This document is the Accepted Manuscript version of a Published Work that appeared in final form in
ACS Energy Lett. 2016, 1, 302-308 copyright © American Chemical Society after peer review and technical
editing by the publisher. To access the final edited and published work see
http://pubs.acs.org/doi/pdf/10.1021/acsenergylett.6b00162

Low Open-Circuit Voltage Loss in Solution
Processed Small Molecule Organic Solar Cells
Sachetan M. Tuladhar,* † Mohammed Azzouzi, †, ‡ Florent Delval, †, ‡ Jizhong Yao, † Anne A. Y.
Guilbert, † Thomas Kirchartz, ǁ ,ǂ Nuria F. Montcada, § Rocio Dominguez, ‼ Fernando Langa, ‼
Emilio Palomares§ and Jenny Nelson* †
†

Department of Physics and Centre for Plastic Electronics, Imperial College London, South
Kensington, London, SW7 2AZ,UK.
‡

ǁ

ǂ

École Polytechnique Universite Paris-Sanclay, 91128 Palaiseau, France
IEK5-Photovoltaics, Forschungszentrum Jülich, 52425 Jülich, Germany

Faculty of Engineering and CENIDE, University of Duisburg-Essen, Carl-Benz-Straße 199,
47057 Duisburg, Germany

§

Institute of Chemical Research of Catalonia (ICIQ), Avinguda del Països Catalans 16, 43007
Tarragona, Spain.

‼

Institute for Nanoscience, Nanotechnology and Molecular Materials (INAMOL), University of
Castilla-la Mancha, Campus de la Fabrica, 45071, Toledo, Spain.

AUTHOR INFORMATION

1

Corresponding Author
* s.tuladhar@imperial.ac.uk ; *jenny.nelson@imperial.ac.uk

ABSTRACT

We analyse the voltage losses at open circuit in solution processed, small molecule: fullerene
blend solar cells, using electroluminescence and external quantum efficiency measurements and
the reciprocity relationship between light absorption and emission. For solar cells made from
oligo-thienylenevinylene based donors and phenyl-C71 butyric acid methyl ester (PC71BM), we
find that the voltage loss due to the finite breadth of the absorption edge is remarkably small, less
than 0.01 eV in the best cases, while the voltage loss due to non-radiative recombination reaches
0.29 eV, one of the smallest values reported for an organic solar cell. As a result the open-circuit
voltage reaches around 1.0 V for an optical gap of 1.6 eV, greatly exceeding the voltage of a
high performance polymer based system with similar optical gap. We assign the remarkably
small absorption broadening loss to a low degree of energetic disorder in the small molecule
system that allows efficient charge separation at a lower driving force than in typical conjugated
polymer blends.

TOC GRAPHICS

2

energy E (eV)

1.6
1.4
1.2
1.0
0.8

Eg

qVOC,SQ

qVOC,SQ

qVOC,rad

broadened
absorption edge

qVOC

non-radiative
recombination

qVOC,rad

qVOC

SML01:PC71BM PTB7:PC71BM

Organic solar cells have attracted intense interest due to the potential for manufacture of solar
cells at low cost using solution processing and flexible substrates.1,2 Research in the field of
solution processed organic photovoltaics (OPV) has mostly focused on blends of conjugated
polymers with fullerenes due to their high efficiency that has reached 11% in multijunctions3 and
above 10% in single junctions.4-6 Recently organic solar cells based on low molecular weight
solution processed semiconductor molecules (so called “small molecule” organic solar cells)
have attracted more attention due to their increasing efficiencies, which have reached record
efficiencies of 9-10% in single junction devices7,8 and 10% in tandem structures9. The growing
interest in small-molecule optoelectronics is also stimulated by several potential advantages over
semiconducting polymers, such as a more reproducible synthesis and the relative ease of
purification,10 whilst maintaining the potential for control of blend microstructure e.g. through
use of additives.11 A further important advantage is the relatively small variation in site energies
that may be expected in a small molecule film compared to a conjugated polymer, as a result of
the smaller range of backbone conformations available to a small molecule. Such a limited
variation in hole or electron energies reduces the so-called energetic disorder of the system;
reduced energetic disorder may positively impact the rate of charge transport and the charge

3

separation yield.12 Previous studies on high performance small molecule devices indicate that
energy losses between the optical gap and the energy equivalent to the open-circuit voltage, eVoc,
can be relatively small 13, 14 and that the penalty for charge pair separation may be low14. Hence,
it is of great interest to analyse and quantify the energy loss processes in small molecule devices
in order to understand their potential performance.
We present here a study of the open-circuit voltage (Voc) loss of solar cells based on two
different small-molecule donors built around a 3,4-ethylenedioxythiophene (EDOT) group and
based on a previously published study of devices made from a family of such materials15 blended
with phenyl-C71-butyric acid methyl ester (PC71BM). One of the molecules led to a Voc of 1.0 V,
which makes these materials especially interesting for our study of the energy losses in small
molecule solar cells. We use methods presented in a recent study16 to evaluate the different
contributions to the loss in Voc. This method uses the concept of a radiative open circuit voltage
Voc,rad as previously defined17, 18 but extends the definition of Voc,rad to respect the actual light
absorption spectrum of the solar cell under study. This allows us to assign an upper radiative
limit to Voc, Voc,SQ, based on the Shockley-Queisser19 condition of a sharp absorption edge at the
optical gap Eopt, and a second limit Voc,rad which replaces the step-function with the measured
shape of the absorptance. These assignments allow the losses to absorption edge broadening and
non-radiative phenomena to be distinguished.18,20 In this work we use the external quantum
efficiency coupled with electroluminescence spectroscopy to determine the Voc,SQ and the Voc,rad
in order to characterise the losses in the open-circuit voltage. The study was extended to different
processing conditions, in order to understand the effect of processing-induced changes in film
microstructure on the energy loss. The results show that these materials have a very low energy
loss compared to polymers with similar optical gap. Moreover, we find that processing

4

treatments that are known to promote phase segregation also appear to retard non-radiative
recombination, as expected. Finally, we attempt to rationalise the low energy loss in these
materials using quantum chemical calculations of the excited state spectra of these blend
systems.
The small molecules used in our study are 2,2'-((5,5'-((1E,1'E)-(2,3-dihydrothieno[3,4b][1,4]dioxine-5,7-diyl)bis(ethene-2,1-diyl))bis(3,4-dihexylthiophene-5,2diyl))bis(methanylylidene))dimalononitrile (SML01) and

(5Z,5'Z)-5,5'-((5,5'-((1E,1'E)-(2,3-

dihydrothieno[3,4-b][1,4]dioxine-5,7-diyl)bis(ethene-2,1-diyl))bis(4-hexylthiophene-5,2diyl))bis(methanylylidene))bis(3-ethyl-2-thioxothiazolidin-4-one) (SML06), both of which are
based on a central EDOT moiety, substituted on both sides by thienylenevinylene moiety, and
terminated at each end by dicyanovinylene groups (in the case of SML01), or rhodanine groups
(for SML06)15. Replacing the dicyanovinylene groups with the rhodanine groups lead to a
smaller optical gap and higher lying HOMO15, as confirmed with time-dependent density
functional (TD-DFT) calculations (Table S1). The molecules also differ in the number of side
chains attached to the thiophenes, which is likely to influence their ability to crystallise. The
chemical structures of these small molecules are shown in Scheme 1.

C6H13

C6H13

C6H13
S
NC

NC
O

CN

C2 H5

N

S

S

O

CN
O

S

S

S

C6H13

C6H13

C6H13

O
S

S
O

smL01
SML01

N

O
S

S

C2H5

smL06
SML06

Scheme 1. Chemical structures of SML01 and SML06.

5

2

(a) SML01:PC71BM
0

Neat
SVA
PDMS
PDMS SVA

current density J/(mAcm-2)

-2
-4
-6
-8
2
0

(b) SML06:PC71BM

-2
-4
-6
-8
0.0

0.2

0.4

0.6

0.8

1.0

voltage V/V

Figure 1. Current vs. voltage (J-V) curves under illumination for (a) SML01:PC71BM solar cell
devices and (b) SML06: PC71BM solar cell devices that have been processed differently, as
follows: ‘neat’: a blend film with no additive or further processing; ‘SVA’: neat film that has
been solvent vapour annealed in CH2Cl2; ‘PDMS’: a blend film cast from a solution with 0.3
mg/ml PDMS added; ‘PDMS SVA’: PDMS containing blend that has also been solvent vapour
annealed.

6

Representative current density – voltage (J-V) curves obtained from blend devices of the
structure glass/ITO/PEDOT:PSS / SML0x:PC71BM (1:1)/Ca/Al are plotted in Figure 1. The
results are in good agreement with those in a previous report on these small molecules.15 In
accordance with the HOMO energies and optical gaps of the donor materials, SML06 results in a
higher short circuit current density, Jsc, whilst SML01 results in a higher Voc. Comparison of
different process routes shows that by exposing the film to a solvent vapour annealing (SVA)
stage (3 to 5 minutes in CH2Cl2 solvent at room temperature) or applying the SVA to a blend
film containing 0.3 mg/ml of the inert polymer additive polydimethylsiloxane (PDMS) leads to
an improvement of fill factor and consequently power conversion efficiency (PCE). In the case
of SML06:PC71BM the effect on Jsc is also substantial, as the untreated device suffers from slow
charge collection relative to charge generation, which we infer from the poor fill factor. The
impact of the processing on Voc is not large in comparison.
To characterise the difference in film microstructure due to processing, we compare
atomic force microscope (AFM) images of the film surfaces, before and after solvent vapour
annealing, in Figure S1. The SVA treated films appear to show larger domain structures,
implying better phase segregation, particularly in the case of SML06. Photoluminescence
intensity data (Figure S2) show the effect of SVA more clearly than the surface images.
Exposure of blend films to solvent vapour reduces the quenching of the donor PL signal,
indicating a growth in the size of pure SML0x domains. In the case of SML06 the domain
growth is accompanied by a red shift in the onset of absorption and of EQE (visible on a linear
scale in Figure S3) suggesting aggregation of the molecule upon SVA. Crystallisation occurs

7

readily only in the case of SML06 because of the single alkyl chain on the flanking thiophene in
that molecule.
To quantify the losses in open-circuit voltage we have measured the external quantum
efficiency (EQE) and electroluminescence (EL) response of the devices. Figure 2 shows the
resulting EQE spectra extended to lower energy over several orders in magnitude using the EL
data and exploiting the reciprocity relation21 as done previously.16,22 Compared to spectra for
polymer:PC71BM or evaporated small molecule solar cells,16, 22-24 the EQE spectra in Figure 2a,
and 2b show a remarkably sharp absorption edge over several orders of magnitude for all
devices. In addition, it is evident that the EQE for the devices subjected to SVA maintain the
sharp gradient over more orders magnitude than the as-fabricated devices for both molecule
types.

8

100
EL

10-1
EQE

0.8

10-3

0.6

10-4

0.4
intensity φem /norm.

10-2

10-5

0.2

-6
(a) SML01:PC71BM 10

EL

1.0
0.8
0.6
0.4
0.2
1.0

100
10-1
EQE
10-2
Neat
10-3
PDMS
PDMS SVA 10-4
SVA
10-5
-6
(b) SML06:PC71BM 10

external quantum efficiency EQE

1.0

1.5
2.0
2.5
energy E/eV

Figure 2. Electroluminescence and external quantum efficiency for (a) SML01:PC71BM and (b)
SML06:PC71BM blend device with different processing conditions. The external quantum
efficiencies are composed of the directly measured quantum efficiency and the quantum
efficiency determined from the EL spectra.

9

Following previously presented methods16, 25, we can separate the voltage difference between
the optical absorption edge Eopt/q where (q is the electric charge) and the measured Voc into three
different terms. For an ideal solar cell with a step-function absorptance and perfect carrier
collection but in the absence of light concentration or any directionality of light emission, we can
define a maximum open-circuit voltage for a material with a given band gap Eg by using the
Shockley-Queisser theory. The resulting open-circuit voltage Voc,SQ is typically about 250 mV
smaller than Eg/q at one sun illumination. This first component in the energy loss, Eg/q-Voc,SQ
cannot easily be associated to any particular loss mechanism but it would be reduced for
concentrator solar cells. In the context of organic solar cells this term can be considered as an
unavoidable loss mechanism. However, in principle the next two loss mechanisms are avoidable
and change substantially from one blend to another.16 The first of these, ∆Voc,abs = Voc,SQ - Voc,rad,
is the loss due to a broadened absorption edge. The so-called radiative open-circuit voltage is,
like Shockley-Queisser theory, based on the principle that the only charge carrier loss process is
radiative recombination, but it uses the measured absorptance rather than the ideal step-function
of Shockley and Queisser to calculate the radiative open-circuit voltage. The more the measured
absorptance mimics a step function, the closer Voc,SQ and Voc,rad will be. In contrast large shifts in
energy between the main onset of absorption and the tail of absorption, which dominates the
luminescence, lead to large values for ∆Voc,abs = Voc,SQ - Voc,rad. In an organic donor-acceptor
blend based solar cell this loss can be substantial because the energy of interfacial charge transfer
states is often strongly red shifted relative to the onset of strong absorption by the
semiconductors. However, if the charge transfer state is barely visible and the luminescence peak
is close in energy to the absorption onset, ∆Voc,abs can in principle also become negative. This is
due to the fact that the step-function like absorptance used for the SQ limit always has an

10

absorptance of 1 at the peak of emission, while the absorptance at the emission peak in the
radiative limit can be substantially smaller than one. Thus, the radiative saturation current
density can in principle be smaller than the SQ saturation current density without violating the
theory. This means that energetic disorder has a small positive effect on Voc and a negative effect
on Jsc, when compared with a SQ solar cell at a given experimentally determined band gap.
The last loss ∆Voc,nr = Voc,rad - Voc represents the reduction in open-circuit voltage due to nonradiative recombination. Table 1 summarises the results of this analysis for the devices discussed
in the present paper and compares the results to data published in several recent publications that
allowed us to perform the same analysis as proposed here.

Table 1. Voltage loss analysis for the small molecule devices and comparison with a range of
devices published in the literature and analyzed in the same way.

In case of the literature data,

we added an abbreviation to show which donor-acceptor combination is used. P represents
polymer, F is fullerene, SM is small molecule, ‘:’ means donor-acceptor blend and ‘/’ indicates a
planar heterojunction between a donor and an acceptor. For the devices presented in this paper,
the optical gap Eg is determined from the onset of the external quantum efficiency as illustrated
in Figure S3.

Blends

Eg

VOC,SQ

VOC, rad

∆VOC, abs VOC

∆VOC, nr

SML01:PC71BM

1.61

1.32

1.28

0.04

0.99

0.29

SML01:PC71BM (PDMS)

1.61

1.32

1.30

0.02

0.99

0.31

SML01:PC71BM (PDMS, SVA)

1.60

1.31

1.31

0.005

0.99

0.32

11

SML06:PC71BM

1.58

1.29

1.21

0.08

0.85

0.36

SML06:PC71BM (PDMS)

1.58

1.29

1.22

0.07

0.85

0.37

SML06:PC71BM (SVA)

1.54

1.26

1.23

0.03

0.88

0.35

SML06:PC71BM (PDMS, SVA)

1.53

1.25

1.24

0.01

0.87

0.37

PTB7:PC71BM [11] (P:F)

1.61

1.33

1.13

0.20

0.74

0.39

PIPCPa:PC61BM [14] (P:F)

1.41

1.15

1.16

-0.01

0.89

0.27

DBP:ZClb [26] (SM:SM)

1.95

1.66

1.58

0.08

1.33

0.25

DTD:N2200c [27] (SM:P)

1.55

1.27

1.16

0.11

0.82

0.34

PTB7-Th:IDTIDT-ICd [28] (P:SM) 1.53

1.25

1.28

-0.03

0.94

0.34

a
PIPCP contains a backbone comprised of CPDT-PT-IDT-PT repeat units (CPDT = cyclopentadithiophene, PT = pyridyl[2,1,3]thiadiazole,
IDT = indacenodithiophene);
b
(tetraphenyldibenzoperyflanthrene : zinc chlorodipyrrin);
c
(diketopyrrolopyrrole-thieno[2,3-f]benzofuran : commercial polymer acceptor);
d
(poly(thieno[3,4-b]thiophene/benzodithiophene) : indacenodithiophenoindacenodithiophene)

Both optical gap and Voc are larger for SML01 than SML06. In the case of SML06 only, the
band gap is reduced by up to 0.05 eV upon SVA. This is due to the greater ability of this
molecule to crystallise upon annealing, also suggested by the domain growth in Figure S1and S2.
In all Voc,SQ lies 0.28 – 0.29 V below the optical gap. The loss in actual Voc compared to this
maximum, i.e. ∆Voc,abs + ∆Voc,nr, varies between 0.32 and 0.44 V. Compared to many typical
polymer:fullerene solar cells11 the avoidable loss ∆Voc,abs due to the non-sharp absorption edge is
very small, between 0.01 and 0.08 eV compared, for example, to 0.20 eV in the case of
PTB7:PC71BM.
The non-radiative voltage loss ∆Voc,nr of these materials ranges from 0.29 to 0.37 V and
in the best cases is lower than the lowest previously published values for polymer:fullerene solar
cells which were 0.32 V for an indacenodithiophene polymer29 and 0.34 V for a
diketopyrrolopyrole based polymer16. However, even lower values can be found in the literature
if the present analysis is applied to published data. The lowest values for ∆Voc,nr that we could

12

find were 250 mV for an evaporated bilayer solar cell reported in Ref [26] and 270 mV for a
polymer:fullerene bulk heterojunction solar cell reported in Ref [14]. Other relatively low values
that have been published recently are 340 mV for both a small molecule:polymer blend and a
polymer:small molecule blend both of which have no fullerene acceptor.27,28 Thus, low voltage
losses can be found for all combinations of donor and acceptor molecules. However, it can be
noted that material combinations including either a small molecule donor or a non-fullerene
acceptor are very well represented among the materials with very low voltage losses. In the
following we consider possible reasons for low voltage losses in the blends based on thje small
molecules SML01 and SML06.
The combination of low ∆Voc,abs and ∆Voc,nr explains the high open-circuit voltages that we
could achieve with these materials. Whilst the origin of the relatively low non-radiative loss is
not known, it is consistent with the higher degree of purification possible for small molecules
than for polymers, which could reduce the defect density, and also with the lower degree of
disorder in electronic state energies that is expected in small molecules due to their limited
conformational phase space . In the case of the SML06 devices, the non-radiative loss is lower
after SVA. This is consistent with the longer charge recombination time observed for the SVA
relative to the non-annealed devices using transient-photovoltage measurements (Figure S4). As
established previously15 SVA appears to lead to larger domains and therefore less interfacial
area. It has been shown elsewhere that reduction in interfacial area in bulk heterojunctions can
improve Voc by reducing non-radiative recombination losses24 and it is consistent with the fact
that the lowest ∆Voc,nr we found was for a bilayer system where the interfacial area is naturally
minimized.

13

We notice as well that the solvent vapour annealing stage decreases ∆Voc,abs for SML06
samples by at least 0.03eV. This is due to the sharpening of the absorption edge that is observed
upon SVA, which was assigned to aggregation of SML06.
Compared to most polymer:fullerene systems (see e.g. Ref [16]), the SM devices show
exceptionally low absorption broadening losses and a Voc that is closer to the optical gap. In
order to understand why these small-molecule devices show a low voltage loss, especially the
one due to the absorption edge, we have used time-dependent density functional theory (TDDFT) to investigate the electronically excited states of the donor : acceptor complex for different
positions of the fullerene (PC71BM) relative to the donor molecule.30,31 We calculated the excited
states and their electronic density distribution using TD-DFT with B3LYP/6-31G* in vacuo with
Gaussian 0932. Whilst we recognize that the B3LYP functional has known limitations that affect
the study of charge transfer (CT) CT states, we selected this functional because it has shown
good agreement with experimental trends in CT state energy with changes in chemical
structure30 and it has been shown to reproduce experimental data as well as more expensive
methods33. For the present study we first confirmed that the method reproduced the relative
HOMO energies and optical gap of the two small molecules correctly. The results of the
calculation for different position of the fullerene (on top of the central EDOT, the flanking
thiophene (labelled T1), or the terminal group (labelled Rhod or CN)) are shown in Figure 3.
These positions are presented in Figure S5. The different states are classified as being CT states,
when most of an electronic charge is transferred to the fullerene, excitonic, when the excitation is
localized largely on either donor or acceptor, or mixed in the intermediate case.

14

(a)
100

(b)
100

SML01

-1

10

10-2

10-2

10-3

10-3

10-4
100

10-4
100

PC71BM

10-2

10-2

10-3

10-3

10-4
100

PC71BM

10-1

10-4
100

oscillator strength

10-1

oscillator strength

SML06

10-1

SML01:PC71BM

10-1

EDOT

10-2
10-3
10-4
100

SML01:PC71BM

10-1

SML06:PC71BM EDOT

10-1
10-2
10-3
10-4
100

SML06:PC71BM 1T1

10-1

1T1

10-2

10-2

10-3

10-3

10-4
100

10-4
100

10-1

SML01:PC71BM

SML06:PC71BM Rhod

10-1

CN

-2

10-2

10

CT states
excitons

10-3
10-4
1.6

1.8

energy E/eV

10-3
10-4
2.0

2.2

1.6

1.8

2.0

2.2

energy E/eV

Figure 3. TD-DFT calculation of excited states of SML01, SML06 and for SML01:PC71BM and
SML06:PC71BM complexes with the PC71BM in three different positions relative to the donor
molecule.

15

When the fullerene is placed above the acceptor (Rhod or CN) part of the molecules, there is
no charge-transfer state at an energy below the exciton state. For both the T1 and EDOT
positions, the lowest state is a charge-transfer state, and although the charge-transfer state for the
EDOT position lies lower in energy, the energies of the CT states for the different positions lie
within some kT (0.025eV) of each other. From these calculations we will consider the CT state
to be the first excited state for the complex where the fullerene is in the EDOT position, i.e. the
lowest excited state of all geometries considered. We hence obtain a (value for the exciton – CT
state energy difference) of 0.15 eV. The relatively small difference between the CT state and
donor exciton energy could explain the low voltage loss observed for these materials. In fact, we
can compute the absorption from the oscillator strength (Figure S6); these plots confirm that a
sharp, approximately exponential absorption edge is expected for these calculated transitions and
that the sharp edge is compatible with the low ∆Voc,abs observed.
In polymer based systems, current generation is observed to depend on the interfacial
driving force, and typically switches off when that the exciton-CT state energy gap drops below
~0.3 eV34. There are exceptions where efficient generation occurs at lower driving force, for
example in an isoindigo polymer35,36 and a diketopyrrolopyrrole terthiophene (DPP3T) polymer37
as well as for the four systems listed at the bottom of Table 1. For DPP3T:PCBM, for instance,
∆Voc,abs = 0.02 eV16 while the calculated exciton-CT state energy difference using a DPP3T
oligomer: PCBM complex is 0.15 eV, using the same methods as used for the SML0x:PCBM
complexes here (Figure S7) This value of 0.15eV is similar to the small molecule systems
studied here and smaller than the calculated exciton-CT state difference for the majority of
polymer:fullerene combinations used in solar cells27 (calculated using the same level of theory).

16

Thus, relative to typical polymer:fullerene systems, the SML01 and SML06 systems show
charge separation at unusually low exciton-CT state energy gaps. The larger threshold for charge
separation usually seen in polymer systems may result from a larger degree of conformational
disorder in the polymer compared to small molecules, leading to disorder in the CT state energies
and correspondingly inefficient separation for the fraction of CT states lying in the low energy
tail. For the small molecule systems we expect less disorder in both exciton and CT energy as a
result of the reduced number of donor: acceptor configurations, relative to a polymer: fullerene
system. We explored the effect of conformational disorder in the small molecule on exciton and
CT state energies and found that distortions of the small molecule backbone conformation that
are expected at room temperature affect the position of CT and exciton by < 20 meV and cause a
relatively minor broadening of the absorption spectrum (Figure S8). We propose that the
efficient charge separation at low driving force in small molecule based systems results, at least
in part, from the relatively low energetic disorder in the system, and that this is what enables the
relatively sharp EQE edge of the small molecule fullerene system, and hence the low losses in
Voc.
In conclusion, we have investigated the voltage loss in a small molecule solar cell using a
quantitative analysis of the loss processes based on quantum efficiency and electroluminescence
measurements. We have found that for SML01 and SML06, the absorption edge is very sharp
thereby considerably reducing the associated losses in open-circuit voltage. Additional
processing steps such as solvent vapour annealing lead to a further reduction in voltage losses,
probably due to phase segregation and a reduction in interfacial area. Time-dependent DFT
calculations on the donor-acceptor interface show that the energy difference between the lowest
charge-transfer state of the donor-acceptor complex and the donor exciton is around 0.15 eV and

17

lower than the threshold for efficient charge generation in typical polymer based systems.

We

suggest that the high charge generation efficiency at low energetic driving force results from
reduced disorder in the electronic energies of small-molecule, compared to polymer, systems.

Experimental Methods

The

devices

were

fabricated

with

the

layer

structure

glass/ITO/PEDOT:PSS/SML0x:PC71BM/Ca/Al. The substrates were cleaned by ultrasonic
treatment in acetone, and isopropyl alcohol and subsequently blow-dried with nitrogen. The
surfaces of the ITO substrates were processed in ozone at 100 W for 5 minutes. Subsequently, a
hole transport layer PEDOT:PSS (Clevios PVP AI 4083, filtered at 0.45 µm) was deposited by
spin-coating at 3500 rpm (thickness around ~30 nm ) onto the ITO surface and then annealed at
150 °C for 20 min. The organic films were deposited from a solution containing SML0x and
PC71BM in a weight ratio of 1:1 at a solid concentration of 20 mg/ml in chloroform. Another
solution was prepared with the same condition but also containing 0.3 mg/ml of PDMS
(molecular weight 14000 g/mol).
The active layer was spin-coated at 8000 rpm with acceleration 4000 rpm/s, yielding a
thickness around 75 nm. Where necessary an additional step was also carried out straight after
deposition of the active layer by exposing the films to a saturated vapour of dichloromethane
solvent (CH2Cl2) in a sealed closed vessel (solvent vapour annealing, SVA). The films were
exposed to the solvent vapour for 3-5 minutes. Finally, the layers of the top electrode, Ca
(~20nm) and Al (~100 nm), were deposited subsequently at a rate of 0.1Å/s through a shadow
mask with an aperture area of 0.045 cm2 under a constant pressure less than 10-6 Torr. Current

18

density–voltage (J–V) characteristics of the devices were measured using a Keithley 236 Source
Measure Unit and a xenon lamp with AM1.5G filters and 100 mW/cm² illumination solar
simulator (Oriel Instruments).
Electroluminescence (EL) was measured using a Shamrock 303 spectrograph combined with
an iDUS InGaAs array detector cooled to –90 °C. The driving injection current is in the range of
1.25 and 1250 mA/cm2. The obtained EL spectra intensity was calibrated with the spectrum from
a calibrated halogen lamp.
External quantum efficiency (EQE) was measured using a grating spectrometer (CVI
DIGIKROM 240) to create monochromatic light combined with a tungsten halogen light source.
The monochromatic light was chopped at 235 Hz and Stanford Research Systems SR380 lock-in
amplifier with an internal transimpedance amplifier of 106 V/A was used to detect the
photocurrent. Long pass filters at 610, 715, 780, 850, and 1000 nm were used to filter out the
scattered light from monochromator. The spectra were taken from 350 to 1100 nm and calibrated
by a silicon photodiode.
PL Steady-state: PL was collected using a Horiba Jobin Yvon Fluorolog-2 PL spectrometer.

TD-DFT: Ground state energies and geometries were calculated using Density Functional
Theory (DFT) with B3LYP functional and 6-31G* basis set in Gaussian 0929.We estimate the
LUMO levels from the first excited state of a time-dependent DFT (TDDFT) calculation using
the same functional and basis set.

19

ASSOCIATED CONTENT
Supporting Information
The supporting information contains morphological study, along with the measurements of
photoluminescence, EQE, TPV lifetimes and TD-DFT calculations

AUTHOR INFORMATION
Notes
The authors declare no competing financial interests.
ACKNOWLEDGMENT
JN acknowledges the Royal Society for a Wolfson Merit Award. JN, SMT and AAYG
acknowledge the Engineering and Physical Sciences Research Council for support via grants
EP/K029843/1, EP/K030671/1, EP/J017361/1. JY acknowledges support via an Imperial College
Rector’s Scholarship. MA acknowledges receipt of a Bourse d’Excellence from Ecole
Polytechnique. TK acknowledges support from the DFG (Grant KI-1571/2-1). SMT
acknowledges Hendrik Faber for AFM measurements.

REFERENCES
(1) Service, R. F. Outlook Brightens for Plastic Solar Cells. Science 2011 332, 293.
(2) Kaltenbrunner, M., White, M. S.; Glowacki; E. D., Sekitani, T.; Someya, T.; Saricifti, N.
S.; Bauer, S. Ultratin and Lightweight Organic Solar Cells with High Flexibility Nature
Commun. 2012, 3, 770.

20

(3) Chen, C., Chang, W., Yoshimura, K., Ohya, K., You, J., Gao, Yang, Y. An Efficient
Triple-Junction Polymer Solar Cell Having a Power Conversion Efficiency Exceeding
11%. Adv. Mater. 2014, 26, 5670-5677.
(4) He, Z. C.; Xiao, B.; Liu, F.; Wu, H.; Yang, Y.; Xiao, S.; Wang, C.; Russel, T. P.; Cao, Y.
Single-Junction Polymer Solar Cells With High Efficiency and Photovoltage. Nature
Photon. 2015, 9, 174-179.
(5) Liu, Y.; Zhao, J.; Li, Z.; Mu, C.; Ma, W.; Hu, H.; Jiang, K.; Lin, H.; Ade, H. Aggregation
and Morphology Control Enables Multiple Cases of High-Efficiency Polymer Solar Cells
Nature Commun. 2014, 5, 5293.
(6) Vohra, V.; Kawashima, K.; Kakara, T.; Koganezawa, T.; Osaka, I.; Takimiya, K.;
Murata, H. Efficient Inverted Polymer Solar Cells Employing Favourable Molecular
Orientation. Nature Photonics. 2015, 9, 403–408.
(7) Zhang, Q, ; Kan, B.; Liu, F.; Long, G.; Wan, X.; Chen, X.; Yi Zuo, Ni, W.; Zhang, H.;
Li, M.; Hu, Z.; Huang, F.; Cao, Y.; Liang, Z.; Zhang, M.; Russell, T. P.; Chen, Y. Smallmolecule solar cells with efficiency over 9%. Nature Photonics 2015, 9, 35.
(8) Kan, B.; Qian Zhang, Q.; Li, M.; Wan, X; Ni, W.; Long, G.; Wang, Y.; Yang, X.; Feng,
H.; Chen, Y. Solution-Processed Organic Solar Cells Based on Dialkylthiol-Substituted
Benzodithiophene Unit with Efficiency near 10%. J. Am. Chem. Soc. 2014, 136, 1552915532
(9) Liu, Y.; Chen, C.; Hong, Z.; Gao, J.; Yang, Y. (Michael); Zhou, H.; Dou, L.; Li, G.;
Yang, Y. Solution-Processed Small-Molecule Solar Cells: Breaking The 10% Power
Conversion Efficiency. Sci. Rep. 2013, 3, 3356.
(10)

Sun, Y.; Welch. G. C.; Leong, W. L.; Takacs, C. J.; Bazan, G. C.; Heeger, A. J.

Solution-Processed Small-Molecule Solar Cells with 6.7% Efficiency. Nature Mater.
2011, 11, 44-48.

21

(11)

Mukherjee, S.; Proctor, C. M.; Tumbleston, J. R.; Bazan, G. C.; Nguyen, T.-Q.;

Ade, H. Importance of Domain Purity and Molecular Packing in Efficient SolutionProcessed Small-Molecule Solar Cells. Adv. Mater. 2015, 27, 1105–1111.
(12)

Nayak, P. K.; Narasimhan, K. L.; Cahen, D. Separating Charges at Organic

Interfaces: Effects of Disorder, Hot States, and Electric Field. J. Phys. Chem. Lett. 2013,
4, 1707-1717.
(13)

Gao, K.; Li, L.; Lai, T.; Xiao, L.; Huang, Y.; Huang, F.; Peng, J.; Cao, Y.; Liu, F.;

Russell, T. P.; Janssen, R. A. J.; Peng, X. Deep Absorbing Porphyrin Small Molecule for
High-Performance Organic Solar Cells with Very Low Energy Losses. J. Am. Chem. Soc.
2015, 137, 7282-7285.
(14)

Ran, N. A.; Kuik, M.; Love, J. A.; Proctor, C. M.; Nagao, I.; Bazan, G. C.;

Nguyen, T.-Q. Understanding the Charge-Transfer State and Singlet Exciton Emission
from Solution-Processed Small-Molecule Organic Solar Cells. Adv. Mater. 2014, 26,
7405–7412.
(15)

Montcada, N. F. ; Domínguez, R. ; Pelado, B. ; Cruz, P. ; Palomares, E. ; Langa,

F. High Photocurrent in Oligo-Thienylenevinylene-Based Small Molecule Solar Cells
with 4.9% Solar-to-Electrical Energy Conversion. J. Mater. Chem. A, 2015, 3, 1134011348.
(16)

Yao, J.; Kirchartz, T.; Vezie, M. S.; Faist, M. A.; Gong, W.; He, Z.; Wu, H.;

Troughton, J.; Watson, T.; Bryant D.; Nelson, J. Quantifying Losses in Open-Circuit
Voltage in Solution-Processable Solar Cells. Phys. Rev. Appl., 2015, 4, 014020.
(17)

Rau, U. Reciprocity Relation Between Photovoltaic Quantum Efficiency and

Electroluminescent Emission of Solar Cells Phys. Rev. B, 2007, 76, 085303.

(18)

Kirchartz, T.; Rau, U.; Kurth, M.; Mattheis, J.; Werner, J. H. Comparative Study

of Electroluminescence from Cu(In,Ga)Se2 and Si Solar Cells Thin Solid Films 2007,
515, 6238-6242.

22

(19)

W. Shockley and H. J. Queisser. Detailed Balanced Limit of Efficiency of p-n

Junction Solar Cells J. Appl. Phys. 1961, 32, 510-519.
(20)

Koster, L.; Shaheen, S. E.; Hummelen, J. C. Pathways to a New Efficiency

Regime for Organic Solar Cells Adv. Energ. Mater, 2012, 2, 1246-1253.

(21)

Rau, U. Reciprocity Relation between Photovoltaic Quantum Efficiency and

Electroluminescent Emission of Solar Cells. Phys. Rev. B, 2007, 76, 085303.

(22)

Vandewal, K.; Albrecht, S.; Graham, K.R.; Widmer, J.;

Douglas, J. D.;

Schubert, M.; Mateker, W. R.; Bloking, J. T.; Burkhard, G. F.; Sellinger, A.; Fréchet,
J. M. J.; Amassian, A.; Riede, M. K.; McGehee, M. D.; Neher, D.; Salleo, A. Efficient
Charge Generation by Relaxed Charge-Transfer States at Organic Interfaces. Nature
Materials 2013, 13, 63-68.

(23)

Hörmann, U.; Kraus, J.; Gruber, M.; Schuhmair, C.;

Linderl. T.; Grob, S.;

Kapfinger, S.; Klein, K.; Stutzman, M.; Krenner, H. J.; Brütting, W. Quantification of
Energy Losses in Organic Solar Cells from Temperature-Dependent Device
Characteristics. Phys. Rev. B 2013, 88, 235307.

(24)

Vandewal, K.; Widmer, J.; Brabec C, J.; Heumueller, T.; McGehee, M. D.; Leo

K.; Riede, M.; Salleo, A. Increased Open-Circuit Voltage of Organic Solar Cells by
Reduced Donor-Acceptor Interface Area'. Adv. Mater. 2014, 26, 3839-3843.

(25)

Rau, U., Paetzold, U.W., Kirchartz, T. Thermodynamics of Light Management in

Photovoltaic Devices Phys. Rev. B, 2014, 90, 035211.
(26)

Bartynski, A. N.; Gruber, M.; Das, S.; Rangan, S.; Mollinger, S.; Trinh, C.;

Bradforth, S. E.; Vandewal, K.; Salleo, A.; Bartynski, R. A.; Bruetting, W.; Thompson,
M. E. Symmetry-Breaking Charge Transfer in a Zinc Chlorodipyrrin Acceptor for High
Open Circuit Voltage Organic Photovoltaics J. Am. Chem. Soc. 2015, 137, 5397-5404.

23

(27)

Tang, Z.; Liu, B.; Melianas, A.; Bergqvist, J.; Tress, W.; Bao, Q.; Qian, D.;

Inganäs, O.; Zhang, F. A New Fullerene-Free Bulk-Heterojunction System for Efficient
High-Voltage and High-Fill Factor Solution-Processed Organic Photovoltaics Adv.
Mater. 2015, 27, 1900-1907.
(28)

Li, Y.; Liu, X.; Wu, F.; Zhou, Y.; Jiang, Z.; Song, B.; Xia, Y.; Zhang, Z.; Gao, F.;

Inganas, O.; Li, Y.; and Liao L. Non-fullerene acceptor with low energy loss and high
external quantum efficiency: towards high performance polymer solar cells. J. Mater.
Chem. A, 2016, 4, 5890.

(29)

Baran, D.; Vezie, M. S.; Gasparini, N.; Deledalle, F.; Yao, J.; Schroeder, B. C.;

Bronstein, H.; Ameri, T.; Kirchartz, T.; McCulloch, I.; Nelson, J.; Brabec, C. J. Role of
Polymer Fractionation in Energetic Losses and Charge Carrier Lifetimes of Polymer:
Fullerene Solar Cells. J. Phys. Chem. C, 2015, 119, 19668–19673.

(30)

Few, S.; Frost, J. M.; Kirkpatrick, J.; Nelson, J. Influence of Chemical Structure

on the Charge Transfer State Spectrum of a Polymer:Fullerene Complex , J. Phys. Chem.
C, 2014, 118, 8253–8261.

(31)

Veldman, D., Meskers, S. C. J.; Janssen, R. A. J. The Energy of Charge-Transfer

States in Electron Donor-Acceptor Blends: Insight into the Energy Losses in Organic
Solar Cells. Adv. Funct. Mater., 2009, 19, 1939-1948.

(32)

Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;

Cheese-man, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A. et al.
Gaussian 09, Revision A.1,.Gaussian, Inc., WallingfordCT, 2009.

(33)

Niedzialek, D.; Duchemin, I.; Branquinho de Queiroz, T.; Osella, S.; Rao, A.;

Friend, R.; Blase, X.; Kümmel, S.; Beljonne, D. First Principles Calculations of Charge

24

Transfer Excitations in Polymer–Fullerene Complexes: Influence of Excess Energy. Adv.
Funct. Mater. 2015, 25, 1972–1984.

(34)

Faist, M. A.; Kirchartz, T.; Gong, W.; Ashraf, R. S.; McCulloch, I.; de Mello, J.

C.; Ekins-Daukes, N. J.; Bradley, D. D. C.; Nelson, J. Competition Between the Charge
Transfer State and the Singlet States of Donor or Acceptor Limiting the Efficiency in
Polymer:Fullerene Solar Cells. J. Am. Chem. Soc. 2012, 134, 685−692.
(35)

Vandewal, K.; Ma, Z.; Bergqvist, J.; Tang, Z.; Wang, E.; Henriksson, P.;

Tvingstedt, K.; Andersson, M. R.; Zhang, F.; Inganäs, O. Quantification of Quantum
Efficiency and Energy Losses in Low Bandgap Polymer:Fullerene Solar Cells with High
Open-Circuit Voltage. Adv. Funct. Mater. 2012, 22, 3480–3490.

(36)

Wang, E.; Zhang, Z.; Vandewal, K.; Henriksson, P.; Zhang, F.; Andersson, M. R.;

Ma, Z.; Inganäs, O. An Easily Accessible Isoindigo- Based Polymer for HighPerformance Polymer Solar Cells. J. Am. Chem. Soc., 2011, 133, 14244−14247.
(37)

Bijleveld, J. C.; Zoombelt, A. P.; Mathijssen, S.G. J.; Wienk, M. M.; Turbiez, M.;

de Leeuw, D. M.; Rene A. J. Janssen, R. A. J. J. Am. Chem. Soc. 2009, 131, 16616–
16617.

25