Solution-based electrical doping of organic photovoltaics with non-fullerene acceptors facilitated by solvent vapor pre-treatment

Molecular doping is an important technique that can enable controlled electrical doping of organic semiconductors. It involves the use of molecular dopants that either accept electrons from (p-type dopants) or donate electrons to (n-type dopants) host organic semiconductors.^1–3 The addition of charge carriers enhances the carrier density and electrical conductivity of the host. Furthermore, electrical doping correlates with a shift in the Fermi level energy of the host that changes energetic barrier heights at interfaces between the host and adjacent layers, such as electrodes in multilayer devices. Therefore, molecular doping of organic semiconductors is widely applied to the fabrication of organic (opto-)electronic devices.^4–6 

In the case of organic photovoltaics (OPVs), molecular p-type (n-type) doping has been used to create a p-doped (n-doped) region adjacent to the electrode to facilitate charge carrier collection.^7–9 Since OPVs are devices with asymmetric electrical characteristics across their thickness, a challenge of using molecular doping is to spatially constrain the electrically doped region to achieve the desired device functionality and stable operation. From a fabrication perspective, achieving a spatially constrained electrically doped region at the interface of an organic semiconducting film, such as an OPV’s photoactive layer, removes the need for using a charge collecting interlayer because it reduces charge recombination in the vicinity of the contact by increasing the hole or electron density in that region. Hence, it allows simplifying the device geometry of an OPV.^10–14 

12-Molybdophosphoric acid (PMA) hydrate is a solution-processable p-type dopant that allows conjugated polymers to be electrically p-doped over a limited depth typically of a few tens of nm from the surface of polymer films via immersion of polymer films in a PMA–nitromethane or PMA–acetonitrile solution. The post-process immersion method has been shown to be applicable to various bulk heterojunction (BHJ) systems, enabling the fabrication of simplified OPV devices without a specific hole-collecting layer that exhibit a comparable photovoltaic performance to reference devices with evaporated MoO₃ as a hole-collecting layer.^10,11

It is worth noting that the BHJ systems processed using the aforementioned PMA doping technique are generally based on polymer donors and fullerene acceptors. However, the power conversion efficiency (PCE) of OPVs has been improved due, in part, to the emergence of non-fullerene acceptors (NFAs).^15–17 In order to apply the PMA doping technique to NFA-based OPVs, different processing approaches and solvents for PMA have been investigated and proposed. On the one hand, Sun et al. presented a technique known as orthogonal liquid–liquid contact (OLLC) doping. In this case, a small amount of a polymer solution was first dropped onto the surface of a PMA–water solution for doping to occur during film formation. Then, the doped film was transferred to the top of a NFA-based photoactive layer as a stand-alone hole-collecting layer.¹⁸ As such, it differs from the previously discussed immersion technique that allows p-type doping of the photoactive layer over a limited depth. On the other hand, Hu et al. presented a post-process immersion method, where a BHJ photoactive layer of PTB7-Th and IEICO-4F was immersed into a PMA–2-propanol solution, leading to hole-collecting layer-free NFA-based OPVs with a PCE of 11.37%.¹² 

Along with non-fullerene OPVs, new conjugated polymers have also been developed as donor materials.^19–21 We have recently reported that the effectiveness of PMA doping in these novel polymer donors is affected by the volume of the dry polymer film. For certain polymer donors, such as PBDB-T, expansion of the film’s volume is necessary for efficient dopant infiltration, which was demonstrated by the following two approaches: The first one was based on heating both the polymer film and the PMA–acetonitrile solution before immersion. An alternative approach was based on solvent swelling, in which a swelling solvent (i.e., chlorobenzene or 1,2-dichlorobenzene) was added to and combined with the PMA–2-propanol solution, followed by spin coating of the blended PMA solution on the polymer film. The doping process was completed after the removal of the dried PMA layer using 2-propanol.²² Although both techniques have been shown to enhance the doping level of PBDB-T using PMA, they could not be applied successfully to device fabrication likely due to the degradation of the BHJ photoactive layer during the doping step.

In this work, we extend the concept of film volume expansion and present a modified PMA doping technique that enables fabrication of hole-collecting layer-free NFA-based OPVs with a BHJ photoactive layer of PBDB-T and IEICO-4F. We modify the previously reported solvent swelling method by optimizing the volume ratio between the PMA solution and the swelling solvent, as well as by implementing an additional step of solvent vapor treatment on the BHJ photoactive layer film before doping. Consequently, the BHJ film is not damaged during PMA doping and the PMA-doped devices yield a PCE of 8.2%, which is comparable to reference devices with an evaporated MoO₃ hole-collecting layer.

The polymer donor poly[(2,6-(4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)-benzo[1,2-b:4,5-b’]dithiophene))-alt-(5,5-(1′,3′-di-2-thienyl-5′,7′-bis(2-ethylhexyl)benzo[1′,2′-c:4′,5′-c’]dithiophene-4,8-dione)] (PBDB-T) and the non-fullerene acceptor 2,2′-[[4,4,9,9-tetrakis(4-hexylphenyl)-4,9-dihydro-s-indaceno[1,2-b:5,6-b']dithiophene-2,7-diyl]bis[[4-[(2-ethylhexyl)oxy]-5,2-thiophenediyl]methylidyne(5,6-difluoro-3-oxo-1H-indene-2,1(3H)-diylidene)]]bis[propanedinitrile] (IEICO-4F) were purchased from 1-Material. 12-Molybdophosphoric acid (PMA) hydrate was purchased from Alfa Aesar. For the zinc oxide (ZnO) precursor solution, zinc acetate dihydrate, ethanolamine, and 2-methoxyethanol were purchased from Sigma-Aldrich. Finally, 2-propanol, 1,2-dichlorobenzene, chlorobenzene, and 1,8-diiodooctane were also purchased from Sigma-Aldrich. All materials were used as received.

Customized 1 × 1 in.² indium tin oxide (ITO)-coated glass substrates with a sheet resistance of 12–15 Ω/sq purchased from MSE Supplies and 1 × 1 in.² glass substrates cut from VWR Micro Slides were used as substrates for PBDB-T films. All substrates were cleaned in sequential ultrasonic baths of Liquinox detergent in deionized water, deionized water, acetone, and 2-propanol at 40 °C for 40 min each. The cleaned substrates were blow-dried with N₂ and transferred into a N₂-filled glovebox for further processing.

PBDB-T was dissolved in chlorobenzene to a concentration of 5 mg/ml. The PBDB-T solution was stirred overnight at 500 rpm at 50 °C inside a N₂-filled glovebox. The PBDB-T films were prepared by spin coating the PBDB-T solution onto glass/ITO and glass substrates at a spinning speed of 1000 rpm, an acceleration of 1000 rpm/s, and a time of 60 s. The films were then solvent-annealed for 1 h.

PMA was dissolved in 2-propanol to a concentration of 0.5M and continuously stirred for 1 h at room temperature inside a N₂-filled glovebox. This concentration was chosen based on our seminal work¹⁰ that showed that electrical doping at that high concentration was not dependent on the exposure time of the semiconductor to PMA. Next, the PMA solution was combined with 1,2-dichlorobenzene at different volume ratios (i.e., 9:1, 19:1, and 99:1). The blended PMA solutions were stirred for 3 h and then spun on top of glass/ITO/PBDB-T and glass/PBDB-T at 1000 rpm, 1000 rpm/s, and 30 s. For solvent vapor treated samples, the PBDB-T films on glass/ITO and glass substrates were first placed inside glass Petri dishes with a few drops of 1,2-dichlorobenzene. Then, the Petri dishes were put on top of a hot plate at 80 °C for 10 min to expose the samples to a vaporized 1,2-dichlorobenzene atmosphere. The blended PMA solution (at a volume ratio of 99:1) was immediately spun on the samples after vapor treatment. All samples were solvent-annealed for 1 h. Finally, the samples were rinsed with 2-propanol to remove PMA residues from the surface of the films.

A profilometer was used to measure the thickness of PBDB-T films. Here, 25 nm was an average over three different spots on glass/PBDB-T.

The UV–Vis–NIR transmittance spectra of pristine and PMA-doped PBDB-T films were measured using a spectroscopic ellipsometer (J. A. Woollam).

The sheet resistance of pristine and PMA-doped PBDB-T films was measured by a four-point probe system using a Keithley 6430 source meter. For each film, the sheet resistance was measured at three different spots.

The work function of pristine and PMA-doped PBDB-T films was measured using a Kelvin probe in air. For each film, the work function was measured at three different spots. The measurements were calibrated using freshly peeled highly ordered pyrolytic graphite (HOPG) with a known work function of 4.6 eV.

Customized 1 × 1 in.² ITO-coated glass substrates with a sheet resistance of 12–15 Ω/sq purchased from MSE Supplies were used as substrates for the solar cells. The substrates were cleaned in sequential ultrasonic baths of Liquinox detergent in deionized water, deionized water, acetone, and 2-propanol at 40 °C for 40 min each and then blow-dried with N₂.

ZnO thin films were derived by using the sol-gel method. The ZnO precursor solution was prepared by dissolving 0.4 g of zinc acetate dihydrate and 0.112 g of ethanolamine in 4 ml of 2-methoxyethanol and stirred overnight at room temperature in air. After treating the cleaned ITO substrates with UV-ozone for 30 min, the ZnO precursor solution was spun on glass/ITO at 4000 rpm, 928 rpm/s, and 30 s through a 0.45 µm PTFE filter. The coated substrates were immediately annealed at 150 °C for 10 min. Then, the hot plate was turned off, while the coated substrates were kept on the plate for another 10 min to allow slow cooling. Then, the coated substrates were transferred into a N₂-filled glovebox for further processing.

PBDB-T was mixed with IEICO-4F in a 1:1 weight ratio and dissolved in chlorobenzene and 1,8-diiodooctane (at a volume ratio of 99:1) to a concentration of 24 mg/ml. The PBDB-T:IEICO-4F solution was stirred overnight at 500 rpm at 50 °C inside a N₂-filled glovebox. The PBDB-T:IEICO-4F films were prepared by spin coating the PBDB-T:IEICO-4F solution onto glass/ITO/ZnO substrates at 4000 rpm, 1000 rpm/s, and 60 s. The films were then solvent-annealed for 5 h.

The PMA solution was prepared as previously described and combined with 1,2-dichlorobenzene at a volume ratio of 99:1, and the solutions were spun on top of glass/ITO/ZnO/PBDB-T:IEICO-4F at 1000 rpm, 1000 rpm/s, and 30 s. For solvent vapor treated samples, glass/ITO/ZnO/PBDB-T:IEICO-4F substrates were placed inside glass Petri dishes with a few drops of 1,2-dichlorobenzene on top of a hot plate at 80 °C for 10 min before spin coating the blended PMA solution (at a volume ratio of 99:1). All samples were solvent-annealed for 1 h and then rinsed with 2-propanol to remove PMA residues from the surface of the films.

The bottom ITO electrode was exposed for contact by removing part of the PBDB-T:IEICO-4F film with chlorobenzene. All samples were then transferred to a high vacuum (1 × 10⁻⁷ Torr) thermal evaporation system (SPECTROS, Kurt J. Lesker). Then, 150 nm of Ag was deposited through a shadow mask as the top electrode. For reference undoped samples, 10 nm of MoO₃ was deposited as the hole-collecting layer before deposition of the top Ag electrode. The effective area of the photoactive layer was 0.1 cm².

The current density–voltage (J–V) characteristics were measured inside a N₂-filled glovebox using a Keithley 2400 source meter controlled using a LabVIEW program. The devices were illuminated by an Oriel solar simulator (model LCS-100) with an air mass 1.5G filter and an intensity of 100 mW cm⁻². The light intensity of the solar simulator was calibrated by a reference silicon solar cell (model 91 150V, Oriel).

Our first set of investigations aimed to extend the scope of the PMA doping technique based on solvent swelling to the realization of OPV devices. Building from our previous study, we first used 2-propanol (IPA) as the main processing solvent for PMA and 1,2-dichlorobenzene (1,2-DCB) as the swelling solvent [Fig. 1(a)] due to its longer drying time compared to chlorobenzene, which may promote infiltration of PMA into the BHJ films.^22,23 Since 1,2-dichlorobenzene is also the processing solvent for many donor polymers and NFAs, including PBDB-T and IEICO-4F, the amount of 1,2-dichlorobenzene added to the PMA solution has to be controlled to produce swelling while not washing away the underlying organic films. Figures S1 and S2 of the supplementary material show the results of optical microscopy and transmittance measurements used to study the film integrity of PMA-doped PBDB-T:IEICO-4F films using different volume ratios of the PMA solution and 1,2-dichlorobenzene. These studies revealed that although we previously showed that PBDB-T films could be treated with a blended solution of PMA–2-propanol and 1,2-dichlorobenzene up to a volume ratio of 9:1,²² BHJ films would still be partially damaged if treated with a similar 9:1 blended solution of PMA–2-propanol and 1,2-dichlorobenzene and up to a 19:1 ratio by volume (Figs. S1 and S2 of the supplementary material). During these preliminary experiments, it was found that in order to preserve the integrity of BHJ films, the ratio of PMA–2-propanol to 1,2-dichlorobenzene needed to be increased to 99:1.

Next, we investigated the impact of reducing the amount of swelling solvent (1,2-dichlorobenzene) in PMA–2-propanol:1,2-dichlorobenzene solutions on the effectiveness of PMA doping. In the past, we established that the bleaching of the main absorption band and the appearance of new absorption bands in the near-infrared (NIR) wavelengths of the absorptance or transmittance spectra can be used as a proxy to qualitatively evaluate if PMA can electrically dope a semiconductor polymer film.^10,11 Consequently, first, the optical properties of PMA-doped PBDB-T films were studied by measuring the UV–Vis–NIR transmittance spectra. Figure S3 of the supplementary material shows that PBDB-T films treated with a blended solution of PMA–2-propanol and 1,2-dichlorobenzene exhibit more pronounced changes in the spectral transmittance than those treated only with PMA–2-propanol. The emergence of new absorption bands at NIR wavelengths correlates with the presence of PMA anions,¹⁰ suggesting effective doping of PMA molecules into PBDB-T films enabled by solvent swelling. Yet, the changes in the transmittance spectra became less noticeable as the amount of 1,2-dichlorobenzene added to PMA–2-propanol was reduced [Fig. 1(b)], which, in principle, corresponds to a reduced degree of electrical doping. To establish this correlation, the electrical properties of PMA-doped PBDB-T films were characterized by four-point probe measurements (Fig. 2). For PBDB-T films treated only with PMA–2-propanol, the sheet resistance measurements yielded a value of 8.5 × 10⁹ Ω/sq, three orders of magnitude lower than that of pristine PBDB-T films (1.1 × 10¹² Ω/sq). The sheet resistance of PMA-doped films would further decrease when 1,2-dichlorobenzene was added to PMA–2-propanol, reaching a value of 2.9 × 10⁸ Ω/sq for PBDB-T films treated with a blended solution of PMA–2-propanol and 1,2-dichlorobenzene at a volume ratio of 9:1. Above this ratio, the sheet resistance slightly increased up to a value of 1.0 × 10⁹ Ω/sq when the amount of 1,2-dichlorobenzene was reduced in the blend to a volume ratio of 99:1.

Furthermore, we characterized the work function of PMA-doped PBDB-T films using Kelvin probe measurements in air (Fig. 3). The work function of pristine PBDB-T films had a value of 4.68 eV. Without solvent swelling, PMA-doped PBDB-T films showed a work function value of 4.79 eV. In comparison, the work function was found to increase up to a value of 4.93 eV when the PBDB-T films were treated with a blended solution of PMA–2-propanol and 1,2-dichlorobenzene at a volume ratio of 9:1 and slightly decreased to 4.88 eV for films treated with a blended solution of PMA–2-propanol and 1,2-dichlorobenzene at a volume ratio of 99:1. The increase in the work function is attributed to a shift in the Fermi level toward the highest occupied molecular orbital (HOMO) bands of PBDB-T, as would be expected from electrical doping of the film.

In summary, these data collectively demonstrate the importance of solvent swelling for the infiltration of PMA molecules into PBDB-T films and to enable electrical doping. These data also show a clear correlation between the decreased amount of swelling solvent in the solvent blends and the decreased electrical doping.

To optimize the electrical doping of these polymer films, we investigated an additional step to promote swelling, using a solvent vapor treatment prior to treating the films with PMA–2-propanol and 1,2-dichlorobenzene.^24,25 To implement this vapor treatment, we first exposed the polymer films to a saturated 1,2-dichlorobenzene atmosphere at 80 °C for 10 min, prior to treating the film with a PMA–2-propanol and 1,2-dichlorobenzene solution at a volume ratio of 99:1. Using this pre-treatment, it was found that the sheet resistance of PMA-doped PBDB-T films could be decreased by nearly an order of magnitude from 1.0 × 10⁹ to 2.5 × 10⁸ Ω/sq, comparable to that of PMA-doped PBDB-T films treated with a blended solution of PMA–2-propanol and 1,2-dichlorobenzene at a volume ratio of 9:1 (2.9 × 10⁸ Ω/sq) (Fig. 2). Note that doping by PMA–2-propanol:1,2-dichlorobenzene at a volume ratio of 9:1 still produced larger changes in the spectral transmittance of PBDB-T films than doping by PMA–2-propanol:1,2-dichlorobenzene at a 99:1 ratio, even with a solvent vapor pre-treatment, while the work function of both of these films was comparable. One possible explanation to this observation is that exposing the films to solvent vapor may lead to a relatively superficial swelling and increased surface doping of the polymer films.

Next, we used these PMA doping techniques to fabricate OPV devices. PMA-doped OPVs were fabricated with a device architecture of glass/ITO/ZnO (30 nm)/PMA-doped PBDB-T:IEICO-4F (100 nm)/Ag (150 nm) [Figs. 4(a) and 4(b)]. For comparison, reference devices were fabricated with a layer of evaporated MoO₃ as a hole-collecting layer. The J–V characteristics of the OPV devices are shown in Fig. 4(c). Reference devices with a MoO₃ hole-collecting layer exhibited a PCE of 8.6%, open-circuit voltage (V_OC) of 703 mV, short-circuit current density (J_SC) of 22.9 mA cm⁻², and fill factor (FF) of 53%. Without any treatments to swell the BHJ films, the PMA-doped devices yielded a low PCE of 1.2%, V_OC of 238 mV, J_SC of 13.4 mA cm⁻², and FF of 35% (Device A). When 1,2-dichlorobenzene was added to PMA–2-propanol (99:1 by volume), the PCE increased to 7.2%, with a V_OC of 653 mV, J_SC of 22.3 mA cm⁻², and FF of 50% (Device B). Finally, devices with solvent vapor pre-treatment showed a PCE of 8.2%, V_OC of 709 mV, J_SC of 21.4 mA cm⁻², and FF of 54% (Device C), comparable to values found in reference devices. The photovoltaic performance of these last PMA-doped devices, comparable to that of reference devices built using an evaporated MoO₃ hole-collecting layer, may be attributed to an increased built-in potential. Such potential arises from an enhanced level of p-type doping due to the use of swelling solvent and solvent vapor. All performance parameters are summarized in Table I.

In summary, we presented a methodology to achieve p-type electrical doping involving a first vapor treatment and a second spin coating of PMA–2-propanol:1,2-dichlorobenzene at a 99:1 ratio. This new technique was found compatible with the fabrication of OPV devices based on blends of a polymer donor and a NFA. Our method differs from the previously reported immersion method^10–12 in that it includes additional treatments, such as binary solvent blends for PMA and solvent vapor pre-treatment, to expand the BHJ film volume for dopant infiltration. Our results confirm that the presence of a swelling solvent in the solvent blend for PMA is critical to facilitate doping of the polymer donor. However, this doping technique by itself leads to damages of the BHJ films. Even though the film integrity of the BHJ films can be preserved by decreasing the amount of swelling solvent in the solvent blends, changes in spectral transmittance and variations in sheet resistance and work function suggest a decreased doping efficiency. Such a limitation is overcome by exposing the BHJ films to a solvent vapor prior to solution-based doping. The combination of treatments simplifies the fabrication and device geometry of a PBDB-T:IEICO-4F based non-fullerene OPV, which exhibits a PCE of 8.2%, almost comparable to that of an OPV with an evaporated MoO₃ hole-collecting layer (8.6%).

See the supplementary material for the optical microscope images and UV–Vis–NIR transmittance spectra of pristine PBDB-T:IEICO-4F films and PBDB-T:IEICO-4F films treated with PMA–2-propanol:1,2-dichlorobenzene at various volume ratios and for the UV–Vis–NIR transmittance spectra of pristine PBDB-T films and PBDB-T films treated with PMA–2-propanol:1,2-dichlorobenzene at various volume ratios.

This work was supported in part by the Pettit Professorship in the School of Electrical and Computer Engineering at the Georgia Institute of Technology and by the Department of Energy/National Nuclear Security Administration (NNSA) Award No. DE-NA0003921 through the Consortium for Enabling Technologies and Innovation (ETI).

The authors have no conflicts to disclose.

Yi-Chien Chang: Conceptualization (equal); Data curation (lead); Formal analysis (lead); Investigation (lead); Methodology (equal); Writing – original draft (lead); Writing – review & editing (equal). Felipe A. Larrain: Conceptualization (equal); Methodology (equal); Writing – review & editing (equal). Canek Fuentes-Hernandez: Conceptualization (equal); Project administration (equal); Supervision (equal); Writing – review & editing (equal). Youngrak Park: Investigation (supporting); Methodology (supporting); Writing – review & editing (equal). Bernard Kippelen: Conceptualization (equal); Funding acquisition (lead); Project administration (equal); Resources (lead); Supervision (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.