Trouble to see the officials click "Follow" in the upper right corner, which is convenient for you to discuss and share, and can bring you a different sense of participation, thank you for your support!
Text | Xiaoxin
Editor|Xiaoxin
Silicon is the primary material used in the large-scale manufacture of photovoltaic devices, MEMS systems, or diode temperature sensors. Recently, molten silicon has also been proposed as a particularly interesting phase change material for ultra-high temperature latent heat thermal energy storage (LHTES) applications due to its extremely high latent heat (1800 J/g) and high melting point (1414°C).
Given this new application, it is also crucial to establish reliable data on the thermophysical properties of silicon in both solid and liquid states, especially at temperatures well above its melting point.
However, most measurement techniques, such as thermogravimetric analysis, expansion or differential scanning calorimetry, are container-aided. The high reactivity of silicon with refractories can lead to questionable confidence in the results obtained and can even seriously damage the instrument.
Therefore, selecting the right container, crucible material for application and analysis at such high temperatures is a challenging task.
●○Materials and methods○●
In this study, high-purity silicon and commercially available h-B N sinters were used for the seat drip wettability test. The silicon material chosen is amorphous polysilicon from the commercial Siemens process, which is used to produce silicon feedstock for the electronics and solar industries.
This ultra-high purity silicon with a purity above 7N is precipitated from a high-purity silane (SiH) or a trichlorosilane (HSiCl3) high-temperature gas. Commercially available hot isostatically pressed hexagonal boron nitride substrate.
The trace element content in the studied h-BN substrate was analyzed using the LECOCS600 sulfur and carbon analyzer (combustion infrared detection technology) and the LECOTCH600 nitrogen/oxygen/hydrogen analyzer (non-dispersive infrared and thermal conductivity technology). The determination of trace elements in h-BN substrate was C:0.101±0.046, O:1.590±0.398, S:0.014±0.003wt.
Small squares (3×3×3 mm3) are cut from Si polycrystals by a precision metallographic cutting machine equipped with diamond grinding wheels. Each surface of the silicon wafer is gently ground on SiC paper to remove the SiO2 layer before loading into the vacuum chamber, then ultrasonically washed with isopropyl alcohol and dried.
Cut the receiving h-BN substrate on a plate with a diameter of 17 mm and a height of 4 mm. Prior to testing, the contact surface of the h-BN substrate is mechanically polished on a sheet of office paper to obtain an initial surface roughness r ~150 nm (surface profile is measured by NT-MDTNTEGRA spectroscopic scanning probe microscopy. Contrary to previous reports, diamond plaster was not used at any stage of polishing the h-BN substrate.
Wetting tests are performed using experimental complexes detailed elsewhere, which are performed by a seat-drop technique combined with a contact heating procedure, i.e., ultra-high purity silicon wafers (initial mass of m0~50 mg) are placed on an h-BN substrate followed by the following heating/cooling procedure for the Si/h-BN pair: preheating at 150 °C/16 h in a load-locked chamber.
Slowly heat from room temperature to 500°C at a rate of 10°C/min; Rapid heating from 500°C to 1450°C at a rate of 15°C/min; Five intervals for isothermal heating: 1450°C/5min(1); 1550 degrees Celsius / 5 minutes; 1650 degrees Celsius/5 minutes; 1700 degrees Celsius / 5 minutes; 1750°C/10 min, cool from 1750°C to room temperature at a rate of 20°C/min.
The test is performed under high vacuum conditions (p = 1 × 10−6 mbar), and for higher temperatures, an inert gas (static argon, p=850-900 mbar) is introduced into the chamber to inhibit the evaporation of silicon. During the wettability test, images of the droplet/substrate pair are recorded using the MC1310 high-speed high-resolution camera at a rate of 100 frames per second.
After testing, the cured couple is removed from the chamber and then subjected to structural evaluation. Characterization was performed using FEIScios field emitting gun scanning electron microscopy (FEGSEM) combined with energy-dispersive X-ray spectroscopy (EDS).
Using an Anton Paar UNHT unit equipped with a Vickers diamond indenter, the instrument hardness −1 in the micro-region of the cross-sectional solidification couple is tested at a maximum loading force of 0.1 N and a loading rate of 0.20 N·min.
In addition, by using the FACT/FactSage7.1 thermochemistry software database, experimental results are supported by thermodynamic assessment of phase stability and associated reactions occurring in the considered Si-B-C-N system.
The thermodynamic descriptions of the liquid and solid solutions of the system were taken from the FTlite database. The thermodynamic properties of pure substances, including gas species, are taken from the FactPS database.
●○Wetting and spreading behavior○●
A fast-forward video of the wettability test is presented as supplementary data. It should be noted that throughout the test, the droplets exhibit erratic behavior. Significant vibrations of the droplets (i.e. oscillations of the three-phase line) are recorded during both the heating and cooling phases. This effect is more pronounced at high temperatures T>1650C. Similar silicon droplet behavior was observed in high-temperature wettability tests of Si/SiO.
The authors conclude that the drop vibrational substrate of silicon on silica is a direct effect of the formation, accumulation, and release of silica gaseous products. On the other hand, given the lack of gas products generated by the reaction in the Si/h-BN system, the droplet vibrations observed so far are likely to be caused by other causes.
Oscillations of the three-phase line during high-temperature exposure should be discussed in terms of the combined effects of evaporation, chemical reactivity, and interfacial tension in the system.
This phenomenon arises from two conflicting effects occurring simultaneously: (1) matrix dissolution and (2) liquid metal evaporation and formation of volatile oxides. Thus, both methods can observe the apparent contact angle – the former decreases in value, while the latter increases in value.
Therefore, in the current situation, it appears that droplet vibration may result from partial dissolution and supersaturation of molten silicon and nitrogen on the substrate at high temperatures (through solid and liquid diffusion), followed by release through the liquid-gas and liquid-solid interface.
●○Wetting kinetics○●
Wetting kinetic plot θt=f(t) calculated heating to 1750°C, as shown in Figure 2. Four stages (I-IV) can be distinguished on the obtained curve: after the first stage of melting, the silicon sample forms a symmetrical droplet shape at 1420 °C, and the corresponding contact angle is θ 1420=129.
This result describes a very similar result to show... of decrease θ 1430=133 to θ 1430=117. As already mentioned in the introduction, this diffusion behavior comes from the chemical change of liquid silicon (its saturation in B) that reacts with the substrate prior to the formation of the 3N4 interface product layer. Given the results of our experiments, we will discuss this proposed mechanism later.
In the second stage, during further heating to 1550°C, the contact angle remains at a level of 115°, with almost constant θ values and their relatively small changes in this temperature range pointing towards a chemical and thermodynamic equilibrium between the solid/liquid/gas phases involved. Similar unwetting behavior in Si/h-BN systems has previously been reported at 1500°C, corresponding to an equilibrium contact angle of 105°.
In the third stage, at a temperature of 1550-1650 °C, the contact angle value was significantly reduced to the transition from non-wetting to wetting, accompanied by the vibration of the droplet and the consequent dispersion of the contact angle value (θ~90±5), and observation.
This behavior can be attributed to further changes in droplet chemistry due to increased dissolution of h-BN substrates from: The increase in the solubility of boron and nitrogen in liquid silicon has demonstrated that the fraction of nitrogen dissolved in silicon increases from 25 ppma to 67 ppma and boron fraction from 137 ppma to 173 ppma with an increase in the long-term annealing temperature of the Si/h-BN system in the range of 1420-1500 °C.
It can be expected that this trend will also remain constant at higher temperatures. In addition, when the solubility of B in liquid Si is T=1550-1650 °C, it is quite high (between 21 and 30% at. — According to the silicon-boron binary phase diagram, the estimated solubility limit is N in this temperature range, which is significantly lower than and not higher than 0.0002%.
In addition, the presence of boron in liquid silicon reduced nitrogen content. Therefore, it can be said that at this stage of the high-temperature interaction, liquid Si is gradually enriched with boron dissolved from the substrate, while overbalanced nitrogen is released at the liquid/gas interface.
The boron nitride lattice begins to degrade by forming nitrogen vacancies at high temperatures, a phenomenon that provides an additional factor that promotes substrate dissolution and enhances B and n diffusion: In the fourth stage, the contact angle remains at the level of 90° during further heating from 1650 °C to 1750 °C, and the θ t = f(t) curve points to the realization of equilibrium conditions in the system.
However, a very extensive dispersion record θ measurement is due to large drop vibrations. As already discussed in Section 3.1, this behavior should be linked to the strengthening processes that lead to chemical changes in liquid metals, in particular with the action of gaseous products.
It is also reasonable to assume that the strong three-line oscillations come from liquid silicon saturated with nitrogen and subsequently released at the liquid/gas interface. In addition, as the temperature increases, the effect of silicon evaporation should also be considered.
In addition, the wetting-dewetting transition of silicon droplets was recorded during cooling, as θ increased from about 90±3 (at 1750 °C) to θ after full curing = 1003. Interestingly, solidification begins at T=1390°C and is completed at T=1356°C. Both the decrease in the onset temperature of solidification and the relatively wide temperature range of this phenomenon indicate a chemical transition of the initial pure Si to a Si-based alloy.
However, the possible influence of the cooled melt on these T values should also be taken into account. One of the dehumidification models proposed by Zemskov can be used to elucidate the behavior recorded by the Si/h-BN system in this study, hypothesizing residual gas action. The model is based on the release of a previously dissolved gas (in this case, nitrogen) at the solid-liquid interface.
The gas is released at the solidification front, then transferred to the surroundings by diffusion and capillary convection, and then ejected in the gap formed between the solidified metal (in this case, Si crystals) and the substrate (in this case, h-BN). This model also seems to be in good agreement with what we discussed earlier about the wetting behavior of couples.
Reactivity of Si/h-BN systems at temperatures up to 1750°C: Microstructural characteristics of solidifiers, reactivity in Si/h-BN systems is assessed by macroscopic observations and SEM/EDS analysis performed on top view and cross-sectional interface regions. Macroscopic top views of the Si/h-BN pairs before and after the wettability test are shown in Figure 3, and macroscopic observations reveal three main findings:
After exposure to high temperatures, the color changes from initially white to yellow; The formation of dehumidification zones is observed; The presence of rings surrounding the solidified droplets is clearly distinguished. We believe that the color shift of h-BN substrates after wettability testing may be related to slight changes in their structure and chemical properties.
During heating in an inert atmosphere, degradation of the h-BN lattice begins at around 1500 °C. For h-BN samples heat treated to 1800 °C in an argon atmosphere, it has been shown that degradation of the BN lattice occurs through the formation of nitrogen vacancies and is also reflected by changes in optical appearance (from white to yellow).
More detailed SEM evaluation results show a well-defined h-BN sheet morphology in the dehumidification zone. During the high-temperature test, this area is covered with molten silicon and then exposed during the cooling step by the dehumidification movement of the droplets. In addition, the presence of a large number of spherical particles
It should be related to the formation effect of small (2-5 μm diameter) "daughter" droplets left on h-BN platelets after the "mother" Si droplet moves.
The roundness of the Si droplets also proves that they do not wet the surface of the h-BN sheets. This also means that the h-BN wafers are not covered by any continuous wettable product layer. Liang et al. also recorded that the high surface roughness of the obtained sheet form additionally increases the non-wettability of h-BN compared to its counterpart with a flat polished surface.
The wetting behavior of Si/h-BN couples during heating to and cooling from UHT to 1750°C is controlled by the substrate dissolution/reprecipitation mechanism.
The dissolution of the h-BN substrate in the Si sample, as well as the gradual saturation of the molten silicon by boron and nitrogen (by solid and liquid diffusion), changes the chemical composition of the initially pure Si droplets to a Si-B-N alloy at high temperatures, followed by the release of superbalanced nitrogen from the droplets through the liquid-gas and liquid-solid interface.