Polymer solar cells with enhanced fill factors - Semantic Scholar

Comment

Report 6 Downloads 60 Views

ARTICLES PUBLISHED ONLINE: 11 AUGUST 2013 | DOI: 10.1038/NPHOTON.2013.207

Polymer solar cells with enhanced ﬁll factors Xugang Guo1,2†, Nanjia Zhou1,3†, Sylvia J. Lou1, Jeremy Smith1, Daniel B. Tice1, Jonathan W. Hennek1, Rocı´o Ponce Ortiz1,4, Juan T. Lo´pez Navarrete4, Shuyou Li3, Joseph Strzalka5, Lin X. Chen1,6 *, Robert P. H. Chang3 *, Antonio Facchetti1,7 * and Tobin J. Marks1 * Recent advances in polymer solar cell (PSC) performance have resulted from compressing the bandgap to enhance the short-circuit current while lowering the highest occupied molecular orbital to increase the open-circuit voltage. Nevertheless, PSC power conversion efﬁciencies are still constrained by low ﬁll factors, typically below 70%. Here, we report PSCs with exceptionally high ﬁll factors by combining complementary materials design, synthesis, processing and device engineering strategies. The donor polymers, PTPD3T and PBTI3T, when incorporated into inverted bulkheterojunction PSCs with a PC71BM acceptor, result in PSCs with ﬁll factors of 76–80%. The enhanced performance is attributed to highly ordered, closely packed and properly oriented active-layer microstructures with optimal horizontal phase separation and vertical phase gradation. The result is efﬁcient charge extraction and suppressed bulk and interfacial bimolecular recombination. The high ﬁll factors yield power conversion efﬁciencies of up to 8.7% from polymers with suboptimal bandgaps, suggesting that efﬁciencies above 10% should be realizable by bandgap modiﬁcation.

B

ulk-heterojunction (BHJ) polymer solar cells (PSCs) based on interpenetrating polymer donor and fullerene acceptor networks offer a compelling approach to efﬁcient, cost-effective solar energy conversion1. Although PSC performance has improved steadily, with the achievement of 6–9% single-junction power conversion efﬁciencies (PCEs)2–4, these metrics lag behind those of inorganic cells5. PCE (he) is proportional to the product of the short-circuit current (Jsc), the open-circuit voltage (Voc) and the ﬁll factor (FF):

he =

Voc × Jsc × FF Pin

(1)

where Pin is the incident solar power. To maximize PCE, researchers have focused on creating narrow-bandgap polymers to maximize Jsc , while lowering the polymer highest occupied molecular orbital (HOMO) to increase Voc (refs 2,6). Note, however, that HOMO lowering will also increase the bandgap, underscoring the problematic tradeoff between Voc and Jsc. As equation (1) shows, raising the ﬁll factor offers an alternative strategy to enhancing PCE, but also presents a challenge because only a few PSCs have achieved ﬁll factors as high as 70–72% (refs 7,8). Indeed, realizing high ﬁll factors has proven elusive, although there is evidence that carrier mobility, active-layer microstructure and also interfacial and bulk charge recombination play a role7,9. After correcting for parasitic resistances, ﬁll factor is primarily depressed by carrier recombination, which reduces carrier lifetime and therefore the current extractable from the device. The commonly accepted mechanisms for non-geminate recombination are bimolecular, mainly Langevin-type10–13, and Shockley–Read–Halltype trap-assisted recombination, involving energy states or tail states within the bandgap14. Recently, it has been reported that

bias and charge-density dependent bimolecular recombination dominates recombination losses in well-optimized PSCs15–18. Factors limiting ﬁll factor are widely discussed as a ﬁeld-dependent competition between non-geminate recombination and charge extraction, with trap states and departure from diode ideality playing important roles14,19. At high internal electric ﬁelds or near short-circuit conditions, enhanced bulk charge mobility can improve transport and reduce space charge build-up9. However, at low internal ﬁelds or near Voc , long carrier lifetimes determine charge extraction and thus the steepness of the J–V curve20,21. Accordingly, both carrier mobility and lifetime are target parameters to enhance, so as to increase PSC ﬁll factors22. For inorganic p–n junction cells, the ﬁll factor can be expressed as

y˜ oc − ln(˜yoc + 0.72 V) y˜ oc + 1 where y˜ oc is the normalized Voc ,

y˜ oc =

Voc nid kT/q

k is the Boltzmann constant, T is temperature, q is the magnitude of the electrical charge on the electron, and nid is an ideality factor relating to an ideal (nid ¼ 1) or non-ideal (nid . 1) diode23. PSCs typically have ideality factors in the range 1.5–2 due to their inherent disorder24, and departures from unity are attributed to various recombination mechanisms inside band-to-band transitions, that is, trapassisted and tail state recombination23. For high-quality monocrystalline inorganic solar cells where nid ≈ 1, ﬁll factors greater than 80% are routine25.

1

Department of Chemistry and the Materials Research Center, the Argonne-Northwestern Solar Energy Research Center, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, USA, 2 Department of Materials Science and Engineering, South University of Science and Technology of China, No. 1088, Tangchang Boulevard, Shenzhen, Guangdong, 518055, China, 3 Department of Materials Science and Engineering and the Materials Research Center, the Argonne-Northwestern Solar Energy Research Center, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, USA, 4 Department of Physical Chemistry, University of Ma´laga, Campus de Teatinos s/n, Ma´laga 29071, Spain, 5 X-ray Science Division, Argonne National Laboratory, Argonne, Illinois 60439, USA, 6 Chemical Sciences and Engineering Division, Argonne National Laboratory, 9700 South Cass Avenue, Argonne, Illinois 60439, USA, 7 Polyera Corporation, 8045 Lamon Avenue, Skokie, Illinois 60077, USA; † These authors contributed equally to this work. * e-mail: [email protected]; [email protected]; [email protected]; [email protected] NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics

© 2013 Macmillan Publishers Limited. All rights reserved.

1

ARTICLES

NATURE PHOTONICS

dione (TPD) and 2,2′ -bithiophene-3,3′ -dicarboximide (BTI) (Fig. 1a) are effective building blocks for high-mobility polymers. By introducing terthiophene (3T; Fig. 1a) with strategically placed alkylation to promote ordering via side-chain interdigitation, we hypothesized that combining BTI or TPD as the in-chain acceptor with 3T as the in-chain donor could imbue PTPD3T and PBTI3T (Fig. 1b) with signiﬁcant order and mobility. PBTI3T and PTPD3T were synthesized by Stille coupling, and exhibit good solubility/processibility and respectable number-average molecular weights (Mn) of 31 and 40 kDa, respectively. From solution to ﬁlm, PTPD3T and PBTI3T show minimal shifts in optical absorption proﬁles, suggesting extensive aggregation, and the sharp vibronic features imply ordered solid-state structures26. Note that PBTI3T has a somewhat greater backbone torsion (Supplementary Fig. S5) due to the absence of planarizing (thienyl)S...O(carbonyl) interactions (Fig. 1b)27 and a smaller PBTI3T optical bandgap (Eopt g , 1.81 eV) compared with that of PTPD3T (1.82 eV). This reﬂects the lower-lying PBTI3T lowest unoccupied molecular orbital (LUMO), in agreement with density functional theory (DFT) computation (Supplementary Fig. S6), and is due to a greater contribution of BTI versus TPD units in the PBTI3T and PTPD3T LUMO orbital topologies, respectively (Supplementary Fig. S7). The cyclic voltammetry-derived HOMOs are 25.55 eV and 25.58 eV for PTPD3T and PBTI3T, respectively, with the slightly lower PBTI3T HOMO attributable to backbone torsion28. Such low-lying HOMOs should favour higher Voc values in

Here, we report two new semicrystalline in-chain donor–acceptor (D–A) polymers, poly[5-(2-hexyldodecyl)-1,3-thieno[3,4c]pyrrole-4,6-dione-alt-5,5-(2,5-bis(3-dodecylthiophen-2-yl)-thiophene)] (PTPD3T) and poly[N-(2-hexyldodecyl)-2,2′ -bithiophene-3,3′ -dicarboximide-alt-5,5-(2,5-bis(3-decylthiophen-2-yl)thiophene)] (PBTI3T; Fig. 1), with close-packed structures. Inverted BHJ solar cells using these polymers as donors and [6,6]-phenylC71-butyric acid methyl ester (PC71BM) as the acceptor achieve ﬁll factors of 76–80%. Although the exact recombination mechanism(s) for PSCs are still not completely resolved, the fact that PTPD3T and PBTI3T exhibit ﬁll factors close to the theoretical limit for nid ¼ 1.5 (Supplementary Fig. S1) suggests minimal recombination and highly efﬁcient charge extraction. Here, we study the relationship between ﬁll factor and macromolecular structure/orientational order, ﬁlm morphology and charge transport. Although the origin of ﬁll factors clearly involves the interplay of a range of factors, we identify the simultaneous importance of three key ﬁll factor-enhancing ingredients: (1) ordered, p-face-on-electrode oriented microstructures with close p–p interplanar spacings for high carrier mobility; (2) ordered bicontinuous networks for reduced charge trapping; and (3) vertical donor/acceptor phase gradation for suppressed carrier recombination.

Design, synthesis and characterization To optimize the ﬁll factor, we sought highly ordered polymers with close p–p stacking to maximize mobility. Thieno[3,4-c]pyrrole-4,6a

R N

O

O

DOI: 10.1038/NPHOTON.2013.207

R N

O

O

S

S S

S

S

b

R

S

TPD

R

BTI

n-C12H25

n-C10H21

n-C12H25

S

S S

3T

S

S

S

n-C10H21 S

S

OCH3

S O

N

O

N

O n-C8H17

n-C6H13

O

n-C6H13

PTPD3T

c

d PTPD3T PBTI3T

Heat flow (W g−1, endo up)

Absorbance normalized

1.0 0.8 0.6 0.4 0.2

1.0

e

4 × 104

3 × 104

PTPD3T PBTI3T

0.5 0.0 −0.5

2 × 104

PTPD3T

1 × 104 ×5

−1.0

PBTI3T

0

0.0 300

PC71BM

PBTI3T

Intensity (c.p.s.)

O n-C8H13

400

500

600

700

Wavelength (nm)

800

900

50

100

150 200 250 300 350 Temperature (°C)

5

10

15

20

25

2θ (deg)

Figure 1 | Macromolecular building blocks, structures, optical absorption spectra, DSC thermograms and X-ray scattering patterns of the semiconducting polymers. a, Chemical structures of polymer building blocks TPD, BTI and 3T. b, Chemical structures of polymer donors PTPD3T and PBTI3T and fullerene acceptor PC71BM. The (thienyl)S...O(carbonyl) interaction, indicated by the dashed line, promotes PTPD3T backbone coplanarity and crystallinity, while greater backbone ﬂexibility lowers the PBTI3T HOMO and enlarges Voc. c, Optical absorption spectra of PTPD3T and PBTI3T pristine ﬁlms spin-coated from chloroform solution (5 mg ml21). d, DSC thermograms of polymers PTPD3T and PBTI3T at a temperature ramp rate of 10 8C min21 under N2. The upper lines are from the heating run and the lower lines from the cooling run. e, u–2u XRD scattering patterns from PTPD3T and PBTI3T ﬁlms spin-coated on octadecyltrichlorosilane (OTS)-modiﬁed Si/SiO2 substrates after annealing at 240 8C for 0.5 h. 2

NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics

© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE PHOTONICS

ARTICLES

DOI: 10.1038/NPHOTON.2013.207

polymer:PCBM PSCs, while .0.4 eV polymer–fullerene LUMO– LUMO offsets should provide sufﬁcient driving force for exciton dissociation6 and lower relaxed charge transfer state recombination, which will also give more favourable ﬁll factors29,30. Polymer crystallinity was ﬁrst investigated by differential scanning calorimetry (DSC), revealing thermal transitions at 281 8C (and 290 8C)/259 8C and 340 8C/315 8C for PTPD3T and PBTI3T, respectively (Fig. 1d), and indicating some degree of crystallinity26. Note that PTPD3T exhibits two close but discrete endotherms during heating, similar to those of the liquid-crystalline polymer poly(2,5-bis(3-alkylthiophen-2-yl)thieno[3,2-b]thiophene) (PBTTT)26. X-ray diffraction (XRD, Fig. 1e) scans of PTPD3T and PBTI3T ﬁlms show distinctive Bragg features, with PTPD3T exhibiting reﬂections up to third order, indicating greater ordering than in PBTI3T31. Together, the diffraction and DSC data suggest highly ordered microstructures, especially in PTPD3T, in marked contrast to many high-performance PSC polymers, which are signiﬁcantly less ordered27,32.

Solar cell fabrication and performance The BHJ solar cells used inverted architectures, indium tin oxide (ITO)/ZnO/polymer:PC71BM/MoOx/Ag (Fig. 2a). The ZnO electron extraction layer was deposited using sol–gel techniques33 and has a 24.40 eV conduction band33, providing a good match with the PC71BM LUMO (24.20 eV; Fig. 2b) for electron extraction. After active layer spin-coating, the MoOx hole extraction layer was vacuum-deposited, followed by the Ag anode. A range of activelayer compositions (polymer:PC71BM ratios), ﬁlm thicknesses,

solvents and thermal annealing conditions were investigated systematically (Supplementary Tables S2, S3). Processing additives, such as 1,8-diiodooctane (DIO) were used to optimize BHJ PSC performance by promoting nanoscale phase separation34. Without DIO, the highest performance for PTPD3T cells was Jsc ¼ 10.1 mA cm22, Voc ¼ 0.785 V, FF ¼ 63.7% and PCE ¼ 5.07%, and for PBTI3T cells was Jsc ¼ 10.7 mA cm22, Voc ¼ 0.831 V, FF ¼ 64.9% and PCE ¼ 5.79%. DIO addition signiﬁcantly enhanced performance, and optimized cells with 1:2 polymer:PC71BM, using CHCl3 as the solvent with 2% DIO produced PTPD3T cells with Jsc ¼ 12.5 mA cm22, Voc ¼ 0.795 V, FF ¼ 79.6% and PCE ¼ 7.90%, and PBTI3T cells with Jsc ¼ 12.9 mA cm22, Voc ¼ 0.859 V, FF ¼ 77.8% and PCE ¼ 8.66% (Table 1 summarizes the average metrics). Thermal annealing was found to diminish performance. Of the performance parameters, Jsc is somewhat lower here than in many high-PCE materials due to the suboptimal bandgaps2,9,32, and the present performance is largely ascribable to the exceptional ﬁll factors. In comparing PBTI3T and PTPD3T PSCs, the Voc of the PBTI3T cells is 60 mV greater than that of the PTPD3T cells, in agreement with the HOMO ordering. Note that the Jsc of the PBTI3T cells is slightly greater than that of the PTPD3T cells, consistent with the greater photocurrent generated in the red region (Fig. 2d). However, the PTPD3T PSCs achieve a higher ﬁll factor, consistent with a higher degree of PTPD3T ordering and p-faceon-electrode orientation. During ﬁlm thickness optimization, a relatively constant ﬁll factor is observed with increasing thickness (from 65 to 250 nm for PTPD3T and from 70 to 150 nm for PBTI3T), before reaching the optimum thickness (Supplementary Tables S4,

a

b −3.0 −3.73

−3.5

ITO ZnO Acceptor Donor MoOx

Energy levels (eV)

Glass

−4.20

−4.0 −4.5 PTPD3T

PBTI3T

−5.0 PC71BM

−5.5 −5.55

−6.0 Ag

−3.77

−5.58

−6.5

−6.10

−7.0

c

d PTPD3T in CF+DIO PBTI3T in CF+DIO

PBTI3T

PTPD3T

PTPD3T in CF PBTI3T in CF 100

2

80

−2 EQE (%)

Current density (mA cm−2)

0 −4 −6 −8 −10

60 40 20

−12 −14

0 0.0

0.2

0.4

0.6

0.8

Voltage (V)

400

500

600

700

800

Wavelength (nm)

Figure 2 | Device architecture and performance of inverted BHJ polymer solar cells. a, Inverted device architecture, ITO/ZnO/polymer:PC71BM/MoOx/Ag. b, Energy level diagrams for PTPD3T, PBTI3T and PC71BM. c, J–V characteristics of PTPD3T:PC71BM and PBTI3T:PC71BM BHJ inverted solar cells fabricated using chloroform (CF) as solvent without and with DIO as the processing additive under 100 W m22 AM 1.5G illumination. d, EQE spectra of optimized PTPD3T and PBTI3T inverted BHJ solar cells fabricated using chloroform (CF) as the solvent and DIO as the processing additive. The EQE integration versus AM1.5 reference spectrum yields Jsc values within +2% of those from the J–V data. NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics

© 2013 Macmillan Publishers Limited. All rights reserved.

3

ARTICLES

NATURE PHOTONICS

DOI: 10.1038/NPHOTON.2013.207

Table 1 | Charge transport and optimized photovoltaic response data for PTPD3T:PC71BM and PBTI3T:PC71BM inverted BHJ solar cells. Polymer

m h (TFT) (cm2 V21 s21)

m h (SCLC) (cm2 V21 s21)

Voc (V)

J cs (mA cm22)

FF (%)

PCE (%)*

PTPD3T PBTI3T

5.87 × 1022 2.74 × 1023

1.2 × 1023 1.5 × 1023

0.786 (0.795) 0.850 (0.859)

12.3 (12.5) 12.8 (12.9)

78.7 (79.6) 76.3 (77.8)

7.72 (7.90) 8.42 (8.66)

*Polymer:PC71BM ¼ 1:2 (wt/wt), CF as solvent, 2% DIO (vol/vol) as processing additive. Performance metrics are average numbers from 40 devices of each type. Numbers in parentheses are for the best performing cells.

S5, Fig. S8). Beyond this thickness, Voc and ﬁll factor fall, corresponding to increased non-geminate recombination with increasing charge carrier path lengths to the electrodes11.

Film microstructure, morphology and phase separation Neat and blend ﬁlm microstructures were characterized by grazing incidence wide-angle X-ray scattering (GIWAXS; for details see Supplementary Section S8). Molecular orientations and stacking distances for neat polymer ﬁlms and blend polymer ﬁlms with and without DIO are summarized in Supplementary Table S7. For neat PTPD3T and PBTI3T ﬁlms, in-plane (Fig. 3g) and out-of-plane linecuts (Fig. 3h) indicate mixed p-edge-on and p-face-on orientations on the ZnO substrates, as indicated by the scattering vector (q) ≈ 1.8 Å21 p–p stacking features in the inplane and out-of-plane linecuts35,36. The ratio of the areas under the out-of-plane p-stacking reﬂection:in-plane p-stacking reﬂection provides a semi-quantitative measure of the preference for p-faceon polymer orientation. For the neat ﬁlms, the face-on:edge-on (f:e) ratios are 1.7 and 2.1 for PTPD3T and PBTI3T, respectively. The p–p stacking distance for the neat PTPD3T ﬁlms is 3.65 Å, slightly larger than the 3.60 Å for PBTI3T. Using the procedure of Lilliu36, neat PTPD3T was determined to be paracrystalline (paracrystallinity: g ¼ 5.1+1.4%) from the lamellar (h00) stacking peaks (q ¼ 0.25, 0.51, 0.73 Å21), corresponding to the distance between neighbouring backbones, determined by side chain intercalation. This metric compares favourably to other ordered polymers35,36. Because PBTI3T does not exhibit multiple diffraction orders, paracrystallinity was not calculated. The Scherrer equation correlation length of the lamellar stacking in PBTI3T (5.7 nm) is signiﬁcantly smaller than for PTPD3T (9.1 nm), suggesting that PTPD3T has longer range order than PBTI3T. After PC71BM addition, PTPD3T becomes less ordered ( g ¼ 11.9+3.8%) and the PBTI3T p–p correlation length decreases from 5.7 to 2.6 nm, indicating decreased p–p stacking domain size (Fig. 3b,d). For the blend ﬁlms without DIO processing, PBTI3T shows slightly increased face-on preference (f:e ratio ¼ 3.7), while PTPD3T still shows a mixed orientation (f:e ratio ¼ 1.5). For DIO-aided blend ﬁlms, p–p stacking distances contract to 3.62 Å and 3.56 Å in PTPD3T:PC71BM and PBTI3T:PC71BM, respectively, possibly reﬂecting the extended drying time using DIO37. In contrast, reported p–p stacking distances for other high-performance BHJ donor polymers are typically somewhat longer at 3.65–3.90 Å (refs 27,32,38). From a previous PTB PSC study31 showing an inverse linear relationship between the p–p stacking distance and ﬁll factor, the present shorter PTPD3T and PBTI3T p–p stacking distances are consistent with high ﬁll factors32. However, other variables are probably also operative, because examples are known of rather different polymers where short p–p stacking distances do not correlate with exceptional ﬁll factors39. Furthermore, although the addition of DIO to the neat PTPD3T and PBTI3T ﬁlms does not promote signiﬁcant f:e changes, the DIO-processed blends exhibit almost complete p-face-on orientation for PTPD3T:PC71BM and predominant p-face-on orientation for PBTI3T:PC71BM (f:e ratio ¼ 14), which should enhance charge transport towards the electrodes. To our knowledge, this is the ﬁrst report of donor polymer preferential 4

p-face-on orientation on ZnO in inverted PSCs. However, note that the active layer is PCBM-enriched near the ZnO electrode. Short donor polymer p–p stacking distances and preferential p-face-on orientation are characteristics that should promote sizeable out-of-plane carrier mobilities and efﬁcient charge collection. Polymer:PC71BM ﬁlm morphologies were next investigated by transmission electron microscopy (TEM), atomic force microscopy (AFM) and X-ray photoelectron spectroscopy (XPS). Figure 4b shows a PBTI3T:PC71BM blend TEM image, indicating lateral nanoscale phase separation with 10–20 nm domains when DIO is used for ﬁlm processing, versus 30–80 nm domain sizes without DIO (Supplementary Fig. S10). Here, DIO addition greatly increases the materials’ miscibility and interfacial area for exciton dissociation40, in addition to promoting bicontinuous network formation. Without DIO, separate toroid-shaped domains are observed that lack effective interconnectivity for efﬁcient charge transport. Furthermore, photoluminescence spectra reveal greater photoluminescence quenching for DIO-processed ﬁlms than for those without DIO (Supplementary Fig. S12), indicating more efﬁcient exciton dissociation in the former. AFM images reveal similar trends, with nanoscale phase separation and bicontinuous networks emerging following DIO-aided processing. Such networks have reduced grain boundaries and should provide continuous pathways for carrier migration to the respective electrodes, favouring high ﬁll factors and Jsc values by suppressing charge accumulation and nongeminate recombination. Cross-sectional TEM images of optimized PBTI3T:PC71BM devices (Fig. 4c) reveal a PC71BM-rich layer near the blend/ZnO interface (the darker region corresponds to PC71BM). In support of this, S and C mapping of these cross-sections by energy-dispersive X-ray spectroscopy (EDS) reveals a C-rich layer at the blend/ZnO interface, distinct from the bulk ﬁlm. Correspondingly, the S content at this interface is signiﬁcantly lower, again indicating PC71BM enrichment near the blend/ZnO interface. Complementary depth-proﬁled XPS was also used to probe the composition, proceeding from the air/BHJ interface, through the active layer, and to the buried active layer/ZnO interface. Figure 4d shows changes in the S/C ratio that also indicate vertical phase gradation, with polymer enrichment at the blend/air interface, and PC71BM enrichment at the blend/ZnO interface. This gradation probably arises from differences in the polymer and PC71BM surface energies41, and the interactions with ZnO42. Contact angles for the polymer and PC71BM solutions and their corresponding spreading parameters on ZnO were measured (Supplementary Table S11), with the spreading parameter DW deﬁned as the ability of a liquid drop to adhere to the substrate surface during spin-coating43. The magnitude of DW is signiﬁcantly smaller for PC71BM than for both polymers, which should promote PC71BM accumulation at the ZnO/blend interface. Figure 4a shows a schematic of the polymer:PC71BM blend microstructure. Note that the polymerenriched and PCBM-enriched layers can function as partial charge-blocking regions to suppress electron leakage to the MoOx-coated anode and hole leakage to the ZnO-coated cathode, respectively, thereby reducing recombination at the contacts. Comparing the inverted cell performance to that of a conventional cell geometry further supports this role of vertical phase segregation.

NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics

© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE PHOTONICS

ARTICLES

DOI: 10.1038/NPHOTON.2013.207

a 2.0

c 2.0

b 2.0

PTPD3T:PC71BM

PTPD3T

1.0

1.5 qz (Å−1)

1.5 qz (Å−1)

qz (Å−1)

1.5

1.0

1.0

0.5

0.5

0.5

0.0

0.0

0.0

0.0

0.5

1.0

1.5

2.0

0.0

0.5

qxy (Å−1)

1.0

1.5

2.0

0.0

f 2.0

1.0

1.0

0.5

0.5

0.0

0.0

0.0

1.0

1.5

2.0

0.0

0.5

qxy (Å−1)

PBTI3T:PC71BM DIO

1.0

1.5

2.0

0.0

qxy (Å−1)

1.0

1.5

2.0

qxy (Å−1)

h

PBTI3T PBTI3T:PC71BM PBTI3T:PC71BM DIO

0.5

PTPD3T PTPD3T:PC71BM PTPD3T:PC71BM DIO

PBTI3T PBTI3T:PC71BM PBTI3T:PC71BM DIO

I(q) (a.u.)

PTPD3T PTPD3T:PC71BM PTPD3T:PC71BM DIO

I(q) (a.u.)

g

2.0

1.0

0.5

0.5

1.5

1.5 qz (Å−1)

1.5 qz (Å−1)

1.5

1.0 qxy (Å−1)

PBTI3T:PC71BM

PBTI3T

0.0

0.5

qxy (Å−1)

e 2.0

d 2.0

qz (Å−1)

PTPD3T:PC71BM DIO

0.5

1.0

1.5

2.0

0.5

qxy (Å−1)

1.0

1.5

2.0

qz (Å−1)

Figure 3 | GIWAXS data for neat and BHJ blend polymer ﬁlms. a, Two-dimensional (2D) GIWAXS image of neat PTPD3T ﬁlm. b, 2D GIWAXS image of PTPD3T:PC71BM blend ﬁlm without DIO. c, 2D GIWAXS image of PTPD3T:PC71BM blend ﬁlm with DIO as the processing additive. d, 2D GIWAXS image of neat PBTI3T ﬁlm. e, 2D GIWAXS image of PBTI3T:PC71BM blend ﬁlm without DIO. f, 2D GIWAXS image of PBTI3T:PC71BM blend ﬁlm with DIO as the processing additive. g, In-plane linecuts of 2D GIWAXS of polymer ﬁlms. h, Out-of-plane linecuts of 2D GIWAXS of polymer ﬁlms.

Thus, smaller ﬁll factors and Jsc values are observed in conventional cells (Supplementary Table S6, Fig. S9). Note that in inverted devices, the ﬁll factor is also enhanced in PTB7 PSCs3,4.

Carrier mobility Thin-ﬁlm transistor (TFT) and space charge-limited current (SCLC) measurements were used to investigate ﬁlm mobilities parallel and perpendicular to the substrate plane, respectively. The TFT hole mobilities (mh) measured in saturation are 0.008 cm2 V21 s21 and 0.038 cm2 V21 s21 for pristine PBTI3T and PTPD3T ﬁlms, respectively. These mobilities are in agreement with the ordered microstructures, and the higher mh of PTPD3T than in PBTI3T

can be associated with the aforementioned greater backbone coplanarity/ordering of PTPD3T. There is no signiﬁcant hole mobility change for both polymers on blending with PC71BM at levels achieving optimal PSC performance. The mh values are 0.003 cm2 V21 s21 and 0.059 cm2 V21 s21 for PBTI3T:PC71BM and PTPD3T:PC71BM, respectively, indicating minimal disruption of polymer ﬁlm microstructures after PC71BM blending. However, these TFT mobilities represent lateral mobilities at carrier concentrations generally greater than in PSCs. SCLC measurements were therefore used to quantify the carrier mobility perpendicular to the substrate plane. For hole-only devices, the blends are not space charge-limited, possibly due to the vertical phase gradation

NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics

© 2013 Macmillan Publishers Limited. All rights reserved.

5

ARTICLES

NATURE PHOTONICS

a

DOI: 10.1038/NPHOTON.2013.207

b MoOx

1 µm

Polymer

PC71BM ZnO

100 nm

c d 1.0

100 nm

C

S

Normalized peak area (a.u.)

Active layer

ZnO

0.8

0.6

0.4

C 1s Zn 2p S/C

0.2

0.0 0

400

800

1,200

1,600

Etching time (s)

Figure 4 | Morphology of a PBTI3T:PC71BM ﬁlm yielding optimal solar cell performance. a, Schematic of the PBTI3T:PC71BM blend ﬁlm showing vertical phase gradation with a polymer-rich layer near the MoOx/blend interface and a PC71BM-rich layer near the blend/ZnO interface. b, TEM image of a PBTI3T:PC71BM blend ﬁlm. Inset: AFM topographical image. c, TEM cross-sectional image and EDS C and S mappings of an optimized PBTI3T:PC71BM inverted solar cell. EDS mappings were performed simultaneously on the TEM cross-sectional sample across the active layer. d, XPS depth proﬁle of a PBTI3T:PC71BM blend ﬁlm showing S:C ratio evolution as a function of etching time, where etching begins at the air/ﬁlm interface.

noted above, and the PC71BM-rich layer on ZnO may block hole transport. Therefore, neat polymer ﬁlms were used to measure mh(SCLC), yielding 1.2 × 1023 and 1.5 × 1023 cm2 V21 s21 for PTPD3T and PBTI3T, respectively, processed without DIO. These mh values are comparable to or greater than those reported for typical high-PCE BHJ ﬁlms3,27, in agreement with the p-face-on orientations, ordered ﬁlm microstructures and close p–p spacings. Such substantial mh values, even in the SCLC regime, should partially account for the high ﬁll factors, although high mh alone cannot guarantee high device ﬁll factors44,45. The charge transport properties in blend ﬁlms with DIO were next examined in electron-only devices and show electron mobility (me) in the SCLC regime of 2.8 × 1025 and 1.2 × 1025 cm2 V21 s21 for PTPD3T:PC71BM and PBTI3T:PC71BM, respectively. These me values are two to ﬁve times higher than in blend ﬁlms without DIO (Supplementary Table S10), probably reﬂecting the suboptimal morphologies in the latter.

Non-geminate recombination dynamics Transient photocurrent (TPC)46 and transient photovoltage (TPV)47 measurements were performed on PBTI3T and PTPD3T PSCs fabricated with and without DIO following known procedures. The carrier density n, obtained from TPC measurements (Fig. 5a)46, and the dependence of n on Voc for different light intensities (from 6

0.025 to 1.0 sun), are illustrated in Fig. 5b. Note that devices prepared with DIO show signiﬁcantly higher carrier densities than those without DIO, consistent with the J–V data already discussed where DIO increases the current density due to enhanced exciton dissociation. Optimized PTPD3T and PBTI3T devices exhibit similar n values under one sun, consistent with their similar Jsc values under steady-state measurements. Carrier lifetimes t at Voc were extracted from TPV measurements using the same pulse light intensities as in the TPC experiments (Supplementary Section S14). The decay dynamics for all devices agree with previous measurements on P3HT:PC61BM blends, namely that mono-exponential ﬁts apply for a wide range of illumination intensities11,20,21,46–48. Figure 5c shows carrier lifetimes as a function of carrier density for varying illumination intensities. Under one sun, the DIO-processed PTPD3T and PBTI3T devices show carrier lifetimes of 2.5 and 1.6 ms, respectively, whereas PTPD3T and PBTI3T devices without DIO show carrier lifetimes of 7.2 and 3.2 ms, respectively. That devices without DIO have longer carrier lifetimes correlates with the signiﬁcantly lower carrier densities, and a similar lifetime dependence on carrier density has been reported previously by Shuttle et al. 46. However, at ﬁxed n, the DIO-processed PTPD3T and PBTI3T devices show much longer carrier lifetimes, consistent with the morphological changes demonstrated already. The bimolecular recombination

NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics

© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE PHOTONICS

ARTICLES

DOI: 10.1038/NPHOTON.2013.207

a 1.2 1.0 0.8

1.8 × 1017

PTPD3T in CF+DIO PBTI3T in CF+DIO

PTPD3T in CF PBTI3T in CF

1.5 × 1017 n (cm−3)

Current (mA)

b

PTPD3T in CF+DIO PBTI3T in CF+DIO PTPD3T in CF PBTI3T in CF

0.6

1.2 × 1017

9 × 1016

0.4 0.2 0.0

6 × 1016 0.0

0.5

1.0

2.0

0.5

0.6

c 10−3

PTPD3T in CF+DIO PBTI3T in CF+DIO

0.7

0.8

0.9

Voc (V)

Time (µs)

PTPD3T in CF PBTI3T in CF

10−12

τ (s)

krec (cm3 s−1)

10−4

d

10−5

PTPD3T in CF+DIO PBTI3T in CF+DIO

PTPD3T in CF PBTI3T in CF

10−13

10−14 10−6 1017

1.5 × 1017 n

8 × 1016

2 × 1017

1.2 × 1017

1.6 × 1017

n (cm−3)

(cm−3)

Figure 5 | TPC and TPV analysis of BHJ inverted solar cells fabricated using CF as the solvent with or without DIO as the processing additive. a, TPC measured under device short-circuit condition without white light bias. b, Average charge density n, extracted from TPC measurements as a function of Voc , corresponding to different bias light intensities from 0.025 to 1.0 sun. c, Charge carrier lifetime t, determined from TPV measured under device open-circuit condition using the same power of pulsed source as the TPC measurement, as a function of charge density. d, Bimolecular recombination rate constant krec , extracted from carrier t and n, as a function of n. All cells were masked to an illuminated area of 0.375 cm2. The pulsed source power was set so that the DV of the TPV measurement did not exceed 3% of the Voc measured at 0.1 sun.

rate constants krec (Fig. 5d) are calculated from the carrier lifetimes and densities according to krec ¼ 1/(l þ 1)nt (n) (ref. 17), where l is an empirical constant determined from Fig. 5c (Supplementary Section S14). At a given light intensity, krec for the DIO devices is invariably larger than for devices without DIO, while for a given carrier density, DIO-processed devices show drastically reduced krec values. Note that the krec from the TPV analysis could be signiﬁcantlyq smaller than that predicted by the Langevin equation, kL = min(mn , mp ), where q is the electronic charge, 1 is the dielec1 tric constant, and mn(p) is the electron (hole) mobility49. Reduced recombination constants are considered critical for efﬁcient PSC performance17,20,21, with t and krec values being direct indicators of charge extraction at Voc and strongly dependent on ﬁlm morphology. The longer carrier lifetimes and smaller krec values for the DIO-processed devices are in good agreement with the GIWAXS analysis; DIO-processed devices show superior charge extraction properties due to their optimized morphology and macromolecular orientation. Comparing DIO-processed PTPD3T and PBTI3T PSCs, note also that the former exhibit slower decay dynamics under open-circuit conditions (longer carrier lifetimes, smaller krec), consistent with the slightly greater ﬁll factor. Compared to the reported krec for P3HT:PC61BM46,47, our data for optimized PTPD3T and PBTI3T PSCs indicate far slower decay dynamics, emphasizing that optimal ﬁlm morphology and favourable phase gradation are critical parameters for efﬁcient charge extraction at low internal ﬁelds.

Conclusions New in-chain donor–acceptor polymeric semiconductors yield high ﬁll factor/PCE bulk-heterojunction PSCs. Imide-functionalized PBTI3T and PTPD3T synergistically integrate speciﬁc structural

and electronic characteristics into single macromolecules that enhance PCE, principally by enlarging Voc and achieving exceptional ﬁll factors. Although PSC mechanisms involve the interplay of many factors, this work shows that ingredients speciﬁc to 76–80% ﬁll factors include highly-ordered, p-face-on electrode oriented microstructures with close p–p interplanar spacings, ordered BHJ bicontinuous networks and vertical phase gradation. These factors contribute to superior charge carrier mobilities and lifetimes, both responsible for large ﬁll factors and PCEs up to 8.7% from polymers with suboptimal bandgaps of 1.8 eV. From equation (1), achieving such ﬁll factors implies that PCEs signiﬁcantly greater than 10% should be realizable by further modifying the macromolecular architecture.

Methods Polymer synthetic and characterization details are provided in Supplementary Section S2. AFM measurements were performed using a Dimension Icon scanning probe microscope (Veeco) in tapping mode. XRD measurements were performed on an 18 kW Rigaku ATXG diffractometer using a multilayer parabolic mirror, NaI scintillation detector, and X-rays with a wavelength of 1.541 Å. TEM measurements were performed using a JEOL JEM-2100F instrument. For top-down images, specimens were prepared as for actual devices, but were dropcast onto polished 2 mm × 2 mm NaCl (100) substrates (MTI Corporation). After drying, substrates were transferred to deionized water and the ﬂoated ﬁlms transferred to lacey carbon TEM grids (Ted Pella). For cross-sectional TEM, samples were prepared as for actual devices, but with a 50 nm ITO substrate as the cathode. TEM specimens were prepared using focused ion beam (FIB) techniques (FEI Helios NanoLab 600). A thin Pt layer was locally deposited on the sample to protect it from damage during FIB processing. The sample was then lifted with an OmniProbe nanomanipulator and transferred to a semi-spherical Cu TEM grid. STEM imaging was conducted with a high-angle annular dark-ﬁeld detector on the JEM-2100F microscope. The two-dimensional EDS analysis was performed in STEM mode with a 1 nm probe size. XPS (Omicron ESCA Probe) depth proﬁling was performed on devices without

NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics

© 2013 Macmillan Publishers Limited. All rights reserved.

7

ARTICLES

NATURE PHOTONICS

the top electrode (ITO/ZnO/polymer:PC71BM). The samples were sputtered with an Arþ gun at 4,000 eV for 30 s intervals and etched from the air/blend interface. The integrated areas from the C 1s, S 2p and Zn 2p peaks were used for data analysis. Solar cell fabrication and characterization. Prepatterned ITO-coated glass (Thin Film Devices) with a sheet resistance of 10 V A21 was used as the substrate after cleaning by sequential sonication in hexane, deionized water, methanol, isopropanol and acetone. After 20 min ultraviolet/ozone treatment (Jelight), an 20 nm ZnO layer was prepared by spin-coating (at 5,000 r.p.m.) a precursor solution prepared from 0.5 M zinc acetate dehydrate in 0.5 M monoethanolamine and 2-methoxy ethanol. After cleaning the electrical contacts, the ZnO substrates were immediately baked in air at 170 8C for 10 min, then rinsed with deionized water, isopropanol and acetone, and dried in a glovebox. Active-layer solutions were prepared from 8 mg ml21 PBTI3T solution or 10 mg ml21 PTPD3T solution in chloroform (CF) or CF/DIO in various ratios. For optimum performance, PBTI3T and PTPD3T devices were spin-coated at 5,000 and 3,000 r.p.m., respectively. Thin 7.5 nm MoOx and 140 nm Ag were then thermally evaporated through a shadow mask at 1 × 1026 torr. For fabricating conventional PSCs, PEDOT:PSS (Clevios P VP Al 4083) was spin-coated at 5,000 r.p.m. for 30 s and annealed at 150 8C. Active layers were prepared under the same conditions as the inverted PSCs. LiF(1.0 nm)/Al(100 nm) were then thermally evaporated through a shadow mask at 1 × 1026 torr. For device characterization, J–V characteristics were measured under AM1.5G light (100 mW cm22) using the xenon arc lamp of a Spectra-Nova Class A solar simulator. Light intensity was calibrated using an NREL-certiﬁed monocrystalline Si diode coupled to a KG3 ﬁlter to bring spectral mismatch to unity. A Keithley 2400 source meter was used for electrical characterization. The area of all devices was 6 mm2, and a 6 mm2 aperture was placed on top of cells during all measurements. External quantum efﬁciencies (EQEs) were characterized using an Oriel model QE-PV-SI instrument equipped with a NIST-certiﬁed Si diode. Monochromatic light was generated from an Oriel 300 W lamp. TPC and TPV measurements. Transient measurements were performed on all the cells using a ﬁbre-optic-coupled white light source providing a steady-state light bias, while pulsed excitation was provided by the tunable output of an optical parametric ampliﬁer driven by a Ti:sapphire laser system50. The white light bias was not corrected for spectral mismatch. The laser repetition rate was reduced to 500 Hz using a chopper to have an experimental window compatible with the TPV kinetics. An excitation wavelength of 600 nm was used, and the beam was passed through a depolarizer and square diffuser to ensure uniform sample cell excitation. The cells were masked to an illuminated area of 0.375 cm2. For the TPC measurements, no white light bias was used, and the cells were directly connected to the input of a transimpedance ampliﬁer. The ampliﬁer voltage output was recorded on an oscilloscope. For the TPV measurements the cells were directly connected to the 1 MV input of the oscilloscope. All transients shown were the average of 128 scans. The power of the pulsed source was set so that the DV of the TPV measurement did not exceed 3% of the Voc measured at 0.1 sun illumination intensity. The power of the pulsed source was held constant for each pair of cells (with and without DIO) and for their respective TPC and TPV measurements.

Received 14 October 2012; accepted 1 July 2013; published online 11 August 2013

References 1. Yu, G., Gao, J., Hummelen, J. C., Wudl, F. & Heeger, A. J. Polymer photovoltaic cells: enhanced efﬁciencies via a network of internal donor–acceptor heterojunctions. Science 270, 1789–1791 (1995). 2. Cheng, Y-J., Yang, S-H. & Hsu, C-S. Synthesis of conjugated polymers for organic solar cell applications. Chem. Rev. 109, 5868–5923 (2009). 3. He, Z. et al. Simultaneous enhancement of open-circuit voltage, short-circuit current density, and ﬁll factor in polymer solar cells. Adv. Mater. 23, 4636–4643 (2011). 4. He, Z. C. et al. Enhanced power-conversion efﬁciency in polymer solar cells using an inverted device structure. Nature Photon. 6, 591–595 (2012). 5. Green, M. A., Emery, K., Hishikawa, Y., Warta, W. & Dunlop, E. D. Solar cell efﬁciency tables (version 40). Prog. Photovolt. 20, 606–614 (2012). 6. Scharber, M. C. et al. Design rules for donors in bulk-heterojunction solar cells— towards 10% energy-conversion efﬁciency. Adv. Mater. 18, 789–794 (2006). 7. Mihailetchi, V. D., Xie, H. X., de Boer, B., Koster, L. J. A. & Blom, P. W. M. Charge transport and photocurrent generation in poly(3hexylthiophene):methanofullerene bulk-heterojunction solar cells. Adv. Funct. Mater. 16, 699–708 (2006). 8. Wagner, J. et al. High ﬁll factor and open circuit voltage in organic photovoltaic cells with diindenoperylene as donor material. Adv. Funct. Mater. 20, 4295–4303 (2010). 9. Li, G., Zhu, R. & Yang, Y. Polymer solar cells. Nature Photon. 6, 153–161 (2012). 10. Clarke, T. M. et al. Charge carrier mobility, bimolecular recombination and trapping in polycarbazole copolymer:fullerene (PCDTBT:PCBM) bulk heterojunction solar cells. Org. Electron. 13, 2639–2646 (2012). 8

DOI: 10.1038/NPHOTON.2013.207

11. Shuttle, C. G., Hamilton, R., O’Regan, B. C., Nelson, J. & Durrant, J. R. Chargedensity-based analysis of the current–voltage response of polythiophene/fullerene photovoltaic devices. Proc. Natl Acad. Sci. USA 107, 16448–16452 (2010). 12. Zhang, Y., Dang, X. D., Kim, C. & Nguyen, T. Q. Effect of charge recombination on the ﬁll factor of small molecule bulk heterojunction solar cells. Adv. Energy Mater. 1, 610–617 (2011). 13. Cowan, S. R., Banerji, N., Leong, W. L. & Heeger, A. J. Charge formation, recombination, and sweep-out dynamics in organic solar cells. Adv. Funct. Mater. 22, 1116–1128 (2012). 14. Kirchartz, T., Pieters, B. E., Kirkpatrick, J., Rau, U. & Nelson, J. Recombination via tail states in polythiophene: fullerene solar cells. Phys. Rev. B 83, 115209 (2011). 15. Cowan, S. R., Roy, A. & Heeger, A. J. Recombination in polymer-fullerene bulk heterojunction solar cells. Phys. Rev. B 82, 245207 (2010). 16. Koster, L. J. A., Mihailetchi, V. D., Ramaker, R. & Blom, P. W. M. Light intensity dependence of open-circuit voltage of polymer:fullerene solar cells. Appl. Phys. Lett. 86, 123509 (2005). 17. Shuttle, C. G. et al. Bimolecular recombination losses in polythiophene: fullerene solar cells. Phys. Rev. B 78, 113201 (2008). 18. Credgington, D., Hamilton, R., Atienzar, P., Nelson, J. & Durrant, J. R. Nongeminate recombination as the primary determinant of open-circuit voltage in polythiophene:fullerene blend solar cells: an analysis of the inﬂuence of device processing conditions. Adv. Funct. Mater. 21, 2744–2753 (2011). 19. Street, R. A., Schoendorf, M., Roy, A. & Lee, J. H. Interface state recombination in organic solar cells. Phys. Rev. B 81, 205307 (2010). 20. Mauer, R., Howard, I. A. & Laquai, F. Effect of nongeminate recombination on ﬁll factor in polythiophene/methanofullerene organic solar cells. J. Phys. Chem. Lett. 1, 3500–3505 (2010). 21. Mauer, R., Howard, I. A. & Laquai, F. Effect of external bias on nongeminate recombination in polythiophene/methanofullerene organic solar cells. J. Phys. Chem. Lett. 2, 1736–1741 (2011). 22. Proctor, C. M., Kim, C., Neher, D. & Nguyen, T-Q. Nongeminate recombination and charge transport limitations in diketopyrrolopyrrole-based solutionprocessed small molecule solar cells. Adv. Funct. Mater. http://dx.doi.org/10. 1002/adfm.201202643. 23. Green, M. A. Solar-cell ﬁll factors: general graph and empirical expressions. Solid-State Electron. 24, 788–789 (1981). 24. Kippelen, B. & Bredas, J. L. Organic photovoltaics. Energy Environ. Sci. 2, 251–261 (2009). 25. Zhao, J., Wang, A. & Green, M. A. 24.5% efﬁciency silicon PERT cells on MCZ substrates and 24.7% efﬁciency PERL cells on FZ substrates. Prog. Photovolt. 7, 471–474 (1999). 26. McCulloch, I. et al. Semiconducting thienothiophene copolymers: design, synthesis, morphology, and performance in thin-ﬁlm organic transistors. Adv. Mater. 21, 1091–1109 (2009). 27. Guo, X. et al. Bithiopheneimide-dithienosilole/dithienogermole copolymers for efﬁcient solar cells: information from structure–property–device performance correlations and comparison to thieno[3,4-c]pyrrole-4,6-dione analogues. J. Am. Chem. Soc. 134, 18427–18439 (2012). 28. Ko, S. et al. Controlled conjugated backbone twisting for an increased opencircuit voltage while having a high short-circuit current in poly(hexylthiophene) derivatives. J. Am. Chem. Soc. 134, 5222–5232 (2012). 29. Dimitrov, S. D. et al. On the energetic dependence of charge separation in lowband-gap polymer/fullerene blends. J. Am. Chem. Soc. 134, 18189–18192 (2012). 30. Bakulin, A. A. et al. The role of driving energy and delocalized states for charge separation in organic semiconductors. Science 335, 1340–1344 (2012). 31. Chabinyc, M. L. X-ray scattering from ﬁlms of semiconducting polymers. Polym. Rev. 48, 463–492 (2008). 32. Szarko, J. M. et al. When function follows form: effects of donor copolymer side chains on ﬁlm morphology and BHJ solar cell performance. Adv. Mater. 22, 5468–5472 (2010). 33. Chou, C. H., Kwan, W. L., Hong, Z., Chen, L. M. & Yang, Y. A metal-oxide interconnection layer for polymer tandem solar cells with an inverted architecture. Adv. Mater. 23, 1282–1286 (2011). 34. Peet, J., Heeger, A. J. & Bazan, G. C. ‘Plastic’ solar cells: self-assembly of bulk heterojunction nanomaterials by spontaneous phase separation. Acc. Chem. Res. 42, 1700–1708 (2009). 35. Rivnay, J. et al. Drastic control of texture in a high performance n-type polymeric semiconductor and implications for charge transport. Macromolecules 44, 5246–5255 (2011). 36. Lilliu, S. et al. Dynamics of crystallization and disorder during annealing of P3HT/PCBM bulk heterojunctions. Macromolecules 44, 2725–2734 (2011). 37. Rogers, J. T., Schmidt, K., Toney, M. F., Kramer, E. J. & Bazan, G. C. Structural order in bulk heterojunction ﬁlms prepared with solvent additives. Adv. Mater. 23, 2284–2288 (2011). 38. Piliego, C. et al. Synthetic control of structural order in N-alkylthieno[3,4c]pyrrole-4,6-dione-based polymers for efﬁcient solar cells. J. Am. Chem. Soc. 132, 7595–7597 (2010).

NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics

© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE PHOTONICS

ARTICLES

DOI: 10.1038/NPHOTON.2013.207

39. Osaka, I. et al. Synthesis, characterization, and transistor and solar cell applications of a naphthobisthiadiazole-based semiconducting polymer. J. Am. Chem. Soc. 134, 3498–3507 (2012). 40. Peet, J. et al. Efﬁciency enhancement in low-bandgap polymer solar cells by processing with alkane dithiols. Nature Mater. 6, 497–500 (2007). 41. Tada, A., Geng, Y., Wei, Q., Hashimoto, K. & Tajima, K. Tailoring organic heterojunction interfaces in bilayer polymer photovoltaic devices. Nature Mater. 10, 450–455 (2011). 42. Bjorstrom, C. M. et al. Vertical phase separation in spin-coated ﬁlms of a low bandgap polyﬂuorene/PCBM blend: effects of speciﬁc substrate interaction. Appl. Surf. Sci. 253, 3906–3912 (2007). 43. Wobkenberg, P. H. et al. Low-voltage organic transistors based on solution processed semiconductors and self-assembled monolayer gate dielectrics. Appl. Phys. Lett. 93, 013303 (2008). 44. Bronstein, H. et al. Thieno[3,2-b]thiophene-diketopyrrolopyrrole-containing polymers for high-performance organic ﬁeld-effect transistors and organic photovoltaic devices. J. Am. Chem. Soc. 133, 3272–3275 (2011). 45. Parmer, J. E. et al. Organic bulk heterojunction solar cells using poly(2,5-bis(3tetradecyllthiophen-2-yl)thieno[3,2,-b]thiophene). Appl. Phys. Lett. 92, 113309 (2008). 46. Shuttle, C. G. et al. Charge extraction analysis of charge carrier densities in a polythiophene/fullerene solar cell: analysis of the origin of the device dark current. Appl. Phys. Lett. 93, 183501 (2008). 47. Maurano, A. et al. Recombination dynamics as a key determinant of open circuit voltage in organic bulk heterojunction solar cells: a comparison of four different donor polymers. Adv. Mater. 22, 4987–4992 (2010). 48. Li, Z., Gao, F., Greenham, N. C. & McNeill, C. R. Comparison of the operation of polymer/fullerene, polymer/polymer, and polymer/nanocrystal solar cells: a transient photocurrent and photovoltage study. Adv. Funct. Mater. 21, 1419–1431 (2011). 49. Mandoc, M. M., Kooistra, F. B., Hummelen, J. C., de Boer, B. & Blom, P. W. M. Effect of traps on the performance of bulk heterojunction organic solar cells. Appl. Phys. Lett. 91, 263505 (2007). 50. McArthur, E. A., Morris-Cohen, A. J., Knowles, K. E. & Weiss, E. A. Charge carrier resolved relaxation of the ﬁrst excitonic state in CdSe quantum dots probed with near-infrared transient absorption spectroscopy. J. Phys. Chem. B 114, 14514–14520 (2010).

Acknowledgements This research was supported as part of the ANSER Center, an Energy Frontier Research Center funded by the US Department of Energy, Ofﬁce of Science, and Ofﬁce of Basic Energy Sciences (award no. DE-SC0001059), by Polyera Corporation, and by AFOSR (FA9550-08-1-0331). The authors acknowledge the NSF-MRSEC programme of the Northwestern University Materials Research Science and Engineering Center for characterization facilities (DMR-1121262) and the Institute for Sustainability and Energy at Northwestern (ISEN) for partial funding for equipment. The microscopy work was performed in the EPIC facility of the NUANCE Center at Northwestern University, which is supported by NSF-NSEC, NSF-MRSEC, the Keck Foundation, the State of Illinois and Northwestern University. X.G. acknowledges ﬁnancial support from an SUSTC start-up fund. Use of the Advanced Photon Source was supported by the US Department of Energy, Ofﬁce of Science, Ofﬁce of Basic Energy Sciences (contract no. DE-AC02-06CH11357). R.P.O. acknowledges the MICINN of Spain for a Ramo`n y Cajal research contract. J.T.L.N. acknowledges ﬁnancial support from the MICINN (project no. CTQ2012-33733) and the Junta de Andalucı´a (project no. PO9-4708). D.B.T. is funded by the NSF-IGERT Program.

Author contributions X.G. designed and performed materials synthesis and characterization. N.Z. fabricated the solar cell device and characterized the ﬁlm morphology using XRD, AFM, TEM and XPS. S.J.L. and J.St. performed two-dimensional GIWAXS ﬁlm characterization. J.W.H. (OFET) and J.Sm. (SCLC) characterized charge carrier mobility. R.P.O. and J.T.L.N. performed DFT calculations. N.Z. and S.L. characterized the ﬁlm morphology by cross-sectional TEM. N.Z., J.Sm. and D.B.T. carried out transient photovoltage and photocurrent measurements. X.G., N.Z. and T.J.M. prepared the manuscript. All authors discussed the results and commented on the manuscript. L.X.C., R.P.H.C., A.F. and T.J.M. supervised the project.

Additional information Supplementary information is available in the online version of the paper. Reprints and permissions information is available online at www.nature.com/reprints. Correspondence and requests for materials should be addressed to L.X.C., R.P.H.C., A.F. and T.J.M.

Competing ﬁnancial interests The authors declare no competing ﬁnancial interests.

NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics

© 2013 Macmillan Publishers Limited. All rights reserved.

9