Chapter 1 Introduction
1.8 Research approach
1.8.1 High valence and high atomic weight metal ion doping
Typically, the doping of metal oxide is conducted by introducing guest ions (dopants) which are aliovalent to replace the cations or anions in the host materials. In ALD technique, the guest oxide layers, in which metal ions are aliovalent compared with the ones in host materials, are periodically inserted into the host metal oxide to offer dopant ions. Take Al doped ZnO (Al:ZnO) as an example, Al2O3 guest layers were periodically inserted into ZnO host layers by alternating one Al2O3 cycle with a number of ZnO cycles. In this case, since the valence of dopants, i.e. Al3+, is higher than that of Zn2+, the free electrons, which are able to make contributions to conducting, were created at the interface of Al2O3 and ZnO; hence the n-type doping was achieved. In addition to inducing free electrical carriers, the insertion of guest layers created a number of host/guest interfaces periodically which can also serve as the phononscattering centers to reduce the thermal conductivity of Al:ZnO by increasing interfacial scattering.
As listed in Table 1.1, the current ALD researches about thermoelectric properties of doped ZnO thin films focused on the usage of boron-group metal oxides, i.e. Al2O3
and Ga2O3, as inserted guest layers to arrive at n-type doping and reduce thermal conductivities by increasing interfacial scattering. However, based on the concept of
aliovalent doping, the usage of much higher valent metal ions as dopants is able to create more free electrons potentially. Besides, the effect of interface scattering on phonons can be enhanced by increasing the mass difference across the interface. Hence, in order to obtain these two advantages, we target on the ⅣB group metal ions doped ZnO to study the influence of the high valence and high atomic weight metal ion dopants on the thermoelectric properties. Because ⅣB group metals, i.e. titanium (Ti), zirconium (Zr) and hafnium (Hf), are tetravalent in compounds generally, they are able to donate more electrons compared to the boron group metal ions. Furthermore, Zr and Hf are much heavier than Zn, so we can systematically investigate the effect of mass difference between dopants and the host material on suppressing the thermal
conductivity.
1.8.2 Isotope superlattice
In the past, the ways to modify the thermal conductivity of materials, such as morphology control, tuning element composition, crystallinity and introduction of micro/nano-structure into matrix material, influenced the electrical conductivity in the meanwhile. Typically, the damage to electrical conductivity is accompanied with reduction of thermal conductivity through above ways. Luckily, isotopically
composition modifying hews out a new way to overcome the dilemma. Isotopes, the
atoms only differ in numbers of neutron, show almost the same chemical and electrical properties; thus, the incorporation of isotopes into the host materials causes nearly no influence on the electrical performance. However, the mass difference changes the vibration frequency of atoms and then affects the transport properties of phonons.
Nowadays, the researches about the effect of isotopically modification on the thermal conductivity and the thermoelectricity were almost limited to elementary substances, e.g. graphene, carbon nanotube, silicon nanowire and lithium rod88–91. Hence, the isotope effect in multi-element compound is still under investigation. In this study, we use oxygen-18, which is the heaviest stable isotope of oxygen, to deposit the isotope-containing ZnO, and then experimentally verify the effect of the concentration of isotope and the period length of superlattice on suppressing the thermal conductivity of doped ZnO while remaining the optimal electrical performance developed by
aliovalent doping.
1.8.3 Novel MLD conducting polymer development and the metal oxide/polymer superlattice deposition
In tradition, conducting polymer thin films were fabricated by solution processes, such as typical chemical synthesis, spin coating and electrochemical polymerization.
However, the usage of solvents has to be considered in various applications owing to the
costs and toxicity. Besides, the solvent residue and orthogonality are also important concerns while fabricating organic based devices. Therefore, the vapor phase synthesis and deposition, which are able to overcome these problems, have drawn much
attention92. Thanks to the promoted monomer transport on the growth surface, the conducting polymer thin films prepared by vapor processes, such as vapor phase polymerization (VPP) and oxidative chemical vapor deposition (oCVD), are more crystalline and show a higher electrical conductivity92–94.
Recently, Parsons et al. successfully demonstrated the deposition of poly (3, 4-ethylenedioxythiophene) (PEDOT) via oxidative molecular layer deposition technique (oMLD) using 3, 4-ethylenedioxythiophene (EDOT) as a monomer and molybdenum pentachloride (MoCl5) as an oxidant95. The reaction mechanism is shown in Figure 1.19. In their results, deposited PEDOT films were coated on nanostructure features (e.g. silica fiber and nanoporous ITO) with precise control of thickness, which is about 20 nm, and the highest electrical conductivity exceeded 3000 S cm-1.
Nonetheless, the high condensability of MoCl5 (mp: 194°C; bp: 268°C) may cause the considerable variation during process. Thus, the explorations of oxidants with
adequate volatility are required to build up a robust and high reproducibility MLD process. In this study, we use two novel oxidants, i.e. SbCl5 and VOCl3, and then follow the similar reaction mechanism to set up versatile MLD conducting polymer deposition
process, such as polythiophene, polyaniline, and robust PEDOT process. After setting up the robust MLD conducting polymer process, we incorporate the MLD conducting polymer with the doped ZnO we developed in previous researches to deposit the metal oxide/polymer superlattices for investigating the dependence of the composition and the structure of superlattice on the thermoelectric properties.
Figure 1.19: The reaction mechanism of MLD PEDOT95.
1.8.4 Anti-hydrolysis polymer development
In our previous work, the polymer we utilized in the deposition of superlattice, PA32, suffered from the serious hydrolysis under humid conditions due to the following reasons. One is that the polymer was easily decomposed into oligomers with amine and/or carboxylic acid groups due to the high hydrolysis rate constant of amide bonds on PA32. The other is that the main chain of PA32 was composed of short aliphatic
components, which are non-hydrophobic. Therefore, water molecules are able to absorb onto the polymer chain easily and then induce the hydrolysis reaction of deposited polymers. For these two reasons, the gas barrier performance of HfO2/PA32 superlattice decayed with time.
To address this issue, two approaches are proposed in this dissertation. The first one is to develop the deposition of polymers with a relatively low hydrolysis reaction rate constant. For convenience, we choose a versatile and commonly used polymer, polyester, to verify the concept in our proposal. In this study, we control the structure of main chain while substituting the functional group only. The second approach is
maintaining the functional group but modifying the main chain of PA32, such as the introduction of aromatic component and/or long alkyl chains, to enhance the
hydrophobicity of polyamide and prevent the occurrence of hydrolysis reaction. It is noteworthy that the deposition temperature is set at 100~130°C, at which the gas barrier performance of inorganic part, HfO2, was excellent. After the development of anti-hydrolysis polymers, the superlattice gas barriers, which are made up of the alternating of anti-hydrolysis polymers and HfO2, are deposited to verify the effect of
anti-hydrolysis polymers on the long-term stability of high performance gas barrier.