Room-Temperature Mechanochemical Synthesis of ... - Springer Link

Room-Temperature Mechanochemical Synthesis of W2B5 Powders SELIM COS¸KUN and M. LU¨TFI O¨VEÇOG˘LU The mechanochemical synthesis of W2B5 powders was successfully carried out at room temperature. WO2.72, B2O3 and Mg powder blends were mixed to form batches according to the metallothermic reduction of WO2.72 and B2O3 with Mg, which were subsequently mechanically alloyed (MA) using a Spex mill at different durations to constitute W2B5 + MgO as final products. Following mechanochemical synthesis, MgO was removed from the system by leaching the powders with HCl. Microstructural and morphological characterizations of boride powders were carried out via scanning electron microscopy (SEM) and X-ray diffraction (XRD) analyses. Moreover, atomic absorption spectroscopy (AAS) and differential scanning calorimetry (DSC) experiments were carried out to monitor the purity of the powders at different stages of the process. After mechanical alloying of the powder batches comprising 50 pct stoichiometrically excess amount of B2O3 for 30 hours and leaching with a 7 M HCl solution, pure W2B5 powders with an average particle size of 226 nm and an average grain size of 55.3 nm were successfully synthesized. DOI: 10.1007/s11661-012-1551-4 The Minerals, Metals & Materials Society and ASM International 2012

I.

INTRODUCTION

TUNGSTEN borides have promising characteristics like a high melting point, high hardness values (>9 Mohs), high abrasion resistance, chemical inertness, magnetic properties, and metal-like electronic conductivity.[1] These characteristics have made tungsten borides interesting candidates for industrial applications at extreme environmental conditions, which demand abrasive, corrosion-resistant, and electrode materials.[1–8] Tungsten borides are resistant to thermal shock and are also good thermal conductors.[9] They are used in high-temperature applications such as crucibles and ingot molds for precision metallurgy.[9,10] Moreover, tungsten boride powders are used in the fabrication of various composite materials, including metal–boride alloys for fillers (borolites).[11] Currently, various kinds of tungsten borides such as W2B, WB, WB2, W2B5, WB4, and WB12 are known to exist in the W–B system.[1,5,12] Among these, W2B5 comes into prominence since it combines the high hardness (2700 kg/mm2) and modulus of elasticity (755 GPa) values with the low electrical resistivity (19 lX9cm) and density (13.1 g/cm3) values.[1,13] W2B5 phase has lately been used to improve the thermoelectric SELIM COS¸KUN, Postdoctoral Research Associate, is with the Department of Materials Science and Metallurgy, University of Cambridge, Cambridge, U.K., and also with the Department of Metallurgy and Materials Engineering, Particulate Materials Laboratories, Istanbul Technical University, ITU Ayazaga Kampusu Maslak, 34469 Maslak, Istanbul, Turkey. Contact e-mail: [email protected] M. LU¨TFI O¨VEÇOG˘LU, Professor, is with the Department of Metallurgy and Materials Engineering, Particulate Materials Laboratories, Istanbul Technical University. Manuscript submitted May 23, 2012. Article published online December 11, 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

properties of B4C ceramics[3] and to improve the electrical conductivity, wear and oxidation resistance, flexural strength, and fracture toughness of carbon composites.[6,7] Recently, W2B5 has been used to increase the wear resistance of carbon composites against steel bearings.[14] Researchers are specifically interested in problems related to tungsten borides because their production is difficult owing to the use of high-temperature equipment and, in some degree, does not meet the practical needs simultaneously.[11] Tungsten borides have been synthesized by solid-state reactions, reduction of oxides by carbon, borothermal reduction of metal oxide, molten salt electrolysis, and chemical vapor deposition.[1,11,15] Fusion processes, such as metallothermy and selfpropagating high-temperature synthesis, are also used in the production of borides by reduction of metal (including tungsten and boron) oxides with aluminum, magnesium, or calcium.[11] Needless to mention that these processes are generally carried out at elevated temperatures and require technically complex and expensive experimental setup. In addition to the above-mentioned processes, there have been several studies on different transition metal borides[16–24] produced by mechanical alloying (MA) followed by leaching to remove any contamination or impurities. MA is a simple, versatile, and economically feasible method possessing significant technical advantages. One of the greatest advantages of MA is in the synthesis of novel alloys and composites as well as alloying of normally immiscible elements, which is not possible by any other technique.[25–27] The reason is its nonequilibrium nature.[28] If chemical reactions and phase transformations take place due to the application of mechanical energy, then mechanochemical synthesis (MS) is the term applied to the process and the VOLUME 44A, APRIL 2013—1805

applications of mechanochemical synthesis include exchange reactions, reduction/oxidation reactions, decomposition of compounds, and phase transformations.[27,28] Titanium boride (TiB2) is the most investigated transition metal boride produced via mechanochemical synthesis.[16,17,19,20] Other borides produced via MS include ZrB2,[18] HfB2,[24] NbB2,[21] MoB2, Mo2B5,[22] and VB2.[23] Despite all these investigations, currently there is hardly any comprehensive study on the production of pure tungsten borides by mechanochemical synthesis. Perhaps the only exception is the study by Stubicar et al.[29] who carried out X-ray diffraction investigations on elemental W–B mixtures following high-energy ball milling. However, their resultant milled powders have always consisted of elemental W, WC, and B2O3 phases beside different tungsten boride mixtures. On the other hand, there have been some more recent encouraging studies, in which Ricceri and Matteazzi[30,31] produced elemental B and elemental W using mechanochemical synthesis from B2O3 and WO3, respectively. Keeping the above concepts in mind, the aim of the current study was to synthesize and characterize pure W2B5 powders via mechanochemical synthesis followed by leaching for the first time in the literature. To achieve that, the effects of MA duration, initial powder stoichiometry, and the molarity of the leaching solution on the purity of the boride powders were investigated. The proposed synthesis route for pure W2B5 powders is simplistic in its nature and was carried out at room temperature. Thus, this process could find potential commercial uses, having lower costs and requiring relatively simple equipment in comparison with the current industrial processes using high-temperature equipment.

II.

EXPERIMENTAL PROCEDURE

In this study, WO2.72 (Alfa Aesar, Ward Hill, MA; 99.99 pct purity), B2O3 (Alfa Aesar, 99.999 pct purity), and Mg (Alfa Aesar, +99 pct purity) powders were used. These powders were mixed to form powder batches according to the metallothermic reduction of WO2.72 and B2O3 with Mg to constitute W2B5 + MgO as final products as shown in Eq. [1]. For mechanical alloying experiments, powder batches of 10 g each were prepared and all samples prepared to synthesize W2B5 powders will be referred to as 2W5BM. 2WO2:72 ðsÞ þ 2:5B2 O3 ðsÞ þ 12:94MgðsÞ ¼ W2 B5 ðsÞ þ 12:94MgOðsÞ

½1

Some preliminary MA experiments were carried out, which revealed that MA durations less than 20 hours did not cause any significant progress in the metallothermic reduction of WO2.72 and B2O3 with Mg. For this reason, the prepared powder blends were mechanically alloyed for 20 and 30 hours in a stainless steel vial using a vibratory ball mill (8000 D mixer/mill; Spex, Metuchen, NJ) with a speed of 1200 rpm. Stainless steel balls with a diameter of 6.25 mm (1/4 in) were used as the milling media and the ball-to-powder weight ratio was 10:1. High-purity argon gas (Linde, Murray Hill, NJ, 99.999 pct purity) was chosen as the milling atmosphere to prevent the oxidation and contamination of powders during loading/unloading and milling. Otherwise, the presence of air in the vial would lead to the formation of oxides and nitrides in the powder, especially if the powders are reactive in nature. Thus, the loading and unloading of powders into and from the vials were done inside an atmosphere-controlled glove box (Plaslabs, Lansing Charter Township, MI) backfilled with highpurity Ar gas. After MA of the powders, leaching experiments were carried out to remove MgO and all other contaminations from the milling medium and impurities under continuous ultrasonication; 5 M and 7 M HCl were used as the leaching solutions. These solutions contain more HCl than the stoichiometric amount since leaching with stoichiometric excess amount of acid solution generally yields better results compared to leaching with the stoichiometric amount.[32,33] The leaching duration and solid-to-liquid ratio of the leaching solution were set to 15 minutes and 1:10, respectively. After that, the solutions were centrifuged in a Zentrifugen Rotofix 32 A device (Andreas Hettich GmbH & Co. KG, Tuttlingen, Germany) for 15 minutes with a rotation speed of 3500 rpm, diluted, and again centrifuged for another 15 minutes. Finally, the solution was vacuum filtered and washed in ethanol as well as in distilled water for several times followed by drying the powders in air at 353 K (80 C) for 12 hours. All leaching experiments were carried out at room temperature. Microstructural characterization of the leached and unleached powders were carried out in a JSM-T330 (JEOL Ltd., Tokyo, Japan) scanning electron microscope (SEM) coupled with an energy-dispersive spectrometer (EDS) and in a D8 Advance X-ray diffractometer (XRD) (Bruker AXS, Karlsruhe, Germany) using CuKa radiation (k = 0.154 nm) at the settings of 40 kV and 40 mA. TOPAS 3 (Bruker AXS) software was used to estimate crystallite sizes using the modified Scherrer’s formula[34] based on the broadening of XRD diffraction peaks Br ¼

In all experiments, stoichiometric excess (about 10 wt pct) amount of reducing agent Mg was used to account for the surface oxidation of the metal and for the residual oxygen that could be absorbed onto reactant surfaces and milling medium. Moreover, 50 wt pct and 100 wt pct excess of B2O3 were used in the synthesis of W2B5 along with stoichiometric amounts. 1806—VOLUME 44A, APRIL 2013

kk g tan h L cos h

½2

where Br is the total broadening due to reduction in crystallite size and lattice strains, k is the X-ray wavelength, L is the crystallite size, g is the strain in the material, k is a constant, and h is the diffraction angle. Nonisothermal differential scanning calorimetry (DSC) experiments of the powders were conducted in a METALLURGICAL AND MATERIALS TRANSACTIONS A

simultaneous DSC-thermogravimetric analysis (SDT) Q600 device (TA Instruments, New Castle, DE) under Ar in a temperature range between room temperature and 1273 K (1000 C) with a heating rate of 283 K/min (10 C/min). Moreover, atomic absorption spectroscopy (AAS) experiments were carried out in an AAnalyst 800 device (Perkin Elmer Life and Analytical Services, Shelton, CT) to identify the presence of Fe, W, and Mg metals in the leach solution after the leaching process to determine whether the leaching process was successful.

During the experiments for W2B5 synthesis using stoichiometric amounts of initial powders, a crust formation on the walls of the stainless steel vial was observed. Approximately 10 pct of all milled 2W5BM powders were in the form of loose powders and the rest were retained in this crust. This crust was so tightly stuck on the vial surface that breaking it without damaging the vials required an extreme effort and even sometimes could not be achieved. Figure 1 shows the XRD patterns of loose 2W5BM powders MA for 20 and 30 hours. As seen in Figure 1, the diffraction peaks of the W2B5 phase (Bravais lattice: simple hexagonal; a = 0.298 nm; c = 1.387 nm; space group: P63/mmc; ICDD Card No: 38-1365), the W phase (Bravais lattice: body-centered cubic; a = 0.316 nm; space group: Im 3m, ICDD Card No: 04-0806) and the MgO phase (Bravais lattice: facecentered cubic; a = 0.556 nm; space group:Fm3m, ICDD Card No: 71-3631) can be identified in both samples. Since the peaks of W2B5 and W phases overlap at a 2h value of around 40 deg, individual peaks of these phases cannot be distinguished. However, it is clear that the intensity of these overlapped peaks increase with MA duration. There is also a ferberite (FeWO4) phase (Bravais lattice: simple monoclinic; a = 0.473 nm; b = 0.570; c = 0.495 nm; b = 90 deg; space group:

P2/c; ICDD Card No: 74-1130) observed in the loose 2W5BM powders, whose peak intensity decreases with increasing MA duration. Furthermore, it is evident that a small amount of the WB phase also formed in these loose powders from the 2W5BM sample MA for 30 hours. Figure 2 shows the XRD patterns of the crust formation in the 2W5BM samples MA for 20 and 30 hours. As seen in Figure 2, the peaks of the WB phase (Bravais lattice: base-centered orthorhombic; a = 0.319 nm; b = 0.840; c = 0.307 nm; space group: Cmcm, ICDD Card No: 06-0541), the W2B phase (Bravais lattice: body-centered tetragonal; a = 0.556 nm; c = 0.474 nm; space group: I4/mcm, ICDD Card No: 89-1991), and those of the W, W2B5, MgO, and FeWO4 phases can be identified in crust formations from both of these 2W5BM samples. The formation of the FeWO4 phase is believed to be associated with the reaction of free Fe resulting from the wear contamination of the milling medium and initial tungsten oxide present in the system. It is also clear from Figures 1 and 2 that the intensities of W2B5 peaks increase, whereas that of the FeWO4 (2h ~ 30.45 deg) substantially decreases with increasing MA duration, indicating a decline in the FeWO4 amount with MA. Similar behavior is observed both in the crust formations and loose powders. The reason for this change in intensities can be attributed to decomposition of FeWO4 with further MA, and the resultant WOx reacts with B2O3 to form W2B5. This type of formation, decomposition, and reformation cycles are possible with nonequilibrium processes such as mechanical alloying.[27] Apart from that, it can also be seen that there is a slight increase in the amount of W with increasing MA for both loose powders and the crust formations. 2W5BM samples MA for 20 and 30 hours were leached with 5 M HCl, where loose powders and crusts were ground together in an agate mortar to prepare samples for leaching. Although 5 M HCl contains twice more HCl than theoretically required to remove all MgO in the samples, according to additional XRD investigations whose graphs are not presented in this

Fig. 1—XRD patterns of loose 2W5BM powders MA for (a) 30 h and (b) 20 h.

Fig. 2—XRD patterns taken from the crust formations in the 2W5BM samples MA for (a) 30 h and (b) 20 h.

III.

RESULTS AND DISCUSSION

METALLURGICAL AND MATERIALS TRANSACTIONS A

VOLUME 44A, APRIL 2013—1807

article, some amount of MgO was still present in both 2W5BM samples after leaching. Table I presents the results of the atomic absorption spectra (AAS) analyses of the filtered leaching solutions derived from the 2W5BM samples MA for 20 and 30 hours. As clearly seen in Table I, the amount of Mg in the filtered leach solution is slightly higher for the 2W5BM sample MA for 30 hours than that MA for 20 hours, indicating that the amount of the remaining MgO in the samples decreased with increasing MA duration. Higher dissolution of MgO with increasing MA durations can be attributed to the presence of finer particles with higher surface areas emerged during MA at longer durations.[33] However, there was hardly any change in the amount of the FeWO4 phase present in both samples indicating that leaching with HCl could not be used to remove FeWO4. In fact, it is almost impossible to remove FeWO4 for such a system via leaching process.[35] Moreover, there is a substantial amount of W present in both 2W5BM samples MA for 20 and 30 hours after leaching, which was expected due to low or no dissolution of W in an HCl solution.[36] In contrast to the Mg amount, the Fe amount in the leaching solution increased by about four folds. Since longer MA duration is expected to cause the decomposition of FeWO4, as stated before, the resultant iron and/or iron oxide might be dissolved during the leaching process, resulting in high Fe contents in the AAS analysis of the 2W5BM sample MA for 30 hours. On the other hand, the Fe amount of the sample MA for 20 hours is much lower than that MA for 30 hours since in the 2W5BM sample MA for 20 hours, most of the Fe resulting from contamination is retained as FeWO4, which does not dissolve during leaching. It is obvious that using stoichiometric amounts of initial powders is not appropriate for the W2B5 synthesis. The presence of FeWO4 is the most important problem, although its amount decreases with increasing MA duration. Increasing the MA duration was not considered as an appropriate solution for this problem since the amounts of W and other borides in the 2W5BM samples do not change with MA duration. The emergence of W and phases with low boron contents like WB and W2B are believed to be due to insufficient amount of B2O3. Thus, further experiments with stoichiometric excess amounts of B2O3 were carried out at the constant MA duration of 30 hours. Similar to other experiments with borides, also in these experiments with stoichiometric excess amounts of

Table I.

B2O3 to produce W2B5, crust formation on the walls of the vials was observed. However, the amount of the crust formation did not exceed more than ~10 pct of the entire 2W5BM sample and almost 90 pct of the samples were obtained as loose powders, quite contrary to the experiments containing stoichiometric amounts of B2O3. It is believed that the excess amount of B2O3 might have acted as a process control agent (PCA) in the mechanochemical synthesis[27] and prevented the crust formation. Thus, use of an external PCA was avoided, which otherwise would alter the purity of the end-product. Figure 3 shows the XRD patterns of the crust formation in the stoichiometric 2W5BM sample MA for 30 hours and those in the 2W5BM samples MA for 30 hours having 50 pct and 100 pct stoichiometric excess of B2O3. As seen in Figure 3, the peaks of W, WB, W2B, W2B5, and MgO can be identified in all three 2W5BM samples. On the other hand, the peak of the FeWO4 phase disappears when B2O3 was used in excess amounts. Moreover, the Mg3(BO3)2 phase (Bravais lattice: simple orthorhombic; a = 0.540 nm; b = 0.842; c = 0.450 nm; space group: Pnmn, ICDD Card No: 75-1807) is observed in the 2W5BM samples with excess B2O3 and its peak intensity increases with increasing amounts of B2O3. It is likely that some of the excess B2O3 has reacted with MgO to form the Mg3(BO3)2 phase. Figure 4 shows the XRD patterns of the crust formations in the 2W5BM samples MA for 30 hours having 50 pct and 100 pct stoichiometric excess of B2O3 after leaching with 5 M HCl. Although most of the MgO powders were removed after the leaching process, Mg3(BO3)2 is still present in the sample since this Mg3(BO3)2 phase cannot be removed from the tungsten boride system by leaching at room temperature.[32] Figures 5(a) through (c) are the XRD patterns taken from the loose powders in the 2W5BM sample having stoichiometric B2O3 with the loose powders in the 2W5BM samples containing 50 pct and 100 pct excess B2O3, all MA for 30 hours. As clearly seen in Figures 5(a)

AAS Analyses of the Filtered Leaching Solutions from the Samples MA for 20 and 30 h Composition of the Leaching Solution (ppm)

MA Duration

W

Fe

Mg

20 30

– –

16.17 62.28

5908 6135

1808—VOLUME 44A, APRIL 2013

Fig. 3—XRD patterns taken from the crust formations in the 2W5BM samples MA for 30 h: (a) with 100 pct of excess B2O3, (b) with 50 pct of excess B2O3, and (c) with stoichiometric B2O3. METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 4—XRD patterns taken from the crust formations in the 2W5BM samples MA for 30 h: (a) with 100 pct of excess B2O3 and (b) 50 pct of excess B2O3 after leaching with 5 M HCl.

Fig. 5—XRD patterns taken from loose powders in the 2W5BM samples MA for 30 h: (a) with 100 pct excess B2O3, (b) with 50 pct excess B2O3, and (c) with stoichiometric B2O3.

and (b), there are no traces of the FeWO4 or Mg3(BO3)2 phases observed when excess amounts of B2O3 were used. In addition to the main W2B5 and MgO phases, there are also small amounts of W and WB present in the loose powders with 100 pct excess of B2O3 (Figure 5(a)), while no phases except W2B5 and MgO are visible in the sample with 50 pct excess B2O3 (Figure 5(b)). On the basis of preliminary MA and leaching experiments, phases like FeWO4, Mg3(BO3)2, W, WB, or W2B cannot be removed from the samples by leaching. So, remaining experiments were focused on purifying only the 30 hours MA 2W5BM loose powders with 50 pct excess B2O3, which did not contain any of the above-mentioned byproducts. Before presenting the results of high-energy milling and leaching experiments on powders containing excess amounts of B2O3, the reasons for the difference in the compositions of crust formations and loose powders METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 6—DSC curves of (a) the unleached 2W5BM loose powders, (b) the 2W5BM loose powders leached with 5 M HCl, and (c) the 2W5BM loose powders leached with 7 M HCl under Ar atmosphere (100 mL/min) at a heating rate of 283 K/min (10 C/min).

should be mentioned. This difference can be explained in the sense of main mechanisms of the mechanochemical synthesis process. A characteristic feature of all solidstate reactions is that the formation of product phases occurs at the interfaces of the reactants. Mechanical alloying can substantially increase the reaction kinetics of the mechanochemical reactions providing repeated welding and fracturing of powder particles, which increases the area of contact between the reactant powder particles and allows fresh surfaces to come into contact repeatedly.[27,28] However, in the case of crust formation, which is a big concern in mechanical alloying and mechanochemical processing, the reaction kinetics may entirely change, preventing true alloying or reaction to occur[27] since there are no more fresh surfaces involved. Reasonably, it is expected that different reaction kinetics and mechanisms may lead to different endproducts. DSC analysis is considered to be a good verification to determine whether a reaction is complete since it should reveal different thermal behaviors when different phases are present in the structure. Figures 6(a) through (c) are the respective nonisothermal DSC thermograms of the 30 hours MA unleached 2W5BM loose powders and those leached with 5 M HCl and 7 M HCl comprising 50 pct excess B2O3. As clearly seen in Figure 6(a), there is a broad and large exothermic peak between the temperatures of ~723 K and ~1023 K (450 C and 750 C) for the unleached loose powders. A detailed interpretation of Figures 6(a) through (c) will be undertaken in accordance with the XRD analyses shown in Figure 7. Nevertheless, it is should be mentioned that increasing HCl molarity in leaching solutions causes the large exothermic peak in the unleached loose powders first to diminish and then finally to disappear (Figures 6(b) and (c)). In Figure 7, XRD patterns of the leached (with 5 M HCl) and unleached 2W5BM loose powders MA for 30 hours containing 50 pct excess B2O3 both after DSC analyses are shown. VOLUME 44A, APRIL 2013—1809

Fig. 7—XRD patterns of the 2W5BM loose powders MA for 30 h with 50 pct excess of B2O3: (a) before leaching and (b) after leaching with 5 M HCl, both after DSC analyses.

Fig. 8—XRD patterns taken from the 2W5BM loose powders MA for 30 h with 50 pct excess of B2O3 after leaching: (a) with 5 M HCl and (b) with 7 M HCl.

As clearly seen in Figure 7(a), the unleached loose powders almost entirely consists of the MgWO4 phase (Bravais lattice: simple monoclinic; a = 0.492 nm; b = 0.567; c = 0.468 nm; b = 90.7 deg; space group: P2/a, ICDD Card No: 27-0789) with complete absence of the W2B5 phase, indicating that almost all of the W2B5 reacted with MgO. Actually, the reaction between W2B5 and MgO was not expected to happen under a pure Ar atmosphere. Although this phenomenon could not be thoroughly understood, O2 trapped in the powders or trace amount of O2 in Ar should be responsible for this reaction since this reaction is highly exothermic in the presence of any O2 even at room temperature according to the enthalpy and the Gibbs free energy calculations (HSC Chemistry software; Outotec Solutions, Espoo, Finland). It has been shown that a self-propagating reaction results in a narrow and very intense exothermic peak during DSC due to near instantaneous combustion of the entire sample.[18] Because of highly exothermic nature of the MgWO4 formation (predicted theoretically using HSC Chemistry software), it is considered that this reaction is probably a self-propagating reaction. However, the DSC thermogram shown in Figure 6(a) exhibits a complex exotherm indicating that the formation of MgWO4 was not a simple process. Moreover, the presence of this type of broad exotherm is indicative of a system, which reacted gradually including several reactions.[18] The fact that peaks of W and Mg3(BO3)2 phases were also identified in Figures 7(a) and (b) implies that reactions other than the MgWO4 formation also occurred in compliance with the DSC curve shown in Figure 6(a). The formation of the Mg3(BO3)2 phase and other magnesium borates in similar systems at temperatures between 873 K to 973 K (700 C to 800 C) were also reported by several other authors.[37–39] Furthermore, it was stated that mechanically activation propels the MgO-B2O3 reaction for the formation of the Mg3(BO3)2 phase.[37] In addition to these, as expected, the peak of the MgO phase is also observed (Figure 7(a)) since there is more

than enough amount of MgO presented in the sample for the formation of MgWO4. On the basis of the DSC and XRD analyses so far, it is clear that no exothermic peak would emerge during DSC analysis of the 2W5BM loose powders MA for 30 hours comprising 50 pct excess B2O3 after leaching with 5 M HCl, if all the MgO phase were successfully removed from the system. Even if a small amount of MgO were still present in the powders, under these conditions, it would react with W2B5 to form MgWO4, which is a highly exothermic reaction according to the enthalpy and the Gibbs free energy calculations. In Figure 6(b), the DSC curve of the 2W5BM loose powders leached with 5 M HCl is given, exhibiting a small exothermic peak around 923 K to 1023 K (650C to 750C). Accordingly, the XRD pattern of the leached 2W5BM loose powders MA for 30 hours with 50 pct excess of B2O3 after DSC analysis shown in Figure 7(b) reveals the formation of MgWO4, whereas there is still a substantial amount of W2B5 present in the powders. The absence of any MgO peak is reasonable, since there should be very small amount of MgO in the loose powders after leaching, and therefore, all of MgO must have been consumed to form MgWO4 leaving a considerable amount of W2B5 unreacted. Figures 8(a) and (b) are the XRD patterns of the 2W5BM loose powders MA for 30 hours comprising 50 pct excess B2O3 after leaching with 5 M HCl and 7 M HCl, respectively. As evident from Figure 8(a), only the peaks of the W2B5 phase can be identified after leaching. There is only a small bump at the 2h value of the strongest peak of MgO, which is consistent with the small endothermic peak in DSC thermogram shown in Figure 6(b) and XRD pattern (Figure 7(b)) of the leached (with 5 M HCl) 2W5BM loose powders MA for 30 hours containing 50 pct excess B2O3 taken after DSC analyses. Since it was evident from DSC and XRD analyses that leaching with 5 M HCl was promising but not sufficient to enable the synthesis of pure W2B5 powders, 7 M HCl solution (~2.5 times more than

1810—VOLUME 44A, APRIL 2013


Fig. 9—(a) SEM-BEI micrograph (400 times) of the as-blended 2W5BM sample with 50 pct excess of B2O3, (b) SEM-SEI micrograph (1000 times and 5000 times inset) of the loose 2W5BM powders MA for 30 h with 50 pct excess of B2O3 before leaching, and (c) SEM-SEI micrograph (1000 times and 5000 times inset) of the pure W2B5 powders obtained after leaching with 7 M HCl.

stoichiometrically required amount) was prepared for the leaching process. Figure 8(b) is the XRD pattern of the loose powders MA for 30 hours with 50 pct excess of B2O3 after leaching with 7 M HCl. It is clearly evident from Figure 8(b) that only the peaks of W2B5 are identified, and there is not any slightest trace of the MgO phase. This result is very well in accordance with the DSC curve shown in Figure 6(c), which does not reveal any thermal event, and this is solid evidence that the loose powders MA for 30 hours with 50 pct excess of B2O3 is pure enough after leaching with 7 M HCl. As a result, the level of any kind of contamination is not significant, and no other reaction products seem to be present in appreciable amounts in the limit of the investigation techniques employed. Figure 9 shows the SEM/backscattered electron image (BEI) micrograph of the as-blended 2W5BM sample containing 50 pct excess B2O3 (Figure 9(a)), the SEM/ secondary electron image (SEI) micrograph of the 2W5BM loose powders MA for 30 hours with 50 pct METALLURGICAL AND MATERIALS TRANSACTIONS A

excess B2O3 before leaching (Figure 9(b)), and the SEM/ SEI micrograph of the pure W2B5 powders obtained after leaching with 7 M HCl (Figure 9(c)). Energy-dispersive spectra (EDS) scans taken from the as-blended 2W5BM powders revealed that the gray particles in Figure 9(a) are Mg, the white particles are WO2.72, and the black/dark gray particles are B2O3. The particle sizes vary between 60 lm and 70 lm for WO2.72 and Mg powders, while the mean particle size of the B2O3 is higher than 150 lm. EDS spectra taken from different regions shown in the SEMSEI micrograph of the 2W5BM loose powders after MA for 30 hours (Figure 9(b)) revealed that light gray small powder particles are W2B5 and dark gray agglomerates consist of the MgO phase. It is clearly seen that the particle sizes decrease drastically below 1 lm for W2B5 powders, whereas MgO can be seen in larger agglomerates with sizes of 4 lm to 5 lm. After leaching with 7 M HCl, only very fine W2B5 powders with no MgO phase can be observed in the SEM-SEI micrograph shown in Figure 9(c). Thus, as expected, the EDS spectra taken VOLUME 44A, APRIL 2013—1811

from this sample (pure W2B5 powders after leaching with 7 M HCl) did not reveal any presence of MgO phase. Moreover, particle sizes are approximately several 100 nm for pure W2B5 powders MA for 30 hours of the as-blended 2W5BM powders followed by leaching with 7 M at room temperature, as evident from Figure 9(c). Average crystallite sizes of 69 nm and 55.3 nm were estimated for the 2W5BM sample MA for 30 hours with 50 pct excess B2O3 before leaching and after leaching with 7 M, respectively. The decrease in the crystallite sizes after leaching with 7 M is believed to be due to strong acid attack preferentially on the grain boundaries.[35] Particle size distributions of B2O3, WO2.72 and Mg powders in as-blended 2W5BM sample as well as the loose powders MA for 30 hours containing 50 pct excess B2O3 before leaching and after leaching with 7 M were determined by a Mastersizer laser particle analyzer (Malvern Instruments, Ltd., Worcestershire, U.K.). According to these results, B2O3, WO2.72, and Mg powders possess average particle sizes about 191 lm, 61 lm, and 72 lm, which are consistent with the SEM micrograph shown in Figure 9(a). The average particle size of the 2W5BM sample with 50 pct excess B2O3 decreases sharply to 851 nm after MA for 30 hours, and it further declines to 226 nm after leaching with 7 M HCl contributed also by the removal of large MgO agglomerates. As a result, pure W2B5 powders with an average particle size of 226 nm and an average grain size of 55.3 nm were successfully produced by mechanochemical synthesis from WO2.72, Mg and 50 pct stoichiometrically excess B2O3 powders after mechanical alloying for 30 hours, followed by leaching with a 7 M HCl solution. The molar yield of the W2B5 synthesis was calculated as 89 pct according to Eq. [1]. Moreover, using excess B2O3 not only compensated the insufficient amount of B2O3 reagent in stoichiometric powder blends to complete the reaction but also acted as a process control agent to prevent the crust formation on the walls of the stainless steel milling medium, which otherwise could cause big problems in practice. Overall, considering the simplicity of the processing route, i.e. mechanochemical synthesis followed by leaching and the utilization of standard equipment in the synthesis of W2B5, it is hoped that this study would serve as an inspiration for new research investigations in the synthesis of similar materials.

IV.

CONCLUSIONS

In this study, pure W2B5 powders were synthesized at room temperature via mechanochemical synthesis followed by the leaching process. The final product is a fine powder with an average particle size of ~200 nm and can be obtained with high yields (89 pct). Based on the results reported in this study, the following conclusions can be drawn: 1. Mechanical alloying stoichiometric amounts of initial powders (WO2.72, B2O3, and Mg) caused a crust formation on the walls of the stainless steel vial in 1812—VOLUME 44A, APRIL 2013

2.

3.

4.

5.

which approximately 10 pct of the initial charge was in the form of loose powders and the rest was retained in this crust. The formation of FeWO4, W, and several other tungsten boride phases were observed during mechanical alloying of initial powders having stoichiometric amounts. These phases could not be removed from the system by the leaching process. Although the FeWO4 amount decreased with increasing MA duration, the amounts of W and other tungsten borides (different than W2B5) in the samples did not change with MA duration. During mechanical alloying experiments of initial powders containing excess amounts of B2O3, crust formation on the walls of the vials was observed. However, the amount of the crust formation did not exceed more than ~10 pct of the entire sample and almost 90 pct of the samples were loose powders, quite contrary to the experiments containing stoichiometric amounts of B2O3. W, WB, W2B, W2B5, and MgO phases with no sign of the FeWO4 phase were identified in the crust, when B2O3 was used in excess amounts. On the other hand, the Mg3(BO3)2 phase was also observed in the crust with excess B2O3 and its peak intensity increased with an increasing amount of B2O3. Although most of the MgO powders could be eliminated after the leaching process, Mg3(BO3)2 phase could not be removed from the tungsten boride system by leaching at room temperature. In case of the loose powders, there were no traces of the FeWO4 or Mg3(BO3)2 phases observed when excess amounts of B2O3 were used. Moreover, W peaks disappeared in the sample with 50 pct excess B2O3, whereas small amounts of W and WB were present in the sample containing 100 pct excess B2O3. Although leaching with 5 M HCl solution removed most of the MgO in the loose powders, pure W2B5 powders with an average particle size of 226 nm and an average grain size of 55.3 nm were only produced when 30 hours MA loose powders containing 50 pct stoichiometrically excess amount of B2O3 were leached with a 7 M HCl solution.

ACKNOWLEDGMENTS The authors wish to express their thanks to the Scientific and Technological Research Council of Turkey _ (TUBITAK) for the financial support under the project number 105M065. The laboratory equipment utilized in this study were purchased through the research grant 2001 K 750-90146 provided by State Planning Organization (DPT) and this is gratefully acknowledged. We would also like to thank Duygu Ag˘aog˘ulları for her help in leaching experiments. REFERENCES 1. E. Lassner and W.D. Schubert: Tungsten: Properties, Chemistry, Technology of the Element, Alloys and Chemical Compounds, Kluwer Academic, New York, NY, 1999.


2. K.A. Khor, L.G. Yu, and G. Sundararajan: Thin Solid Films, 2005, vol. 478 (1,2), pp. 232–37. 3. K.F. Cai and C.W. Nan: Ceram Int, 2000, vol. 26 (5), pp. 523–27. 4. K.B. Kushkhov, V.V. Malyshev, A.A. Tishchenko, and V.I. Shapoval: Powder Metall. Metall. C, 1993, vol. 32 (1), pp. 7–10. 5. Y. Itoh and Y. Ishiwata: JSME Int. J. A-Mech. M, 1996, vol. 39 (3), pp. 429–34. 6. Y.L.G. Wen and T.Q. Lei: J. Eur. Ceram. Soc., 2006, vol. 26 (15), pp. 3477–86. 7. G. Wen, Y. Lv, and T.Q. Lei: Carbon, 2006, vol. 44 (5), pp. 1005–12. 8. E.A. Gorbunov and M.P. Bryksin-Lyamin: Poroshkovaya Metallurgiya, 1971, vol. 4, pp. 91–97. 9. M. Usta, I. Ozbek, M. Ipek, C Bindal, and A.H. Ucisik: Surf. Coat. Technol., 2005, vol. 194 (2,3), pp. 330–34. 10. S. Stadler, R.P. Winarski, J.M. MacLaren, T. Eskildsen, J. van Ek, D.L. Ederer, E.Z. Kurmaev, M.M. Grush, T.A. Callcott, A. Moewes, and M. Lee: J. Electron. Spectrosc., 2000, vol. 110 (1–3), pp. 75–86. 11. V.V. Gostishchev, V.F. Boiko, and N.D. Pinegina: Theor. Found Chem. Ent., 2009, vol. 43 (4), pp. 468–72. 12. P. Peshev, G. Bliznakov, and L. Leyarovska: J. Less-Common Metall., 1967, vol. 13, pp. 241–47. 13. J.F. Shackelford and W. Alexander: Materials Science and Engineering Handbook, 3rd ed., CRC Press LLC, Boca Raton, FL, 2001. 14. Y. Lv, G. Wen, L. Song, and T.Q. Lei: Wear, 2007, vol. 262 (5–6), pp. 592–99. 15. H. Itoh, T. Matsudaira, S. Naka, H. Hamamoto, and M. Obayashi: J. Mater. Sci., 1987, vol. 22 (8), pp. 2811–15. 16. N.J. Welham: Miner. Eng., 1999, vol. 12 (10), pp. 1213–24. 17. R. Ricceri and P. Matteazzi: Mater. Sci. Eng. A-Struct., 2004, vol. 379 (1,2), pp. 341–46. 18. N. Setoudeh and N.J. Welham: J. Alloy Compd., 2006, vol. 420 (1,2), pp. 225–28. 19. Y. Hwang and J.K. Lee: Mater. Lett., 2002, vol. 54 (1), pp. 1–7. 20. W.M. Tang, Z.X. Zheng, W.C. Wu, J. Lu¨, J.W. Liu, and J.M. Wang: Mater. Chem. Phys., 2006, vol. 99 (1), pp. 144–49.


21. K. Iizumi, C. Sekiya, S. Okada, K. Kudou, and T. Shishido: J. Eur. Ceram. Soc., 2006, vol. 26 (4,5), pp. 635–38. 22. K. Kudaka, K. Iizumi, T. Sasaki, and S. Okada: J. Alloy Compd., 2001, vol. 315 (1,2), pp. 104–07. 23. J.W. Kim, J.H. Shim, J.P. Ahn, Y.W. Cho, J.H. Kim, and K.H. Oh: Mater. Lett., 2008, vol. 62 (15), pp. 2461–64. 24. S. Begin-Colin, G. Le Caer, E. Barraud, and O. Humbert: J. Mater. Sci., 2004, vol. 39 (16,17), pp. 5081–89. 25. M.L. O¨veçoglu: Ph.D. Dissertation, Stanford University, Palo Alto, CA, 1987. 26. J.S. Benjamin: Mater. Sci. Forum, 1992, vol. 88, pp. 1–17. 27. C. Suryanarayana: Prog. Mater. Sci., 2001, vol. 46 (1,2), pp. 1– 184. 28. C. Suryanarayana, E. Ivanov, and V.V. Boldyrev: Mater. Sci. Eng. A-Struct., 2001, vol. 304, pp. 151–58. 29. M. Stubicar, A. Tonejc, and N. Stubicar: Fizika A, 1995, vol. 4, pp. 65–72. 30. R. Ricceri and P. Matteazzi: J. Alloy Compd., 2003, vol. 358 (1–2), pp. 71–75. 31. R. Ricceri and P. Matteazzi: Int. J. Powder Metall., 2003, vol. 39 (3), pp. 48–52. 32. S. Yazıcı: Master’s Thesis, Istanbul Technical University, Istanbul, Turkey, 2009. 33. E. Bilgi: Master’s Thesis, Middle East Technical University, Ankara, Turkey, 2007. 34. C. Suryanarayana and M.G. Norton: X-Ray Diffraction: A Practical Approach, Plenum Press, New York, NY, 1998. _ Duman: Istanbul Technical University, Istanbul, Turkey, Per35. I. sonal communication, October 20, 2011. 36. P. Walker and W.H. Tarn: Handbook of Metal Etchants, CRC Press LLC, Boca Raton, FL, 1991. _ Duman, and M.L. O¨veçog˘lu: 37. D. Ag˘aog˘ulları, O¨. Balcı, H. Go¨kçe, I. Metall. Mater. Trans. A, 2012, vol. 43A, pp. 2520–33. _ Girgin: Cent. Eur. J. Chem., 2010, vol. 8, 38. A. U¨çyıldız and I. pp. 758–65. 39. U. Dosler, M.M. Krzmanc, and D. Suvorov: J. Eur. Ceram. Soc., 2010, vol. 30, pp. 413–18.

VOLUME 44A, APRIL 2013—1813