- Email: [email protected]
Extension of α-AgI stabilization range in AgI–Ag2O–MxOy systems by mechanosynthesis processing
Journal of Non-Crystalline Solids 408 (2015) 66–70
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol
Extension of α-AgI stabilization range in AgI–Ag2O–MxOy systems by mechanosynthesis processing P. Grabowski a,⁎, J.L. Nowinski a, M. Holdynski b, J.E. Garbarczyk a, M. Wasiucionek a a b
Faculty of Physics, Warsaw University of Technology, Koszykowa 75, 00-662 Warsaw, Poland Institute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland
a r t i c l e
i n f o
Article history: Received 18 August 2014 Received in revised form 13 October 2014 Accepted 19 October 2014 Available online xxxx Keywords: α-AgI; Stabilization range; Mechanosynthesis; Glasses; Grain size
a b s t r a c t The ultra-fast quenching method and mechanosynthesis processing were employed to prepare the AgI–Ag2O– MxOy glasses (where MxOy = B2O3, WO3, MoO3, V2O5) containing AgI crystalline inclusions. After heating in elevated temperatures the α-AgI superionic phase was formed: regarding the composition and preparation method its stabilization to lower temperatures was investigated. X-ray diffraction, differential scanning calorimetry and scanning electron microscope methods were used. The study showed that employment of both methods outrivals the solely applied ultra-fast quenching or mechanosynthesis methods, giving the final product with a wider α-AgI stability range. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The alpha silver iodide phase is one of the best superionic conductors, therefore it has been a subject to various studies. Unfortunately the structure of the material is stable above 147 °C at normal thermodynamic conditions [1,2]. There were many attempts to stabilize the phase to lower temperatures. Tatsumisago et al. in the seminal paper [3–7] reported the formation of the α-AgI grains to be stable at room temperature, embedded in rapidly quenched superionic glasses. Guo et al. crystallized 50 nm thick AgI nanoplates, for which α-AgI to β/γ-AgI phase transition temperature Tph occurred at 95 °C [8]. The work of Liang et al. describes AgI nanowires grown electrochemically in a porous alumina template: recorded Tph for such prepared materials occurred at 80 °C [9]. However, the most significant decrease of Tph down to nearly room temperatures was reported by Makiura et al. for size-controlled AgI grains coated with poly-N-vinyl-2-pyrrolidone (PVP) [10]. In our previous paper [11] we showed, that there was a possibility to stabilize to lower temperatures the α-AgI phase by means of mechanosynthesis method (MS). Milling a mixture of AgI, Ag2O and B2O3 powders resulted in the formation of a material in which the αAgI phase was stable at some temperature Tph lower than 147 °C. The temperature difference ΔT = 147 °C − Tph was directly related to the milling parameters. Investigations indicated that ΔT was, in fact, a function of a size of AgI grains constituting the material, following the general relation: the smaller the grain the lower the temperature of the ⁎ Corresponding author. E-mail address: [email protected] (P. Grabowski).
http://dx.doi.org/10.1016/j.jnoncrysol.2014.10.017 0022-3093/© 2014 Elsevier B.V. All rights reserved.
α-AgI → β-AgI phase transition. The lowest recorded temperature, 87 °C, was achieved for high energy milling at a rotation speed of 1000 rpm, the maximal available for Fritsch Pulverisette Premium Line mill. Thus, in our efforts to decrease transition temperature in the way of milling we found a technical obstacle. To overcome the problem, we introduced an idea to assist the milling with another process in order to enhance the effectiveness of granulation of the AgI grains. The attention was turned to the ultra-fast quenching method (UFQ) [3,4,12]. We were aware that this process itself, when applied to the glasses with a high silver iodide contents, might lead to the formation of the α-AgI phase stable even at room temperature. But the essence of our goal was to demonstrate that such assisted milling is possible and workable. Then combination of UFQ and subsequent MS methods (UFQ + MS) was applied to the mixture of AgI, Ag2O and B2O3 compounds. High temperature X-ray based investigation revealed that the AgI phase present in a final material exhibits Tph at about 70 °C. In this paper we study in detail, the impact of the UFQ + MS method on Tph of AgI–Ag2O–MxOy (where MxOy is a glass former oxide) including additional MoO3, WO3 and V2O5 glass forming oxides. Differential scanning calorimetry (DSC) method supported by scanning electron microscopy (SEM) analysis was principally used. 2. Experimental The AgI, AgNO3 and MxOy (where MxOy = B2O3, WO3, MoO3, V2O5) compounds of 7 g total mass were taken with appropriate molar proportion: 78AgI–22(Ag2O·MxOy). The ratio of the glass matrix regarded the glass former and was fixed at 3:1, 2:1 and 1:1 for B2O3, MoO3 and
P. Grabowski et al. / Journal of Non-Crystalline Solids 408 (2015) 66–70
WO3 or V2O5 respectively [3,13,14]. Grounded mixtures were loaded into a crucible and then located in a vertical furnace at 350 °C. The reagents melted releasing some vapor byproducts. Thereafter, when the process of decomposition had ended, the temperature was raised up to 800 °C and the melt was annealed for 20 min. Then, the melt was rapidly quenched between stainless steel twin-rollers rotating at the speed of 900 rpm. The flake-like samples were formed. Mechanosynthesis process was performed employing Fritsch Pulverisette Premium Line planetary ball mill. The obtained flakes, 0.75 g mass for each material, were located inside tungsten carbide vials filled with tungsten carbide balls, 1 mm in diameter each. Ballto-powder ratio of 15:1 was kept fixed for all MS processes. The mill operated at a constant 1000 rpm rotation speed for 0.5 h or 1 h. After processing a powder product was formed. The obtained materials were investigated by X-ray diffraction (XRD), DSC and SEM methods. Phillips X'Pert Pro diffractometer with filtered Cu Kα radiation set in a Bragg–Brentano configuration was used for the XRD measurements. High temperature XRD was carried out with the assistance of Anton Paar 1200 X-ray oven. The average size of silver iodide grains was estimated by Scherrer method [15]. The DSC investigations were performed with the help of a TA Instruments Q200 microcalorimeter. The samples were heated and cooled at a rate of 10 °C/min within 20–170 °C. SEM images were recorded by Nova NanoSEM D9983 scanning electron microscope. The standard deviation value was used as a measurement of uncertainty in the grain size estimations.
3. Results 3.1. XRD
67
The results of the high temperature XRD investigations are visualized in Fig. 2, where the example of the B2O3 material is shown. Starting from the bottom of Fig. 2, the recorded X-ray patterns are ordered with increasing temperature up to 180 °C and then with decreasing temperature down to 40 °C. The positions of the strongest β/γ-AgI lines are marked by the full circles, whereas the full triangles point the diffraction lines ascribed to α-AgI. One can notice, that during heating, in the 85– 120 °C range, some β/γ-AgI peaks become stronger. Above 130 °C, the peaks characteristic for the α-AgI phase are visible on the pattern. They disappear below 65 °C.
3.2. DSC Thermal properties of the studied materials were determined by DSC method. Results of the investigations were divided into two subgroups in respect to α-AgI occurrence in UFQ materials: a) B2O3, MoO3 based materials containing α-AgI and b) materials formed with WO3, V2O5 comprising only low temperature phases — β/γ-AgI. The results of DSC investigations for selected materials from each group are presented in Fig. 4.
3.2.1. B2O3 and MoO3 On a DSC trace recorded for the heating run for the as-prepared UFQ B2O3 or MoO3 based materials (solid lines in Fig. 4a) and b)), the following processes are visible: a weak endothermic process in 50–70 °C range, a broad, weak exothermic reaction occurring in the range 70– 120 °C and a strong endothermic one peaking around 150 °C. During cooling, for the MoO3 material only a single exothermic peak at 129 °C was detected; for the B2O3 material the exothermic events were split
The XRD investigations revealed, for the B2O3 and MoO3 based materials obtained by the UFQ method, diffraction peaks characteristic for β/ γ- and α-AgI phases. In the case of the WO3 and V2O5 based materials, only the lines ascribed to β/γ-AgI were detected. For all the materials prepared by UFQ + MS method the X-ray patterns show only lines attributed to the β/γ-AgI phase: there were no lines characteristic for the α-AgI even for the B2O3 and MoO3 materials, which had exhibited these peaks previously, i.e. after the UFQ processing. Fig. 1 shows the exemplary X-ray patterns of the UFQ B2O3 and WO3 materials — before and after MS processing.
Fig. 1. X-ray powder patterns of the B2O3 based material prepared by a) UFQ; b) UFQ + MS; WO3 based material prepared by c) UFQ and d) UFQ + MS.
Fig. 2. High temperature X-ray powder patterns of the B2O3 based material formed by means of UFQ + MS method.
68
P. Grabowski et al. / Journal of Non-Crystalline Solids 408 (2015) 66–70
Fig. 3. SEM images of the B2O3 based material prepared by a) UFQ; b) UFQ + MS; WO3 based material prepared by c) UFQ and d) UFQ + MS.
into two overlapped peaks, the stronger one with maximum at 93 °C (Fig. 4a)). Heating DSC traces for UFQ + MS B2O3 or MoO3 materials is related to the DSC curves of materials prepared by UFQ method (dashed lines in Fig. 4a) and b)). However, exothermic processes in the range 50–140 °C occurred less intense for MoO3 based sample. For B2O3 material, the endothermic reaction near 150 °C was represented by broad two overlapped peaks. On cooling runs, for the MoO3 material a single exothermic peak occurs at 126 °C and for the B2O3 two overlapped peaks at 65 °C and 71 °C are visible. A sequence of eleven heating and cooling runs was applied for the B2O3 based material. The obtained DSC traces are collected in Fig. 5a). On heating runs, the temperature of the maximum of the endothermic peak slightly changes with a number of the cycle from 149 °C to 151 °C. For cooling runs, the changes of the exothermic peak position are more visible: the temperature of the peak maximum linearly increases in 15 °C range with the number of the run, increasing from 71 °C to 86 °C (Fig. 5b)). The maximum temperature difference ΔT between the endothermic events for heating and the exothermic events observed for cooling was estimated to be 75 °C.
Fig. 4. DSC traces of the B2O3 (a and b)) and WO3 (c and d)) based materials after UFQ (dashed lines) and UFQ + MS (solid lines) heating/cooling cycles. The inset graph presents marked area of the signal presented on b) in larger scale.
3.2.2. WO3 and V2O5 In the case of the UFQ WO3 based materials (solid lines in Fig. 4c) and d)), during heating, a weak endothermic event occurred in the 50–70 °C temperature range and some broad exothermic process was detected in the range 70–120 °C. They were accompanied by a small exothermic peak at around 130 °C and a strong endothermic peak around 150 °C. For the UFQ V2O5 based material the positions of the thermal events were as follows: a weak endothermic in the 40–60 °C temperature range, a broad exothermic at 85–105 °C, an exothermic peak at 115 °C and a strong endothermic at 150 °C. During cooling, for both materials, exothermic peaks were detected respectively at 113 °C for WO3 and 126 °C for V2O5. While that for the WO3 based material, some weak intensity shoulder was exhibited (Fig. 4c)). The DSC trace recorded for the UFQ + MS WO3 material (dashed lines in Fig. 4c) and d)) showed, during the first run, an endothermic event in the 50–80 °C range, some exothermic process in the 85–
P. Grabowski et al. / Journal of Non-Crystalline Solids 408 (2015) 66–70
Fig. 5. a) DSC scans of the B2O3 material prepared by means of UFQ + MS. The upper curves are for cooling, the lower traces represent heating processes. b) Phase transition temperature on cooling process as a function of the cycle (heating/cooling) number. The line is drawn as a guide for eye. c) SEM image of B2O3 material after UFQ + MS and 11 heating/cooling cycles.
130 °C and a strong endothermic peak at 150 °C. The similar sequence of the thermal events for the UFQ + MS V2O5 material was observed during the heating, although the positions of the peaks were slightly different: an endothermic one in the 50–70 °C range, an exothermic peak in the 75–125 °C range and a strong endothermic at 149 °C. For cooling, only exothermic peaks were visible on the DSC traces recorded for the both materials: at 80 °C for WO3 (preceded by weaker one) material and at 118 °C for V2O5 based material. 3.3. SEM Fig. 3 presents the SEM images of the WO3 and B2O3 based materials prepared by the UFQ (Fig. 3a) and c)) and UFQ + MS methods (Fig. 3b) and d)). For the as-prepared UFQ B2O3 material, one can notice oval objects. Inside some of them, small grains are visible. The estimated mean size of the oval objects is equal to 182(10) nm. The image of the UFQ WO3 material exhibits also some oval shape objects but there is no detectable trace of the small grains in it. In this case, the estimated mean of the size of the oval objects is equal to 197(10) nm. After milling the size of the grains visible on the relevant images is smaller, as the mean values indicate, respectively 35(10) nm for the B2O3 and 39(10) nm for the WO3 materials. The UFQ + MS B2O3 based material was additionally subjected to eleven heating/cooling cycles in the 20–170 °C temperature range. Fig. 5c) shows the image of the material after thermal cycling. The shape of the visible grains is similar to that observed for the as-received UFQ + MS B2O3, however the mean value equal to 46(10) nm is higher. 4. Discussion The results evidently indicate that the milling destroys the α-AgI phase existing in the material. In result, in expense, the β-AgI phase is formed. However, if one compares the β-AgI diffraction lines (e.g. 23.82° or 39.41°) for the UFQ and UFQ + MS materials, a noticeable
69
widening for the latter is visible. This indicates that in result of the mechanosynthesis processing AgI grains are formed smaller in size. The Scherrer calculation confirms the formation of the AgI grains smaller for UFQ + MS materials in comparison with the grain size estimated for the UFQ processed only material. For the samples with B2O3, the estimated grain size decreased from 35(5) nm to 20(3) nm, while for material containing WO3 from 34(5) nm to 27(4) nm. The conclusions withdrawn from the X-ray study are supported by the SEM investigations visualized in Fig. 3. As it was mentioned earlier, the mean values of a grain size for the B2O3 and WO3 materials (UFQ + MS) estimated from the SEM images were equal respectively to 35(10) nm and 39(10) nm. Thus, the results demonstrate a nice agreement of the values of a grain size determined by both methods. Also the SEM study shows the evident decrease of a size of the grain embedded in the matrix of a material subjected to milling. The low temperature DSC endothermic processes are characteristic for glass transitions of the AgI–Ag2O–MxOy superionic glasses prepared in ternary and also in pseudobinary systems e.g. [16,17]. At a first look, the observed endothermic 50–70 °C range peak could not be attributed to such glass transition, because, as the DSC results suggested, it was not present on successive heating run. However, detailed investigations showed, that when the sample was heated only up to about 90 °C and subsequently cooled down, the peak was observed again in the DSC curve: indicating that it visualizes a reversible process. Therefore, this evidence voted for assigning the endothermic 50–70 °C peak to the glass transition of the amorphous constituent of the UFQ or UFQ + MS materials. The presence of the two exothermic peaks detected at 70–120 °C and 130 °C indicates some processes, in which thermal energy was released, e.g. crystallization. The high temperature XRD investigations support such interpretation: the 70–120 °C peak can be related to crystallization of β-AgI and whereas the one appearing at 130 °C to the αAgI precipitation (Fig. 2) — other works on UFQ materials with high AgI content reported temperatures above 90 °C [18], or in 110–130 °C range [19]. Regarding our B2O3 and MoO3 UFQ materials, the obtained XRD and DSC results suggest that once the α-AgI phase is formed directly in the material during ultra-fast quenching additional thermal processing, like heating, does not produce an extra amount of the α-AgI phase. Then, consequently, the exothermic 130 °C peak is absent, as it is observed for the B2O3 or MoO3 ultra-fast quenched materials. In opposition, when the α-AgI phase is not formed after the UFQ, the further heating could prompt production of the α-AgI precipitates as the exothermic 130 °C peak evidences. An impact of the thermal processing on the ultra-fast quenched glasses was studied in detail and discussed in the papers of a Japanese group from Osaka Prefecture University [14]. They suggested the existence of some embryos embedded in the ultra-fast quenched glass matrix being seeds of precipitation for alpha silver iodide phase. As our study shows, among various observed DSC peaks, only two of them regularly appear for all studied materials: regardless of a chemical composition, UFQ or UFQ + MS method of preparation or even number of a DSC run applied to the investigated sample. The first one, endothermic peak at 150 °C is present during heating runs and its position essentially does not change. The second one is the exothermic peak observed on cooling runs. In contrary, its position depends strongly on the chemical composition, method of preparation and a number of a DSC cycles. In some cases, some splitting of these peaks was observed. Both peaks were attributed to the AgI phase transition: high temperature XRD investigations showed that the 150 °C peak was related to β-AgI → αAgI transformation, whereas the exothermic one to the α-AgI → β-AgI transition. It is worthy to note that we did not record any XRD evidences suggesting that the splitting of these DSC peaks could be the result of some crystallization or other processes different in their nature from silver iodide phase transformation. The result agrees with well-established literature reports quoting 147 °C as the temperature of β-AgI → α-AgI structure transformation
70
P. Grabowski et al. / Journal of Non-Crystalline Solids 408 (2015) 66–70
Fig. 6. Temperature of the α-AgI → β-AgI phase transition during cooling as a function of a silver iodide grain size for the materials after UFQ + MS. The inset shows the influence of milling time on the α-AgI → β-AgI phase transition temperature. Milling time “0 h” stands for materials after the UFQ only. All lines are drawn as a guide for eye. Dashed curve is taken from [10].
[1,2] and with those describing lowering of the temperature of reverse phase transition, which was observed for some polycrystalline AgI or heterogeneous materials containing silver iodide phase but obtained utterly different methods [8–10]. Thus, determined positions of the exothermic peak marks the lowest and limit temperature at which the αAgI is stable when cooling from high temperature. Once it is formed by annealing of the material above 147 °C, then the alpha phase is stable down to the low limit temperature. However, prolonged thermal processing like multiple heating/cooling cycling causes the Tph to increase, in result, narrowing the range of stability of the α-AgI phase. This effect, undesirable for stability of the alpha AgI phase, one could relate to some crystallization processes causing the AgI grains to agglomerate and grow. Fig. 6 presents the dependence between a grain size determined by Scherrer method from the XRD patterns and a temperature Tph found from the DSC traces. The smallest in size β-AgI grains were formed for the B2O3 and WO3 materials after UFQ + MS, which simultaneously demonstrated the lowest Tph temperature values, i.e. the widest α-AgI stability ranges. The inset shows the dependence of milling time on the α-AgI → β-AgI transition temperature. The result indicate that the most significant influence is during the first 0.5 h lasting milling, while prolonged one practically has no mining. The dashed line included in Fig. 6 is taken from the paper of Rie Makiura et al. [10]. Crystallized AgI nanoparticles were coated with poly-N-vinyl-2-pyrrolidone (PVP) in conditions allowing controlling the particle size. Our results, however obtained by utterly different methods and for various materials, very nicely correspond with those from [10], supporting an idea that the particle size of the AgI grain is a crucial parameter for the Tph value. Regarding the UFQ + MS method, we are also demonstrating the role of the matrix in which the grains are embedded. Depending on its chemical composition, the matrix could help or impede formation of the small AgI grains. 5. Conclusions The combination of ultra-fast quenching and subsequent mechanosynthesis methods (UFQ + MS) applied to the AgI–Ag2O–MxOy systems
with high AgI contents, forms materials in which crystalline AgI component allows to stabilize the α-AgI to temperatures far below 147 °C. The exhibited α-AgI stabilization range for materials containing various metal oxides (MxOY = B2O3, WO3, MoO3, V2O5) was found to be wider than for relevant products prepared solely by ultra-fast quenching or mechanosynthesis methods. As presented, even short lasting milling, ca. half an hour, is sufficient enough to extend the stabilization range ΔT. The lowest Tph value 65 °C was found for the B2O3 based material. However, long lasting annealing at elevated temperature or multiple heating/ cooling resulted in decreasing of the α-AgI stability range. Acknowledgments This work is partially sponsored by the European Social Fund in Human Capital Programme (CAS/22/POKL). References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
C. Tubandt, E. Lorenz, Z. Phys. Chem. 87 (1914) 513–543. K. Funke, Prog. Solid State Chem. 11 (1967) 345–402. M. Tatsumisago, Y. Shinkuma, T. Minami, Nature 354 (1991) 217–218. M. Tatsumisago, Y. Shinkuma, T. Saito, T. Minami, Solid State Ionics 50 (1992) 273–279. M. Tatsumisago, N. Torata, T. Saito, T. Minami, J. Non-Cryst. Solids 196 (1996) 193. M. Tatsumisago, T. Saito, T. Minami, Thermochim. Acta 280 (281) (1996) 333. M. Tatsumisago, N. Itakura, T. Minami, J. Non-Cryst. Solids 232–234 (1998) 267. Y. Guo, J. Lee, J. Maier, Solid State Ionics 177 (2006) 2467–2471. C. Liang, K. Terabe, T. Hasegawa, M. Aono, N. Iyi, J. Appl. Phys. 102 (2007) 124308. R. Makiura, T. Yonemura, T. Yamada, M. Yamauchi, R. Ikeda, H. Kitagawa, K. Kato, M. Takata, Nat. Mater. 8 (2009) 476–480. P. Grabowski, J.L. Nowiński, J.E. Garbarczyk, M. Wasiucionek, Solid State Ionics 251 (2013) 55–58. H.S. Chen, C.E. Miller, Rev. Sci. Instrum. 41 (1970) 1237. A. Dalvi, A.M. Awasthi, S. Bharadwaj, K. Shahi, J. Phys. Chem. Solids 66 (2005) 783–792. M. Tatsumisago, K. Okuda, N. Itakura, T. Minami, Solid State Ionics 121 (1999) 193–200. A. Patterson, Phys. Rev. 56 (10) (1939) 978–982. T. Minami, K. Imazawa, M. Tanaka, J. Non-Cryst. Solids 42 (1980) 469-476. J. Kawamura, M. Shimoiji, J. Non-Cryst. Solids 88 (1986) 281–289. M. Tatsumisago, T. Saito, T. Minami, Phys. Chem. Glasses 42 (43) (2001) 215–219. N. Itakura, M. Tatsumisago, T. Minami, J. Am. Ceram. Soc. 80 (12) (1997) 3209–3212.
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol
Extension of α-AgI stabilization range in AgI–Ag2O–MxOy systems by mechanosynthesis processing P. Grabowski a,⁎, J.L. Nowinski a, M. Holdynski b, J.E. Garbarczyk a, M. Wasiucionek a a b
Faculty of Physics, Warsaw University of Technology, Koszykowa 75, 00-662 Warsaw, Poland Institute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland
a r t i c l e
i n f o
Article history: Received 18 August 2014 Received in revised form 13 October 2014 Accepted 19 October 2014 Available online xxxx Keywords: α-AgI; Stabilization range; Mechanosynthesis; Glasses; Grain size
a b s t r a c t The ultra-fast quenching method and mechanosynthesis processing were employed to prepare the AgI–Ag2O– MxOy glasses (where MxOy = B2O3, WO3, MoO3, V2O5) containing AgI crystalline inclusions. After heating in elevated temperatures the α-AgI superionic phase was formed: regarding the composition and preparation method its stabilization to lower temperatures was investigated. X-ray diffraction, differential scanning calorimetry and scanning electron microscope methods were used. The study showed that employment of both methods outrivals the solely applied ultra-fast quenching or mechanosynthesis methods, giving the final product with a wider α-AgI stability range. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The alpha silver iodide phase is one of the best superionic conductors, therefore it has been a subject to various studies. Unfortunately the structure of the material is stable above 147 °C at normal thermodynamic conditions [1,2]. There were many attempts to stabilize the phase to lower temperatures. Tatsumisago et al. in the seminal paper [3–7] reported the formation of the α-AgI grains to be stable at room temperature, embedded in rapidly quenched superionic glasses. Guo et al. crystallized 50 nm thick AgI nanoplates, for which α-AgI to β/γ-AgI phase transition temperature Tph occurred at 95 °C [8]. The work of Liang et al. describes AgI nanowires grown electrochemically in a porous alumina template: recorded Tph for such prepared materials occurred at 80 °C [9]. However, the most significant decrease of Tph down to nearly room temperatures was reported by Makiura et al. for size-controlled AgI grains coated with poly-N-vinyl-2-pyrrolidone (PVP) [10]. In our previous paper [11] we showed, that there was a possibility to stabilize to lower temperatures the α-AgI phase by means of mechanosynthesis method (MS). Milling a mixture of AgI, Ag2O and B2O3 powders resulted in the formation of a material in which the αAgI phase was stable at some temperature Tph lower than 147 °C. The temperature difference ΔT = 147 °C − Tph was directly related to the milling parameters. Investigations indicated that ΔT was, in fact, a function of a size of AgI grains constituting the material, following the general relation: the smaller the grain the lower the temperature of the ⁎ Corresponding author. E-mail address: [email protected] (P. Grabowski).
http://dx.doi.org/10.1016/j.jnoncrysol.2014.10.017 0022-3093/© 2014 Elsevier B.V. All rights reserved.
α-AgI → β-AgI phase transition. The lowest recorded temperature, 87 °C, was achieved for high energy milling at a rotation speed of 1000 rpm, the maximal available for Fritsch Pulverisette Premium Line mill. Thus, in our efforts to decrease transition temperature in the way of milling we found a technical obstacle. To overcome the problem, we introduced an idea to assist the milling with another process in order to enhance the effectiveness of granulation of the AgI grains. The attention was turned to the ultra-fast quenching method (UFQ) [3,4,12]. We were aware that this process itself, when applied to the glasses with a high silver iodide contents, might lead to the formation of the α-AgI phase stable even at room temperature. But the essence of our goal was to demonstrate that such assisted milling is possible and workable. Then combination of UFQ and subsequent MS methods (UFQ + MS) was applied to the mixture of AgI, Ag2O and B2O3 compounds. High temperature X-ray based investigation revealed that the AgI phase present in a final material exhibits Tph at about 70 °C. In this paper we study in detail, the impact of the UFQ + MS method on Tph of AgI–Ag2O–MxOy (where MxOy is a glass former oxide) including additional MoO3, WO3 and V2O5 glass forming oxides. Differential scanning calorimetry (DSC) method supported by scanning electron microscopy (SEM) analysis was principally used. 2. Experimental The AgI, AgNO3 and MxOy (where MxOy = B2O3, WO3, MoO3, V2O5) compounds of 7 g total mass were taken with appropriate molar proportion: 78AgI–22(Ag2O·MxOy). The ratio of the glass matrix regarded the glass former and was fixed at 3:1, 2:1 and 1:1 for B2O3, MoO3 and
P. Grabowski et al. / Journal of Non-Crystalline Solids 408 (2015) 66–70
WO3 or V2O5 respectively [3,13,14]. Grounded mixtures were loaded into a crucible and then located in a vertical furnace at 350 °C. The reagents melted releasing some vapor byproducts. Thereafter, when the process of decomposition had ended, the temperature was raised up to 800 °C and the melt was annealed for 20 min. Then, the melt was rapidly quenched between stainless steel twin-rollers rotating at the speed of 900 rpm. The flake-like samples were formed. Mechanosynthesis process was performed employing Fritsch Pulverisette Premium Line planetary ball mill. The obtained flakes, 0.75 g mass for each material, were located inside tungsten carbide vials filled with tungsten carbide balls, 1 mm in diameter each. Ballto-powder ratio of 15:1 was kept fixed for all MS processes. The mill operated at a constant 1000 rpm rotation speed for 0.5 h or 1 h. After processing a powder product was formed. The obtained materials were investigated by X-ray diffraction (XRD), DSC and SEM methods. Phillips X'Pert Pro diffractometer with filtered Cu Kα radiation set in a Bragg–Brentano configuration was used for the XRD measurements. High temperature XRD was carried out with the assistance of Anton Paar 1200 X-ray oven. The average size of silver iodide grains was estimated by Scherrer method [15]. The DSC investigations were performed with the help of a TA Instruments Q200 microcalorimeter. The samples were heated and cooled at a rate of 10 °C/min within 20–170 °C. SEM images were recorded by Nova NanoSEM D9983 scanning electron microscope. The standard deviation value was used as a measurement of uncertainty in the grain size estimations.
3. Results 3.1. XRD
67
The results of the high temperature XRD investigations are visualized in Fig. 2, where the example of the B2O3 material is shown. Starting from the bottom of Fig. 2, the recorded X-ray patterns are ordered with increasing temperature up to 180 °C and then with decreasing temperature down to 40 °C. The positions of the strongest β/γ-AgI lines are marked by the full circles, whereas the full triangles point the diffraction lines ascribed to α-AgI. One can notice, that during heating, in the 85– 120 °C range, some β/γ-AgI peaks become stronger. Above 130 °C, the peaks characteristic for the α-AgI phase are visible on the pattern. They disappear below 65 °C.
3.2. DSC Thermal properties of the studied materials were determined by DSC method. Results of the investigations were divided into two subgroups in respect to α-AgI occurrence in UFQ materials: a) B2O3, MoO3 based materials containing α-AgI and b) materials formed with WO3, V2O5 comprising only low temperature phases — β/γ-AgI. The results of DSC investigations for selected materials from each group are presented in Fig. 4.
3.2.1. B2O3 and MoO3 On a DSC trace recorded for the heating run for the as-prepared UFQ B2O3 or MoO3 based materials (solid lines in Fig. 4a) and b)), the following processes are visible: a weak endothermic process in 50–70 °C range, a broad, weak exothermic reaction occurring in the range 70– 120 °C and a strong endothermic one peaking around 150 °C. During cooling, for the MoO3 material only a single exothermic peak at 129 °C was detected; for the B2O3 material the exothermic events were split
The XRD investigations revealed, for the B2O3 and MoO3 based materials obtained by the UFQ method, diffraction peaks characteristic for β/ γ- and α-AgI phases. In the case of the WO3 and V2O5 based materials, only the lines ascribed to β/γ-AgI were detected. For all the materials prepared by UFQ + MS method the X-ray patterns show only lines attributed to the β/γ-AgI phase: there were no lines characteristic for the α-AgI even for the B2O3 and MoO3 materials, which had exhibited these peaks previously, i.e. after the UFQ processing. Fig. 1 shows the exemplary X-ray patterns of the UFQ B2O3 and WO3 materials — before and after MS processing.
Fig. 1. X-ray powder patterns of the B2O3 based material prepared by a) UFQ; b) UFQ + MS; WO3 based material prepared by c) UFQ and d) UFQ + MS.
Fig. 2. High temperature X-ray powder patterns of the B2O3 based material formed by means of UFQ + MS method.
68
P. Grabowski et al. / Journal of Non-Crystalline Solids 408 (2015) 66–70
Fig. 3. SEM images of the B2O3 based material prepared by a) UFQ; b) UFQ + MS; WO3 based material prepared by c) UFQ and d) UFQ + MS.
into two overlapped peaks, the stronger one with maximum at 93 °C (Fig. 4a)). Heating DSC traces for UFQ + MS B2O3 or MoO3 materials is related to the DSC curves of materials prepared by UFQ method (dashed lines in Fig. 4a) and b)). However, exothermic processes in the range 50–140 °C occurred less intense for MoO3 based sample. For B2O3 material, the endothermic reaction near 150 °C was represented by broad two overlapped peaks. On cooling runs, for the MoO3 material a single exothermic peak occurs at 126 °C and for the B2O3 two overlapped peaks at 65 °C and 71 °C are visible. A sequence of eleven heating and cooling runs was applied for the B2O3 based material. The obtained DSC traces are collected in Fig. 5a). On heating runs, the temperature of the maximum of the endothermic peak slightly changes with a number of the cycle from 149 °C to 151 °C. For cooling runs, the changes of the exothermic peak position are more visible: the temperature of the peak maximum linearly increases in 15 °C range with the number of the run, increasing from 71 °C to 86 °C (Fig. 5b)). The maximum temperature difference ΔT between the endothermic events for heating and the exothermic events observed for cooling was estimated to be 75 °C.
Fig. 4. DSC traces of the B2O3 (a and b)) and WO3 (c and d)) based materials after UFQ (dashed lines) and UFQ + MS (solid lines) heating/cooling cycles. The inset graph presents marked area of the signal presented on b) in larger scale.
3.2.2. WO3 and V2O5 In the case of the UFQ WO3 based materials (solid lines in Fig. 4c) and d)), during heating, a weak endothermic event occurred in the 50–70 °C temperature range and some broad exothermic process was detected in the range 70–120 °C. They were accompanied by a small exothermic peak at around 130 °C and a strong endothermic peak around 150 °C. For the UFQ V2O5 based material the positions of the thermal events were as follows: a weak endothermic in the 40–60 °C temperature range, a broad exothermic at 85–105 °C, an exothermic peak at 115 °C and a strong endothermic at 150 °C. During cooling, for both materials, exothermic peaks were detected respectively at 113 °C for WO3 and 126 °C for V2O5. While that for the WO3 based material, some weak intensity shoulder was exhibited (Fig. 4c)). The DSC trace recorded for the UFQ + MS WO3 material (dashed lines in Fig. 4c) and d)) showed, during the first run, an endothermic event in the 50–80 °C range, some exothermic process in the 85–
P. Grabowski et al. / Journal of Non-Crystalline Solids 408 (2015) 66–70
Fig. 5. a) DSC scans of the B2O3 material prepared by means of UFQ + MS. The upper curves are for cooling, the lower traces represent heating processes. b) Phase transition temperature on cooling process as a function of the cycle (heating/cooling) number. The line is drawn as a guide for eye. c) SEM image of B2O3 material after UFQ + MS and 11 heating/cooling cycles.
130 °C and a strong endothermic peak at 150 °C. The similar sequence of the thermal events for the UFQ + MS V2O5 material was observed during the heating, although the positions of the peaks were slightly different: an endothermic one in the 50–70 °C range, an exothermic peak in the 75–125 °C range and a strong endothermic at 149 °C. For cooling, only exothermic peaks were visible on the DSC traces recorded for the both materials: at 80 °C for WO3 (preceded by weaker one) material and at 118 °C for V2O5 based material. 3.3. SEM Fig. 3 presents the SEM images of the WO3 and B2O3 based materials prepared by the UFQ (Fig. 3a) and c)) and UFQ + MS methods (Fig. 3b) and d)). For the as-prepared UFQ B2O3 material, one can notice oval objects. Inside some of them, small grains are visible. The estimated mean size of the oval objects is equal to 182(10) nm. The image of the UFQ WO3 material exhibits also some oval shape objects but there is no detectable trace of the small grains in it. In this case, the estimated mean of the size of the oval objects is equal to 197(10) nm. After milling the size of the grains visible on the relevant images is smaller, as the mean values indicate, respectively 35(10) nm for the B2O3 and 39(10) nm for the WO3 materials. The UFQ + MS B2O3 based material was additionally subjected to eleven heating/cooling cycles in the 20–170 °C temperature range. Fig. 5c) shows the image of the material after thermal cycling. The shape of the visible grains is similar to that observed for the as-received UFQ + MS B2O3, however the mean value equal to 46(10) nm is higher. 4. Discussion The results evidently indicate that the milling destroys the α-AgI phase existing in the material. In result, in expense, the β-AgI phase is formed. However, if one compares the β-AgI diffraction lines (e.g. 23.82° or 39.41°) for the UFQ and UFQ + MS materials, a noticeable
69
widening for the latter is visible. This indicates that in result of the mechanosynthesis processing AgI grains are formed smaller in size. The Scherrer calculation confirms the formation of the AgI grains smaller for UFQ + MS materials in comparison with the grain size estimated for the UFQ processed only material. For the samples with B2O3, the estimated grain size decreased from 35(5) nm to 20(3) nm, while for material containing WO3 from 34(5) nm to 27(4) nm. The conclusions withdrawn from the X-ray study are supported by the SEM investigations visualized in Fig. 3. As it was mentioned earlier, the mean values of a grain size for the B2O3 and WO3 materials (UFQ + MS) estimated from the SEM images were equal respectively to 35(10) nm and 39(10) nm. Thus, the results demonstrate a nice agreement of the values of a grain size determined by both methods. Also the SEM study shows the evident decrease of a size of the grain embedded in the matrix of a material subjected to milling. The low temperature DSC endothermic processes are characteristic for glass transitions of the AgI–Ag2O–MxOy superionic glasses prepared in ternary and also in pseudobinary systems e.g. [16,17]. At a first look, the observed endothermic 50–70 °C range peak could not be attributed to such glass transition, because, as the DSC results suggested, it was not present on successive heating run. However, detailed investigations showed, that when the sample was heated only up to about 90 °C and subsequently cooled down, the peak was observed again in the DSC curve: indicating that it visualizes a reversible process. Therefore, this evidence voted for assigning the endothermic 50–70 °C peak to the glass transition of the amorphous constituent of the UFQ or UFQ + MS materials. The presence of the two exothermic peaks detected at 70–120 °C and 130 °C indicates some processes, in which thermal energy was released, e.g. crystallization. The high temperature XRD investigations support such interpretation: the 70–120 °C peak can be related to crystallization of β-AgI and whereas the one appearing at 130 °C to the αAgI precipitation (Fig. 2) — other works on UFQ materials with high AgI content reported temperatures above 90 °C [18], or in 110–130 °C range [19]. Regarding our B2O3 and MoO3 UFQ materials, the obtained XRD and DSC results suggest that once the α-AgI phase is formed directly in the material during ultra-fast quenching additional thermal processing, like heating, does not produce an extra amount of the α-AgI phase. Then, consequently, the exothermic 130 °C peak is absent, as it is observed for the B2O3 or MoO3 ultra-fast quenched materials. In opposition, when the α-AgI phase is not formed after the UFQ, the further heating could prompt production of the α-AgI precipitates as the exothermic 130 °C peak evidences. An impact of the thermal processing on the ultra-fast quenched glasses was studied in detail and discussed in the papers of a Japanese group from Osaka Prefecture University [14]. They suggested the existence of some embryos embedded in the ultra-fast quenched glass matrix being seeds of precipitation for alpha silver iodide phase. As our study shows, among various observed DSC peaks, only two of them regularly appear for all studied materials: regardless of a chemical composition, UFQ or UFQ + MS method of preparation or even number of a DSC run applied to the investigated sample. The first one, endothermic peak at 150 °C is present during heating runs and its position essentially does not change. The second one is the exothermic peak observed on cooling runs. In contrary, its position depends strongly on the chemical composition, method of preparation and a number of a DSC cycles. In some cases, some splitting of these peaks was observed. Both peaks were attributed to the AgI phase transition: high temperature XRD investigations showed that the 150 °C peak was related to β-AgI → αAgI transformation, whereas the exothermic one to the α-AgI → β-AgI transition. It is worthy to note that we did not record any XRD evidences suggesting that the splitting of these DSC peaks could be the result of some crystallization or other processes different in their nature from silver iodide phase transformation. The result agrees with well-established literature reports quoting 147 °C as the temperature of β-AgI → α-AgI structure transformation
70
P. Grabowski et al. / Journal of Non-Crystalline Solids 408 (2015) 66–70
Fig. 6. Temperature of the α-AgI → β-AgI phase transition during cooling as a function of a silver iodide grain size for the materials after UFQ + MS. The inset shows the influence of milling time on the α-AgI → β-AgI phase transition temperature. Milling time “0 h” stands for materials after the UFQ only. All lines are drawn as a guide for eye. Dashed curve is taken from [10].
[1,2] and with those describing lowering of the temperature of reverse phase transition, which was observed for some polycrystalline AgI or heterogeneous materials containing silver iodide phase but obtained utterly different methods [8–10]. Thus, determined positions of the exothermic peak marks the lowest and limit temperature at which the αAgI is stable when cooling from high temperature. Once it is formed by annealing of the material above 147 °C, then the alpha phase is stable down to the low limit temperature. However, prolonged thermal processing like multiple heating/cooling cycling causes the Tph to increase, in result, narrowing the range of stability of the α-AgI phase. This effect, undesirable for stability of the alpha AgI phase, one could relate to some crystallization processes causing the AgI grains to agglomerate and grow. Fig. 6 presents the dependence between a grain size determined by Scherrer method from the XRD patterns and a temperature Tph found from the DSC traces. The smallest in size β-AgI grains were formed for the B2O3 and WO3 materials after UFQ + MS, which simultaneously demonstrated the lowest Tph temperature values, i.e. the widest α-AgI stability ranges. The inset shows the dependence of milling time on the α-AgI → β-AgI transition temperature. The result indicate that the most significant influence is during the first 0.5 h lasting milling, while prolonged one practically has no mining. The dashed line included in Fig. 6 is taken from the paper of Rie Makiura et al. [10]. Crystallized AgI nanoparticles were coated with poly-N-vinyl-2-pyrrolidone (PVP) in conditions allowing controlling the particle size. Our results, however obtained by utterly different methods and for various materials, very nicely correspond with those from [10], supporting an idea that the particle size of the AgI grain is a crucial parameter for the Tph value. Regarding the UFQ + MS method, we are also demonstrating the role of the matrix in which the grains are embedded. Depending on its chemical composition, the matrix could help or impede formation of the small AgI grains. 5. Conclusions The combination of ultra-fast quenching and subsequent mechanosynthesis methods (UFQ + MS) applied to the AgI–Ag2O–MxOy systems
with high AgI contents, forms materials in which crystalline AgI component allows to stabilize the α-AgI to temperatures far below 147 °C. The exhibited α-AgI stabilization range for materials containing various metal oxides (MxOY = B2O3, WO3, MoO3, V2O5) was found to be wider than for relevant products prepared solely by ultra-fast quenching or mechanosynthesis methods. As presented, even short lasting milling, ca. half an hour, is sufficient enough to extend the stabilization range ΔT. The lowest Tph value 65 °C was found for the B2O3 based material. However, long lasting annealing at elevated temperature or multiple heating/ cooling resulted in decreasing of the α-AgI stability range. Acknowledgments This work is partially sponsored by the European Social Fund in Human Capital Programme (CAS/22/POKL). References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
C. Tubandt, E. Lorenz, Z. Phys. Chem. 87 (1914) 513–543. K. Funke, Prog. Solid State Chem. 11 (1967) 345–402. M. Tatsumisago, Y. Shinkuma, T. Minami, Nature 354 (1991) 217–218. M. Tatsumisago, Y. Shinkuma, T. Saito, T. Minami, Solid State Ionics 50 (1992) 273–279. M. Tatsumisago, N. Torata, T. Saito, T. Minami, J. Non-Cryst. Solids 196 (1996) 193. M. Tatsumisago, T. Saito, T. Minami, Thermochim. Acta 280 (281) (1996) 333. M. Tatsumisago, N. Itakura, T. Minami, J. Non-Cryst. Solids 232–234 (1998) 267. Y. Guo, J. Lee, J. Maier, Solid State Ionics 177 (2006) 2467–2471. C. Liang, K. Terabe, T. Hasegawa, M. Aono, N. Iyi, J. Appl. Phys. 102 (2007) 124308. R. Makiura, T. Yonemura, T. Yamada, M. Yamauchi, R. Ikeda, H. Kitagawa, K. Kato, M. Takata, Nat. Mater. 8 (2009) 476–480. P. Grabowski, J.L. Nowiński, J.E. Garbarczyk, M. Wasiucionek, Solid State Ionics 251 (2013) 55–58. H.S. Chen, C.E. Miller, Rev. Sci. Instrum. 41 (1970) 1237. A. Dalvi, A.M. Awasthi, S. Bharadwaj, K. Shahi, J. Phys. Chem. Solids 66 (2005) 783–792. M. Tatsumisago, K. Okuda, N. Itakura, T. Minami, Solid State Ionics 121 (1999) 193–200. A. Patterson, Phys. Rev. 56 (10) (1939) 978–982. T. Minami, K. Imazawa, M. Tanaka, J. Non-Cryst. Solids 42 (1980) 469-476. J. Kawamura, M. Shimoiji, J. Non-Cryst. Solids 88 (1986) 281–289. M. Tatsumisago, T. Saito, T. Minami, Phys. Chem. Glasses 42 (43) (2001) 215–219. N. Itakura, M. Tatsumisago, T. Minami, J. Am. Ceram. Soc. 80 (12) (1997) 3209–3212.