PCCP View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2014, 16, 12848
View Journal | View Issue
Interaction of crystal water with the building block in Y2Mo3O12 and the effect of Ce3+ doping Xiansheng Liu, Yongguang Cheng, Erjun Liang* and Mingju Chao Ce3+ ions are introduced into the lattice of Y2Mo3O12 with a sol–gel method with the aim to reduce its hygroscopicity and pursue the interaction of crystal water molecules with the building block. It is found that Ce3+ ions occupy the positions of Y3+ in the lattice and have the function of expelling crystal water molecules in the microchannels so that the number of crystal water molecules decreases significantly as the Ce3+ content increases and a complete depletion of the crystal water is achieved when the content of Ce3+ is higher than 8 mol%. Based on the binding energy changes of Mo 3d and Y 3d with and without Ce3+ in the lattice, the configuration of the crystal water in the building block is deduced, namely, a crystal water serves as a spring with its O2 pointing to the Y3+ in an octahedron and with its H+ approaching the
Received 11th January 2014, Accepted 7th April 2014 DOI: 10.1039/c4cp00144c
next nearest O2 in the Y–O–Mo bridge. With such a configuration, the effects of the crystal water on the thermal expansion properties of Y2Mo3O12 and the like are explained. It is also shown that the number of crystal water molecules per molecular formula can be quantified by the full width at half maximum of the Raman bands or relative intensity with linear relationships, suggesting that Raman spectroscopy can be a
www.rsc.org/pccp
potential tool in quantifying crystal water molecules at room temperature in this or related materials.
1 Introduction Compounds with negative thermal expansion (NTE) have attracted much attention since the discovery of the NTE of ZrW2O8 in a large temperature range (0.3 to 1050 K).1–12 However, the wide application of NTE materials is not realized thus far due to some critical disadvantages, such as unsuitable phase transition temperature,1,2,4 strong hygroscopicity,3,13–15 unstable mechanical properties,16 too high/low or narrow temperature range for NTE,8,17,18 with poisonous ions,9,12 etc. Y2M3O12 (M: W, Mo) belonging to the A2M3O12 family with an orthorhombic structure has a strong NTE property but it is highly hygroscopic, like other members with a larger A3+ cation size. Its NTE was only observed after the complete release of the crystal water. It has a building block structure with corner-sharing YO6 octahedra and MoO4 tetrahedra. It was found that crystal water molecules residing in the microchannels of the building blocks hinder not only the external librational and translational motions but also the internal stretching and bending vibrations of the polyhedra and hence its NTE property.3,15,16 Crystal water molecules in the microchannels result in a large but different contraction for different axes of the lattice while the release of them leads to abrupt expansion of the lattice differently for different directions.15,19,20 This should School of Physical Science & Engineering and Key Laboratory of Materials Physics of Ministry of Education of China, Zhengzhou University, Zhengzhou 450052, China. E-mail: [email protected]; Fax: +86 371 67766629; Tel: +86 371 67767838
12848 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
be responsible for the amorphous state and poor mechanical properties of the structure.14,16 From both a scientific and application point of view, it is therefore significant to pursue effective methods to reduce the hygroscopicity in Y2M3O12 and related materials. On the other hand, the macroscopic effects of crystal water on the thermal expansion properties in Y2M3O12 and the like is nearly clear, but the microscopic mechanisms of the interaction between the water molecules and the building blocks remain unclear. For example, hydrogen bonding of water with possible oxygen positions in the building block was assumed.17 However, it does not seem possible that one-end hydrogen bonding of water with ‘‘O’’ in the building block could make so large a contraction and distortion of the lattice. Why the crystal water causes such large contraction on the building blocks in Y2Mo2O12 and related materials remains unresolved. Although many efforts have been made to reduce the hygroscopicity in Y2M3O12 and YbM3O12 (M: W, Mo), the problems mentioned above remain unresolved. When the larger A3+ cation is partially replaced by smaller cations, such as Y3+ partially substituted by Fe3+ (64.5 pm),17 the hygroscopicity can be reduced only to some extent if the orthorhombic phase remains. A higher content of substitution results in a monoclinic phase which has a large positive thermal expansion. In some cases, such as in Yb2xAlxMo3O12 and Yb2xCrxMo3O12,20,21 a pure orthorhombic phase is only possible for an even smaller content substitution (0.0 r x r 0.4) so that the materials are still highly hygroscopic. Otherwise, monoclinic or monoclinic–orthorhombic composited
This journal is © the Owner Societies 2014
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
Paper
phases appear. The substitution of La3+ (25 mol%) for Y3+ results in the formation of a monoclinic phase and large thermal expansion below 373 K, suggesting that it is hygroscopic.22 To eliminate the hygroscopicity in orthorhombic Y2Mo3O12 and the like is still challenging. In this paper, Ce3+ ions are introduced into the Y2Mo3O12 lattice with a sol–gel method with the aim to reduce its hygroscopicity, pursue the interaction of crystal water molecules with the building blocks as well as tailor its thermal expansion property. To reach these aims, temperature dependent X-ray diffraction (XRD), high resolution transmission electron microscopy (HRTEM), temperature dependent Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), differential scanning calorimetry (DSC) and thermogravimetric (TG) as well as dilatometry analyses are used. It is found that Ce3+ ions occupy the positions of Y3+ in the lattice and the incorporation of Ce3+ has the function of expelling crystal water molecules in the microchannels of Y2Mo3O12. As the Ce3+ content increases, the number of crystal water molecules decreases significantly so that the crystal water molecules are completely removed when the content of Ce3+ is higher than 8 mol% and a low thermal expansion material can be achieved without hygroscopicity. The crystal water molecules have an overwhelming effect on the changes of the Raman spectra, causing significant broadening on admission and sharpening on the release of crystal water molecules, as well as relative intensity changes of the Raman bands. It is shown that the number of crystal water molecules per molecular formula can be quantified by the full width at half maximum of the Raman bands or relative intensity with linear relationships. Based on the binding energy changes of Mo 3d and Y 3d with and without Ce3+ in the lattice, the configuration of th crystal water in the building block is presented. With such a configuration, the effects of the crystal water on the thermal expansion behaviors can be well understood.
2 Materials and methods Analytical grade Y(NO3)39H2O, (NH4)6Mo7O244H2O and Ce(NO3)36H2O were used as raw materials. C6H8O7 was used as a complexing agent. 30 mL (0.01 mol) aqueous solutions of Y(NO3)39H2O containing 0, 1, 2, 3, 4, 5, 6, 8, 9 and 10 mol% contents of Ce3+ from Ce(NO3)36H2O were prepared with deionized water and stirred with a magnetic stirrer for 1 h at 323 K. Transparent sols were formed after adding C6H8O7 (0.1 mol) (its molar ratio was ten fold that of Y(NO3)39H2O) to the above solutions. (NH4)6Mo7O244H2O was then put into each of the sols to maintain a molar ratio of (Y + Ce) : Mo = 2 : 3 in each sol. Each sol was divided into four parts in different evaporating dishes and their pH values were tuned to 1, 3, 5, and 7, respectively, with NH3H2O (30%) and HNO3 (30%). They were stirred rigorously at 353 K for 6 h and then heated in a blast oven at 408 K for 12 h to become gels. They were then heated at 623 K for 3 h and at 1023 K for another 3 h in a muffle furnace. The products were cooled naturally to room temperature for analysis. Some powders were measured directly and some were pressed into pellets (diameter of 12 mm and thickness of 2 mm)
This journal is © the Owner Societies 2014
PCCP
or cylinders (diameter of 6 mm and height of 10 mm) according to measurement requirements. XRD measurements were carried out with an X-ray diffractometer (Model X’Pert PRO) to identify the crystalline phase. Raman spectra were recorded on a laser Raman spectrometer (Renishaw MR-2000) with an excitation wavelength of 532 nm. The temperature dependence of the Raman spectra was recorded by using a TMS 94 heating/freezing stage from Linkam Scientific Instruments Ltd., with an accuracy of 0.1 K. High resolution transmission electron microscopy (HRTEM JEM-2010, JEOL) was used to observe the lattice fringes. Differential scanning calorimetry (DSC) and thermogravimetric analysis (TG) were done on an Ulvac Sinku-Riko DSC (Model 1500M/L) in the temperature range of 293–873 K with heating and cooling rates of 10 K min1. XPS (Axis Ultra, Kratos, UK) was used to analyze the composition of the samples and oxidation states of the elements. The linear thermal expansion coefficient was measured by a dilatometer (LINSEIS DIL L76).
3 Results and discussion 3.1
XRD and TEM characterizations for Ce3+-doped Y2Mo3O12
The influence of the pH values and the content of Ce3+ on the final products were studied by XRD. Fig. 1a and b shows the XRD patterns of Ce3+-doped Y2Mo3O12 prepared at pH = 1 and 5, respectively. It is shown that both the pH values in the initial sols and the nominal content of Ce3+ influence the final structure of the samples, as revealed by the abrupt change in the XRD patterns occurring between 2 and 3 mol% of Ce3+ for the samples prepared at pH = 1 and between 4 and 5 mol% of Ce3+ for the samples prepared at pH = 5. For the samples with Ce3+ content r2 mol% (pH = 1) or r4 mol% (pH = 5), the XRD patterns resemble that of highly hydrated Y2Mo3O12 3H2O,14,15,19,23 indicating that these samples retain the structure of Y2Mo3O123H2O and are still highly hydrated. The XRD patterns at or above a 3 mol% (pH = 1) or 5 mol% (pH = 5) content of Ce3+ resemble that of orthorhombic Y2Mo3O12 after the release of crystal water.14 It is obvious that the diffraction peaks shift progressively to lower angles on increasing the content of Ce3+, suggesting an increase in the lattice constant due to the larger radius of Ce3+ than Y3+ (effect of crystal water exists also). In order to clarify the sole effect of the introduction of Ce3+, we performed XRD patterns of the samples with 0, 2, 5 and 9 mol% Ce3+ at 413 K (at 413 K, crystal water is released) and indexing of the XRD pattern of Y2Mo3O12 at 413 K, as shown in Fig. 1c and d. Table 1 lists the lattice constants and unit cell volumes of Ce3+-doped Y2Mo3O12 (pH = 1) from structure refinements of the XRD data using the Rietveld method. It is found that, for crystals with the same space group, the lattice constants increase on increasing the content of Ce3+ (0 - 2 mol%, 5 - 9 mol%). For crystals with different space groups, it is not suitable to compare their lattice constants. We have also made use of the data to draw the building blocks of Y2Mo3O12 crystals with 0, 2, 5 and 9 mol% Ce3+ (Fig. 1e). The structures with the same space groups
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12849
View Article Online
Paper
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PCCP
Fig. 1 (a–d) XRD patterns of Ce3+-doped Y2Mo3O12 prepared from the sols with different pH values: (a) pH = 1; (b) pH = 5; (c) XRD patterns of the samples with 0, 2, 5 and 9 mol% Ce3+ at 413 K (pH = 1); (d) indexing of the XRD pattern of Y2Mo3O12 at 413 K (pH = 1). (e) Diagrams of the Ce3+-doped Y2Mo3O12 building blocks with 0, 2, 5 and 9 mol% Ce3+: the space groups of Y2Mo3O12 with 0, 2 mol% Ce3+ are Pbcn and the space groups of Y2Mo3O12 with 5, 9 mol% Ce3+ are Pba2.
change less but notable changes appear after the space group transforms from Pbcn to Pba2. The bonding length and angles change markedly with different space groups.
12850 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
Fig. 2 shows the HRTEM images of the lattice fringes and diffraction patterns for Y2Mo3O12 (Fig. 2a, b) and 2 mol% Ce3+-doped Y2Mo3O12 (Fig. 2c, d) prepared at pH = 1. It is shown
This journal is © the Owner Societies 2014
View Article Online
Paper
PCCP
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
Table 1 Constants (a, b, c, d and volume) of Y2Mo3O12 samples with 0, 2, 5 and 9 mol% Ce3+ at 413 K
Content of Ce3+ (%)
Space group
a (Å)
b (Å)
c (Å)
v (Å3)
d(111) (nm)
0 2 5 9
Pbcn Pbcn Pba2 Pba2
13.8726 13.9028 10.3542 10.3790
9.9249 9.9410 10.3704 10.3797
10.0033 10.0136 10.6016 10.6026
1377.2809 1383.9495 1138.3662 1142.2247
0.6282 0.6291 0.6028 0.6035
Fig. 2 HRTEM images of Y2Mo3O12 (a, b) and 2 mol% Ce3+-doped Y2Mo3O12 (c, d) prepared at pH = 1.
that the interplanar spacing along the (111) direction of Y2Mo3O12 increases primarily on increasing the content of Ce3+. The interplanar spacing increases from 0.6035 nm for Y2Mo3O12 to 0.6045 nm for 2 mol% Ce3+-doped Y2Mo3O12, which are smaller than the samples with 0 and 2 mol% Ce3+ in Table 1 due to shrinkage of the lattice induced by crystal water at room temperature. Fig. 1 and 2 demonstrate that Ce3+ was incorporated into the lattice of Y2Mo3O12 and the Ce3+-doped Y2Mo3O12 crystallized in an orthorhombic structure. The samples are obviously hygroscopic for Ce3+ contents of o3 mol% (pH = 1) or o5 mol% (pH = 5). As we will see later, though the distinctive change in the XRD patterns occur at the Ce3+ content of around 3 mol% (pH = 1)
Fig. 3
or o5 mol%, the samples still contain crystal water until a 8–9 mol% content of Ce3+, as revealed by our Raman spectroscopic, DSC and dilatometric analyses. 3.2 Influence of the content of Ce3+ and crystal water on the Raman spectra of Y2Mo3O12 Fig. 3a and b show the Raman spectra of Ce3+-doped Y2Mo3O12 samples with different Ce3+ content prepared at pH = 1 and pH = 5, respectively. The Raman modes appearing in the ranges of 970–900 cm1 and 900–740 cm1 can be ascribed to symmetric and asymmetric stretching vibrations of the MoO4 tetrahedra, respectively,15,19 while those from 400 to 300 cm1 are attributed to symmetric and asymmetric bending motions in both the YO6 octahedra and MoO4 tetrahedra, and those modes below 300 cm1 are translational and librational vibrations.15,19,24 The common features of the Raman spectra in Fig. 3a and b are the appearance of only three broad Raman bands at around 940, 824 and 333 cm1 for lower contents of Ce3+ and the Raman bands split into sharp modes and increase in intensity on further increasing the content of Ce3+. However, the distinct changes in the Raman spectra start to occur at different contents of Ce3+ for the samples prepared at different pH values, 3 mol% for pH = 1 (see Fig. 3a) and 5 mol% for pH = 5 (see Fig. 3b), in accordance with the XRD analyses. This suggests that the pH values have an obvious effect on the real content of Ce3+ in the final products and a lower pH value seems more favorable for the Ce3+ ions to incorporate into the lattice of Y2Mo3O12. It is interesting to note that the behaviors of the Raman spectra on increasing the content of Ce3+ are quite similar to that of the Raman spectra of Y2Mo3O12 with increasing temperature, in which only three broad Raman bands appear at room temperature and the Raman bands start to split and increase in intensity above 403 K.15,19 Since Y2Mo3O12 is highly hygroscopic, the enormous change in the Raman spectra was attributed to the release of crystal water. The crystal water molecules in the microchannels hinder not only the librational and translational motions, but also the stretching and bending vibrations of the polyhedra.15 It can be inferred that the distinct changes in the Raman spectra on increasing the content of Ce3+ are related to the reduction of crystal water molecules. Namely, the incorporation of Ce3+ in the lattice of Y2Mo3O12 has the ability to inhibit water
Dependence of the Raman spectra of Ce3+-doped Y2Mo3O12 on the molar percent of Ce3+ prepared at (a) pH = 1 and (b) pH = 5.
This journal is © the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12851
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PCCP
Paper
Fig. 4 Temperature dependence of the Raman spectra of Y2Mo3O12 doped with 2 mol% Ce3+.
species entering the microchannels in the building blocks. The continuous increase and sharpening of the Raman bands on increasing the content of Ce3+ (>3 mol% for pH = 1; and >5 mol% for pH = 5) may indicate a successive reduction of the crystal water molecules. In order to confirm the above assumption and gain insight into the effect of Ce3+ on the crystal water, we carried out a series of temperature-dependent Raman spectroscopic studies (the samples prepared at pH = 1 are used in the following measurements). Fig. 4 shows the Raman spectra of Y2Mo3O12 with 2 mol% Ce3+ at different temperatures. On increasing the temperature from 293 K to 383 K, there is no significant change in the Raman spectra except that the Raman bands at about 943, 829 and 342 cm1 shift progressively to 952, 823 and 337 cm1, respectively. The Raman spectra start to change enormously at about 393 K. An intense Raman band appears at about 966 cm1 with a shoulder at about 936 cm1. The asymmetric stretching Raman band at about 823 cm1 becomes sharper and much more intense. The bending mode shifts from 337 to 326 cm1. Above 393 K, both the symmetric and asymmetric Raman modes increase in intensity with slight blue shifts with increasing temperature until 443 K. Generally, Raman bands weaken and broaden with increasing temperature. The abrupt changes in the Raman spectra at about 393 K and the abnormal behavior above this temperature are attributed to the release of crystal water molecules, which form hydrogen bonds with the O2 ions in the Y–O–Mo bridges.19 The hydrogen bonding attracts a partial amount of the electrons from the bridging oxygen toward the hydrogen bonds and hence weakens the O–Mo bonds. This is confirmed by the red shift of the stretching Raman bands with decreasing temperature (more crystal water molecules) or the blue shift of the Raman bands with increasing temperature (less crystal water molecules). A Raman band frequency
Raman mode shift should be included. We therefore give the Raman band frequency as o(T) = o0 + D(T) + D(m), where D(m) depends on the number of crystal water molecules, m, and accounts for the crystal-water induced shift. With temperature increasing, D(T) leads to a red shift of a Raman band while D(m) results in a blue shift due to the reduction of the number of crystal water molecules. From RT to 393 K and from 393 to 443 K, the effects caused by the release of crystal water molecules and by temperature increasing coexist. The blue shifts of the stretching modes are hence the result of the competition of the two effects. From RT to 393 K, only the weakly interacting crystal water molecules are able to be removed, while the crystal water molecules interacting strongly with the bridging oxygens start to be released at about 393 K and above, as indicated by the abrupt change in the Raman spectrum around 393 K. From 443 K to 573 K, the Raman band at 966 cm1 shifts to 962 cm1 and that at 823 cm1 shifts to 820 cm1. The red shifts can be considered as a pure temperature-induced effect. Though the Raman bands already start to split at 3 mol% Ce3+, we demonstrate here that the samples at this or even higher contents of Ce3+ are still hygroscopic. Fig. 5 shows the Raman spectra of Y2Mo3O12 doped with 5 mol% of Ce3+. It is clear that the obvious changes in the Raman spectra occur at about 393–403 K, as indicated by the intensity inverse of the Raman bands at about 823 and 962 cm1 with respect to the Raman bands at 853 and 944 cm1, respectively, as well as the increase of the Raman bands on increasing the temperature to above 403 K. As discussed above for the sample with 2 mol% of Ce3+, the obvious changes in the Raman spectra between 393 and 403 K should be correlated to the release of crystal water. These suggest that this sample still contains a certain amount of the crystal water molecules, which start to be released at about 403 K. Detailed analyses for other samples show that the Raman spectra for the samples with Ce3+ content r8 mol% behave similarly with increasing temperature (not shown here). However, no obvious changes in the Raman spectra are observed for the samples with >8 mol% Ce3+. Fig. 6 shows the Raman spectra of Y3Mo3O12 doped with 9 mol% Ce3+ at different temperatures. The Raman spectra do not exhibit obvious changes between 383 and 403 K, in contrast
can be expressed as o(T) = o0 + D(T), where DðTÞ ¼ C 1þ
2
" þD 1þ
3
þ
3
#
is ðeho=3kB T 1Þ2 the temperature-induced shift of a Raman band, and C and D are anharmonic constants with negative values for normal Raman modes.25 In the present case, a crystal water-induced eho=2kB T 1
eho=3kB T 1
12852 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
Fig. 5 Temperature dependence of the Raman spectra of Y2Mo3O12 doped with 5 mol% Ce3+.
This journal is © the Owner Societies 2014
View Article Online
Paper
PCCP
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
Increasing the content of Ce3+ causes an increase in DGdop and a decrease in the number of crystal water molecules and hence a reduction in DGcw. Fig. 7 suggests that the decrease in crystal water molecules has an overwhelming effect on the bandwidths and relative intensities when the content of Ce3+ is less than 8 mol%. The changes in the FWHMs and relative intensities above 8 mol% Ce3+ may be regarded as a pure effect of Ce3+-doping. 3.3 Correlation of the number of crystal water molecules with features of the Raman spectra Fig. 6 Temperature dependence of the Raman spectra of Y2Mo3O12 doped with 9 mol% Ce3+.
to the temperature-dependent Raman spectra of Y3Mo3O12 doped with 2 and 5 mol% Ce3+. It is reasonable that the crystal water molecules are completely eliminated for the sample with a 9 mol% Ce3+ content. This is an exciting result because complete elimination of crystal water in Y3Mo3O12 and the like at ambient conditions is important for applications and has not really been achieved, though many efforts have been made.19–21,26 A comparison of Fig. 3–5 shows that the Raman bands at about 853 cm1 are mainly influenced by the incorporation of Ce3+ instead of crystal water. In order to correlate the effect of the crystal water molecules on the changes in the Raman bands in Fig. 3, we show in Fig. 7 the changes in the full width at half maximum (FWHM) for the four intense stretching Raman bands at about 823, 853, 944 and 962 cm1 and their relative intensities with respect to the Raman band at about 853 cm1 with the Ce3+ content. (All the Raman spectra were fitted with Lorentzian peak profiles and the data were derived from the curve fittings.) It is shown that the bandwidths decrease and the relative intensity changes obviously on increasing the content of Ce3+ until 8–9 mol%. We attribute the decrease in the FWHM to the decrease in the content of crystal water molecules because a Raman band linewidth generally broadens after doping due to lattice distortion. A Raman band linewidth, G, in our case can be expressed by the sum of the natural linewidth, G0(T), of Y2Mo3O12 at room temperature, Ce3+ doping-induced broadening, DGdop, and crystal water-induced broadening, DGcw, G = G0(T) + DGdop + DGcw.
DSC and TG measurements were performed to confirm the above analyses by Raman spectroscopy (see Fig. 8). The endothermic peaks at around 423 K in the DSC curves (Fig. 8a) and the corresponding weight loss (Fig. 8b) for the samples with a Ce3+ content of r8 mol% confirm the existence of crystal water molecules. However, neither endothermic peaks nor weight losses are observed for the sample with 9 mol% Ce3+. This indicates that the samples with Z9 mol% Ce3+ do not contain crystal water molecules, confirming our Raman analyses above. The number of crystal water molecules per molecular formula is calculated to be 2.48, 2.46, 2.12, 1.46, 0.74, 0.43, 0.12 and 0.0 for the samples with 0, 1, 2, 3, 5, 6, 8 and 9 mol% contents of Ce3+. It reduces markedly on increasing the content of Ce3+ and the crystal water is eliminated completely when the Ce3+ content is increased to 9 mol%. With these results, it is possible for us to correlate the changes in the FWHMs and relative intensities to the number of crystal water molecules, as shown in Fig. 9. By examining Fig. 7, we know that the FWHMs of the Raman bands at 832 and 944 cm1 are most sensitive to the number of crystal water molecules. Therefore, we sum the FWHMs of these two bands (denoted as G1) as well as those of all the four bands (G2) to correlate with the number of crystal water molecules in order to improve the sensitivity. Both can be well fitted with linear equations as G1 = 22.48048 + 17.22256m with R2 = 0.98792 (Fig. 9a); and G2 = 41.97288 + 18.80879m with R2 = 0.98569 (Fig. 9b), where m is the number of crystal water molecules per molecular formula. We can see that using the sum of the most sensitive bandwidths fits a little bit better than using the sum of all four bandwidths. We attribute this to the influence of the Ce3+-doping-induced bandwidth change.
Fig. 7 Dependence of the FWHMs (a) and relative intensities (b) of the selected Raman bands on the content of Ce3+.
This journal is © the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12853
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PCCP
Paper
Fig. 8 (a) DSC and (b) TG curves of Ce3+-doped Y2Mo3O12 with 0, 1, 2, 3, 5, 6, 8 and 9 mol% Ce3+.
As discussed above, the Raman band at about 853 cm1 is most sensitive to the Ce3+ content but the least influenced by the release of crystal water molecules. Keeping this in mind, the relative intensity of the 832 cm1 band with respect to the band at 853 cm1 can also be used to characterize the number of crystal water molecules, as shown in Fig. 9c. It can be expressed Ið832 cm1 Þ as ¼ 0:58669 þ 0:21605m, with R2 = 0.96183. Fig. 3– Ið853 cm1 Þ 7 and 9 indicate that Raman spectroscopy can be used as a powerful tool to quantify the crystal water molecules at room temperature and investigate their influence on the vibrational properties of materials. From the Raman measurements (Fig. 3–6) and discussion above, it is possible for us to identify the stretching Raman modes more specifically. The Raman bands at 944 and 853 cm1 are most influenced by Ce3+-doping. They can be identified as the symmetric and antisymmetric stretching vibrations of the Mo–O bonds correlated mainly with Ce3+, respectively. However, the contribution from the Mo–O bonds correlated mainly with Y3+ cannot be omitted. Similarly, the Raman bands at 960 and 823 cm1 are less influenced by Ce3+-doping. They can be identified
Correlations of the FWHMs G1 (a) and G2 (b) of the Raman bands I 832 cm1 and the relative intensity of (c) with the number of crystal I ð853 cm1 Þ water molecules.
as the symmetric and antisymmetric stretching vibrations of the Mo–O bonds correlated mainly with Y3+, respectively. 3.4 Effects of Ce3+-doping and the reduction of crystal water on the thermal expansion properties of Y2Mo3O12 Fig. 10 shows the relative length changes of Ce3+-doped Y2Mo3O12 containing 0, 1, 3, 5, 8, 9 and 10 mol% contents of Ce3+ with increasing temperature. The materials shrink distinctly from about 350 to 408 K and then expand abruptly until about 460 K for Ce3+ contents of less than 3 mol%, whereas those with 3–8 mol% Ce3+ exhibit less abrupt expansion from about 350 to 430 K. In this temperature range, the thermal expansion behavior is dominated by the release of crystal water. It is clear that the interaction of crystal water molecules with the building block of Y2Mo3O12 is so strong that they make the building blocks contract considerably and the release of them leads to an abrupt expansion of the building blocks. After complete release of the crystal water molecules, the materials show NTE (8.94 106 K1, 9.22 106 K1 and 2.25 106 K1 for 0, 1 and 3 mol% Ce3+, respectively) or low positive expansion (1.67 106 K1 and 2.27 106 K1 for 5 and 8 mol% Ce3+). The samples with Ce3+ contents higher than 8 mol% exhibit low thermal expansions without abnormal changes in the whole temperature range (2.743 106 K1 and 3.53 106 K1 for 9 and 10 mol% Ce3+), suggesting that they do not possess crystal water molecules at all. The thermal
Fig. 9
12854 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
Fig. 10 Relative length change (dL/L0 (%)) with temperature of Ce3+-doped Y2Mo3O12 (pH = 1, Ce3+ content of 0, 1, 3, 5, 8, 9, and 10 mol%).
This journal is © the Owner Societies 2014
View Article Online
Paper
expansion properties agree well with the Raman and DSC analyses above. In a previous paper, we demonstrated that the c-axis and b-axis were more influenced by crystal water than the a-axis in Fe2xYxMo3O12.19
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
3.5 Mechanism of crystal water reduction in Ce3+-doped Y2Mo3O12 To explore the effect of Ce3+ doping on the reduction of crystal water, we measured the binding energies of Y2Mo3O12 and 5 mol% Ce3+-doped Y2Mo3O12 by XPS. Fig. 11a–f shows the survey spectra of Y2Mo3O12 and Ce3+-doped Y2Mo3O12, and the core level spectra of Y 3d, Mo 3d and O 1s, respectively. After the introduction of Ce3+, though the spin–orbit splitting of Y 3d (1.9 eV) and Mo 3d (3.2 eV) do not change, the binding energy of Y 3d5/2 decreases from 158.1 to 156.0 eV,27 which is much more distinct than the binding energy shift of Mo 3d5/2 (232.7 - 232.6 eV) (Fig. 11b and c). Considering this fact together with the valence and cation radii difference, it is reasonable that Ce3+ ions take the positions of Y3+ rather than Mo6+ in Y2Mo3O12. The O 1s peak is obviously asymmetric in Y2Mo3O12 but it becomes more symmetric in Ce3+-doped Y2Mo3O12 (Fig. 11d). Deconvolution of the O 1s region shows three peaks (Fig. 11e). The first, located at 530.6 eV, is assigned to Mo–O–Y in Y2Mo3O12, whereas the second at 531.8 eV and third at 533.4 eV are associated with hydroxyl groups.28 This explains why Y2Mo3O12 is highly hygroscopic. Nevertheless, after introducing Ce3+, the O 1s peak can only be deconvoluted into two peaks, suggesting suppression in the hydroxyl groups. This confirms that the introduction of Ce3+ has the ability to repel crystal water molecules.
PCCP
In a recent paper, we assumed that the protruded ‘‘O’’ anions in the microchannels of the building blocks had a higher probability to interact with water species to form hydrogen bonds.19 The distinct changes in the stretching Raman bands of MoO4 tetrahedra before and after the release of crystal water molecules also support the hydrogen bonding to the bridging oxygen ions. However, if the water molecules interacted with the bridging oxygen ions only, it would be difficult to understand so large a lattice contraction caused by crystal water molecules or expansion on the release of them. Comparisons of the hygroscopicity between Y2Mo3O12 and Ce3+-doped Y2Mo3O12 and the binding energy changes in Y 3d5/2,3/2 and O 1s after the substitution of Ce3+ for Y3+ throw light on the configurations of the crystal water molecules in the microchannels, i.e. the crystal water molecule has its oxygen ion approaching the Y3+ in the octahedra. From these discussions, it is possible for us to speculate on the configurations of the crystal water molecules in the building blocks. That is, a water molecule is oriented with its O2 pointing to the Y3+ in an octahedron and with its H+ approaching the next nearest bridging O2 in the Y–O–Mo bridge to form hydrogen bonding, as depicted in Fig. 12. With this configuration, the effects of crystal water on the thermal expansion properties, as presented in Fig. 10, can be well explained. The crystal water serves as a spring to drag connected octahedra and tetrahedra to rotate, making the c-axis contract more obviously than the b-axis. In such a configuration, the a-axis is least influenced by the crystal water. The release of the crystal water molecules leads to a restoration of the lattice and hence an abrupt expansion around the releasing temperature of the crystal water. As the content of
Fig. 11 The survey spectra of Y2Mo3O12 and 5 mol% Ce3+-doped Y2Mo3O12 (a); and the core level spectra of Y 3d (b), Mo 3d (c) and O 1s (d–f).
This journal is © the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12855
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PCCP
Fig. 12 Schematic of the Y2Mo3O12 building block and configuration of the crystal water molecule.
Ce3+ increases, there are fewer or no crystal water molecules, causing less or no obvious contraction of the lattice. Therefore, the linear thermal expansion coefficients change less abruptly or even change smoothly with increasing temperature. Since Ce has a lower electronegativity (1.12) than Y (1.22), Ce3+ has a lower capacity than Y3+ to attract ‘‘O’’ in the water molecules. Fig. 8 suggests a successive decrease in the average binding energy of hydrogen bonds of crystal water molecules on increasing the content of Ce3+. A first principles calculation shows a decrease in the binding energy of a water molecule with the building block from 0.6 eV to 0.2 eV by Ce-doping (unpublished data from Prof. Qiang Sun, Zhengzhou University). This explains, at least partly, why the Ce3+-doping eliminates the crystal water.
4 Conclusions Ce3+ ions are introduced into the lattice of Y2Mo3O12 with a sol–gel method with the aim to reduce its hygroscopicity, pursue the interaction of the crystal water molecules with the building blocks as well as tailor its thermal expansion property. It is found that Ce3+ ions occupy the positions of Y3+ in the lattice and the incorporation of Ce3+ expels crystal water molecules in the microchannels of Y2Mo3O12. As the Ce3+ content increases, the number of crystal water molecules decreases significantly so that the crystal water molecules can be completely removed when the content of Ce3+ is larger than 8 mol% and a low thermal expansion material can be achieved without hygroscopicity. Based on the binding energy changes of Mo 3d and Y 3d with and without Ce3+ in the lattice, the configuration of the crystal water in the building block is deduced. The crystal water molecule is oriented with its O2 ion pointing to the Y3+ in an octahedron and with its H+ ion approaching the next nearest bridging O2 in the Y–O–Mo bridge to form hydrogen bonds. The crystal water molecules serve as springs to drag the connected octahedra and tetrahedra to rotate, making the c-axis contract more obviously than the b-axis. In such a configuration, the a-axis is the least influenced by the crystal water. Compared to the effect of Ce3+, the crystal water molecules have an overwhelming effect on the FWHM and relative intensity changes of the Raman bands. It is also shown that the number of crystal water molecules per molecular formula can be quantified by the
12856 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
Paper
full width at half maximum of the Raman bands or relative intensity with linear relationships, G1 = 22.48048 + 17.22256m Ið832 cm1 Þ ¼ 0:58669 þ 0:21605m, where m is the number and Ið853 cm1 Þ of crystal water molecules per molecular formula. This suggests that Raman spectroscopy can be used as a potential tool in quantifying crystal water molecules in this and related compounds at room temperature. This work provides a profound insight into the interaction mechanism of crystal water molecules in the A2M3O12 (A = Lu, Er, Yb, Y; M = W, Mo) family of compounds and the abnormal thermal expansion properties caused by crystal water also paves the way towards solving the long standing hygroscopic problem in Y2Mo3O12 and related compounds.
Acknowledgements This work was supported by the National Science Foundation of China (No. 10974183; No.11104252), by the Doctoral Fund of the Ministry of Education of China (No. 20114101110003) and the fund for Science &. Technology innovation team of Zhengzhou (No. 112PCXTD337). We also thank Prof. Qiang Sun for providing the binding energy data for the crystal water in Y2Mo3O12.
References 1 T. A. Mary, J. S. O. Evans, T. Vogt and A. W. Sleight, Science, 1996, 272, 90–92. 2 P. P. Sahoo, S. Sumithra, G. Madras and T. N. G. Row, Inorg. Chem., 2011, 50, 8774–8781. 3 B. A. Marinkovic, M. Ari, R. R. Avillez, F. Rizzo, F. F. Ferreira, K. J. Miller, M. B. Johnson and M. A. White, Chem. Mater., 2009, 21, 2886–2894. 4 K. J. Miller, C. P. Romao, M. Bieringer, B. A. Marinkovic, L. Prisco and M. A. White, J. Am. Ceram. Soc., 2013, 96, 561–566. 5 J. R. Wang, J. X. Deng, R. B. Yu, J. Chen and X. R. Xing, Dalton Trans., 2011, 40, 3394–3397. 6 R. Futamura, T. Iiyama, A. Hamasaki and S. Ozeki, Phys. Chem. Chem. Phys., 2012, 14, 981–986. 7 L. C. Gallington, K. W. Chapman, C. R. Morelock, P. J. Chupas and A. P. Wilkinson, Phys. Chem. Chem. Phys., 2013, 15, 19665–19672. 8 X. Y. Song, Z. H. Sun, Q. Z. Huang, M. Rettenmayr, X. M. Liu, M. Seyring, G. N. Li, G. H. Rao and F. X. Yin, Adv. Mater., 2011, 23, 4690–4694. 9 J. Chen, L. L. Fan, Y. Ren, Z. Pan, J. X. Deng, R. B. Yu and X. R. Xing, Phys. Rev. Lett., 2013, 110, 115901. ˜oz, J. B. Keith, S. J. Tracy, 10 C. W. Li, X. L. Tang, J. A. Mun D. L. Abernathy and B. Fultz, Phys. Rev. Lett., 2011, 107, 195504. 11 M. Azuma, W. Chen, H. Seki, M. Czapski, S. Olga, K. Oka, M. Mizumaki, T. Watanuki, N. Ishimatsu, N. Kawamura, S. Ishiwata, M. G. Tucker, Y. Shimakawa and J. P. Attfield, Nat. Commun., 2011, 2, 347. 12 A. L. Goodwin, M. Calleja, M. J. Conterio, M. T. Dove, J. S. O. Evans, D. A. Keen, L. Peters and M. G. Tucker, Science, 2008, 319, 794–797.
This journal is © the Owner Societies 2014
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
Paper
13 S. Sumithra and A. M. Umarji, Mater. Res. Bull., 2005, 40, 167–176. 14 S. D. Gates and C. Lind, J. Solid State Chem., 2007, 180, 3510–3514. 15 E. J. Liang, H. L. Huo, J. P. Wang and M. J. Chao, J. Phys. Chem. C, 2008, 112, 6577–6581. 16 C. Yuan, Y. Liang, J. P. Wang and E. J. Liang, J. Chin. Ceram. Soc., 2009, 37, 726–732. 17 A. K. Tyagi, S. N. Achary and M. D. Mathews, J. Alloys Compd., 2002, 339, 207–210. 18 T. Chatterji, M. Zbiri and T. C. Hansen, Appl. Phys. Lett., 2011, 98, 181911. 19 Z. Y. Li, W. B. Song and E. J. Liang, J. Phys. Chem. C, 2011, 115, 17806–17811. 20 Q. J. Li, B. H. Yuan, W. B. Song, E. J. Liang and B. Yuan, Chin. Phys. B, 2012, 21, 046501.
This journal is © the Owner Societies 2014
PCCP
21 M. M. Wu, X. L. Xiao, Z. B. Hu, Y. T. Liu and D. F. Chen, Solid State Sci., 2009, 11, 325–329. 22 H. F. Liu, X. C. Wang, Z. P. Zhang and X. B. Chen, Ceram. Int., 2012, 38, 6349–6352. 23 B. A. Marinkovic, P. M. Jardim, R. R. Avillez and F. Rizzo, Solid State Sci., 2005, 7, 1377–1383. 24 L. Wang, F. Wang, P. F. Yuan, Q. Sun, E. J. Liang, Y. Jia and Z. X. Guo, Mater. Res. Bull., 2013, 48, 2724–2729. 25 M. Balkanski, R. F. Wallis and E. Haro, Phys. Rev. B, 1983, 28, 1928–1934. 26 M. M. Wu, Z. B. Hu, Y. T. Liu and D. F. Chen, Mater. Res. Bull., 2009, 44, 1943–1947. ´cs, K. Magyari, M. Neumann and 27 V. Simon, D. Eniu, A. Taka S. Simon, J. Optoelectron. Adv. Mater., 2005, 7, 2853–2857. 28 J. C. Dupin, D. Gonbeau, P. Vinatier and A. Levasseur, Phys. Chem. Chem. Phys., 2000, 2, 1319–1324.
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12857
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2014, 16, 12848
View Journal | View Issue
Interaction of crystal water with the building block in Y2Mo3O12 and the effect of Ce3+ doping Xiansheng Liu, Yongguang Cheng, Erjun Liang* and Mingju Chao Ce3+ ions are introduced into the lattice of Y2Mo3O12 with a sol–gel method with the aim to reduce its hygroscopicity and pursue the interaction of crystal water molecules with the building block. It is found that Ce3+ ions occupy the positions of Y3+ in the lattice and have the function of expelling crystal water molecules in the microchannels so that the number of crystal water molecules decreases significantly as the Ce3+ content increases and a complete depletion of the crystal water is achieved when the content of Ce3+ is higher than 8 mol%. Based on the binding energy changes of Mo 3d and Y 3d with and without Ce3+ in the lattice, the configuration of the crystal water in the building block is deduced, namely, a crystal water serves as a spring with its O2 pointing to the Y3+ in an octahedron and with its H+ approaching the
Received 11th January 2014, Accepted 7th April 2014 DOI: 10.1039/c4cp00144c
next nearest O2 in the Y–O–Mo bridge. With such a configuration, the effects of the crystal water on the thermal expansion properties of Y2Mo3O12 and the like are explained. It is also shown that the number of crystal water molecules per molecular formula can be quantified by the full width at half maximum of the Raman bands or relative intensity with linear relationships, suggesting that Raman spectroscopy can be a
www.rsc.org/pccp
potential tool in quantifying crystal water molecules at room temperature in this or related materials.
1 Introduction Compounds with negative thermal expansion (NTE) have attracted much attention since the discovery of the NTE of ZrW2O8 in a large temperature range (0.3 to 1050 K).1–12 However, the wide application of NTE materials is not realized thus far due to some critical disadvantages, such as unsuitable phase transition temperature,1,2,4 strong hygroscopicity,3,13–15 unstable mechanical properties,16 too high/low or narrow temperature range for NTE,8,17,18 with poisonous ions,9,12 etc. Y2M3O12 (M: W, Mo) belonging to the A2M3O12 family with an orthorhombic structure has a strong NTE property but it is highly hygroscopic, like other members with a larger A3+ cation size. Its NTE was only observed after the complete release of the crystal water. It has a building block structure with corner-sharing YO6 octahedra and MoO4 tetrahedra. It was found that crystal water molecules residing in the microchannels of the building blocks hinder not only the external librational and translational motions but also the internal stretching and bending vibrations of the polyhedra and hence its NTE property.3,15,16 Crystal water molecules in the microchannels result in a large but different contraction for different axes of the lattice while the release of them leads to abrupt expansion of the lattice differently for different directions.15,19,20 This should School of Physical Science & Engineering and Key Laboratory of Materials Physics of Ministry of Education of China, Zhengzhou University, Zhengzhou 450052, China. E-mail: [email protected]; Fax: +86 371 67766629; Tel: +86 371 67767838
12848 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
be responsible for the amorphous state and poor mechanical properties of the structure.14,16 From both a scientific and application point of view, it is therefore significant to pursue effective methods to reduce the hygroscopicity in Y2M3O12 and related materials. On the other hand, the macroscopic effects of crystal water on the thermal expansion properties in Y2M3O12 and the like is nearly clear, but the microscopic mechanisms of the interaction between the water molecules and the building blocks remain unclear. For example, hydrogen bonding of water with possible oxygen positions in the building block was assumed.17 However, it does not seem possible that one-end hydrogen bonding of water with ‘‘O’’ in the building block could make so large a contraction and distortion of the lattice. Why the crystal water causes such large contraction on the building blocks in Y2Mo2O12 and related materials remains unresolved. Although many efforts have been made to reduce the hygroscopicity in Y2M3O12 and YbM3O12 (M: W, Mo), the problems mentioned above remain unresolved. When the larger A3+ cation is partially replaced by smaller cations, such as Y3+ partially substituted by Fe3+ (64.5 pm),17 the hygroscopicity can be reduced only to some extent if the orthorhombic phase remains. A higher content of substitution results in a monoclinic phase which has a large positive thermal expansion. In some cases, such as in Yb2xAlxMo3O12 and Yb2xCrxMo3O12,20,21 a pure orthorhombic phase is only possible for an even smaller content substitution (0.0 r x r 0.4) so that the materials are still highly hygroscopic. Otherwise, monoclinic or monoclinic–orthorhombic composited
This journal is © the Owner Societies 2014
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
Paper
phases appear. The substitution of La3+ (25 mol%) for Y3+ results in the formation of a monoclinic phase and large thermal expansion below 373 K, suggesting that it is hygroscopic.22 To eliminate the hygroscopicity in orthorhombic Y2Mo3O12 and the like is still challenging. In this paper, Ce3+ ions are introduced into the Y2Mo3O12 lattice with a sol–gel method with the aim to reduce its hygroscopicity, pursue the interaction of crystal water molecules with the building blocks as well as tailor its thermal expansion property. To reach these aims, temperature dependent X-ray diffraction (XRD), high resolution transmission electron microscopy (HRTEM), temperature dependent Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), differential scanning calorimetry (DSC) and thermogravimetric (TG) as well as dilatometry analyses are used. It is found that Ce3+ ions occupy the positions of Y3+ in the lattice and the incorporation of Ce3+ has the function of expelling crystal water molecules in the microchannels of Y2Mo3O12. As the Ce3+ content increases, the number of crystal water molecules decreases significantly so that the crystal water molecules are completely removed when the content of Ce3+ is higher than 8 mol% and a low thermal expansion material can be achieved without hygroscopicity. The crystal water molecules have an overwhelming effect on the changes of the Raman spectra, causing significant broadening on admission and sharpening on the release of crystal water molecules, as well as relative intensity changes of the Raman bands. It is shown that the number of crystal water molecules per molecular formula can be quantified by the full width at half maximum of the Raman bands or relative intensity with linear relationships. Based on the binding energy changes of Mo 3d and Y 3d with and without Ce3+ in the lattice, the configuration of th crystal water in the building block is presented. With such a configuration, the effects of the crystal water on the thermal expansion behaviors can be well understood.
2 Materials and methods Analytical grade Y(NO3)39H2O, (NH4)6Mo7O244H2O and Ce(NO3)36H2O were used as raw materials. C6H8O7 was used as a complexing agent. 30 mL (0.01 mol) aqueous solutions of Y(NO3)39H2O containing 0, 1, 2, 3, 4, 5, 6, 8, 9 and 10 mol% contents of Ce3+ from Ce(NO3)36H2O were prepared with deionized water and stirred with a magnetic stirrer for 1 h at 323 K. Transparent sols were formed after adding C6H8O7 (0.1 mol) (its molar ratio was ten fold that of Y(NO3)39H2O) to the above solutions. (NH4)6Mo7O244H2O was then put into each of the sols to maintain a molar ratio of (Y + Ce) : Mo = 2 : 3 in each sol. Each sol was divided into four parts in different evaporating dishes and their pH values were tuned to 1, 3, 5, and 7, respectively, with NH3H2O (30%) and HNO3 (30%). They were stirred rigorously at 353 K for 6 h and then heated in a blast oven at 408 K for 12 h to become gels. They were then heated at 623 K for 3 h and at 1023 K for another 3 h in a muffle furnace. The products were cooled naturally to room temperature for analysis. Some powders were measured directly and some were pressed into pellets (diameter of 12 mm and thickness of 2 mm)
This journal is © the Owner Societies 2014
PCCP
or cylinders (diameter of 6 mm and height of 10 mm) according to measurement requirements. XRD measurements were carried out with an X-ray diffractometer (Model X’Pert PRO) to identify the crystalline phase. Raman spectra were recorded on a laser Raman spectrometer (Renishaw MR-2000) with an excitation wavelength of 532 nm. The temperature dependence of the Raman spectra was recorded by using a TMS 94 heating/freezing stage from Linkam Scientific Instruments Ltd., with an accuracy of 0.1 K. High resolution transmission electron microscopy (HRTEM JEM-2010, JEOL) was used to observe the lattice fringes. Differential scanning calorimetry (DSC) and thermogravimetric analysis (TG) were done on an Ulvac Sinku-Riko DSC (Model 1500M/L) in the temperature range of 293–873 K with heating and cooling rates of 10 K min1. XPS (Axis Ultra, Kratos, UK) was used to analyze the composition of the samples and oxidation states of the elements. The linear thermal expansion coefficient was measured by a dilatometer (LINSEIS DIL L76).
3 Results and discussion 3.1
XRD and TEM characterizations for Ce3+-doped Y2Mo3O12
The influence of the pH values and the content of Ce3+ on the final products were studied by XRD. Fig. 1a and b shows the XRD patterns of Ce3+-doped Y2Mo3O12 prepared at pH = 1 and 5, respectively. It is shown that both the pH values in the initial sols and the nominal content of Ce3+ influence the final structure of the samples, as revealed by the abrupt change in the XRD patterns occurring between 2 and 3 mol% of Ce3+ for the samples prepared at pH = 1 and between 4 and 5 mol% of Ce3+ for the samples prepared at pH = 5. For the samples with Ce3+ content r2 mol% (pH = 1) or r4 mol% (pH = 5), the XRD patterns resemble that of highly hydrated Y2Mo3O12 3H2O,14,15,19,23 indicating that these samples retain the structure of Y2Mo3O123H2O and are still highly hydrated. The XRD patterns at or above a 3 mol% (pH = 1) or 5 mol% (pH = 5) content of Ce3+ resemble that of orthorhombic Y2Mo3O12 after the release of crystal water.14 It is obvious that the diffraction peaks shift progressively to lower angles on increasing the content of Ce3+, suggesting an increase in the lattice constant due to the larger radius of Ce3+ than Y3+ (effect of crystal water exists also). In order to clarify the sole effect of the introduction of Ce3+, we performed XRD patterns of the samples with 0, 2, 5 and 9 mol% Ce3+ at 413 K (at 413 K, crystal water is released) and indexing of the XRD pattern of Y2Mo3O12 at 413 K, as shown in Fig. 1c and d. Table 1 lists the lattice constants and unit cell volumes of Ce3+-doped Y2Mo3O12 (pH = 1) from structure refinements of the XRD data using the Rietveld method. It is found that, for crystals with the same space group, the lattice constants increase on increasing the content of Ce3+ (0 - 2 mol%, 5 - 9 mol%). For crystals with different space groups, it is not suitable to compare their lattice constants. We have also made use of the data to draw the building blocks of Y2Mo3O12 crystals with 0, 2, 5 and 9 mol% Ce3+ (Fig. 1e). The structures with the same space groups
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12849
View Article Online
Paper
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PCCP
Fig. 1 (a–d) XRD patterns of Ce3+-doped Y2Mo3O12 prepared from the sols with different pH values: (a) pH = 1; (b) pH = 5; (c) XRD patterns of the samples with 0, 2, 5 and 9 mol% Ce3+ at 413 K (pH = 1); (d) indexing of the XRD pattern of Y2Mo3O12 at 413 K (pH = 1). (e) Diagrams of the Ce3+-doped Y2Mo3O12 building blocks with 0, 2, 5 and 9 mol% Ce3+: the space groups of Y2Mo3O12 with 0, 2 mol% Ce3+ are Pbcn and the space groups of Y2Mo3O12 with 5, 9 mol% Ce3+ are Pba2.
change less but notable changes appear after the space group transforms from Pbcn to Pba2. The bonding length and angles change markedly with different space groups.
12850 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
Fig. 2 shows the HRTEM images of the lattice fringes and diffraction patterns for Y2Mo3O12 (Fig. 2a, b) and 2 mol% Ce3+-doped Y2Mo3O12 (Fig. 2c, d) prepared at pH = 1. It is shown
This journal is © the Owner Societies 2014
View Article Online
Paper
PCCP
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
Table 1 Constants (a, b, c, d and volume) of Y2Mo3O12 samples with 0, 2, 5 and 9 mol% Ce3+ at 413 K
Content of Ce3+ (%)
Space group
a (Å)
b (Å)
c (Å)
v (Å3)
d(111) (nm)
0 2 5 9
Pbcn Pbcn Pba2 Pba2
13.8726 13.9028 10.3542 10.3790
9.9249 9.9410 10.3704 10.3797
10.0033 10.0136 10.6016 10.6026
1377.2809 1383.9495 1138.3662 1142.2247
0.6282 0.6291 0.6028 0.6035
Fig. 2 HRTEM images of Y2Mo3O12 (a, b) and 2 mol% Ce3+-doped Y2Mo3O12 (c, d) prepared at pH = 1.
that the interplanar spacing along the (111) direction of Y2Mo3O12 increases primarily on increasing the content of Ce3+. The interplanar spacing increases from 0.6035 nm for Y2Mo3O12 to 0.6045 nm for 2 mol% Ce3+-doped Y2Mo3O12, which are smaller than the samples with 0 and 2 mol% Ce3+ in Table 1 due to shrinkage of the lattice induced by crystal water at room temperature. Fig. 1 and 2 demonstrate that Ce3+ was incorporated into the lattice of Y2Mo3O12 and the Ce3+-doped Y2Mo3O12 crystallized in an orthorhombic structure. The samples are obviously hygroscopic for Ce3+ contents of o3 mol% (pH = 1) or o5 mol% (pH = 5). As we will see later, though the distinctive change in the XRD patterns occur at the Ce3+ content of around 3 mol% (pH = 1)
Fig. 3
or o5 mol%, the samples still contain crystal water until a 8–9 mol% content of Ce3+, as revealed by our Raman spectroscopic, DSC and dilatometric analyses. 3.2 Influence of the content of Ce3+ and crystal water on the Raman spectra of Y2Mo3O12 Fig. 3a and b show the Raman spectra of Ce3+-doped Y2Mo3O12 samples with different Ce3+ content prepared at pH = 1 and pH = 5, respectively. The Raman modes appearing in the ranges of 970–900 cm1 and 900–740 cm1 can be ascribed to symmetric and asymmetric stretching vibrations of the MoO4 tetrahedra, respectively,15,19 while those from 400 to 300 cm1 are attributed to symmetric and asymmetric bending motions in both the YO6 octahedra and MoO4 tetrahedra, and those modes below 300 cm1 are translational and librational vibrations.15,19,24 The common features of the Raman spectra in Fig. 3a and b are the appearance of only three broad Raman bands at around 940, 824 and 333 cm1 for lower contents of Ce3+ and the Raman bands split into sharp modes and increase in intensity on further increasing the content of Ce3+. However, the distinct changes in the Raman spectra start to occur at different contents of Ce3+ for the samples prepared at different pH values, 3 mol% for pH = 1 (see Fig. 3a) and 5 mol% for pH = 5 (see Fig. 3b), in accordance with the XRD analyses. This suggests that the pH values have an obvious effect on the real content of Ce3+ in the final products and a lower pH value seems more favorable for the Ce3+ ions to incorporate into the lattice of Y2Mo3O12. It is interesting to note that the behaviors of the Raman spectra on increasing the content of Ce3+ are quite similar to that of the Raman spectra of Y2Mo3O12 with increasing temperature, in which only three broad Raman bands appear at room temperature and the Raman bands start to split and increase in intensity above 403 K.15,19 Since Y2Mo3O12 is highly hygroscopic, the enormous change in the Raman spectra was attributed to the release of crystal water. The crystal water molecules in the microchannels hinder not only the librational and translational motions, but also the stretching and bending vibrations of the polyhedra.15 It can be inferred that the distinct changes in the Raman spectra on increasing the content of Ce3+ are related to the reduction of crystal water molecules. Namely, the incorporation of Ce3+ in the lattice of Y2Mo3O12 has the ability to inhibit water
Dependence of the Raman spectra of Ce3+-doped Y2Mo3O12 on the molar percent of Ce3+ prepared at (a) pH = 1 and (b) pH = 5.
This journal is © the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12851
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PCCP
Paper
Fig. 4 Temperature dependence of the Raman spectra of Y2Mo3O12 doped with 2 mol% Ce3+.
species entering the microchannels in the building blocks. The continuous increase and sharpening of the Raman bands on increasing the content of Ce3+ (>3 mol% for pH = 1; and >5 mol% for pH = 5) may indicate a successive reduction of the crystal water molecules. In order to confirm the above assumption and gain insight into the effect of Ce3+ on the crystal water, we carried out a series of temperature-dependent Raman spectroscopic studies (the samples prepared at pH = 1 are used in the following measurements). Fig. 4 shows the Raman spectra of Y2Mo3O12 with 2 mol% Ce3+ at different temperatures. On increasing the temperature from 293 K to 383 K, there is no significant change in the Raman spectra except that the Raman bands at about 943, 829 and 342 cm1 shift progressively to 952, 823 and 337 cm1, respectively. The Raman spectra start to change enormously at about 393 K. An intense Raman band appears at about 966 cm1 with a shoulder at about 936 cm1. The asymmetric stretching Raman band at about 823 cm1 becomes sharper and much more intense. The bending mode shifts from 337 to 326 cm1. Above 393 K, both the symmetric and asymmetric Raman modes increase in intensity with slight blue shifts with increasing temperature until 443 K. Generally, Raman bands weaken and broaden with increasing temperature. The abrupt changes in the Raman spectra at about 393 K and the abnormal behavior above this temperature are attributed to the release of crystal water molecules, which form hydrogen bonds with the O2 ions in the Y–O–Mo bridges.19 The hydrogen bonding attracts a partial amount of the electrons from the bridging oxygen toward the hydrogen bonds and hence weakens the O–Mo bonds. This is confirmed by the red shift of the stretching Raman bands with decreasing temperature (more crystal water molecules) or the blue shift of the Raman bands with increasing temperature (less crystal water molecules). A Raman band frequency
Raman mode shift should be included. We therefore give the Raman band frequency as o(T) = o0 + D(T) + D(m), where D(m) depends on the number of crystal water molecules, m, and accounts for the crystal-water induced shift. With temperature increasing, D(T) leads to a red shift of a Raman band while D(m) results in a blue shift due to the reduction of the number of crystal water molecules. From RT to 393 K and from 393 to 443 K, the effects caused by the release of crystal water molecules and by temperature increasing coexist. The blue shifts of the stretching modes are hence the result of the competition of the two effects. From RT to 393 K, only the weakly interacting crystal water molecules are able to be removed, while the crystal water molecules interacting strongly with the bridging oxygens start to be released at about 393 K and above, as indicated by the abrupt change in the Raman spectrum around 393 K. From 443 K to 573 K, the Raman band at 966 cm1 shifts to 962 cm1 and that at 823 cm1 shifts to 820 cm1. The red shifts can be considered as a pure temperature-induced effect. Though the Raman bands already start to split at 3 mol% Ce3+, we demonstrate here that the samples at this or even higher contents of Ce3+ are still hygroscopic. Fig. 5 shows the Raman spectra of Y2Mo3O12 doped with 5 mol% of Ce3+. It is clear that the obvious changes in the Raman spectra occur at about 393–403 K, as indicated by the intensity inverse of the Raman bands at about 823 and 962 cm1 with respect to the Raman bands at 853 and 944 cm1, respectively, as well as the increase of the Raman bands on increasing the temperature to above 403 K. As discussed above for the sample with 2 mol% of Ce3+, the obvious changes in the Raman spectra between 393 and 403 K should be correlated to the release of crystal water. These suggest that this sample still contains a certain amount of the crystal water molecules, which start to be released at about 403 K. Detailed analyses for other samples show that the Raman spectra for the samples with Ce3+ content r8 mol% behave similarly with increasing temperature (not shown here). However, no obvious changes in the Raman spectra are observed for the samples with >8 mol% Ce3+. Fig. 6 shows the Raman spectra of Y3Mo3O12 doped with 9 mol% Ce3+ at different temperatures. The Raman spectra do not exhibit obvious changes between 383 and 403 K, in contrast
can be expressed as o(T) = o0 + D(T), where DðTÞ ¼ C 1þ
2
" þD 1þ
3
þ
3
#
is ðeho=3kB T 1Þ2 the temperature-induced shift of a Raman band, and C and D are anharmonic constants with negative values for normal Raman modes.25 In the present case, a crystal water-induced eho=2kB T 1
eho=3kB T 1
12852 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
Fig. 5 Temperature dependence of the Raman spectra of Y2Mo3O12 doped with 5 mol% Ce3+.
This journal is © the Owner Societies 2014
View Article Online
Paper
PCCP
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
Increasing the content of Ce3+ causes an increase in DGdop and a decrease in the number of crystal water molecules and hence a reduction in DGcw. Fig. 7 suggests that the decrease in crystal water molecules has an overwhelming effect on the bandwidths and relative intensities when the content of Ce3+ is less than 8 mol%. The changes in the FWHMs and relative intensities above 8 mol% Ce3+ may be regarded as a pure effect of Ce3+-doping. 3.3 Correlation of the number of crystal water molecules with features of the Raman spectra Fig. 6 Temperature dependence of the Raman spectra of Y2Mo3O12 doped with 9 mol% Ce3+.
to the temperature-dependent Raman spectra of Y3Mo3O12 doped with 2 and 5 mol% Ce3+. It is reasonable that the crystal water molecules are completely eliminated for the sample with a 9 mol% Ce3+ content. This is an exciting result because complete elimination of crystal water in Y3Mo3O12 and the like at ambient conditions is important for applications and has not really been achieved, though many efforts have been made.19–21,26 A comparison of Fig. 3–5 shows that the Raman bands at about 853 cm1 are mainly influenced by the incorporation of Ce3+ instead of crystal water. In order to correlate the effect of the crystal water molecules on the changes in the Raman bands in Fig. 3, we show in Fig. 7 the changes in the full width at half maximum (FWHM) for the four intense stretching Raman bands at about 823, 853, 944 and 962 cm1 and their relative intensities with respect to the Raman band at about 853 cm1 with the Ce3+ content. (All the Raman spectra were fitted with Lorentzian peak profiles and the data were derived from the curve fittings.) It is shown that the bandwidths decrease and the relative intensity changes obviously on increasing the content of Ce3+ until 8–9 mol%. We attribute the decrease in the FWHM to the decrease in the content of crystal water molecules because a Raman band linewidth generally broadens after doping due to lattice distortion. A Raman band linewidth, G, in our case can be expressed by the sum of the natural linewidth, G0(T), of Y2Mo3O12 at room temperature, Ce3+ doping-induced broadening, DGdop, and crystal water-induced broadening, DGcw, G = G0(T) + DGdop + DGcw.
DSC and TG measurements were performed to confirm the above analyses by Raman spectroscopy (see Fig. 8). The endothermic peaks at around 423 K in the DSC curves (Fig. 8a) and the corresponding weight loss (Fig. 8b) for the samples with a Ce3+ content of r8 mol% confirm the existence of crystal water molecules. However, neither endothermic peaks nor weight losses are observed for the sample with 9 mol% Ce3+. This indicates that the samples with Z9 mol% Ce3+ do not contain crystal water molecules, confirming our Raman analyses above. The number of crystal water molecules per molecular formula is calculated to be 2.48, 2.46, 2.12, 1.46, 0.74, 0.43, 0.12 and 0.0 for the samples with 0, 1, 2, 3, 5, 6, 8 and 9 mol% contents of Ce3+. It reduces markedly on increasing the content of Ce3+ and the crystal water is eliminated completely when the Ce3+ content is increased to 9 mol%. With these results, it is possible for us to correlate the changes in the FWHMs and relative intensities to the number of crystal water molecules, as shown in Fig. 9. By examining Fig. 7, we know that the FWHMs of the Raman bands at 832 and 944 cm1 are most sensitive to the number of crystal water molecules. Therefore, we sum the FWHMs of these two bands (denoted as G1) as well as those of all the four bands (G2) to correlate with the number of crystal water molecules in order to improve the sensitivity. Both can be well fitted with linear equations as G1 = 22.48048 + 17.22256m with R2 = 0.98792 (Fig. 9a); and G2 = 41.97288 + 18.80879m with R2 = 0.98569 (Fig. 9b), where m is the number of crystal water molecules per molecular formula. We can see that using the sum of the most sensitive bandwidths fits a little bit better than using the sum of all four bandwidths. We attribute this to the influence of the Ce3+-doping-induced bandwidth change.
Fig. 7 Dependence of the FWHMs (a) and relative intensities (b) of the selected Raman bands on the content of Ce3+.
This journal is © the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12853
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PCCP
Paper
Fig. 8 (a) DSC and (b) TG curves of Ce3+-doped Y2Mo3O12 with 0, 1, 2, 3, 5, 6, 8 and 9 mol% Ce3+.
As discussed above, the Raman band at about 853 cm1 is most sensitive to the Ce3+ content but the least influenced by the release of crystal water molecules. Keeping this in mind, the relative intensity of the 832 cm1 band with respect to the band at 853 cm1 can also be used to characterize the number of crystal water molecules, as shown in Fig. 9c. It can be expressed Ið832 cm1 Þ as ¼ 0:58669 þ 0:21605m, with R2 = 0.96183. Fig. 3– Ið853 cm1 Þ 7 and 9 indicate that Raman spectroscopy can be used as a powerful tool to quantify the crystal water molecules at room temperature and investigate their influence on the vibrational properties of materials. From the Raman measurements (Fig. 3–6) and discussion above, it is possible for us to identify the stretching Raman modes more specifically. The Raman bands at 944 and 853 cm1 are most influenced by Ce3+-doping. They can be identified as the symmetric and antisymmetric stretching vibrations of the Mo–O bonds correlated mainly with Ce3+, respectively. However, the contribution from the Mo–O bonds correlated mainly with Y3+ cannot be omitted. Similarly, the Raman bands at 960 and 823 cm1 are less influenced by Ce3+-doping. They can be identified
Correlations of the FWHMs G1 (a) and G2 (b) of the Raman bands I 832 cm1 and the relative intensity of (c) with the number of crystal I ð853 cm1 Þ water molecules.
as the symmetric and antisymmetric stretching vibrations of the Mo–O bonds correlated mainly with Y3+, respectively. 3.4 Effects of Ce3+-doping and the reduction of crystal water on the thermal expansion properties of Y2Mo3O12 Fig. 10 shows the relative length changes of Ce3+-doped Y2Mo3O12 containing 0, 1, 3, 5, 8, 9 and 10 mol% contents of Ce3+ with increasing temperature. The materials shrink distinctly from about 350 to 408 K and then expand abruptly until about 460 K for Ce3+ contents of less than 3 mol%, whereas those with 3–8 mol% Ce3+ exhibit less abrupt expansion from about 350 to 430 K. In this temperature range, the thermal expansion behavior is dominated by the release of crystal water. It is clear that the interaction of crystal water molecules with the building block of Y2Mo3O12 is so strong that they make the building blocks contract considerably and the release of them leads to an abrupt expansion of the building blocks. After complete release of the crystal water molecules, the materials show NTE (8.94 106 K1, 9.22 106 K1 and 2.25 106 K1 for 0, 1 and 3 mol% Ce3+, respectively) or low positive expansion (1.67 106 K1 and 2.27 106 K1 for 5 and 8 mol% Ce3+). The samples with Ce3+ contents higher than 8 mol% exhibit low thermal expansions without abnormal changes in the whole temperature range (2.743 106 K1 and 3.53 106 K1 for 9 and 10 mol% Ce3+), suggesting that they do not possess crystal water molecules at all. The thermal
Fig. 9
12854 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
Fig. 10 Relative length change (dL/L0 (%)) with temperature of Ce3+-doped Y2Mo3O12 (pH = 1, Ce3+ content of 0, 1, 3, 5, 8, 9, and 10 mol%).
This journal is © the Owner Societies 2014
View Article Online
Paper
expansion properties agree well with the Raman and DSC analyses above. In a previous paper, we demonstrated that the c-axis and b-axis were more influenced by crystal water than the a-axis in Fe2xYxMo3O12.19
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
3.5 Mechanism of crystal water reduction in Ce3+-doped Y2Mo3O12 To explore the effect of Ce3+ doping on the reduction of crystal water, we measured the binding energies of Y2Mo3O12 and 5 mol% Ce3+-doped Y2Mo3O12 by XPS. Fig. 11a–f shows the survey spectra of Y2Mo3O12 and Ce3+-doped Y2Mo3O12, and the core level spectra of Y 3d, Mo 3d and O 1s, respectively. After the introduction of Ce3+, though the spin–orbit splitting of Y 3d (1.9 eV) and Mo 3d (3.2 eV) do not change, the binding energy of Y 3d5/2 decreases from 158.1 to 156.0 eV,27 which is much more distinct than the binding energy shift of Mo 3d5/2 (232.7 - 232.6 eV) (Fig. 11b and c). Considering this fact together with the valence and cation radii difference, it is reasonable that Ce3+ ions take the positions of Y3+ rather than Mo6+ in Y2Mo3O12. The O 1s peak is obviously asymmetric in Y2Mo3O12 but it becomes more symmetric in Ce3+-doped Y2Mo3O12 (Fig. 11d). Deconvolution of the O 1s region shows three peaks (Fig. 11e). The first, located at 530.6 eV, is assigned to Mo–O–Y in Y2Mo3O12, whereas the second at 531.8 eV and third at 533.4 eV are associated with hydroxyl groups.28 This explains why Y2Mo3O12 is highly hygroscopic. Nevertheless, after introducing Ce3+, the O 1s peak can only be deconvoluted into two peaks, suggesting suppression in the hydroxyl groups. This confirms that the introduction of Ce3+ has the ability to repel crystal water molecules.
PCCP
In a recent paper, we assumed that the protruded ‘‘O’’ anions in the microchannels of the building blocks had a higher probability to interact with water species to form hydrogen bonds.19 The distinct changes in the stretching Raman bands of MoO4 tetrahedra before and after the release of crystal water molecules also support the hydrogen bonding to the bridging oxygen ions. However, if the water molecules interacted with the bridging oxygen ions only, it would be difficult to understand so large a lattice contraction caused by crystal water molecules or expansion on the release of them. Comparisons of the hygroscopicity between Y2Mo3O12 and Ce3+-doped Y2Mo3O12 and the binding energy changes in Y 3d5/2,3/2 and O 1s after the substitution of Ce3+ for Y3+ throw light on the configurations of the crystal water molecules in the microchannels, i.e. the crystal water molecule has its oxygen ion approaching the Y3+ in the octahedra. From these discussions, it is possible for us to speculate on the configurations of the crystal water molecules in the building blocks. That is, a water molecule is oriented with its O2 pointing to the Y3+ in an octahedron and with its H+ approaching the next nearest bridging O2 in the Y–O–Mo bridge to form hydrogen bonding, as depicted in Fig. 12. With this configuration, the effects of crystal water on the thermal expansion properties, as presented in Fig. 10, can be well explained. The crystal water serves as a spring to drag connected octahedra and tetrahedra to rotate, making the c-axis contract more obviously than the b-axis. In such a configuration, the a-axis is least influenced by the crystal water. The release of the crystal water molecules leads to a restoration of the lattice and hence an abrupt expansion around the releasing temperature of the crystal water. As the content of
Fig. 11 The survey spectra of Y2Mo3O12 and 5 mol% Ce3+-doped Y2Mo3O12 (a); and the core level spectra of Y 3d (b), Mo 3d (c) and O 1s (d–f).
This journal is © the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12855
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
PCCP
Fig. 12 Schematic of the Y2Mo3O12 building block and configuration of the crystal water molecule.
Ce3+ increases, there are fewer or no crystal water molecules, causing less or no obvious contraction of the lattice. Therefore, the linear thermal expansion coefficients change less abruptly or even change smoothly with increasing temperature. Since Ce has a lower electronegativity (1.12) than Y (1.22), Ce3+ has a lower capacity than Y3+ to attract ‘‘O’’ in the water molecules. Fig. 8 suggests a successive decrease in the average binding energy of hydrogen bonds of crystal water molecules on increasing the content of Ce3+. A first principles calculation shows a decrease in the binding energy of a water molecule with the building block from 0.6 eV to 0.2 eV by Ce-doping (unpublished data from Prof. Qiang Sun, Zhengzhou University). This explains, at least partly, why the Ce3+-doping eliminates the crystal water.
4 Conclusions Ce3+ ions are introduced into the lattice of Y2Mo3O12 with a sol–gel method with the aim to reduce its hygroscopicity, pursue the interaction of the crystal water molecules with the building blocks as well as tailor its thermal expansion property. It is found that Ce3+ ions occupy the positions of Y3+ in the lattice and the incorporation of Ce3+ expels crystal water molecules in the microchannels of Y2Mo3O12. As the Ce3+ content increases, the number of crystal water molecules decreases significantly so that the crystal water molecules can be completely removed when the content of Ce3+ is larger than 8 mol% and a low thermal expansion material can be achieved without hygroscopicity. Based on the binding energy changes of Mo 3d and Y 3d with and without Ce3+ in the lattice, the configuration of the crystal water in the building block is deduced. The crystal water molecule is oriented with its O2 ion pointing to the Y3+ in an octahedron and with its H+ ion approaching the next nearest bridging O2 in the Y–O–Mo bridge to form hydrogen bonds. The crystal water molecules serve as springs to drag the connected octahedra and tetrahedra to rotate, making the c-axis contract more obviously than the b-axis. In such a configuration, the a-axis is the least influenced by the crystal water. Compared to the effect of Ce3+, the crystal water molecules have an overwhelming effect on the FWHM and relative intensity changes of the Raman bands. It is also shown that the number of crystal water molecules per molecular formula can be quantified by the
12856 | Phys. Chem. Chem. Phys., 2014, 16, 12848--12857
Paper
full width at half maximum of the Raman bands or relative intensity with linear relationships, G1 = 22.48048 + 17.22256m Ið832 cm1 Þ ¼ 0:58669 þ 0:21605m, where m is the number and Ið853 cm1 Þ of crystal water molecules per molecular formula. This suggests that Raman spectroscopy can be used as a potential tool in quantifying crystal water molecules in this and related compounds at room temperature. This work provides a profound insight into the interaction mechanism of crystal water molecules in the A2M3O12 (A = Lu, Er, Yb, Y; M = W, Mo) family of compounds and the abnormal thermal expansion properties caused by crystal water also paves the way towards solving the long standing hygroscopic problem in Y2Mo3O12 and related compounds.
Acknowledgements This work was supported by the National Science Foundation of China (No. 10974183; No.11104252), by the Doctoral Fund of the Ministry of Education of China (No. 20114101110003) and the fund for Science &. Technology innovation team of Zhengzhou (No. 112PCXTD337). We also thank Prof. Qiang Sun for providing the binding energy data for the crystal water in Y2Mo3O12.
References 1 T. A. Mary, J. S. O. Evans, T. Vogt and A. W. Sleight, Science, 1996, 272, 90–92. 2 P. P. Sahoo, S. Sumithra, G. Madras and T. N. G. Row, Inorg. Chem., 2011, 50, 8774–8781. 3 B. A. Marinkovic, M. Ari, R. R. Avillez, F. Rizzo, F. F. Ferreira, K. J. Miller, M. B. Johnson and M. A. White, Chem. Mater., 2009, 21, 2886–2894. 4 K. J. Miller, C. P. Romao, M. Bieringer, B. A. Marinkovic, L. Prisco and M. A. White, J. Am. Ceram. Soc., 2013, 96, 561–566. 5 J. R. Wang, J. X. Deng, R. B. Yu, J. Chen and X. R. Xing, Dalton Trans., 2011, 40, 3394–3397. 6 R. Futamura, T. Iiyama, A. Hamasaki and S. Ozeki, Phys. Chem. Chem. Phys., 2012, 14, 981–986. 7 L. C. Gallington, K. W. Chapman, C. R. Morelock, P. J. Chupas and A. P. Wilkinson, Phys. Chem. Chem. Phys., 2013, 15, 19665–19672. 8 X. Y. Song, Z. H. Sun, Q. Z. Huang, M. Rettenmayr, X. M. Liu, M. Seyring, G. N. Li, G. H. Rao and F. X. Yin, Adv. Mater., 2011, 23, 4690–4694. 9 J. Chen, L. L. Fan, Y. Ren, Z. Pan, J. X. Deng, R. B. Yu and X. R. Xing, Phys. Rev. Lett., 2013, 110, 115901. ˜oz, J. B. Keith, S. J. Tracy, 10 C. W. Li, X. L. Tang, J. A. Mun D. L. Abernathy and B. Fultz, Phys. Rev. Lett., 2011, 107, 195504. 11 M. Azuma, W. Chen, H. Seki, M. Czapski, S. Olga, K. Oka, M. Mizumaki, T. Watanuki, N. Ishimatsu, N. Kawamura, S. Ishiwata, M. G. Tucker, Y. Shimakawa and J. P. Attfield, Nat. Commun., 2011, 2, 347. 12 A. L. Goodwin, M. Calleja, M. J. Conterio, M. T. Dove, J. S. O. Evans, D. A. Keen, L. Peters and M. G. Tucker, Science, 2008, 319, 794–797.
This journal is © the Owner Societies 2014
View Article Online
Published on 08 April 2014. Downloaded by Vanderbilt University on 04/10/2014 15:04:55.
Paper
13 S. Sumithra and A. M. Umarji, Mater. Res. Bull., 2005, 40, 167–176. 14 S. D. Gates and C. Lind, J. Solid State Chem., 2007, 180, 3510–3514. 15 E. J. Liang, H. L. Huo, J. P. Wang and M. J. Chao, J. Phys. Chem. C, 2008, 112, 6577–6581. 16 C. Yuan, Y. Liang, J. P. Wang and E. J. Liang, J. Chin. Ceram. Soc., 2009, 37, 726–732. 17 A. K. Tyagi, S. N. Achary and M. D. Mathews, J. Alloys Compd., 2002, 339, 207–210. 18 T. Chatterji, M. Zbiri and T. C. Hansen, Appl. Phys. Lett., 2011, 98, 181911. 19 Z. Y. Li, W. B. Song and E. J. Liang, J. Phys. Chem. C, 2011, 115, 17806–17811. 20 Q. J. Li, B. H. Yuan, W. B. Song, E. J. Liang and B. Yuan, Chin. Phys. B, 2012, 21, 046501.
This journal is © the Owner Societies 2014
PCCP
21 M. M. Wu, X. L. Xiao, Z. B. Hu, Y. T. Liu and D. F. Chen, Solid State Sci., 2009, 11, 325–329. 22 H. F. Liu, X. C. Wang, Z. P. Zhang and X. B. Chen, Ceram. Int., 2012, 38, 6349–6352. 23 B. A. Marinkovic, P. M. Jardim, R. R. Avillez and F. Rizzo, Solid State Sci., 2005, 7, 1377–1383. 24 L. Wang, F. Wang, P. F. Yuan, Q. Sun, E. J. Liang, Y. Jia and Z. X. Guo, Mater. Res. Bull., 2013, 48, 2724–2729. 25 M. Balkanski, R. F. Wallis and E. Haro, Phys. Rev. B, 1983, 28, 1928–1934. 26 M. M. Wu, Z. B. Hu, Y. T. Liu and D. F. Chen, Mater. Res. Bull., 2009, 44, 1943–1947. ´cs, K. Magyari, M. Neumann and 27 V. Simon, D. Eniu, A. Taka S. Simon, J. Optoelectron. Adv. Mater., 2005, 7, 2853–2857. 28 J. C. Dupin, D. Gonbeau, P. Vinatier and A. Levasseur, Phys. Chem. Chem. Phys., 2000, 2, 1319–1324.
Phys. Chem. Chem. Phys., 2014, 16, 12848--12857 | 12857
