Enhanced Performance of Inverted Perovskite Solar Cells Employing NiO_x and Cu-Doped NiO_x Nanoparticle Hole Transport Layers

Abstract

In this study, p-type NiO_x and Cu-doped NiO_x nanoparticles (NPs) were synthesized by a simple chemical precipitation method and used as hole transport layers (HTLs) for inverted perovskite solar cells (PSCs). The microstructural property, surface morphology, elemental composition, optical property, charge recombination, and surface topography of the NiO_x and Cu-NiO_x HTLs were comprehensively characterized. The results showed that the NiO_x and Cu-NiO_x NPs were uniformly coated on the substrates without pinholes or voids. Cu incorporation into NiO_x did not change its crystalline nature and considerably improved its electrical conductivity. The Cu-NiO_x HTLs exhibited superior photoluminescence quenching and the least lifetime decay, which indicated that Cu-NiO_x exhibited higher charge transport than NiO_x HTLs. The fabricated PSC performances were further analyzed using current density–voltage characteristics, external quantum efficiency, and electrochemical impedance spectroscopy. The PSCs with PEDOT:PSS, NiO_x, and 2% Cu-NiO_x HTLs exhibited power conversion efficiencies of 11.93%, 13.72%, and 15.54%, respectively. The 2% Cu-NiO_x HTL-based device showed the best performance compared with the PEDOT:PSS- and NiO_x-based devices. Academic Editors: Chunyang Zhang, Dou Zhang

1. Introduction

Organic and inorganic metal halide perovskites have emerged as pivotal materials in optoelectronic applications due to their high charge-carrier mobility, strong light absorption, and tunable bandgap. Moreover, they have the potential to expand their applications into diverse fields such as photocatalysis [1], energy storage [2], and advanced computing and memory (memristors) [3]. Over the past decade, perovskite solar cells (PSCs) have attracted significant attention owing to their rapidly increasing power conversion efficiency and potential for commercialization [4,5,6]. Structurally, PSCs are generally categorized into two configurations: conventional (n–i–p) and inverted (p–i–n). Although conventional PSCs are efficient, they are highly susceptible to hysteresis and thermal instability and are expensive. Inverted PSCs exhibit limited hysteresis and excellent stability but inferior PCE. Nevertheless, inverted PSCs have attracted considerable research interest because of their simple fabrication processes and suitability for use in flexible devices due to their low-temperature processing. The performance of inverted PSCs can be improved through hole transport layer (HTL) engineering. The requirements of PSCs are high optical transparency, hole mobility, and energy alignment with the perovskite layers.

In inverted PSCs, poly(3,4-(ethylenedioxy)thiophene)/poly(styrene sulfonate) (PEDOT:PSS) HTLs are common. However, these HTLs are hygroscopic and acidic, which severely affects PSC stability. Therefore, metal oxides are an ideal candidate for replacing organic HTLs. NiO_x is the most widely studied HTL material because of its excellent p-type semiconducting property, high transparency, and numerous deposition methods. In 2014, Jeng et al. first introduced NiO_x HTLs in PSCs as a replacement for PEDOT:PSS [7]. Several researchers have since developed NiO_x thin films for PSCs by using various methods such as those using the sol–gel process, combustion, nanoparticle (NP) ink, chemical bath deposition, electrodeposition, spray pyrolysis, sputtering, electron-beam physical vapor deposition process, and pulsed-laser deposition [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23]. Low-temperature fabrication methods are attractive because they are suitable for fabricating flexible devices. NiO_x HTLs are typically prepared using a low-temperature (150 °C) combustion method. NiO_x HTLs prepared from NP ink are easy to fabricate and exhibit higher crystallinity than those prepared with other techniques. This method can be readily adapted for the doping of other metals to improve the hole mobility of the NiO_x layer. This improvement is crucial because NiO_x is limited because of its electrical property and exhibits lower hole mobility, valence band maximum, and more surface defects than (PEDOT:PSS) HTLs. Techniques such as doping and surface modification have been proposed to improve the conductivity and charge collection of NiO_x HTLs [24,25,26,27]. Metal doping is the best surface modification approach for improving the electrical properties of NiO_x HTLs. Many researchers have reported doping NiO_x with metals such as Li, Cu, Mg, Co, Cs, and Fe [28,29,30,31,32].

Cu doping has attracted considerable research attention because of its high carrier concentration, mobility, and work function [33]. Several reports have demonstrated PCEs exceeding 23% in inverted perovskites solar cells through strategies such as multi-cation incorporation at the A-site, the use of two-dimensional (2D) perovskites, and interface engineering at both the HTL and the ETL [34,35]. In this work, we focus on the incorporation of Cu at the Ni site to enhance the electrical properties and overall device performance by modifying the HTL without any heating or annealing during the fabrication process, except for the perovskite film. We fabricated inverted PSCs with presynthesized NiO_x NP and Cu-doped NiO_x (Cu-NiO_x) NP HTLs. The device structure was Glass/FTO/NiO_x or Cu-NiO_x or PEDOT:PSS/perovskite/PCBM/BCP/Ag, as shown in Figure 1a, and its equivalent energy level diagram is displayed in Figure 1b. A simple chemical precipitation technique was used to prepare NiO_x NPs. The synthesized NPs were characterized through X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), field emission scanning electron microscopy (FESEM), ultraviolet (UV)–visible (Vis) spectroscopy, photoluminescence (PL) spectroscopy, and atomic force microscopy (AFM). The performance of the fabricated device was characterized using current density–voltage (J–V) analysis, electrochemical impedance spectroscopy (EIS), and external quantum efficiency (EQE). The fabricated Cu-NiO_x PSCs exhibited improved performance. The PSCs fabricated using 2% Cu-NiO_x NP inks, with a PCE of 15.54%, exhibited the highest efficiency. The device fabricated using pristine NiO_x and PEDOT:PSS exhibited PCEs of 13.72% and 11.93%, respectively.

2. Experimental Details

2.1. Synthesis of NiO and Cu-Doped NiO NPs

First, 0.4 M nickel nitrate hexahydrate (Ni(NO₃)₂ 6H₂O) (CHONEYE PURE CHEMICALS, Miaoli County, Taiwan, Purity > 98%) was dissolved in deionized (DI) water to obtain a green solution. Then, 10 M sodium hydroxide (NaOH) was added dropwise till the solution reached a pH of 10. The precipitate was stirred for 30 min, centrifuged at 16,000 rpm for 10 min, and washed with DI water three times. The green powders were dried at 80 °C overnight and calcined at 270 °C for 2 h at a 5 °C/min ramping rate. Finally, NiO_x NPs in the form of black powder were obtained. The NPs were dispersed in DI water at a 20 mg/mL concentration. For Cu doping, various mole percentages of nickel nitrate were replaced by Copper(II) nitrate trihydrate Cu(NO₃)₂ 3H₂O (CHONEYE PURE CHEMICALS, Miaoli County, Taiwan, Purity > 98%). Other procedures were the same as those for the undoped NiO_x NPs.

2.2. PSC Device Fabrication

The schematic of PSC device fabrication is represented in Figure 2. FTO substrates (Ruilong Optoelectronics Co., Ltd, Miaoli County, Taiwan) were patterned and cleaned in an ultrasonic bath using detergent, acetone, and isopropyl alcohol (IPA) for 10 min. Then, the substrates were dried using a nitrogen gun and UV–ozone treated for 15 min. The HTL production method was as follows: NiO_x (or Cu-NiO_x) NP ink was spin-coated onto the FTO substrates at a speed of 3000 rpm for 30 s. We also fabricated devices with a PEDOT:PSS HTL for comparison. The PEDOT:PSS (UNI-ONWARD Corp, New Taipei City, Taiwan) was spin-coated at 4000 rpm for 30 s and annealed at 140 °C for 10 min. Subsequently, the perovskite active layer was produced through the following procedures: 1.4 M MAI (UNI-ONWARD Corp, New Taipei City, Taiwan) and 1.4 M PbI₂ (UNI-ONWARD Corp, New Taipei City, Taiwan) were dissolved in a mixed solvent of GBL/DMSO in a 7:3 ratio. The perovskite solution was stirred at 60 °C overnight and filtered using a 0.45 µm PTFE syringe filter before deposition. The solution was spin-coated onto the FTO/NiO_x substrates at 1000 rpm for 15 s and 5000 rpm for 25 s. Toluene was used as an antisolvent and was added exactly 5 s before the end of the second step. The substrates were annealed at 100 °C for 10 min. The electron transport layer (ETL) production process was as follows: 20 mg/mL of phenyl-C61-butyric acid methyl ester (PCBM) in 1,4-dichlorobenzene was spin-coated on the substrates at 2000 rpm for 30 s, and a saturated solution of BCP in methanol was spin-coated at 6000 rpm for 30 s. No annealing process was performed during fabrication, except for the perovskite film. Finally, a Ag electrode was evaporated through thermal evaporation.

2.3. Characterization

NiO_x and Cu-NiO_x were characterized through XRD (Rigaku D/Max-2500V, Rigaku Corp, Tokyo, Japan); their surface morphology was observed through FESEM (JEOL JSM-7000F, JEOL Inc., Tokyo, Japan), and their surface topography was measured through AFM (AutoProbe CP, Thermomicroscopes, Sunnyvale, CA, USA). Moreover, elemental analysis was performed through XPS (Thermo K-Alpha, Thermo Fisher Scientific, Waltham, MA, USA). UV-Vis absorption and transmission were measured through a UV-Vis spectrophotometer (Jasco V-650, Jasco, Tokyo, Japan). PL and lifetime measurements were performed using a solid-state laser (λ = 405 nm, PDL 800-D, PICOQUANT) equipped with a homemade laser scanning confocal microscope and a spectrometer equipped with a thermoelectrically cooled charge-coupled detector (DV420A-BU2, Andor, Belfast, Northern Ireland). The performance of the fabricated PSC device was analyzed using J–V curves (CH 6116D, CH Instruments, Bee Cave, TX, USA), EQE (QE-R, Enlitech, Kaohsiung, Taiwan), and EIS. The J–V characteristics of the PSCs were measured under the illumination of a simulated AM 1.5G solar spectrum from a 550-W Xe lamp solar simulator (Abet Technologies Sun 3000 Class AAA, Abet Technologies, Milford, CT, USA).

3. Results and Discussion

3.1. Chemical Reaction Mechanism of NiO_x and Cu-NiO_x NPs

The NiO_x and Cu-NiO_x NPs were synthesized using a simple chemical precipitation method according to procedures reported in studies [36,37,38,39,40,41,42] with some modifications (see details in the experimental section). We synthesized pristine NiO_x NPs and 2%, 4%, and 8% Cu-NiO_x NPs. HTL development using simple synthesis methods is promising for the large-scale industrial production of PSCs. The chemical reaction mechanism of nonstoichiometric NiO_x NPs and Cu-NiO_x NPs is expressed in Equations (1)–(4), respectively.

$${Ni\left(NO\right)}_3+NaOH\to {Ni\left(OH\right)}_2\downarrow + {Na\left({NO}_3\right)}_2$$

(1)

$${Ni\left(OH\right)}_2\stackrel{270 °\mathrm{C}}{\to }{NiO}_x+H_{2}O$$

(2)

$${Ni\left(NO\right)}_3+ {Cu\left(NO\right)}_3+ NaOH\to {Ni\left(OH\right)}_2\downarrow + {Cu\left(OH\right)}_2\downarrow + {Na\left({NO}_3\right)}_2$$

(3)

$${Ni\left(OH\right)}_2+{Cu\left(OH\right)}_2\stackrel{270 °\mathrm{C}}{\to }{Cu-doped NiO}_x+ H_{2}O$$

(4)

The calcination temperature was set at 270 °C to ensure the thermal decomposition of Ni(OH)₂ to nonstoichiometric NiO_x NPs. However, Ni(OH)₂ cannot completely decompose to NiO at a temperature lower than the calcination temperature of 270 °C. Because Ni(OH)₂ is electrically insulated, it cannot be easily dispersed in solvents. If the calcination temperature exceeds 270 °C, the particle size increases. Both cases are not suitable for making an effective HTL for PSCs.

3.2. Characterization of the HTLs

XRD patterns for pristine NiO_x and various percentages of Cu-NiO_x NPs are illustrated in Figure 3. The diffraction peaks at 2θ were approximately 37°, 43°, and 62°, which corresponded to the (hkl) planes of (111), (002), and (022), respectively. Two other small peaks were obtained at approximately 75° and 79°, which corresponded to the (113) and (222) planes, respectively. The (002) peak domination in all the NPs and broad peaks indicated the presence of small crystals in the synthesized NPs. The crystallite size calculated using the Debye–Scherrer formula was approximately 4–6 nm. The XRD data is presented in

Table 1. Microstructural properties of NiO_x and Cu-NiO_x nanoparticles.

Sample	2θ	(hkl)	Interplanar Distance (Å)	FWHM (β) (radians)	Crystallite Size (D) (nm)	Dislocation Density (1/D2) (×1016)	Microstrain (β/4tanθ)	Lattice Parameter (a)
NiOx	37	(111)	2.43	0.03	4.94	4.12	1.27	4.21
	43.09	(002)	2.1	0.04	4.73	4.49	1.15	4.2
	62.54	(022)	1.49	0.04	4.52	4.92	0.85	4.21
2% Cu-NiOx	37.09	(111)	2.43	0.03	6.51	2.37	0.96	4.2
	43.16	(002)	2.1	0.03	5.67	3.13	0.96	4.2
	62.59	(022)	1.49	0.03	6.25	2.57	0.62	4.2
4% Cu-NiOx	36.96	(111)	2.44	0.03	4.91	4.17	1.28	4.22
	43.07	(002)	2.11	0.04	4.21	5.65	1.29	4.21
	62.55	(022)	1.49	0.04	4.06	6.07	0.95	4.2
8% Cu-NiOx	36.96	(111)	2.44	0.04	4.53	4.88	1.39	4.22
	43.11	(002)	2.1	0.04	4.28	5.49	1.27	4.2
	62.57	(022)	1.49	0.05	3.98	6.34	0.97	4.2
ICDD 98-000-5229	37.25	(111)	2.42
	43.28	(002)	2.09
	62.86	(022)	1.48

. The XRD results were consistent with the values of the International Center for Diffraction Data (reference number: 98-000-5229), which confirmed the existence of the cubic NiO_x phase. No other peaks related to metallic copper or copper oxide were observed, which revealed that the incorporation of Cu into nickel oxide did not change its crystal structure [43].

The optical properties were studied using UV-Vis spectroscopy. We measured the absorbance spectra of equally concentrated NiO_x and Cu-NiO_x NPs dispersed in DI water (Figure 4) in the wavelength range from 300 to 800 nm. The absorbance increased with an increase in the Cu concentration. The optical bandgaps were calculated using the standard Tauc plot method. The bandgap values were 3.8, 3.74, 3.71, and 3.66 eV for undoped NiO_x, 2%, 4%, and 8% Cu-NiO_x NPs, respectively. The bandgap values of synthesized NPs decreased with an increase in the Cu concentration. The incorporation of Cu atoms slightly narrowed the bandgap. The transmittance spectra of spin-coated NiO_x and Cu-NiO_x NPs on glass is illustrated in Figure 5, which revealed more than 85% of transmission in the visible region. The transmittance value decreased with an increase in the Cu concentration because of the narrowing bandgap.

The surface morphologies of NiO_x and various percentages of Cu-NiO_x NPs were observed through FESEM. Figure 6 clearly reveals that the NPs completely covered the substrate, and no pinholes or voids occurred, which confirmed the presence of some nanoclusters in the 4% and 8% Cu-NiO_x NP films. The size of the nanoclusters increased with an increase in the doping concentration of Cu. The conversion of Ni(OH)₂ to NiO_x occurred in the temperature range 250–305 °C, but the conversion of Cu(OH)₂ to CuO occurred from 150 °C, which resulted in the formation of these small nanoclusters. So far, the formation of these clusters because of copper doping has not been reported. The FESEM images of NiO_x and 2% Cu-NiO_x were similar, and they almost completely dispersed in DI water. In similar studies, the thickness of the HTL was reported to be approximately 50 nm, which is typically sufficient to ensure uniform coverage and efficient charge transport without increasing series resistance [44]. Therefore, they were selected for device fabrication.

XPS was performed to examine the elemental composition of NiO_x NPs and 2% Cu-NiO_x films. The XPS spectra of the Ni 2p, O 1s, and Cu 2p core level are illustrated in Figure 7. The Ni 2p spectra consisted of two peaks, namely, the main peak and the shoulder peak. The deconvoluted peaks at 853, 856, and 858 eV corresponded to NiO (Ni²⁺), Ni₂O₃ (Ni³⁺), and NiOOH, respectively. The Ni³⁺ ions confirmed the presence of Ni vacancies in the film, which rendered the film a p-type semiconductor. The peak at 861 eV was attributed to the modification of the NiO structure [45,46]. There was no significant change in the Ni 2p peaks between the NiO_x and Cu-NiO_x thin films. The O 1s spectra exhibited three peaks at 529, 531, and 533 eV, which corresponded to NiO (Ni²⁺), Ni₂O₃ (Ni³⁺), and NiOOH ions, respectively. In Cu 2p spectra, the peaks located at 933 and 953 eV corresponded to Cu 2p_3/2 and Cu 2p_1/2 transitions, respectively. The Cu²⁺ peak was dominant in Cu 2p_3/2 spectra, whereas the Cu⁺ peak was not significant [47]. The surface topography of NiO_x and 2% Cu-NiO_x NP thin films was investigated through AFM. Figure 8 displays the AFM images of NiO_x and 2% Cu-NiO_x NP thin films. The average root-mean-square roughness values of the NiO_x and 2% Cu-NiO_x NPs films were 4.068 ± 1.067 and 3.527 ± 0.9376 nm, respectively. The roughness decreased with Cu doping, which was consistent with reports.

PL and time-resolved PL (TRPL) were used to evaluate charge extraction and recombination from the devices. Typically, PL quenching and a short PL lifetime are expected for perovskites in contact with charge transport layers. We measured the PL spectra and the TRPL profile of the perovskite layers deposited on various HTLs (Figure 9). The PL quenching was better for the perovskite film deposited on Cu-NiO_x than for that on pristine NiO_x, which indicated that Cu incorporation in NiO_x improved the charge extraction property from the perovskite film. The TRPL profile was measured through time-correlated single-photon counting to analyze charge dynamics. PL lifetime decay tails were fitted with biexponential fitting function $y=y_0+A_1{e}^{-(x-x_0)/{\tau }_1}+A_2{e}^{-(x-x_0)/{\tau }_2}$, where (τ₁) is related to the diffusion of photogenerated excitons into defects, and (τ₂) is related to intrinsic electron–hole recombination. The average PL lifetime value was calculated using the formula ${\tau }_{avg}=\left({\mathsf{\Sigma }}_i{A}_i{\tau }_i^2\right)/({\mathsf{\Sigma }}_i{A}_i{\tau }_i)$. The obtained fitting values are summarized in

Table 2. Summary of the PL lifetime parameters from the fitting curves of the PL decay.

Sample	A1 (%)	τ1 (ns)	A2 (%)	τ2 (ns)	τavg (ns)
FTO/PSK	61.8	9.6	38.2	37.6	29.4
FTO/NiOx/PSK	64.9	3.3	35.1	11.4	8.6
FTO/2% Cu-NiOx/PSK	55.9	2.2	44.1	8.6	7.0

. The PL lifetime for the perovskite film deposited on Cu-NiO_x was shorter than that of pristine NiO_x, which indicates higher efficient charge transfer in Cu-NiO_x.

3.3. Characterization of the PSC Devices

The J–V curve, EQE, and EIS analyses were performed to investigate the PSC device performance of the synthesized NiO_x and 2% Cu-NiO_x NP HTLs. The device structure of the inverted PSCs was Glass/FTO/NiO_x or Cu-NiO_x or PEDOT:PSS/perovskite/PCBM/BCP/Ag. We also fabricated PSCs with a PEDOT:PSS HTL for comparison. The J–V analysis was first performed on the devices to investigate the effects of Cu doping in NiO_x NP HTLs on the short-circuit current density (J_SC), open-circuit voltage (V_OC), fill factor (FF), and PCE of the PSC device. The J–V curves of fabricated inverted PSCs with PEDOT:PSS, NiO_x, and 2% Cu-NiO_x HTLs are displayed in Figure 10. Both NiO_x and 2% Cu-NiO_x exhibited better performance than the PEDOT:PSS HTL devices. The detailed device parameters are listed in

Table 3. Device performance of PEDOT:PSS, NiO_x, and Cu-NiO_x HTL PSCs (the average values were calculated from 12 devices for each HTL material).

HTL		VOC (V)	JSC (mA/cm2)	Fill Factor	PCE (%)
PEDOT:PSS	Average	0.900 ± 0.020	17.919 ± 0.863	0.697 ± 0.015	11.239 ± 0.464
	Best	0.903	18.66	0.71	11.93
NiO	Average	1.015 ± 0.019	19.014 ± 0.456	0.702 ± 0.011	13.539 ± 0.239
	Best	1.015	18.83	0.71	13.72
2% Cu-NiO	Average	1.028 ± 0.003	20.172 ± 0.273	0.744 ± 0.005	15.417 ± 0.118
	Best	1.026	20.57	0.74	15.54

. The 2% Cu-NiO_x device exhibited the highest J_SC value of 20.57 mA/cm², whereas the PEDOT:PSS and pristine NiO_x devices exhibited J_SC values of 18.83 and 18.66 mA/cm², respectively. The J_SC value was higher in the Cu-based device than that in the pristine NiO_x device. This phenomenon was attributed to the increased electrical conductivity and better energy alignment of the HTL to the perovskite layer because of the incorporation of Cu in nickel oxide. The 2% Cu-NiO_x device exhibited a V_OC of 1.026, an FF of 0.74, and a PCE of 15.54%. The 2% Cu-NiO_x NP-based device showed the best performance compared with the pristine NiO_x and PEDOT:PSS devices.

The EQE spectra of the PEDOT:PSS-, NiO_x-, and 2% Cu-NiO_x-based devices are displayed in Figure 11. The EQE spectra of NiO_x and Cu-NiO_x devices were similar, but the Cu-doped device exhibited a higher EQE value than the pristine NiO_x device did. A slight difference in EQE spectra at higher wavelengths for the PEDOT:PSS device was observed. The fabricated devices exhibited satisfactory EQE values in the entire visible region. The higher EQE values in the Cu-doped device indicated that Cu-NiO_x HTLs exhibited better charge transport properties.

The results of EIS performed on the fabricated devices are displayed in Figure 12. The Nyquist plot for the devices fabricated using PEDOT:PSS, NiO_x, and 2% Cu-NiO_x were analyzed in the frequency range from 10 Hz to 1 MHz, and the AC amplitude was 0.01 V under one solar illumination at a bias of V_OC. Typically, the higher-frequency regime corresponded to the charge transport resistance of the MAPbI₃/HTL interface and the lower-frequency regime was attributed to a low-frequency dielectric response of the perovskite materials [48,49]. A low impedance value is more favorable to carrier transport. The EIS results indicated that the Cu-NiO_x devices exhibited the lowest impedance value. This result was consistent with the J–V and EQE results.

The Cu doping in NiO_x increased the electrical conductivity, which reduced the charge transport resistance at the MAPbI₃/HTL interface and enhanced the band alignment between perovskites and HTLs.

4. Conclusions

We fabricated inverted PSCs using NiO_x and Cu-NiO_x HTLs through spin-coating NP inks. The prepared spin-coated films were uniform and did not exhibit any voids. Cu incorporation into NiO_x did not change its crystalline nature and considerably improved its electrical conductivity. The optical bandgap decreased with an increase in the Cu concentration. Few nanoclusters occurred at higher doping concentrations because of the differences in the calcination temperature between Ni(OH)₂ and Cu(OH)₂. The Cu-NiO_x HTLs exhibited superior PL quenching and the least lifetime decay, which indicated that Cu-NiO_x exhibited superior charge transport compared with pristine NiO_x HTLs. The EQE in the Cu-doped device throughout the visible region was higher than that in other devices. EIS analysis revealed the incorporation of Cu in NiO_x HTL reduced the interface impedance. The fabricated NiO_x and 2% Cu-NiO_x PSCs exhibited PCEs of 13.72% and 15.54%, respectively.