Demonstrating the Simplicity and In Situ Temperature Monitoring of the Mechanochemical Synthesis of Metal Chalcogenides Suitable for Thermoelectrics

Mechanically induced self-propagating reactions (MSR) are an immediate transformation of precursors into products via an exothermic combustion-like process activated by mechanical action (similar to self-heat sustaining reactions where similar processes are activated by heat). The occurrence of MSR can often be identified by changes in the physical appearance of the product, a distinct smell at the moment of the reaction, or a scratching sound from the milling jar. However, empirical evidence suggests that these sensory indicators are not always reliable, and in many cases, the MSR event may not have occurred despite the presence of these signs¹⁶.

Gas temperature and pressure sensors have been developed and mounted into milling jar lids in the past to address these limitations, and numerous studies on the MSR events in chemical synthesis have been reported^20,21,22. However, the valuable equipment is expensive and not interchangeable between jars. Therefore, we developed an economical temperature sensor for observing MSR during planetary ball milling. This sensor provides a cost-effective and practical alternative to a more expensive and less sensitive sensor system since, in the present case, the sensor is directly in contact with the gas and sometimes also powder. The sensor, placed inside a small hole in the jar's lid close to the powder and milling balls, transfers the heat generated inside the jar into a readable temperature measurement. This principle allows for real-time monitoring of the MSR through the conversion of heat into an electrical signal, enabling precise and continuous temperature tracking. The fastest scan rate of the temperature is 80 ms. The sensor demonstrated high sensitivity to local temperature changes inside the jar, as evidenced by the real-time monitoring of temperature fluctuations during steps 4.7 and 4.8 in the protocol.

The MSR occurred much faster in the Cu-S system than in the case of Sn-Se. This is in agreement with our previous works^14,19. The difference between the repeated experiments in the case of ignition time is not large and is within an experimental error. Notably, the transformation to the SnSe phase was rapid, occurring within the range between 87 and 97 seconds for all three experiments, respectively. In the case of Cu_1.8S synthesis, the MSR event was observed consistently in all four trials, always within the first second of milling. However, there were big differences in terms of the temperature increase in both systems. This is most probably related to the fact that the hot particles will get close to the sensor or not, which might always be different. Thus, the temperature increase cannot be considered as a relevant parameter, at least with the current setup. Nevertheless, it seems that the temperature increase is milder in the Cu-S system. The obtained products exhibit distinct physical characteristics compared to the precursors. Cu_1.8S, appearing blue and agglomerated (Figure 2A,B), suggests the occurrence of the MSR. The dark blue color is typical for covellite, CuS and the photograph was taken for the sample Cu_1.8S-1 where, according to XRD, the content of covellite was around 10 %. For the samples where the amount of covellite was lower, the color of the product was black,e.g., Cu_1.8S-4. The SnSe products were always gray.

The X-ray diffraction, in combination with proper data treatment, is an extremely powerful tool to describe the phase composition of the products. Copper sulfides produced via the mechanochemical synthesis can possess different phase compositions, depending on the precursors used. For example in Godočíková et al.²³, when the copper acetate (CH₃COO)₂Cu.4H₂O and Na₂S were milled, covellite and chalchantite (CuSO₄.5H₂O) were obtained. When the synthesis was performed on a larger scale in an eccentric vibratory mill, utilizing the Cu:S stoichiometric ratio 1:1 yielded the formation of CuS and Cu₂S phases²⁴. Our group was the first to report the MSR event for Cu-S system¹⁶, and it was found that CuS and Cu_1.8S are formed as a result of MSR when treating 1:1 stoichiometry, with covellite becoming more pronounced with prolonged milling²⁵. It has to be noted that copper has to have a specific needle-like morphology to react via a MSR^14,16. As mentioned in the results part, there are differences in the phase composition of the XRD patterns of Cu_1.8S samples. These differences are most probably caused by the time how fast we were able to stop the milling process after the MSR event. According to Baláž et al.³, almost pure digenite is formed after MSR, which is in accordance with the sample Cu_1.8S-3. We most probably succeeded in stopping the mill quite early in this case. However, further milling leads to a partial transformation to covellite, and in the case of samples Cu_1.8S-1 and Cu_1.8S-2, we managed to stop the mill slightly later. According to the reports on the mechanochemical synthesis of SnSe from corresponding elements, the phase-pure orthorhombic SnSe phase was almost exclusively obtained^16,26,27. In our case also the presence of SnSe₂ phase was observed. Again, its amount differs among the three experiments, and thus, we are of the opinion that there is some interchange between the SnSe and SnSe₂ depending on the time that elapsed from the MSR until the stopping of the mill. Nevertheless, the Sn-Se system seems to be more stable than Cu-S one with regards to the phase composition obtained after MSR, as the appearance of all three XRD patterns is very similar (Figure 3A), unlike in the Cu-S system (Figure 3B). The answers regarding phase compositions during milling could be provided by in situ X-ray monitoring^28,29 and we are planning to study these reactions in this way in the future. However, only oscillation milling can be monitored in this way for now (in situ monitoring of planetary ball milling setup is not available yet). The Rietveld refinement of the two selected XRD patterns from both systems confirmed the nanocrystalline character of the products. Metal chalcogenides are usually prepared with the crystallites on the nanoscale via mechanochemistry¹⁶. However, MSR processes are specific, and due to combustion, the crystallite sizes are usually larger than when performing the reaction in a gradual manner, e.g., the ones in Baláž et al.¹⁶ are much larger than the ones in Tsuzuki et al.³⁰. In the case of Cu_1.8S, the determined crystallite size of the main digenite phase was either around 63 nm or 39 nm. In the former case, the product was analyzed sooner after the MSR, so the crystallite size is larger. The crystallite sizes reported in this study are larger than 15 nm reported for the CuS obtained after the MSR in²⁵. The SnSe particles were larger than that of Cu_1.8S, namely the values larger than 84 nm were observed. These values are significantly larger than 12 nm reported for SnSe synthesized by mechanical alloying for 1 h³¹. In general, continuing milling after the MSR seems to rapidly diminish the crystallite size. After the MSR event, the crystallite size decreases and can reach similar values to the ones obtained via gradual processes after some time.

The nanostructuring being brought about via the mechanochemical approach is very beneficial for the TE performance of the materials³². Despite the fact, that they are densified by a subsequent step (such as spark plasma sintering), the materials still keep the memory of their original synthetic procedure³³. However, it is beyond the scope of this study to analyze and discuss the TE performance of the prepared products.

This protocol shows a simple, environmentally and user-friendly methodology to obtain nanocrystalline materials and even monitor temperature during the process. The critical steps in the protocol are to check if the monitoring device is working properly (step 3.2) and to stop the milling process as soon as possible after the MSR occurrence (step 4.8). The rest is just about characterization and data analysis. For now, there is a limitation about the quick destruction of the sensor, but this can be easily replaced with a new one, as these are very cheap. It would be great if the mill could automatically stop at the moment of MSR, but this has not been solved yet. The solution proposed in this study offers much more detailed temperature monitoring than is available on the market for now. From the synthesis point of view, by just introducing elements, we are capable of producing the powder suitable for TE applications within the second range. Utilizing mechanochemical synthesis to ignite MSRs for diverse materials preparation via simultaneous temperature monitoring might be an interesting topic for future research on a more general level.