Hydrogen-incorporation stabilization of metallic VO2(R)
Supporting Information for
Hydrogen-incorporation stabilization of metallic VO2(R) phase to room temperature displaying promising low-temperature thermoelectric effect
Changzheng Wu†, Feng Feng†, Jun Feng†, Jun Dai†, Lele Peng†, Jiyin Zhao†, Jinlong Yang†, Cheng Si ‡, Ziyu Wu‡ and Yi Xie *, † †Hefei National Laboratory for Physical Sciences at Microscale and ‡National Synchrotron Radiation Laboratory, University of Science & Technology of China, Hefei, Anhui 230026, P.R. China
S1
S1. Experimental section 1．Synthesis of hydric paramontroseite VO2 precursor The hydric paramontroseite VO2 was prepared from post-treatment of the synthetic goethite VOOH sample similar to the reported literature by our group 1. In a typical synthesis procedure, the hydric paramontroseite VO2 sample was prepared by the post heating-treatment of goethite VOOH at a lower temperature of 80 oC. And then the hydric paramontroseite VO2 sample was collected for the following characterizations and synthesis procedures. The XRD pattern of the as-obtained hydric paramontroseite VO2 (Figure S1) matches well with the standard JCPDS card NO. 73-0514, which can be readily indexed into paramontroseite VO2 with space group Pbnm. The thermo-gravimetric (TG) curve in air atmosphere for the as-obtained hydric paramontroseite VO2 reveals the hydric feature and provides a clue to know the hydrogen contents of the hydric sample. As shown in Figure S2, a small platform exists at around 100 oC could be attributed to loss of absorbed water molecules, indicating 0.5% water content in the sample. And then the continuous weight lose ranging from 100 o
C to 280 oC could be assigned to the lose of hydrogen ingredients in the hydric paramontroseite VO2.
And thus it could be found that the hydrogen contents of hydric paramontroseite VO2 precursor is about 0.9%.
S2
301
112
131 230
310
230 220
240
130
021
120 110
Relative Intensity (a.u.)
Experimental Pattern
JCPDS Card No. 73-0514
10
20
30
40
50
60
70
2θ (degrees)
Figure S1. Experimental XRD pattern of hydric paramontroseite VO2 precursor and standard pattern in JCPDS card NO. 73-0514. 1.002 1.000
Weight (%)
0.998 0.996 99.5%
0.994 0.992 0.990 0.988 0.986
98.6% 50
100
150
200
250
300
350
Temperature ( OC)
Figure S2.
TG curve of hydric paramontroseite VO2 precursor in air atmosphere.
S3
2. Heating transformation of hydric paramontroseite VO2 precursor into hydric VO2 and monoclinic VO2(M) The as-prepared black hydric paramontroseite VO2 precursor was pressed into bulk sample with a cuboid shape (8 mm × 6mm × 3mm) under a pressure of 10 Mpa for 20 min in order to decrease the porosity of the bulk pallets. And then the pallet samples of hydric paramontroseite VO2 were further heated in a nitrogen gas flow at 250 oC, 300 oC, and 600 oC for 3h to produce hydric rutile VO2(R), hydric VO2(M-R), and monoclinic VO2(M), respectively. In our case, the heating post-treatment of the pellets could effectively remove the grain boundary forming the well-defined pellet sample with high crystalline that was vital for the following transport experiments. The color of hydric rutile VO2(R) and VO2(M-R) is black, while the color of monoclinic VO2(M) product slightly changed to blue-black. In our case, based on the calculation results of the bulk elemental analysis, the hydrogen content in the sample obtained at heating 250 oC (hydric rutile VO2(R)), 300 o
C (hydric monoclinic-and-rutile VO2(M-R)), and 600 oC (monoclinic VO2(M)) are 3.8 ‰, 1.3 ‰,
about 0 ‰, respectively. 3. Characterizations The as-prepared samples were characterized by X-ray powder diffraction (XRD) with a Philips X’Pert Pro Super diffractometer with Cu Kα radiation (λ=1.54178 Å). Infared (IR) spectra analyses were operated on samples palletized with KBr powders in the range 4000-400 cm-1, using an infared Fourier transform spectrophotometer (Nicolet, ZOSX). Elemental analysis of hydrogen element
concentrations for VO2 samples were measured on a VARIO ELIII (German) element
S4
analytical instrument. The magnetization was characterized by a superconducting quantum interference device (SQUID, quantum design MPMS XL-7) magnetometer from 200 and 360 K. DSC cycling curves were measured by the NETZSCH DSC 200 F3 with a heating/cooling rate of 5 o
C min-1 between -50 oC and 90 oC. Raman spectra were recorded at room temperature with a
LABRAM-HR Confocal Laser MicroRaman Spectrometer 750 K with a laser power of 0.5 mW. The HRTEM image and ED pattern were obtained on a JEOL-2010 transmission electron microscope at an acceleration voltage of 220 kV. The 1H solid NMR spectra were obtained with a 500.132 MHz NMR spectrometer (Bruker, Germany) at 298.2 K. The experimental data with involving eight transients were collected over 2048 complex points. A mixing time of 300 ms, a relaxation delay of 2s, an acquisition time of 205 ms, and a 90° pulse width of 8.2 μs were used. The data were calculated with a Lorentzianto-Gaussian window function and zero filling in both dimensions to display data on a 2048 × 2048 2D-matrix. All the export data for solid NMR results were performed by NMRPipe software on a Linux workstation. The thermal and electrical transport properties of the bulk sample were measured using a Physical Property Measurement System (PPMS) from 210 to 360 K. 4．Calculation details For VO2(R), we adopted the experimental lattice parameters and atom position for calculations. For hydric VO2(R), since its crystal data from the structural characterizations were almost same with
S5
that of pure rutile VO2(R). Therefore, H atoms were incorporated in the void space of VO2(R) keeping rutile V-O lattice framework (Please see the hydric rutile VO2(R).cif in supporting information for detail structure). The calculation of density of states (DOS) was carried out using projected augmented wave (PAW) method as implemented in Vienna Ab-initio Simulation Package (VASP) 2 . The HSE06 form of hybrid density function
3
was used to describe the electron
exchange-correlation interaction with a 8×8×14 Monkhorst-Pack grid of Brillouin-zone k-point sampling.
S2. XRD patterns of the as-obtained hydric vanadium dioxides The XRD pattern for the product obtained at higher heating temperature of 600 oC for 3h was shown in Figure S3c, of which all of the reflection peaks can be readily indexed to the typical monoclinic VO2(M) (JCPDS card 82-0661, a=5.751 Å, b=4.528 Å, c = 5.380 Å). And the room-temperature XRD pattern for the sample obtained at a lower 250 oC was shown in Figure S3a, which could be readily indexed into the tetragonal rutile VO2(R) (JCPDS card 79-1655, a=4.552 Å, c =2.856 Å), indicating a room-temperature rutile phase is indeed achieved in our case. Moreover, the absence of the typical XRD peaks (100)M (26.94o) and (-102)M(33.40o), the position shifting of VO2(R) (110) peak into lower 2θ degrees compared with VO2(M) (011) peaks in 26o ≤ 2θ ≤29o (Figure S3d), and the splitting of VO2(R) (310) and (002) in 64
o
≤ 2θ ≤66
o
(Figure S3e) that
evolved from VO2(M) (130) peak, clearly verified that the heating treatment of the as-prepared hydric paramontroseite VO2 in a N2 gas flow produces the vanadium oxides with gradient hydrogen
S6
contents of 3.8 ‰, 1.3 ‰, nearly zero, which could be indexed to hydric rutile VO2(R), hydric VO2(R-M), and monoclinic VO2(M) at room temperature, respectively.
Relative Intensity (a.u.)
VO2(M) JCPDS 82-0661
Monoclinic VO2(M)
(c) Hydric VO2(M-R)
(b)
Hydric VO2(R)
(a)
VO2(R) JCPDS 79-1655
VO2(M)
60
(e)
(110) R
VO2(R) with 3.8 ‰ H
28
29
2θ (degrees)
64
(130) M
2θ (degree)
VO2(M-R) with 1.3 ‰ H
27
50
70
VO2(M)
Hydric VO2(M-R) with 1.3 ‰ H
(002) R
40
(310) R
(100) M
(d) Relative Intensity (a.u.)
30
Intensity (a.u.)
20 (011) M
10
Hydric VO2(R) with 3.8 ‰ H
65 66 2θ (degrees)
Figure S3. (a-c) Room-temperature XRD patterns of the as-obtained vanadium dioxide samples obtained by heating at 250 oC for 3h (a), 300 oC for 3h (b), and 600 oC for 3 h (c), which can be readily indexed to rutile VO2(R), intermediate VO2(M-R) and monoclinic VO2(M), respectively. (d)
S7
The partial XRD pattern with 2θ value ranging from 26.5 to 29o. (e) Partial XRD patterns indexing with 2θ value ranging from 64 o to 66o.
S3. HRTEM image and ED pattern of hydric VO2(R) The HRTEM image shows the typical rutile structure character in the 3.8 ‰-hydrogen including VO2 sample. Figure S4 shows the representative SAED pattern taken a typical edge of well-dispersed hydric VO2 particle. Our experimental ED pattern gives the symmetrical tetragonal dotted lattice. The lattice fringes in Figure S4 show two sets of perpendicularly lattice spacing of about 3.216 Å that correspond to the (110) and (1-10) planes of rutile VO2, which shows the typical rutile structure’s HRTEM image and SAED pattern along the [001] axis, confirming the existence of rutile structure in the 3.8 ‰-hydrogen including sample. Furthermore, the atomic model of rutile VO2 projected along the [001] axis also exhibits tetragonal symmetry, and the [110] and [1-10] axes are indicated by the arrows in Figure S4b, giving further evidence for the rutile structure of sample with 3.8 ‰-hydrogen content. In a word, the as-obtained HRTEM result provided more diagnostic and direct evidence for the rutile phase of the VO2 sample with 3.8 ‰-hydrogen content.
S8
(a)
(b)
Figure S4. (a) HRTEM image and SAED pattern taken from a typical edge of well-dispersed hydric VO2(R) particle projected along [001] direction. (b) The corresponding atomic model illustration for the HRTEM images in (a) shows the vertical crystal planes of (110) and (1-10), giving the representive rutile structural feature
S9
S4. The atomic crystal structure of monoclinic VO2(M) and rutile VO2(R)
Figure S5. Structural illustration of monoclinic VO2(M) and rutile VO2(R). (a) Monoclinic VO2(M) supercell structure. (b) The local environment illustration of the vanadium atom in crystal structure of monoclinic VO2(M). (c) Rutile VO2(R) supercell structure. (d) The V local environment illustration in monoclinic VO2(M) structure. For sake of simplicity, the V atom is fixed relatively to the adjacent O and V atoms.
S10
Figure S6. The atomic model of the V-V chains in the rutile VO2(R) supercell structure, showing the infinite vanadium-vanadium chains along z-axis, i.e. the c-direction of rutile structure.
S5. FTIR spectra and Raman spectra of hydric VO2(R) and monoclinic VO2(M) 1. FTIR spectra
749
(2) Hydric VO2(M-R)
1625
Relative Intensity (a.u.)
530 500
(1) VO2(M)
(3) Hydric VO2(R)
1000 1500 2000 2500 -1 Wavenumber (cm )
Figure S7. Room temperature FTIR spectra for the as-obtained hydric vanadium dioxide samples.
S11
FTIR spectrum could reveal the structural and electrical information according to unique spectral characteristics of semiconducting VO2(M) and metallic VO2(R). As is known, the rutile VO2(R) with higher structural symmetry has the less vibration modes than the monoclinic VO2(M), which would significantly depress the fingerprint region in FTIR spectrum for rutile phase.4,5 In the typical IR spectrum of monoclinic VO2(M) sample as shown in Figure S7, the IR absorption peaks at 749 cm-1 and 530 cm-1 could be attributed to the stretching vibrations (V-O-V)4, giving the monoclinic typical feature. While for the IR spectrum of the sample with 3.8‰, the depressed fingerprint peaks provide a clue for indexing the sample to metallic VO2(R). Also, in FTIR spectrum, the absorption band at 1625 cm-1 in the FTIR spectrum signify the evidence of the isolated hydroxyl groups with the deformation vibrations (O-H).6 Since strict same experimental conditions, the peak intensities with the sequence of (a)>(b)>(c) here reflects the hydrogen concentration in a sequence of hydric VO2(R), hydric VO2(M-R), and monoclinic VO2, the results of which were consistent with the elemental analysis and the 1H solid-state MAS NMR results.
S12
2. Raman spectra (a) Monoclinic VO2(M)
360 340
Intensity (a.u.)
50
191
Relative Intensity (a.u.)
380
613
320 300 280
222 401
260
304
240 220 200
40
400
600
800
1000
hydric Rutile VO2(R)
30
20
10
180 200
(b)
200
1200
400
600
-1
800
1000
Raman Shift ( cm )
-1
Raman shift (cm )
Figure S8. Raman spectra of the as-obtained monoclinic VO2(M) without H incorporation and hydric rutile VO2(R) with 3.8 ‰ hydrogen content. Raman spectra provide a fingerprint for identifying the micro-structural characteristics of bonding and coordinations for inorganic solids. As a supplementary evidence, Raman scattering was used to distinguish between monoclinic VO2(M) and rutile VO2(R). In our case, Raman spectra of monoclinic VO2(M) and hydric Rutile VO2(R) were recorded at room temperature using a Ar+ ion laser (514 nm) excitation source and the laser power was set at a low value of 0.5 mW to minimize the local heating of two samples, ensuring them not to be oxidized during the Raman test. As is shown, the sample without H content shows Raman peaks at 191 (Ag), 222 (Ag), 260 (Bg), 307 (Bg), 340 (Ag), 386 (Ag), 613 (Ag), which could exclusively assigned to the monoclinic VO2(M) phase.7, 8 And the sample with 3.8 ‰ hydrogen content shows a broad band from 300 to 800 cm-1, giving the typical features for the rutile structure as confirmed by previous
S13
studies.9 That is to say, Raman spectra revealed that the sample without H involving is in VO2(M) phase, and with 3.8 ‰ hydrogen content is in the rutile VO2(R) phase, respectively, which not only provides the supplementary evidence for the results from systematically characterizations based on XRD, XAFS, FTIR etc., but also provide a solid evidence to confirm the successful stabilization of high-temperature phase of rutile VO2(R) to room temperature.
S6. The hydrogen concentration of the as-obtained hydric vanadium dioxides The hydrogen concentrations of the as-obtained vanadium dioxides could be given by the elemental analysis results as well as the 1H solid-state MAS NMR spectrum. Based on the calculation results of the bulk elemental analysis by a VARIO ELIII (German) element analytical instrument, the hydrogen contents of the samples obtained by heating at 250 oC (hydric rutile VO2(R)), 300 oC (mixture of hydric monoclinic and rutile sample VO2(M-R)), and 600 oC (monoclinic VO2(M)) are 3.8‰, 1.3‰, about 0 ‰, respectively. These results were further verified by the 1H solid-state MAS NMR spectrum of vanadium dioxides. Figure S9 shows the 1H solid-state MAS NMR spectrum of hydric VO2(R), hydric VO2(R-M), and monoclinic VO2(M). Except the signal at around 1.5 ppm arises from the background of the employed NMR rotor10, the signals at 6.9 ppm and 10.6 ppm could be assigned to the presence of hydroxyl group.11 Keeping the experimental condition constant especially for the mass of VO2 samples, the integral area at the signal 6.9 ppm and 10.6 ppm are correlated with the hydrogen content of the concerned samples. Although the 1H
S14
solid-state NMR couldn’t give the exact value of the hydrogen concentration, it clearly gives the trend of hydrogen concentration with the following order: hydric VO2(R) > hydric VO2(R-M) > monoclinic VO2(M). In effect, the infinite 1×1 octahedra tunnel structure of rutile VO2(R) provides the accommodated space for the incorporated hydrogen ions, and the H+ incorporation would significantly influence the phase-transition behavior of VO2.
Hydric VO2(R) phase
Intensity (a.u.)
800
600 Hydric VO2(M-R) phase
400 Monoclinic VO2
200
0
0
10
20
Chemical Shift (ppm) Figure S9.
1
H solid-state MAS NMR spectrum of hydric VO2(R), hydric VO2(R-M), and
monoclinic VO2(M).
S15
S7. DSC thermograms of hydric VO2(R), hydric VO2(M-R) and monoclinic VO2(M)
Figure S10. DSC thermograms of monoclinic VO2(M) (a), hydric VO2(M-R) (b) and hydric VO2(R) (c) measured on both heating and cooling ways. Since monoclinic VO2(M) and rutile VO2(R) were well-known as the two states in a fully-reversible phase transition, the differential scanning calorimeter (DSC) and temperature dependent of zero-field cooled (ZFC) magnetization were performed to identify the phase kind at a given temperature. Figure S10 shows the differential scanning calorimetry (DSC) thermograms of as-prepared monoclinic VO2(M), hydric VO2(R), and hydric VO2(M-R) measured on both heating and cooling processes. For VO2(M), the lattice distortion and changes of conduction electrons correlations would answer for the entropy change when passing through the phase transition
12
and
thus DSC curves providing the phase-transition behaviors gives the material state at a given
S16
temperature. As shown in Figure S10, for the hydric VO2(R) sample, no signs of any abnormity in DSC curve, indicating there were no any structural phase transition in whole temperature range. The monoclinic VO2 sample shows the endothermic and exothermal transition temperature were 65.3 oC and 59.9 oC, respectively, revealing the semiconducting state at room temperature. However, the intermediate hydric VO2(M-R) shows a broad feeble peak in the wide temperature range from -20 oC to 60 oC for both heating and cooling curves, indicating that the coexistence of metallic rutile VO2(R) and semiconducting monoclinic VO2(M) in a wide temperature range covering the ambient temperature range.
S8. ZFC curves of the as-obtained hydric vanadium dioxides samples The phase-transition process from metallic rutile VO2(R) to insulating monoclinic VO2(M) accompanies with the orientation changes from sharing to localized V d-orbit electron clouds, which would bring the magnetic transformation from magnetic moment disordering to ordering that leads to the magnetic susceptibility change on passing through the phase transition. Figure S11 shows the M(T) around the transition temperature for the VO2 samples in a magnetic field of 200 Oe. For the monoclinic VO2(M) sample, the susceptibility is smaller and temperature independent below the transition temperature, and then the obvious discontinuity with abrupt increase curves presents near the phase transition point. The discontinuity in monoclinic VO2(M) were contributed by the relatively higher density of free carriers for the appearance of metallic phase, as well as the changes
S17
of Van Vleck contributions by the shifting of atomic cores.13 For hydric VO2(M-R) sample, the changes of magnetic susceptibility exhibits more moderate and the discontinuity covered a wide temperature range from 260 K to 330 K, indicating the phase transition continuous a very large temperature region including the ambient temperature range. However, for the hydric VO2(R) samples, there were no any signs of discontinuity, revealing that the as-obtained sample has no any phase transition over the concerned temperature range. 0.44
(a)
0.40
χ (emu/mol)
0.36
(b)
0.32
0.28
(c)
0.24
0.20 220
240
260
280
300
320
340
360
Temperature (K)
Figure S11. ZFC magnetization as a function of temperature in an applied magnetic field of 200 Oe for monoclinic VO2(M) (a), hydric VO2(M-R) (b) and hydric VO2(R) (c).
S18
S9. Density of states (DOS) of hydric rutile VO2(R) and monoclinic VO2(M) hybrid density function calculations
according to the
Figure S12. Density of states (DOS) of hydric rutile VO2(R) (a) and monoclinic VO2(M) (b) according to the hybrid density function calculations.
References:
1. Wu, C.Z.; Feng, F.; Feng, J.; Dai, J.; Yang, J.-L.; Xie, -Y. J. Phys. Chem. C 2011, 115, 791. 2. Kresse, G.; Joubert, D. Phys. Rev. B 1999, 59, 1758. 3. Heyd, J.; Scuseria, G. E.; Ernzerhof, M. J. Chem. Phys. 2006, 124, 219906. 4. Wu, C.-Z.; Zhang, X.-D.; Dai, J.; Yang, J.-L.; Wu, Z.-Y.; Wei, S.-Q.; Xie, Y. J. Mater. Chem. 2011, 21, 4509. 5. Valmalette, J. C.; Gavarri, J. R. Mater. Sci. Engineer. B 1998, 54, 168.
S19
6. Quantitative Chemical Analysis, 4th ed; Harris, D. C., Ed.; W. H. Freeman and Company: New York, 1995. 7. Chou, J. Y.; Lensch-Falk, J. L.; Hemesath, E. R. and Lauhon, L. J. J. Appl. Phys. 2009, 105, 034310-1-6. 8. Zhang, S.X., Kim, I.S., Lauhon, L.J. Nano. Lett. 2011,Dx.doi.org/10.1021/nl103925m 9. Schilbe, P. Physica B 2002, 316, 600. 10. Fang, J.; Shi, F.; Bu, J.; Ding, J.; Xu, S.; Bao, J.; Ma, Y.; Jiang, Z.; Zhang, W.; Gao, C.; Huang, W. J. Phys. Chem. C 2010, 114, 7940. 11. (a) Crocker, M.; Herold, R. H. M.; Wilson, A. E.; Mackay, M.; Emeis, C. A.; Hoogendoorn, A. M.
J. Chem. Soc., Faraday Trans. 1996, 92, 2791. (b) Yang, J.; Zhang, M. J.; Deng, F.; Luo, Q.;
Yi, D. L.; Ye, C. H. Chem. Commun. 2003, 884. (c) Zhang, H. L.; Yu, H. G.; Zhang, A. M.; Li, S. H.; Shen, W. L.;(d) Deng, F. Environ. Sci. Technol. 2008, 42, 5316. 12. Berglund, C. N.; Guggenheim, H. J. Phys. Rev. 1969, 185, 1022. 13. (a) Imada, M.; Fujimori, A.; Tokura, Y. Rev. Mod. Phys. 1998, 70, 1039. (b) Holman, K. L.; McQueen, T. M.; Williams, A. J.; Klimczuk, T.; Stephens, P. W.; Zandbergen, H. W.; Xu, Q.; Ronning, F.; Cava, R. J. Phys. Rev. B 2009, 79, 245114.
S20
Hydrogen-incorporation stabilization of metallic VO2(R) phase to room temperature displaying promising low-temperature thermoelectric effect
Changzheng Wu†, Feng Feng†, Jun Feng†, Jun Dai†, Lele Peng†, Jiyin Zhao†, Jinlong Yang†, Cheng Si ‡, Ziyu Wu‡ and Yi Xie *, † †Hefei National Laboratory for Physical Sciences at Microscale and ‡National Synchrotron Radiation Laboratory, University of Science & Technology of China, Hefei, Anhui 230026, P.R. China
S1
S1. Experimental section 1．Synthesis of hydric paramontroseite VO2 precursor The hydric paramontroseite VO2 was prepared from post-treatment of the synthetic goethite VOOH sample similar to the reported literature by our group 1. In a typical synthesis procedure, the hydric paramontroseite VO2 sample was prepared by the post heating-treatment of goethite VOOH at a lower temperature of 80 oC. And then the hydric paramontroseite VO2 sample was collected for the following characterizations and synthesis procedures. The XRD pattern of the as-obtained hydric paramontroseite VO2 (Figure S1) matches well with the standard JCPDS card NO. 73-0514, which can be readily indexed into paramontroseite VO2 with space group Pbnm. The thermo-gravimetric (TG) curve in air atmosphere for the as-obtained hydric paramontroseite VO2 reveals the hydric feature and provides a clue to know the hydrogen contents of the hydric sample. As shown in Figure S2, a small platform exists at around 100 oC could be attributed to loss of absorbed water molecules, indicating 0.5% water content in the sample. And then the continuous weight lose ranging from 100 o
C to 280 oC could be assigned to the lose of hydrogen ingredients in the hydric paramontroseite VO2.
And thus it could be found that the hydrogen contents of hydric paramontroseite VO2 precursor is about 0.9%.
S2
301
112
131 230
310
230 220
240
130
021
120 110
Relative Intensity (a.u.)
Experimental Pattern
JCPDS Card No. 73-0514
10
20
30
40
50
60
70
2θ (degrees)
Figure S1. Experimental XRD pattern of hydric paramontroseite VO2 precursor and standard pattern in JCPDS card NO. 73-0514. 1.002 1.000
Weight (%)
0.998 0.996 99.5%
0.994 0.992 0.990 0.988 0.986
98.6% 50
100
150
200
250
300
350
Temperature ( OC)
Figure S2.
TG curve of hydric paramontroseite VO2 precursor in air atmosphere.
S3
2. Heating transformation of hydric paramontroseite VO2 precursor into hydric VO2 and monoclinic VO2(M) The as-prepared black hydric paramontroseite VO2 precursor was pressed into bulk sample with a cuboid shape (8 mm × 6mm × 3mm) under a pressure of 10 Mpa for 20 min in order to decrease the porosity of the bulk pallets. And then the pallet samples of hydric paramontroseite VO2 were further heated in a nitrogen gas flow at 250 oC, 300 oC, and 600 oC for 3h to produce hydric rutile VO2(R), hydric VO2(M-R), and monoclinic VO2(M), respectively. In our case, the heating post-treatment of the pellets could effectively remove the grain boundary forming the well-defined pellet sample with high crystalline that was vital for the following transport experiments. The color of hydric rutile VO2(R) and VO2(M-R) is black, while the color of monoclinic VO2(M) product slightly changed to blue-black. In our case, based on the calculation results of the bulk elemental analysis, the hydrogen content in the sample obtained at heating 250 oC (hydric rutile VO2(R)), 300 o
C (hydric monoclinic-and-rutile VO2(M-R)), and 600 oC (monoclinic VO2(M)) are 3.8 ‰, 1.3 ‰,
about 0 ‰, respectively. 3. Characterizations The as-prepared samples were characterized by X-ray powder diffraction (XRD) with a Philips X’Pert Pro Super diffractometer with Cu Kα radiation (λ=1.54178 Å). Infared (IR) spectra analyses were operated on samples palletized with KBr powders in the range 4000-400 cm-1, using an infared Fourier transform spectrophotometer (Nicolet, ZOSX). Elemental analysis of hydrogen element
concentrations for VO2 samples were measured on a VARIO ELIII (German) element
S4
analytical instrument. The magnetization was characterized by a superconducting quantum interference device (SQUID, quantum design MPMS XL-7) magnetometer from 200 and 360 K. DSC cycling curves were measured by the NETZSCH DSC 200 F3 with a heating/cooling rate of 5 o
C min-1 between -50 oC and 90 oC. Raman spectra were recorded at room temperature with a
LABRAM-HR Confocal Laser MicroRaman Spectrometer 750 K with a laser power of 0.5 mW. The HRTEM image and ED pattern were obtained on a JEOL-2010 transmission electron microscope at an acceleration voltage of 220 kV. The 1H solid NMR spectra were obtained with a 500.132 MHz NMR spectrometer (Bruker, Germany) at 298.2 K. The experimental data with involving eight transients were collected over 2048 complex points. A mixing time of 300 ms, a relaxation delay of 2s, an acquisition time of 205 ms, and a 90° pulse width of 8.2 μs were used. The data were calculated with a Lorentzianto-Gaussian window function and zero filling in both dimensions to display data on a 2048 × 2048 2D-matrix. All the export data for solid NMR results were performed by NMRPipe software on a Linux workstation. The thermal and electrical transport properties of the bulk sample were measured using a Physical Property Measurement System (PPMS) from 210 to 360 K. 4．Calculation details For VO2(R), we adopted the experimental lattice parameters and atom position for calculations. For hydric VO2(R), since its crystal data from the structural characterizations were almost same with
S5
that of pure rutile VO2(R). Therefore, H atoms were incorporated in the void space of VO2(R) keeping rutile V-O lattice framework (Please see the hydric rutile VO2(R).cif in supporting information for detail structure). The calculation of density of states (DOS) was carried out using projected augmented wave (PAW) method as implemented in Vienna Ab-initio Simulation Package (VASP) 2 . The HSE06 form of hybrid density function
3
was used to describe the electron
exchange-correlation interaction with a 8×8×14 Monkhorst-Pack grid of Brillouin-zone k-point sampling.
S2. XRD patterns of the as-obtained hydric vanadium dioxides The XRD pattern for the product obtained at higher heating temperature of 600 oC for 3h was shown in Figure S3c, of which all of the reflection peaks can be readily indexed to the typical monoclinic VO2(M) (JCPDS card 82-0661, a=5.751 Å, b=4.528 Å, c = 5.380 Å). And the room-temperature XRD pattern for the sample obtained at a lower 250 oC was shown in Figure S3a, which could be readily indexed into the tetragonal rutile VO2(R) (JCPDS card 79-1655, a=4.552 Å, c =2.856 Å), indicating a room-temperature rutile phase is indeed achieved in our case. Moreover, the absence of the typical XRD peaks (100)M (26.94o) and (-102)M(33.40o), the position shifting of VO2(R) (110) peak into lower 2θ degrees compared with VO2(M) (011) peaks in 26o ≤ 2θ ≤29o (Figure S3d), and the splitting of VO2(R) (310) and (002) in 64
o
≤ 2θ ≤66
o
(Figure S3e) that
evolved from VO2(M) (130) peak, clearly verified that the heating treatment of the as-prepared hydric paramontroseite VO2 in a N2 gas flow produces the vanadium oxides with gradient hydrogen
S6
contents of 3.8 ‰, 1.3 ‰, nearly zero, which could be indexed to hydric rutile VO2(R), hydric VO2(R-M), and monoclinic VO2(M) at room temperature, respectively.
Relative Intensity (a.u.)
VO2(M) JCPDS 82-0661
Monoclinic VO2(M)
(c) Hydric VO2(M-R)
(b)
Hydric VO2(R)
(a)
VO2(R) JCPDS 79-1655
VO2(M)
60
(e)
(110) R
VO2(R) with 3.8 ‰ H
28
29
2θ (degrees)
64
(130) M
2θ (degree)
VO2(M-R) with 1.3 ‰ H
27
50
70
VO2(M)
Hydric VO2(M-R) with 1.3 ‰ H
(002) R
40
(310) R
(100) M
(d) Relative Intensity (a.u.)
30
Intensity (a.u.)
20 (011) M
10
Hydric VO2(R) with 3.8 ‰ H
65 66 2θ (degrees)
Figure S3. (a-c) Room-temperature XRD patterns of the as-obtained vanadium dioxide samples obtained by heating at 250 oC for 3h (a), 300 oC for 3h (b), and 600 oC for 3 h (c), which can be readily indexed to rutile VO2(R), intermediate VO2(M-R) and monoclinic VO2(M), respectively. (d)
S7
The partial XRD pattern with 2θ value ranging from 26.5 to 29o. (e) Partial XRD patterns indexing with 2θ value ranging from 64 o to 66o.
S3. HRTEM image and ED pattern of hydric VO2(R) The HRTEM image shows the typical rutile structure character in the 3.8 ‰-hydrogen including VO2 sample. Figure S4 shows the representative SAED pattern taken a typical edge of well-dispersed hydric VO2 particle. Our experimental ED pattern gives the symmetrical tetragonal dotted lattice. The lattice fringes in Figure S4 show two sets of perpendicularly lattice spacing of about 3.216 Å that correspond to the (110) and (1-10) planes of rutile VO2, which shows the typical rutile structure’s HRTEM image and SAED pattern along the [001] axis, confirming the existence of rutile structure in the 3.8 ‰-hydrogen including sample. Furthermore, the atomic model of rutile VO2 projected along the [001] axis also exhibits tetragonal symmetry, and the [110] and [1-10] axes are indicated by the arrows in Figure S4b, giving further evidence for the rutile structure of sample with 3.8 ‰-hydrogen content. In a word, the as-obtained HRTEM result provided more diagnostic and direct evidence for the rutile phase of the VO2 sample with 3.8 ‰-hydrogen content.
S8
(a)
(b)
Figure S4. (a) HRTEM image and SAED pattern taken from a typical edge of well-dispersed hydric VO2(R) particle projected along [001] direction. (b) The corresponding atomic model illustration for the HRTEM images in (a) shows the vertical crystal planes of (110) and (1-10), giving the representive rutile structural feature
S9
S4. The atomic crystal structure of monoclinic VO2(M) and rutile VO2(R)
Figure S5. Structural illustration of monoclinic VO2(M) and rutile VO2(R). (a) Monoclinic VO2(M) supercell structure. (b) The local environment illustration of the vanadium atom in crystal structure of monoclinic VO2(M). (c) Rutile VO2(R) supercell structure. (d) The V local environment illustration in monoclinic VO2(M) structure. For sake of simplicity, the V atom is fixed relatively to the adjacent O and V atoms.
S10
Figure S6. The atomic model of the V-V chains in the rutile VO2(R) supercell structure, showing the infinite vanadium-vanadium chains along z-axis, i.e. the c-direction of rutile structure.
S5. FTIR spectra and Raman spectra of hydric VO2(R) and monoclinic VO2(M) 1. FTIR spectra
749
(2) Hydric VO2(M-R)
1625
Relative Intensity (a.u.)
530 500
(1) VO2(M)
(3) Hydric VO2(R)
1000 1500 2000 2500 -1 Wavenumber (cm )
Figure S7. Room temperature FTIR spectra for the as-obtained hydric vanadium dioxide samples.
S11
FTIR spectrum could reveal the structural and electrical information according to unique spectral characteristics of semiconducting VO2(M) and metallic VO2(R). As is known, the rutile VO2(R) with higher structural symmetry has the less vibration modes than the monoclinic VO2(M), which would significantly depress the fingerprint region in FTIR spectrum for rutile phase.4,5 In the typical IR spectrum of monoclinic VO2(M) sample as shown in Figure S7, the IR absorption peaks at 749 cm-1 and 530 cm-1 could be attributed to the stretching vibrations (V-O-V)4, giving the monoclinic typical feature. While for the IR spectrum of the sample with 3.8‰, the depressed fingerprint peaks provide a clue for indexing the sample to metallic VO2(R). Also, in FTIR spectrum, the absorption band at 1625 cm-1 in the FTIR spectrum signify the evidence of the isolated hydroxyl groups with the deformation vibrations (O-H).6 Since strict same experimental conditions, the peak intensities with the sequence of (a)>(b)>(c) here reflects the hydrogen concentration in a sequence of hydric VO2(R), hydric VO2(M-R), and monoclinic VO2, the results of which were consistent with the elemental analysis and the 1H solid-state MAS NMR results.
S12
2. Raman spectra (a) Monoclinic VO2(M)
360 340
Intensity (a.u.)
50
191
Relative Intensity (a.u.)
380
613
320 300 280
222 401
260
304
240 220 200
40
400
600
800
1000
hydric Rutile VO2(R)
30
20
10
180 200
(b)
200
1200
400
600
-1
800
1000
Raman Shift ( cm )
-1
Raman shift (cm )
Figure S8. Raman spectra of the as-obtained monoclinic VO2(M) without H incorporation and hydric rutile VO2(R) with 3.8 ‰ hydrogen content. Raman spectra provide a fingerprint for identifying the micro-structural characteristics of bonding and coordinations for inorganic solids. As a supplementary evidence, Raman scattering was used to distinguish between monoclinic VO2(M) and rutile VO2(R). In our case, Raman spectra of monoclinic VO2(M) and hydric Rutile VO2(R) were recorded at room temperature using a Ar+ ion laser (514 nm) excitation source and the laser power was set at a low value of 0.5 mW to minimize the local heating of two samples, ensuring them not to be oxidized during the Raman test. As is shown, the sample without H content shows Raman peaks at 191 (Ag), 222 (Ag), 260 (Bg), 307 (Bg), 340 (Ag), 386 (Ag), 613 (Ag), which could exclusively assigned to the monoclinic VO2(M) phase.7, 8 And the sample with 3.8 ‰ hydrogen content shows a broad band from 300 to 800 cm-1, giving the typical features for the rutile structure as confirmed by previous
S13
studies.9 That is to say, Raman spectra revealed that the sample without H involving is in VO2(M) phase, and with 3.8 ‰ hydrogen content is in the rutile VO2(R) phase, respectively, which not only provides the supplementary evidence for the results from systematically characterizations based on XRD, XAFS, FTIR etc., but also provide a solid evidence to confirm the successful stabilization of high-temperature phase of rutile VO2(R) to room temperature.
S6. The hydrogen concentration of the as-obtained hydric vanadium dioxides The hydrogen concentrations of the as-obtained vanadium dioxides could be given by the elemental analysis results as well as the 1H solid-state MAS NMR spectrum. Based on the calculation results of the bulk elemental analysis by a VARIO ELIII (German) element analytical instrument, the hydrogen contents of the samples obtained by heating at 250 oC (hydric rutile VO2(R)), 300 oC (mixture of hydric monoclinic and rutile sample VO2(M-R)), and 600 oC (monoclinic VO2(M)) are 3.8‰, 1.3‰, about 0 ‰, respectively. These results were further verified by the 1H solid-state MAS NMR spectrum of vanadium dioxides. Figure S9 shows the 1H solid-state MAS NMR spectrum of hydric VO2(R), hydric VO2(R-M), and monoclinic VO2(M). Except the signal at around 1.5 ppm arises from the background of the employed NMR rotor10, the signals at 6.9 ppm and 10.6 ppm could be assigned to the presence of hydroxyl group.11 Keeping the experimental condition constant especially for the mass of VO2 samples, the integral area at the signal 6.9 ppm and 10.6 ppm are correlated with the hydrogen content of the concerned samples. Although the 1H
S14
solid-state NMR couldn’t give the exact value of the hydrogen concentration, it clearly gives the trend of hydrogen concentration with the following order: hydric VO2(R) > hydric VO2(R-M) > monoclinic VO2(M). In effect, the infinite 1×1 octahedra tunnel structure of rutile VO2(R) provides the accommodated space for the incorporated hydrogen ions, and the H+ incorporation would significantly influence the phase-transition behavior of VO2.
Hydric VO2(R) phase
Intensity (a.u.)
800
600 Hydric VO2(M-R) phase
400 Monoclinic VO2
200
0
0
10
20
Chemical Shift (ppm) Figure S9.
1
H solid-state MAS NMR spectrum of hydric VO2(R), hydric VO2(R-M), and
monoclinic VO2(M).
S15
S7. DSC thermograms of hydric VO2(R), hydric VO2(M-R) and monoclinic VO2(M)
Figure S10. DSC thermograms of monoclinic VO2(M) (a), hydric VO2(M-R) (b) and hydric VO2(R) (c) measured on both heating and cooling ways. Since monoclinic VO2(M) and rutile VO2(R) were well-known as the two states in a fully-reversible phase transition, the differential scanning calorimeter (DSC) and temperature dependent of zero-field cooled (ZFC) magnetization were performed to identify the phase kind at a given temperature. Figure S10 shows the differential scanning calorimetry (DSC) thermograms of as-prepared monoclinic VO2(M), hydric VO2(R), and hydric VO2(M-R) measured on both heating and cooling processes. For VO2(M), the lattice distortion and changes of conduction electrons correlations would answer for the entropy change when passing through the phase transition
12
and
thus DSC curves providing the phase-transition behaviors gives the material state at a given
S16
temperature. As shown in Figure S10, for the hydric VO2(R) sample, no signs of any abnormity in DSC curve, indicating there were no any structural phase transition in whole temperature range. The monoclinic VO2 sample shows the endothermic and exothermal transition temperature were 65.3 oC and 59.9 oC, respectively, revealing the semiconducting state at room temperature. However, the intermediate hydric VO2(M-R) shows a broad feeble peak in the wide temperature range from -20 oC to 60 oC for both heating and cooling curves, indicating that the coexistence of metallic rutile VO2(R) and semiconducting monoclinic VO2(M) in a wide temperature range covering the ambient temperature range.
S8. ZFC curves of the as-obtained hydric vanadium dioxides samples The phase-transition process from metallic rutile VO2(R) to insulating monoclinic VO2(M) accompanies with the orientation changes from sharing to localized V d-orbit electron clouds, which would bring the magnetic transformation from magnetic moment disordering to ordering that leads to the magnetic susceptibility change on passing through the phase transition. Figure S11 shows the M(T) around the transition temperature for the VO2 samples in a magnetic field of 200 Oe. For the monoclinic VO2(M) sample, the susceptibility is smaller and temperature independent below the transition temperature, and then the obvious discontinuity with abrupt increase curves presents near the phase transition point. The discontinuity in monoclinic VO2(M) were contributed by the relatively higher density of free carriers for the appearance of metallic phase, as well as the changes
S17
of Van Vleck contributions by the shifting of atomic cores.13 For hydric VO2(M-R) sample, the changes of magnetic susceptibility exhibits more moderate and the discontinuity covered a wide temperature range from 260 K to 330 K, indicating the phase transition continuous a very large temperature region including the ambient temperature range. However, for the hydric VO2(R) samples, there were no any signs of discontinuity, revealing that the as-obtained sample has no any phase transition over the concerned temperature range. 0.44
(a)
0.40
χ (emu/mol)
0.36
(b)
0.32
0.28
(c)
0.24
0.20 220
240
260
280
300
320
340
360
Temperature (K)
Figure S11. ZFC magnetization as a function of temperature in an applied magnetic field of 200 Oe for monoclinic VO2(M) (a), hydric VO2(M-R) (b) and hydric VO2(R) (c).
S18
S9. Density of states (DOS) of hydric rutile VO2(R) and monoclinic VO2(M) hybrid density function calculations
according to the
Figure S12. Density of states (DOS) of hydric rutile VO2(R) (a) and monoclinic VO2(M) (b) according to the hybrid density function calculations.
References:
1. Wu, C.Z.; Feng, F.; Feng, J.; Dai, J.; Yang, J.-L.; Xie, -Y. J. Phys. Chem. C 2011, 115, 791. 2. Kresse, G.; Joubert, D. Phys. Rev. B 1999, 59, 1758. 3. Heyd, J.; Scuseria, G. E.; Ernzerhof, M. J. Chem. Phys. 2006, 124, 219906. 4. Wu, C.-Z.; Zhang, X.-D.; Dai, J.; Yang, J.-L.; Wu, Z.-Y.; Wei, S.-Q.; Xie, Y. J. Mater. Chem. 2011, 21, 4509. 5. Valmalette, J. C.; Gavarri, J. R. Mater. Sci. Engineer. B 1998, 54, 168.
S19
6. Quantitative Chemical Analysis, 4th ed; Harris, D. C., Ed.; W. H. Freeman and Company: New York, 1995. 7. Chou, J. Y.; Lensch-Falk, J. L.; Hemesath, E. R. and Lauhon, L. J. J. Appl. Phys. 2009, 105, 034310-1-6. 8. Zhang, S.X., Kim, I.S., Lauhon, L.J. Nano. Lett. 2011,Dx.doi.org/10.1021/nl103925m 9. Schilbe, P. Physica B 2002, 316, 600. 10. Fang, J.; Shi, F.; Bu, J.; Ding, J.; Xu, S.; Bao, J.; Ma, Y.; Jiang, Z.; Zhang, W.; Gao, C.; Huang, W. J. Phys. Chem. C 2010, 114, 7940. 11. (a) Crocker, M.; Herold, R. H. M.; Wilson, A. E.; Mackay, M.; Emeis, C. A.; Hoogendoorn, A. M.
J. Chem. Soc., Faraday Trans. 1996, 92, 2791. (b) Yang, J.; Zhang, M. J.; Deng, F.; Luo, Q.;
Yi, D. L.; Ye, C. H. Chem. Commun. 2003, 884. (c) Zhang, H. L.; Yu, H. G.; Zhang, A. M.; Li, S. H.; Shen, W. L.;(d) Deng, F. Environ. Sci. Technol. 2008, 42, 5316. 12. Berglund, C. N.; Guggenheim, H. J. Phys. Rev. 1969, 185, 1022. 13. (a) Imada, M.; Fujimori, A.; Tokura, Y. Rev. Mod. Phys. 1998, 70, 1039. (b) Holman, K. L.; McQueen, T. M.; Williams, A. J.; Klimczuk, T.; Stephens, P. W.; Zandbergen, H. W.; Xu, Q.; Ronning, F.; Cava, R. J. Phys. Rev. B 2009, 79, 245114.
S20
