Properties of Aluminosilicate Refractories with Synthesized Boron-Modified TiO2 Nanocrystals

An efficient microwave supported synthesis, with a reaction time of only one and a half minute, to prepare boron-modified titania nanocrystals TiO2:(B), was developed. The nanocrystals were obtained by hydrolysis of titanium tetraisopropoxide (TTIP) together with benzyl alcohol and boric acid, and the approach did not need surfactants use and a final calcination step. The produced TiO2:(B) nanocrystals were characterized in detail by low magnification Transmission Electron Microscopy (TEM), Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES), X-Ray Diffractometry (XRD), and a Micro Raman Spectroscopy. One of the obtained samples was then tested as an additive in various amounts in a typical aluminosilicate refractory composition. The effects of these additions in bricks were evaluated, according to UNI EN 196/2005, in terms of thermo-physical and mechanical properties: diffusivity, bulk density, apparent density, open and apparent porosity and cold crushing strength. Bricks' microstructure was analysed by Scanning Electron Microscopy (SEM) and energy dispersion spectroscopy (EDS). The bricks obtained with nanoadditives presented improved mechanical characteristics with respect to the typical aluminosilicates, presumably because of a better compaction during the raw materials' mixing stage.


Introduction
Aluminosilicate bricks have been used as refractory in industrial furnaces in constructing and restoring, in coke ovens, glass melting furnaces, and hot blast furnaces [1,2]. The high temperature volume stability and creep features of aluminosilicate bricks [2] are the unique properties of the heavy-duty refractory products. The conventional silica bricks generally exhibit some disadvantages such as degradation [1,3], low refractoriness [4], and thermal expansion due to their poor alkali content and thermal resistance [5,6,7]. The incorporation of nano-metal oxides, such as TiO 2 , enhances the physico-chemical and thermomechanical properties of refractories and their composites [8,9,10].
The processing of nanomaterials for the construction industry is complex in nature and requires careful monitoring to achieve the required performances. In the case of nanocomposites a major difficulty lies in the tendency of nanoparticles to aggregate into micrometer-sized crystalline forms in the attempt to reduce their surface free energy.
The present investigation deals with the eco-friendly production of nanoscale boron-modified TiO 2 particles and their application in aluminosilicate refractories. It is necessary, however, that the nanoparticles arrange uniformly in the matrix to effectively contribute to the composite's final properties; the nanoparticles' shape, moreover, may be important in increasing the compound mechanical strength according to preferential directions.
Knowledge of the thermo-mechanical properties of nanocrystals modified in an aluminosilicate refractory is relatively limited, and potentially of great importance, because such materials in service are typically subjected to constant stress over long periods, and therefore a design that ensures components' optimal performance over time requires a careful estimation of their lifetime.
In this paper concerned with improving thermal/insulating quality, resistance, and lifetime of standard concrete, boron-modified titanium dioxide nanoparticles were added to aluminosilicate specimens and the resultant thermo-physical and mechanical characterizations are presented. Tests were conducted also using anatase titanium dioxide nanocrystals synthesized in our previous works [11] but the results with boron modified anatase nanocrystals of the present study gave better results in terms of bricks' properties.
Physical properties, such as thermal diffusivity, apparent porosity, water absorption, and density were measured and compression mechanical resistance has been studied.
Finally, the open pore distribution and their size were estimated by mercury porosimetric analysis supporting the results from thermal and mechanical characterization. The refractories with and without additive (standard) were compared to highlight the changes in the composites' behaviour. Microstructure study was performed with SEM-EDS analysis.
In a typical preparation, TTIP (1 mL, 3.36 mmol) was added to boric acid dispersed in benzyl alcohol (10 mL, 96.4 mmol), while stirring in a Teflon vessel. The reaction vessel was sealed and maintained at 200° C for 1 minute and 30 seconds, without the requirement of a post-synthetic calcination treatment of the nanocrystals [12]. The microwave-assisted synthesis took place in a single and quick step, by using a microwave digestion system [13,14,15]. The system uses 2.45 GHz microwaves and it is controlled in both temperature (200 °C) and pressure (300 psi). The reaction vessel has been connected to a pressure transducer that monitors and controls the pressure during synthesis. The resulting milky suspension was centrifuged; the white precipitate was twice washed with diisopropylether and dried overnight at room temperature. Different TTIP: H 3 BO 3 molar ratios (1:1, 1:2, 1:3, 1:4, 1:5) were tested and samples were named BT1, BT2, BT3, BT4, and BT5, respectively.

Preparation of aluminosilicate bricks
The BT2 nanocrystals were tested as bricks' additives. They were prepared in three different suspensions of benzyl alcohol in concentration percentages of 0.5 wt.%, 1 wt.%, and 2 wt.%.
The bricks were obtained mixing 2.2 kg of commercial refractory material (Al 2 O 3 55 wt.% and SiO 2 42 wt.%) with water at 6 wt.% in which the additive suspensions were emulsified previously. The mixture was "thrown" in the mould in two different phases spaced by a compaction process and placed in the curing chamber at 20 °C, exposed to not less than 90% relative humidity for 24 hours; then each brick was carried out of the mould, and put back in the chamber to mature for 28 days, as described in UNI EN 196/2005. Later all the specimens were fired at 1300 °C for 24 hours.

Nanocrystals' characterization
Morphological characterizations of the TiO 2 :(B) nanopowders were performed using a JEOL JEM 1011 Transmission Electron Microscope (TEM). Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES) was performed on a Varian Vista AX ICP-AES instrument, with samples digested using a mixture of HNO 3 and HF 1:1. The ICP-AES measurements were accomplished by calibration using aqueous standards of Ti and B after appropriate dilution; the calibration ranges were selected according to the expected concentrations of the elements of interest.
The Raman spectra were acquired with a Renishaw inVia Raman microscope, with 633 nm excitation wavelength, a 50X microscope objective giving a spot diameter of 2μm, a resolution of 4 cm -1 , and a power of about 0.8 mW. Powders of the prepared samples were placed on a glass slide, and spectra were collected at room temperature.
Powder X-ray diffraction (XRD) for phase analysis of the nanocrystals was performed on a Rigaku diffractometer in Bragg-Brentano reflection geometry, using filtered Cu Kα radiation. The XRD patterns were recorded in the range of 2Θ = 20° -90° by step scanning, using 2Θ increments of 0.03°a nd a fixed counting time of 2 s/step.
The XRD patterns revealed only the presence of the anatase phase and were analysed by using a whole-profile Rietveld-based fitting program [16] according to the following procedure, fully described elsewhere [11][12][13][14][15]: 1. The instrumental resolution function (IRF) was evaluated by fitting the XRD pattern of a LaB6 NIST standard recorded under the same experimental conditions as those used for measuring the samples. The IRF data file was provided separately to the program in order to allow subsequent refinement of the XRD patterns of the samples.

2.
The phase composition of the samples was determined by fitting the XRD patterns with the crystal structure models of tetragonal TiO 2 anatase (space group I41/ amd; cell parameters: a = b = 3.7835430 Å and c = 9.614647 Å; α = β = γ = 90°). The weight percentage accuracy of the estimation was determined to be 5-7% w/w.

3.
The inhomogeneous peak broadening of the anatase TiO 2 reflections was described by a phenomenological model based on the modified Scherrer formula [11][12][13][14][15]. The refinement procedure allowed the calculation of the coherent crystal apparent domain size along each reciprocal lattice vector (h,k,l) direction. Other refinable parameters were the unit cell parameters, whereas the background was linearly interpolated and unrefined.

Aluminosilicate bricks' characterization
All bricks were characterized by different techniques. Density, apparent porosity, and water absorption were measured by the Archimedes method using water. Thermal diffusivity was determined using the Hot Disk technique: a disk sensor was placed between two pieces of the sample material, then heated by a constant electrical current for about 40 s. The generated heat, dissipated from the sensor into the surrounding unknown sample material, causes a temperature variation both in the sensor and in the surrounding material. The sensor average transient temperature is simultaneously measured monitoring a change in its electrical resistance. The sensor resistivity change correlates with the corresponding change in temperature through a temperature coefficient of resistivity.
The refractory mechanical performances were estimated by compression breaking tests, according to the Standard in force (UNI EN 196/2005); prismatic specimens were used (4x4x4 cm) and a 65-L12G2 Controls programmed control press was employed. The load was increased uniformly at a speed of 2400 N / s until specimen failure, which is of the brittle type.
The bricks' microstructure was analysed through Scanning Electron Microscopy with a JEOL 5410 LV SEM microscope equipped for energy dispersion spectroscopy (EDS). EDS allows detecting X-photons emitted by the sample in response to the incident electrons from the SEM microscope giving evidence about the specimen elemental composition. The samples were fractured under liquid nitrogen and successively metalized to ensure a good electrical conductivity.
Porosimetry measurements were accomplished with a Pascal 140 and 240 Thermo Finnigan mercury porosimeters. The penetration pressure relates directly to the pores' size according to the Washburn equation [17]: where γ is the mercury surface tension (generally 0.480 N/ m), θ is the contact angle between the mercury and the solid (in average 140°), Pc is the penetration pressure, and R the pore radius. The hypothesis of the Washburn equation is that of cylindrical pores.

Results and discussion
TEM images of the boron-modified titania nanoparticles are reported in Figure 1.
TEM analysis, reported in Table 1, revealed that the axes' average lengths were comparable in all the samples, with an estimated average length for the short axis of about 5-6 nm, and of 7-8 nm for the long axis. The calculated aspect ratios were comparable under the error in all the samples, thus boron addition did not influence appreciably the size of the resultant nanocrystals. Boron content evaluated by ICP-AES increased with the nominal increment of TiO 2 : B ratio ( XRD was carried out to investigate the crystal structure of boron-modified TiO 2 nanoparticles and the corresponding XRD patterns are shown in Figure 2.     [12,18]. Micro Raman spectra were collected on TiO 2 :(B) samples and are reported in Figure 3. All the detected peaks in Raman spectra belong to the anatase TiO 2 phase and lie in the region 100-700 cm -1 : Fig. 3 shows E g (1) around 144cm -1 , E g (2) at 197 cm -1 , B 1g (1) at 400 cm-1, A 1g (1) and B 1g (2) at 520 cm -1 , and E g (3) at 640 cm -1 . Table 3 shows the data for unmodified and modified bricks, obtained varying suspensions weight percentages (from 0.5 wt.% to 2 wt.%). Tests have shown that nanocrystals increasing additions, under the error left the bulk density and the apparent density, water absorption and the apparent porosity were left almost unchanged. The resulting bulk densities were a little higher in respect to the reference brick. Bending failure tests were conducted on three points at room temperature, using a programmed press. Each dried specimen was placed horizontally between two supports distanced by 100 mm, and divided into two half prisms. Compression breakage tests were conducted according to standard regulations: the two prisms settled previously for the bending test were used, and in this case a press with a programmed control was also employed. The load was uniformly increased at a speed of 2400 N/s, until the specimen failure. The failure mode denoted fragile composites.
The breaking strength (Table 3) increased appreciably with the nanoadditive concentrations with respect to the reference brick. Thermal Diffusivity, determined by the hot disk method slightly decreased with nanofillers amount (Table 3). Results reported in Table 4 show that the cumulative volume, which is indicative of the matrix compactness, does not correlate with the nanocrystals' amounts and is only slightly lower than in the reference brick. The surface areas increased as well as the calculated pores' size, while the calculated open porosity was lower than in the refer-ence brick (which is in agreement with results reported in Table 3). To study the mechanism improving the breaking strength, SEM analysis of refractory bricks has been conducted and the microstructures at low and high magnification are shown in Fig. 4. In the low magnification mode, large aggregates with smooth surfaces could be observed in both samples; however, in the refractory without nanoparticles there are areas with clear internal defects and larger micropores. The addition of TiO 2 :(B) nanoparticles causes some differences in the microstructures: the microcracks' and micropores' sizes were greatly reduced. High magnification SEM images confirmed a denser microstructure with fewer voids in TiO 2 :(B) modified refractory (Fig. 4: c and d).
Moreover, EDS through the elemental analysis evidenced that in nanocrystals-modified specimens the fillers were present along the entire scanned surface.
To explain SEM results and the overall increase in the breaking strength with nanocrystals' additions it is to be assumed that nanoadditives possibly induced a better particles' aggregation, through a thickening effect during the formation of the liquid phase. Previous literature studies [19] demonstrated that anatase containing boron, at least at the surface of the nanocrystals, has peculiar characteristics in bricks: boron favours the persistence of the anatase phase when thermal treatments are performed [19]. It can qualitatively explain the higher compression breaking resistance with the nanofillers obtained in the present study. The results of the SEM analysis are in contradiction with results collected in Table 4, where an increase in the pore radii calculated in bricks with modified nanocrystals is observed: we speculate that it may be because of a failure of the hypothesis of the Washburn equation, which is valid for cylindrical pores.

Conclusions
In the present work, a microwave-assisted synthesis of boron-modified TiO 2 anatase nanocrystals with about one minute and a half of reaction time was developed. The effective boron content was tuned by the addition of boric acid and increased with its nominal amount. It was found that boron addition does not influence appreciably the nanocrystals' size and morphology. One of the obtained samples was tested as additive in various quantities in a typical aluminosilicate refractory composition. The effects of nanocrystals' addition on bricks were evaluated according to UNI EN 196/2005 in terms of thermophysical, mechanical properties, and microstructure. The thermal conductivity decreases and breaking strength increases with the nanofillers' addition. According to SEM images the TiO 2 :(B) nanoparticles' addition reduced the microcracks' and micropores' sizes.
SEM-EDS spectroscopy confirmed the presence of nano fillers along the scanned surfaces. Possibly, nanoadditives induced better particles' aggregation, through a thickening effect during the formation of the liquid phase; in this way the nanocrystals-modified bricks presented improved mechanical characteristics due to a better compaction of the components.

Acknowledgements
This work was partially supported by the FIRB 2009/2010 project "Rete integrata per la Nano Medicina (RINAME)" -RBAP114AMK_006. Giuseppe Chita and Giovanni Filograsso are acknowledged for their technical and administrative support.