Low-Temperature (<40 °C) Atmospheric-Pressure Dielectric-Barrier-Discharge-Jet Treatment on Nickel Oxide for p-i-n Structure Perovskite Solar Cells.

A scan-mode low-temperature (<40 °C) atmospheric-pressure helium (He) dielectric-barrier discharge jet (DBDjet) is applied to treat nickel oxide (NiO) thin films for p-i-n perovskite solar cells (PSCs). Reactive plasma species help reduce the trap density, improve the transmittance and wettability, and deepen the valence band maximum (VBM) level. A NiO surface with the lower trap density surface of NiO allows better interfacial contact with the MAPbI3 layer and increases the carrier extraction capability. MAPbI3 can better crystallize on a more hydrophilic NiO surface, thereby suppressing charge recombination from the grain boundary and the interface. Further, the deeper VBM allows better band alignment and reduces the probability of nonradiative recombination. NiO treatment using He DBDjet with a scan rate of 0.3 cm/s can improve PSC efficiency from 13.63 to 14.88%.

Introduction

The feasibility of low-cost solution processes for realizing high-efficiency perovskite solar cells (PSCs) has made this photovoltaic technology attractive. Such processes have also enabled large-area deposition by a roll-to-roll printing process to further reduce the fabrication cost.[1] In 2009, the first PSC was fabricated with a power conversion efficiency (PCE) of 3.8%.[2] A few years later, a 2,2′,7,7′-tetrakis(N,Ndi-p-methoxyphenylamine)9,9′-spirobifluorene (spiro-OMeTAD) was introduced as a solid-state hole transport layer (HTL).[3] Because of the long diffusion length of the perovskite layer, a high-performance planar PSC without the mesoporous structure of the transport layer was developed.[4−6] In 2013, the first p–i–n structure PSC with a PCE of 3.9% was fabricated; it had a structure similar to that of an organic photovoltaic cell.[7] Owing to benefit including easy fabrication, flexibility, and low hysteresis, many studies have focused on developing p–i–n structure PSCs.[8,9] Several organic molecules were introduced for significantly improving the performance of p–i–n structure PSC.[10−16] However, to improve stability, nickel oxide (NiO) as the HTL of p–i–n structure PSC was launched.[17−20] Nowadays, the PCE of PSCs has been improved to 25.2% with a single-junction architecture and 28.0% with a silicon-based tandem architecture.[21]

Plasma contains abundant ions and electrons that are chemically reactive. It has applied extensively for surface modification and syntheses of materials. Low-pressure plasma (LPP) technology has been used to perform material deposition and surface treatment for PSCs. Plasma-enhanced chemical vapor deposition was used to deposit the electron transport layer (ETL) for an n–i–p structure PSC.[22] Later, a PSC with an ETL deposited by plasma-enhanced atomic layer deposition exhibited a PCE of 19%.[23−25] LPPs have also been used for depositing transparent bottom electrodes.[26,27] For material surface modification, an LPP was used for metal oxide transport layer treatment of an n–i–p structure PSC; an air LPP treatment increased the concentration of oxygen vacancies in TiO2 and improved the PCE.[28] A CO2 LPP was used to improve the surface hydrophilicity and conductivity of a TiO2 layer and to reduce the defect-state density, thereby improving the PCE of the PSC to 15.4%.[29] Another study used a CO2 LPP to treat TiO2 and realized a PCE of 18.1%.[30] Room-temperature nitrogen LPP processing for treating SnO2 electrodes improved PSCs made on glass and flexible substrates.[31] An oxygen LPP was used to treat a NiO film for improving its wettability and deepening its valence band maximum (VBM) level, leading to improved performance of the p–i–n structure PSC.[32] Ecofriendly mild oxygen plasma-treated poly(3,4-ethylenedioxythiophene:poly(styrene sulfonate)) (PEDOT:PSS) has been used as an electrode for indium tin oxide-free PSCs with a PCE of 10.5%.[15] LPP also been used to treat spiro-OMeTAD to realize better band alignment with the perovskite layer in an n–i–p structure PSC.[33] The surface energy and work function of various conjugated polymers were ameliorated by oxygen LPP treatment, and the PCE was improved to a maximum of 19%.[34] Spiro-OMeTAD HTL was shown to be rapidly oxidized by Ar/O2 LPP treatment.[35] The conductivity of treated PEDOT:PSS was increased by mild Ar LPP treatment, leading to a better PSC performance.[36] Ar LPP treatment has also been used for surface composition tuning for fabricating high-efficiency PSCs and fast photodetectors.[37]

LPP is undoubtedly a stable and mature technology that has been used extensively in the fabrication of electronic devices. However, LPP is a vacuum process, making it disadvantageous for integration with solution processes of PSCs. Atmospheric pressure plasma (APP) technology can be operated at regular pressures without using vacuum systems; therefore, it is particularly suitable for integration into the nonvacuum fabrication processes of dye-sensitive solar cells[38−42] and PSCs.[43−45] Typical APP technology includes transfer arc, dielectric barrier discharge (DBD), corona discharge, and APP jet (APPJ). Recent developments have overcome issues such as high breakdown voltage, continuous arcing, and instability, making this technology ready for practical usage.[46−48] APP has a much shorter mean free path and thinner sheath than LPP, and therefore its ion bombardment effect is limited; it is particularly useful for the surface modification of materials that are sensitive to ion bombardment damages. APPs with different electrode designs and plasma excitation methods can be used to realize plasmas with various heavy particle temperatures and electron densities. Caution should be exercised to specify the APP properties when discussing APP technology. Several APP techniques have been applied in PSC fabrication processes. An air APPJ was used to convert scalable and robust perovskite films for PSCs to achieve a remarkably consistent PCE of 15.7% without hysteresis. In this case, an APPJ with a temperature of ∼160 °C was applied to convert the perovskite precursor solution right after it was sprayed.[49] In addition, 5 min atmospheric-pressure Ar/O2 plasma treatment under an almost room temperature environment (<50 °C) has been used for converting the spin-coated SnCl2 precursor into an SnO2 film that was then used as the ETL of an n–i–p PSC with a PCE of 19.56%.[50] APPs were also used for cleaning transparent conducting oxide glass substrates and performing surface modifications of the ETL, HTL, and perovskite layer of PSCs.[43−45,51−53]

Previously, our research team has applied atmospheric-pressure surface-diffusion DBD (SDDBD) for the material modification of MAPbI3 used as the absorbing layer of p–i–n and n–i–p planar PSCs. Proper SDDBD treatment time can enhance the PSC performance, owing to the grain growth of MAPbI3 and slight PbI2 precipitation after SDDBD treatment.[43,45] In these studies, the SDDBD device, a portable plasma generator, was placed inside a nitrogen-filled glovebox to perform the plasma treatment, as illustrated in our previous publication.[44] Surface diffusion-type APP without a proper cooling mechanism is slightly disadvantageous from the viewpoint of temperature control. Further, during the diffusion process, reactive plasma species could react with gas molecules in the environment and lose reactivity. In this study, we instead use a jet-type DBD with two metal-ring electrodes, named DBDjet (see Supporting Information, Figure S1). High-flow rate gas flow cools the plasma to maintain the working temperature below 40 °C (Figure a). This He DBDjet is scanned over the NiO film used as the HTL of a p–i–n PSC. The scan-mode He-DBDjet-treated NiO films are carefully characterized, and the resultant PSC reproducibly shows improved performance.

Figure 1

(a) Working temperature and (b) OES of He DBDjet.

Results and Discussion

In this study, a DBDjet with two metal-ring electrodes is used for scanning over a NiO surface with speeds of 0.1, 0.2, and 0.3 cm/s. The higher the scan rate, the shorter is the plasma-influencing period. Figure S1 shows the DBDjet configuration. He gas was used as the plasma gas, owing to its low breakdown voltage. Figure a shows the temperature time course after igniting the plasma. The temperature increased slightly after plasma ignition, and the steady-state working temperature was below 40 °C. Therefore, in this study, the major effects on the property adjustment of materials are caused mainly by reactive plasma species without substantial heating. Figure b shows the optical emission spectra (OES) of the He DBDjet when treating the NiO film. The peaks in the range of 500–750 nm correspond to He emissions. The peaks in the ranges of 220–270, 280–310, 310–360, and 390–430 nm correspond to CO+ (first negative system, B2Σ – X2Σ), OH (3064 Å system, A2Σ+ – X2Π), N2 (second positive system, C3Πμ – B3Πg), and N2+ (first negative system, B2Σμ+ – X2Σg+), respectively.[54−57] The He DBDjet was operated in an atmospheric environment. The plasma jet also reacted with environmental air, resulting in emissions from species other than He.

Figure shows scanning electron microscopy (SEM) images of the NiO film with/without He DBDjet treatment. Some cracks are observed on the NiO film; these seemed to heal by He DBDjet treatment. As the scan rate decreases, the crack size decreases. The NiO film treated by the NiO DBDjet with a scan speed of 0.1 cm/s was the densest; this benefited the follow-up deposition of the perovskite layer. Figure S2 shows the corresponding atomic force microscopy (AFM) images and roughness values. The roughness was mainly determined by the textured fluorine-doped tin oxide (FTO) substrate; therefore it was not changed by the He DBDjet treatment. Figure S3 shows the grazing incidence X-ray diffraction (GIXRD) results of NiO on the FTO substrate. The sharp peaks are from the FTO glass substrate, and NiO diffraction shows only a weak signal at around 2θ = 43–44°.

Figure 2

Top-view SEM images of NiO films with He DBDjet treatment.

Figure a–c shows the total transmittance spectra, specular transmittance spectra, and absorption spectra of NiO-coated FTO. After He DBDjet treatment, both transmittances increased slightly, and the absorption decreased. The more light the perovskite layer can absorb, the higher is the PCE the solar cell. Therefore, higher transmittance and lower absorption of the NiO-coated FTO are beneficial for PSCs. The overall total transmittances were above 75% at 400–800 nm. Figure S4a,b shows the reflectance and haze spectra. Based on the absorption data, the band gap is estimated to be ∼4.0 eV using the Tauc plot, as shown in Figure d; this agrees well with a previous study.[58]

Figure 3

UV–Vis spectra of the NiO film: (a) total transmittance spectra, (b) specular transmittance spectra, (c) absorption spectra, and (d) Tauc plot.

Figure shows the X-ray photoelectron spectroscopy (XPS) of the O 1s spectra of NiO film with/without He DBDjet treatment, and Table shows the contents of the corresponding deconvolution peaks. The peaks at 529.2, 531.1, 530.8, 531.8, and 532.4 correspond to NiO, Ni2O3, Ni(OH)2, O–C=O, and C–O, respectively. He DBDjet treatment at a scan rate of 0.1 cm/s increased hydrophilic O–C=O and C–O functional groups from 3.51 to 6.48% and from 0.00 to 1.13%, respectively, and this trend is confirmed by the C 1s spectra (Figure S5). This suggests that after DBDjet treatment, some hydrophilic functional groups can attach on the surface NiO to enhance its wettability.[59] After He DBDjet treatment, the hydrophobic Ni(OH)2 content decreased from 17.78 to 15.20%, and the peak in XPS N 1s spectra was also expanded (Figure S7). The ambient nitrogen diffused into the quartz tube of the He DBDjet, and nitrogen-containing free radicals were produced to react with materials.

Figure 4

XPS of O 1s spectra of the NiO film that is (a) as-deposited and with He DBDjet treatment with scan rates of (b) 0.3, (c) 0.2, and (d) 0.1 cm/s.

Table 1

XPS Deconvolution for O 1s Orbital

%	NiO	Ni2O3	Ni(OH)2	O–C=O	C–O
NiO as-deposited	63.84	14.87	17.78	3.51	0.00
NiO He DBDjet 0.3 cm/s	61.77	14.11	17.26	6.86	0.00
NiO He DBDjet 0.2 cm/s	61.38	14.83	16.38	6.65	0.76
NiO He DBDjet 0.1 cm/s	60.43	16.76	15.20	6.48	1.13

The band alignment of each layer of the PSCs can influence the open circuit voltage (Voc).[60] An aligned band structure can suppress carrier recombination at the interfaces.[61] Therefore, ultraviolet photoelectron spectroscopy (UPS) was used to determine the VBM of the NiO film. The VBM of the as-deposited NiO film was estimated to be 4.7 eV. As the He DBDjet influencing time increased (scan rate decreased), the VBM values increased to 4.87, 4.94, and 4.95 with scan rates of 0.3, 0.2, and 0.1, respectively. This agrees with previous studies, indicating that plasma treatment deepens the VBM of metal oxide layers.[28,31,32] Because the VBM of the perovskite layer is ∼5.4 eV,[62] a NiO film with deeper VBM can have better band alignment (Figure ).

Figure 5

UPS spectra of the NiO film with He DBDjet treatment.

Figure shows the water contact angle results of NiO films with/without He DBDjet treatment. Without He DBDjet treatment, the NiO film shows a higher water contact angle of 35°, partly because of the hydrophobic Ni(OH)2, which had been used to construct superhydrophobic nanostructures.[63,64] After He DBDjet treatment, the reduced Ni(OH)2 content and increased O–C=O and C–O contents decreased the water contact angle of the NiO film from 35 to 22°. Higher wettability of the NiO film can improve the crystallinity of the follow-up deposited perovskite layer and ameliorate the interface.[65]Figure S8 shows top-view SEM images of the perovskite layer. Improved perovskite crystallinity can suppress the nonradiative recombination caused by the grain boundaries.[66,67]

Figure 6

Water contact angle images of NiO films with He DBDjet treatment.

Figure a shows the current density–voltage (J–V) curves of PSCs (reverse scan). Table shows the corresponding photovoltaic parameters. Figure S9 shows the forward scan J–V curves. After He DBDjet treatment, the short-circuit current density (Jsc), fill factor (FF), and PCE were improved. The best improvement and PCE are seen in the PSC with NiO treated by He DBDjet at a scan rate of 0.3 cm/s; the PCE was improved from 13.63 to 14.88%. This could be attributed to the better wettability of the NiO film that leads to better crystallinity of the perovskite layer and, in turn, suppresses the charge recombination. The deeper VBM after He DBDjet treatment could also lead to better band alignment for improved PSC performance. The increased Jsc could be attributed to the higher transmittance and lower absorption of NiO-coated FTO substrates after He DBDjet treatment. Figures S10 and S11 show the photovoltaic parameters of six batches of PSCs, and Table S1 lists the statistics of photovoltaic parameters. The PCE of PSCs indeed improved reproducibly after He DBDjet treatment. The most significant improvement was seen for a scan rate of 0.3 cm/s. The average PCE increased from 12.62% (without He DBDjet treatment) to 13.32% (with He DBDjet treatment at a scan rate of 0.3 cm/s).

Figure 7

PSC characteristics with He DBDjet treatment on NiO films. (a)J–V curves with the reverse scan from Voc to Jsc, (b) Nyquist plot, (c) steady-state PL spectra, and (d) TRPL decay profiles.

Table 2

Photovoltaic Parameters of PSCs

		Voc (V)	Jsc (mA/cm2)	FF (%)	PCE (%)
no plasma treatment	forward	1.05	17.35	69.53	12.68
	reverse	1.06	17.16	74.97	13.63
He DBDjet 0.3 cm/s	forward	1.06	18.23	67.56	13.06
	reverse	1.07	18.23	76.19	14.88
He DBDjet 0.2 cm/s	forward	1.06	17.69	69.74	13.11
	reverse	1.07	17.91	76.35	14.65
He DBDjet 0.1 cm/s	forward	1.05	17.65	68.61	12.74
	reverse	1.06	17.68	75.94	14.25

Figure b shows the Nyquist plot of electrochemical impedance spectroscopy (EIS), and the inset of Figure b shows the equivalent circuit to analyze the data. Table S2 shows the EIS fitting parameters. R1, R2, and R3 correspond to the series resistance, charge transporting resistance, and recombination resistance, respectively.[68,69] With the He DBDjet treatment at a scan rate of 0.3 cm/s, R2 decreased from 1073 to 867 Ω; and R3 increased from 7819 to 8151 Ω. With He DBDjet treatment to enhance wettability and to heal cracks, the contact between the NiO film and the perovskite layer was improved; therefore, R2 decreased, and the charge extraction efficiency of the NiO film improved. The increased R3 can be attributed to the deeper VBM of the NiO film that allows better band alignment and to the improved perovskite crystallization that suppresses the nonradiative recombination.

Figure c shows the steady-state photoluminescence (PL) spectra of NiO/perovskite. A better charge extraction effectiveness can suppress carrier recombination at the interface. Before charge recombination, the carrier was extracted into the HTL, causing the quenching effect of the PL spectra.[70,71] The steady-state PL spectra indicate that the NiO film treated by the He DBDjet with a scan rate of 0.3 cm/s has the best carrier extraction capability, whereas the as-deposited NiO film showed poorer carrier extraction capability. Figure d shows the time-resolved PL (TRPL) spectra. Two recombination mechanisms of NiO/perovskite are involved: Shockley–Hall–Read recombination at low injection and nongeminate carrier recombination at high injection.[71] Therefore, the data were fitted with a biexponential function, f(t) = A0 + A1·exp(−t/τ1) + A2·exp(−t/τ2). Table S3 lists the fitting parameters for the TRPL data. Without He DBD jet treatment, the PL decay times of NiO/perovskite were τ1 = 13.60 ns and τ2 = 84.44 ns. After He DBD jet treatment, the PL decay times of NiO/perovskite are τ1 = 9.27 ns and τ2 = 98.18 ns (scan rate of 0.3 cm/s), τ1 = 10.89 ns and τ2 = 77.95 ns (scan rate of 0.2 cm/s), and τ1 = 11.67 ns and τ2 = 75.88 ns (scan rate of 0.1 cm/s). The average recombination lifetime can be estimated by the equation listed in literature.[71] The average lifetime was 24.34 ns for the as-deposited NiO film and decreased to 13.66 ns for the NiO film treated by He DBDjet with a scan rate of 0.3 cm/s. As the scan rate decreases (plasma influencing time increases), the lifetime decreased to 14.70 ns (scan rate of 0.2 cm/s) and 17.16 ns (scan rate of 0.1 cm/s). This result suggests that the He DBDjet-treated NiO can extract carriers faster from the perovskite to the HTL. This confirms that DBDjet treatment on NiO can indeed improve the PSC performance.

Conclusions

An atmospheric-pressure He DBDjet with two metal-ring electrodes is used for treating NiO that is used for p–i–n structure PSCs. He DBDjet treatment improves the wettability and slightly heals cracks of sol–gel-derived NiO. The treatment also deepens the VBM, as evidenced by UPS, to allow better band alignment between the NiO and the perovskite layer. Steady-state PL and TRPL experiments indicate better charge extraction efficiency of the NiO/perovskite. The transmittance of NiO-coated FTO also increases slightly after DBDjet treatment. An EIS experiment indicates reduced transporting resistance and increased recombination resistance after DBDjet treatment. All of these findings support the results of improved PSC performance with NiO treated by a scan-mode low-temperature (<40 °C) atmospheric-pressure He DBDjet.

Experimental Section

He DBDjet setup: Figure S1 shows a schematic of He DBDjet with two-metal ring electrodes. The inner diameter of the quartz tube was 8 mm, and the outer diameter was 10 mm. The two electrodes were fixed outside the quartz tube and separated by a 10 mm gap. The gap between the sample and the bottom of the quartz tube was fixed at 1 mm. The plasma emerged inside the quartz tube in space between the two electrodes and was carried to the sample surface by the He jet flow of 3 slm (stand liter per minute). The power electrode was connected to a high ac voltage with a peak voltage of 10 kV and frequency of 20 kHz. Every sample was treated by the He DBDjet 20 times with scan rates of 0.3, 0.2, or 0.1 cm/s.

PSC fabrication

The FTO glass substrate (TEC7, ∼8 Ω/sq) was used as the substrate and sequentially cleaned in deionized water, acetone, isopropanol, and UV–ozone cleaner for 15 min. The NiO precursor solution was prepared by dissolving 0.5 M nickel acetate (99.998%, trace metal basis, Sigma-Aldrich) and ethanolamine (99.5%, Sigma-Aldrich) in ethanol and stirring overnight at 60 °C; this was the same as that in our previous study.[72] The solution was spin-coated on a FTO substrate at 6000 rpm for 40 s and annealed at 325 °C for 10 min. Then, the NiO film was treated by the He DBDjet. Next, the sample was immediately transferred into a nitrogen-filled glovebox, and the perovskite film was deposited on the NiO film by a one-step process.[44] The perovskite precursor was prepared by dissolving 1.2 mM PbI2 (99.999%, metals basis, Alfa Aesar) and CH3NH3I (MAI, 98%, Dyesol) in dimethylformamide (99.8%, Sigma-Aldrich). After perovskite deposition, PC61BM, doped with DMOAP, and BCP were deposited on the perovskite layer, which served as ETL.[43] Finally, 85 nm Ag with the area of 0.09 cm2 was deposited using an e-beam evaporator as the top electrode.

Characterization

The working temperature of the He DBDjet was monitored by a thermography camera (FLIR, E63900). The surface morphology of NiO and perovskite was inspected by SEM (JOEL, JSM-7800Prime). The surface morphology was probed by AFM (Bruker, BioScope Resolve). The crystallization was surveyed by GIXRD (Bruker, D8 DISCOVER SSS). The transmittance and reflectance spectra were measured by an ultraviolet–visible–near infrared (UV–vis–NIR) spectrophotometer (JASCO, V-670). The surface chemical bonding status was investigated by XPS (Thermo Fisher Scientific, ESCALAB Xi+), and the C–C bond at 284.6 eV was used as the reference peak. The VBM level was measured by UPS (Thermo VG-Scientific/Sigma Probe) with a sample bias of −10 V. The water contact angle was obtained by a contact angle goniometer (Sindatek, model 100SB). The J–V curve of the PSC was measured using a sourcemeter (Agilent, B2902A) under illumination of simulated AM1.5 light (ABET, Sun 2000 Solar Simulator). The electrochemical performance was evaluated by EIS (Metrohm-Autolab, PGSTAT204). To measure the PL and TRPL, a 450 nm laser was emitted using a pulsed diode laser (PicoQuant, PDL 200-B), and an optical measurement system (PSH 3G system) was mounted to detect the spectra.