Tomoyuki
Matsuda
a,
Jungeun
Kim
b and
Yutaka
Moritomo
*a
aGraduated School of Pure and Applied Science, University of Tsukuba, Tsukuba, Ibaraki 305-8571, Japan. Fax: +81 29 853 4337; Tel: + 81 29 853 4337E-mail: moritomo@sakura.cc.tsukuba.ac.jp
bJASRI/SPring-8, 1-1-1 Kouto, Sayo-cho, Sayo-gun, Hyogo 679-5198, Japan
First published on 1st November 2011
We investigated the thermal expansion properties of metal nitroprussides, MA[Fe(CN)5NO]·zH2O (MA = Mn, Fe, Co, Ni, Cu, Zn, and Cd), which have various network structures, i.e. three-dimensional (3D) cubic (Fmm for MA = Fe, Co, and Ni), distorted three-dimensional (distorted-3D) orthorhombic (Pnma for MA = Mn, Zn, and Cd), and two-dimensional (2D) orthorhombic (Amm2 for MA = Cu) structures. In the cubic system, in which the unbridged NO groups are randomly distributed, isotropic positive thermal expansion was observed. On the contrary, in the orthorhombic system, in which the unbridged NO groups are uniaxially distributed, positive (negative) thermal expansion was observed along (perpendicular to) the uniaxial direction. We interpreted the characteristic thermal responses in terms of the rotational vibrations of the [Fe(CN)5NO] units as well as their steric hindrances due to the network dimensionality.
Prussian blue type cyanides, AxMA[MB(CN)6]y·zH2O (A = alkali metal ion, MA = Mn, Fe, Co, Ni, Cu, Zn and Cd, and MB = Fe, Co, Pt), show characteristic thermal expansion properties.4–6 The compounds belong to face-centered cubic (Fmm; Z = 4) structures, in which MA and MB are bridged by the cyano groups and form a three-dimensional (3D) network structure.10 The alkali metal cations and water molecules are accommodated in the interstitial sites of the lattice. The thermal expansion coefficient (β ≡ dlna/dT) of the lattice constant a systematically decreases with an increase in a: β = +1.46 × 10−5 K−1 for Rb0.49Co[Fe(CN)6]0.80·3.8H2O (a = 9.9431(3) Å) and β = −0.88 × 10−5 K−1 for Cs1.00Cd[Fe(CN)6]1.00·0.5H2O (a = 10.7551(1) Å). The systematic change in β is well explained by the thermally-induced rotational vibrations of the [Fe(CN)6] units as well as their steric hindrances. According to this scenario, we may significantly control the thermal expansion properties by the modification of the dimensionality of the cyano-bridged network.
In this work, we systematically studied the thermal expansion properties of metal nitroprussides, MA[Fe(CN)5NO]·zH2O (MA = Mn, Fe, Co, Ni, Cu, Zn, and Cd), in which one of the CN groups are replaced by the NO group. The unbridged NO group modified the dimensionality of the cyano-bridged network structure. Based on the structural analysis, we classified the network structure into three categories, i.e., 3D cubic (Fmm for MA = Fe, Co, and Ni), distorted three-dimensional (distorted-3D) orthorhombic (Pnma for MA = Mn, Zn, and Cd), and two-dimensional (2D) orthorhombic (Amm2 for MA = Cu) structures. We found that the variation of the network dimensionality crucially influences the thermal expansion properties. We interpreted the dimensionality-dependence of the thermal expansion in terms of the rotational vibrations of the [Fe(CN)5NO] units as well as their steric hindrances.
In order to precisely determine a and the space groups of the above-synthesized metal nitroprussides, X-ray powder-diffraction patterns were measured at the synchrotron-radiation facility, SPring-8. First, powder samples were filled into a 0.3 mmϕ glass capillary. The capillary was put on a Debye–Scherrer camera at the BL02B2 beamline of SPring-8.11 The wavelength of the X-ray (λ ∼ 0.5 Å) was calibrated by the lattice constant of standard CeO2 powder. The sample temperature was controlled by the cooled nitrogen gas. The exposure time was 5 min. The lattice constants of each compound were refined by the RIETAN-FP program,12 and the 2θ range used in the Rietveld analyses was 3°–30°
![]() | ||
Fig. 1 X-Ray powder diffraction patterns at 100 K and Rietveld analysis of Co (a), Mn (b), and Cu (c). Red dots, black line, and blue line are the observed plots, calculated pattern, and their difference, respectively. Green bars represent the calculated positions of the Bragg reflections. (inset) Schematic illustration of the unit cell. |
Space group | Lattice constant / Å | |||
---|---|---|---|---|
a | b | c | ||
Mn[Fe(CN)5NO]·2.1H2O | Pnma | 14.0371(5) | 7.5464(2) | 10.4127(3) |
Fe[Fe(CN)5NO]·5.0H2O |
Fm![]() |
10.3144(3) | — | — |
Co[Fe(CN)5NO]·5.4H2O |
Fm![]() |
10.2501(2) | — | — |
Ni[Fe(CN)5NO]·5.8H2O |
Fm![]() |
10.1537(2) | — | — |
Cu[Fe(CN)5NO]·2.2H2O | Amm2 | 7.2231(4) | 6.9700(4) | 10.0850(6) |
Zn[Fe(CN)5NO]·2.0H2O | Pnma | 13.8463(3) | 7.4099(2) | 10.3789(2) |
Cd[Fe(CN)5NO]·2.3H2O | Pnma | 14.2012(5) | 7.6830(3) | 10.5029(4) |
The cubic compounds (Fe, Co, and Ni) show 3D network structures similar to the case of the Prussian blue analogues. Here, we note that the five cyano groups of [Fe(CN)5NO] are bridged to the neighboring MA while the NO group is unbridged.27
On the contrary, the orthorhombic compounds (Mn, Cu, Zn, and Cd) have rather complicated network structures (Fig. 2). The Mn, Zn, and Cd compounds show distorted-3D network structure. Within the ab plane (Fig. 2a, left), MA atoms are surrounded by four cyanide nitrogen atoms of [Fe(CN)5NO], forming cyano-bridged 2D layers (MA—N—C—Fe). Along the c axis (Fig. 2a, right), one side of MA is coordinated to cyanide nitrogen atom of [Fe(CN)5NO], while the other side is coordinated to the oxygen atom of the water ligand. That is, the five cyano groups of [Fe(CN)5NO] are bridged to the neighboring MA, while the NO group is unbridged. Due to the steric confinements between the free NO group of [Fe(CN)5NO], and water ligand of MA, the cyano-bridged MA-Fe layers along the ab plane are significantly distorted. The Cu compound shows 2D network structure. Within the ab plane (Fig. 2a, left), Cu atoms are surrounded by four cyanide nitrogen atoms of [Fe(CN)5NO]. Along the c axis (Fig. 2a, right), however, both sides of Cu atoms are coordinated to the oxygen atoms of water ligands. That is, the four CN groups of [Fe(CN)5NO] are bridged to the neighboring Cu in the ab plane while the remaining CN group and the NO group are unbridged along the c axis. The cyano-bridged Cu-Fe layers stack alternately along the c axis, causing 2D network structure.
![]() | ||
Fig. 2 Schematic crystal structures of distorted-3D network for Mn, Zn, and Cd (a), and 2D network for Cu (b). (left) The projection in the ab plane. (right) Network structure perpendicular to the ab plane. |
![]() | ||
Fig. 3 Temperature dependences of the normalized lattice constants for Fe (a), Co (b), and Ni (c). The lines are the results of the least-squares fittings. |
Next, we show in Fig. S2 (ESI†) typical examples of the magnified X-ray diffraction patterns around 020, 311, and 212 reflections in Mn, Zn, and Cd, which have distorted-3D networks. The 311 and 212 reflections shift to the high-angle side with a decrease in temperature, reflecting positive thermal expansion (PTE). On the contrary, the 020 reflection shifts to the low-angle side with a decrease in temperature, reflecting NTE along the b axis. The lattice constants are refined against temperature. The reliable factor RB ranges from 1.01 to 2.80%. To evaluate the thermal expansion of cyano-bridged ab plane, we use the geometrical average of the lattice constants (a' ≡ a1/2 × b1/2). We show the temperature dependences of a' and c values in Fig. 4. We determined β by the least-squares fittings. The β values are −9.8 × 10−7 / K (a'), and 5.35 × 10−5 / K (c) for Mn, −4.3 × 10−6 / K (a'), and 5.19 × 10−5 / K (c) for Zn, and −3.4 × 10−6 / K (a'), and 6.63 × 10−5 / K (c) for Cd.
![]() | ||
Fig. 4 Temperature dependences of the normalized a' (≡ a1/2 × b1/2) and c for Mn (a), Zn (b), and Cd (c). The lines are the results of the least-squares fittings. |
We show in Fig. S3 (ESI†) a typical example of magnified X-ray diffraction patterns around 200, 020, 004, and 220 reflections in Cu, which have a 2D network structure. The 020 and 004 reflections shift to the high-angle side with decrease in temperature, reflecting PTE along the b and c axes. On the contrary, 200 and 220 reflections shift to the low-angle side with decrease in temperature, reflecting NTE along the a axis and ab plane. The lattice constants of Cu are refined against temperature, and the reliable factor RB ranges from 3.08 to 3.41%. To evaluate the thermal expansion of cyano-bridged ab plane, we use the geometrical average of the lattice constants (a' ≡ a1/2 × b1/2). We show in Fig. 5 the normalized values of a' and c against temperature. We determined β by the least-squares fitting, and the β values are −5.2 ×10−6 / K (a'), and 1.20 × 10−4 / K (c).
![]() | ||
Fig. 5 Temperature dependence of the normalized a' (≡ a1/2 × b1/2) and c for Cu. The lines are the results of the least-squares fittings. |
In the in-plane components (Fig. 6a), the distorted-3D and 2D systems show NTE, while the 3D system shows PTE. This dimensionality dependences of the thermal expansion is well understood by the steric hindrance effect. The origin of the NTE can be ascribed to the rotational vibrations of the [Fe(CN)5NO] units, analogous to the case of the Prussian blue analgoues.9 In the distorted-3D and 2D systems, there exist enough spaces for the [Fe(CN)5NO] units to show thermally-induced rotational vibrations (Fig. 2). Then, the rotational vibration causes the NTE as observed. In the 3D system, however, the steric hindrance suppress the rotational vibration. Then, the inherent PTE overcomes the NTE behavior.
![]() | ||
Fig. 6 (a) In-plane, and (b) out-of-plane components of the thermal expansion against MA for MA[Fe(CN)5NO]·zH2O. The Fe, Co, and Ni compounds with 3D network (closed squares), the Mn, Zn, and Cd compounds with distorted- 3D network (closed diamonds), and the Cu compound with 2D network (closed circles). |
In the out-of-plane components (Fig. 6b), the magnitude of PTE increases with a decrease in the dimensionality. This is because the chemical bonds between the cyano-bridged layers become weaker as the dimensionality decreases. Actually, the number N of the cyano-bridge per one [Fe(CN)5NO] between the layers decreases from 1.67 (= 2 × (5/6)) for the 3D system, 1 for the distorted-3D system, and 0 for the 2D system. The reduction of the cyano-bridge weakens the inter-layer interaction, and enhances the PTE behavior.
Footnote |
† Electronic Supplementary Information (ESI) available: Temperature dependences of magnified X-ray powder diffraction patterns. See DOI: 10.1039/c1ra00547b/ |
This journal is © The Royal Society of Chemistry 2011 |