Chapter 5. Scaled-up Chemical Synthesis of Zn-Sb Nanoparticles and Thermoelectric
5.3. Results and Discussion
5.3.4. Thermoelectric properties of the Zn-Sb nanostructured pellet
101
shown in Fig. 5.8 indicate that both Zn and Sb exist in the sample, but segregated from each other to in different areas, with dark color for Sb and lighter for Zn in the dark field image. The elemental mapping images further indicate that oxygen exists in the same area with Zn rather than Sb which is consistent with the existence of ZnO and Sb. The EDS results for Zn area and Sb area (Appendix V, Fig. A5.3) are also consistent with the results indicated in elemental mapping images. But, it is noted that, Zd and Sb are not completely segregated and ZnSb phases can exist in the pellet, where both Zn and Sb were found (Fig. 5.8 and Appendix V, Fig. A5.4).
The TEM image (Fig. 5.8) also indicates that the pellet contains space between nanoparticles which contributed to the low density of the sample.
In summary, the pellet display Sb/ZnSb and ZnO segregated phases especially near the surface. The pellet contains the large (in order of micron) and small (hundred nanometer) grain sizes with porosity exist in the sample. Even though the optimal Zn-Sb phases were not found as the pure phase in the pellet, the maintain of small grain size under the pressing condition is important for further using of NPs as the building block for the nanostructured TE materials.
The following investigation of the TE properties of pellet is necessary to understand relation of the structure, composition and TE properties which will help further progress in this field.
5.3.4. Thermoelectric properties of the Zn-Sb nanostructured pellet
102
Figure 5.9. Temperature dependence of the thermoelectric properties of Zn-Sb nanostructured pellet prepared from Sample III: (A) electrical conductivity, (B) Seebeck coefficient, (C) thermal conductivity and (D) the dimensionless thermoelectric figure of merit.
Seebeck coefficient versus temperature (Fig. 5.9B) is nearly constant from near room temperature to 200 oC then decreases as the temperature increases. It can be explained due to the increase of carrier concentration due to the increases of temperature. The positive sign of Seebeck coefficient over the entire measured temperature range indicates of the p-type conduction with holes as the dominant carrier in the sample. The value of α is of about 17 µVK
-1 from near room temperature to 200 oC which is larger than the Seebeck coefficient of Zn (2.5 µVK-1)18 but in the same order with the Seebeck coefficient of Sb (~24 µVK-1).19
Thermal conductivity as function of temperature was plotted in Fig. 5.9C. к slightly increases along temperature range from near room temperature to 250 oC, then increases more significantly when the temperature rises up to 400 oC. Additionally, it is possibility that the small value of κ (~0.5-0.6 WK-1m-1 at near room temperature to 250 oC) can be contributed partly by the low density of the sample. The к versus T curve indicates the phonon scattering by grain boundary. The huge amount of grain boundary can be offered by the nanostructure still maintained in the pellet.
0.2 0.25 0.3 0.35 0.4
0 50 100 150 200 250 300 350 400 450
ρ(mΩ.m)
Temperature (oC) A)
0 5 10 15 20
0 50 100 150 200 250 300 350 400 450 α (µVK-1)
Temperature (oC) B)
0.4 0.5 0.6 0.7 0.8
0 50 100 150 200 250 300 350 400 450 κ(Wm-1K-1)
Temperature (oC) C)
0.0 0.2 0.4 0.6 0.8
0 50 100 150 200 250 300 350 400 450
ZT x 10-3
Temperature (oC) D)
103 Sample IV pellet
After the hot pressing at 100 MPa and 500 oC for 5 hours, the pellet was used to measure the thermoelectric properties on a ZEM and Laser flash. The results were shown in Fig. 5.10.
The temperature dependence of the thermal conductivity, electrical resistivity and Seebeck coefficient of the pellet does not show any significant different in the cooling curve and heating curve, which indicates the stability of the sample and its homogenous thermal expansion.
Figure 5.10. Temperature dependence of the thermoelectric properties of Zn-Sb nanostructured pellet prepared from Sample IV: (A) electrical conductivity, (B) Seebeck coefficient, (C) thermal conductivity and (D) the dimensionless thermoelectric figure of merit.
Fig. 5.10A is the temperature dependence of the pellet’s resistivity in which the decrease of ρ along with the increasing of temperature (dρ/dT<0) for the entire measurement temperature range exhibits the semiconductor like behavior of the pellet. The resistivity of 0.026-0.029 mΩ.m which is 3 order larger than the value of Zn (metal, 6x10-5 mΩ.m) and 2 order higher than that of Sb (semimetal, 40x10-5 mΩ.m) also suggests that the Zn-Sb pellet is a semiconductor. When compared to the case of Zn42Sb58 sample without addition of Zn NPs, the resistivity of this sample is much smaller even its density is lower than that of the former one. Additionally, in Sample IV pellet, the content of zinc oxides may be larger than that in Sample III pellet (the XRD peaks of ZnO in Sample IV pellet are clear with higher relative
0.025 0.026 0.027 0.028 0.029 0.03
0 50 100 150 200 250 300 350 400 450
ρ(mΩ.m)
Temperature (oC) A)
15 20 25 30
0 50 100 150 200 250 300 350 400 450 α (µVK-1)
Temperature (oC) B)
1.5 2 2.5 3
0 50 100 150 200 250 300 350 400 450 κ(Wm-1K-1)
Temperature (oC) C)
2.0 3.0 4.0 5.0 6.0
0 50 100 150 200 250 300 350 400 450
ZT x 10-3
Temperature (oC) D)
104
intensity than that of Sample III pellet. Therefore, it is possible that the electrical contact between NPs in this case was improved due to the higher temperature (500 oC) during the hot pressing process. It is reasonable because the increase of the thermal conductivity of 3 times compared to that of Sample III pellet also observed for the entire measured temperature range.
The Seebeck coefficient (Fig. 5.10B) is positive in entire measurement temperature range which again indicates the p-type conduction with hole as the dominant carrier in the sample. Moreover α is found to decrease with increasing of the measurement temperature.
Combination with the increase of thermal conductivity as well as the decrease of electrical conductivity along with increasing of measurement temperature, those tendencies of α, κ, and ρ can originate from the increase of carrier concentration of semiconductor due to the increasing of temperature. The value of Seebeck coefficient is higher than 20 µVK-1 and reaches the maximum value of 26 µVK-1 at 150 oC. Even the Seebeck coefficient is improved compared to that of the Sample III pellet, it is still similar to that of Sb (~24 µVK-1).19
The overall zT is very small (~ 10-3) but was improved one order compared to that of Sample III pellet mainly due to the improvement of the electrical conductivity in compensation with the increase of thermal conductivity. In both cases, the small zT is due to the low Seebeck coefficient and electrical conductivity. Both samples show that the maximum of zT falls in range of 200-250 oC which is different than the normal peak position at 400 oC for Zn50Sb50.20 Discussion on the TE properties and the micro-nanostructure of the pellet
Figure 5.11. Schematic illustration of the micro-structure of the pellet
Based on the results of TE properties, it is clear that both of the pellets display the semiconducting characteristic. The Seebeck coefficient close to the value of Sb, and which is far from the value of ZnO (the sign of α is negative for ZnO) as well as ZnSb phases indicates the primary contribution of Sb in the pellet which is consistent with the result of composition
105
and structure analysis as illustrated in Fig. 5.11. The Zn-Sb and ZnO phase may distribute in the matrix of Sb, and the spaces between NPs/phases can exist in the sample.
Because the grain size of ZnO phase is smaller than 100 nm (of 55 nm in Sample IV pellet and even smaller than 20 nm in Sample III pellet), the thermal conductivity of ZnO with small grain size can be reduced to 3 Wm-1K-1 (for 20 nm grain size of ZnO in bulk sample)32 compared to the bulk ZnO (~100 Wm-1K-1) (Table 5.3). The ZnSb and/or Sb phase also have grain size less than 100 nm. Moreover, the density of the pellets is lower than 90% of the theoretical value which can enhance the phonon scattering and reduce thermal conductivity.
These reasons contribute to the fact that the thermal conductivity of both pellets is small. The electrical resistivity of these pellets is similar to ZnSb, even though the sample rich in Sb. There are several reasons for this result such as: the presence of ZnO phase with high electrical resistivity can inhibit the electrical conduction of Sb and/or ZnSb main phase in the pellet, the porosity of the pellet can also offer the electron scattering. The increase of electrical and thermal conductivity in the Sample IV pellet compared to Sample III pellet can be explained due to the increase in the grain size of the main ZnSb and/or Sb phases which is mainly because of the higher hot-pressing temperature used for Sample IV pellet.
Table 5.3 Thermoelectric properties of related materials. The data if not mentioned more were given at RT.
Materials ρ [mΩ.m]
σ [µVK-1]
κ [WK-1m-1]
zT d
[gcm-3]
Reference, note
Zn47Sb53 0.11 196 2.3 4.56x10
-3
> 92%
theoretical value
[21]
Zn50Sb50
bulk 0.2 300-500 1.41 0.6-0.8
(300 oC) 6.36 [20]
Zn50Sb50
single crystal
0.012 ~200 3.7 [22], at 0 oC
106 Materials ρ
[mΩ.m]
σ [µVK-1]
κ [WK-1m-1]
zT d
[gcm-3]
Reference, note Zn50Sb50
polycrystal ~200 1.3-2.6 [23,24]
Zn 6x10-5 2.5 116 7.134 [18]
Sb 40x10-5 24.4 20-50 6.697 [19]
ZnO 75x107 -1.5x106
[25], n-type SC ZnO bulk
single crystal
100 5.607 [26],29
ZnO bulk
polycrystal 30-40 [30,31]
ZnO, grain
size<100 nm 10-500
-60~-100 (25-500
oC)
3 5x10-2
(300 oC) 90% [32]
Sb2O3
thin film
5x109
-2.5x104 [27, 28]