Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/264537279

Sol-gel synthesized, low-temperature processed, reduced molybdenum


peroxides for organic optoelectronics applications

Article in Journal of Materials Chemistry C · May 2014


DOI: 10.1039/C4TC00301B

CITATIONS READS

27 949

10 authors, including:

Antonios M Douvas Maria Vasilopoulou


National Center for Scientific Research Demokritos National Center for Scientific Research Demokritos
69 PUBLICATIONS 1,560 CITATIONS 159 PUBLICATIONS 2,598 CITATIONS

SEE PROFILE SEE PROFILE

Dimitra Georgiadou A. Soultati


Imperial College London National Center for Scientific Research Demokritos
64 PUBLICATIONS 1,554 CITATIONS 51 PUBLICATIONS 1,014 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Fabrication of microfluidic devices by polydimethylsiloxane (PDMS) View project

Active surfaces View project

All content following this page was uploaded by Dimitra Georgiadou on 04 September 2014.

The user has requested enhancement of the downloaded file.


Journal of
Materials Chemistry C
PAPER

Sol–gel synthesized, low-temperature processed,


reduced molybdenum peroxides for organic
Cite this: J. Mater. Chem. C, 2014, 2,
6290 optoelectronics applications†
Antonios M. Douvas,*a Maria Vasilopoulou,a Dimitra G. Georgiadou,a
Anastasia Soultati,ab Dimitris Davazoglou,a Nikolaos Vourdas,a
Konstantinos P. Giannakopoulos,a Athanassios G. Kontos,a Stella Kennouc
and Panagiotis Argitisa

Reduced molybdenum peroxides with varying degrees of reduction were synthesized following a modified
sol–gel peroxo method and the respective films were employed as anode interfacial layers in organic
optoelectronics applications, such as organic light emitting diodes (OLEDs) and organic photovoltaics
(OPVs). The degree of reduction was controlled through both the synthesis route and the thermal
treatment protocol of the obtained films. The films were thoroughly investigated with a variety of
spectroscopic, diffraction, and electron microscopy methods (UV-Vis, FT-IR, XPS, UPS, Raman, XRD,
SEM, and TEM). These films were found to be considerably sub-stoichiometric with a relatively high
Received 13th February 2014
Accepted 23rd May 2014
content of hydrogen. When they were used as anode interfacial layers in OLED and OPV devices, high
efficiencies and adequate temporal stability were achieved. The enhanced hole injection/extraction
DOI: 10.1039/c4tc00301b
properties of the reduced molybdenum peroxide films were attributed to the improved charge transport
www.rsc.org/MaterialsC facilitated through the gap states present in these materials.

layers (HTLs), whereas TiO2, ZnO and ZrO2 have been used as
1. Introduction electron transport layers (ETLs).8 For those applications,
Transition metal oxides (TMOs)1 and their molecular however, the current trend is to synthesize TMO lms using
analogues, polyoxometalates (POMs),2,3 have been extensively solution-based methods owing to lower cost and more conve-
used in a wide variety of applications ranging from catalysis4 to nient implementation in an industrial process, as compared
molecular electronics5 and organic optoelectronic devices.6,7 with vacuum deposition-based methods (such as thermal
More specically, in organic electronic devices, such as organic vacuum evaporation, laser vapor deposition, etc.).
light emitting diodes (OLEDs) and organic photovoltaics Especially for molybdenum oxide – one of the most popular
(OPVs), TMOs have been widely used as charge selective inter- materials for use as an anode interlayer in organic optoelec-
facial layers between the electrodes and the active polymeric tronic devices – four main wet chemistry-based synthetic
layer to enhance charge transport from/to the electrodes. Thus, methods have been hitherto reported in the literature. The rst
tungsten (WO3), molybdenum (MoO3), vanadium (V2O5) and method is based on an acidied aqueous dispersion of
nickel (NiO) oxides have been mainly used as hole transport ammonium molybdate leading to the formation of MoO3 lms,
though with considerable surface roughness.9 The second route
is based on MoO3 nanoparticles dispersed in xylene along with a
a
Institute of Nanoscience and Nanotechnology (INN), National Center for Scientic block copolymer, followed by treatment with oxygen plasma
Research “Demokritos”, 15310 Aghia Paraskevi, Athens, Greece. E-mail: adouvas@ and thermal annealing at relatively high temperature (200  C) of
imel.demokritos.gr; Fax: +30 210 6511723; Tel: +30 210 6503231 the MoO3 lms.10 The third method is based on MoO2(acac)2
b
Department of Chemical Engineering, National Technical University of Athens, 15780
(acac ¼ acetylacetonate) precursors, which undergo hydrolysis
Athens, Greece
c
and condensation to Mo2O5(OH)2(H2O)4 and in a limiting case
Department of Chemical Engineering, University of Patras, 26504 Patra, Greece
† Electronic supplementary information (ESI) available: (1) UV-Vis monitoring of
to [Mo6O19]2 clusters leading nally to the formation of MoOx
ox/red Mo peroxides in solution, (2) FT-IR changes in water content during lms.11,12 Finally, the fourth route is based on a sol–gel method
thermal treatment of ox/red Mo peroxide lms, (3) FT-IR changes during involving dissolution of MoO3 in hydrogen peroxide, and then
thermal treatment of a solution-reduced Mo peroxide lm, (4) Raman study of dispersion with polyethylene glycol and 2-methoxyethanol,
ox/red Mo peroxide lms, (5) XRD characterization of ox/red Mo peroxide lms,
resulting in smooth MoO3 lms aer high temperature (275  C)
(6) optical characterization of ox/red Mo peroxide lms using spectroscopic
ellipsometry, and (7) morphological characterization of ox/red Mo peroxide
thermal treatment.13,14
lms using SEM/TEM. Fig. S1–S7. See DOI: 10.1039/c4tc00301b

6290 | J. Mater. Chem. C, 2014, 2, 6290–6300 This journal is © The Royal Society of Chemistry 2014
Paper Journal of Materials Chemistry C

Three weak points can be discerned in the above solution- 2. Results and discussion
based synthetic methods for molybdenum oxides lms, which
may ultimately decrease their hole-transporting efficiency 2.1 Sol–gel synthesis of ox/red Mo peroxide lms
when they become part of an organic optoelectronic device. The Mo peroxide lms were synthesized through a modied
The rst point is the inadequately homogeneous surface of previously reported sol–gel peroxo method.13,14 The initial per-
molybdenum oxide lms that are usually formed. The second oxo route was introduced by Kurusu,18 but it was established
point is the high temperature thermal treatment along with and subsequently modied by Kudo et al. to synthesize stoi-
additional process steps (such as oxygen plasma treatment) chiometric molybdenum peroxides.19–22 Recently, it has been
that are generally needed aer spin coating of the precursor combined with the sol–gel process to obtain more homoge-
solution. The third and most critical point is that the majority neous lms, while keeping at the same time the advantages of a
of the aforementioned synthetic routes aim to the formation solution-based synthesis method, i.e. composition control,
of oxidized MoO3 lms (a requirement that is usually inter- large area deposition and low cost.13 In the present work, MoO3
related to the high temperature thermal treatment). However, powder is dissolved in H2O2 rich (31%) aqueous solution under
as it has been reported in the recent literature, solution-pro- reux at 80  C. Subsequently, polyethylene glycol (PEG) is mixed
cessed hydrogen molybdenum bronze lms and hydrogen with the molybdenum peroxide solution for the adjustment of
vanadium bronze lms exhibit higher hole-transporting effi- the homogeneity/viscosity of the initial solution. Up to this
ciency than their corresponding oxidized lms when used as stage, the solution exhibits a clear yellow colour, an indication
anode interfacial layers in bulk heterojunction (BHJ) OPVs.15 that the synthesized molybdenum peroxide is in a fully or nearly
Similar results have also been reported recently by our group stoichiometric oxidation state. Then, 2-methoxyethanol (2ME)
for oxygen decient hydrogen molybdenum bronze lms is added to promote the reduction of the synthesized molyb-
deposited by thermal vacuum evaporation.16 In that case, the denum peroxide (aer having removed the excess of H2O2) and
superior hole transporting efficiency of the hydrogen molyb- turning therefore the solution colour into blue. It will be shown
denum bronzes was attributed to their favourable energy level that the amount of added 2ME is critical, since it may promote
alignment with the highest occupied molecular orbital the reduction of molybdenum peroxide in the precursor solu-
(HOMO) of the organic semiconductor owing to the intro- tion. Therefore, oxidized (ox) or reduced (red) Mo peroxides can
duction of gap states near the Fermi level without a signi- be formed by controlling the amount of 2ME. Films are then
cant decrease of molybdenum oxide's work function. spin-coated on different substrates from the synthesized solu-
Furthermore, we showed recently that solution-processed tions and subjected to various annealing protocols to produce
hydrogen molybdenum bronzes resulted in improved power fully oxidized or reduced molybdenum peroxides depending on
conversion efficiency and stability when they were used as the type of the solution (ox/red) and the annealing temperature,
anode interfacial layers in BHJ OPV devices.17 However, no as discussed below. More details on the exact synthesis proce-
detailed characterisation of the properties of the obtained dure and lm preparation are given in the Experimental
material (dened as hydrogen molybdenum bronzes) was section.
presented in that work.
In this article, we aim to make a clear correlation between
the exact synthesis route and the subsequent processing 2.2 Reduction of Mo peroxides
conditions with the optoelectronic, crystalline and morpho- 2.2.1 UV-Vis monitoring of ox/red-Mo peroxide lms. The
logical properties of the obtained lms. It is shown that the reduction of oxidized molybdenum peroxide lms can be easily
degree of reduction, stoichiometry and, consequently, the monitored using UV-Vis absorption spectroscopy through the
electronic characteristics of the synthesized lms can be appearance of a new band in the visible/near IR region in the
precisely tuned both through simple modications in the reduced lms.23 Fig. 1 and 2 depict the absorption spectra of
solution synthesis and also by applying specic lm 100 nm thick Mo peroxide lms, before and aer thermal
annealing protocols. To this end, a combination of spectro- treatment, in order to compare the effect of thermal annealing
scopic, diffraction and electron microscopy characterization in lms derived by two slightly different synthetic routes. In
techniques (UV-Vis, FT-IR, Raman, ellipsometry, XPS, UPS, Fig. 1, the initial product of the synthetic route followed was an
XRD, SEM, and TEM) were employed to provide insight oxidized (ox) Mo peroxide lm, which was reduced only aer
into the properties of the obtained material. Finally, OLED being thermally treated. The second process (Fig. 2) involved
and OPV devices were fabricated to highlight the importance the addition of an excess amount of 2ME in the solution, which
of controlling the exact material composition in optoelec- resulted in the formation of a reduced (red) Mo peroxide lm,
tronic applications. It was found that the devices imple- and then the effect of further annealing was studied. In both
menting as anode interfacial layers reduced molybdenum cases, the reduced lms were reoxidized aer annealing at high
peroxide lms, which were thermally treated at relatively low temperatures.
temperature, demonstrated enhanced efficiency and More specically, Fig. 1a depicts the UV-Vis/NIR absorption
adequate stability. Therefore, it could be substantiated that region of an ox-Mo peroxide lm, which was annealed from RT
solution-processed reduced molybdenum peroxide lms up to 130  C. The rst observation is that the non-treated lm
constitute a viable choice for large area, low-cost organic and the one treated at temperatures up to 60  C are completely
optoelectronic devices. transparent below 500 nm. Furthermore, the band at 356 nm

This journal is © The Royal Society of Chemistry 2014 J. Mater. Chem. C, 2014, 2, 6290–6300 | 6291
Journal of Materials Chemistry C Paper

octahedrally coordinated Mo ions.23–26 Thus, it seems that the


thermal reduction of the ox-Mo peroxide lm at 130  C is
possibly accompanied by a tetrahedral-to-octahedral change in
the coordination of Mo ions.
Fig. 1b shows that upon further increase of annealing
temperature at T $ 160  C, the IVCT band gradually decreases
in intensity and red-shis. The latter changes indicate that the
(thermally) reduced Mo peroxide lm is gradually being ther-
mally reoxidized at T $ 160  C, whereas at 230  C it has been
almost completely reoxidized in accordance with the litera-
ture.18,24 It is also noteworthy that the peroxo group does not
reform upon thermal reoxidation of the (thermally) reduced Mo
peroxide lm. That nding indicates that the thermal reduction
of the peroxo group in ox-Mo peroxide lms is irreversible.
A similar thermal reoxidation was observed in a Mo peroxide
Fig. 1 UV-Vis absorption spectra of an ox-Mo peroxide film upon
lm formed by a reduced solution (i.e. with the addition of
thermal treatment: (a) from RT to 130  C (thermal reduction of the
film); inset: magnification of the UV region and (b) from 100 to 230  C excess of 2ME) and the spectra are shown in Fig. 2. In particular,
(thermal reoxidation). the red-Mo peroxide lm shows that the characteristic IVCT
band considerably blue-shied, i.e. 757 / 728 nm (Fig. 2a),
indicating that it is possibly much more reduced than in the
previous case of thermal reduction of ox-Mo peroxide lms.
Moreover, the characteristic peroxo band at 367 nm vanishes
(Fig. 2a), indicating that the reduction of ox-Mo peroxide in
solution proceeds here also through the reduction of the peroxo
group. Upon thermal treatment of the red-Mo peroxide lm at
200  C, the IVCT band signicantly red-shis (i.e. 728 / 787
nm; Fig. 2b), indicating that the lm has been completely
reoxidized at T > 200  C. Analogous observations have also been
made during thermal treatment of ox/red-Mo peroxides in
solution (ESI, Fig. S1†).
2.2.2 FT-IR monitoring of ox/red-Mo peroxide lms. FT-IR
spectroscopy was performed to investigate the chemical bonds
between molybdenum, oxygen, carbon and other atoms present
in the synthesized materials and to study the structural changes
in Mo peroxide lms induced by reduction through either the
Fig. 2 UV-Vis absorption spectra of a red-Mo peroxide film: (a) in solution synthesis route or the thermal treatment of lms.
comparison with an ox-Mo peroxide film and (b) upon thermal treat-
Initially, an ox-Mo peroxide lm, as obtained from the synthesis
ment from 150 to 300  C (thermal reoxidation of the film). In (b), the
spectra are normalized to demonstrate the red-shift of the NIR peak route, presents the following peaks (spectrum (1), Fig. 3): (a) a
(IVCT band). broad band at 2877 cm1 and a shoulder at 2930 cm1 attrib-
uted to symmetric and antisymmetric stretching vibration of
CH2 groups of PEG, respectively. (b) A strong broad band at 1099
in the oxidized state, which is assigned to the peroxo (O–O) cm1, ascribed to stretching vibration of both C–OH and C–O–C
group,24 gradually decreases in intensity and red-shis to groups of PEG.27 (c) Two strong peaks at 983 and 955 cm1,
375 nm. This is the rst indication that the thermal reduction of attributed possibly to two stretching modes (symmetric and
the ox-Mo peroxide lm proceeds through the reduction of the antisymmetric) of terminal Mo]O bonds indicating the pres-
peroxo group. Upon thermal annealing of the ox-Mo peroxide ence of a MoO2 (O]Mo]O) group similar to the one observed
lm at 100  C, a new broad band at 778 nm appears, which blue- in dioxomolybdenum(VI) complexes.26,28,29 (d) A strong band at
shis by 20 nm (to 757 nm) when further annealed at 130  C. 919 cm1, attributed to the stretching vibration of the peroxo
That band is attributed to the intervalence charge transfer group (O–O), although slightly red-shied (10–15 cm1) in
(IVCT, MoV / MoVI) band indicating that the Mo peroxide lm comparison with molybdenum peroxo complexes and vana-
is thermally reduced. Furthermore, upon thermal treatment at dium peroxo complexes reported in the literature, possibly
T $ 100  C, a new big shoulder at 305 nm appears, whereas because of intermolecular H-bonding interactions.24,34 (e) A
the shoulder at 225 nm increases in intensity (see the inset in weak band at 802 cm1, attributed to Mo–O–Mo bridges, indi-
Fig. 1a). Both of these bands have been ascribed to charge cating that a small percentage of Mo–O–Mo bridges exist in
transfer (CT) bands of the Mo6+]O2 bond with the 305 nm the initial Mo peroxide lm.24,26,30–33 (f) Two bands at 633 and
shoulder being indicative of octahedrally coordinated Mo ions, 520 cm1, attributed possibly to symmetric and antisymmetric
whereas the 225 nm band indicative of both tetrahedrally and stretching modes of the Mo(O2) peroxo group, respectively.

6292 | J. Mater. Chem. C, 2014, 2, 6290–6300 This journal is © The Royal Society of Chemistry 2014
Paper Journal of Materials Chemistry C

spectra (1) and (2) suggests progressive change in conguration


(possibly from tetrahedral to octahedral) of Mo ions up to
220  C.18,24,26 (f) Finally, the number of coordinated water
molecules in the Mo peroxide lm seems to progressively
increase, especially upon heating at a high temperature (T $
170  C; ESI, Fig. S2a and b†).35–37
Analogous structural changes to those observed during the
thermal reduction of an ox-Mo peroxide lm were also detected
in a red-Mo peroxide lm (i.e. reduced through the synthesis
route) and the IR spectra are also presented in Fig. 3. In brief, by
comparing the spectra (3) and (1) (Fig. 3), the following changes
are observed: (a) signicant decrease of the peroxo group bands
at 917, 633, and 517 cm1 and formation of two new bands at
673 and 565 cm1 indicating that the reduction in solution (red-
Fig. 3 FT-IR spectra of an oxidized and a reduced Mo peroxide film Mo peroxide) proceeds also through reduction/destruction of
before and after annealing at 220  C: (1) ox-Mo peroxide as synthe- the peroxo group, and (b) the appearance of a shoulder at 1123
sized, (2) ox-Mo peroxide after annealing at 220  C, (3) red-Mo cm1 indicates that PEG is affected by the reduction of Mo
peroxide as synthesized, and (4) red-Mo peroxide after annealing at
peroxide in solution (possibly PEG is also oxidized in solution).
220  C.
The structural changes that were observed during thermal
treatment of the red-Mo peroxide lm are also depicted in Fig. 3
(see spectrum (4) and Discussion in the ESI†). It is noteworthy
Those bands are also signicantly shied (30–40 cm1) to lower
that both lms obtained at the end of thermal treatment (i.e. at
and higher energy respectively, possibly because of intermo-
220  C) seem to be molybdenum trioxide regardless of the
lecular H-bonding interactions of peroxo groups.34 (g) Infrared
initial lm that was formed (compare spectra (2) and (4), Fig. 3).
peaks, attributed to water (ESI, Fig. S2a and b†), indicate that
Similar results were obtained by using Raman spectroscopy
coordinated water molecules mostly exist in the Mo peroxide
during thermal treatment of ox/red-Mo peroxide lms (ESI,
lm, whereas a small percentage of adsorbed water molecules is
Fig. S3†). Further characterisation (crystal structure, optical
also present. From the above FT-IR and UV-Vis results, the
properties, surface and bulk morphology) of the above lms as a
structure of Mo peroxide compounds synthesised in the present
function of annealing temperature is presented in the ESI
work could be written as: MoO2(O2)$H2O or, generally,
(Fig. S4–S7†).
MoO3$H2O2$H2O, dispersed in PEG.18,22,24
Upon annealing an ox-Mo peroxide lm from RT to 220  C
(spectrum (2), Fig. 3) the following infrared changes are 2.3 Chemical structure and electronic characterization of ox/
observed: (a) almost vanishing of the PEG vibration bands at red-Mo peroxide lms
2930, 2877, and 1099 cm1 indicating that PEG is removed upon
2.3.1 Determination of stoichiometry using XPS. Next,
heating. (b) Disappearance of the peroxo group bands at 919,
X-ray photoelectron spectroscopy (XPS) analysis of lms
633, and 520 cm1 giving clear evidence that the peroxo group is
deposited from either the oxidized or the reduced solution and
destroyed upon thermal treatment. These results are in agree-
annealed at different temperatures was performed in order to
ment with the previous UV-Vis absorption ndings showing
investigate the chemical composition of the outermost lm
that the thermal reduction of an ox-Mo peroxide lm proceeds
layers. In the wide scan spectra (not shown), no appreciable
mainly through the reduction/destruction of the peroxo group
charging effects were observed and only carbon, oxygen and
and that during thermal reoxidation of the lm at a higher
molybdenum related signals were detected on the lms'
temperature (T > 130  C), the peroxo group does not reform. (c)
surface. The presence of carbon may be due to surface
Vanishing of one of the two Mo]O stretching mode bands
contamination but it might be also related to incomplete
(preferably the symmetric stretching mode band at 983 cm1),
decomposition (such as PEG) of the precursor solution.
resulting in the formation of a broad Mo]O stretching band
Despite the presence of contaminants, for the lms depos-
with intermediate position (967 cm1). (d) The weak band of
ited from the oxidized solution and heat treated at 220  C
Mo–O–Mo bridges at 802 cm1 slowly vanishes and a very broad
(Fig. 4a and Table 1), the Mo 3d surface signal was typical of Mo
band at 852 cm1 appears, indicating that Mo–O–Mo bridges
in the VI oxidation state. In fact, the Mo 3d5/2 and Mo 3d3/2
possibly form during annealing of the lm. (e) Two new bands
binding energy (BE) values at 232.7 (FWHM ¼ 1.40 eV) and
at 679 and 563 cm1 appear during thermal treatment possibly
235.8 eV (FWHM ¼ 1.47 eV) can be attributed to molybdenum
coming from the corresponding bands of the peroxo group (at
cations being at the highest oxidation state (Mo6+) in accor-
633 and 520 cm1 respectively), indicating that possibly new
dance with the literature.16,17,38–41 It is evident that in this case
Mo–O–Mo bridges form as a result of peroxo groups' cleavage.
the lm exists almost exclusively as MoO3. In the case of lms
The gradual decrease of the number of Mo]O stretching bands
deposited also from the oxidized solution and heat treated at a
(from 2 to 1) in combination with the increase of Mo–O–Mo
lower temperature (i.e. at 170  C; Fig. 4b), the energy peaks were
bridges observed upon annealing of an ox-Mo peroxide lm in
lower and broader than before. The lower and broadened energy

This journal is © The Royal Society of Chemistry 2014 J. Mater. Chem. C, 2014, 2, 6290–6300 | 6293
Journal of Materials Chemistry C Paper

percentage of Mo5+ cations was estimated to be 31% of the


overall Mo states, whereas no evidence for the existence of the
lowest oxidation states Mo4+ was obtained.
The O 1s peaks of the XPS spectra of the above lms are also
presented in Fig. 5. All lms present a broad peak which is
further broadened to higher BEs for those deposited from the
reduced solutions. Deconvolution of this peak allowed to
identify the presence of an intense peak at 530.8 eV, which can
be attributed to the O 1s binding energy of Mo–O bonds,16,17,40,42
and of two smaller broad peaks at 531.6 and 532.3 eV.
Regarding the origin of these peaks, they can be assigned to

Fig. 4 Mo 3d XPS spectra of (a and b) an ox-Mo peroxide film and (c


and d) a red-Mo peroxide film. The films were thermally treated at (a
and c) 220  C and (b and d) 170  C.

peaks were associated with the lower oxidation state of the Mo


cations, which tends to result in lower BE.41–44 From the
deconvolution of these peaks it was proven that a new doublet
appeared at 231.4 eV (FWHM ¼ 1.45 eV) and 234.6 eV (FWHM ¼
1.50 eV) with a spin–orbit split of 3.2 eV, which was attributed to
Mo 3d5/2 and Mo 3d3/2, respectively, of Mo cations at a lower
oxidation state (Mo5+). The percentage of the Mo5+ oxidation
state to the total Mo cations (Mo5+/Mo5+/6+) was estimated to
be 20%.
For lms deposited from the reduced (blue) solutions and
thermally treated at 220  C, the Mo 3d signal resulted more
broadened towards lower BEs (Fig. 4c), again suggesting the
presence of Mo oxidation states lower than VI. Indeed, decon-
volution of the latter spectrum allowed identifying the presence
of two distinct doublets as in the case of the ox-Mo peroxide lm
annealed at 170  C, attributed to Mo species being at 6+ and 5+
oxidation states. In that case, the percentage of Mo5+ cations
was estimated to be 27% of the overall Mo states. Finally,
when the red-Mo peroxide lm was thermally treated at 170  C
(Fig. 4d), the Mo 3d peaks were further decreased and broad- Fig. 5 O 1s XPS spectra of (a and b) an ox-Mo peroxide film and (c and
d) a red-Mo peroxide film. The films were thermally treated at (a and c)
ened evidence for further reduction of the Mo species. The
220  C and (b and d) 170  C.

Table 1 Electronic properties of ox/red-Mo peroxide films synthesized using a slightly modified sol–gel peroxo method

Mo peroxide formula Synthesized solutiona Tb ( C) Mo5+/(Mo5+/6+)c (%) Peroxo + OH/Ototald (%) WFe (eV) VBMe (eV)

MoO3 Ox 220 0 17 5.6 3.0


HxMoO2.9 Ox 170 20 21 5.3 3.0
HxMoO2.75 Red 220 27 29 5.3 3.0
HxMoO2.45 Red 170 31 42 5.2 2.7
a
Oxidized or reduced Mo peroxide in the precursor solution. b Temperature of lm thermal treatment. c Results obtained from Mo 3d XPS spectra.
d
Results obtained from O 1s XPS spectra. e Results obtained from UPS spectra.

6294 | J. Mater. Chem. C, 2014, 2, 6290–6300 This journal is © The Royal Society of Chemistry 2014
Paper Journal of Materials Chemistry C

hydroxyl (–OH) and to peroxo groups, respectively, bound to Mo VBM and WF (3.0 and 5.3 eV, respectively) were also obtained,
atoms,42 while the former peak may have a contribution from when a red-Mo peroxide lm was thermally treated at 220  C
adsorbed water molecules. The relative intensity of this peak is (Fig. 6c). The only difference in that case is the signicant
quite large and becomes larger in the lms obtained from the increase in the density of occupied gap states located at 2.2 eV,
reduced solution and/or heat treated at lower temperatures, near the Fermi level, indicating that molybdenum oxide is
which argues against surface contamination and represents further reduced. Finally, when a red-Mo peroxide lm was
rather strong evidence of stable Mo-peroxide bonds. From the annealed at 170  C (Fig. 6d) a considerable decrease in both
Mo 3d and O 1s XPS spectra the materials’ stoichiometry was VBM and WF (2.7 and 5.2 eV, respectively), accompanied by a
estimated and is presented in Table 1. further increase of the density of gap states located near the
2.3.2 Valence band prole analysis with UPS. Next, the UPS Fermi level, was observed.
photoemission spectra of ox/red-Mo peroxide lms were
analyzed (Fig. 6). The UPS spectrum of an ox-Mo peroxide lm
annealed at 220  C is structureless with a single peak indicating
an amorphous (defect-free) MoO3 lm (Fig. 6a and Table 1).16,17 2.4 Mo peroxide lms as anode interfacial layers in organic
The valence band of MoO3 consists mainly of O 2p orbitals with optoelectronic devices
a maximum (VBM) at 3.0 eV below the Fermi level as expected 2.4.1 Performance of OLEDs. Aer having established a
for a fully stoichiometric Mo oxide (right panel).16,17,43,44 The direct correlation of the preparation conditions of the Mo
high binding energy cut-off of MoO3 is 15.6 eV, corresponding peroxide lms with their electronic structure and physical
to a work function (WF) of 5.6 eV, a value which is very close to properties, those Mo peroxide lms were employed as anode
the values reported in the literature for solution prepared MoO3 interlayers in OLEDs. The fabricated device structure was ITO/
lms.15,17 Thus, the ionization energy (IE) is estimated to be Mo-peroxide/F8BT (emissive layer)/PW12/Al, where PW12 is 12-
8.6 eV. When an ox-Mo peroxide lm was annealed at a lower phosphotungstic acid (H3PW12O40), a polyoxometalate (POM)
temperature (i.e. 170  C; Fig. 6b) the VBM was practically previously reported by our group as a highly efficient electron
constant at 3.0 eV below the Fermi level, whereas the high BE injection/extraction layer (EIL/EEL) in OLEDs and OPVs,7,45
cut-off increased (at 15.9 eV) resulting in a WF value of 5.3 eV. It which can be cast from alcohol-based solvents, that are
is noteworthy that in this case, the formation of a broad set of orthogonal with regard to the polymeric layer underneath.
gap states near the Fermi level, located at 2.2 eV, can be Reference devices based on poly(3,4-ethylenedioxythio-
observed. Those gap states may be attributed to intervalence phene) : poly(styrene sulfonate) (PEDOT-PSS) for hole injection
electrons transferred to the initially empty Mo 4d orbitals of the were also fabricated for comparison purposes.
antibonding p* band near the Fermi level during the thermal In Fig. 7a, the current density–voltage–luminance ( J–V–L)
reduction of peroxo groups.16,43 Similar results concerning the and in Fig. 7b, the current efficiency–voltage characteristics for
devices with an optimized thickness of 20 nm Mo-peroxide lm
or 40 nm PEDOT-PSS as the anode interlayer and a 2 nm thin
polyoxometalate (PW12) as the cathode interlayer are shown.
The main results are summarized in Table 2. In the device based
on the ox-Mo peroxide lm (MoO3, thermally processed at
220  C), an overall low light emission (the peak luminance was
7000 cd m2 at current density equal to 1800 A m2) and device
efficiency (3.9 cd A1) is observed. Slightly higher efficiency
(4.2 cd A1) is obtained for the material processed from the
oxidized solution that was thermally treated at 170  C (namely
HxMoO2.9).
On the other hand, for the devices incorporating reduced Mo
peroxide lms, high peak current densities and luminances are
obtained. The device with the reduced HxMoO2.75 (thermally
treated at 220  C) reaches luminance values up to 34 000 cd m2
and current densities of 3200 A m2. As a result, the peak
current efficiency rises to 10.6 cd A1, representing a more than
100% improvement relative to the device with the ox-Mo
peroxide interlayer. Moreover, the device with the highest
degree of reduction and hydrogen content, namely HxMoO2.45
(i.e. red-Mo peroxide heated at 170  C), exhibits a slightly
increased performance with efficiency reaching 10.9 cd A1.
Note that the reference device with PEDOT-PSS at its anode
Fig. 6 UPS photoemission spectra of (a and b) an ox-Mo peroxide film
interface exhibits luminance values up to 9200 cd m2 and
and (c and d) a red-Mo peroxide film. The films were thermally treated current densities of 2300 A m2, resulting in a maximum
at (a and c) 220  C and (b and d) 170  C. current efficiency of 4 cd A1.

This journal is © The Royal Society of Chemistry 2014 J. Mater. Chem. C, 2014, 2, 6290–6300 | 6295
Journal of Materials Chemistry C Paper

exhibit a work function (WF) of the order of 5.2–5.6 eV which is


sufficient enough to reduce the barrier for hole injection from
the metallic anode to the organic active layer (namely F8BT
having an ionization energy IE in the order of 5.0 eV)16 through
the formation of an interfacial dipole with its positive pole
pointing towards the anode contact.8 However, WF alone cannot
explain the efficiency enhancement. For instance, the device
based on ox-MoO3 with the higher WF values exhibits poor
performance in comparison with the devices based on red-Mo
peroxides. To explain this, we have to take into account that
fully stoichiometric metal oxides are wide band-gap semi-
conductors, adding a large series resistance to the device. In
contrast, the red-Mo peroxides exhibit occupied gap states near
the energy level of the highest occupied molecular orbital
(HOMO) of the active layer and thus may facilitate hole trans-
port. The increased performance of devices based on the red-Mo
peroxides may thus be attributed to the large density of occu-
pied gap states near their Fermi level.
Furthermore, the stability of the devices operating in air was
tested in a period of 700 hours. Fig. 7c depicts that all the Mo
peroxide lm-based devices showed better stability when
compared with the device using a PEDOT-PSS layer for hole
injection. This is a very promising result, making those mate-
rials attractive candidates for future industrial implementation,
since the temporal stability is an important issue in organic
optoelectronic devices.
2.4.2 Hole only devices. For an in-depth understanding on
how the Mo-peroxide’s chemical composition affects the device
operation and supports our argument about the enhanced hole
transport within the device, we next fabricated hole only
devices. To this end, we replaced the PW12/Al cathode with a
simple Au electrode, which causes a large barrier for electron
injection/extraction to/from the LUMO of F8BT from/to the
electrode. Therefore, the resulting current is expected to be
dominated by holes. In Fig. 8, the corresponding dark J–V
Fig. 7 (a) Current density–voltage (solid symbols) and luminance–
characteristics of devices with the structure ITO (anode)/Mo-
voltage (open symbols) characteristic curves of OLEDs with the
structure ITO/Mo-peroxide (20 nm) or PEDOT-PSS (40 nm)/F8BT/ peroxide (20 nm) or PEDOT-PSS (40 nm)/F8BT (70 nm)/Au
PW12 (2 nm)/Al devices. (b) Current efficiency–voltage curves of the (cathode) are shown in a semi-log scale. From the J–V charac-
same OLED devices. (c) Stability measurements for non-encapsulated teristics, it is evident that the hole current density for the
OLEDs based on Mo-peroxides as HILs are presented as plots of devices with the red-Mo peroxide layers is about one order of
luminance versus time of operation in air at room temperature under a
magnitude higher than the corresponding current in the
constant driving current density of 200 A m2.
devices with either the ox-Mo peroxide or the PEDOT-PSS layer.
Most signicantly, the devices with the red-Mo peroxide layers
The efficiency enhancement of the devices using Mo-perox- show space-charge limited current (SCLC) behaviour, as veried
ides, and especially the reduced ones at the anode interface, can by the overlap of the experimental data with the calculated J–V
be explained as follows: as discussed above these materials characteristics using the equation J ¼ 93o3rmV2/8L3 (where J is

Table 2 Device characteristics (and statistics obtained from a batch of 24 devices) of OLEDs having the structure ITO/Mo-peroxide or PEDOT-
PSS (HIL)/F8BT/PW12 (EIL)/Al

Anode interfacial Vturn-on (V) Max. current efficiency


layer (at 10 cd m2) Jmax (A m2) Lmax (cd m2) (cd A1) [at voltage]

MoO3 2.7 1800 (0.5) 7000 (0.5) 3.9 (0.2) (5.0 V)


HxMoO2.9 2.6 2500 (0.5) 10 500 (0.5) 4.2 (0.2) (5.0 V)
HxMoO2.75 2.6 3200 (0.5) 34 000 (0.5) 10.6 (0.2) (4.5 V)
HxMoO2.45 2.6 3400 (0.5) 37 000 (0.5) 10.9 (0.2) (4.5 V)
PEDOT-PSS 2.6 2300 (0.5) 9200 (0.5) 4.0 (0.2) (4.5 V)

6296 | J. Mater. Chem. C, 2014, 2, 6290–6300 This journal is © The Royal Society of Chemistry 2014
Paper Journal of Materials Chemistry C

Fig. 8 Current density–voltage (J–V) curves of hole only devices ITO/


Mo-peroxide (20 nm) or PEDOT-PSS (40 nm)/F8BT (70 nm)/Au
(cathode). The SCLC was also calculated.

the current density, V is the applied voltage, 3o3r is the dielectric


permittivity, m is the carrier mobility, and L is the thickness of
the F8BT active layer) of the SCLC model at low applied voltages.
From Fig. 8, it can also be seen that the hole mobility is
increased in devices with the red-Mo peroxide layers. The
formation of an ohmic contact at the anode interface, which
highly facilitates hole injection/transport, can also be deduced
and attributed to the energy level alignment due to both the
formation of a favourable interfacial dipole and the occupation
of gap states near the device Fermi level.
2.4.3 Performance of OPVs. To further highlight the impact
of the above ndings on the operation of organic electronic
devices, OPV cells incorporating ox/red-Mo peroxide layers as
hole-selective interlayers were fabricated. The device architec- Fig. 9 (a) Photocurrent–voltage and (b) dark current–voltage (J–V)
ture was ITO/Mo-peroxide (20 nm)/active layer (100 nm)/PW12 curves of organic cells with the structure ITO/Mo-peroxide (20 nm)/
(5 nm)/Al. The active layer was based on the bulk heterojunction P3HT:PC71BM (100 nm)/PW12 (5 nm)/Al. (c) The IPCEs of the same
devices.
blend of the P3HT polymer donor and the PC71BM fullerene
acceptor.46 In Fig. 9a, the current density versus voltage ( J–V)
characteristics of the OPV devices under AM 1.5G illumination
processed from the reduced solution exhibits a larger turn-on
are presented. It is evident that the incorporation of Mo
voltage (in the range of 0.55–0.65 V) relative to that with the
peroxide layers induces large improvement in short-circuit
Mo peroxide layers processed from the oxidized solution (0.45–
current density (Jsc) from 8.3 mA cm2 for the devices with the
0.55 V). This implies that the built-in potential (Vbi) across the
ox-MoO3 anode interlayer to 10.2 and 10.4 mA cm2 (25%
device, which is the voltage responsible for exciton dissociation
improvement) for the devices with the reduced HxMoO2.75 and
and the upper limit for the attainable Voc, is increased upon the
HxMoO2.45 peroxide layers, respectively (the maximum recorded
insertion of the Mo peroxide layer. In addition, a large increase
and mean values of devices' operational characteristics are
of the diode forward current and a small suppression of the
summarized in Table 3). Similarly, the ll factor (FF) is
reverse current aer the incorporation of the Mo peroxide lms
improved from 0.52 to a maximum of 0.58 and the open-circuit
are observed. This proves that the introduction of these layers
voltage (Voc) increases from 0.59 V to 0.66 V in the devices with
creates a highly conductive path for hole-extraction, which can
the reduced Mo-peroxide layers. As a consequence, a high PCE
also explain the enhanced FF of these devices.
of 3.9% and slightly higher of 4.0% was estimated for the
Fig. 9c compares the incident photon-to-current collection
devices based on the reduced HxMoO2.75 and HxMoO2.45
efficiency (IPCE) spectra of the above devices. The devices with
peroxides, respectively. Again, the devices with the red-Mo
the MoO3 and HxMoO2.9 layers show the typical spectral
peroxide lms exhibit the largest increase in performance,
response of P3HT:PC71BM blends with a maximum IPCE of
probably due to the increase of the density of occupied gap
57% and 59%, respectively, which are consistent with previous
states, as discussed above.
studies for P3HT:PC71BM devices exhibiting similar short-
The above ndings are further supported by the dark current
circuit currents and overall external quantum efficiencies.46 For
versus voltage characteristics presented in Fig. 9b. It can be
the devices with the reduced HxMoO2.75 and HxMoO2.45 inter-
clearly seen that the device with the Mo peroxide layers
layers, the results demonstrate an enhancement of 25%

This journal is © The Royal Society of Chemistry 2014 J. Mater. Chem. C, 2014, 2, 6290–6300 | 6297
Journal of Materials Chemistry C Paper

Table 3 Device characteristics of P3HT:PC71BM OPV cells with the structure ITO/Mo-peroxide/P3HT:PC71BM/PW12/Ala

Anode interfacial layer Jsc (mA cm2) Voc (V) FF PCE (%) Rs (U cm2)

MoO3 8.3 (7.95  0.25) 0.59 (0.57  0.02) 0.52 (0.51  0.026) 2.6 (2.4  0.18) 18
HxMoO2.9 8.8 (8.61  0.2) 0.63 (0.61  0.02) 0.55 (0.54  0.021) 3.1 (3.0  0.14) 15
HxMoO2.75 10.2 (10.14  0.10) 0.66 (0.65  0.01) 0.58 (0.58  0.014) 3.9 (3.8  0.11) 5
HxMoO2.45 10.4 (10.32  0.16) 0.66 (0.65  0.01) 0.58 (0.57  0.014) 4.0 (3.9  0.13) 4
a
Data and statistics based on 24 cells of each type. Numbers in bold are the maximum recorded values.

(reaching maximum values of 78% at the wavelength of 510 the sol–gel peroxo route proposed by Lin.13 In detail, 3 g of
nm in both cases) in the IPCE over the entire spectral region, molybdenum oxide powder (99%, Panreac) was dissolved in
where the photoactive layer absorbs. These values are correlated 20 mL of hydrogen peroxide (H2O2, 31%, Merck). The solution
with the corresponding enhancements in short-circuit currents was reuxed at 80  C for 2 h and then cooled to room temper-
measured in the red-Mo peroxide based devices relatively to the ature (RT) for 24 h. The MoO3 powder was completely dissolved
ox-Mo peroxide based device. The increased IPCEs in the devices and a clear yellow solution was obtained. Subsequently, poly-
incorporating such red-Mo peroxide layers are mainly attributed ethylene glycol (PEG, Mn 400, Merck) was added in a varying
to corresponding enhanced extraction rates through the occu- volume ratio of MoO3-sol : PEG (optimal ratio, 1 : 0.25) in order
pied gap states of those layers, which are not present in the ox- to improve the lm forming properties due to better homoge-
Mo peroxide. Thus, it is evident that a direct correlation between neity of the resultant solution. The solution was reuxed at 70  C
the Mo peroxide composition/electronic structure and the for 0.5 h and then cooled to RT. Next, 2-methoxyethanol (2ME,
device performance can be accomplished. 98%, Ferak Berlin) was added in a varying volume ratio of MoO3-
sol : 2ME (1 : 1, 1 : 2, 1 : 3, 1 : 4, and 1 : 6; optimal ratio: 1 : 3),
3. Conclusions the solution was reuxed at 60  C for 0.5 h and then cooled to
RT. The resultant solutions displayed intense blue colour for
Reduced molybdenum peroxide lms with varying reduction increased contents of 2ME aer having evaporated the excess of
degrees were synthesized using a modied sol–gel peroxo H2O2. The molybdenum peroxide solutions were spin coated on
method with the critical steps being the addition of an appro- the substrate at a spinning rate of 1500–6000 rpm for 40 s giving
priate amount of 2-methoxy ethanol to the precursor solution approximately 20–200 nm thick lms (optimal thickness for
and the annealing of deposited lms in a broad temperature device operation is approximately 20 nm). Prior to that step, spin
range. Those lms were then characterized in detail and nally coating of 2ME (at 6000 rpm) was found to improve adhesion
used as anode interfacial layers in organic optoelectronics and quality of Mo peroxide lms. Coated Mo peroxide lms were
applications, such as F8BT based-OLEDs and OPVs based on heated on a hot plate at 60–360  C for 5–15 min. The heat
P3HT:PC71BM as the photoactive layer. From spectroscopic treatment was performed in air since there was no indication
measurements it was found that the reduction of molybdenum that oxygen affects the reduction of Mo compounds.
peroxides both in solution (during the synthesis route aer the
addition of an excess of alcohol) and in lms (during annealing)
4.2 Device fabrication
proceeds mainly through the reduction of the peroxo group. The
gradual removal of PEG during the thermal treatment of the Mo For both OLED and OPV fabrication, glass/indium tin oxide
peroxide lms was also evidenced. The chemical composition of (ITO) substrates (20 U cm1) were used as anode electrodes. The
the surface of each Mo peroxide lm was determined by XPS substrates were ultrasonically cleaned with a sequence of
based on Mo 3d and O 1s core level spectra and showed that a solutions/solvents (0.026 M aqueous solution of tetramethyl-
higher degree of reduction can be obtained through the addi- ammonium hydroxide (TMAH), water, acetone, and isopropyl
tion of an appropriate amount of alcohol during synthesis, alcohol, for 15 min each). In OLEDs, on top of the ox/red-Mo
whereas ne tuning of the stoichiometry can be made by peroxide lm, a 70 nm thick emissive layer (EML) of the green-
decreasing the lm annealing temperature. Finally, it was found emitting copolymer, poly[(9,9-dioctyluorenyl-2,7-diyl)-co-1,4-
that when solution-reduced Mo peroxide lms were used as benzo-{2,1,3}-thiadiazole] (F8BT, American Dye Source) was
anode interfacial layers, a signicant improvement of OLED spin coated from a chloroform solution (6 mg mL1) and
and OPV device characteristics was achieved. This is due to the thermally treated at 80  C for 10 min. Then, a thin electron
enhancement in charge transport induced by the occupation of injection/transport layer (EIL/ETL) of polyoxometalate (POM,
gap states in the reduced Mo peroxide lms. 12-tungstophosphoric acid, H3PW12O40, Sigma-Aldrich) was
spin-coated from a methanol solution (10 mg mL1). Subse-
4. Experimental quently, a 150 nm thick Al cathode layer was deposited in a
dedicated thermal evaporator under a high vacuum of 106
4.1 Materials Torr. Reference devices were also prepared with poly(3,4-ethyl-
4.1.1 Synthesis of ox/red-molybdenum peroxide lms. Ox/ enedioxythiophene) : poly(styrene sulfonate) (PEDOT-PSS, Her-
red molybdenum peroxide lms were synthesized by modifying aus) as anode interfacial layer. The active area of the fabricated

6298 | J. Mater. Chem. C, 2014, 2, 6290–6300 This journal is © The Royal Society of Chemistry 2014
Paper Journal of Materials Chemistry C

devices was 0.1256 cm2. In OPVs, the active layer was a 100 nm unit. Luminance and electroluminescence (EL) spectral char-
thick lm of a blend (1 : 0.8 wt% ratio) of poly(3-hexyl thio- acteristics were recorded with an Ocean Optics USB 2000 ber
phene-2,5-diyl) (P3HT, 98% RR, Mw 54 000–75 000, Rieke optic spectrophotometer, assuming a Lambertian emission
Metals) and [6,6]-phenyl-C71-butyric acid methyl ester (PC71BM, prole (for luminance measurements). For the stability
Sigma-Aldrich), spin-coated from a chloroform solution (10 mg measurements the devices were stored in nitrogen between
mL1) and annealed at 135  C for 10 min. measurements. Current density–voltage characteristics of the
fabricated solar cells were measured with a Keithley 2400
4.3 Methods source-measure unit. Cells were illuminated with a Xe lamp and
an AM 1.5G lter to simulate solar light illumination conditions
4.3.1 Spectroscopic methods. UV-Vis absorption spectra of
with an intensity of 100 mW cm2 (1 sun), as was recorded with
(ox/red) molybdenum peroxides in solution and lms, along
a calibrated silicon photodiode. Incident photon-to-current
with polymer lms, were obtained on quartz cuvettes (with 1 cm
efficiencies (IPCE) were measured on an Autolab PGSTAT-30
path length) and quartz slides (2  2 cm2), respectively, using a
potentiostat, using a 300 W Xe lamp in combination with an
Perkin-Elmer UV-Vis Lambda 40 spectrophotometer. FT-IR
Oriel 1/8 monochromator for dispersing the light in an area of
transmittance spectra of Mo peroxide lms were obtained on
0.5 cm2. A Thorlabs silicon photodiode was used for the cali-
silicon wafers (at 4 cm1 resolution and 128 scans) using a
bration of the IPCE spectra. Film preparation and device
Bruker Tensor 27 FT-IR spectrometer with a DTGS detector.
measurements were performed in air.
Chemical analysis of all Mo peroxide lms was performed by
XPS measurements in an ultra-high vacuum VG ESCALAB210.
The spectra were obtained aer excitation using Mg Ka (1253.6 Acknowledgements
eV) radiation of a twin anode in a constant analyzer energy mode
This research was co-nanced by the European Union (Euro-
with a pass energy of 30 eV. All binding energies were referred to
pean Social Fund – ESF) and Greek national funds through the
the C 1s peak at 284.8 eV and to the O 1s peak at 530.2 eV of the
Operational Program “Education and Lifelong Learning” of the
surface adventitious carbon and oxygen, respectively. The Mo-
National Strategic Reference Framework (NSRF) – Research
peroxides' stoichiometry was estimated using the XPS-measured
Funding Program: ARCHIMEDES III. Investing in knowledge
Mo 3d core levels, respectively, and the corresponding O 1s core-
society through the European Social Fund.
level spectra. To this extent, the areas under the photoemission
peaks were integrated by tting the O 1s and Mo 3d spectra with
asymmetric Gaussian–Lorentzian curves. The peroxo and –OH References
groups versus total oxygen content were estimated by tting the
1 S. W. Boettcher, J. Fan, C.-K. Tsung, Q. Shi and G. D. Stucky,
O 1s core levels. The raw data, aer a Shirley background
Acc. Chem. Res., 2007, 40, 784.
subtraction, were tted by a non-linear least squares routine
2 D.-L. Long, E. Burkholder and L. Cronin, Chem. Soc. Rev.,
using peaks with a mix of Gaussian–Lorentzian shape. The error
2007, 36, 105.
is estimated at 10% in all the XPS-derived atomic percentages.
3 M. T. Pope, Inorg. Chem., 1972, 11, 1973.
The valence band spectra of the metal oxides were evaluated
4 D. P. Debecker, R. Delaigle, K. Bouchmella, P. Eloy,
aer recording the UPS spectra of about 10 nm thick lms
E. M. Gaigneaux and P. H. Mutin, Catal. Today, 2010, 157,
deposited on an ITO substrate. For the UPS measurements, the
125.
He I (21.22 eV) excitation line was used. A negative bias of 12.28
5 A. M. Douvas, E. Makarona, N. Glezos, P. Argitis,
V was applied to the samples during UPS measurements in order
J. A. Mielczarski and E. Mielczarski, ACS Nano, 2008, 2, 733.
to separate sample and analyzer high binding energy (BE) cut-
6 J. Meyer, S. Hamwi, M. Krögger, W. Kowalsky, T. Riedl and
offs and estimate the absolute work function value from the
A. Kahn, Adv. Mater., 2012, 24, 5408.
high BE cut-off region of the UPS spectra. The analyzer resolu-
7 L. C. Palilis, M. Vasilopoulou, A. M. Douvas,
tion is determined from the width of the Au Fermi edge to be
D. G. Georgiadou, S. Kennou, N. A. Stathopoulos,
0.16 eV. Regarding the UPS measurements, because the high
V. Constantoudis and P. Argitis, Sol. Energy Mater. Sol.
intensity UV photons used for UPS measurements may alter the
Cells, 2013, 114, 205.
surface of lms, a certain protocol for the measurement of the
8 M. T. Greiner, L. Chai, M. G. Helander, W.-M. Tang and
work function was adopted: rst the core levels and work
Z.-H. Lu, Adv. Funct. Mater., 2013, 23, 215.
function of the sample were measured using the low intensity
9 F. Liu, S. Shao, X. Guo, Y. Zhao and Z. Xie, Sol. Energy Mater.
X-rays, then the work function was measured using UPS, and
Sol. Cells, 2010, 94, 842.
nally the core levels and work function were measured again
10 J. Meyer, R. Khalandovsky, P. Görrn and A. Kahn, Adv.
using XPS to verify that the samples remained unaffected. The
Mater., 2011, 23, 70.
thickness of both Mo peroxide and active polymer lms was
11 M. R. Pedrosa, R. Aguado, V. Dίez, J. Escribano, R. Sanz and
measured using an Ambios technology XP2 prolometer.
F. J. Arnáiz, Eur. J. Inorg. Chem., 2007, 3952.
12 J. J. Jasieniak, J. Seier, J. Jo, T. Mates and A. J. Heeger, Adv.
4.4 Device characterization Funct. Mater., 2012, 22, 2594.
Current density–voltage (J–V) characteristics of the OLED 13 S.-Y. Lin, C.-M. Wang, K.-S. Kao, Y.-C. Chen and C.-C. Liu,
devices were measured with a Keithley 2400 source-measure J. Sol-Gel Sci. Technol., 2010, 53, 51.

This journal is © The Royal Society of Chemistry 2014 J. Mater. Chem. C, 2014, 2, 6290–6300 | 6299
Journal of Materials Chemistry C Paper

14 C. Girotto, E. Voroshazi, D. Cheyns, P. Heremans and 31 C. Rocchiccioli-Deltcheff, M. Fournier, R. Franck and


B. P. Rand, ACS Appl. Mater. Interfaces, 2011, 3, 3244. R. Thouvenot, Inorg. Chem., 1983, 22, 207.
15 F. Xie, W. C. H. Choy, C. Wang, X. Li, S. Zhang and J. Hou, 32 N. Mizuno, K. Katamura, Y. Yoneda and M. Misono, J. Catal.,
Adv. Mater., 2013, 25, 2051. 1983, 83, 384.
16 M. Vasilopoulou, A. M. Douvas, D. G. Georgiadou, 33 K. Eguchi, Y. Toyozawa, N. Yamazoe and T. Seiyama, J.
L. C. Palilis, S. Kennou, L. Sygellou, A. Soultati, I. Kostis, Catal., 1983, 83, 32.
G. Papadimitropoulos, D. Davazoglou and P. Argitis, J. Am. 34 H. Mimoun, L. Saussine, E. Daire, M. Postel, J. Fischer and
Chem. Soc., 2012, 134, 16178. R. Weiss, J. Am. Chem. Soc., 1983, 105, 3101.
17 A. Soultati, A. M. Douvas, D. G. Georgiadou, L. C. Palilis, 35 S. E. Horsley, D. V. Nowell and D. T. Stewart, Spectrochim.
J. M. Feckl, S. Gardelis, M. Fakis, S. Kennou, P. Falaras, Acta, Part A, 1974, 30, 535; L. Bertsch and H. W. Habgood,
T. Stergiopoulos, D. Davazoglou, P. Argitis and J. Phys. Chem., 1963, 67, 1621.
M. Vasilopoulou, Adv. Energy Mater., 2014, 4, DOI: 10.1002/ 36 M. F. Daniel, B. Desbat, J. C. Lassegues, B. Gerand and
aenm.201300896. M. Figlarz, J. Solid State Chem., 1987, 67, 235.
18 Y. Kurusu, Bull. Chem. Soc. Jpn., 1981, 54, 293. 37 M. F. Daniel, B. Desbat, J. C. Lassegues and R. Garie, J. Solid
19 T. Kudo, Nature, 1984, 312, 537. State Chem., 1988, 73, 127.
20 K. Yamanaka, H. Oakamoto, H. Kidou and T. Kudo, Jpn. J. 38 W. Grünert, A. Y. Stakheev, R. Feldhaus, K. Anders,
Appl. Phys., Part 1, 1986, 25, 1420. E. S. Shpiro and K. M. Minachev, J. Phys. Chem., 1991, 95,
21 K. Hinokuma, A. Kishimoto and T. Kudo, J. Electrochem. Soc., 1323.
1994, 141, 876. 39 M. T. Greiner, M. G. Helander, W.-M. Tang, Z.-B. Wang,
22 K. Hinokuma, K. Ogasawara, A. Kishimoto, S. Takano and J. Qiu and Z.-H. Lu, Nat. Mater., 2012, 11, 76;
T. Kudo, Solid State Ionics, 1992, 53–56, 507. M. T. Greiner, L. Chai, M. G. Helander, W.-M. Tang and
23 N. Giordano, A. Castellan, J. C. J. Bart, A. Vaghi and Z.-H. Lu, Adv. Funct. Mater., 2012, 22, 4557.
F. Campadelli, J. Catal., 1975, 37, 204. 40 T. H. Fleisch and G. J. Mains, J. Chem. Phys., 1982, 76, 780.
24 K. Segawa, K. Ooga and Y. Kurusu, Bull. Chem. Soc. Jpn., 41 A. Katrib, A. Benadda, J. W. Sobczak and G. Maire, Appl.
1984, 57, 2721. Catal., A, 2003, 242, 31.
25 Y. Iwasawa, Y. Sato and H. Kuroda, J. Catal., 1983, 82, 289. 42 S. Sladkevich, V. Gutkin, O. Lev, E. A. Legurova,
26 N. Giordano, J. C. J. Bart, A. Vaghi, A. Castellan and D. F. Khabibulin, M. A. Fedotov, V. Uvarov,
G. Martinotti, J. Catal., 1975, 36, 81. T. A. Tripol'skaya and P. V. Prikhodchenko, J. Sol-Gel Sci.
27 M. Rozenberg, A. Loewenschuss and Y. Marcus, Spectrochim. Technol., 2009, 50, 229.
Acta, Part A, 1998, 54, 1819. 43 J. B. Goodenough, Prog. Solid State Chem., 1971, 5, 145.
28 F. W. Moore and R. E. Rice, Inorg. Chem., 1968, 7, 2510. 44 J. B. Goodenough, Phys. Rev., 1960, 117, 1442.
29 F. A. Cotton and R. M. Wing, Inorg. Chem., 1965, 4, 867. 45 L. C. Palilis, M. Vasilopoulou, D. G. Georgiadou and
30 C. Rocchiccioli-Deltcheff, R. Thouvenot and R. Franck, P. Argitis, Org. Electron., 2010, 11, 887.
Spectrochim. Acta, Part A, 1976, 32, 587. 46 G. Dennler, M. C. Scharber and C. J. Brabec, Adv. Mater.,
2009, 21, 1323.

6300 | J. Mater. Chem. C, 2014, 2, 6290–6300 This journal is © The Royal Society of Chemistry 2014

View publication stats

You might also like