Atomic Layer Deposition of SnO 2 as an Electron Transport Material for Solid-State P ‑ type Dye-Sensitized Solar Cells

: Tin oxide (SnO 2 ) as an electron transport material was prepared by atomic layer deposition in dye-sensitized NiO films to fabricate solid-state p-type dye-sensitized solar cells using two organic dyes PB6 and TIP as photosensitizers. Due to the excellent electron mobility and satisfactory penetration of SnO 2 material into the NiO film, a record photocurrent density over 1 mA cm − 2 was achieved with a power conversion efficiency of 0.14%. The effect of an inserted Al 2 O 3 layer between the dye-sensitized NiO and SnO 2 layer on photovoltaic performance of the devices was also investigated. The results suggest that the charge recombination between NiO and SnO 2 can be significantly suppressed, showing prolonged charge lifetime and enhanced photovoltage.


■ INTRODUCTION
Since the report of pioneer work by O'Regan and Graẗzel and co-authors in 1991, 1 dye-sensitized solar cells (DSSCs) have attracted great attention. 2The traditional DSSCs, also called ntype DSSCs, are based on dye-sensitized TiO 2 photoanodes, which have experienced rapid progress during the past three decades. 3,4In 1999, Lindquist and co-authors reported p-type DSSCs with a liquid electrolyte in which dye-sensitized NiO films were used as a photocathode. 5The photocathode plays a vital role in tandem DSSCs, which could reach a theoretical power conversion efficiency (PCE) of more than 30%.However, the development of p-type DSSCs is still far behind that of n-type DSSCs. 6Various studies have been focused on improving the performance of p-type DSSCs 7−11 because of their good potential in tandem solar cells and solar fuel devices. 12,13he traditional liquid p-type DSSCs consist of a dyesensitized mesoporous NiO photocathode, an I − /I 3 − redox couple, and a counter electrode.Due to the small energy difference between the valence band (VB) of NiO and reduction potential of I − /I 3 − , the open-circuit voltage (V oc ) of p-type DSSCs is normally very low, less than 200 mV. 14,15In addition, the severe charge recombination in liquid p-type DSSCs also leads to the low V oc and poor fill factor (FF). 16,17 Moreover, the stability of the liquid device due to potential electrolyte leaking is always a concern.Besides the good stability, the solid-state p-type DSSCs can achieve high voltage because of high energy difference between the VB of NiO and the conduction band (CB) of electron transport material (ETM).In 2014, our group 18 reported the first example of solid-state p-type DSSCs using P1 dye as a photosensitizer and PCBM as ETM, which achieved a V oc of 620 mV.A slow charge recombination between holes in NiO and electrons in PCBM − was observed.However, the photocurrent density (J sc , 50 μA cm −2 ) of such a device was extremely low because of the low electron mobility and thick layer of PCBM.Other diketopyrrolopyrrole derivatives have also been used in the PCBM-based solid-state p-type DSSCs, showing an enhanced J sc of 0.45 mA cm −2 because of their strongly reducing ability and long lifetime of the charge-separated state. 19−23 Atomic layer deposition (ALD) of TiO 2 has been proved to be able to enhance ultrafast charge separation and dye regeneration (t 1/2 ≤ 500 fs) in a core−shell NiO-dye−TiO 2 film. 21Although the photocurrent density of the solar cell using ZnO as ETM has made a big improvement, 23 the J sc of solid-state p-type DSSCs was still lower than 1 mA cm −2 , which was much lower than that of liquid p-type DSSCs.According to our previous study, the fact that a majority of dyes are not entirely electrically connected with the ETM contributes to the low J sc of solidstate p-type DSSCs. 24These results encouraged us to further explore other suitable metal oxides as ETMs and improve photovoltaic parameters. 25in oxide (SnO 2 ) is an n-type metal oxide semiconductor with high electron mobility and stability, which has been used as ETM in perovskite solar cells. 26,27Its electron mobility can reach as high as 410 cm 2 V −1 s −1 . 28In this work, we employed SnO 2 as ETM in solid-state p-type DSSCs.To make good penetration of SnO 2 into dye-sensitized NiO films, the ALD method was employed.By extending the exposure and purge time in the process of ALD of SnO 2 , good contact between dye-sensitized mesoporous NiO films and SnO 2 can be achieved, which is helpful for efficiently extracting electrons from the dyes after hole injection into NiO.A record J sc over 1 mA cm −2 was reached for solid-state p-type DSSCs.In addition, introduction of ALD Al 2 O 3 between the dyesensitized NiO and SnO 2 layer can suppress the charge recombination, resulting in an obvious enhancement of the photovoltage.
■ EXPERIMENTAL SECTION Materials.PB6 21 and TIP 23 dyes were reported in our previous work.Fluorine-doped tin-oxide coated glass (FTO, TEC15) was purchased from Pilkington.tert-Butanol, acetonitrile, zinc powder, and hydrochloric were purchased from Sigma-Aldrich.Aluminum (Al) wire was purchased from Umicore.
Device Fabrication.The FTO conductive glass substrates were etched with 2 M hydrochloric and zinc powder.The etched FTO substrates were washed with detergents, deionized water, acetone, and ethanol using an ultrasonic bath.Then, the cleaned FTO were dried with flow N 2 .A compact NiO blocking layer (80 nm) was deposited on the FTO substrate by reactive magnetron sputtering.The deposition condition is the same as our previous work. 21Before screen printing the mesoporous NiO film, the sputtered NiO blocking layer was annealed in an oven at 450 °C for 0.5 h in an ambient atmosphere.The NiO paste was prepared according to the previous literature. 22Then, one layer of mesoporous NiO films with a thickness of 750 nm was deposited on the blocking layer by the screen printing method.The screen-printed NiO film was transferred into an oven and sintered at 450 °C for 30 min.After cooling to room temperature, the mesoporous NiO films were dipped into a 0.2 mM dye (PB6 or TIP) tert-butanol/acetonitrile (v/v = 1:1) solution for 16 h.Subsequently, the sensitized NiO film was rinsed with ethanol and dried with N 2 flow.The ETM SnO 2 was deposited in the NiO film by the ALD method.The back-contact (Al) electrode was deposited on the ALD SnO 2 film with a thickness of 120 nm.For NiO/dye/Al 2 O 3 / SnO 2 -based devices, the point Al 2 O 3 was inserted after NiO sensitization but before ALD of SnO 2 .
ALD of Al 2 O 3 and SnO 2 .The ALD employed a stop-flow process in the project.The precursors for ALD of Al 2 O 3 were composed of trimethylaluminum (TMA) and deionized water (H 2 O).The temperature of the chamber was 120 °C with a stability time of 1 h before the ALD process.In total, 10 cycles of Al 2 O 3 were deposited on the NiO film.
The stop flow involved a sequence of (1) reduced carrier gas flow, (2) reduced pumping speed through a bypass constriction, (3) precursor injection, (4) precursor exposure, (5) restored pumping speed, and (6) restored carrier gas flow and purging.For the TMA precursor, the duration of steps 1 to 6 were 2, 2, 0.2, 21.8, 2, and 2 s.During this sequence, the carrier gas flow was reduced from 40 to 10 standard cubic centimeters per minute (sccm) in each of the five unused precursor lines of the system.The TMA line used a flow of 150 sccm during the injection step.In the H 2 O half-cycle, the duration of steps 1 to 6 was the same, that is, 2, 2, 0.

■ RESULTS AND DISCUSSION
The device structure of the solid-state p-type DSSCs, which is composed of electrodes (FTO/NiO), dyes (PB6 or TIP), ETM (ALD SnO 2 ), and back-contact material (Al), is shown in Figure 1.The energy levels of components used in solidstate p-type DSSCs are illustrated in Figure 1a.Under light illumination, the excited dye injects hole into the NiO VB, and then, the reduced dye injects electron into the SnO 2 CB.Two dyes, PB6 and TIP, with a D-π-A structure (Figure 1b), but different in the π linkers, are used as the photosensitizers.The energy levels of NiO, dyes, and Al 2 O 3 are taken from our previous work. 20,23For SnO 2 , the energy levels are taken from the literature, 29 which has used a similar method to prepare ALD SnO 2 .
The detailed process of ALD of SnO 2 into mesoporous NiO films is shown in Figure S1.According to our previous work, 20 the good core−shell structure of NiO and ALD TiO 2 can show ultrafast hole and electron injection, which is beneficial for the dye regeneration.In order to have good penetration of SnO 2 into the mesoporous NiO film, an extended precursor exposure time was applied in this work (see details in the Experimental Section).The process of ALD of Al 2 O 3 is similar to ALD of SnO 2 .
In order to observe the morphology of the films and check the status of SnO 2 penetration, scanning electron microscopy (SEM) is used to characterize the morphology of different layers.The SEM image of the sputter-deposited NiO blocking layer film is shown in Figure 2a.One can see a very dense and uniform NiO film, which is beneficial to effectively block the unwanted contact between FTO and SnO 2 , avoiding short circuit from the contact of the back-contact and FTO surface.A more intimate contact between SnO 2 and NiO (near to dye) can not only promote charge separation but also increase charge recombination, while use of a thin Al 2 O 3 layer can suppress such charge recombination.From our previous study, 20 a thin ALD Al 2 O 3 layer between dye and NiO did not influence the hole injection and a thin ALD Al 2 O 3 layer between NiO and ETM did not influence fast electron injection from the dye to ETM.In this work, we sensitized dye first and then used ALD Al 2 O 3 , so the electronic coupling between NiO and the dye should not be significantly changed.The satisfactory penetration of SnO 2 and Al 2 O 3 into mesoporous NiO films is therefore expected to accelerate charge separation and suppress charge recombination, thereby benefitting photovoltaic parameters.
To evaluate the photovoltaic performance of the solid-state p-type DSSCs, two kinds of devices, with or without Al 2 O 3 layers, were assembled.For the device of FTO/NiO/dyes/ SnO 2 /Al, the photocurrent density−photovoltage (J−V) curves are shown in Figure 3a and their photovoltaic parameters are listed in Table 1.A PCE of 0.086% was achieved with a V oc of 0.199 V, a J sc of 1.082 mA cm −2 , and a FF of 0.397 when PB6 dye was used in the device.Using TIP dye instead of PB6 dye, the FTO/NiO/TIP/SnO 2 /Al-based device reached a PCE of 0.141% with a V oc of 0.269 V, a J sc of 1.141 mA cm −2 , and a FF of 0.458.The J sc of the FTO/NiO/ TIP/SnO 2 /Al-based device is a record for the solid-state ptype DSSCs.When the dyes were varied in FTO/NiO/dyes/ SnO 2 /Al-based devices, the TIP dye-based device attained higher V oc and FF values than those of the PB6 dye-based device.Such results can be attributed to the π bridge in TIP dye having a stronger ability to suppress the charge recombination between holes in the NiO VB and electrons in the SnO 2 CB in the solar cell than the PB6 dye does.The IPCE spectra of FTO/NiO/dyes/SnO 2 /Al-based devices are shown in Figure 3b.The maximum value of IPCE is beyond 8%, which is a big progress for solid-state p-type DSSCs.Although the J sc values of FTO/NiO/dyes/SnO 2 /Al-based devices are largely improved, the V oc of such devices is still lower than what we reported previously using other metal oxides as ETMs. 20,22The positive CB of SnO 2 and unwanted charge recombination may contribute to the low V oc .To suppress the charge recombination, an ALD Al 2 O 3 layer with 10 cycles (about 1 nm) was deposited on the NiO/dyes film prior to ALD of SnO 2 to form the FTO/NiO/dyes/Al 2 O 3 / SnO 2 /Al-based device.For the FTO/NiO/PB6/Al 2 O 3 /SnO 2 / Al-based device (Figure 3c), a V oc of 0.381 V was reached, which is much higher than that of the FTO/NiO/PB6/SnO 2 / Al-based (V oc = 0.199 V) device without Al 2 O 3 .A similar phenomenon can be observed for the TIP dye-based device, which achieved a V oc of 0.423 V with Al 2 O 3 in the device.The theoretical photovoltage of the device based on SnO 2 ETM is 0.5 V determined from the potential difference between the VB of NiO and the CB of SnO 2 .These results indicate that the ALD of Al 2 O 3 can effectively suppress the charge recombination and therefore enhance the V oc of solid-state p-type DSSCs.In addition, the FF (0.541) of the TIP dye-based device is also enhanced.However, the J sc of the FTO/NiO/dyes/Al 2 O 3 / SnO 2 /Al-based device is much lower than that of the device without Al 2 O 3 , and the maximum IPCE (Figure 3d) value of FTO/NiO/dyes/Al 2 O 3 /SnO 2 /Al is reduced to about 1%.In our previous work, 20 we did not observe that the hole injection from the dye was significantly affected by introduction of a thick Al 2 O 3 layer.
In our previous work, we found that the electron injection from the reduced dye to NiO surface states is an interface charge transfer pathway. 30Considering the energy levels of NiO surface states (ca.+0.6 to −0.4 vs NHE) and SnO 2 CB (0 vs NHE), the electron transfer from the injected electrons in NiO surface states to the SnO 2 CB is also thermodynamically feasible.Therefore, besides direct electron injection from the Table 1.Photovoltaic Parameters of Solid-State p-Type DSSCs (Light Intensity: AM 1.5G, 100 mW cm −2 ; Active Area: 0.125 cm −2 ) reduced dye to SnO 2 , the electron injection via NiO surface states to SnO 2 could be another pathway.However, when one cycle of Al 2 O 3 is applied to passivate the majority of the NiO surface states, the direct electron transfer from the reduced dye to SnO 2 should dominate the charge separation process.A mechanism scheme of interface charge separation is shown in Figure 4.In addition, the conductivity of the NiO film is decreased obviously after inserting Al 2 O 3 , 30 which could have a negative effect on the J sc as well.The reduced J sc could be therefore explained by the removal of NiO surface states by Al 2 O 3 which influences interface charge transfer and the hole transport in NiO films. 30o elucidate the variation of V oc and J sc after ALD of Al 2 O 3 , transient photovoltage/photocurrent techniques are used to analyze the process of charge transport.The relation between charge lifetime and photovoltage is shown in Figure 5a.The V oc is adjusted with different light intensities.As shown in Figure 5a, the FTO/NiO/PB6/Al 2 O 3 /SnO 2 /Al-based device has a much longer charge lifetime than that of the FTO/NiO/ PB6/SnO 2 /Al-based device, which suggests that inserting Al 2 O 3 into the device can effectively suppress the charge recombination and then improve photovoltage of the device. 24,30The same trend can be observed in Figure 5a for the TIP dye-based device.In addition, the charge lifetime of FTO/NiO/TIP/Al 2 O 3 /SnO 2 /Al is longer than that of the FTO/NiO/PB6/Al 2 O 3 /SnO 2 /Al-based device.These results indicate that the TIP dye with indacenodithieno[3,2-b]thiophene as a π linker has a stronger ability in suppressing the charge recombination than the PB6 dye with thiophene as a π linker. 23As a result, the TIP dye-based device can reach higher V oc than that of the PB6 dye-based device under the same condition (with or without Al 2 O 3 ), which is consistent with previous reports. 23,31To explain why the FTO/NiO/ dyes/SnO 2 /Al-based device has a much higher J sc than the FTO/NiO/dyes/Al 2 O 3 /SnO 2 /Al-based device, the dependence of charge transport time on photocurrent is measured. 32,33As displayed in Figure 5b, the FTO/NiO/dyes/ SnO 2 /Al-based device has a much shorter charge transport time than that of the FTO/NiO/PB6/Al 2 O 3 /SnO 2 /Al-based device, which can explain the higher photocurrent density obtained from the devices without Al 2 O 3 layers.After tracking the PCE of the FTO/NiO/PB6/SnO 2 /Al-based device for above 1300 h under the air condition in dark, the device can maintain over 88% of the highest PCE (Figure S7).
To assess effects of thickness of the Al 2 O 3 layer on the performance, three different thicknesses of the Al 2 O 3 layer (0.1, 0.5, and 1 nm) were prepared.The J−V curve is shown in Figure 6 and the corresponding photovoltaic data are collected in Table S3.In this paper, we used the ALD method we reported 30 previously to remove NiO surface states.As a result, a very obvious decrease in J sc is observed by inserting one cycle (0.1 nm) of Al 2 O 3 because of NiO surface states, which has been suggested to be the one of hole transport channels in the previous studies. 34With the increased thickness of the Al 2 O 3 layer, the J sc is decreased by a low degree.In our previous work, we used transient absorption spectroscopy to measure the ultrafast regeneration of the NiO/dye/Al 2 O 3 /TiO 2 -based system.We found that the introduction of one cycle of Al 2 O 3 did not influence the fast electron injection from dye to the   TiO 2 CB. 20,21Therefore, in the NiO/dye/Al 2 O 3 /SnO 2 -based system, the electron injection in the CB of SnO 2 should be also efficient enough.However, the V oc is improved with the increased thickness of Al 2 O 3 due to suppression of charge recombination between the injected holes in NiO and injected electrons in SnO 2 .The trend is different from our previous result when we used TiO 2 as ETM. 20There could be two possibilities: (i) The ALD method we used previously in TiO 2 work did not completely remove the NiO surface states, which is evident from the charge transport time in our previous work, 20 where we observed that the charge transport time is constant and not dependent on light intensity.With the removal of NiO surface states, the charge transport time is dependent on light intensity as shown in Figure 5 in this work.The same trend can be also observed in our previous work with the similar ALD method. 30(ii) In SnO 2 -based devices, the injected electrons in NiO surface states by excited dye transfer to the SnO 2 CB play a crucial role in charge separation.Introduction of Al 2 O 3 passivates the majority of NiO surface states and therefore significantly decreases J sc .The decrease in NiO conductivity should be also responsible for the decreased J sc .Also, according to our previous work, 30 the dye sensitization can protect some surface states from ALD Al 2 O 3 .When the thickness of Al 2 O 3 is increased, the electron transfer from the protected NiO surface states to the SnO 2 CB will be slowed down, causing decreased J sc .This is different from the TiO 2 -based device, where the injected electrons in NiO surface sates cannot be efficiently used due to more negative potential of the TiO 2 CB than that of NiO surface states.Prasittichai and Hupp reported SnO 2 /Al 2 O 3 in liquidelectrolyte DSSCs 35 and used one cycle of ALD of Al 2 O 3 to passivate the SnO 2 -based photoanode, which improved the V oc , FF, and electron lifetime very obviously.Their result is in consistence with our result in this work.

■ CONCLUSIONS
In summary, SnO 2 ETM is prepared by the ALD method in a dye-sensitized mesoporous NiO film for solid-state p-type DSSCs.By extending the exposure time in the ALD process, good penetration of SnO 2 into the mesoporous NiO film can be realized, which in turn permits good contact formation with dye-sensitized NiO particles.As a result, a photocurrent density above 1 mA cm −2 is achieved using the FTO/NiO/ dyes/SnO 2 /Al-based device, which is a record for the solidstate p-type DSSCs.The FTO/NiO/TIP/SnO 2 /Al-based device can reach a PCE of 0.14%.In addition, introduction of Al 2 O 3 into the device can largely reduce the charge recombination to enhance the V oc .This work shows the potential to further improve the efficiency of solid-state p-type DSSCs by optimizing ETMs and dyes.

* sı Supporting Information
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsaem.2c01328.Process of ALD; device structure; SEM; absorption of dye on NiO films; J−V curves; stability of the device; and photovoltaic parameters of solid-state p-type DSSCs (PDF) 2, 21.8, 2, and 2 s, while the H 2 O carrier gas flow was increased to 150 sccm and the other precursor line flows were reduced to 10 sccm.The precursor of ALD SnO 2 was composed of tetrakis-(dimethylamido)tin (TDMASn) and deionized water (H 2 O) and used the same stop-flow time sequence as for the Al 2 O 3 ALD.The carrier gas (N 2 ) flow was 200 and 200 sccm during the TDMASn and H 2 O injection steps, respectively.In all, 100 cycles of SnO 2 were deposited on the NiO film.A slide of Si wafers was used to evaluate the thickness of SnO 2 and Al 2 O 3 with a spectroscopic ellipsometer.Device Characterization.The sputtering of NiO films was performed on a Von Ardenne magnetron sputter system (CS730S) with a Ni target.The ALD was performed on a new PICOSUN R-200 advanced ALD system.The back-contact layer (Al) was deposited on a Lesker PVD 75 evaporator.Current density−voltage (J−V) curves were measured using a Newport solar simulator (model 91160, AM 1.5 G) equipped with a Keithley 2400 source meter.Before measurement, the light intensity (100 mW cm −2 ) was calibrated with a calibrated filter Si cell.The area of the mask was 0.125 cm 2 during the measurement.The incident photon-to-current efficiency (IPCE) spectrum was measured with a setup consisting of a xenon lamp (Spectral Products ASB-XE-175), a monochromator (Spectral Products CM110), and a Labjack U6 DAQ board.The IPCE device was calibrated with a Si solar cell before measurement.Transient Photocurrent and Photovoltage Measurements.The transient photocurrent and photovoltage measurements were conducted using a Dyenamo tool-box DN-AE01 system.The light intensity (white LED light, Luxeon Star 1 W) was adjusted by

Figure 1 .
Figure 1.(a) Energy levels of components used in p-type solid-state DSSCs with ALD SnO 2 as ETM.(b) Molecular structures of PB6 and TIP dyes.

Figure
Figure 2b displays the SEM image of the mesoporous NiO film that is helpful for loading more dyes.After 100 cycles of ALD of SnO 2 (about 16 nm) layers into the mesoporous NiO film, the morphology of the NiO/ALD SnO 2 film still remains porous from the SEM image in Figure 2c but becomes denser.In order to check if the ALD SnO 2 and Al 2 O 3 have good penetration into the mesoporous NiO film, energy-dispersive X-ray spectroscopy (EDX) of the cross-section SEM image (Figure 2d) of FTO/NiO/Al 2 O 3 /SnO 2 has been carried out.The result shows a uniform distribution of SnO 2 in the mesoporous film, indicating good penetration of ALD SnO 2 into the NiO film.Due to the presence of Sn in the FTO substrate, Sn can be also observed in the part of FTO in Figure 2f.The uniform distribution of Al 2 O 3 in the mesoporous NiO film indicates better penetration than that in our previous work, 20 likely due to longer exposure time in the process of ALD.A more intimate contact between SnO 2 and NiO (near to dye) can not only promote charge separation but also increase charge recombination, while use of a thin Al 2 O 3 layer can suppress such charge recombination.From our previous study, 20 a thin ALD Al 2 O 3 layer between dye and NiO did not influence the hole injection and a thin ALD Al 2 O 3 layer between NiO and ETM did not influence fast electron injection from the dye to ETM.In this work, we sensitized dye first and then used ALD Al 2 O 3 , so the electronic coupling between NiO and the dye should not be significantly changed.The satisfactory penetration of SnO 2 and Al 2 O 3 into mesoporous NiO films is therefore expected to accelerate charge separation and suppress charge recombination, thereby benefitting photovoltaic parameters.To evaluate the photovoltaic performance of the solid-state p-type DSSCs, two kinds of devices, with or without Al 2 O 3 layers, were assembled.For the device of FTO/NiO/dyes/ SnO 2 /Al, the photocurrent density−photovoltage (J−V)

Figure 2 .
Figure 2. (a) SEM image of the sputter-deposited NiO blocking layer on the FTO substrate, (b) SEM image of mesoporous NiO films; (c) SEM image of atomic layer-deposited SnO 2 film on the FTO/NiO film; (d) cross-section SEM image of FTO/NiO/PB6/ALD Al 2 O 3 /ALD SnO 2 ; and (e−g) EDX element mapping of Ni, Sn, and Al.

Figure 4 .
Figure 4. Possible process of electron transfer before and after ALD of Al 2 O 3 .

Figure 5 .
Figure 5. (a) Relation between charge lifetime and voltage for NiO/dye/SnO 2 /Al-and NiO/dye/Al 2 O 3 /SnO 2 /Al-based devices and (b) relation of charge transport time with current.