Synthesis of Monoclinic Vanadium Dioxide via One-pot Hydrothermal Route

: Pure monoclinic vanadium dioxide nanoparticles (VO 2 NPs) with a controlled uniform size are considered essential for the preparation of thermochromic smart window coatings on desired substrates. Herein, we report a facile one-step hydrothermal synthesis of VO 2 (M) NPs without post-treatment of annealing, which may induce unwanted aggregation of NPs. In contrast with the annealed sample, the one-step processed VO 2 (M) NPs exhibit superior thermochromic performance with the solar modulation efficiency of 11.8% and luminous transmittance of 37.3%.

The key to realizing these intensive applications depends on the controlled preparation of high-quality VO2 single/polycrystals powders or thin-film due to the multivalent nature of vanadium element, making it tough work to synthesize high purity VO2. Compared with some high-cost or complex methods, such as electrodeposition [13], sputtering [14], pulse laser [15], and sol-gel [16], hydrothermal synthesis is probably the most effective method to control the morphology, size, and phase structures of VO2. Recently, Li et al. summarized a detailed review on the hydrothermal synthesis of the VO2 phase and its application in thermochromic smart windows [17], from which we know that most of the reported one-step hydrothermal syntheses require a high temperature and long reaction time. Meanwhile, in most cases, certain metal doping, such as tungsten [18,19], is necessary to obtain VO2(M) nanoparticles (NPs). However, the doping elements that substitute vanadium elements normally deteriorate the crystallinity and hinder the thermochromic performance of VO2. Son et al. prepared the VO2(M) phase with different morphologies by changing hydrothermal temperatures, pH value, and duration [20]. However, the product was generally a mixture of the two phases, which is not suitable for the actual applications. Chen et al. synthesized the fine crystalline VO2(M) NPs via one-step hydrothermal, and they found both a high latent heat of 43 J/g and good thermochromic property of the VO2 in comparison to the bulk one [21]. However, the used high hydrothermal temperature, up to around 400 °C, leads to a high cost and uneven size distribution of VO2 NPs. Alie et al. also reported the direct hydrothermal synthesis of VO2 without doping via modifying the molar ratio of oxalic acid and V2O5 [22]. The obtained VO2 microcrystals showed a mixture of star shape and hollow sphere shape. The use of seed-assisted onestep hydrothermal to form VO2 composite has also been recently reported by Jin et al. They reported a novel VO2-ZnO dandelion structure with the help of the TiO2 seeds [23]. Li et al. synthesized TiO2 seeds assisted Mo-doped VO2 zero-dimensional NPs due to the low nucleation energy [24]. A novel one-step hydrothermal was invented by Guo et al., in which they separated the hydrogen peroxide from the precursor solution, providing a rational oxidizing hydrothermal system [25]. However, the duration of hydrothermal synthesis has been doubled at the same time. Powell et al. and Malarde et al. introduced a continuous hydrothermal flow synthesis method for the direct preparation of a large amount VO2(M) without a post-annealing process, while the mixing temperature synthesis was more than 400 °C [26,27]. Although the continuous hydrothermal synthesis showed a great value towards the industrial procedure, the as-prepared VO2 NPs exhibited inferior thermochromic performance, proving the poor crystallinity and quality of the product in general.
In this paper, we successfully synthesize non-doping VO2(M) NPs via one-step hydrothermal at a relatively low temperature within a short time by adjusting hydrothermal parameters, including the amount of the reduction, reaction time, and precursor concentration. The obtained VO2(M) has evenly distributed uniform size and morphology without further annealing treatment, which displays high crystallinity and good thermochromic performance.

Materials and Synthesis
In the experiments, all chemicals of V2O5(Alfa Aesar, Ward Hill, MA, USA, 99.6%), N2H4·H2O (Alfa Aesar, Ward Hill, MA, USA, 99.9%), and H2O2 (30 wt%, Sigma-Aldrich, St. Louis, MI, USA) were used as received without any further purification. In a typical hydrothermal process, 1 mmol commercial V2O5 powders were added into 5 mL preheated H2O2 (30 wt%) at 90 °C with 200 rpm stirring. After a vigorous reaction, a brownish V2O5-H2O2 precursor was obtained. A certain amount of deionized (DI) water was added subsequently. Afterward, the brownish precursor was chemically reduced by adding N2H4·H2O solution dropwise with stirring for 10 min. After the precursor turned blueblack, it was transferred into a 50 mL PTFE liner and sealed in a stainless steel autoclave. The hydrothermal temperature was set at 260 °C for a certain period (12-168 h), and then the samples were cooled to room temperature in the air. The black precipitate was collected by centrifugation and washed several times in turn with DI water, acetone, and ethanol. After being dried at 50 °C in a vacuum oven, the final black powders were obtained. No further annealing process was performed on any of the products.

Preparation of VO2-Based Thermochromic Film
The 50 mg of as-prepared VO2 NPs were evenly dispersed into 5 mL ethanol and 5 mL concentrated Si-Al solution by one-hour ultrasonication bath. The synthesis of the Si-Al solution has been reported in our previous work [8]. The VO2 films were further prepared by spin coating on the soda-lime glass substrate with the same speed and coating times. The coated substrate was dried in the vacuum oven for one hour at 90 °C, then another one hour at 150 °C to remove the solvent and form the uniform coatings.

Characterization
The X-ray diffraction (XRD) patterns of the obtained powders were collected from a Shimazu Powder XRD-6000 X-ray diffractometer with Cu Kα radiation (λ = 0.15406 nm) under 40 KV voltage and 30 mA current. The surface morphology of the samples was characterized by field emission scanning electron microscope (FESEM, JSM-7600F, JEOL, Akishima, Tokyo, Japan). Differential Scanning Calorimeter (DSC Q10, TA instrument, New Castle,DE, USA) was used to characterize the phase transition behavior of the samples upon the heating and cooling rate of 10 °C/min under N2 atmosphere. Variable temperature optical transmission spectra of the prepared VO2 films were measured on a UV-Vis-NIR spectrophotometer (Cary 5000, Agilent Ltd., Santa Clara, CA, USA) in the wavelength of 250-2500 nm at 20 and 90 °C, respectively. A Linkam PE120 system Peltier heating and the cooling stage was attached to regulate the temperature of the substrates. The integrated luminance transmittance and solar transmittance were calculated using the below expression based on the transmittance data: where T(λ) denotes the film transmittance data at wavelength λ, ϕ is the solar irradiance spectrum for air mass 1.5 (corresponding to the sun standing 37° above the horizon). A more detailed discussion on the evaluation method was reported in [28].

Effect of the Amount of the Reduction
As discussed in our previous paper [29], the amount of reduction is the crucial factor for the synthesis of VO2 phase. Thus, in a series of hydrothermal experiments, the amount of N2H4·H2O solution was optimized first.
The XRD patterns of the hydrothermal samples under the same condition, but with different amounts of the reduction, are shown in Figure 1. The crystallography data are listed in Table 1. The phase evolution of the obtained vanadium oxide could be observed from the increment of the reduction. When the initial amount of N2H4·H2O was 30 μL, the hydrothermal product was still V2O5, which is the same phase as the raw material. This is due to the fact that the reduction was mainly consumed by the residual H2O2 in the precursor. After the amount of N2H4·H2O increased to 50 μL, the V2O5 was reduced to pure V6O13. When 60 μL of N2H4·H2O was selected, a high crystalline pure phase of VO2(M) (JCPDS #43-1051) was synthesized. The pure VO2(M) was still retained when the amount of reduction increased to 70 μL, but from the XRD pattern, the crystallinity of the product became much lower. As shown in Figure 1e,f, the powders were reduced to V6O11 and V3O5, upon which the amount of reduction was further raised to 80 and 90 μL. Additionally, the crystallinity of the powders gradually decreased with increment of N2H4·H2O no matter what kind of vanadium oxide phase it was. Therefore, to synthesize the VO2 NPs with high crystallinity, 60 μL reduction was chosen as the standard amount for the consequent hydrothermal synthesis.  The morphology of the above samples is shown in Figure 2a-e. With the phase evolution of the obtained vanadium oxides, the morphology of these samples changes from nanowires/microwires dramatically into sub-100 nm NPs and further growth into larger particles. In Figure 2a, the length of V2O5 nanowires is up to several micrometers, which is different from the sheet structure of V2O5-H2O2 precursor both in size and shape reported before [30]. Figure 2b shows the microbelt structure of the obtained V6O13 powder, which is caused by the aggregation of the initially formed nanowires under hydrothermal conditions. Simultaneously, small amounts of the NPs could be observed at the edge of the microwire. From Figure 2c, the morphology of the VO2 product is the uniform NPs with a diameter around 100 nm. No nanowire or micrometer was observed. Upon further adding the reduction, the size of the particles is getting larger and irregular, from NPs to micrometer bulks, as shown in Figure 2d,e. The irregular morphology also causes a decrease in the crystallinity, as discussed in Figure 1 of the XRD results. Based on Figures 1 and 2, a hydrothermal synthesis mechanism could be speculated. With the high temperature and pressure, the sheet structure of the precursor turned into nanowires at the initial stage. The reduction process began with the increased amount of N2H4·H2O. The nanowires grew together into the microwires via an oriented attachment mechanism, and then the NPs gradually appeared at the edge of the microwires. The final morphology using N2H4·H2O as the reduction of the VO2 was NPs, which is similar to the previous reports. Herein, N2H4·H2O plays a crucial role in the hydrothermal process of VO2 (M), which provides a reducing environment and controls the phase and morphology of the NPs. Gui et al. speculated a molecular coordination mechanism, and the hydrazine and vanadium could be assembled as a building block and promote the growth of the particle-shape VO2 nanocrystalline [31]. The same group further synthesized nanorod VO2(R) via adding hydrazine, and they further postulated that hydrazine could serve as a soft template. Both Santulli et al. and Son et al. found the potential of N2H4 as the coordination agent in central vanadium ions into one-dimensional nanorod, nanoneedle morphologies [20]. Additionally, Ji et al. postulated that with the existence of N2H4, the [V2O7] 4− was the main vanadium complex, while with the presence of H2C2O4, the dominant vanadium complex was larger volume [V10O28] 6− , in which micro-size particles were favorably produced [32].

Effect of the Hydrothermal Duration
The second parameter discussed in the series of the hydrothermal preparation was the hydrothermal duration while keeping N2H4·H2O at 60 μL. The XRD results in Figure 3 show that the hydrothermal samples with different duration increased from 12 h to 168 h. It can be found that all samples are single-phase VO2(M) when the reduction amount was kept constant. The XRD peaks of the VO2(M) gradually intensified with the increase of the hydrothermal duration.  Figure 4a shows a large amount of the hollow half-sphere shape within 12 h hydrothermal synthesis, which indicates that a short time less than 12 h was far from enough for the growth of uniform NPs, and the size of the particles were rather uneven, ranging from 20 nm to 130 nm. With the increase in hydrothermal duration to 24 h, the uniform NPs morphology is shown in Figure 4b, with an average size of 60.2 nm. As shown in Figure 4c,d, the hydrothermal duration continued to be extended, the shape of the NPs remained the same, and the size grew larger. The average size was 75.9 nm for the 72 h sample, while the size of NPs grew to nearly 90 nm at an even longer duration at 168 h.

Effect of the Precursor Concentration
Besides the reduction amount and hydrothermal duration, the precursor concentration is the last hydrothermal parameter we were concerned with in our experiment. The preparation of the precursor was introduced in the experimental section. To adjust the precursor concentration, the volume of the solvent, including H2O2 and DI water, was kept unchanged, while the amount of the raw material of V2O5 powder was tuned from 1 mmoL up to 2 mmoL accordingly. Figure 5 shows the XRD patterns of the hydrothermal samples with different precursor concentrations. The crystallography data are listed in Table 2. Note that the concentration of V2O5 powder was kept at 1 mmol in our experiment as mentioned above. As shown in Figure 5b, all of the XRD peaks are attributed to the standard VO2(M) peaks. Two uncertain peaks marked with the star, around 18.7° and 37.8°, turned up as we increased the concentration of V2O5 up to 1.25 mmol (Figure 5b). Upon further increasing the V2O5 concentration to 1.5 mmol, Figure 5c exhibits that most of the peaks are assigned to VO2(B), although a small number of peaks attributed to VO2(M) could still be observed. When the concentration of V2O5 powder was increased to 2 mmol, V6O13 was synthesized, as shown in Figure 5d. However, there were still VO2(M) peaks. This suggests that the reduction amount is significant to synthesize the VO2(M), while the precursor concentration must be controlled within the corresponding range.  The SEM images in Figure 6 shows morphologies of the above samples, which highly coincided with the XRD results. Figure 6a shows the highly uniform NPs, which is the pure VO2(M) phase. In Figure 6b, two shapes of the particles are clearly shown. The NPs are VO2(M), while the thicker nanobeams mostly belong to the unknown phase indicated in the XRD pattern (Figure 5b). Two different kinds of morphology in NPs and ultralong nanowires are shown in Figure 6c. According to the XRD results, the nanowires with lengths up to several micrometers belong to VO2(B), while the NPs are VO2(M). In Figure  6d, it is found that besides the VO2(M) NPs, microbelts are observed. The microbelts comprise V6O13, based on the XRD result, which is completely identified as the V6O13 in Figure  2b.  Figure 7 shows the DSC of the synthesized VO2 NPs from different hydrothermal durations in the hydrothermal process. In general, endothermic peaks are observed in the heating segment, while exothermal peaks are found during the cooling process, and the thermal hysteresis indicates that the VO2 possesses first-order and fully reversible phase transition. The transition temperature and enthalpy of the samples were integrated by the collected data and determined by the TA analysis software. In Figure 7a, the 12 h hydrothermal sample barely exhibits any peaks. The bump appears at 66.2 °C during the heating ramp and at 43.4 °C during the cooling ramp. The transition temperature was 54.8 °C, and the latent heat is 7.6 J/g, much lower than the reported data [17]. This suggests a large number of defects in the samples, and it further proves that the insufficient hydrothermal time for the VO2 synthesis, as discussed in Figure 4a. As the duration increased to 24 h, the endothermic and exothermal peaks were both shifted towards the higher temperature, that is, 70.2 °C and 56.5 °C, which brought the τc to 63.4 °C, and the enthalpy increased to 37.3 J/g. As for the 72 h hydrothermal sample, the position of the peaks barely shifted, while the enthalpy further increased to 43.3 J/g. A more significant change for DSC peaks could be observed for the 168 h duration sample. The endothermic peak appeared at 77.4 °C and the exothermal bump at 59.2 °C, which puts the τc at 68.3 °C. The obtained enthalpy value was 49.4 J/g. In contrast, the endothermic or exothermal peaks for 72 h and 168 h are sharper, indicating the larger particle size for these samples (Figure 4). The trend that ΔH and τc both increased with the longer hydrothermal duration is further shown in Figure 7b.

Thermochromic Performance of VO2(M) NPs
The thermochromic property is evaluated through the temperature-dependent UV-Vis-NIR spectra, as shown in Figure 8. The solid and dotted lines demonstrate the high temperature and low-temperature optical transmittance, respectively, of the prepared VO2 films. Figure 8a shows all three samples with different concentrations of VO2 dispersion prepared from the non-doped single-phase VO2 NPs. The concentration of the prepared VO2 dispersion was adjusted by the added amount of the VO2 powders, and the volume of the solvent, including alcohol and Si-Al gel, was kept the same. The thickness of the three samples was also kept the same by controlling the same speed and time of the spin-coating process. For the 1% film, the integrated luminous transmittance (Tlum) was calculated to be 44.4%, and the solar modulation (ΔTsol) was 10.4%. With the higher concentration of the dispersion, the solar modulating ability was improved alongside the degradation of the Tlum (2% sample, Tlum ~ 37.3%, ΔTsol ~ 11.8%, and 3% sample, Tlum ~ 13.8%, ΔTsol ~ 13.6%). The thermochromic performances of the hydrothermal samples of different durations are exhibited in Figure 8b. All samples were made with the same concentration of VO2 dispersion, as well as the same film thickness. With the duration increased to 72 h and 168 h, the blueshift of the crosspoint between the low and high temperature can be observed in the transmittance spectra, as shown in the circles of Figure 8b. This is caused by the different optical scattering properties between the room temperature VO2(M) and high-temperature phase VO2(R). The higher transmittance in the visible wavelength range is due to the less light scattered in the high-temperature VO2(R). With the longer hydrothermal duration, crosspoints appear for the 72 h sample with particle size at 75.9 nm and the 168 h sample with 86.1 nm. This phenomenon deteriorates the thermochromic performance of the longer duration sample (72 h sample, Tlum ~ 31.7 ΔTsol ~ 11.0%, and 168 h sample, Tlum ~ 35.7%, ΔTsol ~ 5.7%) compared with the 24 h sample at the same thickness and dispersion concentration (Tlum ~ 37.3%, ΔTsol ~ 11.8%).

Conclusions
In this paper, pure VO2(M) NPs with uniform morphology were prepared via onestep hydrothermal synthesis without doping or annealing treatment. The raw material of V2O5 was gradually reduced to microbelt V6O13 and then to desired VO2 NPs, and eventually to larger and irregular morphology V6O11 and V3O5 with the increment of the N2H4·H2O addition. Furthermore, hydrothermal duration and precursor concentration have been studied carefully to obtain the pure and high crystallinity NPs. The obtained VO2 powders were dispersed and spin-coated onto glass substrates to prepare VO2 films, which exhibited comparable thermochromic performance. This provides an alternative way to understand the hydrothermal process mechanism for synthesizing the multi-phase and multi-valence vanadium oxides. The proposed hydrothermal method could be extended for various vanadium oxides synthesis in applications beyond smart window coatings.