Structural, optical, and dielectrically analyses.
After applying the D8 Advance X-ray diffractometer, the resulting XRD patterns of the synthesized Ca2−xCuxP2O7 powders (x = 0.00−2.00) are displayed in Fig. 2. The structures of Ca2−xCuxP2O7 were analyzed through the Rietveld refinement analytic technique 30 using the FullProf package 31. A pseudo-Voigt function (a linear combination between the Lorentzian and the Gaussian functions) was adequate at all time for obtaining the good fits of the experimental data. The initial model for the refinement of the single phase structure (Ca2P2O7 and Cu2P2O7) was took from parameters described well in the Calvo research 32.
In addition, Fig. 3 showed the corresponding Rietveld refinement results of the Ca2−xCuxP2O7, when x = 0.00, 1.00, and 2.00. Fig. 3 showed the calculated (Ycal) and observed (Yobs) diffraction patterns as well as the different values between them (Yobs−Ycal) of samples. The refinement plots gave the evolution of the XRD patterns in the various ratios between Ca and Cu (Ca2−xCuxP2O7, x = 0.00, 1.00 and 2.00). The Rietveld refinement analysis and the XRD data of powders confirmed the formation of metal pyrophosphate compounds (Ca2−xCuxP2O7).
The crystallographic information of the synthesized compounds was described briefly. When x = 0.00, the single metal pyrophosphate phase, β-Ca2P2O7, was obtained with the tetragonal crystal system, space group of P41, space group number of 76, Schoenflies symbol of C42 33, and number of formula units per unit cell or Z = 8. When x = 1.00, the binary metal pyrophosphate phase, CaCuP2O7, was obtained with the monoclinic crystal system, space group of P21/c, space group number of 14, Schoenflies symbol of C2h5, and Z = 4. Finally, when x = 2.00, another single metal pyrophosphate phase, α-Cu2P2O7, was obtained with the monoclinic crystal system, space group of C2/c, space group number of 15, Schoenflies symbol of C2h6, and Z = 4. For other Ca2−xCuxP2O7 samples, when x = 0.50, exhibited two phases between CuCaP2O7 and α-Cu2P2O7, whereas, when x = 1.50, two phases between CuCaP2O7 and β-Ca2P2O7 were then observed. The P−O−P bond angles (of the O3P−O−PO3 bridge of P2O74−) and M−O6 bond lengths (M = Ca or Cu) were determined by using the refinement analysis, and the obtained values were summarized in Table 1.
Table 1. Bond angles and bond lengths from the Rietveld refinement analytic technique for the synthesized Ca2−xCuxP2O7 samples; x = 0.00, 1.00, and 2.00
x values
|
Compounds
|
P−O−P angles / °
|
M−Ox bond lengths / Å
|
χ2
|
Bonding
|
Sample 1
|
Sample 2
|
Sample 3
|
Sample 4
|
2.00
|
α-Cu2P2O7
|
154.60(10)
|
Cu−Oeq
|
1.886(5)
|
ND
|
ND
|
ND
|
1.2
|
Cu−Oeq
|
1.942(6)
|
ND
|
ND
|
ND
|
Cu−Oeq
|
1.980(5)
|
ND
|
ND
|
ND
|
Cu−Oeq
|
1.995(5)
|
ND
|
ND
|
ND
|
Cu−Oax
|
2.354(9)
|
ND
|
ND
|
ND
|
Cu−Oax
|
2.920(9)
|
ND
|
ND
|
ND
|
1.00
|
CaCuP2O7
|
159.00(1)
|
Cu−Oeq
|
1.853(3)
|
ND
|
ND
|
ND
|
2.86
|
Cu−Oeq
|
2.004(2)
|
ND
|
ND
|
ND
|
Cu−Oeq
|
2.106(1)
|
ND
|
ND
|
ND
|
Cu−Oeq
|
2.129(5)
|
ND
|
ND
|
ND
|
Cu−Oax
|
2.245(3)
|
ND
|
ND
|
ND
|
Cu−Oax
|
2.811(1)
|
ND
|
ND
|
ND
|
0.00
|
β-Ca2P2O7
|
116.51(7)
140.96(1)
|
Ca−O1
|
2.53(5)
|
2.52(6)
|
2.55(4)
|
2.49(5)
|
1.31
|
Ca−O2
|
2.33(4)
|
2.60(5)
|
2.65(6)
|
2.19(4)
|
Ca−O3
|
2.50(5)
|
2.89(4)
|
2.61(5)
|
2.36(5)
|
Ca−O4
|
2.80(7)
|
2.16(6)
|
2.26(10)
|
2.56(7)
|
Ca−O5
|
2.36(4)
|
2.27(4)
|
2.26(4)
|
2.28(4)
|
Ca−O6
|
2.18(6)
|
2.30(4)
|
2.34(5)
|
2.53(5)
|
Ca−O7
|
2.91(4)
|
2.83(5)
|
2.50(4)
|
2.81(5)
|
when eq and ax subscripts are the equatorial and axial (or apical) positions, respectively, whereas the χ2 is goodness of fit. ND is not detected.
Since X-ray absorption near-edge structure (XANES) is very sensitive to both the change in the local geometry (especially the ligand environment of the metal) and the oxidation state 34. Therefore, the spectra were collected at both Ca and Cu K-edges. They could help to understand the Fourier transform evolutions 34. The X-ray absorption edge energies (E0) of the synthesized Ca2−xCuxP2O7 compounds at the Ca and Cu K-edge are listed in Table 2.
Table 2. X-ray absorption edge energies (E0) of the synthesized Ca2−xCuxP2O7 compounds, when x = 0.00−2.00
x values
|
Compounds
|
X-ray absorption edge energies (E0) / electron Volt, eV
|
Ca K-edge
|
Cu K-edge
|
Cu0
|
Cu0
|
ND
|
8978.45
|
Cu1+
|
Cu2O
|
ND
|
8979.52
|
Cu2+
|
CuO
|
ND
|
8987.72
|
Ca2+
|
CaO
|
4033.09
|
−
|
2.00
|
Cu2P2O7
|
ND
|
8987.88
|
1.50
|
Ca0.5Cu1.5P2O7
|
4033.14
|
8987.08
|
1.00
|
CaCuP2O7
|
4033.42
|
8987.09
|
0.50
|
Ca1.5Cu0.5P2O7
|
4033.42
|
8987.13
|
0.00
|
Ca2P2O7
|
ND
|
ND
|
ND is not detected.
The E0 values of the various Cu valences (Cu0, Cu1+, and Cu2+) obtained in this work are in line well to the information reported by Yano and Yachandra 34. They reported that the E0 values raise with raising oxidation state. They also described that an electron in an atom experiences the full charge of the positive nucleus. In contrast, in case of many electrons, the electrons in an outer layer are simultaneously repelled by the negative-charged electrons and attracted to the positive nucleus. The lower the oxidation state of metals, the less positive the overall charge of the atom. Consequently, to excite an electron from an orbital, more energy is required. In summary, when the metal has more positive charge, the E0 values (XANES spectra) shifts to a higher energy 34. According to Table 2, in the Cu K-edge, the E0 values of the synthesized Ca2−xCuxP2O7 samples (x = 0.50−2.00) were similar to the E0 values of the Cu2+O, pointing out that Cu2+ was in a monoclinic symmetry. In addition, the XANES spectra of samples in the Ca K-edge showed the E0 values similar to Ca2+O, indicating that there were Ca2+ in the crystal structure of Cu2P2O7 host, resulting in the formations of Ca2−xCuxP2O7. Fig. 4 presented the local environment of Ca atoms when they entered the Cu2P2O7 structure. The spectra of Ca and Cu in CuCaP2O7 compound were difference. These demonstrated that the coordinated environments of the divalent Ca in the CuCaP2O7 are statistically significant difference 35.
The coordinated complex with different properties have different colors such as blue for Cu(NH3)4H2O)22+, red for Co(NH3)5H2O3+, and green for CoF63−. This different color phenomenon was well explained by the crystal field theory (CFT), that described by El Jazouli et al. and Chen et al. 36, 37. The optical properties and the corresponding CIE chromatic coordinates 36, 38, 39 of Ca2−xCuxP2O7 samples (x = 0.00−2.00) are showed in Fig. 5. All Ca/Cu ratio compounds, except the composition with x = 0.00 (Ca2P2O7), showed a greenish color, in which Ca2P2O7 exhibited a colorless powder. The colors of samples were dictated from elongation or compression of the z ligand bonds of Cu2+ ion. The result of composition with x = 2.00 (Cu2P2O7) illustrated a yellowish-green color, while the binary metal compounds (x = 0.50−1.50) presented color tones that changed from blue-green to bluish-green.
The mean static atomic dielectric constants (εr) of the synthesized Ca2−xCuxP2O7 compounds were estimated using the well-known Clausius-Mossotti relation 40 as following equation (Eq. (3)):
(3)
where αD is the sum of the dielectric polarizabilities of individual ions, whereas Vm is the molar volume.
The effect of porosity on the permittivity was eliminated by applying the Bosman and Havinga’s correction 41 as showed in Eq. (4), which can be used for the some materials, i.e., dense ceramics, having porosity lower than 5%:
εr,corrected = εr,measured(1+1.5P) (4)
where εr,measured and εr,corrected are the measured and corrected relative permittivity, respectively, and P is fractional porosity.
After applying the Clausius-Mossotti relation (Eq. (3)), the dielectric constant (εr) values as a function of the composition x of the synthesized Ca2−xCuxP2O7 (x = 0.00−2.00) were presented in Fig. 6, which showed the combination values between the calculated data (atomic polarization part, red bars), and measured results (atomic polarization part + ionic polarization part, red and purple bars). The single metal pyrophosphates (Ca2P2O7 and Cu2P2O7) showed the εr of 15.6 and 10.5, which were higher than the εr of some binary metal pyrophosphates (i.e., CaCuP2O7, εr = 9.8), except other binary metal pyrophosphates of x = 0.50 (Ca1.50Cu0.50P2O7, εr = 11.6) and 1.50 (Ca0.50Cu1.50P2O7, εr = 10.7). The Clausius-Mossotti equation focused on only the dielectric constant from atomic polarization (electron cloud bias in the electric fields). Indeed, the samples were measured at a frequency of 1 MHz for the decreasing extrinsic factor, and the polarization caused the moving of both cations (Cu2+, Ca2+, and P5+) and anions (O2−) in the crystal Ca2−xCuxP2O7 structure. The moving of the ions in the electric field were cause of increasing dielectric constant, which compared to the calculated data using the Clausius-Mossotti equation. The equation used in this study considered dielectric constant, using the bond angle, bond length, and volume of the MO6 octahedral.
The extended X-ray absorption fine structure (EXAFS) spectra of the synthesized Ca2−xCuxP2O7 samples are showed in Fig. 7. An environment around Cu atoms was investigated. The primitive EXAFS model was taking from parameters, which obtained from the Rietveld refinement of each sample.
Table 3. Bond length from EXAFS fitting for Ca2−xCuxP2O7 samples; x = 1.00 and 2.00
x values
|
Compounds
|
Path
|
Shell
|
CN
|
R / Å
|
σ2 / Å2
|
R-factor
|
2.00
|
Cu2P2O7
|
Cu−Oeq
|
1
|
2
|
1.90505
|
0.00508
|
0.00369
|
Cu−Oeq
|
1
|
2
|
1.96427
|
0.00499
|
Cu−Oax
|
1
|
1
|
2.30075
|
0.03223
|
Cu−Oax
|
2
|
1
|
2.91358
|
0.00796
|
1.00
|
CaCuP2O7
|
Cu−Oeq
|
1
|
1
|
1.88239
|
0.05892
|
0.01308
|
Cu−Oeq
|
1
|
2
|
1.94469
|
0.00529
|
Cu−Oeq
|
1
|
1
|
1.99797
|
0.06324
|
Cu−Oax
|
1
|
1
|
2.15606
|
0.33359
|
Cu−Oax
|
2
|
1
|
2.88853
|
0.00839
|
when eq and ax subscripts are equatorial and axial (or apical) positions, respectively. CN is coordination number, R is radial distance, σ2 is mean squared displacement, and R-factor is the fitting statistic factor.
Details of EXAFS spectroscopic fitting of Ca2−xCuxP2O7 samples were summarized in Table 3, which exhibited the distortion of CuO6 octahedral. The spectra of x = 0.00 was undetectable because limitation of instrument in beamline 8 of the National Synchrotron Research Center (Thailand). As presented in Table 3, the samples, when x = 1.00 and 2.00, showed three main shells. The first shell of spectrum from the model was consisting of four equatorial oxygen atoms, Cu−O1eq, Cu−O2eq, Cu−O3eq, and Cu−O4eq of CuO6 octahedral. Then, the second shell was detected only one axial oxygen atom, Cu−O5ax. The last one axial oxygen atom, Cu−O6ax was observed in the third shell. The Cu of Cu−O6 were also combined to the phosphorus atom, Cu−P. Different radial distances (R / Å) between the Rietveld refinement and EXAFS fitting may be cause of the measurement type of each technique. X-ray diffraction (Rietveld refinement) investigated the global structure, while X-ray absorption (EXAFS fitting) probed on the detail of Cu/Ca local structure 42, 43. The fitting statistic factor (R-factor) of x = 1.00 were worse than that of x = 2.00 because of two important factors. First, the crystal structure of α-CaCuP2O7 (x = 1.00) was lower symmetry than another sample (Cu2P2O7 (x = 1.00)). Second, the α-CaCuP2O7 (x = 1.00) exhibited four difference types of atomic position in the unit cell.
Vibrational spectroscopy.
The FTIR and Raman spectroscopies are good methods for identifying the chemical bonding of rotational, vibration, and other low-frequency modes in the phosphate group 44. After applying the Spectrum GX FTIR spectrometer, the FTIR spectra of the synthesized Ca2−xCuxP2O7 samples are presented in Fig. 8, whereas the corresponding assignments were tabulated in Table 4. The FTIR spectra observed in this research closely like to the spectral results reported in the literatures 12, 13, 15-17, 45. They successfully synthesized and investigated the vibrational spectroscopy of various single, double, and triple metal pyrophosphates, i.e., Mg2P2O7, Mn1.8Co0.2P2O7, Mn1.8Co0.1Mg0.1P2O7, Co1.6Zn0.2Mn0.2P2O7, and CoFeP2O7.
The vibrational characteristics of the synthesized Ca2−xCuxP2O7 were described in detail. The strong vibrational bands at about 1190 and 1060 cm−1 were attributed to an asymmetric (νas PO3) vibrational mode of PO3 unit of the pyrophosphate (O3P−O−PO34− or P2O74−) ions, whereas a vibrational band at about 1100 cm−1 was attributed to the symmetric stretching (νs PO3) of the PO3 unit. The asymmetric (δas PO3) and symmetric (δs PO3) bending modes are observed at the vibrational positions at about 580 and 540 cm−1, respectively. The asymmetric (νas P−O−P) and symmetric stretching (νs P−O−P) modes of the P−O−P bridge of the O3P−O−PO34− group were observed at the vibrational positions of about 960 and 740 cm−1, respectively. However, in case of the Ca2−xCuxP2O7 samples with x = 0.50 (Ca1.5Cu0.5P2O7) and x = 1.5 (Ca0.5Cu1.5P2O7), the P−O−P symmetric stretching mode appeared two peaks in the range of 776−693 cm−1, which corresponded to the vibrational characteristics (symmetric stretching) of the P−O−P bridge. These detected peaks may be due to the mixing phases of the metal pyrophosphate compounds, i.e., Ca2P2O7 and CuCaP2O7. In addition, the rocking mode of the P−O−P deformations, the torsional, and external modes were found in the 450−410 cm−1 regions.
The Raman spectroscopic technique was additionally applied to investigate and support the FTIR results, especially the vibrational spectroscopy of the metal oxide (M−O) bond as well as the lattice vibration by observing at the low frequency range (650−100 cm−1). Furthermore, the phase characteristics (α-, β-phases) of the metal pyrophosphate compounds can be observed from this spectroscopic technique. After applying the DXR Raman microscope, the Raman spectra of samples are shown in Fig. 9, and the corresponding vibrational assignments are listed in Table 4. It was observed that the result corresponded well to the FTIR result. The Raman results showed the specific phase, which formed at high temperature in the pyrophosphate with x = 1.00 (CaCuP2O7) as described in the literature 46 through undetectably weak peak at about 1210 cm−1. The three distinct peaks of the Ca2−xCuxP2O7, where x = 0.00, 0.50, 1.50 and 2.00, that originated from the νas PO3 vibrational characteristics, were observed and found to be at about 1210, 1140 and 1080 cm−1. The Raman spectra was clearly noticeable that the studied metal pyrophosphates displayed very sharpness and splitting, especially in the investigated frequency region (1300−100 cm−1). The vibrational analysis of the P2O74− ion, which contained the O−P−O radical (PO2− of O2O−P−OPO34−) and the P−O−P bride (of O3P−O−PO34−) were exhibited in Raman spectra. Moreover, the M−O stretching and the phase characteristics were also observed. The Raman spectra observed in this research were similar to the spectra reported by Sronsri et al. 12, 13, 15-17 and Boonchom et al. 45. The strong vibrational band around 1100 cm−1 was recognized to the stretching of the PO3 unit of the O3P−O−PO34−. The asymmetric (νasym POP) and symmetric (νsym POP) stretching types of the P−O−P bridge of the O3P−O−PO34− were detected at about 960 and 730 cm−1, respectively. While the asymmetric (δasym PO3) and symmetric (δsym PO3) vibrational bending modes of the O3P−O−PO34− were observed at about 600 and 520 cm−1, respectively.
Table 4. Vibrational positions (wavenumber / cm−1) and vibrational assignments (modes) of the synthesized Ca2−xCuxP2O7 samples obtained from the FTIR and Raman techniques
Compounds
|
Wavenumber / cm−1
|
FTIR
|
Raman
|
Assignments
|
Ca2−xCuxP2O7;
x = 0.00−2.00
|
1250 − 1200
|
NO
|
weak
|
α-phase characteristics
|
1200 − 1100
|
very strong
|
very strong
|
νas PO3
|
1100 − 1050
|
very strong
|
very strong
|
νs PO3
|
1050 − 1000
|
very strong
|
very strong
|
νas PO3
|
980 − 950
|
strong
|
very weak
|
νas P−O−P
|
760 − 730
|
medium
|
weak
|
νs P−O−P
|
650 − 280
|
strong
|
strong
|
δ OPO + νs M−O
|
260 − 160
|
NO
|
weak
|
νs M−O
|
200 − 100
|
NO
|
medium
|
lattice vibration
|
NO is not observed.
Dielectric and optical properties.
Structural-dielectric relation.
Bond angle and bond length were successfully investigated by using the Rietveld refinement technique as shown in Table 1. The obtained refinement results were then used to describe the phenomena of the dielectric properties of samples. In general, the dielectric properties of the metal pyrophosphate (M2P2O7) group occurred from two important effects, which consisted of O-atom shifting in the collinear P−O−P bridge, and M2+-ion movement in the MO6 octahedral. According to the previous works, due to the shifting of O atom in the collinear P−O−P bridge, the P−O−P bond angles of Ca2P2O7 and Cu2P2O7 of 130° 47 and 157° 48 were reported, respectively. In this section, only three synthesized Ca2−xCuxP2O7 samples, when x = 0.00, 1.00, and 2.00, were considered. The sample, when x = 0.00 (Ca2P2O7), showed two different P−O−P bond angles. First, the bond angle of 116.52o appeared for 4 clusters per a unit cell with the asymmetric P−O bond lengths of 1.765 Å and 1.887 Å. Second, the P−O−P bond angle of 140.96° appeared for 4 clusters per unit cell with the symmetric P−O bond lengths of 1.536 Å and 1.827 Å. The sample, when x = 2.00 (Cu2P2O7), had a P−O−P bond angle of 154.6°, and appeared of 4 clusters per a unit cell with the symmetric P−O bond length of 1.574 Å. Pogorzelec-Glaser et al. 46 reported that at high temperature, the binary metal pyrophosphate (CuM2+P2O7) compounds crystallized in the monoclinic crystal system with the space group of C2/m, and the P−O−P angle was linear (180°). The sample, when x = 1.00 (CaCuP2O7), exhibited the space group of P21/n. The refinement result showed the P−O−P bond angle of 159.00o and appeared of 4 clusters per unit cell with asymmetric P−O bond lengths of 1.592 Å and 1.521 Å. However, the number of P−O−P clusters in each Ca2−xCuxP2O7 sample (x = 0.00, 1.00, and 2.00) was equal (4). Based on these obtained results, the P−O−P cluster number did not affect the polarization of samples.
The single metal pyrophosphate, when x = 0.00 (Ca2P2O7), showed outstanding the dielectric constant (15.6, as shown in Fig. 6). It was very high polarization; it therefore caused and made the narrow P−O−P bond angle. In addition, the long P−O bond length of the sample of x = 0.00 (Ca2P2O7), resulting in the weak bonding, was better polarization than the samples of x = 1.00 (CaCuP2O7) and x = 2.00 (Cu2P2O7). Additionally, the volume of the octahedral coordination was calculated using the method reported by Swanson et al. 49 to present the relationship between the polarization and metal oxide bonding. Besides, the distortion index (D) was used to describe the distortion of the sample crystal structure, Baur 50 described the calculation of D value based on the bond lengths as shown in Eq. (5).
(5)
where lav is the average bond length, and li is the atomic distance from the central atom to the ith coordinating atom.
The refinement analysis results also showed the changing of the average M−O bond lengths in the MO6 octahedral site, which caused the molecular polarization. As demonstrated in Table 5, both the average bond lengths and octahedral volumes decreased with increasing x values. However, the different result was observed for the distortion index. The distortion index values increase with increasing x values, which then decreased the molecular polarization, resulting in the decreasing of the dielectric constant (εr). These analyses showed that the polarization of the Ca2−xCuxP2O7 occurred due to the O shifting in the collinear P−O−P bridge, which is the main factor of the generation of narrow bond angle that caused the high polarization and high dielectric constant. Moreover, the movement of M2+ ions in the MO6 octahedral was supplementary factor, in which the longer average M−O bond length and larger octahedral volume, leading to the high polarization and high dielectric constant of the materials.
Table 5. Average bond length, octahedral volume, and distortion index of Ca2−xCuxP2O7 samples (x = 0.00, 1.00, and 2.00)
x values
|
Compounds
|
Average bond lengths / Å
|
Octahedral volumes / Å3
|
Distortion index
|
2.00
|
Cu2P2O7
|
2.1794
|
12.6198
|
0.1400
|
1.00
|
CaCuP2O7
|
2.4032
|
15.9050
|
0.0886
|
0.00
|
Ca2P2O7
|
2.4479
|
18.6696
|
0.0655
|
Structural-optical relation.
The distortion of the MO6 octahedral can increase the Cu−O6 bond lengths of the Ca2−xCuxP2O7, resulting in the increasing of octahedral crystal field splitting energy (D0, please see Fig. 10). The D0 values of the synthesized Ca2−xCuxP2O7 samples (x = 0.50−2.00) are listed in Table 6.
Table 6. Approximate wavelength of the energy absorption
x values
|
Compounds
|
Wavelengths / nm
|
D0 / kJ·mol−1
|
Cu−Oax bond lengths / Å
|
Observed
|
Absorbed
|
XRD
|
EXAFS
|
2.00
|
Cu2P2O7
|
561
|
403
|
297
|
2.92
|
2.91
|
1.50
|
Ca0.5Cu1.5P2O7
|
488
|
642
|
186
|
ND
|
ND
|
1.00
|
CaCuP2O7
|
489
|
644
|
185
|
2.81
|
2.89
|
0.50
|
Ca1.5Cu0.5P2O7
|
492
|
648
|
184
|
ND
|
ND
|
0.00
|
Ca2P2O7
|
ND
|
ND
|
ND
|
ND
|
ND
|
ND is not detected.
As presented in Table 6, the D0 increased with increasing of the Cu2+ fraction in the Ca2−xCuxP2O7 compound, and when x = 2 (Cu2P2O7), the highest D0 value was obtained. The compounds illustrated the change of color from blue-green to bluish-green. The colorless compound, when x = 0.00 (Ca2P2O7), was due to the fulfillment state in the octet rule of Ca2+ ion in the structure, despite the distortion appearing in the CaO6 octahedral site. The octahedral splitting diagram of Ca2−xCuxP2O7; x = 0.50−2.00 were summarized and presented in Fig. 10. Total interpretations showed that the MO6 octahedral distortion affected both color of the sample and the polarization of octahedral unit, as reflected on the dielectric constant of the compounds.