Solution processing of n-type doped ZnO

Abstract

After setting the framework of this manuscript by a overview of the available literature it appears that ZnO is a very attractive and cheap material as transparent conductive oxide. The conductivity can even further be improved by n-type doping. (p-type doping for ZnO is up to today not possible.) Generally trivalent elements such Al, Ga and In are used for n-type doping, although Cl-doping of ZnO is even done, but only by electro-deposition. Within this invesitgated field of solution processing, different pathways can be distinguished: a nanoparticulate route, molecular precursors or direct synthesis. Although direct synthesis through chemical bath and electro-deposition is able to obtain Al-doped ZnO and Cl-doped ZnO at low temperature (< 100 ℃) this method is neglected.[111, 227] From an upscaling point of view, this method is quite cumbersome due to the need of large vessels of reagents and the production of a lot of waste. Therefore the focus is put on nanoparticles and molecular precursors, which can be directly applied in more divers deposition techniques. The versatility of these routes is more widely spread. Although the deposition technique of choose here is spin coating, which is easy on a laboratory scale, the suspensions/solutions could be even applied in spray coating or ink-jet printing, provided with adaptations towards viscosity. In the first part different synthesis routes for Al-doped ZnO nanoparticles have been explored from co-precipitation to thermal decomposition synthesis. Although each method certainly has its advantages, some shortcomings need some additional care. Co-precipitation is able to synthesize nanoparticles with a high yield. Nevertheless the morphology obtained here is not so clear: no real spherical particles are observed, with here and there an exception. For the incorporation of Al inside the lattice, different atmospheres during calcination could steer Al allocation in a minimalistic way. The calcination in nitrogen seems to be able to push the Al more in the substitutional tetrahedral place, although no real high values have been obtained. None of this favors the use of these nanoparticles. It might be of interest to precipitate the nanoparticles under elevated temperatures, giving the dopant more energy to be substitionial tetrahedral incorporated. Further on, the thermal decomposition of Al-doped ZnO nanoparticles is studied in-depth. Either dibenzyl ether or benzyl amine was used as a solvent for the dissolution of Zn and Al acetylacetonates. In dibenzyl ether, an in-situ stabilization method was developed where surfactants are used to tune the size and shape during synthesis. The mechanism from reagents to Al-doped ZnO nanoparticles is elaborated, showing the versatility of obtaining doped oxides through this method. After synthesis, the used ligands are chemisorbed to the nanoparticle surface to ensure steric stabilization of nanoparticles. In that manner a stable apolar suspension in hexane (or other apolar solvents) can be obtained, which can be used directly for deposition. In the end, a conductive AZO film was obtained from these building blocks. Nevertheless, still elevated temperatures are needed to remove the ligands. Further on, increased temperatures might be needed to induce grain growth and significantly reduce grain boundary scattering. Although these ligands can be exchanged to electrostatic stabilization, which can be removed at lower temperature, the questions asked are if this is sufficient to ensure conductivity throughout the films. So far every effort towards low temperature conductive AZO films has been unsuccessful. The thermal decompostion in benzyl amine generated a high amount of rod-like nanoparticles. The obtained suspension in ethylene glycol was stable without any additives. The nanorods were intensively studied by 27Al MAS NMR analysis and the largest fraction of substitutional tetrahedral Al allocation was found. This, together with the morphology, make these NRs the ideal building blocks for TCO films. Nevertheless after deposition and impregnation with a molecular precursor, that served as a glue, no improved conductivity (in correlation to pure molecular precursors) was obtained. Probably the grain boundary scattering is the dominating effect and charge carriers, which should be present in high quantities, can’t be extracted from the NRs nor are mobilized through the film. Although stable suspensions could be obtained, additional high temperature treatments are necessary each time to introduce measurable conductivity. Sintering of nanoparticles is very important and can only happen close to the melting point of the material. To reduce this temperature, very small particles are necessary. Whether these nanoparticles have a future to obtain good conductive TCOs is really the question and could be resolved by exploring further low temperature sintering methods. Laser sintering is often put forward in literature as a low temperature sintering method, although simulations show the development of local high temperatures. If low temperature sintering is really possible needs to be investigated by future work. Other sintering methods such as combustion processing might also improved the conductivity of films from NPs. A key point in figuring this out might be resolving how and why chemical bath deposition of Al-doped ZnO and electro-deposition of Cl-doped ZnO is able to obtain conductive layers.[111, 227] Nevertheless a lot of new funda mental insights in the chemistry behind NP synthesis have been presented and could improve the way of synthesizing building blocks for TCOs. The second part focused on molecular or sol(ution)-gel precursors, which are obtained by the dissolution of proper metal ion containing reagents. A holistic approach has been taken to synthesis n-type doped ZnO and to understand the fundamentals behind synthesis and the functional characteristics. Hereby the focus was mainly put on Al-doped ZnO since this is to most attractive way of doping ZnO. Not only is Al abundantly available it is non-toxic and cheap. Further on the characteristics of Al-doped ZnO exceed the conductivity and transparency of In and Ga-doped ZnO in most cases. First of all, the focus has been put on organic precursors. Butoxyethanol based precursors have been developed, which have proven to be a good alternative for teratogenic solvents such as methoxyethanol. A minimal sheet resistance of only 57 Ohm/ is obtained. In addition, it has been obvious that the TCO layer needs to have a minimal thickness to avoid surface effects. This thickness is strongly correlated to the molarity of the applied precursor and resulted in an improved conductivity for lower precursor concentrations. Alternatively, aqueous citrato precursors were developed to obtain n-type doped ZnO films. Al, In, Ga and Cl have been tested. In the end a 2% ZnO:Al film results in the lowest obtained resistivity of 1.77 10−3 Ohm cm with the highest transparency within the visible range after an anneal at 450 ℃ in a 5 % H2 / 95 % Ar atmosphere. The good characteristics of Al-doped ZnO (after RA), together with the abundant availability, non-toxicity and the reduced price of both Al and Zn, make AZO the ideal candidate as TCO. In addition synthesis of very good conductive ZnO with a sheet resistance as low as 676 Ohm and a resistivity of 4.93 10−3 Ohm cm is achieved. Nevertheless, through the reduced transparency the presence of metallic Zn can be hypothesized. Although In and Cl-doping don’t lead to an improved conductivity in comparison to ZnO, the transparency is significantly better. Nevertheless, as Cl-doping leads to a reduced thickness of the ZnO:Cl layer, the presence of Cl inside the layer can be questioned. Probably this method, due to the reductive treatment, is nonideal to guarantee Cl-doping. Also Ga-doping improves the conductivity and the transparency but Al-doping is certainly the most elegant dopant of aqueous chemical solution deposition method. Because of the establishment of Al-doping to be the most adequate aqueous dopant, the mechanism behind the conductivity was investigated. The key to obtaining a competitive resistivity, which is 10−3 Ohm cm in the order of typical sputter and solution processed films (though still higher than the best sputter deposited films), is the application of a reductive anneal at moderate temperatures in a hydrogen containing atmosphere. Therefore, the effect of the reductive anneal was investigated and the following in depth understanding was obtained. The temperature of the reductive anneal plays a role in solution processing of ZnO and can easily remove the complete layer when it is too high or the treatment takes too long. Depending on the dopant and on the equipment/ inert gas, this temperature can vary. This is achieved through oxygen removal through reaction with hydrogen of the RA. Although there are a lot of controversial thoughts on the role of oxygen in solution processed (doped) ZnO, it is without doubt that oxygen has an influence. The reductive anneal leads to the desorption of oxygen and hole formation and activates the Al-doping by a shift of octahedral to tetrahedral coordination, also a small fraction of penta-coordinated Al is observed. Additionally the observed oxygen species at the grain boundaries and surface are removed. This surface has a big influence on the resistivity as observed before, but for denser layers a thinner layer is sufficient to counter these effects. In addition to the Al reorganization, a H reservoir was found after RA, which was larger for Al-doped films. As oxygen is being removed during RA, metallic Zn is left behind. This can easily vaporize at high temperatures, although for pure ZnO (treated at lower temperature) this is hypothesized to attribute to the improved conductivity. For doped films, it has not been found so far. Good transparent and conductive films have been obtained through (aqueous) solution processing, from which the organic precursor seem to produce more stable films in function of time. For the aqueous deposited films this has been monitored in function of time. In function of time it can be stated that, although charge carriers are still present after one year shelf life, the negative effects overtake the mechanism empowering the conductivity and destroy the mobility in the system. The exact origin of the reduced conductivity is expected to be a coherence of several factors. Shifts in Al coordination, refill of oxygen vacancies and the lifting of the passivation of defects are linked to this decreasing conductivity. Nevertheless, all these molecular precursors still needs high thermal treatments to obtain good conductive layers. As an alternative a low temperature decomposing precursor in 2-methoxyethanol was developed with metal acetylacetonates and ammonium nitrate. Insights in the thermal decomposition of combustion chemistry and the optimization of the morphology have rendered conductive In-doped ZnO layers at a maximum temperature of only 240 ℃. Nevertheless the conductivity decayed easily and requested further improvement in morphology. The applied temperature of the hot plate, the molarity and the use of dopants generates significantly different looking films. Although no conductive Al-doped ZnO films could be obtained, general insights in the combustion chemistry are presented to improve further investigations. Further explorations on the deposition temperature, with maybe UV treatment, could render conductive layers. Nevertheless, this should be further investigated. In addition the development of low temperature decomposing precursors was further explored for aqueous precursors. The citrato precursors is able to decompose at low temperature, either by UV or by adaptation through combustion processing with NH4NO3 . Alternatively a glycine based combustion precursor was developed either with or without additional NH4NO3 . Although, this extra oxidizer is not necessary for a low temperature decomposition according to thermogravimetric analysis and it even destroys the wettability in a multi-layer deposition process like spin coating. By saturating the one with NH4NO3 (and the precursor with only glycine) with additional (Al-doped) ZnO powders, multiple layers can be deposited. In addition to a macroscopically better morphology, microscopically still improvements are necessary to have a fully closed layer. Nevertheless if NH4NO3 is present inside the saturated precursors, crystalline ZnO can be detected in the film, indicating an improved nucleation. Saturating the precursor as well as NH4NO3 results in improved characteristics. Alternatively, Weber et al. published an aqueous based method to deposit on low temperature crystalline In and Ga-doped ZnO. [288] This method seems promising and is of interest to further explore as a facile route for low temperature deposition of Al-doped ZnO as TCO in PV devices. The obtained insights for high temperature processed conductive AZO currently have not contributed to the understanding of the non-conductive low temperature processed AZO, but a first step is set towards fundamental investigations. In the future, this study will be able to contribute to the optimization of low temperature processed conductive AZO and can be compared to recent publications [288] [111] [227] on low temperature deposition of crystalline ntype doped ZnO, which may or may not render conductive films.