June 28, 2011
I am working at Academia Sinica in Taipei, Taiwan in a research group focused on producing bulk thermoelectric materials, as opposed to thin film thermoelectric materials. Thermoelectric materials allow for the direct conversion of heat into electrical energy. They are a means of increasing efficiency as waste heat from an engine, chemical reaction, or industrial process that can be used to produce energy instead of being simply waste. Thermoelectric devices have been used to cool car seats in luxury vehicles and can be used as solar panels using the sun’s rays to heat one side of the material.
In early June a theoretical paper was published that predicted Ce3Te4 would have a figure of merit (ZT value) somewhere between 13 and 14. The paper is available here. So we have been working to produce this material and confirm or deny the assertion that the material would have such a remarkable efficiency. Producing this material has been very interesting. Bulk cerium was purchased and then purified at this institute and high purity tellurium is available as a powder. So the powder element is cold compacted. These compacted cylinders and the bulk cerium are then melted together by arc welding in an argon chamber to prevent oxidation of the sample. The sample’s properties are then measured. The density is measured first because it is the simplest measurement to make. The sample’s thermal diffusivity is then measured in an LFA. This property is measured by hitting the sample on one side with a very few photons and measuring the time for the other side to heat up. This procedure is done at a many different temperatures. Next the sample's heat capacity is measured in a differential scanning calorimeter (DSC).
This instrument works by holding a reference material and the sample and determining the differential in heat that is added to each to increase the temperature of each as a function of temperature. The electrical resistivity would be measured next. After arc welding not all of the sample forms into a single bulk there are usually one or two very small beads of sample in addition to the main bulk sample. These small beads are ground to a powder by hand and submitted into the x-ray diffraction instrument to determine the contents of the powder. Much like NMR there are certain peak signals that are representative of certain compounds such as Ce3Te4, or other stoichiometric combinations of these elements as well as their oxides, such as CeO2. We have made four of these bulk samples so far and are in the process of measuring their properties.
The people I work with in lab here are very interesting, fun, and accommodating, which is especially appreciated by a foreigner like me. It is summer here, so there are other students working in the lab and they are willing to do activities after work or on the weekends. The research is taken very seriously, but the people I work with make it very fun and I don’t feel stressed in the lab, the atmosphere is quite relaxed.
July 11, 2011
We have run into quite a few problems during our experiments. Our first problem was the vast differences between the melting and vaporization temperatures of the two elements we are using, cerium and tellurium. Tellurium melts at 450° Celsius and vaporizes near 1000° Celsius, while cerium melts at 800° C and vaporizes above 3000° C. And according to our phase diagram (Gschneider, K.A., Jr. and Verkade, M.E. “Selected Cerium Phase Diagrams.”) the melting point of Ce3Te4 is 1645° C. It is necessary to melt the sample to produce a product with the most desirable properties since melting the sample greatly increases the electrical conductivity. So the first attempt to create a bulk sample of Ce3Te4 by arc welding produced a thick black lining inside the vacuum chamber. This was alleged to be tellurium and the ensuing x-ray diffraction measurements validated this hypothesis. There was a 20% mass loss comparing the sample before arcing, to the sample after arc. The XRD revealed that a prominent component in the bulk sample was CeTe, so almost all of the mass lost was tellurium. The reason tellurium was lost was because it has a lower melting point and boiling point so it enters the gas phase more easily than cerium and then condenses on the cooler walls of the chamber, which it cannot be reclaimed.
Our next attempt at arc welding was a slight improvement, we added an excess of tellurium because we thought tellurium would be lost in the process, which it was. Our larger problem though was the brittleness of the sample. After arcing the materials together, as the sample cooled it would fracture! This was very frustrating and it took 12 welds in total to completely fuse the sample (the first sample only took 6). When we took sample 2 out of the arc chamber after the first 6 welds and attempted to measure its thermal diffusivity, the sample turned to powder near 600 Kelvin. The sample was removed from the Laser Flash Apparatus (LFA – device that measures thermal diffusivity) and was re-arc welded 6 more times. When the sample was placed in the LFA again it again fractured at the same temperature. We thought the sample was too brittle and the force provided by the sample holder forced the sample to crack as it is was heated and therefore expanded.
For sample 3 we tried to make the sample less brittle by provided a longer time period for the sample to cool. This was done by attempting not to heat the sample to as high of a temperature (tried to remain just above the melting point of the sample ~ 1700° C) and the arc was kept in the vicinity of the sample while it cooled. These procedures allowed the copper plates around the sample to heat up and thereby reduced the rate of heat loss from the sample and decrease the cooling rate. It turned out that this sample also fractured at a similar temperature despite the steps taken to decrease the brittleness. A few good things were occurring though, with each successive sample the XRD results were revealing that the sample were coming closer to the ideal stoichiometry of Ce3Te4. This was due to decreasing weight loss percentages and better estimates for the amount of excess tellurium that needed to be added. So now we think that the material undergoes a phase transition near 600 Kelvin. It is most likely that the transition causes the material to expand dramatically, so it pushes against the sample holder, and the force is just too great and breaks the sample. To confirm or deny the phase transition hypothesize we will be doing XRD diffraction measurements at various temperatures to see if the lattice constants of the compound change around 600 Kelvin. Each compound can be classified as being in a certain 3-dimensional geometry (let’s say a rectangular prism for simplicity). The lattice constants would be the lengths of the sides of this shape, usually denoted as “a”, “b”, and “c.”
From these two shapes you can see that the lengths of the sides of the prism are not equivalent. We think the sample’s parameters are (qualitatively) going from a’, b’, c’ to a, b, c and that this expansion against a force causes it to break. I will give you the update on the results in the next post!
July 27, 2011
So we have confirmed that our sample of Ce3Te4 undergoes a temperature induced phase transition between 200 and 400 Celsius. We confirmed our hypothesis using powder x-ray diffraction with varying temperatures. It can be seen very clearly that the lattice shifts in this temperature range. The peaks on the diffractogram are detected at different angles, which means that the atoms are positioned differently and that the lattice structure changes in this temperature range. The phase transformation makes measuring any of the properties of the material over 300° C almost impossible because the sample breaks while crossing that temperature. So we come up with a plan to try to minimize the phase transition. Numerous experiments involving La3Te4 have been conducted and many papers on this research have been published. We will try to synthesize La3Te4 because it does not have a phase transition in the temperature range from 300 K to 1250 K and the thermoelectric properties of this material are also excellent (zT > 1.0). The properties of lanthanum (La) and cerium are also very similar: similar atomic mass, and most importantly both are commonly trivalent, which means that they, at least theoretically, are good substitutes for each other. So we have started synthesizing just a “simple” La3Te4 with the plan to start replacing lanthanum with stoichiometric amounts of cerium to produce La2.5Ce0.5Te4 and La1.5Ce1.5Te4. We are having problems synthesizing even the material with only lanthanum and tellurium. It is complicated for a number of reason but our main problem has been lanthanum’s ease of oxidation which severely limits the properties we are trying to maximize, especially electrical conductivity. Oxidation produces valence balanced formula units, so there are no free electrons. Without free electrons, it is difficult to conduct electricity.