Photochemical stability of π-conjugated polymers for polymer solar cells: A rule of thumb

Article abstract of DOI:10.1039/c0jm03105d

A comparative photochemical stability study of a wide range of π-conjugated polymers relevant to polymer solar cells is presented. The behavior of each material has been investigated under simulated sunlight (1 sun, 1000 W m^-2, AM 1.5G) and ambient atmosphere. Degradation was monitored during ageing combining UV-visible and infrared spectroscopies. From the comparison of the collected data, the influence of the polymer chemical structure on its stability has been discussed. General rules relative to the polymer structure-stability relationship are proposed.

Full text of DOI:10.1039/c0jm03105d

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2011, 21, 4132
www.rsc.org/materials
PAPER
Photochemical stability of p-conjugated polymers for polymer solar cells:
a rule of thumb†
ꢀ
Matthieu Manceau, Eva Bundgaard, Jon E. Carle, Ole Hagemann, Martin Helgesen, Roar Søndergaard,
*
Mikkel Jørgensen and Frederik C. Krebs
Received 15th September 2010, Accepted 16th December 2010
DOI: 10.1039/c0jm03105d
A comparative photochemical stability study of a wide range of p-conjugated polymers relevant to
polymer solar cells is presented. The behavior of each material has been investigated under simulated
sunlight (1 sun, 1000 W mꢀ2, AM 1.5G) and ambient atmosphere. Degradation was monitored during
ageing combining UV-visible and infrared spectroscopies. From the comparison of the collected data,
the influence of the polymer chemical structure on its stability has been discussed. General rules relative
to the polymer structure–stability relationship are proposed.
From a simplified chemical point of view, polymers for organic
Introduction
solar cells can be described as the combination of a rigid p-
conjugated backbone regularly substituted by side-chain groups
ensuring their solution processability. Previously published
papers already identified the critical role of the side-chain in the
polymer degradation processes.14–16 A large difference in terms of
the stability between MDMO–PPV and P3HT—two of the most
studied polymers in the field—was also reported, P3HT being
much more stable whatever the ageing conditions.17 However, to
our knowledge, no detailed studies have yet been dedicated to the
influence of the backbone nature on the polymer stability.
In this work, we present a photochemical stability study in air
on 24 different polymers (34 including the thermo-cleaved
derivatives) relevant to PSCs. Samples were selected to cover
a very broad range of polymer types (purely donor, donor/
acceptor, thermo-cleavable, etc.) and chemical structures. Many
of the moieties commonly used in the PSCs field are thus
included in this paper. As all the experiments were conducted
under the same conditions, comparison of the collected data was
possible and the influence of different points is discussed (donor
and acceptor group nature, side-chain type). This screening
finally allowed for the description of general rules for the p-
conjugated polymer photochemical stability.
Polymer-based solar cells (PSCs) have the potential to become
one of the future’s renewable and environmentally friendly
energy sources. Combining several attractive properties—flexi-
bility, low manufacturing costs, low capital investment in
equipment, a low thermal budget and the use of only abundant
elements in the active layer—they open up a variety of new
market opportunities and applications and have thus been under
intense research focus during the last decade.1–6 This develop-
ment has led to a significantly improved device power conversion
efficiency, that now exceeds 8%.7 To reach this, many classes of
new polymers have been designed, synthesized, characterized and
incorporated into photovoltaic devices.1,2,8–11 Consequently,
a very broad range of material families have already been
investigated in order to create polymers with good solubility,
small band-gap, strong absorbance, appropriate HOMO and
LUMO energy levels and high charge carrier mobilities.
However efficient and promising these materials are, their
practical use in large-scale PSC production can only be successful
if they also provide a good processability, a sufficient photo-
chemical stability and device stability.12,13 So far, this last point
has received limited attention and the literature is still scarce. As
a result, the relationship between the polymer chemical structure
and the expected device efficiency is rather well explored, the
chemical structure–photochemical stability relationship,
however, remains largely unknown.
Experimental
Samples preparation
Synthetic procedures and characterization data for the materials
have either been described in detail elsewhere18–24 or are given in
the ESI†. Molecular weights and optical band-gaps of the
samples are collected in Table S1, ESI† (where available, power
conversion efficiencies have been added).
Risø National Laboratory for Sustainable Energy, Technical University of
Denmark, P.O. Box 49, DK-4000 Roskilde, Denmark. E-mail: frkr@risoe.
dtu.dk; Fax: +45 46 77 47 91; Tel: +45 46 77 47 99
† Electronic
supplementary
information
(ESI)
available:
Characterization data of the polymers, list of abbreviations, thermal
cleavage conditions and IR spectra recorded along ageing. See DOI:
10.1039/c0jm03105d
Pure polymer samples were spin-coated on KBr plates from
chlorobenzene solutions. The polymer concentration in the
4132 | J. Mater. Chem., 2011, 21, 4132–4141
This journal is ª The Royal Society of Chemistry 2011

View Online
spin-coating solutions were adjusted to get a maximum peak
absorbance of about 0.8 for each material.
provide a list of the studied chemical units along with their full
IUPAC names and the abbreviations in the ESI†.
Thermal cleavage was performed in the inert atmosphere of
a glove box and the reaction progress was checked by IR spec-
troscopy. The heating step was kept as short as possible to avoid
undesirable thermal degradation reactions. For cleavage
temperatures and durations, see Table S2, ESI†.
Results and discussion
Pure donor polymers
During the last decade, two polymer families have played a major
role in the development of PSCs: polyphenylenevinylene
derivatives and polythiophene. For example, polymers such as
poly(2-methoxy-5-(20-ethylhexyloxy)-1,4-phenylenevinylene)
Ageing and characterization
Samples were illuminated under 1 sun and ambient air with
monitoring of the relative humidity (but no control) using
a standard solar simulator from Steuernagel Lichttechnik (KHS
575, AM 1.5G, 1000 W mꢀ2, 85 ꢁC, 30 ꢂ 10% RH). The samples
were removed periodically and UV-visible absorbance and IR
spectra were recorded to monitor the degradation. UV-visible
absorbance spectra were recorded from 200 to 1100 nm using
a UV-1700 spectrometer from Shimadzu. IR spectroscopy was
conducted with a Spectrum One from PerkinElmer operating in
transmission mode (4 cmꢀ1 resolution, 32 scans summation).
(MEH–PPV),
poly(2-methoxy-5-(30,70-dimethyloctyloxy)-1,4-
phenylenevinylene) (MDMO–PPV) and poly(3-hexylthiophene)
(P3HT) have been widely studied as they used to give the best
device performances. A few reports on the photochemical
stability of these materials have been published.14–17 Some of
these studies revealed that polyphenylenevinylene derivatives are
extremely unstable under photo-oxidative conditions.14 This
behavior has been attributed both to the presence of the vinylene
bond and of the alkoxy substituents. Conversely, P3HT has been
shown to be much more stable.17
In this first part of the study, we compare the photochemical
behavior in air of two donor conjugated polymers: MEH–
PPV and poly(2,20-(2,5-bis(2-hexyldecyloxy)-1,4-phenylene)-
dithiophene) (JC1) (Fig. 1).
Stability evaluation
To quantitatively compare all the materials, the total amount of
absorbed photons (NTott) was monitored versus ageing time over
the range l1 ꢀ l2 by summation over the polymer absorption
peaks (Table S1, ESI†). This value was calculated according to
the following equation:
Both materials are comprised of a dialkoxybenzene unit
alternating either with
a vinylene bond (MEH–PPV) or
a bithiophene group (JC1). P3HT data were also added and used
as a benchmark. From the results presented in Fig. 1, it is very
clear that JC1 is much more stable than MEH–PPV. This result
confirms that the exocyclic double bond has a very strong
detrimental effect on the MEH–PPV stability. This very low
stability is due to the fact that vinylene bonds can be easily
saturated by the radicals formed after side-chain cleavage.14
Owing to its aromaticity, thiophene is much more difficult to
saturate. Replacing the vinylene unit by a bithiophene thus
greatly enhances the whole polymer stability.
l2
ꢀ
ꢁ
X
tðlÞ
NTt ot
¼
N0ðlÞ ꢃ 1 ꢀ 10ꢀA
l1
where At(l) is the absorbance at a given wavelength (l) and time
(t), and N0(l) is the incident photon flux. At(l) was directly
extracted from the UV-visible absorbance spectra of the sample
at the corresponding ageing time (t). The ASTM G173 standard
was used as a reference for the incident photon flux.25
At the end of the degradation, the quantity of absorbed
photons systematically reached a constant value (NTotN) after
which no absorbance evolution followed. This value was always
above zero due to the absorption, reflection and scattering of the
KBr substrate. Finally, the normalized number of photons
absorbed by the polymer was calculated by:
Very interestingly, JC1 is observed to be as stable as P3HT
which could be surprising as alkoxy side-chains are well-known
for their poor stability. It should, however, be recalled that P3HT
instability has been ascribed to the hexyl side-chains. The pres-
ence of two unsubstituted thiophene rings in JC1 then probably
balances the presence of the alkoxy substituents on the phenyl
ring and explains our finding.
NTt ot ꢀ NTNot
NPt hotons
¼
NT0ot ꢀ NTNot
Donor–acceptor copolymers, backbone composition effect
t
For a great majority of the experiments, NPhotons exhibited
a linear decay. In that case, the experimental data were then fitted
with a straight curve. Quantitative comparisons of the respective
stability of different samples were then established using the
slopes of these curves. In some cases a logarithmic time scale was
used for the sake of clarity.
Presently, strong efforts are directed towards the synthesis of
materials absorbing at longer wavelengths (i.e. with lower band-
gaps) that can harvest a larger fraction of the solar spectrum. The
most common strategy to control the band-gap is to alternate the
electron-rich (donor) and electron-poor (acceptor) groups in the
main chain of the polymer.8,26–28 An Internal Charge Transfer
(ICT) from the donor to the acceptor occurs and a reduction of
the band-gap is achieved. In this part of the study, attention will
be focused on these so-called donor/acceptor polymers.
Naming of compounds
Different abbreviations can often be found in the literature for
the same material due to the very complex IUPAC names that
these materials generally present. We have employed the most
commonly employed abbreviations for these materials and also
Influence of the donor group. Five series of polymers were aged
to investigate the influence of the donor group on the stability of
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4132–4141 | 4133

View Online
Fig. 1 (Left) Chemical structure of the investigated samples, (right) evolution of the normalized amount of absorbed photons during photochemical
ageing. Note how MEH–PPV degrades very quickly.
the polymer. For a given series of compounds, the only difference
between the samples was the nature of the donor moiety.
that the substitution of this quaternary carbon atom by a silicon
atom results in a significant improvement in the stability. One
could imagine that this increase originates from a lowering of the
HOMO level when carbon is replaced by silicon, as a deeper
HOMO enhances the oxidative stability.31 However, Scharber
et al. and Chen et al. previously showed that the nature of the
bridging atom (C or Si) had almost no influence on the HOMO
and LUMO levels of the polymer.32,33 This implies that the
observed stability improvement cannot result from a lower
HOMO level. But, it could be explained by the presence of the
silicon atom which is known to be less easily oxidized than the
carbon. Finally, using unsubstituted thiophene as a donor group
highly improves the photochemical stability. More than 600
hours of irradiation are necessary to achieve full degradation. It
is likely that this increase comes from the absence of both the
quaternary site and side-chain in this donor moiety.
Dithienylthienopyrazine series. Materials belonging to this first
class are copolymers based on dithienylthienopyrazine bearing
thermo-cleavable tertiary esters on the pyrazine ring, alternating
with different donor groups: fluorene, carbon-bridged cyclo-
pentadithiophene
(CPDT),
silicon-bridged
cyclo-
pentadithiophene (Si-CPDT) and thiophene (Fig. 2).
Fig. 3 presents the evolution of the normalized amount of
absorbed photons versus ageing time for each polymer. As clearly
observed, the stability ranking from the lowest to the highest is as
follows: fluorene–CPDT–Si-CPDT–thiophene. It is also worth
noting that an identical ranking is obtained with the cleaved
materials but on a very different timescale (Fig. 3). This second
point will be discussed in detail later. Different comments can be
made based on this ranking. First of all, the two least stable
polymers contain a quaternary carbon atom in their backbone.
As this type of site can be readily oxidized,29,30 this can explain
the poor stability of these two polymers that are completely
bleached after less than 100 hours. Secondly, it can be noticed
In parallel, the behavior of the non-cleaved samples was
monitored by IR spectroscopy all along the ageing process (see
Fig. S1, ESI†). Interestingly, very similar modifications are
observed for the different polymers albeit on different timescales.
Indeed, one can systematically notice: (i) a decrease in the
intensity of the signals coming from the alkyl side-chains (3000–
2850 cmꢀ1); (ii) the development of a broad signal in the carbonyl
range (1800–1600 cmꢀ1) and (iii) the appearance of signals
characteristic for sulfinic esters (1115 and 620 cmꢀ1). As reported
in the case of P3HT, the observation of these latter bands indi-
cates that the final degradation stages of the thiophene rings is
reached.15 Their appearance is then a good indicator of how
advanced the sample degradation is. Here we noticed that such
signals appeared very quickly for the fluorene and CPDT-based
compounds (about 20 hours). On the contrary, they were only
detected after 200 hours for the thiophene.
As for cleaved materials, the behavior was very similar to the
pristine polymers, but changes in the IR spectrum were much
slower after cleavage. This is in good agreement with our
previously published results.34
Fig. 2 Chemical structure of the materials in the series of materials using
a thermo-cleavable dithienylthienopyrazine (shown left). The X in the
polymer backbone designates one of four donor groups shown to the
right of the broken line.
Dithienylbenzothiadiazole (Series 1). To check the consistency
of the previous results, three of the four previous donor
4134 | J. Mater. Chem., 2011, 21, 4132–4141
This journal is ª The Royal Society of Chemistry 2011

View Online
Fig. 3 Evolution of the normalized amount of absorbed photons during photochemical ageing. (Left) Pristine polymers, (right) cleaved polymers.
groups—thiophene, CPDT and Si-CPDT—were investigated in
a second set of experiments. Here, samples were based on an
electron-deficient benzothiadiazole group with two flanking
thiophene rings substituted by a cleavable ester moiety (Fig. 4).
From Fig. 4, it is obvious that the stability ranking remains
identical to the one reported in the previous section.
This means that even though different parameters can influ-
ence the stability of a polymer (e.g. molecular weight, regior-
egularity, purity) the chemical nature remains the most
important one.
The thienoimidazolone-based sample appeared to be as
unstable as the one based on CPDT. This is related to the pres-
ence of the imide group that is photochemically unstable.35 As
evidenced by Arnaud et al., irradiation of this unit causes the
homolysis of the C–N bonds which leads to the degradation of
the whole unit through oxidation of the formed radicals. This is
confirmed here, as the IR bands characteristic for the imide
group (1725 and 1560 cmꢀ1) disappeared after only a few hours
of irradiation. As the donor unit is quickly degraded, the ICT is
prevented and a rapid absorbance loss takes place.
Attention should also be drawn to the fact that the polymers
containing the carbazole and dialkoxybenzene groups were more
stable than polymers containing the CPDT moiety. However,
they all degrade relatively quickly as complete photo-bleaching
was achieved after less than 100 hours. A rapid decrease in the IR
bands coming from carbazole moieties (e.g. 1600 cmꢀ1) was
evidenced confirming the limited stability of this moiety. This can
be ascribed to various phenomena. First, it is due to the presence
of the Csp3–N bond that can be easily cleaved as previously
reported.36 A carbazoyl radical is generated, that can further
Dithienylbenzothiadiazole (Series 2). In a third step, three
new electron-rich moieties were introduced—dialkoxybenzene,
carbazole, thienoimidazolone—and compared to thiophene and
CPDT. These donor moieties were copolymerized with
a central benzothiadiazole flanked by two unsubstituted thio-
phene rings (Fig. 5). Here again, the evolution of the amount of
absorbed photons has been recorded all along ageing and
results are shown in Fig. 5. IR data are also provided in the
ESI (Fig. S2†).
Fig. 4 Chemical structures of the investigated polymer samples (left). The X denotes one of the three donor units shown to the right of the broken line
(middle). The evolution of the normalized amount of absorbed photons during photochemical ageing of all materials is shown in the plot (right).
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4132–4141 | 4135

View Online
Fig. 5 Chemical structure of the investigated polymer samples (left). The X denotes one of the five donor units shown to the right of the broken line
(middle). The evolution of the normalized amount of absorbed photons during photochemical ageing of all materials is shown in the plot (right).
react with oxygen. This ends up in the degradation of the
carbazole group, and thus in the interruption of the p-conju-
gated system. The C–N bond cleavage was confirmed by the IR
monitoring, as signals characteristic for alkyl side-chains
decreases all along ageing. However, the breaking of this bond
does not affect the conjugated backbone of the polymer directly
as it is the case for C–N homolysis in the carbazole moiety. This
could explain why the dialkoxybenzene unit gives a photochem-
ical stability comparable to that of the carbazole.
(nC–H z 3000 to 2850 cmꢀ1) and Csp3–N bonds (nC–N
z
1330 cmꢀ1) quickly vanished. Pfister and Williams also suggested
that irradiation of the carbazole group leads to the formation of
quinonic oxidized structures after reaction with the superoxide
anion O2cꢀ.37 This second pathway can also contribute to the
photodegradation of the sample.
Finally, and as one could have expected, the thiophene-based
polymer was once again shown to be the most stable among the
investigated materials although it was substituted by a carboxylic
acid. This is due to the simultaneous absence of breakable bonds
(C–O, C–N), of the quaternary carbon and of cleavable side-
chain.
The dialkoxybenzene-based polymer appeared to be approxi-
mately as stable as the carbazole-based sample. This finding can
seem rather surprising as alkoxy side-chains are usually known
for their very negative impact on the polymer stability. The C–O
bond is indeed readily cleavable under irradiation38 and as
expected, the intensity of the IR bands coming from the side-
chains (nC–H z 3000 to 2850 cmꢀ1 and nC–O 1385 cmꢀ1) gradually
Dithienylbenzothiadiazole (Series 3). This fourth class is
similar to the previous one except that the benzothiadiazole unit
bears two solubilizing alkoxy side-chains (Fig. 6). Three donor
units have been studied—CPDT, Si-CPDT, carbazole—and the
results are presented in Fig. 6. First, it is further confirmed that
Fig. 6 Chemical structure of the investigated polymer samples (left). The X denotes one of the three donor units shown to the right of the broken line
(middle). The evolution of the normalized amount of absorbed photons during photochemical ageing of all materials is shown in the plot (right).
4136 | J. Mater. Chem., 2011, 21, 4132–4141
This journal is ª The Royal Society of Chemistry 2011

View Online
the use of Si-CPDT provides a greater stability than CPDT. This
finding is consistent with the previously discussed data.
Secondly, we observe that carbazole is less stable than Si-
CPDT. As deduced from these results, the stability ranking from
the lowest to the highest for this series of compounds is: CPDT–
Carbazole–Si-CPDT.
From Fig. 9, it is very clear that the sample containing the TPz
unit is significantly more stable. This effect cannot be explained
by the presence of side-chains or by a difference in the samples
HOMO position. Blouin et al. reported that the substitution of
a BTD unit by a TPz should theoretically lead to an increase in
the polymer HOMO energy level.39 This was confirmed experi-
mentally by Bijleveld et al. and should engender a lower oxidative
stability of the TPz-based compound.40 The superior stability of
the sample based on the TPz moiety must then come from
a greater intrinsic photochemical stability of this unit.
Benzothiadiazole series. So far the unsubstituted thiophene
ring has been shown to be the most stable moiety especially
because this moiety is side-chain free. So, in a last series of
experiments, different unsubstituted aromatic donor groups were
investigated: thiophene, benzodithiophene and dithienothio-
phene. Samples were based on a benzothiadiazole unit bearing
solubilizing alkoxy side-chains without flanking thiophenes
(Fig. 7).
Si-bridged cyclopentadithiophene series. A second set of poly-
mers was then studied. This one was based on the Si-CPDT unit
associated either with a benzothiadiazole (BTD) unit or an ester-
substituted thienothiophene (Fig. 10). This latter has recently
given very good results in terms of efficiencies.10,11
As observed in Fig. 7, the change of the thiophene by a ben-
zodithiophene results approximately in a twofold stability
increase. One can also easily notice that the dithienothiophene
derivative provides by far the highest stability. Obviously, the use
of unsubstituted polycyclic aromatic units is thus beneficial in
terms of the photochemical stability.
According to the results presented in Fig. 10, the BTD-based
polymer appears to be much more stable than the thienothio-
phene. However, it should be emphasized that, whereas the BTD
group is not substituted, the thienothiophene unit bears
a primary ester side-chain. It is thus very likely that the rather
fast degradation originates from this group as IR monitoring of
the degradation shows a decrease in the signals pertaining to this
ester moiety (e.g. nC–O z 1350 cmꢀ1, see Fig. S3, ESI†). In
addition, signals characteristic for the degradation of the sulfur-
containing rings appeared quickly (around 1115 and 620 cmꢀ1).
As previously stated, the observation of such signals implies an
opening of the ring and thus an advanced degradation level. The
presence of this side-chain is, however, required to adjust the
position of the energy levels.10 It should be added that very good
performance has also been obtained using the thienothiophene
group substituted by a ketone.11 Ketones are well-known to be
highly unstable under irradiation as they readily evolve through
Norrish reactions.41 It can be anticipated that the photochemical
stability of the whole sample will be rather limited.
In summary, a combination of all the results collected so far
enables the formulation of a global stability ranking for the
investigated donor groups as shown in Fig. 8. One can conclude
that the presence of a quaternary carbon atom or an easily
cleavable bond leads to a rather low stability. Conversely, donor
units that provide the highest stability are those without any side-
chain.
Influence of the acceptor group. In a similar fashion, the
influence of the acceptor group on the photochemical stability of
conjugated polymers was investigated.
Dithienocyclopentadithiophene series. To begin, we studied
benzothiadiazole (BTD) and thienopyrazine (TPz) units. As
illustrated in Fig. 9, samples were based on a dithienocyclo-
pentadithiophene electron-rich group. In order to minimize the
effects of the side-chain, we chose to study a TPz unit substituted
by the carboxylic acid (i.e. thermo-cleaved).
As for the previous section dedicated to the donor group, the
different data were combined to establish the stability ranking
given in Fig. 11. Here again the most stable moieties are the ones
without side-chains or readily cleavable bonds.
Fig. 7 Chemical structure of the investigated polymer samples (left). The X denotes one of the three donor units shown to the right of the broken line
(middle). The evolution of the normalized amount of absorbed photons during photochemical ageing of all materials is shown in the plot (right).
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4132–4141 | 4137

View Online
Fig. 8 Donor group stability ranking.
Fig. 9 Chemical structure of the investigated polymer samples (left). The X denotes one of the two acceptor units shown to the right of the broken line
(middle). The evolution of the normalized amount of absorbed photons during photochemical ageing of all materials is shown in the plot (right).
Fig. 10 Chemical structure of the investigated polymer samples (left). The X denotes one of the two acceptor units shown to the right of the broken line
(middle). The evolution of the normalized amount of absorbed photons during photochemical ageing of all materials is shown in the plot (right).
4138 | J. Mater. Chem., 2011, 21, 4132–4141
This journal is ª The Royal Society of Chemistry 2011

View Online
Side-chain effect
the backbone chemical structure, it was noticed that thermo-
cleaved samples are approximately between 2 and 20 times more
stable than corresponding pristine materials. This set of results
further confirms the higher potential stability offered by thermo-
cleaved conjugated polymers. Among the investigated materials,
the beneficial effect of thermal cleavage surprisingly appeared to
be more pronounced for those which are the most stable before
cleavage. This is of course a very interesting and fortunate result.
It should finally be mentioned that some of the thermo-cleaved
polymers exhibited a very high photochemical stability. For
example, 1000 hours irradiation of the polymer based on a thie-
nopyrazine unit and three thiophene rings only lead to a 20%
decrease in the amount of absorbed photons.
Influence of the nature of side-chains. The presence of side-
chains is necessary to ensure a sufficient solubility of the sample
and thus allows the solution-processability of the PSCs active
layer. However, side-chains can also be used to tune other
properties of the materials such as the positions of the HOMO
and LUMO levels or the packing of the macromolecular
chains.10
To begin, we studied different alternating polymers based on
a dithienocyclopentadithiophene unit associated with a benzo-
thiadiazole group. This latter was either unsubstituted or
substituted with one of the following side-chains: ether or ester
(Fig. 12). As in the previous experiments, the total number of
absorbed photons was recorded all along ageing and the results
are presented on Fig. 12.
Summary
According to these results, the stability ranking from the
lowest to the highest is as follows: CO2R < OR < H. As expected,
the unsubstituted BTD demonstrates the highest stability among
the three samples although the durability difference with the
other samples is surprisingly low. On the contrary, ester and
alkoxy substituents gave a lower stability, the latter being slightly
more stable.
As clearly demonstrated in the present work, slight changes in the
material’s chemical structure can result in huge variations in the
photochemical stability. Indeed, polymer durability was shown
to cover a very broad range of values, from very few hours (e.g.
MEH–PPV) to several thousands of hours (e.g. thermo-cleaved
samples). Several crucial parameters influencing the stability
have been identified through this study and the main findings
have been summarized in the following basic rules.
1. The use of exocyclic double bonds in the main backbone
(MEH–PPV, MDMO–PPV) leads to a poor stability and should
be avoided,
Thermo-cleavable polymers. As described in previously pub-
lished papers14–16 and also shown before in this study, side-chains
have a negative influence on the polymer photochemical stability.
In addition, in a previous paper we showed that side-chain
thermal cleavage systematically lead to a strong increase in the
sample photochemical stability.34 Here, the study was extended
to various new thermo-cleavable polymers, and a total of 8
different samples were investigated.
2. Moieties containing a quaternary site are very unstable (e.g.
fluorene, cyclopentadithiophene) because of the oxidazability of
this site,
3. The presence of readily cleavable bonds (such as C–N or C–
O) also limits stability,
Whatever the sample nature, thermo-cleavage systematically
led to an increase in the stability as exemplified on Fig. 3. In every
case, the amount of absorbed photons exhibited a quasi-linear
decay and the experimental data were thus fitted with a straight
curve. The stability improvement provided by thermal cleavage
was then estimated quantitatively for each sample by comparing
the slope of the curves before and after side-chain cleavage.
These results are reported in the ESI (Table S2†). Depending on
4. Side-chains play a key role in conjugated polymer degra-
dation and their cleavage largely improves stability,
5. Aromatic polycyclic units generally exhibit a good photo-
chemical stability.
Indirectly, it was also shown that the position of the HOMO
level for the polymers is not a sufficient criterion to conclude on
the photochemical stability.
Fig. 11 Acceptor group stability ranking.
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4132–4141 | 4139

View Online
Fig. 12 Chemical structure of the investigated polymer samples (left). The X denotes one of the three substituents shown to the right of the broken line
(middle). The evolution of the normalized amount of absorbed photons during photochemical ageing of all materials is shown in the plot (right).
References
ꢀ
1 S. Beaupre, P.-L. T. Boudreault and M. Leclerc, Adv. Mater., 2010,
22, E6–E27.
2 M. Helgesen, R. Søndergaard and F. C. Krebs, J. Mater. Chem., 2010,
20, 36–50.
3 T. D. Nielsen, C. Cruickshank, S. Foged, J. Thorsen and F. C. Krebs,
Sol. Energy Mater. Sol. Cells, 2010, 94, 1553–1571.
4 F. C. Krebs, T. D. Nielsen, J. Fyenbo, M. Wadstrøm and
M. S. Pedersen, Energy Environ. Sci., 2010, 3, 512–525.
5 F. C. Krebs, J. Fyenbo and M. Jørgensen, J. Mater. Chem., 2010, 20,
8994–9001.
6 F.C.Krebs, T.TromholtandM. Jørgensen,Nanoscale, 2010, 2, 873–886.
7 Konarka http://www.konarka.com/, accessed on 19/01/2011.
8 E. Bundgaard and F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2007,
91, 954–985.
9 X. Zhan and D. Zhu, Polym. Chem., 2010, 1, 410–419.
10 Y. Liang and L. Yu, Acc. Chem. Res., 2010, 43, 1227–1236.
11 H.-Y. Chen, J. Hou, S. Zhang, Y. Liang, G. Yang, Y. Yang, L. Yu,
Y. Wu and G. Li, Nat. Photonics, 2009, 3, 649–653.
12 F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2009, 93, 394–412.
13 M. Jørgensen, K. Norrman and F. C. Krebs, Sol. Energy Mater. Sol.
Cells, 2008, 92, 686–714.
Fig. 13 Most stable donor (top) and acceptor units (bottom) identified
within this study.
Fig. 13 provides a list of the most stable donor and acceptor
building blocks we identified. Assembly of these units into larger
aromatic ones is also expected to lead to stable blocks as evi-
denced by the behavior of the dithienothiophene group. Finally,
for the side-chains, a good rule of thumb is to keep their amount
as low as possible whatever their chemical nature.
14 S. Chambon, A. Rivaton, J.-L. Gardette, M. Firon and L. Lutsen, J.
Polym. Sci., Part A: Polym. Chem., 2007, 45, 317–331.
15 M. Manceau, A. Rivaton, J.-L. Gardette, S. Guillerez and
^
N. Lemaitre, Polym. Degrad. Stab., 2009, 94, 898–907.
ꢀ
16 H. Hintz, H.-J. Egelhaaf, H. Peisert and T. Chasse, Polym. Degrad.
Stab., 2010, 95, 818–825.
17 M. Manceau, S. Chambon, A. Rivaton, J.-L. Gardette, S. Guillerez
^
and N. Lemaitre, Sol. Energy Mater. Sol. Cells, 2010, 94, 1572–1577.
Conclusions
ꢀ
18 F. C. Krebs, J. E. Carle, N. Cruys-Bagger, M. Andersen,
M. R. Lilliedal, M. A. Hammond and S. Hvidt, Sol. Energy Mater.
Sol. Cells, 2005, 86, 499–516.
In conclusion, we have mapped the photochemical stability in air
for a wide range of p-conjugated polymers relevant to PSCs. By
comparing the data collected for more than 25 different samples,
various points critical for the polymer stability have been iden-
tified. This enabled us to rationalize how variations in the
chemical structure of p-conjugated polymers impact the photo-
chemical stability. Our results thus provide a better description
of the structure–stability relationship, as well as important
insight that will prove useful to anyone in the process of
designing new materials for PSCs. We are of course fully aware
that the photochemical stability is only one aspect of the complex
PSC stability problem and that stable but inefficient polymers are
rather useless. However, we believe that this study can provide
meaningful help if ones aim is to synthesize new good candidates
for PSCs that unite efficiency and stability.
ꢀ
19 J. E. Carle, J. W. Andreasen, M. Jørgensen and F. C. Krebs, Sol.
Energy Mater. Sol. Cells, 2010, 94, 774–780.
20 M. Helgesen and F. C. Krebs, Macromolecules, 2010, 43, 1253–1260.
21 M. Helgesen, S. A. Gevorgyan, F. C. Krebs and R. A. J. Janssen,
Chem. Mater., 2009, 21, 4669–4675.
22 M. Helgesen, F. C. Krebs, M. Bjerring and N. C. Nielsen, Chem.
Mater., 2010, 22, 5617–5624.
23 E. Bundgaard, O. Hagemann, M. Manceau, M. Jørgensen and
F. C. Krebs, Macromolecules, 2010, 43, 8115–8120.
24 J. Liu, E. N. Kadnikova, Y. Liu, M. D. McGehee and
ꢀ
J. M. J. Frechet, J. Am. Chem. Soc., 2004, 126, 9486–9487.
25 NREL’s AM1.5 Standard Dataset: http://rredc.nrel.gov/solar/
spectra/am1.5/, accessed on 15/09/2010.
26 E. E. Havinga, W. Hoeve and H. Wynberg, Polym. Bull., 1992, 29,
119–126.
27 E. E. Havinga, W. Hoeve and H. Wynberg, Synth. Met., 1993, 55,
299–306.
4140 | J. Mater. Chem., 2011, 21, 4132–4141
This journal is ª The Royal Society of Chemistry 2011

View Online
28 P. M. Beaujuge, C. M. Amb and J. R. Reynolds, Acc. Chem. Res.,
2010, 43, 1396–1407.
35 R. Arnaud, E. Fanton and J.-L. Gardette, Polym. Degrad. Stab.,
1994, 45, 361–369.
€
29 L. J. Lindgren, X. Wang, O. Inganas and M. R. Andersson, Synth.
Met., 2005, 154, 97–100.
36 A. Rivaton, B. Mailhot, G. Derderian, P. O. Bussiere and J.-
L. Gardette, Macromolecules, 2003, 36, 5815–5824.
30 A. Rivaton, B. Mailhot, J. Soulestin, H. Varghese and J.-L. Gardette,
Eur. Polym. J., 2002, 38, 1349–1363.
37 G. Pfister and D. J. Williams, J. Chem. Phys., 1974, 61, 2416–2426.
38 S. Chambon, A. Rivaton, J.-L. Gardette and M. Firon, Sol. Energy
Mater. Sol. Cells, 2008, 92, 785–792.
31 D. M. de Leeuw, M. M. J. Simenon, A. R. Brown and
R. E. F. Einerhand, Synth. Met., 1997, 87, 53–59.
32 M. C. Scharber, M. Koppe, J. Gao, F. Cordella, M. A. Loi, P. Denk,
M. Morana, H.-J. Egelhaaf, K. Forberich, G. Dennler, R. Gaudiana,
D. Waller, Z. Zhu, X. Shi and C. J. Brabec, Adv. Mater., 2009, 21,
1–4.
33 C.-H. Chen, C.-H. Hsieh, M. Dubosc, Y.-J. Cheng and C.-S. Hsu,
Macromolecules, 2010, 43, 697–708.
34 M. Manceau, M. Helgesen and F. C. Krebs, Polym. Degrad. Stab.,
2010, 95, 2666–2669.
39 N. Blouin, A. Michaud, D. Gendron, S. Wakim, E. Blair, R. Neagu-
^
Plesu, M. Belletete, G. Durocher, Y. Tao and M. Leclerc, J. Am.
Chem. Soc., 2008, 130, 732–742.
40 J. C. Bijleveld, M. Shahid, J. Gilot, M. M. Wienk and
R. A. J. Janssen, Adv. Funct. Mater., 2009, 19, 3262–3270.
41 C. H. Bamford and R. G. W. Norrish, J. Chem. Soc., 1935, 1504;
R. G. W. Norrish and C. H. Bamford, Nature, 1936, 138, 1016;
R. G. W. Norrish and C. H. Bamford, Nature, 1937, 140, 195;
R. G. W. Norrish, Trans. Faraday Soc., 1937, 33, 1521.
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4132–4141 | 4141