Surpassing the Exciton Diffusion Limit in Single-Walled Carbon

Nov 22, 2016 - Semiconducting single-walled carbon nanotube (s-SWNT) light sensitized devices, such as infrared photodetectors and solar cells, have r...
0 downloads 14 Views 3MB Size
Surpassing the Exciton Diffusion Limit in Single-Walled Carbon Nanotube Sensitized Solar Cells Ghada I. Koleilat,§,† Michael Vosgueritchian,§ Ting Lei,§ Yan Zhou,§ Debora W. Lin,§ Franziska Lissel,§ Pei Lin,‡ John W. F. To,§ Tian Xie,§ Kemar England,§ Yue Zhang,‡ and Zhenan Bao*,§ §

Department of Chemical Engineering, Stanford University, Stanford, California 94305, United States School of Materials Science and Engineering, University of Science and Technology Beijing, Beijing 100083, People’s Republic of China



S Supporting Information *

ABSTRACT: Semiconducting single-walled carbon nanotube (s-SWNT) light sensitized devices, such as infrared photodetectors and solar cells, have recently been widely reported. Despite their excellent individual electrical properties, efficient carrier transport from one carbon nanotube to another remains a fundamental challenge. Specifically, photovoltaic devices with active layers made from s-SWNTs have suffered from low efficiencies caused by three main challenges: the overwhelming presence of highbandgap polymers in the films, the weak bandgap offset between the LUMO of the s-SWNTs and the acceptor C60, and the limited exciton diffusion length from one SWNT to another of around 5 nm that limits the carrier extraction efficiency. Herein, we employ a combination of processing and device architecture design strategies to address each of these transport challenges and fabricate photovoltaic devices with sSWNT films well beyond the exciton diffusion limit of 5 nm. While our solution processing method minimizes the presence of undesired polymers in our active films, our interfacial designs led to a significant increase in current generation with the addition of a highly doped C60 layer (n-doped C60), resulting in increased carrier separation efficiency from the s-SWNTs films. We create a dense interconnected nanoporous mesh of s-SWNTs using solution shearing and infiltrate it with the acceptor C60. Thus, our final engineered bulk heterojunction allows carriers from deep within to be extracted by the C60 registering a 10-fold improvement in performance from our preliminary structures. KEYWORDS: single-walled carbon nanotubes, diffusion length, n-doped C60, nanoporous matrix, solution shearing

S

easy access to the desired infrared regions with strong absorption features.8,14−19 The as-synthesized SWNTs are typically mixtures of tubes with various chiralities, a third of which are metallic and the rest semiconducting. In recent years, many have reported various methods15,17,20 of separating the semiconducting types from their metallic counterparts, which would potentially allow a wider range of applications in electronics where only semiconducting SWNTs are required. Whereas the metallic SWNTs are better suited for conductive electrode applications, semiconducting SWNTs (s-SWNTs) have been used recently as sensitizers in optically active devices such as solar cells.15,18,21−26

olution processing is an attractive avenue in photovoltaics, enabling the potential for low cost, low temperature, flexible, stretchable, and most importantly, large area applications.1−6 One class of materials that can be solution processed is single-walled carbon nanotubes (SWNTs). SWNTs consist of a variety of chiralities and diameters corresponding to specific bandgaps and optical transitions, some of which are ideal for photovoltaic applications.7−10 Half of the sun’s radiant energy lies in the infrared; it is not surprising that the optimal bandgap for a single junction solar cell is in the near-infrared (NIR) region at 950 nm.11−13 Multijunction solar cells offer the potential of reaching high efficiencies through the inclusion of infrared-bandgap materials. In fact, for a double and triple junction solar cell, the smallest bandgap junctions optimally lie at 1250 and 1750 nm, respectively.13 By virtue of their strong quantum confinement and thus bandgap tunability, solution-processed SWNTs offer © 2016 American Chemical Society

Received: September 20, 2016 Accepted: November 22, 2016 Published: November 22, 2016 11258

DOI: 10.1021/acsnano.6b06358 ACS Nano 2016, 10, 11258−11265

Article

www.acsnano.org

Article

ACS Nano

Figure 1. Schematics of (a) s-SWNT/C60/n-doped C60 structure and (b) s-SWNT surrounded by the sorting polymer P3DDT, which selectively wraps around s-SWNTs and roams freely in solution. Two major carrier transport problems hindering s-SWNT contribution to the performance are illustrated in panel c and d. (c) With the abundant presence of P3DDT, the photogenerated carriers inside the s-SWNTs are trapped by the high bandgap of P3DDT. (d) We illustrate the small band offset between the LUMO of s-SWNTs and the C60, which is weaker than the exciton binding energy in the nanotubes.

Recently, polyfluorene17 and polythiophene16,20,27 sorted semiconducting SWNTs dispersed in nonpolar solvents have been used for photovoltaic generation.16,18,22,26,27 Generally, the s-SWNT film, comprised of either a single chirality or multiple chiralities, is simply interfaced with the fullerene acceptor C60 in a flat bilayer heterojunction as depicted in Figure 1a.15,16,22 The efficiency of those first-generation architectures has been relatively low. The major hurdle toward improved performance was cited to be the dominance of carrier recombination in the films due to the limited exciton diffusion length from one SWNT to another, which is estimated to be approximately 5 nm.15,16,22,27 Recently, to overcome the bilayer heterojunction restrictions, bulk heterojunctions have been attempted: Ye et al. report a structure where an aerogel-based template of randomly oriented s-SWNTs is filled with the fullerene derivative [6,6]-phenyl-C71-butyric acid methyl ester (PC70BM),26 while Gong et al. report polychiral SWNTs mixed with PC70BM with poly(3-hexylthiophene-2,5-diyl) (P3HT) used as a surfactant.25 The challenge remains in the highly dominant visible contribution of the polymers, including P3HT and PC70BM, overshadowing the less than 20% carbon nanotube’s influence on the photocurrent generation. In recent years, we published reports on regioregular poly(3dodecylthiophene-2,5-diyl) (P3DDT) solution sorting of semiconducting HiPCo16 and CoMoCAT27 SWNTs and their application in a bilayer architecture with the fullerene acceptor C60. Similarly to other reports in literature, the performance of our solar cells was dominated by the

contribution (>85%) of the absorption from the polymer P3DDT present in the film.16,27 Thus, the challenge in SWNT based solar cells is to effectively extract charge carriers from the SWNTs themselves, which absorb light from the visible to the infrared. In our P3DDT sorted s-SWNT system, we attribute the weak SWNT contribution to the total performance to the following three main carrier transport problems: (a) The dominant presence of the high bandgap polymer (P3DDT) in the film causing issues in two forms: the free-roaming polymer species suspended between the sSWNTs in solution and the wrapping polymer species enveloping the s-SWNT tubes. With the high bandgap P3DDT separating the small-bandgap donor (SWNTs) from the acceptor (C60), infrared photogenerated carriers are trapped within the polymer−SWNT matrix, which prevents the carriers from coming in contact with the acceptor as depicted in Figure 1b,c. (b) The small bandgap offset between the LUMO levels of the acceptor (C60) and the donor (SWNTs): The photogenerated exciton (Figure 1d) at the donor− acceptor interface requires a driving force to separate into an electron and a hole; without a proper band offset, recombination occurs at the interface. (c) The vertical limited exciton diffusion length from one carbon nanotube to another: The exciton diffusion length is estimated to be in the 5 nm range, which means a film of SWNTs thicker than 5 nm forming a bilayer 11259

DOI: 10.1021/acsnano.6b06358 ACS Nano 2016, 10, 11258−11265

Article

ACS Nano

concentration of P3DDT and SWNTs in the initial solution using a ratio of 1:1:0.8 to obtain Solution A. Figure 2 depicts

junction with C60 would be dominated by carrier recombination. A SWNT film less than 5 nm is not close to the optimal thickness to absorb enough light for photovoltaic generation. Herein, we exclusively employ CoMoCAT SWNTs because they contain chiralities, predominately the (6,5) chirality, that are ideal for single junction solar cells. We describe the combination of processing and device engineering strategies to overcome the above transport challenges and realize s-SWNT sensitized photovoltaic devices. First, to minimize the polymer content and reduce carrier trapping, we employed a more rigorous selective solution sorting process. This process produced 40 times more concentrated (6,5) s-SWNT content in the solution than methods previously reported27 with a significant reduction of the polymer content. Second, we redesigned the device interfaces to improve the carrier separation and extraction at the donor/acceptor interface. Because the band offset between the LUMO levels of the acceptor (C60) and the donor (SWNTs) is inadequate for efficient exciton separation, especially from the smallbandgap SWNTs, we sought to significantly enhance the internal field bias inside the active junction as a complementary driving force to move carriers away from the donor/acceptor interface. As previously reported, the exciton diffusion limit in a SWNTs bulk film is approximately 5 nm;15,16,22,27 however, that estimation does not account for possible field-driven transport of carriers. We employed a heavily doped region of C60 to increase the effect of field-driven carrier extraction in the devices. We observed an enhancement in overall performance, which we believe is due to the improvement in carrier extraction efficiency and reduction in bulk recombination. Third, we improved the active junction architecture to provide more donor contact areas. The flat bilayer junction that we employed so far was an ineffective structure with minimal interfacial area between the donor and acceptor. Hence, we engineered a bulk heterojunction with interpenetrating carrier pathways within the donor and acceptor regions. To enable this structure, we used the solution shearing method28−30 rather than the customary spin-coating or doctor-blading methods to deposit our s-SWNTs solution onto the substrate. With careful control of the deposition process, we were able to create a dense nanoporous mesh of s-SWNTs that would allow the acceptor (C60) to infiltrate deeply into the film and extract photocarriers from beyond the 5 nm limit within the s-SWNT region. With the above improvements, we observed a 10-fold improvement in overall performance of our final devices. Furthermore, the SWNTs in those devices accounted for approximately 30% of the total short-circuit current of the solar cell, and their presence in the devices resulted in a direct enhancement in the C60’s visible response compared to the cell containing only C60 and P3DDT.

Figure 2. Absorption of the various solutions described in the text. Solution C is the final solution employed in the remaining text and is 40 times more concentrated in SWNT species than the previously reported sorted CoMoCAT solution, with the P3DDT content significantly more reduced.27

the absorption spectrum of the various dispersions employed. Solution A contains over 10 times more SWNTs and P3DDT than the previously reported solution.27 By employing CoMoCAT SWNTs with a higher content of semiconducting SWNTs (see Methods) and adjusting the relative amount of species in the mixture to 1:0.4:0.8, we increased significantly the s-SWNT content in Solution B, specifically the (6,5) chirality, as shown in Figure 2. Our optimal Solution C with significantly reduced P3DDT content is obtained after adjusting the ultrasonication time to 1 h and increasing the centrifuge temperature to 20 °C. Our final suspension, Solution C, has over 40 times higher concentration of (6,5) species than our previously reported solution. With the free-roaming P3DDT presence significantly diminished in Solution C, the first major hurdle to carrier transport is alleviated with the sorted sSWNTs now having a higher probability to directly contact the acceptor C60 in our films. Consequently, excitons generated from the s-SWNTs have a higher probability of reaching the C60 interface and separating into free carriers to generate external current. Next, we looked into redesigning the device interfaces and the active layer junction architecture to improve (1) carrier separation and extraction from the donor/acceptor interface and (2) carrier transport from deep within the bulk of the s-SWNT films. The sorted s-SWNTs are deposited via the spin-coating process described in the Methods section in this report unless otherwise stated. Field-Enhanced Carrier Separation and Extraction. In a solar junction, at the interface where the P- and N-type materials are in contact, two opposing processes affecting the density of the photogenerated bound excitons are constantly competing: exciton recombination and exciton dissociation, which is then followed by carrier transport and current generation. The driving forces behind efficient exciton dissociation can either be due to a deep enough band offset between the LUMOs of the donor (P-type) and acceptor (Ntype) materials larger than the exciton binding energy or due to the presence of a strong internal electric field able to

RESULTS AND INTERPRETATION Increase of s-SWNT Content in Solution Sorting. Recently, we reported the P3DDT sorting process of semiconducting CoMoCAT carbon nanotubes and evaluated its potential in photovoltaic generation.27 In our previous work, we employed a 1:1:5 ratio of P3DDT to CoMoCAT to toluene (w/w/v), respectively. We first ultrasonicated the mixture for 30 min followed by 1 h centrifugation at 16 °C.27 To increase the carbon nanotube active layer thickness, we increased the 11260

DOI: 10.1021/acsnano.6b06358 ACS Nano 2016, 10, 11258−11265

Article

ACS Nano

Figure 3. (a) Energy levels of the materials used in our photovoltaic devices. We illustrate here the diversity of carbon nanotubes present in the films by drawing a range of bandgaps with the (6,5) chirality being dominant. (b) Schematic illustrating the effect of adding n-doped C60 on the depletion width of our device. The depletion width inside the s-SWNTs film increases from W1 to W2 enabling the extraction of more carriers from deep within the films. The main transport mechanism inside the depletion region is drift transport whereby the carriers are swept by an internal electric field. On the other hand, in the quasi-neutral region (QNR), the dominant mechanism is diffusion, and there the diffusion length is limited to around 5 nm. With the addition of n-doped C60, for a similar thickness film, the QNR decreases from QNR1 to QNR2, and recombination is thus limited.

tion due to the widely reported vertical limited exciton diffusion length from one carbon nanotube to another15,16,22 significantly hindered the performance of our first-attempted bilayer structures. With addition of a heavily n-doped C60 layer, the effective area for the internal electric field increases the depletion width inside the active layer and more specifically inside the s-SWNT film. Therefore, the excitons at the interface of the s-SWNTs/ C60/n-doped C60 are now more easily dissociated, and the electron and hole are transported in opposing directions by the electric field as illustrated in Figure 3b. This heavily doped C60 layer was prepared by coevaporating C60 and 2-(2-methoxyphenyl)-1,3-dimethyl-1H-benzoimidazol-3-ium iodide (oMeO-DMBI-I) (see Methods). The dopant was previously reported by our group for n-doping of organic semiconductors.33,34 As shown in Table 1, the interfacial carrier extraction enhancement translates directly to an increase in short circuit current generated by our solar cells. Table 1 details the average performance parameters of our solar cells with and without ndoped C60. With the increase in depletion width and the reach of the drift-based transport of carriers within the active junction, the diffusion-based transport and thus the bulk recombination inside the QNR of the s-SWNTs is minimized, which translates into a superior open-circuit voltage and fill factor. As detailed in Table 1, the efficiency of our structures

disassociate the exciton and sweep carriers in opposing directions. After carrier separation, transport to the external contacts is governed by the electric field based drift transport inside the depletion region followed by carrier diffusion based transport in the bulk of the film, that is, the quasi-neutral region (QNR). It is usually a combination of all the latter phenomena, drift and diffusion, occurring in synchronization that leads to a superior solar junction. In a classical P−N junction, the depletion width, where carrier transport is driven by an internal field, is heavily dependent on the doping level of the p-type and n-type regions. For example, when an N+ region is included forming a P−N:N + structure, the depletion width inside the N and P regions widens. Considering that the N region is made very thin and thus succumbs entirely to the internal field, more of the Pregion becomes subjected to the internal field force, extending the carrier extraction and drift-based transport from within the P-region. We illustrate the latter concept with our s-SWNT/C60 structure in Figure 3: the band diagrams of our junction at equilibrium, with and without an added heavily n-doped layer of C60, are shown. First, the LUMO band offset between the s-SWNTs and the C60 (Figure 3a) is far lower than the widely reported exciton binding energy of 0.3−0.4 eV,31,32 and thus our initial junctions are dominated by interfacial recombination. In fact, the recombination at the interface in addition to bulk recombina11261

DOI: 10.1021/acsnano.6b06358 ACS Nano 2016, 10, 11258−11265

Article

ACS Nano Table 1. Average Performance Parameters of Solar Cellsa processing spin-coated (5 nm) spin-coated (5 nm) sheared

Jsc (mA)

Voc (V)

FF (%)

ηb (%)

without n-doped C60 with n-doped C60

0.91

0.26

0.21

0.05

1.56

0.29

0.40

0.18

with n-doped C60

2.38

0.45

0.46

description

increases, the dependence on diffusion-based transport in the quasi-neutral region of the s-SWNT film dominates. Thicker films of s-SWNTs are thus severely limited by the short exciton diffusion limit: the excitons that are generated far away from the C60 interface and are not subject to the internal electric field in the depletion region are not being efficiently extracted before succumbing to recombination. We therefore directed our efforts into engineering a bulk heterojunction structure to allow the extraction of excitons from deep within a thicker film of s-SWNTs. The blending and solution processing widely employed to form organic bulk heterojunctions cannot be easily applied here because mixing sSWNTs with PCBM leads to aggregation.25 Essentially, we want to create a continuous interconnected s-SWNT nanoporous matrix with nanometer-scale thickness that can be infiltrated by the C60 acceptor without disturbing the SWNT film morphology. To that end, we employed a solution shearing deposition method to allow us to have control over the SWNT film properties. Solution shearing has previously been used to control the morphology of organic bulk heterojunctions and various organic films.28−30 Figures 5b,c and SI-1 depict atomic force microscopy (AFM) images of s-SWNT films deposited using spin-coating versus the solution shearing method. The spin-coated films appear to be relatively flat with an average roughness of 2 nm by AFM. The scanning electron microscopy (SEM) image (Figure SI-2) and AFM showed no visible pores. Solar cells made with spin-coated s-SWNT films covered with C60 are thus similar to donor/acceptor bilayer junctions. In contrast, a nanoporous s-SWNT mesh is apparent with the films fabricated via solution shearing (average roughness of 8.5 nm; Figure 5b and Figures SI-1 and SI-2). With easy control of the shearing deposition conditions, thicknesses of over 25 nm are achieved. In addition to the nanometer pores, nanoscale trenches along the shearing

0.49

The figures of merit were averaged for 5 devices in each case. Solar cell power conversion efficiency. a

b

with an n-doped C60 layer was at least 3.5 times higher than our initial bilayer junction. Surpassing the Diffusion Limit in a Nanoporous SWNT Matrix. Thus far, we have enhanced the interfacial interaction between the SWNTs and the acceptor C60 first by minimizing the high-bandgap free-roaming polymer that acts as a carrier-trapping agent that engulfs the s-SWNTs and by fortifying the reach of the internal electric field that sweeps the carrier in opposite directions inside the depletion region. Next, we proceed to increase the s-SWNT film thickness, which enhances the amount of photons absorbed and current generated. Note that with our default spin-coating process, the maximum thickness we can achieve is close to 9 nm due to the bottleneck of spin-coating layers of SWNTs on top of another, which effectively leads to the slow dissolution of the bottom layers. Figure 4a,b summarizes the short-circuit current density (Jsc) and the open-circuit voltage (Voc) thickness trends of various structures, all with a heavily n-doped C60 layer. The Jsc increases with thickness of spin-coated films due to the increase in film absorption and enhanced extraction from the n-doped C60 presence. On the other hand, the Voc steadily decreases (blue line). The open-circuit voltage of a solar cell is dependent on the saturation current in the solar junction and is thus heavily reliant on recombination losses in the film. As thickness

Figure 4. Plot of the (a) short-circuit current density vs thickness of spin-coated and sheared s-SWNT devices and (b) open-circuit voltage vs thickness of spin-coated and sheared s-SWNTs devices. (c) Current−voltage plot of a s-SWNT sheared device (red curve) and a s-SWNT spincoated device (blue curve), both topped with the doped n-doped C60 layer. The blue dotted line represents the performance of a s-SWNT spin-coated device without the doped n-doped C60 layer. (d) EQE spectra of a s-SWNT sheared device (red) showing a significant contribution from the s-SWNTs, with the black curve depicting the spectra of a device made exactly the same but without the s-SWNTs in it. As observed, the visible contribution of the C60 at 450 nm is enhanced from the presence of the s-SWNTs in the film. Note that the figures of merit were averaged for 5 devices in each case. 11262

DOI: 10.1021/acsnano.6b06358 ACS Nano 2016, 10, 11258−11265

Article

ACS Nano

Figure 5. (a) Absorption spectra of 9 nm spin-coated and 25 nm sheared s-SWNT films. AFM images of (b) a sheared s-SWNT film depicting the nanoporous nature of the structure and (c) a flat spin-coated film.

the QNR. Thicker and higher performing devices might be possible when we are able to overcome the bottleneck of the nanoporous s-SWNTs deposition process: controlling the substrate temperature while shearing has proven to affect thickness in polymers.29 We also believe that higher performance will be achieved when the P3DDT content is entirely eliminated and thick s-SWNTs films are interpenetrated by appropriate materials; a donor material forming a type II interface with the s-SWNTs along with solution processed fullerene PC60BM topped with our current n-doped C60 interface would potentially form a high efficiency photovoltaic junction.

direction are seen in the 3D AFM image shown in Figure SI-1. Figure 5a presents the absorption spectrum of a 9 nm thick spin-coated s-SWNT film and a 25 nm thick sheared nanoporous s-SWNT film. Both thicknesses represent the approximate maximum thickness achieved with both deposition processes (see Methods). The absorption of the sheared porous film is 1.6 times stronger than the spin-coated one, which directly translates to an improved current generated by a similar factor (Table 1 and Figure 4c). However, the overall performance of the sheared film is at least 2.7 times higher than the spin-coated devices. As a matter of fact, the opencircuit voltage drastically improves as well as the fill factor: this is a direct indication of the reduction in the bulk recombination in the quasi-neutral region. With increased sheared film thickness, the short-circuit current of the devices steadily increases while the open-circuit voltage remains unchanged (Figure 4a,b, red lines). With this improved structure, the highest performance was observed for the thickest film (∼25 nm) achievable, which is 5 times thicker than the widely reported ∼5 nm exciton diffusion limit.16,22,26 Unfortunately, with higher thicknesses, the films become very rough and cause shunting in the devices. Our best device architecture and performance is depicted in Figure 4c,d. With our processing and design strategies, the external quantum efficiency of our device reached 11% in the NIR, which is at least 3 times higher than our previously published performance.16,27 We also show the EQE of a pure P3DDT/C60/n-doped C60 film, illustrating that the visible contribution of C60 increases drastically with the presence of the s-SWNTs. As previously reported by Ye et al.,26 the sSWNTs act as hole transport facilitators. With the use of the ndoped C60 to enhance the internal electric field, more excitons out of the s-SWNTs and the C60 are extracted. The s-SWNT contribution computed from 650 nm to the NIR reaches 30% of the total performance, doubling our last reported s-SWNT contribution27 and greater than the recently reported contributions of SWNTs in bulk heterojunctions.25,26

METHODS Preparation of SWNT Solutions. Solution A was made by mixing 25 mg of CoMoCAT SWNTs (SG65) from Sigma-Aldrich with 25 mg of P3DDT (from Sigma-Aldrich) in 20 mL of toluene and sonicating (Sonics & Materials, Inc.,VCX 750) for 30 min at an amplitude level of 70%. The solution was then centrifuged (Thermo Scientific, Sorvall LYNX 4000, F21-8x50y Rotor) at 17000 rpm (20000g) for 1 h at 16 °C. Solution B was made by mixing 10 mg of CoMoCAT SWNTs (SG65i) from Sigma-Aldrich with 25 mg of P3DDT in 20 mL of toluene; sonication and centrifugation was identical to solution A. Solution C was made similarly to solution B but sonicated for 1 h at an amplitude level of 70%. Solution C was then centrifuged at 17000 rpm for 1 h at 20 °C. The supernatant was collected before making devices. SWNT Structure Characterization. A Veeco Multimode AFM instrument was used to take the AFM topography images using the tapping mode regime. The absorption measurements were done in 1 mm path-length quartz cells using a Cary 6000i spectrophotometer (Varian) with toluene as a background solvent. SEM images in the Supporting Information were taken using an FEI Magellan 400 XHR microscope with a 0.5 kV accelerating voltage and 5 pA current. Solar Cell Fabrication. Clean, patterned ITO (∼15 Ω/sq) on glass was used as the substrate. The substrates were treated for 20 min with UV-O3. A poly(3,4-ethylenedioxythiophene)/polystyrene sulfonate (PEDOT/PSS) solution (CLEVIOS AL4083, Heraeus) was then spin-coated on the substrates at 4500 rpm for 1 min followed by annealing at 120 °C in air for 30 min. The substrates were then transferred to an N2 environment. A poly[(9,9-dioctylfluorenyl-2,7diyl)-co-(4,4′-(N-(p-butylphenyl))diphenylamine)] (TFB) solution was prepared (0.5% weight mixed with toluene) and then spin-coated at 6000 rpm for 30 s inside the glovebox. The substrates were then annealed at 80 °C for 1 h after which a toluene wash was done to remove the excess TFB. Spin-Coated s-SWNT Based Devices. s-SWNT C solutions (150 μL) were deposited on the ITO substrates for 1 min and then spincoated at 700 rpm for 30 s followed by 4000 rpm for 10 s. This process was repeated 3−9 times with a 1 min annealing step at 70 °C in the glovebox.

CONCLUSION With the combination of processing and improved device structure designs, we were able to improve the carrier transport in s-SWNT sensitized solar junctions. We were able to fabricate consistently 5 times thicker s-SWNT films with 10-fold improved performance on our initial flat bilayer junction. We were able to greatly improve performance with our design while minimizing (1) exciton recombination due to the presence of high-bandgap polymers and (2) exciton bulk recombination in 11263

DOI: 10.1021/acsnano.6b06358 ACS Nano 2016, 10, 11258−11265

Article

ACS Nano Sheared s-SWNT Based Devices. The Si blade used for printing was tilted at 8° and separated by 300 μm from the substrate. Si blade was treated with OTS/trichloroethylene (anhydrous) to form a hydrophobic surface. A total of 50−90 μL of solution was used for each film fabrication. Shearing speeds varied from 30 to 7.5 μm/s depending on the thickness needed. Shearing speeds below 7.5 μm/s resulted in inhomogeneous films. Films were fabricated in ambient conditions at room temperature. After the film deposition, the substrates were stored in the glovebox overnight. After SWNT deposition, 70 nm of commercially available C60 (fullerene powder, sublimed, 99.9+%, Alfa Aesar) was deposited thermally under vacuum (5 × 10−5 Torr) at 0.02 nm/s (Angstrom Engineering evaporator). For devices with n-doped C60, we first deposited 25−35 nm of commercially available C60 followed by a 35− 45 nm of n-doped C60. 2-(2-Methoxyphenyl)-1,3-dimethyl-1Hbenzoimidazol-3-ium iodide (o-MeO-DMBI-I) was used to dope the C60 film as previously reported.33 Briefly, we coevaporated o-MeODMBI-I with C60 at a doping concentration of 10.0 wt %. Finally, ∼1 nm of LiF followed by 70 nm of Ag was vacuum evaporated on top of the device at 0.02 nm/s (Thermionics Laboratory, Inc. evaporator). Solar Cell Characterization. Solar spectra were obtained with a Newport solar simulator with a flux of 100 mW/cm2 that approximated the solar spectrum under AM1.5G simulated conditions in an N2 environment. The active area was measured from the common area of the patterned ITO with the top electrode. The area of the devices was 0.04 cm2, and the performance was verified by aperturing the illumination area (slightly smaller area than the active area). In addition, the current generated was verified by measuring the EQE spectra (and thus integrated current generated) in an independent second group system with an apertured/limited illumination source. In particular, we employed PEDOT/PSS 4083 and patterned ITO substrates to minimize measurement artifact effects that undoubtedly arise (up to 40% overestimation) from using unpatterned ITO on its own without aperturing, especially with small area devices. The incident photon to current efficiency (IPCE) and double-pass absorption were measured under monochromic illumination, and the calibration of the incident light intensity was performed with a calibrated silicon photodiode.

(NSERC) in the form of the Banting Postdoctoral Fellowship. F.L. acknowledges the Swiss National Science Foundation for an Early Mobility Postdoc grant. P.L. acknowledges the support of Program of Introducing Talents of Discipline to Universities (B14003) to do short-term research in Stanford University.

REFERENCES (1) Chen, C.-C.; Dou, L.; Zhu, R.; Chung, C.-H.; Song, T.-B.; Zheng, Y. B.; Hawks, S.; Li, G.; Weiss, P. S.; Yang, Y. Visibly Transparent Polymer Solar Cells Produced by Solution Processing. ACS Nano 2012, 6, 7185−7190. (2) Graetzel, M.; Janssen, R. A. J.; Mitzi, D. B.; Sargent, E. H. Materials Interface Engineering for Solution-Processed Photovoltaics. Nature 2012, 488, 304−312. (3) Gomez De Arco, L.; Zhang, Y.; Schlenker, C. W.; Ryu, K.; Thompson, M. E.; Zhou, C. Continuous, Highly Flexible, and Transparent Graphene Films by Chemical Vapor Deposition for Organic Photovoltaics. ACS Nano 2010, 4, 2865−2873. (4) Kapur, V. K.; Fisher, M.; Roe, R. Fabrication of Light Weight Flexible CIGS Solar Cells for Space Power Applications. MRS Online Proc. Libr. 2001, 668, H2.6. (5) Lipomi, D. J.; Tee, B. C.-K.; Vosgueritchian, M.; Bao, Z. Stretchable Organic Solar Cells. Adv. Mater. 2011, 23, 1771−1775. (6) Chortos, A.; Koleilat, G. I.; Pfattner, R.; Kong, D.; Lin, P.; Nur, R.; Lei, T.; Wang, H.; Liu, N.; Lai, Y.-C.; Kim, M.-G.; Chung, J. W.; Lee, S.; Bao, Z. Mechanically Durable and Highly Stretchable Transistors Employing Carbon Nanotube Semiconductor and Electrodes. Adv. Mater. 2016, 28, 4441−4448. (7) Iijima, S. Helical Microtubules of Graphitic Carbon. Nature 1991, 354, 56−58. (8) Baughman, R. H.; Zakhidov, A. A.; de Heer, W. A. Carbon Nanotubes–the Route Toward Applications. Science 2002, 297, 787− 792. (9) Dinadayalane, T. C.; Leszczynski, J. Remarkable Diversity of Carbon−carbon Bonds: Structures and Properties of Fullerenes, Carbon Nanotubes, and Graphene. Struct. Chem. 2010, 21, 1155− 1169. (10) Kataura, H.; Kumazawa, Y.; Maniwa, Y.; Umezu, I.; Suzuki, S.; Ohtsuka, Y.; Achiba, Y. Optical Properties of Single-Wall Carbon Nanotubes. Synth. Met. 1999, 103, 2555−2558. (11) Brown, A. S.; Green, M. A. Detailed Balance Limit for the Series Constrained Two Terminal Tandem Solar Cell. Phys. E 2002, 14, 96− 100. (12) Henry, C. H. Limiting Efficiencies of Ideal Single and Multiple Energy Gap Terrestrial Solar Cells. J. Appl. Phys. 1980, 51, 4494−4500. (13) Shockley, W.; Queisser, H. J. Detailed Balance Limit of Efficiency of P-n Junction Solar Cells. J. Appl. Phys. 1961, 32, 510− 519. (14) Itkis, M. E.; Borondics, F.; Yu, A.; Haddon, R. C. Bolometric Infrared Photoresponse of Suspended Single-Walled Carbon Nanotube Films. Science 2006, 312, 413−416. (15) Jain, R. M.; Howden, R.; Tvrdy, K.; Shimizu, S.; Hilmer, A. J.; McNicholas, T. P.; Gleason, K. K.; Strano, M. S. Polymer-Free NearInfrared Photovoltaics with Single Chirality (6,5) Semiconducting Carbon Nanotube Active Layers. Adv. Mater. 2012, 24, 4436−4439. (16) Ramuz, M. P.; Vosgueritchian, M.; Wei, P.; Wang, C.; Gao, Y.; Wu, Y.; Chen, Y.; Bao, Z. Evaluation of Solution-Processable CarbonBased Electrodes for All-Carbon Solar Cells. ACS Nano 2012, 6, 10384−10395. (17) Arnold, M. S.; Green, A. A.; Hulvat, J. F.; Stupp, S. I.; Hersam, M. C. Sorting Carbon Nanotubes by Electronic Structure Using Density Differentiation. Nat. Nanotechnol. 2006, 1, 60−65. (18) Bindl, D. J.; Wu, M.-Y.; Prehn, F. C.; Arnold, M. S. Efficiently Harvesting Excitons from Electronic Type-Controlled Semiconducting Carbon Nanotube Films. Nano Lett. 2011, 11, 455−460. (19) Wang, H.; Mei, J.; Liu, P.; Schmidt, K.; Jiménez-Osés, G.; Osuna, S.; Fang, L.; Tassone, C. J.; Zoombelt, A. P.; Sokolov, A. N.; Houk, K. N.; Toney, M. F.; Bao, Z. Scalable and Selective Dispersion

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.6b06358. AFM images of s-SWNT sheared films and SEM images of spin-coated and sheared s-SWNT films (PDF)

AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. ORCID

Ghada I. Koleilat: 0000-0002-0338-8472 Pei Lin: 0000-0002-3300-5241 Present Address †

G.I.K.: Department of Electrical Engineering, Dalhousie University, Halifax, NS, Canada.

Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This work was funded by Global Climate and Energy Project (GCEP) at Stanford University. G.I.K. acknowledges the support of the Government of Canada and the Natural Sciences and Engineering Research Council of Canada 11264

DOI: 10.1021/acsnano.6b06358 ACS Nano 2016, 10, 11258−11265

Article

ACS Nano of Semiconducting Arc-Discharged Carbon Nanotubes by Dithiafulvalene/Thiophene Copolymers for Thin Film Transistors. ACS Nano 2013, 7, 2659−2668. (20) Wang, H.; Bao, Z. Conjugated Polymer Sorting of Semiconducting Carbon Nanotubes and Their Electronic Applications. Nano Today 2015, 10, 737−758. (21) Arnold, M. S.; Zimmerman, J. D.; Renshaw, C. K.; Xu, X.; Lunt, R. R.; Austin, C. M.; Forrest, S. R. Broad Spectral Response Using Carbon Nanotube/Organic Semiconductor/C60 Photodetectors. Nano Lett. 2009, 9, 3354−3358. (22) Shea, M. J.; Arnold, M. S. 1% Solar Cells Derived from Ultrathin Carbon Nanotube Photoabsorbing Films. Appl. Phys. Lett. 2013, 102, 243101. (23) Xie, Y.; Gong, M.; Shastry, T. A.; Lohrman, J.; Hersam, M. C.; Ren, S. Broad-Spectral-Response Nanocarbon Bulk-Heterojunction Excitonic Photodetectors. Adv. Mater. 2013, 25, 3433−3437. (24) Bernardi, M.; Lohrman, J.; Kumar, P. V.; Kirkeminde, A.; Ferralis, N.; Grossman, J. C.; Ren, S. Nanocarbon-Based Photovoltaics. ACS Nano 2012, 6, 8896−8903. (25) Gong, M.; Shastry, T. A.; Xie, Y.; Bernardi, M.; Jasion, D.; Luck, K. A.; Marks, T. J.; Grossman, J. C.; Ren, S.; Hersam, M. C. Polychiral Semiconducting Carbon Nanotube−Fullerene Solar Cells. Nano Lett. 2014, 14, 5308−5314. (26) Ye, Y.; Bindl, D. J.; Jacobberger, R. M.; Wu, M.-Y.; Roy, S. S.; Arnold, M. S. Semiconducting Carbon Nanotube Aerogel Bulk Heterojunction Solar Cells. Small 2014, 10, 3299−3306. (27) Wang, H.; Koleilat, G. I.; Liu, P.; Jiménez-Osés, G.; Lai, Y.-C.; Vosgueritchian, M.; Fang, Y.; Park, S.; Houk, K. N.; Bao, Z. High-Yield Sorting of Small-Diameter Carbon Nanotubes for Solar Cells and Transistors. ACS Nano 2014, 8, 2609. (28) Diao, Y.; Tee, B. C.-K.; Giri, G.; Xu, J.; Kim, D. H.; Becerril, H. A.; Stoltenberg, R. M.; Lee, T. H.; Xue, G.; Mannsfeld, S. C. B.; Bao, Z. Solution Coating of Large-Area Organic Semiconductor Thin Films with Aligned Single-Crystalline Domains. Nat. Mater. 2013, 12, 665− 671. (29) Diao, Y.; Zhou, Y.; Kurosawa, T.; Shaw, L.; Wang, C.; Park, S.; Guo, Y.; Reinspach, J. A.; Gu, K.; Gu, X.; Tee, B. C. K.; Pang, C.; Yan, H.; Zhao, D.; Toney, M. F.; Mannsfeld, S. C. B.; Bao, Z. FlowEnhanced Solution Printing of All-Polymer Solar Cells. Nat. Commun. 2015, 6, 7955. (30) Reinspach, J. A.; Diao, Y.; Giri, G.; Sachse, T.; England, K.; Zhou, Y.; Tassone, C.; Worfolk, B. J.; Presselt, M.; Toney, M. F.; Mannsfeld, S.; Bao, Z. Tuning the Morphology of Solution-Sheared P3HT:PCBM Films. ACS Appl. Mater. Interfaces 2016, 8, 1742−1751. (31) Ma, Y.-Z.; Valkunas, L.; Bachilo, S. M.; Fleming, G. R. Exciton Binding Energy in Semiconducting Single-Walled Carbon Nanotubes. J. Phys. Chem. B 2005, 109, 15671−15674. (32) Maultzsch, J.; Pomraenke, R.; Reich, S.; Chang, E.; Prezzi, D.; Ruini, A.; Molinari, E.; Strano, M. S.; Thomsen, C.; Lienau, C. Exciton Binding Energies in Carbon Nanotubes from Two-Photon Photoluminescence. Phys. Rev. B: Condens. Matter Mater. Phys. 2005, 72, 241402. (33) Wei, P.; Menke, T.; Naab, B. D.; Leo, K.; Riede, M.; Bao, Z. 2(2-Methoxyphenyl)-1,3-Dimethyl-1H-Benzoimidazol-3-Ium Iodide as a New Air-Stable N-Type Dopant for Vacuum-Processed Organic Semiconductor Thin Films. J. Am. Chem. Soc. 2012, 134, 3999−4002. (34) Wei, P.; Liu, N.; Lee, H. R.; Adijanto, E.; Ci, L.; Naab, B. D.; Zhong, J. Q.; Park, J.; Chen, W.; Cui, Y.; Bao, Z. Tuning the Dirac Point in CVD-Grown Graphene through Solution Processed N-Type Doping with 2-(2-Methoxyphenyl)-1,3-Dimethyl-2,3-Dihydro-1HBenzoimidazole. Nano Lett. 2013, 13, 1890.

11265

DOI: 10.1021/acsnano.6b06358 ACS Nano 2016, 10, 11258−11265