The study of solvent additive effects in efficient polymer photovoltaicsvia impedance spectroscopy

En-Ping Yao a,b, Chun-Chao Chen a, Jing Gao a, Yongsheng Liu a, Qi Chen a, Min Cai a,Wei-Chou Hsu b,n, Ziruo Hong a,n, Gang Li a,n, Yang Yang a,n

a Department of Materials Science and Engineering, University of California-Los Angeles, Los Angeles, CA 90095, USAb Institute of Microelectronics, National Cheng Kung University, No.1, University Road, Tainan City 70101, Taiwan

a r t i c l e i n f o

Article history:Received 13 March 2014Received in revised form12 May 2014Accepted 24 May 2014Available online 12 July 2014

Keywords:OPVImpedance spectroscopyBulk heterojuctionInterface

a b s t r a c t

1, 8-Diiodooctane (DIO) has been known for its role of improving the polymer morphology andenhancing performance of polymer bulk heterojunction (BHJ) solar cell. In this work, the impedancespectroscopy was used to investigate the interface of poly(4,8-bis-alkyloxybenzo(1,2-b:4,5-b0)dithio-phene-2,6-diyl-alt- (alkyl thieno(3,4-b) thiophene-2-carboxylate)-2,6-diyl) (PBDTTT-C):PC70BM in BHJwith DIO as additive. Based on our results, we were able to simulate the device into an equivalent circuitmodel, which allows us to conveniently analyze the organic/organic interfacial contact in the organicphotovoltaic (OPV) device. Thus, we demonstrate that the impedance spectroscopy can an effectiveapproach in characterizing the donor/acceptor interfaces, such that a direct correlation can beestablished between the morphology and the device performance of BHJ devices.

& 2014 Published by Elsevier B.V.

1. Introduction

Organic photovoltaics (OPVs) have attracted considerableattention in the past decade because of their unique advantagesover their inorganic counterparts, such as low-cost fabrication,compatible with flexible substrates, and roll-to-roll processing [1].Polymer photovoltaic cells are based on the donor/acceptor con-cept that Frankel excitons with large binding energy are split at thedonor/acceptor interfaces due to chemical potential difference [2].The BHJ structure has been widely used, in which the donor andacceptor are blended, forming intimate contact on molecular levelthat allows excitons to reach the D/A interfaces for dissociationinto free charges. Normally polymers are used as electron donor,while fullerene C60, C70, and their derivatives, such as PCBM, aregiven as electron acceptors [3–10]. Therefore, some high efficientOPVs have been published in single junction [11–14] and tandemdevices [15–19]. In order to maximize power conversion efficiency(PCE), it is important to form interpenetrating networks for boththe purpose of exciton dissociation and charge transport [20]. Themorphology on microscale is strongly depending on the proces-sing parameters, such as blending ratio, solvent, and annealing,through which an optimal morphology can be reached.

One effective method is to use high boiling-point solventadditive to control morphology. For example, 1,8-octanedithiol(OT) [21], a high-boiling-point minor solvent with favorablesolubility only for PCBM, is able to cause a phase separation inboth P3HT:PCBM and PCPDTBT:PCBM BHJ films at optimal scale.Its poor solubility for polymers forces the precipitation of polymercomponent during the evaporation of major solvent. Similarsolvent additives [22–26] also works well, causing dramaticimprovement in the efficiency.

Hence the understanding of how the additive works and howto make the connection between film morphology and electricalproperties becomes much important. Morphology study on micro-and nano-scales involves various tools in order to gain insight inthe phase separation, and aggregation behaviors of individualmolecules or conjugated unites [27–32]. On the other hand,morphological characteristics dictate the optical and electricalproperties in the polymer:fullerene blend films. Electrical char-acteristics of OPV devices are normally characterized by severaltools [33–37]. These methods reveal detail information on thecharge dynamics in the OPV cells, giving empirical guidelines formorphology modification [38]. However it is still difficult todirectly probe the charge generation and recombination at thedonor/acceptor interfaces on the molecular level. The correlationbetween the film morphology and the electric properties of OPVdevices therefore remains unclear.

In OPV devices, several interfaces form upon direct contacts ofindividual components, and each interface plays an important role[39]. Assuming Ohmic contacts for charge collection by both anode

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/solmat

Solar Energy Materials & Solar Cells

http://dx.doi.org/10.1016/j.solmat.2014.05.0490927-0248/& 2014 Published by Elsevier B.V.

n Corresponding authors.E-mail addresses: wchsu@eembox.ncku.edu.tw (W.-C. Hsu),

zrhong@ucla.edu (Z. Hong), gangl@ucla.edu (G. Li), yangy@ucla.edu (Y. Yang).

Solar Energy Materials & Solar Cells 130 (2014) 20–26

and cathode in optimal device architectures, the donor/acceptorinterfaces and film morphology in polymer blend films are there-fore of significant importance to maximize photovoltaic efficiency.Here, we employ impedance spectroscopy (IS) to investigate theorganic/organic interfaces under various morphological scenariosinduced by solvent additives. It allows us to link the morphologyto the electrical properties of the OPV cells.

IS has been widely used in determining the electric character-istics through simple equivalent circuit models in many optoelec-tronic devices like dye sensitized solar cells (DSSC) [40], organiclight emitting diodes (OLEDs) [41], organic thin film transistors(OTFTs) [42], and OPVs [43,44]. In our case, the interfaces of theOPV devices can be considered as simple elements such asresistances and capacitances in circuit models. With the represen-tation of the elements, the electric properties of the BHJ can beconnected to current density–voltage (J–V) characteristics of thedevices. More importantly, the interface between the donor andacceptor in the BHJ can directly relate to the morphology of theBHJ. Hence, IS can be a powerful method to connect the morphol-ogy of the BHJ and the performance of the devices.

PBDTTTT-C is among the earliest polymers achieving over 6%efficiency in OPVs [45]. It has clear response to solvent additive,making it an ideal model system to study both morphology andelectrical properties in polymer films. Yet it is not well understood,even neglected, due to the overwhelming progress of the field ofpolymer solar cells. In this work, we used IS method andequivalent circuit model to characterize photovoltaic cells basedon PBDTTT-C, and scrutinize the relationship between themorphology and the device performance under the influence of1, 8-Diiodooctane (DIO) solvent additive.

2. Experimental setting

The OPV devices have a structure of Indium Tin Oxide (ITO)/poly(3,4-ethylenedioxythiophene): poly(styrene sulfonate) (PEDOT:PSS)/PBDTTT-C:P70BM/Calcium (Ca)/Aluminum (Al). First, ITO-coated glasssubstrates were cleaned sequentially in acetone, isopropanol, and DIwater by ultrasonic cleaner. The cleaned ITO substrates were exposedto ozone for 15 min in a UV ozone system. Then, the PEDOT:PSS wasdropped on each substrate (about 30 nm) and annealed at 120 1C for15 min. The PBDTTT-C:PC70BM layer was deposited about 90 nm onthe PEDOT:PSS by spin-coating the blend chlorobenzene and 1,2-dichlorobenzene solution with 10mg/mL PBDTTT-C and 15mg/mLPC70BM with or without 3% volume ratio of DIO in nitrogen filledglove box. At last, the Ca and Al were deposited by thermal evapora-tion with the rate of 0.05 nm/s and 0.2 nm/s and the thickness of20 nm and 100 nm under �10�6 Torr, respectively. The device activearea was 0.1 cm2 for all the solar cell devices discussed in this work.Device characterizationwas measure under simulated AM1.5G irradia-tion (100 mW/cm2) using a xenon-lamp-based solar simulator. The ISmeasurements were implemented using a Hewlett-Packard precisionLCR meter 4284 A. The frequency range was from 100 Hz to 1 MHz,and the DC bias of each device was set at the value of the voltage thatdrives the device at 1 mA/cm2 in the dark with the magnitude of thealternative signal at the value of 30 mV. The obtained IS data werefitted with EIS spectrum Analyzer in terms of appropriate equivalentcircuits.

3. Results and discussion

3.1. J–V characteristics

The J–V curves of the PBDTTT-C:PC70BM-based devices incholorobenzene (CB) or dicholorobenzene (DCB) solution and with

or without and with DIO additive are shown in Fig. 1, and thecorresponding performance parameters of the devices are listedon the table inserted in Fig. 1. Using CB only as solvent toprocessing the polymer:fullerene blend films, the device shows aPCE of 3.08%. DCB has a much higher boiling point of 180 1C thanthat of CB (130 1C), yielding a better PCE of 4.2%, with improvedshort circuit current density (Jsc) and fill factor (FF). The morpho-logical difference between the devices made with CB and DCB dueto the different boiling points can determine the phase separation,the domain formation and the intermolecular packing of polymerchains and PCBM molecules during the drying process of the films.It is obvious to see the dramatic improvement in PCE, which canbe explained from the morphological changing by DIO. Also byadding DIO in both CB and DCB, the devices processed from bothsolvents show almost identical J–V characteristics and PCE, withlower open circuit voltage (Voc) but higher Jsc and better FF, incomparison with those of OPV devices without any additive.

3.2. TEM analysis and phase images of AFM

Fig. 2 shows the transmission electron microscopy (TEM)images and atomic force microscopy (AFM) phases image ofPBDTTT-C:PC70BM film in CB or DCB solution and with or without3% DIO. In the images, the prominent area should be the crystal-lization of PC70BM, and the other area is the series of PBDTTT-Cchains. The prominent units become smaller by doping 3% DIOinto the CB or DCB implying that the aggregation of the PC70BMdecreases in scale apparently. Moreover, since the domain size ofPC70BM decreases, the improved phase separation was expectedby doping with 3% DIO in both CB and DCB. Hence, adding DIOreduces the size of the domains in the film, and it gives longertime for the molecules to reorganize and assemble into smaller butfiner phase separation. With better phase separation, the electron–hole pair would be easily dissociated since more interface areabetween donor and acceptor molecules is obtained. Besides, withreduced domain size and less phase separation, it would be easierfor the excitons to diffuse to the donor/acceptor interfaces, and togenerate free carriers. It also results in the fine interpenetratingnetworks in the whole bulk heterojunction and forms pathwaysfor efficient charge transport and collection. Hence, charge extrac-tion and injection are significantly enhanced, as strongly indicatedthe improved Jsc and FF.

With the confirmed photovoltaic performance which is con-sistent with previous reports [46] on PBDTTT-C:PC70BM with andwithout DIO, we thus started electrical characterization of the

Fig. 1. J–V curve of PBDTTT-C:PC70BM-based OPV devices in different conditions.The inset listed the performance of each device.

E.-P. Yao et al. / Solar Energy Materials & Solar Cells 130 (2014) 20–26 21

Fig. 2. The AFM phase image of the PBDTTT-C:PC70BM film made from the solution of (a) DCB without DIO, (b) DCB with DIO, (c) CB without DIO, and (d) CB with DIO, andthe corresponding TEM image from the solution of (e) DCB without DIO, (f) DCB with DIO, (g) CB without DIO, and (h) CB with DIO.

E.-P. Yao et al. / Solar Energy Materials & Solar Cells 130 (2014) 20–2622

polymer:fullerene blend films, so as to understand the effects ofDIO additive.

3.3. Equivalent circuit model and impedance analysis

According to the sandwiched device structure, a circuit modelis defined as shown in Fig. 3(a). The shunt pair with R1 and C1 inFig. 3(a) corresponds to the BHJ layer. The capacitance of the activelayer, C1, is usually called “diffusion capacitance” [47] or “chemicalcapacitance” [48] as the devices are under forward bias. The shuntpair with R2 and C2 in Fig. 2(a) corresponds to the two electricalcontacts of the interfaces between PEDOT:PSS/PBDTTT-C:PC70BMand PBDTTT-C:PC70BM/Ca. The R3 in Fig. 2(a) corresponds to theelectrodes include the resistance of ITO, PEDOT:PSS, Ca, and Al.

As measuring impedance spectroscopy at varied frequency, theimpedance of the device shall be a complex number consisting ofthe real and imaginary parts, which are also called the resistance

and the reactance, respectively. Fig. 3(b) shows the resistance vsfrequency characteristics, Fig. 3(c) shows the reactance vs fre-quency, and Fig. 3(d) shows the combination of the resistance vscorresponding reactance at different frequencies, called Cole–Coleplot. In Fig. 3(b), the resistance of each device decreases as thefrequency increases from 100 Hz to 1 MHz. The result implies thatthere should be at least one capacitance shunted with theresistances in the equivalent model so that the circuit is moreclose to short circuit condition while frequency gets higher.Besides, the Cole–Cole plot in Fig. 3(d) is not really a smoothsemicircle, which indicates that there is more than one pair ofresistance–capacitance-shunt in the equivalent model. The domi-nant resistance causing the large semicircle in low frequency rangeof the Cole–Cole plot should be the shunt pair belonging to BHJpart, and the interface represents the other smaller semicircle ofthe Cole–Cole plot. Moreover, there are some resistances due tothe whole capacitance-shunted pairs, which represent the

0 100 200 300 400 500 600 700 800 9000

50

100

150

200

250

300

350

400

450

In CB-No DIO

In CB-3% DIO

In DCB-No DIO

In DCB-3% DIO

Bulk Heterojunction

Interface junction

100 1k 10k 100k 1M0

100

200

300

400 DCB-No DIOCB-No DIODCB-3% DIOCB-3% DIO

100 1k 10k 100k 1M0

100200300400500600700800900

DCB-No DIOCB-No DIODCB-3% DIOCB-3% DIO

Electrodes and wires

Frequency (Hz)

Frequency (Hz)

- Rea

ctan

ce, -

Z” (Ω

)

- Rea

ctan

ce, -

Z” (Ω

)

Resistance, Z’ (Ω)

Res

ista

nce,

Z’ (

Ω)

Fig. 3. (a) The equivalent circuit model of the devices and the (b) resistance–frequency, (c) reactance–frequency, and (d) Cole–Cole plots of the devices in differentconditions.

Table 1The fitted parameter of each element in the equivalent model of PBDTTT-C: PC70BM OPV devices with and without DIO in DCB and CB.

R1(Ω) C1(F) R2(Ω) C2(F) R3(Ω) τ(s) PCE(%) Error (%)

In DCB-No DIO 671.65 9.37E-09 109.46 9.33E-09 13.42 6.29E-06 4.20 1.6–7.4In CB-No DIO 773.89 7.69E-09 100.6 9.85E-09 26.42 5.95E-06 3.08 1.4–4.9In DCB-3% DIO 542.37 1.34E-08 58.62 1.23E-08 17.01 7.27E-06 6.21 1.8–6.1In CB-3% DIO 526.46 1.40E-08 70.69 8.61E-09 15.75 7.37E-06 6.14 3.1–7.3

E.-P. Yao et al. / Solar Energy Materials & Solar Cells 130 (2014) 20–26 23

resistance of electrodes, since the resistance does not decrease tozero even at the high frequency (1 MHz).

Fitting the Cole–Cole plot of Fig. 3(d) into the equivalent modelgives us the fitted value of each element listed in Table 1. Theresistance (R1) of the BHJ decreases when blending DIO into CBand DCB, however, the capacitance (C1) of the BHJ increasessimultaneously. To look into the variation of R1 and C1 of theequivalent circuit, assuming that each interface between PBDTTT-Cchain and PC70BMmolecules could be regarded as a unit of a shuntpair with a resistance (Ru) and a capacitance (Cu) as shown inFig. 4. The R1 and C1 represent the combination of all the shuntpairs of the Rus and Cus in the interfaces between PBDTTT-C chainand PC70BM molecules. The following equations are the basicdefinition of the capacitance and resistance

C ¼ εAd

ð1Þ

R¼ ρLA

ð2Þ

where C is the capacitance, R is the resistance, εis the permittivity, d isthe thickness, ρ is the resistivity, L is the length (or the thickness), andA of each equation is the area. In the BHJs, the holes and the electronsmainly transport through PBDTTT-C chain and PC70BM molecules,respectively. Following the equations, it indicates that the interfacearea between PBDTTT-C chain and PC70BM molecules increases byblending DIO [22], the capacitance will increase and the resistance willdecrease as the data shown in Table 1.

3.4. Carrier transition time calculation

The capacitive time constant of the equivalent circuit is directlyrelated to the carrier transition time in the following equation:

C ¼ τavgR

ð3Þ

where τavg is the average of the carrier transition time. As the BHJcan be described as a shunt pair of a resistance and a capacitancein the equivalent circuit, the average carrier transition time atshunt circuit condition is equal to recombination time of chargecarriers, i.e. carrier lifetime. Table 1 shows the τavg of the BHJ layer

and the PCE of OPV devices. The longer τavg there is, the lower therecombination is, and the better chance that carriers could reachthe electrodes. It is consistent with the improved PCE as clearlyshown in the J–V characteristic in the light state in Fig. 1. Theresistance decreases dramatically upon adding DIO into the majorsolvents, suggesting depressed charge recombination in the bulkfilm. Thus low charge recombination rate prolongs the carrierlifetime.

In previous discussion, we see that the additive solventdecreases the domain size and increase the effective donor/acceptor interface. It is commonly believed that recombinationoccurs at the effective interface, and more interfaces result instronger recombination. Nevertheless longer carrier lifetime andlower recombination are confirmed from both IS and J–V measure-ment. We thus explain the discrepancy as part of effects of DIOadditive. It has been realized that DIO is a good solvent for PCBM,and poor for PBDTTT-C. Therefore during the film drying process,DIO of high boiling point tends to stay in the film even after themajor solvent is gone, and it then develops relatively pure polymerdomain first, before precipitation of PCBM. In this case, comparedwith single solvent solution, the phase separation of polymer andfullerene is more complete. The incomplete phase separation fromsingle solvent coating results in strong recombination loss, whileDIO additive effectively depresses the recombination by increasingpurity in individual polymer and PCBM domains [49].

3.5. Thickness difference to IS

To make further confirmation of the simulate model, the devices indifferent thickness of PBDTTT-C:PC70BM BHJ deposited from DCB withDIO are analyzed by impedance spectroscopy as shown in Fig. 5(a). Tokeep the morphology comparable, we tuned solution concentration toadjust film thickness, while keeping spin-coating condition the same[50]. The Cole–Cole plot of the device was larger since the PBDTTT-C:PC70BM layer became thicker. Basically, the enlargement of the Cole–Cole plot can be related to both of the shunt pairs in the equivalentcircuit. However, it is easy to observe that the enlargement of theCole–Cole plot is owing to the increase of the radius of the dominantsemicircle corresponding to the shunt pair of the BHJ since the shapeof the smaller semicircle is still not obvious. Fig. 5(b) shows thevariation of the resistance part and the capacitance part of BHJ (R1 andC1) with varied thickness of PBDTTT–C:PC70BM layer, and theirsimulated data are listed in Table 2. Since the PBDTTT-C:PC70BM layerbecomes thicker, the R1 increases and the C1 decreases. The result canbe explained by Eq. (1) and Eq. (2) where the thickness stands for thed and the L. However, the resistance part (R2) of the interfacesbetween PEDOT:PSS/PBDTTT-C:PC70BM and PBDTTT-C:PC70BM/Cadoes not change dramatically (�60Ω), neither the capacitance part(C2) (�12 nF). This is an evidence supporting that in the simulatecircuit model the interfaces remain unchanged, while changing thethickness of the PBDTTT-C:PC70BM layer. Besides, the thickness of thephotoactive layer is very critical to the performance of the device withthe consideration of not only the photon absorption but also thecarrier diffusion length, which relates to the lifetime of the carriers.Since the R1 and C1 vary with the thickness of PBDTTT-C:PC70BMlayer, the thickness with the longer average carrier transition time, aslisted in Table 2, directly corresponds to the better performance of thedevice. Therefore, the impedance spectroscopy can be a method toderive the best thickness of the active layer in OPV devices.

3.6. Additive concentration difference to IS

Blending DIO can lead to more favorable morphology of thePBDTTT-C:PC70BM BHJ due to decreasing the size of PC70BM andincreasing the interfaces area between the PBDTTT-C chains andPC70BM molecules, and the morphology will be different even

Fig. 4. The corresponding equivalent circuit of the interface between PBDTTT-C andPC70BM.

E.-P. Yao et al. / Solar Energy Materials & Solar Cells 130 (2014) 20–2624

varying the DIO concentration slightly [51]. Fig. 5(c) shows theCole–Cole plot of PBDTTT-C:PC70BM-based OPV devices withdifferent concentration of DIO as 1%, 3%, and 5% in DCB solutionwith the same thickness of PBDTTT-C:PC70BM layer (�65 nm). Thecurve becomes smaller with higher concentration of DIO, and thedifference mostly comes from the larger semicircle, which stoodfor the PBDTTT-C:PC70BM layer. By fitting the Cole–Cole plot withthe simulate circuit model, the variation of R1 and C1 were shownin Fig. 5(d). The R1 decreases since the concentration of DIOincreases, and the C1 increases simultaneously. As the result, themorphology of PBDTTT-C:PC70BM layer exactly got changed withdifferent concentration of DIO. Without the variation of thethickness of the PBDTTT-C:PC70BM layer, the only varied factorwould be the interface area between PBDTTT-C chains and PC70BMmolecules as mentioned above. According to the inference, theinterface area between PBDTTT-C chains and PC70BM moleculesshould be increased so that the resistance decreases and thecapacitance increases corresponding to Eq. (1) and Eq. (2) follow-ing result of Fig. 5(d). Hence, the interface between PBDTTT-C andPC70BM increased with more concentration reasonably and the

result can be confirmed according to an easy way with the simpleequivalent model. Consequently, we can derive the macroscopicview of the interaction between the donor and the acceptor withdifferent morphologies of the BHJ in OPV by fitting the impedancespectroscopy.

4. Conclusion

In this work, we applied impedance spectroscopy on DIO-doped system of OPV to discuss the correlation of the morphologyand the equivalent circuit of the PBDTTT-C:PC70BM BHJ. Theresistance part of the shunt circuit corresponding to the BHJdecreases since the DIO was added to CB and DCB solution, alongwith the increasing in capacitance simultaneously. The resultcould be explained by the increment of molecular interface areabetween the PBDTTT-C and PC70BM due to the domain size of thePC70BM becoming smaller with the addition of DIO. As the result,IS could be an effective way to determine the morphology of theBHJ by the electrical method. Besides, the simulated circuit model

0 100 200 300 400 500 600 700 8000

50

100

150

200

250

300

350

160nm 120nm 90nm 75nm 50nm

40 60 80 100 120 140 160 180

8.00E-009

1.00E-008

1.20E-008

1.40E-008

1.60E-008

1.80E-008

350

400

450

500

550

600

650

700

750

Thickness (nm)

C1 R1

)

Res

ista

nce

()

Res

ista

nce

()

Capacitance (F)

0 100 200 300 400 500 600 7000

50

100

150

200

250

300

1% DIO 3% DIO 5% DIO

0 1 2 3 4 58.00E-009

1.00E-008

1.20E-008

1.40E-008

1.60E-008

1.80E-008

420440460480500520540560580600620

Concentration of DIO (%)

C1 R1

-Rea

ctan

ce, -

Z)

-Rea

ctan

ce, -

Z)

Capacitance (F)

)Fig. 5. The Cole–Cole plots and the variation of R1 and C1 of the devices with (a), (b) different thickness and (c), (d) different concentrations of DIO of PBDTTT-C:PC70BMlayer respectively.

Table 2The values R1, R2, C1, C2, lifetime, and the power conversion efficiency of PBDTTT-C:PC70BM OPV devices in different thickness of PBDTTT-C:PC70BM layer.

Thickness of activelayer (nm)

R1(Ω) C1(F) R2(Ω) C2(F) τ1(s) PCE(%) Error (%)

160 722.47 8.52E-09 58.76 1.12E-08 6.16E-06 3.76 2.7–3.9120 627.56 1.03E-08 60.69 1.24E-08 6.46E-06 4.45 3.2–7.990 542.37 1.34E-08 58.62 1.23E-08 7.27E-06 6.21 1.8–6.175 468.03 1.48E-08 59.28 1.37E-08 6.92E-06 5.42 1.2–5.750 366.29 1.78E-08 58.82 1.38E-08 6.51E-06 4.70 2.8–6.9

E.-P. Yao et al. / Solar Energy Materials & Solar Cells 130 (2014) 20–26 25

was verified by comparing the variation of active layer thickness tothe value of the elements in the simulate circuit. Moreover, thepower conversion efficiency of the device increases while relaxa-tion time derived by the simulate circuit increases. By using the ISto analyze OPV devices, we can obtain deeper electrical informa-tion of the devices of investigating more details of the BHJ.

References

[1] L. Blankenburg, K. Schultheis, H. Schache, S. Sensfuss, M. Schrödner, Reel-to-reel wet coating as an efficient up-scaling technique for the production ofbulk-heterojunction polymer solar cells, Sol. Energy Mater. Sol. Cells 93 (2009)476.

[2] C.W. Tang, Two-layer organic photovoltaic cell, Appl. Phys. Lett. 48 (1986) 183.[3] G. Li, V. Shrotriya, J.S. Huang, Y. Yao, T. Moriarty, K. Emery, Y. Yang, High-

efficiency solution processable polymer photovoltaic cells by self-organizationof polymer blends, Nat. Mater. 4 (2005) 864–868.

[4] W. Ma, C. Yang, X. Gong, K. Lee, A. Heeger, Thermally stable, efficient polymersolar cells with nanoscale control of the interpenetrating network morphol-ogy, Adv. Funct. Mater. 15 (2005) 1617–1622.

[5] G. Zhao, Y. He, Y. Li, 6.5% efficiency of polymer solar cells based on poly(3-hexylthiophene) and indene-C60 bisadduct by device optimization, Adv. Mater.22 (2010) 4355–4358.

[6] J.-L. Wu, F.-C. Chen, M.-K. Chuang, K.-S. Tan, Near-infrared laser-drivenpolymer photovoltaic devices and their biomedical applications,, EnergyEnviron. Sci. 4 (2011) 3374.

[7] T.-Y. Chu, S. Alem, P.G. Verly, S. Wakim, J. Lu, Y. Tao, S. Beaupré, M. Leclerc,F. Bélanger, D. Désilets, S. Rodman, D. Waller, R. Gaudiana, Highly efficientpolycarbazole-based organic photovoltaic devices, Appl. Phys. Lett. 95 (2009)063304.

[8] G. Namkoong, J. Kong, M. Samson, I.-W. Hwang, K. Lee, Active layer thicknesseffect on the recombination process of PCDTBT:PC71BM organic solar cells,Org. Electron. 14 (2013) 74.

[9] L. Dou, J. You, Z. Hong, Z. Xu, G. Li, R.A. Street, Yang Yang, 25th anniversaryarticle: a decade of organic/polymeric photovoltaic research, Adv. Mater. 25(2013) 6642–6671.

[10] Gang Li, rui Zhu, Yang Yang, Polymer solar cells, Nat. Photon. 6 (2012) 180–185.[11] Z.C. He, C.M. Zhong, S.J. Su, M. Xu, H.B. Wu, Y. Cao, Enhanced power-conversion

efficiency in polymer solar cells using an inverted device structure, Nat.Photon. 6 (2012) 591.

[12] X. Xia, G. Wei, S. Wang, J.D. Zimmerman, C.K. Renshaw, M.E. Thompson,S.R. Forrest, Small-molecule photovoltaics based on functionalized squarainedonor blends, Adv. Mater. 24 (2012) 1956.

[13] Y.-H. Chen, L.-Y. Lin, C.-W. Lu, F. Lin, Z.-Y. Huang, H.-W. Lin, P.-H. Wang,Y.-H. Liu, K.-T. Wong, J. Wen, D.J. Miller, S.B. Darling, Vacuum-deposited small-molecule organic solar cells with high power conversion efficiencies byjudicious molecular design and device optimization, J. Am. Chem. Soc. 134(2012) 13616.

[14] R. Fitzner, E. Mena-Osteritz, A. Mishra, G. Schulz, E. Reinold, M. Weil, C. Körner,H. Ziehlke, C. Elschner, K. Leo, M. Riede, M. Pfeiffer, C. Uhrich, P. Bäuerle,Correlation of π-conjugated oligomer structure with film morphology andorganic solar cell performance, J. Am. Chem. Soc. 134 (2012) 11064.

[15] L. Dou, J. You, J. Yang, C.-C. Chen, Y. He, S. Murase, T. Moriarty, K. Emery, G. Li,Y. Yang, Tandem polymer solar cells featuring a spectrally matched low-bandgap polymer, Nat. Photon. 6 (2012) 180.

[16] J.B. You, L.T. Dou, K. Yoshimura, T. Kato, K. Ohya, T. Moriarty, K. Emery, C.-C. Chen, J. Gao, G. Li, Y. Yang, A polymer tandem solar cell with 10.6% powerconversion efficiency, Nat. Commun. 4 (2013) 1446.

[17] ⟨http://www.heliatek.com/⟩ (accessed Mar 2013).[18] J. Xue, S. Uchida, B.P. Rand, S.R. Forrest, Asymmetric tandem organic photo-

voltaic cells with hybrid planar-mixed molecular heterojunctions, Appl. Phys.Lett. 85 (2004) 5757.

[19] M. Riede, C. Uhrich, J. Widmer, R. Timmreck, D. Wynands, G. Schwartz,W.-M. Gnehr, D. Hildebrandt, A. Weiss, J. Hwang, S. Sundarraj, P. Erk,M. Pfeiffer, K. Leo, Efficient organic tandem solar cells based on smallmolecules, Adv. Funct. Mater. 21 (2011) 3019.

[20] G. Li, Y. Yao, H. Yang, V. Shrotriya, G. Yang, Y. Yang, Solvent annealing effect inpolymer solar cells based on poly(3-hexylthiophene) and methanofullerenes,Adv. Funct. Mater. 17 (2007) 1636.

[21] T. Salim, L.H. Wong, B. Bräuer, R. Kukrej, Y.L. Foo, Z. Baod, Y.M. Lam, Solventadditives and their effects on blend morphologies of bulk heterojunctions, J.Mater. Chem. 21 (2011) 242.

[22] Y. Liang, Z. Xu, J. Xia, S.-T. Tsai, Y. Wu, G. Li, C. Ray, L. Yu, For the bright future—bulk heterojunction polymer solar cells with power conversion efficiency of7.4%, Adv. Mater. 22 (2010) E135.

[23] S.J. Lou, J.M. Szarko, T. Xu, L. Yu, T.J. Marks, L.X. Chen, Effects of additives on themorphology of solution phase aggregates formed by active layer componentsof high-efficiency organic solar cells, J. Am. Chem. Soc. 133 (2011) 20661.

[24] A.J. Moulé, K. Meerholz, Controlling morphology in polymer–fullerene mix-tures, Adv. Mater. 20 (2008) 240.

[25] F.-C. Chen, H.-C. Tseng, C.-J. Ko, Solvent mixtures for improving deviceefficiency of polymer photovoltaic devices, Appl. Phys. Lett. 92 (2008) 103316.

[26] J. Zhou, X. Wan, Y. Liu, Y. Zuo, Z. Li, G. He, G. Long, W. Ni, C. Li, X. Su, Y. Chen,Small molecules based on benzo[1,2-b:4,5-b0]dithiophene unit for high-performance solution-processed organic solar cells, J. Am. Chem. Soc. 134(2012) 16345.

[27] X. Yang, J. Loos, S.C. Veenstra, W.J.H. Verhees, M.M. Wienk, J.M. Kroon, M.A.J. Michels, R.A.J. Janssen, Nanoscale morphology of high-performance polymersolar cells, Nano Lett. 5 (2005) 579.

[28] S.S. van Bavel, M. Bärenklau, G. de With, H. Hoppe, J. Loos, P3HT/PCBM bulkheterojunction solar cells: impact of blend composition and 3D morphologyon device performance, Adv. Funct. Mater. 20 (2010) 1458.

[29] M. Dante, J. Peet, T.-Q. Nguyen, Nanoscale charge transport and internalstructure of bulk heterojunction conjugated polymer/fullerene solar cells byscanning probe microscopy, J. Phys. Chem. C 112 (2008) 7241.

[30] H. Hoppe, M. Niggemann, C. Winder, J. Kraut, R. Hiesgen, A. Hinsch,D. Meissner, N.S. Sariciftci, Nanoscale morphology of conjugated polymer/fullerene-based bulk- heterojunction solar cells, Adv. Funct. Mater. 14 (2004)1005.

[31] V. Palermo, M. Palma, P. Samorì, Electronic characterization of organic thinfilms by Kelvin probe force microscopy, Adv. Mater. 18 (2006) 145.

[32] Y. Kim, S. Cook, S.M. Tuladhar, S.A. Choulis, J. Nelson, J.R. Durrant, D.D.C. Bradley, M. Giles, I. McCulloch, C.-S. Ha, M. Ree, A strong regioregularityeffect in self-organizing conjugated polymer films and high-efficiency poly-thiophene:fullerene solar cells, Nat. Mater. 5 (2006) 197.

[33] G. Dennler, A.J. Mozer, G. Juska, A. Pivrikas, R. Osterbacka, A. Fuchsbauer,N.S. Sariciftci, Charge carrier mobility and lifetime versus composition ofconjugated polymer/fullerene bulk-heterojunction solar cells, Org. Electron. 7(2006) 229.

[34] C. Arndt, U. Zhokhavets, M. Mohr, G. Gobsch, M. Al-Ibrahim, S. Sensfuss,Determination of polaron lifetime and mobility in a polymer/fullerene solarcell by means of photoinduced absorption, Synth. Met. 147 (2004) 257.

[35] H. Ohkita, S. Cook, Y. Astuti, W. Duffy, S. Tierney, W. Zhang, M. Heeney,I. McCulloch, J. Nelson, D.D.C. Bradley, J.R. Durrant, Charge carrier formation inpolythiophene/fullerene blend films studied by transient absorption spectro-scopy, J. Am. Chem. Soc. 130 (2008) 3030.

[36] K.H. Chan, S.W. Tsang, K.H. Lee, F. So, S.K. So, Charge injection and transportstudies of poly(2,7-carbazole) copolymer PCDTBT and their relationship tosolar cell performance, Org. Electron. 13 (2012) 850.

[37] J. Peet, M.L. Senatore, A.J. Heeger, G.C. Bazan, The role of processing in thefabrication and optimization of plastic solar cells, Adv. Mater. 21 (2009) 1521.

[38] L.-M. Chen, Z. Xu, Z. Hong, Y. Yang, Interface investigation and engineering-achieving high performance polymer photovoltaic devices, J. Mater. Chem. 20(2010) 2575.

[39] G. Sliauzys, G. Juska, K. Arlauskas, A. Pivrikas, R. Osterbacka, M. Scharber,A. Mozer, N.S. Sariciftci, Recombination of photogenerated and injected chargecarriers in π-conjugated polymer/fullerene blends, Thin Solid Films 511 (2006)224.

[40] F. Fabregat-Santiago, J. Bisquert, G. Garcia-Belmonte, G. Boschloo, A. Hagfeldt,Influence of electrolyte in transport and recombination in dye-sensitized solarcells studied by impedance spectroscopy, Sol. Energy Mater. Sol. Cells 87(2005) 117.

[41] C.-C. Chen, B.-C. Huang, M.-S. Lin, Y.-J. Lu, T.-Y. Cho, C.-H. Chang, K.-C. Tien,S.-H. Liu, T.-H. Ke, C.-C. Wu, Impedance spectroscopy and equivalent circuits ofconductively doped organic hole-transport materials, Org. Electron. 11 (2010)1901.

[42] V.S. Reddy, S. Das, S.K. Ray, A. Dhar, Studies on conduction mechanisms ofpentacene based diodes using impedance spectroscopy, J. Phys. D: Appl. Phys.40 (2007) 7687.

[43] B.J. Leever, C.A. Bailey, T.J. Marks, M.C. Hersam, M.F. Durstock, in situcharacterization of lifetime and morphology in operating bulk heterojunctionorganic photovoltaic devices by impedance spectroscopy, Adv. Energy Mater. 2(2012) 120.

[44] G. Garcia-Belmonte, A. Guerrero, J. Bisquert, Elucidating operating modes ofbulk-heterojunction solar cells from impedance spectroscopy analysis, J. Phys.Chem. Lett. 4 (2013) 877.

[45] Y. Liang, D. Feng, Y. Wu, S.-T. Tsai, G. Li, C. Ray, L. Yu, Highly efficient solar cellpolymers developed via fine-tuning of structural and electronic properties, J.Am. Chem. Soc. 131 (2009) 7792.

[46] S.A. Hawks, F. Deledalle, J. Yao, D.G. Rebois, G. Li, J. Nelson, Y. Yang, T. Kirchartz,J.R. Durrant, Relating recombination, density of states, and device perfor-mance in an efficient polymer:fullerene organic solar cell blend, Adv. EnergyMater. 3 (2013) 1201.

[47] K. Hess, Advanced Theory of Semiconductor Devices, Wiley, New York, 2000.[48] J. Bisquert, Chemical capacitance of nanostructured semiconductors: its origin

and significance for nanocomposite solar cells, Phys. Chem. Chem. Phys. 5(2003) 5360.

[49] Y. Yao, J. Hou, Z. Xu, G. Li, Y. Yang, Effects of solvent mixtures on the nanoscalephase separation in polymer solar cells, Adv. Funct. Mater. 18 (2008) 1783.

[50] C.-W. Chu, H. Yang, W.-J. Hou, J.S. Huang, G. Li, Y. Yang, Control of thenanoscale crystallinity and phase separation in polymer solar cells, Appl. Phys.Lett. 92 (2008) 103306.

[51] G. Ren, E. Ahmed, S.A. Jenekhe, Non-fullerene acceptor-based bulk hetero-junction polymer solar cells: engineering the nanomorphology via processingadditives, Adv. Energy Mater. 1 (2011) 946.

E.-P. Yao et al. / Solar Energy Materials & Solar Cells 130 (2014) 20–2626