High temperature preparation
First of all, as seen on Figure 4.8, the samples of the 1stdomain are exclusively
samples which were prepared at high temperature (> 200 °C). As discussed in Chapter 10, a high temperature of deposition favours a more crystalline and more homogeneous thin film structure.
The first domain is characterized by a high valence band maxima, which tend to decrease with increasing O 1s(Def.) weight (Figure 4.8). It means, following Fermi statistics, the concentration of holes in the valence band maximum increases. Also, higher O 1s(Def.) weights are obtained at high oxygen concentration during thin film fabrication. This would support the assumption that the thin films are intrinsically doped by an excess of oxygen. Thus, the experiments suggest that in the first domain higher hole concentrations are attained with increasing O 1s(Def.) weight. Thus charge compensation of the dopant, formed at high oxygen concentration, is realized through the formation of holes.
As explained in the introduction, the dopant can be either a nickel vacancy or an oxygen interstitial. In bulk NiO, nickel vacancies are the only dopant but, as no detailed literature treated this case yet, it is assumed that oxygen interstitials can be present at grain boundaries. Chapter 5 will highlight that the grain boundaries accumulate oxygen species if the thin film is prepared at room temperature, but the temperature erases the oxygen-rich
grain boundaries and might convert the thin film to a more stoichiometric crystallographic structure. Therefore, we assert that holes in the valence band of the NiO thin films prepared at high temperature originate from nickel vacancies and not from oxygen interstitial.
Meanwhile, the Ni 2p region is not particularly affected by the increase of the O 1s(Def.) weight as the RN i ratio is constant as seen in Figure
4.8. Supported by the fact that the difference of spectra of the samples having an O 1s(Def.) weight of 2.5% and 4.9 % with the reference sample (O 1s(Def.)=0%) do not evidence an obvious change in the electronic structure in the Ni 2p region (Figure 4.2, right), that the cyclic voltammetry realized on a sample prepared at high temperature discards the presence of Ni3+ on the
surface (see Section 4.3.5) and finally that the low workfunction value of the samples from the 1st domain indicates that the formation of Ni3+ or O− is
limited (see Section 4.3.4). Therefore, it is assumed that the NiO thin films produced at high temperature do not contain the charge compensating species Ni3+.
It has to be mentionned that at the minimum, in the 1st domain, the
Fermi energy is 0.6 eV when determined by linear extrapolation (or 0.35 eV in taking the threshold where no photo-emission is detected, Figure 4.3) for the samples from the 1st domain. This means that the ratio h /Nv, which
can be determined from equation 2.1.2 is at best 10−6-10−3. Therefore, the holes/oxygen interaction cannot be detected by in-situ XPS (the accuracy of the XPS is at least 1%). Thus the O 1s(Def.) cannot be related to holes but could be associated with peroxo species present in small quantity. Thus, it is assumed that the O 1s(Def.) observed for sample prepared at high oxygen concentration and high temperature of deposition might be related to the formation of a peroxo species which would accumulate primarily at the grain boundaries (see Chapter 5).
Moreover, as detailed in Section 2.2.1, NiO is a charge transfer material. It means that charges are compensated through Ni2+/O2− interaction. As
a nickel vacancy (V00Ni) replaces a lattice nickel atom, the charges of nickel
vacancies might also be compensated by V00Ni/O2− interactions. Therefore,
it would make more sense that missing nickel atoms are compensated by positive charges localized on the surrounding oxygen atoms. Thus, the negatively charged nickel vacancies V00Niwould interact electronically with the
six surrounding oxygen atoms and so charge compensation by holes would be primarily localized on the oxygen atoms.
As the Fermi energy decreases with increasing O 1s(Def.), and as it is very unlikely that Ni3+ can be created in the samples prepared at high
temperature, we propose that, for the first domain, the compensation of nickel vacancies is realized by the formation of holes:
Ni×Ni+ O×O+
1
2O2→ NiO + O
×
and also by peroxo species:
Ni×Ni+ O×O+ O2→ NiO + 2 OO+ V00Ni (4.6)
Room temperature deposition
The 2nd domain (Figure 4.8) is characterized by an almost constant Fermi
level position, an increasing RN i ratio and a high workfunction. Although the
O 1s(Def.) increases with oxygen concentration during sample preparation, the almost stable Fermi level position would suggest the addition of defects does not necessarily create more free charge carriers. A convincing demonstration that holes are not created when the NiO thin film is prepared at room temperature is given in Chapter 5.
Contrary to samples prepared at high temperature, samples prepared at room temperature of the 2nd domain are highly defective where an oxygen- rich secondary phase accumulate at the grain boundaries (see Chapter 5). Therefore, also with respect to what has been written in the previous part, the implication of oxygen interstitials in the non-stoichiometry of these sample cannot be totally discarded.
Regarding the charge compensating species in the 2nd domain, it can be
noticed that the RN i ratio increases with increasing O 1s(Def.) weight. In
particular, the differences of the Ni 2p XP spectra with the Ni 2p reference spectra (O 1s(Def.)=0 %) evidence an additional electronic component (Figure 4.2). The additional component might be attributed to Ni3+. Indeed, the XPS
spectra of NiOOH, where nickel is formally in the +III state, as for lithium doped NiO, provides a more prominent Ni 2p(Sat.) peak in comparison to the Ni 2p(Main) peak [173, 174]. Thus we attribute the swelling of the Ni 2p(Sat.) peak with increasing oxygen concentration in the process for samples prepared at room temperature, as a fingerprint of the presence of Ni3+ species.
This might be in line with the electrochemical measurements (Figure 4.5) and, to a lesser extent, with the optical measurements (Figure 4.6), which would suggest that triply ionized nickel atoms (Ni3+) are present on
the surface and in the bulk of the thin films prepared at room temperature. Also the high workfunction value reported for the samples of the second domain (Figure 4.8) can be related to both Ni3+ and O− species (see Section
4.3.4). Finally, as detailed in Chapter 5, the thin films prepared at room temperature would accumulate oxygen species at the grain boundaries that would eventually lead to the formation of nickel peroxide NiO2. An XAS
study on NiO thin films prepared at room temperature suggested that charge compensation of dopants is dominantly carried by peroxo species (O−) and to a lesser extent to Ni3+ [83]. Therefore, we propose the charge compensation
mechanism in RT-NiO thin films to be: 5 Ni×Ni+ 4 O×O+ O2→ N iO + 4 (1 − δ)(N i•N i+ O × O) + 4 δ(O • O+ N i × N i) + V 00 N i+ O 00 i (4.7)
with δ a parameter comprised between 0 and 1 to take into account charge compensation discrepancy over oxygen and nickel atoms.
Looking at the O 1s region for the samples of the second domain, the deconvolution reveals that the O 1s(Def.) shoulder is composed of at least two electronics states (Figure 4.1). As the Ni 2p spectra, the electrochemical measurements and the high workfunction value suggest the presence of Ni3+ in the NiO thin films prepared at room temperature, it is assumed that the additional O 1s(Def.) component at high binding energy, for the samples of the second domain, is indirectly related to the presence of Ni3+. In crystalline
structure, the emergences of Ni3+ is very unlikely as the Ni 3d7 state is too
deep to be raised to the Fermi energy. Thus, we assume that the presence of the Ni3+ is rendered possible due to the amorphousness of the NiO thin films
prepared at room temperature.
To summarize, for the samples from the second domain, the charge carrier concentration (h+) does not necessarily increase with increasing number
of nickel vacancies and oxygen interstitials, but instead the results might suggest that nickel vacancies and oxygen interstitials are compensated by the formation of peroxo species (O−) and by the oxidation of nickel atoms into Ni3+.