Runtao
Wang
,
Lin
Xie
*,
Tai
Wu
,
Chenghao
Ge
and
Yong
Hua
*
School of Materials and Energy, Yunnan University, Kunming 650091, China. E-mail: l.xie@ynu.edu.cn; huayong@ynu.edu.cn
First published on 8th February 2023
The interface between the absorber and transport layers is shown to be critical for highly efficient perovskite solar cells (PSCs). The undesirable physical and chemical properties of interfacial layers often cause unfavorable band alignment and interfacial states that lead to high charge-carrier recombination and eventually result in lower device performance. Herein, we demonstrate a simple and effective strategy to improve the performance of PSCs by constructing a conduction band offset (CBO) with a small spike, through the inductive effect induced by an organic small molecule. As a result, the modified devices show an enhancement in all photovoltaic performance characteristics with a power conversion efficiency (PCE) increase of 10.6% and retaining more than 94% of its initial PCE after 1800 h of exposure to N2. Importantly, we find that a moderate spike-like CBO at the interface between the perovskite film and hole transport layer facilitates rapid charge-carrier injection in devices and reduces charge recombination at the interface, thereby increasing the open-circuit voltage and fill factor. Furthermore, a large spike barrier at the interface increases device resistance, leading to a reduced fill factor. Our present work provides valuable information for understanding the influence of a spike-like CBO on charge-carrier dynamics to further improve the performance and stability of PSCs.
Interface-induced non-radiative recombination is mainly caused by misaligned energy levels, back charge-carrier transfer, and surface defects. The incorporation of organic molecules containing N-, S-, or O-donors has been universally acknowledged to show peculiarities in boosting efficiency and stability based on Lewis acid–base chemistry.21–30 Recently, Hagfeldt et al. discovered that a dipolar interlayer of azetidinium lead iodide (AzPbI3) could increase device performance.31 Yang et al. found that tailoring the dipolar molecules to induce an extra electric field at the heterointerface can improve the performance and stability of the device.32 Our group also found that incorporating a dipolar interlayer that enhances the built-in electric field can enhance charge-carrier injection and minimize non-radiative recombination in PSCs.33–36 Although attempts through interface engineering have been successfully demonstrated to reduce the defect density,37–39 it has certain limitations in terms of inevitably massive and diverse defects.
Apart from defect passivation, effective interface engineering should also take care of the selective extraction of majority carriers and block minority carriers. An optimal interface should be designed to ensure efficient extraction of majority carriers while blocking minority carriers. For instance, the different bandgaps of the hole transport layer (HTL) and perovskite realize an energy cliff or spike at the interface. While the holes of HTL could easily overcome the barrier to return to the interface. Furthermore, the spike in the conduction band at the interface can cause a desirable increase in the barrier to the back injection of electrons of the ETL into the interface under forward bias. Therefore, it is imperative to develop strategies to optimize the heterointerface in perovskite films to reduce the charge recombination and defects to further increase the device performance. Moreover, an understanding of the energy band alignment at the interface and the charge-carrier dynamics between the perovskite film and the charge-transporting layers is still lacking.
In this work, we present a simple modification strategy of a moderately spiked CBO for highly efficient PSCs using triflic anhydride ((CF3SO2)2O, Tf2O) via an inductive effect. Through prudent control over the concentration of Tf2O, the interfacial energetics and interfacial defects can be effectively managed. This systematic study demonstrates the influence of the inductive effect on manipulating the spike-like CBO of modified perovskite films and highlights the importance of a spike-like CBO in reducing the charge-carrier recombination at the interface compared to a cliff CBO. Our strategy yields a significant enhancement in open-circuit voltage (VOC) and fill factor (FF) in the resultant PSCs with a PCE exceeding 21%, retaining more than 90% of its initial PCE despite exposure to N2 for more than 1800 h. Our findings offer a feasible and effective strategy to construct spike-like energy band alignment at the heterointerface in highly efficient perovskite solar cells.
To gain insight into the intermolecular interaction between Tf2O and FAI, 1H nuclear magnetic resonance spectroscopy (1H NMR) was performed, as shown in Fig. 1f. For FAI, two singlet peaks at δ = 7.91 and 8.85 ppm in the NMR spectra correspond to the protons of –CH and –NH2 of pure FAI. These two single peaks imply one uniform chemical environment for the –CH and –NH2 groups in FAI. When FAI was mixed with Tf2O, the two peaks show chemical shifts compared to the proton peaks of FAI, suggesting an inductive effect due to F⋯NH2 hydrogen-bonding interaction between Tf2O and FAI. Intriguingly, the signals for –CH and –NH2 in the FAI spectra are split into multiplets and two peaks by adding Tf2O, which can be ascribed to the positive charges in FAI distributed uniformly in the N–C–N region being polarized due to the formation of a hydrogen bond between CNH and –CF3, resulting in a change in the electronegativity of the two nitrogens. The advantage of the inductive effect is that the polarized FAI can anchor hydrogen iodide to the secondary amine, which can stably move the iodide ion. The intermolecular interaction between Tf2O and FAI is illustrated in Fig. 1g. The interfacial interaction of Tf2O toward the uncoordinated Pb cations and organic cations in the perovskite film would be favorable to the mitigation of non-radiative recombination at the interface, as depicted in Fig. 1h. These results suggest that the inductive effect can stabilize iodide ions on the perovskite film surface, change the interfacial energetic of the perovskite film, and efficiently suppress the degradation of the perovskite film to a certain extent.
To investigate the influence of Tf2O on device performance, we fabricated PSCs with Tf2O-modified perovskite films on a mesoporous device architecture consisting of the following layers: fluorine-doped tin oxide (FTO)/titanium oxide (TiO2)/(CsPbI3)0.05(FA0.85MA0.15Pb(I0.85Br0.15)3)0.95/Tf2O/Spiro-OMeTAD/Au. Considering the insulating property of Tf2O, we carefully regulated the concentration of Tf2O to control the Tf2O residue on top of the perovskite films, as shown in Fig. 2a. The corresponding photovoltaic parameters are shown in Fig. S2 (ESI†). The statistical distribution of PCE values from 18 devices each for the control, Tf2O-1, and Tf2O-5 modified devices confirms that the enhancement in photovoltaic performance arises from the introduction of Tf2O to the perovskite film surface. The optimal concentration for the Tf2O-treated PSCs is obtained at 1 mM. Fig. 2b presents the current density–voltage (J–V) curves of the control and Tf2O-modified PSCs measured under standard AM 1.5G illumination in ambient conditions. The champion control device shows VOC, current density (JSC), FF, and PCE values of 1.088 V, 24.13 mA cm−2, 72.84%, and 19.14%, respectively (Table 1). 1 mM Tf2O-modified PSCs (referred to as Tf2O-1) achieved an overall enhancement in JSC up to 24.68 mA cm−2, a VOC of 1.111 V, and FF of 78.92%, resulting in a remarkable higher PCE of 21.64%. Devices treated with 5 mM Tf2O (referred to as Tf2O-5) yield a lower PCE of 20.45% with a JSC of 24.12 mA cm−2, a VOC of 1.118 V, and a reduced FF of 75.84%. The incident photon-to-electron conversion efficiency (IPCE) spectra of the fabricated devices are shown in Fig. 2c. The integrated JSC for the control, Tf2O-1, and Tf2O-5 modified PSCs are found to be 23.88 mA cm−2, 24.35 mA cm−2, and 23.93 mA cm−2, respectively. Subsequently, we subjected the control and Tf2O-1 modified devices to a long-term stability test, which was monitored by periodically measuring the photovoltaic characteristics of devices that were stored in an N2-filled glovebox. As shown in Fig. 2d and S3 (ESI†), the Tf2O-1 modified device shows outstanding tolerance, retaining 94% of its initial PCE after 1800 h of storage. The control device shows a 27% reduction in PCE only after 500 h. Additionally, the maximum power point (MPP) tracking for 200 s at a voltage of 0.88 V reveals a stable power output, with a stabilized PCE of 21.40% (Fig. S4, ESI†).
Device | J SC (mA cm−2) | V OC (V) | FF (%) | PCE (%) | ||
---|---|---|---|---|---|---|
Control | Best | 24.13 | 1.088 | 72.84 | 19.14 | |
Average | 24.81 | 1.061 | 71.04 | 18.70 | ||
Tf2O-1 | Best | 24.69 | 1.111 | 78.92 | 21.64 | |
Average | 24.73 | 1.109 | 77.17 | 21.17 | ||
Tf2O-5 | Best | 24.12 | 1.118 | 75.84 | 20.45 | |
Average | 24.33 | 1.109 | 72.28 | 19.50 |
To understand the underlying reason behind the photovoltaic performance of Tf2O-modified PSCs, various characterizations were conducted on the control and Tf2O-modified perovskite films. Surface photovoltage (SPV) measurement was carried out to monitor the changes in interfacial energetics caused by the generation and redistribution of photoexcited charge carriers in semiconductors. As illustrated in Fig. 3a, bandgap-related knees at 1.54 eV, 1.54 eV, and 1.55 eV can be found in control, Tf2O-1, and Tf2O-5 modified films prepared on glass, respectively. It should be noted that the perovskite material used in our study is a weak n-type film:43 the photogenerated electrons spontaneously transfer from bulk to the buried interface, leaving holes and positive ions on the free surface and generating an internal electric field pointing to the buried interface. Fig. 3b depicts the light-intensity-dependent SPV of control, Tf2O-1, and Tf2O-5 modified perovskite films. Notably, the difference in SPV response is attributed to the different concentrations of Tf2O on the perovskite film. The absolute values of SPV in Tf2O-1 modified film (∼90 eV) were significantly higher than in the control film without Tf2O modification (∼40 eV), indicating a greater built-in electric field (BIEF) in Tf2O-1 modified film. By raising the concentration of Tf2O to 5 mM, the absolute value of the SPV is further increased to ∼150 eV. As can be seen in Fig. 3c, there is a clear positive association between SPV and the VOC of Tf2O-modified devices. Thus, the SPV is intimately associated with the related VOC. With Tf2O-1 and Tf2O-5 modification, the VOC of the control device gradually increased from 1.088 V to 1.119 V. Furthermore, we can visualize the Tf2O distribution with nanoscale resolution using Kelvin probe force microscopy (KPFM). As shown in Fig. 3d–f and S5 (ESI†), the WF of control, Tf2O-1, and Tf2O-5 films are calculated to be −375.6 mV, −331.3 mV, and −264.3 mV, respectively. Accordingly, the change in WF in transition metal oxide and perovskite materials is correlated to the cation state and point defect concentrations.44 As shown in Fig. 3d–f, a uniform WF distribution was found in the Tf2O-1 modified film, indicating more effective passivation in comparison to Tf2O-5. Whereas the non-uniform WF of Tf2O-5 film might be attributed to the formation of monohydrate at high concentration levels, as illustrated in the atomic force microscopy (AFM) and scanning electron microscopy (SEM) (Fig. S6, ESI†). These results indicate that adding Tf2O at the perovskite surface raises the valence-band edge of perovskite with respect to the Fermi level. Overall, the surface potential reveals larger band bending at the Tf2O-1 modified interface and a significantly larger total charge depletion width at the heterointerface compared to the control.
According to the conclusions drawn from SPV and WF measured above, a schematic in Fig. 3g is drawn to understand the role of the spike structure in the complete device. It is known that a cliff does not impede the photogenerated electrons; however, it lowers the activation energy for carrier recombination at the interface, which results in a decrease in VOC. The cliff can act as a barrier against injected electrons under forward bias, resulting in the recombination of a majority of carriers via defects at the interface. In contrast, a spike-like CBO was introduced when perovskite was treated with Tf2O, which increases the energy for carrier recombination at the interface, improving the VOC and overall PCE. Interestingly, we also observed a considerable up-shift in WF in Tf2O-5 modified film, which can cause a large CBO at the heterointerface, forming a potential well to trap charges/holes at the heterointerface that traps charges and increases the resistance to hole extraction, thereby reducing the FF. This result is very consistent with a recent study by Yang's group, in which substantial band bending might result in increased resistance and a lower FF.45 Considering the inductive effect of Tf2O, we believe that the changes in spike-like CBO in Tf2O-1 and Tf2O-5 modified films can be attributed to the polarization of the FAI, as demonstrated by the NMR and XPS results.
To investigate the effect of Tf2O on the carrier recombination kinetics and trap states, we subjected the control and Tf2O-1 films to a series of tests. Fig. 4a provides a comparison of the PL intensity between the control and Tf2O-1 modified films. Clearly, the higher photoluminescence (PL) intensity noted for the Tf2O-1 modified film compared to the control implies a reduction in defect density with Tf2O treatment. Lengthening of carrier lifetime, in the absence of a charge extraction layer, is determined through time-resolved photoluminescence (TRPL) studies (Fig. 4b and Table S1†), where the average lifetime of the Tf2O-1 modified film is 587.9 ns, up from 181.1 ns for the control. The trap densities of the control and Tf2O-modified devices determined from space-charge-limited-current (SCLC) (Fig. 4c) are calculated to be 1.2 × 1016 cm−3 for the control and 5.5 × 1015 cm−3 for the Tf2O-1 modified devices, respectively. Additionally, the light-intensity-dependent J–V characteristics were also evaluated. The VOC and the seminatural logarithm of light intensity exhibit a linear fitting relationship. In general, the device exhibits negligible trap-assisted Shockley–Read–Hall (SRH) recombination when the slope is close to 1 kT/q. The slopes are calculated to be 1.55 kT/q and 1.22 kT/q for the control and Tf2O-1 modified devices, respectively, where k is the Boltzmann constant, q is the elemental charge, and T is the temperature. The lower slope of the Tf2O-1 modified device implies a reduction in trap-assisted SRH recombination compared to the control device. Consequently, the treatment of perovskite film with Tf2O-1 can effectively suppress SRH recombination and enhance the carrier lifetime due to the chemical interaction between Tf2O and the perovskite film. In line with the TRPL and SCLC results, transient photovoltage (TPV) decay curves of the modified devices also show an increasing trend from 0.057 ms to 0.085 ms when using Tf2O modification (Fig. S7, ESI†). X-ray diffraction (XRD) measurements were conducted, as shown in Fig. S8 (ESI†). The full width at half-maximum (FWHM) of the XRD peak is estimated from a Gaussian curve fitting. The FWHM of both control and Tf2O-modified film is calculated to be 0.119, indicating that the chemical interaction with Tf2O does not influence the crystallinity of the perovskite films.
We performed nanosecond-(ns) and femtosecond-(fs) transient absorption (TA) measurements to understand the photoinduced carrier dynamics of Tf2O-1 modified perovskite film, including charge-carrier injection and hot hole carrier (HC) extraction. We employed fs-TA to understand the influence of Tf2O on hole transfer at the interface of TiO2/perovskite, as revealed in Fig. 5a–d. All samples of the perovskite film were excited at 475 nm under an excitation energy of 14 μJ cm−2. Fig. S9† exhibits the 2D pseudo-color plots of the fs-TA spectra of perovskite/HTL and perovskite/Tf2O/HTL films. Notably, a photobleaching (PB) band at 766 nm indicates the band state filling between the valence band edge and the conduction band edge of the perovskite film. Scans at various delays extracted from pseudo-color plots are presented in Fig. 5a and b. The high-energy tails gradually become narrower with time due to the hole-carrier cooling process, which can be evaluated with the carrier temperature (Tc). Accordingly, Tc can be extracted by fitting the high-energy tail above the bandgap of the TA spectra with the Maxwell–Boltzmann distribution function of exp (Ef − E/kBTc), where Ef is the Fermi level, and kB is the Boltzmann constant.46,47 As revealed in Fig. 5c, the perovskite/HTL shows a slower HC temperature decay of 3.3 fs, whereas the perovskite/Tf2O/HTL film presents a faster HC temperature decay of 2.3 fs. The faster HC cooling process presented in perovskite/Tf2O/HTL can break the thermal photons of hot carriers, thereby improving the device's performance. The HC relaxation dynamics contain a dominant three-exponential decay component due to the recombination of charge carriers. As shown in Fig. 5d, the fs-TA curves were well fitted by a three-exponential-function equation, I(t) = A1exp(−x/τ1) + A2exp(−x/τ2) + A3exp(−x/τ3), where τ1 refers to charge-carrier trapping at the perovskite interface, τ2 is HC transport from the perovskite to the hole transport layer (HTL), and τ3 is the carrier recombination time.48,49
The fitting details and values for fs-TA are summarized in Table S3.† Clearly, the large portion of τ1 implies that the trap filling process is dominant in the HC relaxation dynamics. The decreased τ1 in the Tf2O-modified films compared to the film without Tf2O indicates a faster trap filling process, suggesting a larger splitting in the quasi-Fermi energy levels and a larger charge depletion width in the perovskite film, which is consistent with the results of SPV measurement. Furthermore, the decreased τ2 and τ3 indicate that the entire recombination process can be effectively suppressed with the help of Tf2O. Therefore, these results suggest that the dipolar layer of Tf2O can provide a thermodynamic driving force by tuning the heterointerface, thus promoting hole transport dynamics. To understand the modified heterointerface on the charge-carrier transport, we also studied the ns-TA measurement perovskite with and without Tf2O prepared on glass/TiO2 substrates. Similarly, a photobleaching peak at 766 nm is obtained in the ns-TA spectra (Fig. 5e and f). The corresponding kinetic decay traces of PB at 766 nm were studied. The lifetime of the perovskite films with electron transporting layers (ETLs) was fitted by a single-exponential function, as shown in Fig. 5g. The decay lifetime values of control and Tf2O-modified film are calculated to be 577.6 ns and 515.6 ns, respectively. Therefore, the ETL/perovskite interfacial charge recombination velocity can be evaluated to be 1.7 × 107 and 1.94 × 107 s−1 based on κ = 1/τ.50 The decreased charge recombination velocity of Tf2O-modified film demonstrates the fast charge-carrier injection from the perovskite film to ETLs and thus effectively reduces the charge aggregation and non-radiative recombinations at the ETL interface. As a result, both the fs-TA and ns-TA measurements confirm that the Tf2O-induced spike-like CBO can efficiently promote the photoinduced charge/hole transfer at the interfaces, which could provide a reasonable explanation for the improved VOC, FF, and PCE of the Tf2O-modified device. Fig. 5h depicts the proposed mechanism of spike-like CBO offered by Tf2O. A modest favorable spike-like CBO rises in the Tf2O-modified perovskite surface, acting as a barrier to suppress the interfacial non-radiative recombination and enhancing the hot hole carrier extraction at the interface, thus enhancing the device performance and stability.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2sc06499e |
This journal is © The Royal Society of Chemistry 2023 |