Organic Electronics 9 (2008) 847–851

Contents lists available at ScienceDirect

Organic Electronics
journal homepage: www.elsevier.com/locate/orgel

Charge carrier mobility and lifetime of organic bulk heterojunctions

analyzed by impedance spectroscopy
Germà Garcia-Belmonte a,*, Antoni Munar a, Eva M. Barea a, Juan Bisquert a, Irati Ugarte b,
Roberto Pacios b
a
Departament de Física, Universitat Jaume I, 12071 Castelló, Spain
b
Department of Microsystems, IKERLAN-IK4 S. Coop., P° J.M. Arizmendiarrieta, 2, 20500 Arrasate-Mondragón, Gipuzkoa, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Charge carrier diffusion and recombination in an absorber blend of poly(3-hexylthiophene)
Received 24 April 2008 (P3HT) and [6,6]-phenyl C61-butyric acid methyl ester (PCBM) with indium tin oxide (ITO)
Received in revised form 4 June 2008 and aluminium contacts have been analyzed in the dark by means of impedance spectros-
Accepted 8 June 2008
copy. Reverse bias capacitance exhibits Mott–Schottky-like behavior indicating the forma-
Available online 17 June 2008
tion of a Schottky junction (band bending) at the P3H:PCBM-Al contact. Impedance
measurements show that minority carrier (electrons) diffuse out of the P3HT:PCBM-Al
depletion zone and their accumulation contributes to the capacitive response at forward
PACS:
73.40.Ei
bias. A diffusion–recombination impedance model accounting for the mobility and lifetime
73.40.Lq parameters is outlined. Electron mobility results to be 2  103 cm2 V1 s1 and lifetime
73.61.Wp lies within the milliseconds timescale.
84.60.Jt Ó 2008 Elsevier B.V. All rights reserved.

Keywords:
P3HT:PCBM heterojunctions
Organic solar cells
Capacitance
Impedance spectroscopy
Mobility
Lifetime

1. Introduction When using derivatives of poly(phenylene vinylene)

(PPV) as hole transporting material (HTM), some aspects
Organic conducting materials are at the heart of bulk of the BHJ device operation (particularly the internal elec-
heterojunction (BHJ) solar cells, a promising alternative tric field distribution [2] and the transport mechanisms
to silicon-based solar cells due to their optical, electronic [3,4]) have been properly interpreted. Blends formed of
and mechanical properties [1]. Bulk heterojunctions are PPV derivatives and fullerenes are considered to form un-
formed by an interpenetrating blend of an optically active doped systems. Therefore, the basic device model resem-
polymer and electron accepting molecules. At the optically bles that proposed for a-Si thin film p–i–n structures [3].
active polymer excitons are created by light absorption. Transport is mainly determined by charge carrier drift
This exciton dissociates at the polymer/molecule interface, upon the electrical field driving force. Recently, BHJ organic
thus spatially separating the electrons and holes which solar cells using P3HT as HTM have shown high perfor-
therefore can be transported to the electrodes by hopping. mance [5]. P3HT exhibits relatively high hole mobility
(104–103 cm2 V1 s1) [6] and can easily undergo
p-doping when exposed to the air or moisture [7,8],
* Corresponding author. Tel.: +34 964 728040; fax: +34 964 729218.
E-mail address: [email protected] (G. Garcia-Belmonte). with the consequent formation of Schottky contacts.

1566-1199/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.orgel.2008.06.007
848 G. Garcia-Belmonte et al. / Organic Electronics 9 (2008) 847–851

Models which regard polymer semiconductors as undoped quent ultrasonic baths of acetone, methanol and isopropa-
materials should be then revised to include the effect of nol. They were later introduced in oxygen plasma for
band bending (depletion zones) and minority carrier injec- 5 min. Baytron P CH8000 from HC Starck was spin-coated
tion and storage in the diode bulk. on top of the ITO substrate giving a film thickness of
The central magnitude that informs about carrier accu- 50 nm. The resulting film was annealed at 120 °C for
mulation is the capacitance, which can be readily deter- 5 min in order to remove any possible water residual on
mined by impedance spectroscopy. This is because in all the film. The P3HT:PCBM film was then spin-cast from
solar cells the generation of positive and negative carriers the previously prepared solution at 800 rpm. The active
creates a splitting of Fermi levels that is ultimately respon- layer thickness was measured to be L  200 nm. The film
sible for the photovoltage [9]. In silicon solar cells, for in- was pre-annealed at 85 °C for 5 min to remove any solvent
stance, the capacitance shows two main components as a remaining on the film after deposition. Samples were then
function of the bias [10,11]: a Mott–Schottky characteris- taken to a vacuum chamber to thermally evaporate the
tic, due to the modulation of the Schottky barrier, at re- aluminium contacts through a shadow mask. The evapora-
verse and moderate forward bias, and a chemical tion was done at 3  106 atm at rates between 2 and
capacitance [9], that increases exponentially for intense 10 Å/s. The active area of the diode, given by the overlap
forward bias. The first characteristic indicates the presence between ITO and aluminium, is 9 mm2. The finalized de-
of doping whereby the solar cell device is able to accumu- vices were immediately encapsulated with an epoxy resin
late substantial minorities. Such carrier storage is manifest
in the second characteristic, the chemical capacitance,
which directly reflects the carrier statistics [9]. In standard
dye-sensitized solar cells (DSC) based on nanostructured
TiO2, the shape of the voltage dependence of the capaci-
tance is somewhat different, since the ‘‘hole conductor” is
a liquid electrolyte with high ionic concentration, so that
the Schottky barrier in the active semiconductor layer is
not found. However, the chemical capacitance is very
clearly observed and shows the density of states of elec-
trons accumulated in the electron-transporting material
(TiO2) [12,13]. The identification of the voltage-dependent
capacitance has become then a major tool for assessing the
energetics in a DSC [14] and for interpreting the recombi-
nation lifetime [15]. The strong accumulation associated
with the unconstrained rise of the Fermi level with bias ap-
pears in high performance solar cells [13], while in many
other cases, the charge storage is inhibited by additional
mechanisms and the solar cell capacitance may become
negative at strong forward bias [10].
Given the general significance of the capacitance for
interpretation of the fundamental transport and recombi-
nation mechanism governing the operation of new classes
of solar cell devices, in this paper we present direct mea-
surements and interpretation of the capacitance under re-
verse and forward voltages of a BHJ structure of the type
ITO/PEDOT:PSS/P3HT:PCBM/Al. Reverse bias capacitance
exhibits Mott–Schottky-like behavior signaling the
formation of a Schottky junction (band bending) at the
P3HT:PCBM-Al contact. Impedance modeling allows to
extract both the recombination time and mobility of
the minority carriers (electrons) at forward bias in the
Fig. 1. (a) Band structure of the P3HT:PCBM heterojunction in equilib-
dark. rium (V = 0). Band bending appears near the cathode and holes can
occupy HOMO states of the P3HT within the neutral region. (b) Main
dynamic processes occurring in the diode bulk layer for V > Vbi. Holes
2. Experimental (majority carriers) are able to occupy the whole active layer. Electrons
(minority carriers) are injected (1) into the LUMO of the PCBM, diffuse
along the diode bulk (2), and eventually recombine with holes (3). (c)
We have built diodes with structure ITO/PEDOT:PSS/
Equivalent circuit accounting for the diffusion–recombination mecha-
P3HT:PCBM/Al following the next procedure: RR-P3HT nism used for fitting. Modulation of stored excess minority carriers gives
from Rieke Materials and PCBM from Nano-c were dis- rise to distributed chemical capacitances cn. Bimolecular recombination
solved in xylene in a weight ratio of (1:0.75). The solution of conduction band electrons and valence band holes are modeled by
was heat up to 65 °C and continuously stirred for 3 h. In resistive elements rrec. Transport of electrons is represented by means of
rt. An additional series resistance is needed to model contact and wire
parallel, ITO substrates from Diamond Coatings with a effects, Rs. Finally a capacitor Cg  ee0A/L represents dielectric contribu-
sheet resistance of 40 X/h were cleaned in 5 min subse- tion of the diode.
 G. Garcia-Belmonte et al. / Organic Electronics 9 (2008) 847–851 849

and glass. Even though encapsulation took place right after 2ðV bi  VÞ
C 2 ¼ ð1Þ
evaporation, it has to be stressed that devices were briefly A2 eee0 NA
in contact with air. Nevertheless, repetitive measurements
where Vbi is the built-in potential, V is the applied voltage,
of dark J–V characteristics (inset of Fig. 2a) gave identical
A corresponds to the device active surface (9 mm2), e ac-
results over several weeks. The impedance measurements
counts for the elementary charge, e is the relative dielectric
were carried out in the air and room temperature, using
constant, e0 the permittivity of the vacuum, and NA the
an Autolab PGSTAT-30 equipped with a frequency analyzer
concentration of acceptor impurities. Our experimental re-
module in the frequency range between 1 MHz and 1 Hz.
sults follow Eq. (1) (Mott–Schottky curve) which exhibits a
Ac oscillating amplitude was as low as 10 mV (rms) in or-
straight line over a wide bias voltage range (Fig. 2a) yield-
der to maintain the linearity of the response.
ing Vbi = 0.43 V and NA = 3.5  1016 cm3, assuming e = 3
for P3HT:PCBM. In this case the depletion region for a re-
3. Results and discussion
verse bias voltage between 0 and 1 V results to be varying
between 63 and 115 nm, compatible with an active layer
It is well-known that P3HT is a conjugated polymer that
thickness of 200 nm. At V = Vbi (flat-band conditions) the
in exposure to oxygen and/or moisture results p-doped
depletion layer adjacent to the P3HT:PCBM/Al contact dis-
[7,8]. Under such conditions, it has been suggested that
appears and the neutral, doped region extends along the
the P3HT-Aluminum contact shows a Schottky diode
whole diode bulk (Fig. 1b).
behavior. Band bending with a corresponding depletion
As commented in the Introduction, when the Schottky
zone is formed at the contact as indicated in Fig. 1a
barrier vanishes at V > Vbi it is expected that the device
[16,17]. For a Schottky diode [18] the junction capacitance,
capacitance is governed by a chemical capacitance due to
which appears as a consequence of the modulation of the
the excess minority carriers Cl [11,19,20], that is related
depletion layer, exhibits a bias dependence according to
to the change of the occupancy of electrons in the LUMO
the Mott–Schottky relation
of the PCBM, as follows
dn
C l ¼ e2 ð2Þ
dEFn

Here the capacitance is given per unit volume, n corre-

sponds to the electron concentration, and EFn accounts for
the electron quasi-Fermi level (see Fig. 1b). According to
Eq. (2), the capacitance should increase as a function of for-
ward bias. For ideal statistics, the increase is exponential
[10]. However, it has been shown in classic Schottky diodes
[21] that under moderate to large forward bias conditions,
minority carrier injection can be limited depending on the
barrier heights at the metal–semiconductor junction. Oxi-
dation of the aluminium cathode (due to the preparation
conditions) leads to the formation of an Al2O3 layer be-
tween the aluminium and the P3HT:PCBM. Electron injec-
tion from aluminium to PCBM might be blocked to some
extent by this tunneling barrier, and could limit charge col-
lection under illumination This injection-limiting effect
yields a peak in the capacitance–voltage characteristics,
and eventually a strong negative capacitance component
[21].
Fig. 2b shows the measured capacitance at low fre-
quency (100 Hz) as a function of the bias. The capacitance
increases in the forward direction, reaching a peak for volt-
ages of approximately 0.6–0.7 V. At this point the capaci-
tance decreases caused by the limitation upon the
concentration of minority carriers in the neutral regions
of the device. Alternative explanations of a large low-fre-
quency capacitance at forward bias point to the occurence
of charge blocking caused by slow injection/extraction at
the contacts [22]. Although these effects cannot be com-
pletely discarded, such interpretations are however based
Fig. 2. (a) Mott–Schottky curve (100 Hz) which exhibits a straight line on increasing capacitance values at forward bias in opposi-
yielding Vbi = 0.43 V and NA = 3.5  1016 cm3, assuming e = 3 for tion to our observations in Fig. 2.
P3HT:PCBM. In the inset: J–V characteristics measured in the dark. (b)
Fig. 3 displays the measured electrical impedance in the
Capacitance (100 Hz) as a function of bias. Vertical line separates voltage
regions for which capacitance is determined either by the depletion layer
complex plane for different forward bias voltages from 0.5
modulation (V < Vbi) or the storage of minority carriers (V > Vbi). to 1 V. It can be seen that the impedance spectra are com-
850 G. Garcia-Belmonte et al. / Organic Electronics 9 (2008) 847–851

Fig. 4. Electron mobility determined from fitting parameters, using Eq.

Fig. 3. Impedance spectra measured in the dark at different bias as
(3) and the approximation ln = eDn/kBT. Electron recombination time
indicated. Experimental points and fitting results (solid line) using the
(effective lifetime) resulting from fits.
equivalent circuit model of Fig. 1c. Characteristic times are marked. Inset:
detail of the high-frequency interval which signals the transition between
diffusion and recombination responses. mobility extracted from fits exhibits a nearly constant va-
lue approximately equal to 2  103 cm2 V1 s1. This va-
lue is in good agreement with that derived using PCBM
posed of a large semicircle at low frequencies together electron-only devices from J–V measurements, for which
with an almost straight line in the high-frequency part (in- current is considered space-charge-limited, and hence
set of Fig. 3), which is rather independent of the applied electrical field-driven rather than diffusion-determined
voltage. This type of impedance pattern belongs to the re- [27]. Electron effective lifetime results within the time
sponses usually encountered in systems in which carrier interval 0.3–0.1 ms, decreasing as the bias increases. Such
transport is determined by diffusion–recombination be- time presumably refers to losses produced by electron–
tween non-absorbing contacts [23]. Injected minority car- hole bimolecular recombination along the absorber layer,
riers (electrons) from the Al contact are able to diffuse although surface recombination routes cannot be ex-
within the bulk neutral region eventually reaching the cluded. Alternative methods to investigate charge carrier
PEDOT:PSS film which acts as selective contact. Transmis- lifetime such as photoinduced absorption [28,29] find val-
sion line models able to represent such diffusion–recombi- ues of similar order.
nation impedance response are well-known [23,24], and
have been integrated in standard software. The impedance 4. Conclusion
model (Fig. 1c) consists of an equivalent circuit which
comprises distributed resistors rt, standing for the electron In summary, it has been shown that the P3HT:PCBM-Al
transport, the distributed chemical capacitance cn = Cl and contacts show a behavior consistent with a Schottky diode
rrec accounting for the electron recombination resistance. in the dark and under reverse bias. Furthermore, the capac-
An additional series resistance is needed to model contact itance values at forward bias are viewed as a strong indica-
and wire effects, Rs. Finally a capacitor Cg  ee0A/L repre- tion that minority charge carriers play a significant role.
sents dielectric contribution of the diode. The impedance With such assumption and measuring ac electrical imped-
model described contains two characteristic times related ance spectra we propose a model to interprete the diffu-
to the electron diffusion sd = rtcn (transit time), and the sion and recombination processes which allows us to
effective lifetime sn = rreccn, respectively [23]. determine both the diffusion and recombination time con-
By examining Fig. 3 one can observe that fitting curves stants. Electron mobility and effective lifetime can there-
match the impedance spectra in the whole measuring fre- fore be easily computed which permits to reach a deeper
quencies. The position of the characteristic times is marked understanding of the mechanisms governing doped BHJ
in Fig. 3. Rs  100 X, and Cg  1.0 nF in agreement with the devices.
permittivity value assumed. The electron diffusivity
(chemical diffusion coefficient) can be calculated by means Acknowledgment
of the relation [23]
We thank financial support from Ministerio de Educa-
L2
Dn ¼ ð3Þ cion y Ciencia under project HOPE CSD2007-00007 (Con-
sd solider-Ingenio 2010).
Here L = 200 nm is the thickness of the diffusion layer.
Assuming that the electron statistics slightly departs from References
dilute concentration conditions, the electron mobility
[1] C.J. Brabec, N.S. Sariciftci, C. Hummelen, Adv. Mater. 11 (2001) 15–
might be readily calculated by using the Nernst–Einstein
26.
relationship as ln = eDn/kBT (being kBT the thermal energy) [2] C. Lungenschmied, G. Dennier, H. Neugebauer, N.S. Saricifti, E.
[25,26]. Results are summarized in Fig. 4. The electron Ehrenfreund, Appl. Phys. Lett. 89 (2006) 223519.
 G. Garcia-Belmonte et al. / Organic Electronics 9 (2008) 847–851 851

[3] L.J.A. Koster, E.C.P. Smits, V.D. Mihailetchi, P.W.M. Blom, Phys. Rev. B [17] M. Glatthaar, N. Mingirulli, B. Zimmermann, T. Ziegel, R. Kern, M.
72 (2005) 085205. Niggermann, A. Hinsch, A. Gombert, Phys. Stat. Sol. (a) 202 (2005)
[4] P.W.M. Blom, V.D. Mihailetchi, L.J.A. Koster, D.E. Markov, Adv. Mater. R125–R127.
19 (2007) 1551–1566. [18] E.H. Rhoderick, R.H. Williams, Metal–Semiconductor Contacts,
[5] W. Ma, C. Yang, X. Gong, K.-S. Lee, A.J. Heeger, Adv. Funct. Mater. 15 Clarendon Press, Oxford, 1988.
(2005) 1617–1622. [19] S.M. Sze, Physics of Semiconductor Devices, John Wiley and Sons,
[6] C. Tanase, E.J. Meijer, P.W.M. Blom, D.M. de Leeuw, Phys. Rev. Lett. 91 New York, 1981.
(2003) 216601. [20] J. Bisquert, D. Cahen, G. Hodes, S. Rühle, A. Zaban, J. Phys. Chem. B
[7] M.S.A. Abdou, F.P. Orfino, Y. Son, S. Holdcroft, J. Am. Chem. Soc. 119 108 (2004) 8106–8118.
(1997) 4518. [21] M.A. Green, J. Shewchun, Solid State Electron. 16 (1973) 1141–
[8] S. Hoshino, M. Yoshida, S. Uemura, T. Kodzasa, N. Takada, T. Kamata, 1150.
K. Yase, J. Appl. Phys. 95 (2004) 5088. [22] M. Glatthaar, M. Riede, N. Keegan, K. Sylvester-Hvid, B.
[9] J. Bisquert, Phys. Chem. Chem. Phys. 5 (2003) 5360. Zimmermann, M. Niggermann, A. Hinsch, A. Gombert, Sol. Energy
[10] I. Mora-Seró, J. Bisquert, F. Fabregat-Santiago, G. Garcia-Belmonte, G. Mater. Sol. Cells 91 (2007) 390–393.
Zoppi, K. Durose, Y. Proskuryakov, I. Oja, A. Belaidi, T. Dittrich, R. [23] J. Bisquert, J. Phys. Chem. B 106 (2002) 325–333.
Tena-Zaera, A. Katty, C. Lévy-Clement, V. Barrioz, S.J.C. Irvine, Nano [24] A. Pitarch, G. Garcia-Belmonte, I. Mora-Seró, J. Bisquert, Phys. Chem.
Lett. 6 (2006) 640. Chem. Phys. 6 (2004) 2983–2988.
[11] I. Mora-Seró, Y. Luo, G. Garcia-Belmonte, J. Bisquert, D. Muñoz, C. [25] G. Garcia-Belmonte, J. Bisquert, G. Popkirov, Appl. Phys. Lett. 83
Voz, J. Puigdollers, R. Alcubilla, Sol. Energy Mater. Sol. Cells 92 (2008) (2003) 2178–2180.
505–509. [26] J. Bisquert, Phys. Chem. Chem. Phys. 10 (2008), doi:10.1039/
[12] F. Fabregat-Santiago, J. Bisquert, G. Garcia-Belmonte, G. Boschloo, A. b719943k.
Hagfeldt, Sol. Energy Mater. Sol. Cells 87 (2005) 117–131. [27] V.D. Mihailetchi, J.K.J. van Duren, P.W.M. Blom, J.C. Hummelen, R.A.J.
[13] Q. Wang, S. Ito, M. Grätzel, F. Fabregat-Santiago, I. Mora-Seró, J. Janssen, J.M. Kroon, M.T. Rispens, W.J.H. Verhees, M.M. Wienk, Adv.
Bisquert, T. Bosshoa, H. Imaic, J. Phys. Chem. B 110 (2006) 19406. Funct. Mater. 13 (2003) 43–46.
[14] Q. Wang, J.-E. Moser, M. Grätzel, J. Phys. Chem. B 109 (2005) 14945. [28] C. Arndt, U. Zhokhavets, M. Mohr, G. Gobsch, M. Al-Ibrahim, S.
[15] J. Bisquert, A. Zaban, M. Greenshtein, I. Mora-Seró, J. Am. Chem. Soc. Sensfuss, Synthetic Metals 147 (2004) 257–260.
126 (2004) 13550. [29] G. Dennler, A.J. Mozer, G. Juska, A. Pivrikas, R. Österbacka, A.
[16] G. Dennler, C. Lungenschmied, N.S. Saricifti, R. Schwödiauer, S. Fuchsbauer, N.S. Saricifti, Organic Electron. 7 (2006) 229–234.
Bauer, H. Reiss, Appl. Phys. Lett. 87 (2005) 163501.