Lorenzo
Bottiglieri
*a,
João
Resende
b,
Matthieu
Weber
a,
Odette
Chaix-Pluchery
a,
Carmen
Jiménez
a and
Jean-Luc
Deschanvres
a
aUniv. Grenoble Alpes, CNRS, Grenoble INP, LMGP, F-38000 Grenoble, France. E-mail: Lorenzo.bottiglieri@grenoble-inp.fr; Jean-luc.deschanvres@grenoble-inp.fr
bAlmaScience, Campus da Caparica, 2829-516 Caparica, Portugal
First published on 3rd June 2021
In this work, we report the enhancement of the functional properties of CuCrO2, a promising p-type transparent conductive oxide, achieved in out of stoichiometry CuCrO2 thin films synthesized by aerosol-assisted chemical vapor deposition. Out of stoichiometry films consisting of Cr-deficient CuCrO2, i.e. Cu-rich CuCrO2 phase, have a resistivity value of 0.05 Ω cm and an average transmittance of 58% in the visible range, resulting in a Gordon's figure of merit of 2200 μS. This is the highest ever published figure of merit among Cu-rich CuCrO2 films synthesized by chemical methods. A remarkable result is that when further increasing the Cu/(Cu + Cr) ratio, the formation of CuO was not detected, allowing the synthesis of composite films formed by Cu2O and CuCrO2 p-type oxides, which are more conductive than the Cu-rich CuCrO2 phase. These nanocomposite films present an improved carrier mobility, with a resistivity value of around 0.02 Ω cm, and a reduced energy gap, with a transmittance of 52%, resulting in a figure of merit of 1400 μS. Both these thin films can find applications as a hole transport layer in various transparent optoelectronic devices where p-type TCOs are required, especially when synthesized by a solution-based process at a low temperature and ambient pressure over large surface areas.
(1) |
The massive breakthrough of transparent electronics is hindered by the lack of p-type TCOs with properties comparable to those of their n-type counterparts, the latter being characterized by an optical transmittance of around 80% in the visible range and a resistivity value as low as 0.001 Ω cm leading to a FoMG value of more than 200 mS.2 The development of p-type TCOs with high FoMG values allows the fabrication of fully transparent p–n junctions, as previously reported for Li-doped ZnO/ZnO,3 and Sb-doped p-type SnO2/Sb-doped n-type SnO2,3,4 although these devices were fabricated by physical deposition methods working at a high temperature.
In the hot topic of the synthesis of highly performing p-type TCOs, copper-based delafossite oxides with the general formula Cu1+M3+O2 (M = Al,5 Fe,6 Ge,7 Ga,8 Cr,9 Y10 or Sc11) arise as promising materials since the first report on CuAlO25 was characterized by p-type conductivity and good transparency. Among these, copper chromium oxide, CuCrO2, emerges as an ideal p-type TCO with a band gap of around 3.1–3.3 eV,12 ensuring good transparency and a relatively good electrical resistivity (ρ = 1 Ω cm) for stoichiometric films.13 These optical and electrical properties make it suitable for a wide variety of optoelectronic devices, such as photovoltaic devices,14,15 light16 and gas sensors,17 transparent diodes18 and transistors.19 Furthermore, this material is also attractive for other peculiar features such as its magnetic properties,20 photocatalytic ability,21 and thermoelectric properties.22 CuCrO2 films can be synthesized through different techniques such as sputtering,23 chemical vapour deposition (CVD),9,24 spray pyrolysis,25,26 atomic layer deposition (ALD),13 sol gel,27 hydro-thermal synthesis,28 molecular-beam epitaxy29 and pulsed laser deposition (PLD).30 Among them, chemical methods allow the deposition over large surface areas at a relatively low temperature (<400 °C), compatible with glass and even plastic substrates. These features are extremely appealing for industrial applications.
Concerning the transport mechanisms in CuCrO2, a small polaron hopping among Cu1+/Cu2+ sites12 was suggested as the conduction mechanism. Theoretical calculations12 showed that the p-type conductivity is related to the formation of defects, such as Cu vacancies (VCu), which have the lowest formation energy as confirmed by the experimental reports.24 Furthermore, various studies30,31 highlighted the possible formation of Cu anti-site defects (CuCr), i.e. Cu atoms occupying Cr sites. Intrinsic dopants are demonstrated to be responsible for the enhancement of the charge carrier density and, consequently, of the electrical properties. The kind of defects formed in the material is believed to be dependent on the growth conditions and finally on the deviation from the stoichiometry of CuCrO2.24–26,32 For instance, the high conductivity of Cu-deficient CuCrO2 films is attributed to the presence of Cu vacancies as detected by using scanning transmission electron microscopy (STEM).33
Several published studies concern the characterization of Cu-poor CuCrO2 films. Crêpellière et al.24 reported the synthesis by the pulsed-injection metal–organic CVD (PI-MOCVD) of films with a copper content Cu/(Cu + Cr) of 33%, still preserving the crystalline delafossite phase. These films presented a resistivity value of around 0.06 Ω cm and an average transmittance of around 50% in the visible range, yielding the highest FoMG value of 2300 μS, to date, for this material. Lunca Popa et al.33 reported the lowest resistivity value of 0.009 Ω cm achieved for the same Cu-poor composition, Cu0.66Cr1.33O2, synthesized by direct liquid injection MOCVD (DLI-MOCVD). Farrell et al.25 reported a slightly lower FoMG value of 350 μS for Cu-deficient CuCrO2 films with a Cu/(Cu + Cr) of ≈ 30–35% and oxygen in excess (Cu0.4CrO2.5) by using a non-vacuum technique, the spray pyrolysis. They reported a resistivity value of around 0.08 Ω cm and an average transmittance of 55%.
On the other hand, Cu-rich CuCrO2 films, i.e. Cr-poor CuCrO2 (CuCr1−xO2), have also been reported as a promising p-type TCO. Ling et al.34 reported the synthesis of CuCr1−xO2 by the solid-state reaction at a high temperature, showing a variation of the resistivity of two orders of magnitude for x = 0.1 compared to the stoichiometric sample.34 They attributed this enhancement to hybridization between Cu 3d and O 2p orbitals and the presence of a mixed-valence band Cu1+/Cu2+. Sidik et al.30 reported the synthesis of CuCr1−xO2 by PLD with a resistivity value of 0.04 Ω cm and an average optical transmittance of 60% for x = 0.03. The enhancement of the conductivity with Cr deficiency was linked to the presence of Cu atoms in Cr-vacancies to form Cu anti-site defects. Nevertheless, it was also reported that a high Cu content leads to the formation of a parasitic CuO24,27 phase, detrimental for the optoelectronic properties of the film.
To date, there is no report on the enhancement of the functional properties of Cu-rich CuCrO2 thin films synthesized by aerosol-assisted metal–organic CVD (AA-MOCVD), a chemical method at a low temperature and ambient pressure, with no post-treatment. Therefore, the present study aims to evaluate the impact of the cationic ratio, Cu/(Cu + Cr), on the electrical and optical properties. In this work, Cu-rich CuCrO2 thin films are highlighted as the best compromise between conductivity and transparency. To our knowledge, this is the first report that Cu-rich CuCrO2 films, synthesized by a chemical method, display a figure of merit comparable to those of their Cu-poor counterparts. Furthermore, the film growth by AA-MOCVD creates an oxygen-poor environment that prevents the formation of the parasitic CuO phase. This allows the synthesis of composite materials formed by Cu2O and CuCrO2, displaying a conductivity higher than that of Cu-rich CuCrO2 thin films due to the enhancement of the carrier mobility. In the hot topic of the synthesis of efficient p-type TCOs, the present work offers two appealing candidates for transparent electronic devices, where p-type TCOs deposited at low temperature and on a large surface area are required.
AA-MOCVD was performed in a home-made vertical flux cold wall reactor as depicted in the study by de Oliveira et al.35 The deposition took place at atmospheric pressure ensured by an extracting system. A nitrogen trap was used for the condensation of the organic vapors after the reaction. Dry compressed air was used as the process gas with a total flow rate of 4800 sccm. The oxygen partial pressure was 0.21 × 105 Pa. The substrate temperature was 350 °C for all depositions. Two series of samples were synthesized. Series 1: the deposition time was set to 60 minutes and the solution consumption rate was fixed at 2 ml min−1. Two total molar concentration solutions of (Cu + Cr) of 10 mM and 20 mM with Cu/(Cu + Cr) varying between 40% and 100% were used. Series 2: after calibration of the deposition rate in series 1, 4 samples with a thickness in the range of 75–85 nm were synthesized by tuning the deposition time. The total molar concentration was fixed to 10 mM with the Cu/(Cu + Cr) values of 40%, 50%, 70% and 100%.
The composition of the films was analysed by using energy-dispersive X-ray spectroscopy (EDS) measurements with an energy beam of 15 KeV and using an Oxford Inca Energy detector in a FEI Quanta 250 field-emission scanning electron microscope (FESEM). The reported values are the average of 5 points over a surface area of 6.25 cm2. Two scanning electron microscopes (SEM) were used for the top-view and cross-section SEM observations. Series 1 was analysed using the FEI Quanta 250 FESEM microscope, and the samples of series 2 were observed using a FEG-ZEISS-Gemini 300 microscope.
X-ray diffraction (XRD) patterns were obtained using a Bruker D8 Advance diffractometer in the Bragg–Brentano configuration (θ–2θ) with Cu Kα1 radiation (λ = 0.15406 nm). Raman spectroscopy was performed at room temperature using a Jobin Yvon/Horiba LabRam spectrometer equipped with a liquid nitrogen-cooled charge-coupled detector. The excitation source was the 488 nm line of an Ar+ laser; it was focused to a spot size close to 1 μm2 by using a 100× objective. The laser power at the sample surface was around 80 μW.
High-resolution transmission electron microscopy (HRTEM) observations were performed using a JEOL 2010 LaB6 instrument operating at 200 kV with a 0.19 nm point-to-point resolution. The sample lamellae were prepared in cross-section by tripod mechanical polishing and Argon ion milling until the perforation of the interface. The superficial roughness was measured by using atomic force microscopy (AFM) performed using a Veeco D3100 AFM on 1 μm2 surface and the raw data were fitted by Gwydion software.
X-ray photoelectron spectroscopy (XPS) measurements were performed using a K-alpha spectrometer from Thermo Fisher Scientific with an Al Kα1,2 (1486.6 eV) X-ray source. The core levels of Cu, Cr and O in the film were probed in the Cu 2p, Cr 2p, and O 1s energy ranges after Ar milling at 2 KeV for 3 minutes. The experimental data were fitted with the Advantage software from Thermo Fisher Scientific. The valence band spectrum was acquired by using XPS (VB-XPS) in the constant analyser energy mode using a step size of 0.1 eV in the range of −1 to 10 eV. Previously, the samples were subjected to Ar milling under the same conditions listed before. The binding energy (BE) scale of the spectrometer was calibrated by the positions of the peaks of Au 4f7/2 (83.9 ± 0.1 eV) and Cu 2p3/2 (932.8 eV ± 0.1 eV) core levels of the pure gold and copper metals.
The optical properties of the films were investigated by UV–vis–IR spectroscopy, using a Lambda 950 spectrophotometer from PerkinElmer equipped with an integration sphere, using a wavelength step of 5 nm. The total reflection was measured on the same instrument using PTFE/BaSO4 and a black trap light for the 100% and 0% references, respectively. The electrical properties were measured by using a linear 4-probe system with a distance of 1 mm between the tips. The sheet resistance values correspond to the average of 5 points for each sample. The electronic transport properties were measured by the Hall effect at room temperature in a homemade setup with a magnetic field of 0.5 T.
The film thickness measured from the SEM cross-section images allows calculating the deposition rate for different compositions. The growth rate as a function of the film composition is reported in Fig. 1b for the two total molar concentrations used for these experiments. In these deposition conditions, it is evidenced that the Cu-rich films grow faster, and generally, a higher total molar concentration leads to the synthesis of thicker films but not at twice the thickness. When increasing the concentration of the Cu precursor, the higher growth rate suggests that the decomposition of Cr(acac)3 and its reaction with Cu and O are catalysed by Cu(acac)2 and its by-products, in agreement with the deposition of CuCrO2 by PI-MOCVD,24 where a similar trend was found. Moreover, these results are coherent with the growth mechanism proposed for CuCrO2 deposited by spray pyrolysis25 as the film formation occurs by the consecutive stacking of CrO6 octahedra and O–Cu–O dumbbells. Initially, Cr is adsorbed until the surface is saturated. Successively, Cu(acac)2 decomposes and reacts with the Cr terminated surface until the surface is adequately Cu covered. Finally, the Cr atoms react with the Cu terminated surface allowing the film growth. A lack of Cu inhibits the formation of O–Cu–O dumbbells and the subsequent film growth, explaining the increasing growth rate with the Cu content.
These analyses allowed us to control the composition and thickness of the films and were used to synthesize the samples of series 2. The films were grown from the solution concentrations of Cu/(Cu + Cr) of 40%, 50%, 70% and 100%. Several runs lead to film compositions corresponding to 50 ± 5%, 63 ± 5%, 77 ± 5% and 100% of Cu. These samples will be labelled as CuCrO2:X in the following, with X corresponding to the Cu content (in %) in the films as measured by EDS.
Raman spectroscopy was used on the same samples to confirm this hypothesis. Raman spectra are shown in Fig. 2b. Our stoichiometric and Cu-rich CuCrO2 samples are characterized by the three Raman modes at 101 cm−1, 460 cm−1, and 709 cm−1 (triangles in Fig. 2b(1) and (2)), assigned to the Eu, Eg, and A1g modes of CuCrO2, respectively, in agreement with the previous report on this material.38 Additional modes are attributed to the presence of defects able to relax the Raman selection rules. The first one labelled with * and observed in the range of 500–670 cm−1 is attributed to the presence of intrinsic defects.39 Its intensity increases with Cu content, suggesting a greater amount of defects for Cu-rich CuCrO2 when compared to the stoichiometric CuCrO2. The second one, at 770 cm−1, labelled as ■, has not been reported before and further investigations are required for its assignment. Cu2O modes are expected at 108 cm−1, 149 cm−1, 216 cm−1, 495 cm−1 and 649 cm−140 and the CuO modes at 297 cm−1, 347 cm−1, and 632 cm−1.41 The absence of the CuO and Cu2O Raman modes in these samples confirms that Cu-based oxides are either absent or amorphous and dispersed in CuCrO2 films, in agreement with the XRD results. The successful deposition of the crystallized non-stoichiometric CuCrO2 without any detectable secondary phase in the films of the Cu composition of up to 63% is achieved. Further results presented in the next section (Fig. 4c) confirm the presence of Cu2O in the films with a Cu/(Cu + Cr) of > 65%.
For the Cu2O + CuCrO2 composite film (Fig. 2b(3)), the Raman modes at 99 cm−1 is attributed to the Eu mode of the delafossite phase. Additional modes are assigned to Cu2O, as confirmed by the spectrum of the Cu2O film deposited under the same conditions (Fig. 2b(4)). The formation of Cr oxides is excluded due to the Cu-rich/Cr-poor environment used for the growth.
Thus, the classification of the samples of series 2 is well confirmed, allowing us to label as the stoichiometric CuCrO2, films with a Cu composition of ∼50%, Cu-rich CuCrO2 with a Cu composition of ∼65%, Cu2O + CuCrO2 with a Cu composition of ∼77%, and Cu2O.
The films corresponding to the out of stoichiometry composition of CuCr0.58OX (Cu/(Cu + Cr) = 63%) are characterized by a greater amount of defects in comparison with the stoichiometric compound. The preservation of the crystalline delafossite phase within a wide variation in the cationic ratio is consistent with previous reports on the Cu0.66Cr1.33O2 (Cu/(Cu + Cr) of 33%) by DLI-MOCVD42 and PI-MOCVD.24 Moreover, it is important to note that CuO is not detected in any of our films, despite Cu2+ being present in the starting precursor. As described in the work by Lim et al.,43 the gas phase decomposition of Cu(acac)2 results in the formation of metallic copper condensing on the substrate. Then, Cu0 oxidizes to Cu2O. When the oxygen partial pressure is too high during the growth, Cu2O undergoes a secondary oxidation reaction to form CuO. The direct oxidation of Cu0 to Cu2+ is excluded because the Cu oxidation state transforms successively from Cu0 to Cu1+ and then to Cu2+.44 Thus, the absence of Cu0 and CuO in our films demonstrates that the used oxygen partial pressure is adequately high to completely oxidize Cu0 to Cu1+, while still sufficiently low to prevent the formation of CuO. The reaction with Cr can also be a stabilising factor for Cu1+. Furthermore, the deposition of a composite material formed solely by Cu2O and CuCrO2 is in agreement with the phase diagram of the Cu–Cr–O system under these deposition conditions.45 As shown by the thermodynamic results, the use of a low oxygen partial pressure allows the coexistence of Cu2O and CuCrO2 phases, preventing the formation of CuO, which would be detrimental for the synthesis of transparent conductive materials.
Cross-section images in insets clearly show the similar thickness of the three films and a variation of the superficial roughness with the Cu/(Cu + Cr) cationic ratio, with a maximum roughness for Cu-rich CuCrO2. The AFM measurements confirmed the trend of the roughness with values of root mean square roughness of 0.99 ± 0.15 nm, 2.08 ± 0.35 nm and 1.42 ± 0.18 nm for CuCrO2:59%, CuCrO2:65% and Cu2O + CuCrO2:77% films, respectively.
The microstructure of two Cu-rich films, CuCrO2:59% and CuCrO2:65%, and of the Cu2O + CuCrO2:73% composite film was analysed using TEM. HRTEM images and SAED patterns are presented in Fig. 4. The Cu-rich CuCrO2 thin films present a nanocolumnar microstructure with a broad orientation distribution, as reported previously by our group.16 The increase of the Cu content leads to the growth of wider grains with higher vertical alignment, as visible by the comparison between CuCrO2:59% (Fig. 4a) and CuCrO2:65% (Fig. 4b). These results are in good agreement with SEM observations, as the growth of bigger nanocolumnar grains is likely related to the increase in the superficial roughness for the Cu-rich films. SAED patterns confirm that both films crystallize in the CuCrO2 structure without any detectable secondary phase, in agreement with the Raman and XRD results. This result corroborates that there is a wide compositional range, up to Cu/(Cu + Cr) = 65%, in which the crystalline delafossite phase is preserved despite the variation of stoichiometry.
For the Cu2O + CuCrO2:73% nanocomposite film (Fig. 4c), the formation of the secondary Cu2O phase results in more rounded grains. The film is composed of both Cu2O and CuCrO2 as confirmed by SAED, which is consistent with the Raman and XRD results.
Fig. 5 XPS spectra in the (a) Cu 2p3/2 and (b) Cr 2p3/2 orbitals energy ranges for CuCrO2:59%, CuCrO2:65% and Cu2O + CuCrO2 (Cu/(Cu + Cr) = 73%). The spectra were shifted for clarity. |
The Cr 2p3/2 core level (Fig. 5b) is centred at a binding energy of 576 eV corresponding to Cr3+ in agreement with the previous report on CuCrO2.24 Nevertheless, the distinction among the Cr species is complicated due to the presence of the Cu Auger peak, Cu LMM, centred at 569.9 eV. Again, the position of this Auger peak suggests the presence of Cu1+ rather than the metallic Cu,46 which should be centred at 568.3 eV.47 These results indicate that Cu and Cr are mainly in the 1+ and 3+ oxidation states, respectively, despite the stoichiometry variation. The O 1s peak (not shown here) was centred at around 530 eV.
The existence of Cu2+ does not necessarily imply the formation of CuO, not detected in our films, because it has been reported that Cu2+ can be present in the planar triangular network of the lattice, balanced by oxygen interstitial.18 This is in agreement with the results obtained by the combustion synthesis of CuCrO2+x,18 where the pure delafossite phase was detected despite the higher amount of Cu2+ species (37% Cu1+vs. 63% Cu2+). The presence of Cu2+ in our samples can be induced by the Cr deficiency, leading to an excess of holes at the Cu sites. This is supported by Ling et al.34 in the case of films obtained by the solid state reaction due to the fact that the Cu2+ content increases when the Cr content decreases. In our case, because of the O-poor environment used during the growth, the formation of Cu2+ is not favoured.
Table 1 presents the contribution in % of Cu1+ and Cu2+ to the Cu 2p3/2 spectra. The FWHM of the fitted peaks is also included in brackets. This procedure leads to a maximum value of the Cu1+/Cu2+ ratio of CuCrO2:65%. As proposed by Ling et al.,34 this fact suggests a favourable electronic structure with an improvement of hybridization between Cu 3d and O 2p orbitals. A more recent work by Singh et al.48 confirmed that the conduction mechanism takes place among Cu1+–O–Cu2+ sites, rather than directly between Cu1+ and Cu2+ sites. We can speculate that in both cases, with the conduction taking place via direct Cu1+–Cu2+ or via Cu1+–O–Cu2+ sites, the maximization of the Cu1+/Cu2+ ratio is required to increase the number of hopping sites and, thus, the film conductivity.
Cu/(Cu + Cr) in the film by EDS (%) | Cu1+ (%) (FWHM) | Cu2+ (%) (FWHM) |
---|---|---|
59 | 80 (1.26 eV) | 20 (2.05 eV) |
65 | 87 (1.25 eV) | 13 (2.05 eV) |
73 | 79 (1.22 eV) | 21 (2.04 eV) |
VB-XPS was performed on the Cu-rich film with a Cu/(Cu + Cr) of 65% (Fig. S2 in the ESI†) and was compared with the curve resulting from the band theory, adapted by the work of Yokobori et al.49 The theoretical spectra shows that the maximum of the valence band is dominated by the Cr 3d states (labelled with α) with a contribution attributed to the Cu 3d states (indicated with β), responsible for the peak with the highest intensity. Additional contributions at higher binding energies (labelled with γ) are attributed to the O 2p orbital. We can distinguish the 3 contributions in the experimental spectra, although the resolution of the spectrometer was too low to obtain additional information. The reduced intensity of the Cr states compared to that of the Cu ones can be attributed to the lower cross-section of the trivalent cation compared to Cu.50 However, we may speculate that the variation in intensities can also be linked to the Cu abundance/Cr deficiency in our films, leading to the limited presence of the shoulder of the Cr 3d states at about 1.5 eV. Consequently, we can assert that the Fermi level remains close to the top of the valence band, in agreement with the high charge density of delafossite oxides and with the previous report of VB-XPS for this compound.29 The use of a higher resolution XPS spectrometer will be essential to obtain a proper quantification of the energy difference between the Fermi level and the valence band edge.
Electrical measurements during annealing up to 200 °C under N2/O2 with 20% O2 (not shown here) were performed on a 200 nm thick sample with a Cu/(Cu + Cr) of 61%. The resistivity variation with the temperature was fitted using the Arrhenius equation, leading to an activation energy of 76 meV. This value ranges around 3 KbT, highlighting a highly degenerate semiconductor. This result is in good agreement with the proximity of the Fermi level with the top of the valence band as confirmed by our VB-XPS spectra.
Fig. 6 Variation of the resistivity with the cationic ratio in the out of stoichiometry CuCrO2 thin films. The values corresponding to the samples of series 2 are indicated by the arrows. |
The low conductivity of samples in region 1 can be attributed to their low crystallinity of the films for this composition. Although this is only a qualitative explanation, the enhancement of the conductivity of two orders of magnitude for samples with a composition in the 50–65% range can be confirmed through simultaneous factors as established by Chen et al.51 On the one hand, the increase in the cationic ratio can result in the formation of defects such as Cu vacancy and Cu antisite that have a lower formation energy than others.52 In our Cu-rich/Cr-deficient conditions, we infer that the additional Cu atoms might be placed in the Cr vacancies to form Cu antisite defects (CuCr). This will be balanced by the generation of holes, leading to an increase of the charge carrier density in the films,51 thus resulting in a lower resistivity of the film. Other defect species can play a major role in the electrical properties.
In region 2, the films within the 60–65% compositional range show an improved crystallinity as indicated by the higher intensity of the (012) reflection in XRD and the growth of bigger grains, as observed by SEM and TEM. This is congruent with the work by Sidik et al.30 where Cr deficiency was demonstrated to enhance the crystallinity of the films. Furthermore, as stated before, the enhancement of the conductivity can be attributed to the favourable electronic structure of these films through the maximization of the Cu1+/Cu2+ ratio, as proved by XPS. Nevertheless, the lowest resistivity value of 0.02 Ω cm was achieved in region 3 for the Cu2O + CuCrO2 composite film with 67% of Cu. Increasing the Cu content to above 65% favours, in our deposition conditions, the formation of the Cu2O phase in the delafossite phase that is more conductive, thus increasing the electrical resistivity up to the values obtained for the pure Cu2O. This increase can be attributed to the change in grain shape and degradation of the film crystallinity. In comparison, the films deposited by PI-MOCVD at low pressure with a similar composition resulted in the synthesis of CuO + Cu2O + CuCrO2,24 with a resistivity higher than that of the delafossite phase out of stoichiometry. This finally demonstrates that the fact to prevent the formation of CuO, as reported in our study, is promising for the synthesis of highly conductive films.
Out of stoichiometry CuCrO2 films with no detectable secondary phase are characterized by a low carrier mobility, being under the detection limit of our experimental setup. This allows setting an upper limit of 0.1 cm2 V−1 s−1 for these films. The Hall effect measurements were successfully performed for Cu2O + CuCrO2:73% and pure Cu2O, confirming the p-type behaviour of these films. The mobility values of 0.65 cm2 V−1 s−1 and 5 cm2 V−1 s−1 were obtained for these two samples, respectively. The corresponding charge carrier density values were 9.3 × 1018 cm−3 and 1.5 × 1016 cm−3, respectively. These results reveal an improvement in mobility by the formation of Cu2O. This finding allowed us to conclude that the lowest resistivity achieved for Cu2O + CuCrO2 with 67% of Cu is due to the combination of the good mobility of Cu2O and the high charge density achieved for CuCrO2.
Despite thickness variations, the FoMG value was calculated from eqn (1) to compare the properties of our films with the literature; the results are reported in Fig. 7d. The FoMG values from the literature were estimated with a 0% reflectivity when this value is not reported in the study. The AA-MOCVD Cu-rich CuCrO2:65% films showed a promising FoMG value of 2200 μS, comparable with the value of 2300 μS, the highest for this material, obtained by the Cu-deficient CuCrO2 films deposited by PI-MOCVD.24 The AA-MOCVD growth method presents many advantages as it is performed at atmospheric pressure and at lower temperature than the PI-MOCVD method. Besides, the Cu2O + CuCrO2:67% composite films with higher conductivity achieved a good FoMG value of 1400 μS, despite a reduction in their optical performances.
The deposition of the out of stoichiometry CuCrO2 films without any detectable secondary phase is achieved up to a composition of Cu/(Cu + Cr) of 65%. Single-phase Cu-rich CuCrO2 films present a nanocolumnar microstructure with a resistivity value lower than 0.1 Ω cm, and a wide band gap of around 3.1–3.2 eV. The optimal composition was found for Cu-rich CuCrO2:65% with a resistivity value of 0.05 Ω cm and an average transmittance of 55% in the 400–800 nm range, resulting in a high p-type FoMG value of 2200 μS. The enhancement of p-type conductivity is attributed to the formation of the intrinsic defects with Cu, occupying the Cr vacancies and creating Cu antisite defects, and to the further improvement of the film crystallinity.
In summary, we highlight the successful deposition of the Cr- deficient CuCrO2 by AA-MOCVD with a figure of merit comparable with the one of the Cu-deficient CuCrO2 films. Besides, the growth in an oxygen-poor environment has been demonstrated to impede the formation of CuO, allowing the synthesis of nanocomposite films composed exclusively of Cu2O and CuCrO2. These films are characterized by a carrier mobility of 0.65 cm2 V−1 s−1, higher than the Cu-rich CuCrO2 films. In this study, the lowest resistivity value of 0.02 Ω cm was achieved for the Cu2O + CuCrO2 films with 67% of Cu. Despite a reduced average transmittance of 52%, the performance of the nanocomposite films led to a FoMG value of 1400 μS.
As revealed in the present work, the modulation of the optical and electrical properties, such as mobility, resistivity, total transmittance, and band gap, is achieved by tuning the stoichiometry of the films. This allowed us to synthesize two appealing candidates for hole transport layers in the thin film, organic, and perovskite solar cells, as well as for organic light-emitted diodes (OLEDs) and thin film transistors (TFTs) as these Cu-rich CuCrO2 thin films can be incorporated into these devices.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1ma00156f |
This journal is © The Royal Society of Chemistry 2021 |