Polaron Recombination in Pristine and Annealed Bulk Heterojunction Solar Cells
PPolaron Recombination in Pristine and Annealed Bulk Heterojunction Solar Cells
C. Deibel, a A. Baumann, and A. Wagenpfahl
Experimental Physics VI, Julius-Maximilians-University of W¨urzburg, D-97074 W¨urzburg
V. Dyakonov
Experimental Physics VI, Julius-Maximilians-University of W¨urzburg, D-97074 W¨urzburg andFunctional Materials for Energy Technology, Bavarian Centre for Applied Energy Research (ZAE Bayern), D-97074 W¨urzburg (Dated: October 25, 2018)The major loss mechanism of photogenerated polarons was investigated in P3HT:PCBM solar cells by thephoto-CELIV technique. For pristine and annealed devices, we find that the experimental data can be explainedby a bimolecular recombination rate reduced by a factor of about ten (pristine) and 25 (annealed) as comparedto Langevin theory. Aided by a macroscopic device model, we discuss the implications of the lowered loss rateon the characteristics of polymer:fullerene solar cells.
PACS numbers: 71.23.An, 72.20.Jv, 72.80.Le, 73.50.Pz, 73.63.BdKeywords: organic semiconductors; polymers; photovoltaic effect; charge carrier recombination
Organic bulk heterojunction (BHJ) solar cells have shownan increasing performance in the recent year, and also scien-tific progress concerning the fundamental understanding hasbeen made . However, the dominant loss mechanism of thephotocurrent is still under discussion. In polymer:fullerenesolar cells, usually bimolecular recombination processes areobserved by charge extraction techniques, whereas the shortcircuit current of state-of-the-art devices shows a monomolec-ular signature. Koster et al. approached an explanation of thisdiscrepancy by applying a macroscopic device model, statingthat bimolecular recombination at short circuit accounted foronly a few percent loss, therefore being latent in the current–voltage measurements. In order to contribute to this discus-sion, we apply photo-induced charge extraction by linearlyincreasing voltage (photo-CELIV) on pristine and annealedP3HT:PCBM (poly(3-hexyl thiophene):[6,6]-phenyl-C bu-tyric acid methyl ester) solar cells in order to investigate thepolaron recombination dynamics.We prepared organic bulk heterojunction solar cells byspin coating 1:1 blends of poly[3-hexyl thiophene-2,5-diyl](P3HT) with [6,6]-phenyl-C61 butyric acid methyl ester(PCBM), 20mg/ml dissolved in Chlorobenzene, on PE-DOT:PSS covered ITO/glass substrates. The active layer wasabout 105nm thick. Al anodes were thermally evaporated. Weobtained P3HT from Rieke Metals and PCBM from Solenne.The photo-CELIV method was applied on pristine and an-nealed samples. Charge carriers were generated by a shortlaser pulse (Nitrogen laser with dye unit, 5ns, 500 µ J/cm ).After a delay time at zero internal field, the remaining chargesare extracted by a voltage ramp in reverse bias. The chargecarrier mobility and the concentration of extracted charge car-riers are obtained simultaneously. Furthermore, we use amacroscopic simulation program implemented by us whichsolves the differential equation system of the Poisson, conti-nuity and drift–diffusion equations by an iterative approach,as described in Ref. . a Electronic address: [email protected] j [ x - A / c m ] -3 s]P3HT:PCBM 1:1pristine, T=180K delay time 10ms 100ms 1ms 10ms FIG. 1: Photo-CELIV spectrum of a pristine P3HT:PCBM solar cellin dependence of the delay time betwen laser pulse and extractionvoltage pulse at 180K. The second extraction peak shows the negli-gible influence of injection currents for the pristine sample.
A photo-CELIV measurement of a pristine P3HT:PCBMsolar cell is shown in Fig. 1. The evaluation of the extractedcharge concentration in dependence on the delay time gives usdirect insight into the polaron recombination dynamics. Nei-ther monomolecular nor Langevin-type bimolecular recombi-nation are able to fit the experimental data well. Instead, thedata is to be described by a reduced Langevin recombinationrate R = ζγ ( np − n i ) (1)with a prefactor ζ < ζ = ), where n and p are electron and hole con-centration, respectively, n i is the intrinsic carrier concentra-tion, and γ is the Langevin recombination parameter. The lat-ter is linearly proportional to the charge carrier mobility. Wenote that a trimolecular fit ( dn / dt ∝ n ) gave an even lowerdeviation for the 180K pristine P3HT:PCBM sample. Despite a r X i v : . [ c ond - m a t . m t r l - s c i ] J u l -80-60-40-20020406080 C u rr e n t D e n s i t y [ A / m ] ff 60 58 48 %h 2.8 2.6 2.0 % FIG. 2: Simulated current–voltage characteristics of a poly-mer:fullerene solar cell under one sun for three different values of therecombination prefactor ζ . The pronounced influence of the reducedbimolecular recombination on the field-dependent photocurrent canbe clearly seen. a very recent report with similar findings , further investiga-tions are necessary to verify this unexpected result. Therefore,and in analogy to literature and our previous results onannealed P3HT:PCBM devices only, we interpret the experi-mental data in view of reduced Langevin rates for both pristineand annealed samples. At 180K, the pristine solar cells show ζ of around 0.1, which is further reduced to ζ = .
04 for an-nealed devices. At higher temperatures, the pristine samplestays at approximately 0.1. For the annealed sample, the de-termination of the prefactors is strongly influenced by chargeinjection, increasing the error margin; within its bounds, wesee no temperature dependent variation of ζ . Previously re-ported was an even weaker Langevin recombination rate, thusa lower ζ , at higher temperatures .In order to clarify the impact of a reduced bimolecularrecombination rate on working polymer solar cells, we per-formed macroscopic simulations. The calculated current–voltage characteristics under one sun for three different valuesof ζ , 1 (Langevin), 0.1, and 0.01, are shown in Fig. 2 . Clearly,the field-dependent photocurrent is improved by a lowering ofthe bimolecular recombination rate. In order to quantify theinfluence of the latter, we use the recombination yield as ameasure. It is defined asrecombination yield = − UPG (2)where U = PG − ( − P ) R is the net generation rate, G is theexciton generation rate, and P the polaron pair dissociationyield. The recombination yield is shown in Fig. 3. We notethat the recombination term R includes photogenerated andinjected carriers, but is normalised to the polaron photogen-eration rate PG . If normal Langevin recombination ( ζ = R e c o m b i n a t i o n Y i e l d FIG. 3: Simulated bimolecular recombination yield (Eqn. 2) of poly-mer:fullerene solar cells in dependence on the recombination pre-factor ζ . C a rr i e r C o n c e n t r a t i o n [ m - ] FIG. 4: Simulated carrier concentration in the active layer of poly-mer:fullerene solar cells under open-circuit conditions. The reducedbimolecular recombination leads to a significant increase of the car-rier concentration. the losses start only at higher voltages. Thus, mainly the fillfactor and the open-circuit voltage are negatively affected bythe nongeminate recombination. The low recombination rate,however, limits its impact on the solar cell performance, sothat the charge extraction properties are governing the solarcell efficiency rather than the bimolecular recombination.The effect of the recombination yield on the carrier concen-tration under open-circuit conditions is depicted in Fig. 4. Thelower the recombination rate, the higher the steady-state car-rier concentration in the active area of the polymer:fullerenesolar cell. Actually, this difference in carrier concentra-tion is directly reflected in the solar cell characteristics: thehigher carrier concentration indicates that the quasi-Fermi lev-els move closer to their respective bands. This leaves moreroom for the open-circuit voltage, which is thus increased forraised electron and hole concentrations due to lower nongem-inate loss rates.In conclusion, by performing charge extraction experimentson pristine and annealed P3HT:PCBM solar cells, we haveshown that the bimolecular loss rate is reduced significantlyas compared to Langevin theory. The fill factor and the open-circuit voltage are reduced by the nongeminate recombina-tion, but due to the low recombination rate, the solar cellperformance is mainly determined by the charge extractionproperties. Nevertheless, the finding of a reduced Langevin recombination rate has an impact on the understanding andmodelling of polymer solar cells, as was shown by perform-ing macroscopic device simulations, and therefore needs to beaccounted for in future studies.
Acknowledgments
V.D.’s work at the ZAE Bayern is financed by the BavarianMinistry of Economic Affairs, Infrastructure, Transport andTechnology. U. Scherf, C. Brabec, and V. Dyakonov, eds.,
Organic Photo-voltaics. Materials, Device Physics, and Manufacturing Tech-nologies (Wiley VCH, 2008). A. J. Mozer, G. Dennler, N. S. Sariciftci, M. Westerling,A. Pivrikas, R. ¨Osterbacka, and G. Juˇska, Phys. Rev. B , 035217(2005). I. Riedel, J. Parisi, V. Dyakonov, L. Lutsen, D. Vanderzande, andJ. C. Hummelen, Adv. Funct. Mater. , 38 (2004). L. J. A. Koster, E. C. P. Smits, V. D. Mihailetchi, and P. W. M.Blom, Phys. Rev. B , 085205 (2005). G. Juˇska, K. Arlauskas, M. Vili¯unas, and J. Koˇcka, Phys. Rev.Lett. , 4946 (2000). C. Deibel, A. Wagenpfahl, and V. Dyakonov, phys. stat. sol. (RRL) (2008). P. Langevin, C. R. Acad. Sci. , 530 (1909), translated byD.S. Lemons and A. Gythiel, Am. J. Phys. , 1079 (1997). G. Juˇska, private communications (2008). C. Shuttle, B. O’Regan, A. Ballantyne, J. Nelson, D. Bradley,J. D. Mello, and J. Durrant, Appl. Phys. Lett. , 093311 (2008). G. Juˇska, K. Arlauskas, J. Stuchlik, and R. ¨Osterbacka, J. Non-Cryst. Sol. , 1167 (2006). A. Pivrikas, R. ¨Osterbacka, G. Juˇska, K. Arlauskas, and H. Stubb,Synth. Met. , 242 (2005). C. Deibel, A. Baumann, J. Lorrmann, and V. Dyakonov, in