of the Cd atoms intersect perpendicularly the c axis at c/4 and 3c/4 with direction cosines of angles of the a, b, and c axes between these planes [0, 0, 1] in both 1 and 2. The planes of the Br atoms of the organic layers intersect the c axis nearly at c/2 with direction cosines of [0.026, 0.009, 1.000] in 1 and also nearly at c/2 with those of [0.144, 0.084, 0.986] in 2. The space group of the crystals is, thus, different from the room temperature (RT) structure of 3, in which a similar layer structure exists : orthorhombic, Cmca, a=783.4(4), b=2480.2(10), c=806.5(11) pm, T=293 K. There is a slight difference in the stacking manner of the layers, although there is no deviation in the overlapping of each plane of the inorganic layers in the RT structure of 3. The corresponding inorganic planes in the structures of 1 and 2 deviate in the b direction as is shown in Figs. 1 (left) and 2 (left), respectively. The deviation is larger in 2 than in 1 with ca. 0.39 b and 0.18 b, respectively, indicating the dependence of the deviations on the size of the cation. The [CdBr6]4− octahedra are completed through the coordination of two equivalent terminal Br atoms (Br1) and four equivalent bridging ones (Br2) to the central Cd atoms (Table 2). The size of the octahedron is somewhat larger in 1 than in 2; the bond lengths Cd–Br1, Cd–Br2, and Cd–Br2#1 are 263.40(15), 274.63(13), and 307.41(14) pm in 1, and 261.82(8), 270.45(9), and 318.17(10) pm in 2, respectively. The delicate differences in the structures of the octahedra between 1 and 2 originate from the different alkyl chain lengths, i.e. from the presence or absence of only one additional methylene group (–CH2–) in the cations. The elongation of the aliphatic part affects the heat of formation and dipole moments of the alkyl ammonium cations. The difference of the heat of formation estimated by molecular orbital (MO) calculations between the optimized conformation (OPC) of the cation and the conformation of the cation in the crystal (CRC) is 43.43 kcal mol−1 for 1 (heat of formation of OPC: 157.37 and CRC: 200.80 kcal mol−1) and 59.97 kcal mol−1 for 2 (heat of formation of OPC: 148.14 and CRC: 208.11 kcal mol−1). This result suggests that the cations in 2 are more affected by intermolecular interactions than in 1 and that more deformed cations become more stable in the crystals of 2. In addition, the difference of the calculated dipole moments of the cations is 0.03 Debye (dipole moment of OPC: 13.57 and CRC: 13.60 Debye) for 1 and 1.22 Debye (dipole moment of OPC: 14.55 and CRC: 15.77 Debye) for 2. This result also supports the above-mentioned suggestions. The hydrogen bond schemes observed in 1 and 2 are depicted in Figs. 3 and 4, respectively. Table 3 lists the relevant short contacts. However, these distances do not seem to allow a judgment of the hydrogen bond strength in 1 and 2. The NH3 groups will take part in hydrogen bonding more strongly with the terminal Br1 atoms than the bridging Br2 atoms in both crystals.
Crystal structure data for 1 and 2.
|Crystal size, mm3||0.280×0.200×0.080||0.300×0.300×0.200|
|Dcalcd, g cm−3||2.96||2.78|
|[(sinθ)/λ]max, ×103 pm−1||6.482||6.488|
|Refl. measured||11 478||8075|
|R(F)/wR(F2)a,b [F2>2 σ(F2)]||0.0619/0.1221||0.0410/0.0850|
|R(F)/wR(F2)a,b (all refl.)||0.0715/0.1314||0.0478/0.0871|
|Δρfin(max/min), ×10−6e pm−3||2.44/−2.36||1.47/−3.02|
Selected bond lengths (pm) and angles (degrees) for 1 and 2 with estimated standard deviations in parentheses.a
|Bond lengths||Bond lengths|
|Bond angles||Bond angles|
Short contact lengths (pm) and angles (degrees) in 1 and 2.a
2.2 NQR measurements
Our previous investigation of 3 has shown the existence of two 81Br NQR lines of 54.098 and 61.060 MHz at T=77 K . The low-frequency line has been assigned to the bridging Br atoms situated at the equatorial positions and the high-frequency one to the terminal Br atoms at the axial positions of the [CdBr6]4− octahedra. It has been suggested that the successive phase transitions of this crystal may be correlated with the cation motions. Thus, it is of interest if the intriguing fade-out phenomena as well as the successive phase transitions can also be observed in crystals with different cations.
The temperature dependence of the 81Br NQR frequencies (ν/T) of the Br atoms of the anion of 1 is shown in Fig. 5. Two lines were observed in the low-temperature region, which is consistent with its crystal structure at T=113 K. The low-frequency line of 51.39 MHz and the high-frequency line of 61.51 MHz at T=77 K can be assigned to the bridging Br atoms (Br2) and the axial Br atoms (Br1) of the [CdBr6]4− octahedra, respectively, according to the assignment in 3. Discontinuous changes are seen in both ν/T curves, showing the existence of phase transition. The low-frequency line decreased its frequency with increasing temperature up to around 168 K, at which it disappeared. In turn, another line appeared with a frequency of ca. 2 MHz lower. The new line showed an unusual positive temperature dependence up to around 320 K, above which it disappeared. This line could be observed down to 143 K on cooling from the high temperature side. This hysteresis phenomenon shows that the phase transition is first-order type. By referring to the DTA measurements as inserted in the figure, the lower- and higher-temperature phases correspond to phases VI and V, respectively. On the other hand, the high-frequency line kept its frequency almost constant, and it disappeared at around T=156 K in phase VI with increasing temperature. In its place, two lines situated above and below the original line appeared, showing two non-equivalent terminal Br atoms in phase V. These lines approached each other, decreasing their frequencies monotonously with increasing temperature and disappeared at around T=303 K before almost coalescing. On cooling, both lines could be observed down to 139 K below the transition temperature of 158 K, showing a super-cooling phenomenon. The unique feature of the ν/T curves may be understood by considering the hydrogen bonds N–H…Br to the anion and the reorientation motions of the cations. While both ν/T curves exhibit rather small negative temperature dependence owing to the formation of the hydrogen bonds in phase VI, the hydrogen bonds of the bridging Br atoms concerned are gradually broken under the reorientation motion, resulting in the unusual positive temperature dependence in phase V. The positive temperature dependences of the ν/T curves appear when the increase in NQR frequency due to the breaking of the N–H···Br hydrogen bonds exceeds the decrease due to general torsional vibration of the Cd–Br bonds. A similar positive temperature dependence was also observed for the bridging Br atom of 3 at temperatures above T=208 K . In accordance with the presence of only one non-equivalent cation in phase VI according to the X-ray structure determination, one 81Br NQR line has been observed with the frequency of 230.95 MHz at T=77 K as shown in Fig. 7, and the temperature dependence of this line could not be traced because of the restriction of the NQR spectrometer used.
The temperature dependence of the 81Br NQR frequencies assigned to the anion of 2 is shown in Fig. 6. Two lines, with 54.57 and 62.10 MHz at T=77 K, are consistent with the crystal structure at 113 K. The high- and the low-frequency lines are assigned to the terminal Br (Br1) and the bridging Br (Br2) atoms, respectively, of the [CdBr6]4− octahedra as is described in the case of 1. The frequencies of both resonance lines decreased monotonously with increasing temperature up to around 210 K, and then disappeared at around 233 K, exhibiting a drastic frequency decrease and increase in the high- and the low-frequency lines, respectively. This disappearance phenomenon may be related to the transition of phase IV to phase III as described later. On the other hand, two resonance lines were observed in limited ranges in phase III; the high-frequency line exhibited a positive coefficient between ca. 296 and 321 K, but the low-frequency line exhibited a negative coefficient between ca. 273 and 321 K. The disappearance and discontinuous changes of the ν/T curves between phase IV and phase III show that the phase transition is first-order type. As the respective high- and low-frequency lines of both phases probably connect when they are extrapolated toward the transition point, a minor structure change at the transition point is expected. The high- and the low-frequency lines of phase III may, thus, be assigned to the terminal and bridging Br atoms of the anion layers as well. The positive temperature dependence may be attributed to the partial destruction of the hydrogen bonds to the terminal Br atom of the anions with increasing temperature as described above. Upon heating, both lines of phase III disappeared again at around T=322 K, corresponding to another phase transition from III to II. No NQR line was observed for phase II. Only one line could be detected between T=345 and 371 K for phase I, which may be attributable to the bridging Br atom. The missing of the high-frequency line appears to be due to the increased fluctuation of the electric field gradient (EFG) at the terminal Br nuclei caused by the motion of the cations in phase I, resulting in a line-broadening to an unobserved level. A 81Br NQR line was observed with the frequency of 214.50 MHz at 77 K for the Br atom in the cation BrCH2CH2CH2NH3+. This line decreased in its frequency with increasing temperature up to around 175 K, at which it disappeared (Fig. 7). This disappearance indicates an onset of rotation of the cation about its long axis. In that case, the NQR frequency of the relevant Br atom is expected to be averaged down to νav=νo(3cos2θ−1)/2, where νo is the frequency in the case of no motion, and θ is the angle between the z component of the principal axes of the EFG (z-EFG) of the Br atom and the rotational axis. Using the angle of ~45° between the C–Br bond (probably aligned with the z-EFG) and the principal axis of the minimum component of the moment of inertia of CRC, νav is calculated to be ~0.25 νo=53 MHz, although this averaged line could not be observed. The 81Br NQR frequency of the Br atom in the cation in 1 is higher than the one in 2. This observation is in accordance with the results of the MO calculations. The calculated frequency for the Br atom of CRC in 1 (336.2 MHz) is higher than that in 2 (304.4 MHz), because the Br atom of CRC in 1 has a +0.084e charge instead of a −0.016e charge in 2.
2.3 DTA measurements
The DTA measurements revealed the existence of successive phase transitions for both crystals of 1 and 2. The phase names, the transition temperatures, and the orders of phase transition are listed in Table 4. On the 1st cooling run, the crystals of 1 showed a hardly recognizable exothermic peak at T=144 K because of the phase transition V→VI, as shown in Fig. 8a. A corresponding small endothermic peak for the transition VI→V on the heating run was observed at 158 K. This transition is first-order type by judging the shape as well as the hysteresis phenomenon of the thermal anomaly. This is in accordance with the temperature dependence of the 81Br NQR frequencies described above. It is noted that the peak anomalies of the VI↔V transition were not observed in the 2nd run, while the anomalies due to the other transitions were reproduced well, as shown in Fig. 8b. On the successive heating run, a sharp endothermic peak for V→IV appeared with first-order-type features at T=325 K. It is noted that all the 81Br NQR lines became unobserved above this temperature. A further increase in temperature showed the IV→III transition anomaly at 334 K. This transition can be assigned to be second-order type because the only change of slope in the DTA curve appeared as the thermal anomaly. Successively, two sharp endothermic peaks first order in character were observed at T=397 and 404 K, which showed the existence of the III→II and II→I transitions, respectively. These peaks converged into one exothermic peak at 401 K in the successive cooling run owing to a probable super-cooling for the I→II transition. The remaining exothermic peak for IV→V was observed with almost no hysteresis.
Transition points of 1 and 2.
|Compounds||Transition||Ttr (in K) (ΔTtr)||Order|
The crystals of 2 underwent three phase transitions between T=100 and 400 K as is shown in Fig. 9. All peak anomalies appeared with first-order-type features in accordance with the conclusions from the NQR measurements described above. On the 1st cooling run, the crystals of 2 showed a small exothermic anomaly for the III→IV transition at 236 K. The corresponding endothermic peak IV→III appeared at 233 K on the successive heating run. Further heating revealed the other successive endothermic peaks due to the phase transitions of III→II and II→I at T=322 and 340 K, respectively. The respective exothermic peaks for the I→II and II→III transitions appeared at 336 and 320 K on the cooling run. It is eye-catching that the transitions between III and II were accompanied by quite large peaks compared to the other two transitions, showing the appearance of larger thermal events. On this phase transition, the higher-frequency 81Br NQR line due to the terminal Br atoms of the anions showed a positive temperature dependence in phase III, and NQR lines were not observed in phase II. This may indicate that the cation motion is closely correlated to the phase transition. The hydrogen bonds between the cations and the terminal Br atoms of the anions may, thus, be substantially weakened due to probable reorientation motions of the NH3 group of the cations with increasing temperature in phase III. Any further increased motion of the cations may cause a fluctuation of the EFG around the nucleus of the terminal Br atoms, which results in the extinction of their 81Br NQR signals due to the broadening in phase II.
Both crystals of 1 and 2 show isomorphic layer structures (Pbcn) at T=113 K, which are slightly different from the structure (Cmca) of 3 determined at 293 K . Although the RT structures of 1 and 2 have not been determined and two non-equivalent terminal Br atoms in the RT phase or phase V of 1 exist, the difference between the RT structure of 1 and that of 3 may be small considering that the space group Pbcn is a non-isomorphic subgroup of Cmca. The NQR and DTA results for 1 and 2 indicate the existence of similar cation motions, which may be responsible for their phase transitions as follows. Both crystals undergo several phase transitions between T=77 and ca. 400 K according to the results of the DTA and NQR measurements, with two anomalies appearing close to 330 K. The peaks on the lower-temperature side of these two anomalies are quite large, showing that a significant thermal event happens in both crystals. The higher temperature peaks are second-order-type anomalies in the case of 1 and first-order type in the case of 2, and small peaks show the occurrence of only a minor change in the structure. These successive structural changes may be triggered by the cation motions in accordance with the fact that the NQR lines of both crystals disappear above these transition temperatures. The motions of the cations are expected to be around their longitudinal axes. The difference of minimum moments of inertia about the longitudinal axis, Ixx’s, between 3.14×10−46 kg m2 for CRC in 1 and 8.94×10−46 kg m2 for CRC in 2 cause a small difference in behavior in their phase transitions; the lowest-temperature transitions of 1 and 2 show rather different features, as is seen in both the NQR and DTA performance. The partial motion of the cation segments such as the reorientational motion of the NH3 group, which was observed in (ClCH2CH2NH3)2CdCl4 , should be related to these transitions in the lower-temperature regions. The phase transitions of 3, of which the CRC has an Ixx of 1.98×10−46 kg m2, happen at lower temperatures (T=208, 165, 135, and 121 K), while the corresponding phase transitions around 300 K found in 1 and 2 are absent. The present results, thus, indicate that the layer structures are maintained in the crystals of the type (RNH3)2CdBr4 with a series of alkylammonium cations, the physical properties of which, such as phase transitions, are quite dependent on the shapes and functions of the cations.
4 Experimental section
4.1 Sample preparation
Crystals of 1 and 2 were prepared by mixing the respective bromoalkylammonium bromides and CdCO3 with molar ratio of 2:1 in an aqueous solution with a slight excess of hydrobromic acid. Colorless rectangular parallelepiped crystals that appeared from the solutions were dried over P2O5 in a desiccator at RT. – Elemental analyses: for 1: calcd. C 2.07, H 7.04, N 4.12; found C 2.05, H 7.13, N 4.15; for 2: calcd. C 2.55, H 10.15, N 3.95; found C 2.38, H 10.08, N 3.97.
4.2 NQR and DTA measurements
The 79,81Br NQR spectra were recorded by a home-made super-regenerative type NQR spectrometer at temperatures above 77 K. The resonance frequencies were determined by the counting method. The accuracy of the frequency measurements is estimated to be within ±0.05 MHz. DTA was performed by a home-made apparatus at temperatures above 100 K.
4.3 Crystal structure determinations
All measurements were made on a Rigaku Saturn724 CCD X-ray detector (MoKα radiation, λ=0.71075 Å). Data were collected at 113 ± 1 K and processed using CrystalClear . The structures were solved by Direct Methods  and expanded using Fourier techniques. All calculations were performed using the CrystalStructure crystallographic software package , except for the refinement, which was performed using Shelxl-2013 . Crystals of 1 suitable for X-ray measurements were obtained by cutting them off from large crystals, which may result in quality losses causing slightly larger R values. In addition, the first-order transition of VI↔V at around T=158 K in 1 with an accompanying strong hysteresis may also affect the crystal quality.
CCDC 1405118 (for 1) and CCDC 1405119 (for 2) contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
4.4 MO calculations
MO calculations with PM3 basis were carried out using Gamess (Windows, version 11, 2013) and a quantum chemistry package . The heats of formation and dipole moments were calculated on the OPC of the cations and their conformation CRC in the crystals of 1 and 2. Details of the calculations of the NQR frequencies are reported elsewhere .
We are grateful to Prof. Y. Yamada of the Graduate School of Science and Engineering, Saga University, for the determination of the crystal structures of 1 and 2.
J. Burschka, N. Pellet, S. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin, M. Grätzel, Nature2013499, 316.
H. Ishihara, S. Dou, K. Horiuchi, V. G. Krishnan, H. Paulus, H. Fuess, A. Weiss, Z. Naturforsch.199651b, 1216.
K. Suzuki, H. Fujimori, T. Asaji, S. Ishimaru, R. Ikeda, Z. Naturforsch. 200257a, 451.
M. A. White, N. W. Granville, N. J. Davies, L. A. K. Staveley, J. Phys. Chem. Solids198142, 953.
R. Mokhlisse, M. Couzi, N. B. Chanh, Y. Haget, C. Hauw, A. Meresse, J. Phys. Chem. Solids198546, 187.
CrystalClear, Data Collection and Processing Software, Rigaku Corporation, Tokyo (Japan) 1998–2014.
M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini, G. L. Cascarano, L. De Caro, C. Giacovazzo, G. Polidori, D. Siliqi, R. Spagna, J. Appl. Crystallogr.200740, 609.
CrystalStructure (version 4.1), Crystal Structure Analysis Package, Rigaku Corporation, Tokyo, (Japan) 2000–2014.
G. M. Sheldrick, Acta Crystallogr.2008A64, 112.
M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S. Gordon, J. H. Jensen, S. Koseki, N. Matsunaga, K. A. Nguyen, S. J. Su, T. L. Windus, M. Dupuis, J. A. Montgomery, J. Comput. Chem.199314, 1347.
T. M. Gesing, E. Lork, H. Terao, H. Ishihara, Z. Naturforsch. 201671b, 241.