A Low Temperature Growth of Cu2O Thin Films as Hole Transporting Material for Perovskite Solar Cells

Copper oxide thin films have been successfully synthesized through a metal–organic chemical vapor deposition (MOCVD) approach starting from the copper bis(2,2,6,6-tetramethyl-3,5-heptanedionate), Cu(tmhd)2, complex. Operative conditions of fabrication strongly affect both the composition and morphologies of the copper oxide thin films. The deposition temperature has been accurately monitored in order to stabilize and to produce, selectively and reproducibly, the two phases of cuprite Cu2O and/or tenorite CuO. The present approach has the advantages of being industrially appealing, reliable, and fast for the production of thin films over large areas with fine control of both composition and surface uniformity. Moreover, the methylammonium lead iodide (MAPI) active layer has been successfully deposited on the ITO/Cu2O substrate by the Low Vacuum Proximity Space Effusion (LV-PSE) technique. X-ray diffraction (XRD), field emission scanning electron microscopy (FE-SEM), and atomic force microscopy (AFM) analyses have been used to characterize the deposited films. The optical band gap (Eg), ranging from 1.99 to 2.41 eV, has been determined through UV-vis analysis, while the electrical measurements allowed to establish the p-type conductivity behavior of the deposited Cu2O thin films with resistivities from 31 to 83 Ω cm and carrier concentration in the order of 1.5–2.8 × 1016 cm−3. These results pave the way for potential applications of the present system as a hole transporting layer combined with a perovskite active layer in emergent solar cell technologies.


Introduction
Recently, copper-oxide compounds represent one of the most studied classes of semiconducting materials. The main advantages of these materials are related to the exceptional possibility of tuning the optical and electronic properties within their semiconducting behavior [1,2]. Copper oxide-based materials play a significant role in many technological applications, ranging from sensing [3] to catalysis [4], from photodetector (e.g., in combination with ZnO) [5,6] to the photo-electrochemical splitting of water [7,8]. In recent years, copper-based materials in the form of thin films have become objects of interest also in solar cell devices [9][10][11].
Indeed, together with the other classes of p-type layers applied in solar cell technology, organic materials such as PEDOT and spiro-OMeTAD [12] or 2,4,6-triarylpyridine [13], and transition metal oxides such as MoO 3 , V 2 O 5 , WO 3 , NiO and Cu 2 O have been explored as efficient hole transporting layers [14,15]. In particular, among the three most common and stable phases, Cu 2 O, Cu 4 O 3 , and CuO, named cuprite, paramelaconite and tenorite, respectively, the cuprite phase (Cu 2 O) has been intensively studied and applied as a p-type semiconducting material in solar cell devices [9][10][11]16,17]. Among the advantages of Cu 2 O with respect to the organic hole transporting layers are efficient hole collection, optical transparency, stability, a wide variety of synthetic strategies and low-cost production. Besides, Cu 2 O is environmentally stable and its main component, Cu, is earth-abundant and non-toxic. Other interesting Cu-based hole-transport materials, such as CuI or CuSCN layers, have been applied to inverted planar perovskite solar cells [18,19].
Cu 2 O as a thin film layer has been synthesized by several methods, including vapor and solution approaches. The most studied approaches are solution-phase methods [20,21], hydrothermal [22], electrodeposition [23], sputtering [24], molecular beam epitaxy [25], atomic layer deposition [26], thermal oxidation of copper [27], and chemical vapor deposition [28,29]. Among these approaches, metalorganic chemical vapor deposition (MOCVD) represents one of the most promising techniques due to its tunability processes, easy scaling up and industrial applicability. Moreover, the CVD approach offers the possibility of fine-tuning the composition of the Cu-O phases by easily altering the operating conditions and the chemical nature of the precursors [30].
To the best of our knowledge, only a few works in the literature report the thermal-CVD fabrication of pure Cu 2 O thin films without the use of H 2 flow as a reducing agent at relatively low temperatures. Maruyama  In the present work, we propose an in-depth study of the metalorganic chemical vapor deposition process for the reproducible and selective fabrication of both cuprite Cu 2 O and tenorite CuO copper oxide thin films starting from a β-diketonate copper (II) precursor, i.e., the bis(2,2,6,6-tetramethyl-3,5-heptanedionate) copper, Cu(tmhd) 2 . The MOCVD process has been tested in the temperature range of 250-400 • C, allowing the selective and reproducible fabrication of Cu 2 O on a large area at the lowest temperature of 250 • C, and a mixture of Cu 2 O-CuO or the pure CuO at higher temperatures. A sequential deposition of the methylammonium lead iodide (MAPI) layer allowed us to test copper oxide as the substrate for the vapor deposition of MAPbI 3 film to realize the first part of a solar cell with a planar inverted structure. The present approach represents a simple, easily scalable, and industrially appealing process for the production of compact and homogeneous copper oxide films at relatively low temperatures. X-ray diffraction (XRD), field-emission scanning electron microscopy (FE-SEM), and atomic force microscopy (AFM) analyses allowed an accurate determination of the physicochemical properties of the deposited layers, while Hall Effect measurements enabled verifying the p-type conductivity of the deposited films. Finally, UV-vis spectra have been carried out for the determination of the optical band gap.

Cu 2 O Synthesis
The Cu(tmhd) 2 compound was purchased from Sigma-Aldrich and used without further purification. The depositions were performed in a horizontal, hot-wall reactor under reduced pressure, using argon (150 sccm) as a carrier gas, and oxygen (200 sccm) as a reacting gas. The Ar and O 2 flows were controlled using MKS 1160 flow controller units and were introduced in proximity to the reaction zone. The vacuum inside the reactor was maintained through a scroll pump unit and monitored at the value of 4 Torr using MKS Baratron 122AAX. The films were deposited on Si (001) and quartz/ITO substrates in the 250-400 • C temperature range. The precursor source was kept at 130-140 • C for an efficient vaporization process. Each section was heated independently, with ±2 • C accuracy, using K-type thermocouples and computer-controlled hardware.

MAPI Deposition
The MAPbI 3 films were deposited by a Low Vacuum Proximity Space Effusion (LV-PSE) technique with specifically customized vacuum deposition equipment, provided by Kenosistec s.r.l. Lead iodide powders (99.99% purity) were purchased from Sigma Aldrich (St. Louis, MO, USA). Methyl ammonium iodide was purchased from Dyenamo AB. All materials were used as received without any further purification. The LV-PSE technique consists of a sequential deposition of PbI 2 and MAI via physical sublimation from powders at a base pressure of~2 × 10 −2 mbar with the crucibles taken at 350 • C and 135 • C, respectively. The substrate was posed at a medium-range distance with respect to the sources (~2 cm). During the first step, PbI 2 was deposited, then the conversion into MAPbI 3 occurred through an adsorption-incorporation-migration mechanism fully described in previous works [40][41][42].

Characterization
Structural characterization was performed using a Smartlab Rigaku diffractometer in grazing incident mode (0.5 • ) operating at 45 kV and 200 mA equipped with a rotating anode of Cu K α radiation. Film morphologies were investigated using field emission scanning electron microscopy (FE-SEM) ZEISS SUPRA 55 VP. The films deposited on glass were Au-coated before FE-SEM characterization. Topographic characterization was performed through Atomic Force Microscopy (AFM) adopting an Au-coated silicon probe with a nominal 35 nm tip curvature radius and a typical force constant of 0.1 N. The AFM images were obtained in contact mode. Before and after each measurement the noise level was 0.01 nm. The UV-Visible absorption spectra were recorded using a Jasco V-650 spectrophotometer. The spectra were recorded in the wavelength range from 250 to 700 nm for Cu 2 O thin films deposited on ITO-quartz substrates. Electrical characterization of the material was carried out by Hall effect measurements at room temperature using MMR H50 equipment. For this purpose, Van der Paw structures were fabricated, defining Cu/Au Ohmic contacts at the four corners of 1 cm × 1 cm Cu 2 O samples grown on a SiO 2 substrate.

Results
An MOCVD approach has been successfully applied to the synthesis of copper oxide in form of a thin film starting from Cu(tmhd) 2 complex. All the depositions allow the fabrication of compact and homogeneous thin films in the area of 2 cm × 2 cm on Si and quartz/ITO substrates. The effect of deposition temperature has been accurately studied in terms of both phase composition and morphology structures.
A complete overview of the X-ray diffraction (XRD) analysis of copper oxide thin films deposited by MOCVD in the 250-400 • C temperature range is reported in Figure 1. At the lowest temperature of 250 • C, the pattern (red line) exhibits peaks at 29 111/200 reflections of CuO traces. In fact, at higher temperatures (350 together with the previous one, the additional peak at 35.58° points to mixture of Cu2O and CuO phases. Finally, at 400 °C, the pattern in formation of the pure tenorite phase CuO, with the presence of signa and 48.80° related to the 002, 111/200, and -202 reflections, respectivel 045-0937), while no peaks related to Cu2O are present. The observed trend indicates a strong effect of the deposition t selective formation of Cu(I) and Cu(II) oxides, with the lowest tem stabilizing the Cu2O phase and the highest temperature (400 °C) stabili At even lower temperatures of 200 °C, only very small, isolated nuclei S1a), thus indicating that 250 °C is the lowest temperature to produce sig decomposition. This finding is also supported by the EDX data (Figure show any Cu peak; the Cu amount is lower than the detection limit of t The Cu(I) and/or Cu(II) oxide formation can be rationalized con between two aspects: (i) the decomposition mechanism of the Cu(tmhd reactor, which involves the organic component of the complex acting as (ii) the oxidant atmosphere under the present deposition conditions lower temperatures, the first aspect is prevalent and tends to stabiliz phase due to the reducing action of the organic component; on the c temperatures, the second aspect prevails, resulting in the mere formati Differing from the present findings, Gupta et al. [39] reported the f films from the same Cu(tmhd)2 precursor at a much higher deposition t °C, while in present experiments at a temperature of 300 °C the Cu(I oxidize to Cu(II) and at 400 °C pure CuO forms. The reason why Gup the Cu2O at higher temperatures is that in addition to N2, used in their gas, a mixture of N2O and oxygen was used during deposition. The N The observed trend indicates a strong effect of the deposition temperature in the selective formation of Cu(I) and Cu(II) oxides, with the lowest temperature (250 • C) stabilizing the Cu 2 O phase and the highest temperature (400 • C) stabilizing the CuO one. At even lower temperatures of 200 • C, only very small, isolated nuclei are found ( Figure S1a), thus indicating that 250 • C is the lowest temperature to produce significant precursor decomposition. This finding is also supported by the EDX data ( Figure S1b), which do not show any Cu peak; the Cu amount is lower than the detection limit of the technique.
The Cu(I) and/or Cu(II) oxide formation can be rationalized considering a balance between two aspects: (i) the decomposition mechanism of the Cu(tmhd) 2 precursor in the reactor, which involves the organic component of the complex acting as a reducing agent; (ii) the oxidant atmosphere under the present deposition conditions [29]. Therefore, at lower temperatures, the first aspect is prevalent and tends to stabilize the Cu(I) oxide phase due to the reducing action of the organic component; on the contrary, at higher temperatures, the second aspect prevails, resulting in the mere formation of Cu(II) oxide.
Differing from the present findings, Gupta et al. [39] reported the fabrication of Cu 2 O films from the same Cu(tmhd) 2 precursor at a much higher deposition temperature of 750 • C, while in present experiments at a temperature of 300 • C the Cu(I) already starts to oxidize to Cu(II) and at 400 • C pure CuO forms. The reason why Gupta et al. stabilized the Cu 2 O at higher temperatures is that in addition to N 2 , used in their study as a carrier gas, a mixture of N 2 O and oxygen was used during deposition. The N 2 O was acting as a reducing agent in regard to the Cu(II), which would have stabilized for the effect of temperature. In our experiments, starting from a Cu(II) precursor, we succeeded in growing highly homogeneous Cu 2 O layers at very low temperatures, with the tmhd organic ligand acting as a reducing agent. Finally, the nature of the films deposited by Gupta et al. [39] had a reported thickness of only 15 nm and TEM observed islands in the order of 100-200 nm.
It is worth noting that in the literature only a few works report the stabilization of the cuprite crystalline phase at low temperatures, with the lowest reported temperature of 300 • C for vapor deposition approaches [28,29,32,33]. In the present case, the formation of pure Cu 2 O films has been achieved at even lower temperatures and actually, the thermal budget related to a 50 • C difference is a great advantage. In fact, the present optimized process, using a deposition temperature as low as 250 • C, paves the way for a wide application of the current Cu 2 O synthetic approach also on a variety of polymeric substrates.
Morphology features of CuO and Cu 2 O thin films have been studied by field emission scanning electron microscopy (FE-SEM). The FE-SEM image of the Cu 2 O films deposited at 250 • C on Si (100) (Figure 2a) shows the formation of a very compact and homogeneous layer in which, probably due to the low fabrication temperature, a nanostructured feature is barely visible. Indeed, at higher deposition temperatures, i.e., 300 • C and 350 • C in Figure 2b,c, respectively, the films show a porous structure, probably caused by the coalescence of small grains of the order of tens of nanometers during the film growth. Finally, at 400 • C, the morphology appears much more nanostructured with the formation of plate-like grains of the order of 200-400 nm assembled into column-like structures (Figure 2d). The evident change in morphologies as a function of fabrication temperatures may be attributed on the one hand to the different contributions of nucleation and growth processes, on the other hand to the different Cu-O crystalline phase arrangements. The cross-section images display a growth trend, with an increase from 220 nm to 420 nm at 250 • C and 300 • C (Figure 2e,f), and up to 740 nm for the layer obtained at 350 • C (Figure 2g). The film deposited at 400 • C shows, indeed, a thicker and rugged profile, with a thickness of about 1250 nm (Figure 2h). Specifically, growth rates vary from 3, 7, 12 and 21 nm/min for 250, 300, 350 and 400 • C, respectively. To explain the dependence of the growth rate as a function of temperature, the Arrhenius plot ( Figure 3) has been derived, even though the formed Cu-O phases differ in the investigated temperature range.
It is worth noting that in the literature only a few works report the stabilization of t cuprite crystalline phase at low temperatures, with the lowest reported temperature 300 °C for vapor deposition approaches [28,29,32,33]. In the present case, the formation pure Cu2O films has been achieved at even lower temperatures and actually, the therm budget related to a 50 °C difference is a great advantage. In fact, the present optimiz process, using a deposition temperature as low as 250 °C, paves the way for a wi application of the current Cu2O synthetic approach also on a variety of polyme substrates.
Morphology features of CuO and Cu2O thin films have been studied by fie emission scanning electron microscopy (FE-SEM). The FE-SEM image of the Cu2O fil deposited at 250 °C on Si (100) (Figure 2a) shows the formation of a very compact a homogeneous layer in which, probably due to the low fabrication temperature nanostructured feature is barely visible. Indeed, at higher deposition temperatures, i 300 °C and 350 °C in Figure 2b,c, respectively, the films show a porous structure, probab caused by the coalescence of small grains of the order of tens of nanometers during t film growth. Finally, at 400 °C, the morphology appears much more nanostructured w the formation of plate-like grains of the order of 200-400 nm assembled into column-l structures (Figure 2d). The evident change in morphologies as a function of fabricati temperatures may be attributed on the one hand to the different contributions nucleation and growth processes, on the other hand to the different Cu-O crystall phase arrangements. The cross-section images display a growth trend, with an increa from 220 nm to 420 nm at 250 °C and 300 °C (Figure 2e   The absorbance spectra in the UV-vis range of the ITO/quartz su Cu2O_230 nm, Cu2O_140 nm and Cu2O_90 nm films are displayed in absorbance spectra show high transparency up to 300 nm for the ITO su absorption edge localized between 450 and 500 nm for the Cu2O films, wh to the excitonic band gap of Cu2O films. The spectra of the Cu2O_230 nm and Cu2O_90 nm films display an increase in absorbance as a consequence increase, a behavior already observed for films grown with different Transmission spectra of ITO/quartz and Cu2O thin films are reported in F in Figure S3 the transmission of the ITO/quartz annealed at 250 °C is c untreated ITO/quartz substrate. Regardless of thickness, all Cu2O films ab Despite the increase in thickness variation with increasing temperature, the growth rate dependence does not suggest the presence of a kinetic regime and points to a mass transport-limited regime. This evidence is supported by the derived low activation energy, obtained from the slope of the linear fitting analysis, equal to 2.0 ± 0.1 kJ/mol.
Hence, optical absorptions have been evaluated in order to determine the band gap of the semiconductor Cu 2 O thin films. Particularly, UV-vis spectra have been measured for Cu 2 O thin films of various thicknesses, deposited on ITO/quartz substrate.
With the aim of applications as a hole transporting layer (HTL) in solar cell devices, the thickness of the Cu 2 O films has been finely tuned and optimized. Therefore, different depositions have been carried out at 250 • C and with different vaporization temperatures and deposition times, i.e., (I) T vap : 140 • C, t: 30 min; (II) T vap : 140 • C, t: 15 min and (III) T vap : 130 • C t: 15 min, in order to have different-thickness Cu 2 O layers. In particular, under these conditions, three samples have been obtained with thicknesses of 230, 140, and 90 nm, respectively. The cross-section and the morphologies are reported in Figure S2.
The absorbance spectra in the UV-vis range of the ITO/quartz substrate and the Cu 2 O_230 nm, Cu 2 O_140 nm and Cu 2 O_90 nm films are displayed in Figure 4a. The absorbance spectra show high transparency up to 300 nm for the ITO substrate, and an absorption edge localized between 450 and 500 nm for the Cu 2 O films, which corresponds to the excitonic band gap of Cu 2 O films. The spectra of the Cu 2 O_230 nm, Cu 2 O_140 nm and Cu 2 O_90 nm films display an increase in absorbance as a consequence of the thickness increase, a behavior already observed for films grown with different techniques [43]. Transmission spectra of ITO/quartz and Cu 2 O thin films are reported in Figure 4b, while in Figure S3 the transmission of the ITO/quartz annealed at 250 • C is compared to the untreated ITO/quartz substrate. Regardless of thickness, all Cu 2 O films absorb strongly in the UV-vis region between 250 and 450 nm, resulting in a drop in the percentage of transmission. As the thickness of the hole transporting layer is significantly reduced, the transmittance in the visible range increases. In fact, up to λ = 480 nm, the transmittance for the thickest Cu 2 O_230 nm and Cu 2 O_140 nm is minimal and then begins to rise, reaching only 20% for the Cu 2 O_230 nm and 40% for the intermediate Cu 2 O_140 nm at λ = 500 nm. Cu 2 O_90 nm thin film, on the other hand, achieves acceptable transmittance levels (about 60%) at λ = 450 nm and reaches a higher 70% at λ= 650 nm. The present results agree with the literature data [44,45], and point to the potential applicability of the presently grown Cu 2 O film as an HTL layer for photovoltaic applications [46].  Tauc's plot obtained from the relation of (αhν) 2 versus (hν) is calculated f optical measurements reported in Figure S4 for the ITO/quartz substrate and in for the Cu2O films. Using a linear extrapolation, the values of the direct optical b of the Cu2O_230 nm, Cu2O_140 nm and Cu2O_90 nm films are evaluated to be 1 and 2.41 eV, respectively. In order to evaluate the reproducibility of the proces have statistically significant results, the band gaps from three different samples deposition condition of the films Cu2O_230 nm, Cu2O_140 nm and Cu2O_90 nm h extrapolated yielding average values of 1.97 ± 0.02, 2.03 ± 0.03 and 2.37 ± respectively. It is interesting to note that the band gap increases upon decreas thickness. This band gap spreading effect has been previously observed for Cu [47] and in similar thin film studies [48,49].
These values compare well with literature data of Cu2O samples deposited spatial atomic layer deposition [31,50,51]. Tauc's plot obtained from the relation of (αhν) 2 versus (hν) is calculated from the optical measurements reported in Figure S4 for the ITO/quartz substrate and in Figure 5 for the Cu 2 O films. Using a linear extrapolation, the values of the direct optical band gap of the Cu 2 O_230 nm, Cu 2 O_140 nm and Cu 2 O_90 nm films are evaluated to be 1.99, 2.04 and 2.41 eV, respectively. In order to evaluate the reproducibility of the process and to have statistically significant results, the band gaps from three different samples for each deposition condition of the films Cu 2 O_230 nm, Cu 2 O_140 nm and Cu 2 O_90 nm have been extrapolated yielding average values of 1.97 ± 0.02, 2.03 ± 0.03 and 2.37 ± 0.06 eV, respectively. It is interesting to note that the band gap increases upon decreasing film thickness. This band gap spreading effect has been previously observed for Cu 2 O films [47] and in similar thin film studies [48,49]. Tauc's plot obtained from the relation of (αhν) 2 versus (hν) is calculated from optical measurements reported in Figure S4 for the ITO/quartz substrate and in Figu for the Cu2O films. Using a linear extrapolation, the values of the direct optical band of the Cu2O_230 nm, Cu2O_140 nm and Cu2O_90 nm films are evaluated to be 1.99, and 2.41 eV, respectively. In order to evaluate the reproducibility of the process an have statistically significant results, the band gaps from three different samples for e deposition condition of the films Cu2O_230 nm, Cu2O_140 nm and Cu2O_90 nm have b extrapolated yielding average values of 1.97 ± 0.02, 2.03 ± 0.03 and 2.37 ± 0.06 respectively. It is interesting to note that the band gap increases upon decreasing thickness. This band gap spreading effect has been previously observed for Cu2O f [47] and in similar thin film studies [48,49].
These values compare well with literature data of Cu2O samples deposited thro spatial atomic layer deposition [31,50,51]. In order to determine the type of conductivity of the deposited Cu2O films, Van Pauw and Hall Effect measurements have been performed at room temperature.
As an example, Figure 6 shows a typical current-voltage (I-V) measurement car These values compare well with literature data of Cu 2 O samples deposited through spatial atomic layer deposition [31,50,51].
In order to determine the type of conductivity of the deposited Cu 2 O films, Van der Pauw and Hall Effect measurements have been performed at room temperature.
As an example, Figure 6 shows a typical current-voltage (I-V) measurement carried out on a Van der Pauw structure fabricated on the Cu 2 O sample. The inset in Figure 6 shows schematically the geometry of the sample and the I-V measurement setup. From the linear fit of the experimental data, the sheet resistance R SH of the Cu 2 O films has been determined. als 2022, 15,7790 the linear fit of the experimental data, the sheet resistance RSH of t determined. Thereafter, Hall effect measurements have been performed b field of 0.1 T perpendicular to the sample surface. These measu determine the type of carriers responsible for the conduction and t Then, combining the values of the sheet resistance, determin Pauw measurements, with the carrier density, extracted under magnetic field, the carrier mobility has been determined.
Noteworthy, from the sign of the Hall voltage it can be deposited films are p-type. Moreover, no significant differences in have been observed by varying the Cu2O film thickness from 90 n The hole concentration p is in the order of 1.5-2.8 × 10 16 cm −3 , Hall mobility are in the range 4-7 cm 2 V −1 s −1 . Table 1 summarize extracted by the Van der Pauw and Hall effect measurements. Th table have an error of about ±20% arising from the average of d Thereafter, Hall effect measurements have been performed by applying a magnetic field of 0.1 T perpendicular to the sample surface. These measurements allowed us to determine the type of carriers responsible for the conduction and their concentration [52].
Then, combining the values of the sheet resistance, determined through the Van der Pauw measurements, with the carrier density, extracted under the application of the magnetic field, the carrier mobility has been determined.
Noteworthy, from the sign of the Hall voltage it can be concluded that all the deposited films are p-type. Moreover, no significant differences in the electrical properties have been observed by varying the Cu 2 O film thickness from 90 nm to 230 nm.
The hole concentration p is in the order of 1.5-2.8 × 10 16 cm −3 , while the values of the Hall mobility are in the range 4-7 cm 2 V −1 s −1 . Table 1 summarizes the main parameters extracted by the Van der Pauw and Hall effect measurements. The values reported in the table have an error of about ±20% arising from the average of different measurements acquired on the Van der Pauw structure in the different orientations. These differences are likely due to the high contact resistance of the metal pads at the four corners of the Van der Pauw structures, but a slight non-uniformity in the film's electrical properties cannot be excluded.
The observed values are comparable to or even better than literature data which give resistivity in the range of 20-10 3 Ω cm [53,54] and carrier concentrations of 10 15 -10 18 cm −3 [53,54].
Afterward, pure cuprite thin films are deposited on a conductive layer, i.e., ITO on a quartz substrate, starting from the optimized MOCVD parameter conditions. The Cu 2 O_ITO_quartz system has been used for the sequential deposition of the MAPI active layer as part of the stack of a solar cell with a planar inverted architecture [13,40,55,56]. An accurate study has been conducted for each step of the multilayer MAPI_Cu 2 O_ITO assembly. The system has been grown on a surface of 2 cm × 2 cm.
X-ray diffraction patterns of the different layers have been reported in Figure 7. The pattern in brown, related to the MAPI/Cu 2 O/ITO system, displays several peaks pointing to the formation of a crystalline multilayer and associated with: (i) the pure Cu 2 O phase; (ii) the MAPI reflection peaks attributed through comparison with literature data [57]; (iii) CuI impurities. No peaks associated with the ITO substrate are detected due to the thickness of the MAPI_Cu 2 O layers. The detection of the CuI phase is likely a result of some interaction of the MAPI layer with the Cu 2 O thin film. This phenomenon is supported by the study of the interaction between the MAPI and the Cu electrode, which has been demonstrated to produce CuI [58]. Cu2O_ITO_quartz system has been used for the sequential deposition of the MAPI active layer as part of the stack of a solar cell with a planar inverted architecture [13,40,55,56]. An accurate study has been conducted for each step of the multilayer MAPI_Cu2O_ITO assembly. The system has been grown on a surface of 2 cm × 2 cm. X-ray diffraction patterns of the different layers have been reported in Figure 7. The pattern in brown, related to the MAPI/Cu2O/ITO system, displays several peaks pointing to the formation of a crystalline multilayer and associated with: (i) the pure Cu2O phase; (ii) the MAPI reflection peaks attributed through comparison with literature data [57]; (iii) CuI impurities. No peaks associated with the ITO substrate are detected due to the thickness of the MAPI_Cu2O layers. The detection of the CuI phase is likely a result of some interaction of the MAPI layer with the Cu2O thin film. This phenomenon is supported by the study of the interaction between the MAPI and the Cu electrode, which has been demonstrated to produce CuI [58].  Figure 8 show the different morphologies of the multilayer structure from the ITO substrate ( Figure 8a) and the Cu2O film (Figure 8b) to the MAPI layer (Figure 8c). The ITO substrate is a compact homogeneous nanostructured layer with regular grains of about 50 nm and a thickness of about 305 nm. The Cu2O thin film, deposited on top of the ITO substrate, displays a flat morphology (Figure 8b) similar to the one found for the Cu2O film grown on Si (see Figure 2a), with a whole thickness of 525 ± 10 nm. This thickness, considering the previous estimation of 305 nm for the ITO, leads to a net Cu2O layer of about 220 ± 10 nm. Finally, the MAPI film in Figure 8c shows a nanostructured compact surface with smooth grains of hundreds of nanometers in size. The estimated thickness of the MAPI layer is around 275 ± 10 nm, derived from the difference in the cumulative thickness of the ITO_Cu2O_MAPI system of 800 ± 10 nm.

FE-SEM images in
Finally, the atomic force microscopy (AFM) characterization of the three layers (Figure 8a'-c') confirms the homogeneity of the films as well with fully coalesced grains on a larger area of 10 μm × 10 µ m, and a root mean square (RMS) roughness of about 2.9 nm, 8.9 nm and 26.6 nm (measured on areas of 4 μm × 4 μm) for ITO, Cu2O and MAPI films, respectively. The RMS roughness values of three different Cu2O samples, deposited under the same conditions of the Cu2O_230 nm, range from 7.8 to 10.1 nm. The low RMS roughness of the Cu2O layer is comparable to literature data on films grown on Si, further confirming these films as suited for the growth of the perovskite layer.

Conclusions
In summary, a simple approach has been optimized to produce a stack composed of MAPI/Cu2O/ITO/quartz. The extremely tunable MOCVD process allows the selective and reproducible fabrication of pure, uniform and highly compact Cu2O films on various substrates. In addition, the very low deposition temperature of the Cu2O layer makes this process appealing for deposition on temperature-sensitive substrates, such as polymers. In fact, the low operating temperature is a crucial issue and the capability to operate at 250 °C, using a process that is already industrially applied on a large scale, represents a breakthrough for the production of solar cells on plastic flexible supports. The electrical measurements and band gap values confirm the potentiality of the Cu2O layers as p-type semiconducting materials for solar cell devices. Finally, preliminary studies have been carried out to produce the MAPI/Cu2O/ITO stack, allowing us to scrutinize problematic issues regarding the MAPI/Cu2O interface.

Supplementary Materials:
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ma15217790/s1, Figure S1: FE-SEM plan-image and EDX spectrum of the Cu2O film deposited at 200 °C; Figure S2: FE-SEM images in plain and cross of Cu2O thin films; Figure S3: Transmittance spectra of the ITO/quartz substrate; Figure S4  Finally, the MAPI film in Figure 8c shows a nanostructured compact surface with smooth grains of hundreds of nanometers in size. The estimated thickness of the MAPI layer is around 275 ± 10 nm, derived from the difference in the cumulative thickness of the ITO_Cu 2 O_MAPI system of 800 ± 10 nm.
Finally, the atomic force microscopy (AFM) characterization of the three layers (Figure 8a'-c') confirms the homogeneity of the films as well with fully coalesced grains on a larger area of 10 µm × 10 µm, and a root mean square (RMS) roughness of about 2.9 nm, 8.9 nm and 26.6 nm (measured on areas of 4 µm × 4 µm) for ITO, Cu 2 O and MAPI films, respectively. The RMS roughness values of three different Cu 2 O samples, deposited under the same conditions of the Cu 2 O_230 nm, range from 7.8 to 10.1 nm. The low RMS roughness of the Cu 2 O layer is comparable to literature data on films grown on Si, further confirming these films as suited for the growth of the perovskite layer.

Conclusions
In summary, a simple approach has been optimized to produce a stack composed of MAPI/Cu 2 O/ITO/quartz. The extremely tunable MOCVD process allows the selective and reproducible fabrication of pure, uniform and highly compact Cu 2 O films on various substrates. In addition, the very low deposition temperature of the Cu 2 O layer makes this process appealing for deposition on temperature-sensitive substrates, such as polymers. In fact, the low operating temperature is a crucial issue and the capability to operate at 250 • C, using a process that is already industrially applied on a large scale, represents a breakthrough for the production of solar cells on plastic flexible supports. The electrical measurements and band gap values confirm the potentiality of the Cu 2 O layers as p-type semiconducting materials for solar cell devices. Finally, preliminary studies have been carried out to produce the MAPI/Cu 2 O/ITO stack, allowing us to scrutinize problematic issues regarding the MAPI/Cu 2 O interface.