The development of a new approach toward lanthanide-based OLED fabrication: new host materials for Tb-based emitters

Article abstract of DOI:10.1039/C8DT02911C

To develop the recently proposed approach toward host selection for lanthanide-based emitters, four phosphine oxides PO = PO1-PO4 were investigated which are able to both increase the electron mobility and to sensitize terbium luminescence. New highly sol

Full text of DOI:10.1039/C8DT02911C

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
The development of a new approach toward
lanthanide-based OLED fabrication: new host
materials for Tb-based emitters†
Cite this: Dalton Trans., 2018, 47,
16350
Andrey N. Aslandukov, a Valentina V. Utochnikova,
*
a,b Dmitry O. Goriachiy,c
Andrey A. Vashchenko,c Dmitry M. Tsymbarenko,a Michael Hoffmann,d
Marek Pietraszkiewicz e and Natalia P. Kuzminaa
To develop the recently proposed approach toward host selection for lanthanide-based emitters, four
phosphine oxides PO = PO1–PO4 were investigated which are able to both increase the electron mobility
and to sensitize terbium luminescence. New highly soluble and brightly luminescent terbium complexes
TbCl3(PO)·H2O and Tb(pobz)3(PO)·(CH3)2CO (pobz− = phenoxybenzoate) with a quantum yield of up to
100% were synthesized and thoroughly characterized. To study the electroluminescence properties of
these materials, a series of solution-processed OLED devices were fabricated and their heterostructures
were selected based on the HOMO and LUMO energies of PO1–PO4, which were carefully assessed by
the combination of DFT and TDDFT methods. Thus, the effectiveness of the proposed approach was
proved, and the influence of the anionic ligand was shown. The maximum OLED luminance reached
75 Cd m−2, which is a high value for solution-processed OLEDs based on terbium complexes.
Received 17th July 2018,
Accepted 19th October 2018
DOI: 10.1039/c8dt02911c
rsc.li/dalton
ing the device lifetime is the film morphology and the packing
density. Since the molecules in the films obtained by solution
Introduction
Organic light-emitting diodes (OLEDs) have drawn intense processing are not as closely packed as those in the vacuum-
attention during the past few decades due to their potential based layers, solvent impurities may remain inside the solu-
applications in solid-state lighting and flat-panel displays. The tion-processed films and oxygen can diffuse into the film more
standard method for fabricating organic light-emitting diodes easily.4
(OLEDs) based on small molecules is vapor deposition of
However, due to the further optimization of the film depo-
materials through a series of physical shadow masks.1,2 sition process, e.g. selection of an optimal solvent, spin-
However, this process has critical drawbacks including con- coining speed and temperature, for certain materials, it is
siderable loss of the expensive materials during evaporation already possible to obtain OLEDs that are only slightly inferior
and high manufacturing costs. The best way to improve the to vacuum-deposited analogues.5 Thus, a further search for
efficiency of the process and reduce the production costs is to new materials for solution-processed OLEDs is an important
use solution processing of OLED materials.2
task.
As materials, we propose coordination compounds of
At the same time, the efficiencies of solution-processed
OLEDs are generally lower than those of vacuum-deposited lanthanides. Because of the special mechanism of lumine-
OLEDs.3,4 This is primarily related to the fast degradation of scence,6 the triplet excitons are involved in the luminescence
the device. Lee et al. showed that the crucial factor determin- processes, and lanthanide narrow emission bands (down to
FWHM = 10 nm) guarantee a high purity of light. At present,
the best results in OLEDs based on lanthanide complexes were
obtained using mixed-ligand beta-diketonate complexes de-
aM.V. Lomonosov Moscow State University, 1/3 Leninskye Gory, Moscow 119991,
Russia. E-mail: valentina.utochnikova@gmail.com
posited from the gas phase.7,8 However, beta-diketonates are
unstable under current. Therefore, a promising class of
anionic ligands are aromatic carboxylates, since they are
stable9 and exhibit quantum yields of up to 100%.10 However,
they usually do not possess the intrinsic mobility of charge car-
riers. To solve this problem, an additional host material that
must demonstrate high mobility, for example electronic, and
bEVOLED Ltd, 1A-24 Puškina iela, Riga LV-1050, Latvia
cP.N. Lebedev Physical Institute, Leninsky prosp. 53, Moscow 119992, Russia
dFraunhofer Institute for Organic Electronics, Electron Beam and Plasma Technology
FEP, Maria-Reiche-Straße 2, 01109 Dresden, Germany
eInstitute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44/52,
01-224 Warsaw, Poland
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c8dt02911c
16350 | Dalton Trans., 2018, 47, 16350–16357
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
Fig. 1 Investigated host materials.
effectively transfer energy to the emitter is widely utilized. The
classical approach to the choice of the host material is based
on the host → ligand energy transfer, but in the case of lantha-
nide compounds, this transfer implies a large Stokes shift,
which leads to a loss of the OLED efficiency. Thus, we have
recently proposed a new approach to the selection of such a
host that will transfer energy directly to the luminescent ion
due to coordination.11
In the present work, we investigate four new host materials,
able to serve as ligands (Fig. 1). Due to the presence of a elec-
tron-depleted system, we expect them to possess electronic
mobility; indeed, PO2 and PO4 were already tested as host
materials in iridium-based OLEDs.12,13 While the presence of
the PvO fragment makes them capable of forming complexes
with the lanthanide ion,14,15 particularly PO1 and PO4 were
used to form complexes with Tb3+.16,17 As the emitting ion,
Tb3+ was selected because phosphine oxides are known to sen-
sitize the terbium luminescence efficiently.
Since the electroexciton will be generated on a host material
that is directly capable of transferring energy to the lanthanide
ion, the role of the anionic ligand becomes unobvious. To
determine the role of the anionic ligand, we have chosen two
different anions: the chloride anion, which is not capable of
sensitizing the terbium luminescence, and the o-phenoxy-
benzoate anion (pobz−), which is known for its very
effective sensitization of the terbium luminescence
(QY(Tb(pobz)3(H2O)2) = 96%) and at the same time it does not
possess the intrinsic charge carrier mobility.18
Fig. 2 (a) TGA curve of TbCl3(PO1)·H2O and (b) TGA curve and simul-
taneously detected evolution of the gaseous decomposition products of
Tb(pobz)3(PO4)·(CH3)2CO. Normalized MS profiles are given for m/z 18
(H2O), 44 (CO2), and 58 (acetone).
the contribution of complex thermal decomposition in the
temperature range of 325–740 °C. No additional temperature
points corresponding to the local concentration maxima at
lower temperature were detected in the H2O evolution profile
(Fig. 2b), indicating no coordinated water molecules. Unlike
water, the acetone evolution profile (m/z = 58) demonstrated a
pronounced concentration maximum at 155 °C, corresponding
to its removal according to the reaction:
TbðpobzÞ3ðPOÞ ꢀ ðCH3Þ2CO ! TbðpobzÞ3ðPOÞ þ ðCH3Þ2CO:
Such a high temperature point might be the consequence
of the chemical bond between the acetone molecule and
Tb(pobz)3(PO). Thus, the nature of the solvent was determined
from the MS profiles, while the corresponding weight loss
allowed determining the number of the coordinated acetone
molecules.
The TGA data also indicated the absence of the PO impur-
ity, which is indirectly supported by the PXRD data, at least for
PO2 and PO4. Indeed, all the Tb(pobz)3(PO)·(CH3)2CO com-
plexes are amorphous (Fig. S22†), while PO complexes (PO =
Results and discussion
Synthesis and complex composition
TbCl3(PO)·H2O complexes were obtained by a reaction between PO1–PO4) are crystalline (Fig. S19–21†), which proves the
an excess of TbCl3·6H2O and the corresponding PO in ethanol, absence of the free neutral ligand PO impurity.
followed by evaporation of the reaction mixture to dryness and
In the case of both TbCl3(PO)·H2O and Tb(pobz)3
washing with water to avoid the impurity of TbCl3·6H2O. This (PO)·(CH3)2CO, only one neutral ligand molecule (PO) is co-
resulted in the formation of monohydrated complexes, which ordinated, which is confirmed from the TGA data. In the case
were analyzed using the thermal analysis data. While of Tb(pobz)3(PO)·(CH3)2CO, this conclusion is also supported
Tb(pobz)3(PO)·(CH3)2CO (PO = PO1–PO4) complexes were by the 1H NMR spectroscopy data, based on the ratio of the
obtained as a precipitate from a water–acetone mixture, which integrated intensities of the non-overlapping signals of neutral
leads to a different solvate composition.
ligands PO1–PO4 and the anionic ligand pobz−. Since the Tb3+
To determine this solvate composition, the thermal analysis ion has strong paramagnetic ion effects19 and the NMR
data with the simultaneously recorded MS profiles were used spectra of its complexes are usually not quantitative due to
(Fig. 2). The profiles of carbon dioxide (m/z = 44) and water strong signal broadening,20 1H NMR analysis was carried out
(m/z = 18) evolution were very similar in shape and reflected for La(pobz)3(PO)·(CH3)2CO complexes, synthesized by the
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16350–16357 | 16351

View Article Online
Paper
Dalton Transactions
Table 1 Luminescence properties of terbium complexes
Complex
PLQY, % ( 5)
τobs, ms ( 0.05)
TbCl3(PO1)·H2O
TbCl3(PO2)·H2O
TbCl3(PO3)·H2O
TbCl3(PO4)·H2O
Tb(pobz)3(PO1)·(CH3)2CO
Tb(pobz)3(PO2)·(CH3)2CO
Tb(pobz)3(PO3)·(CH3)2CO
Tb(pobz)3(PO4)·(CH3)2CO
29
54
39
69
100
87
1.25
2.77
2.12
1.89
1.86
1.47
1.73
1.76
Fig. 3 1H NMR spectra of La(pobz)3(PO)·(CH3)2CO, PO and La
(pobz)3(H2O)2, PO = (a) PO1 and (b) PO2 in DMSO-d6.
95
100
same reaction (Fig. 3). The ionic radius of La3+ is larger than
that of Tb3+, which suggests that if the La3+ coordination
sphere is saturated due to the coordination of only one neutral
PO ligand in the La(pobz)3(PO)·(CH3)2CO complexes, then the
coordination sphere of the terbium ion will also be saturated
with only one molecule of PO.
The 1H NMR spectra of the La(pobz)3(PO)·(CH3)2CO com-
plexes represent the superposition of the 1H NMR spectra of
La(pobz)3(H2O)2 and of the corresponding PO. The ratio of the
integrated intensities of the signals indicate the 3 : 1 ratio of
the pobz− and PO ligands, which confirms the assigned com-
position. It is also worth noting that due to the coordination
to the La3+ ion, the PO signals in the spectrum of the complex
are slightly shifted in comparison with the spectrum of the
free ligand (i.e. from 7.75 ppm for PO1 to 7.73 ppm for La
(pobz)3(PO1)·(CH3)2CO), indicating the coordination of the
neutral ligand by La3+ in DMSO, while the signals of pobz−
protons are almost not shifted from La(pobz)3(H2O)2 to La
(pobz)3(PO)·(CH3)2CO.
Fig. 4 (a) Excitation (λem
320 nm) spectra of TbCl3(PO)·H2O complexes.
= 545 nm) and (b) luminescence (λex =
substantially from the other excitation spectra, which can be
attributed to the fundamentally different structure of the PO1
ligand, since it is the only ligand in which the triphenyl-
phosphine oxide fragments are connected through an ali-
phatic, and not aromatic, bridge (Fig. 1).
The Tb(pobz)3(PO)·(CH3)2CO (PO = PO1–PO4) complexes
also exhibit only terbium ionic luminescence (Fig. 5b). The
excitation spectra of Tb(pobz)3(PO)·(CH3)2CO are very similar
to each other and to the excitation spectrum of the complex Tb
(pobz)3(H2O)2 (Fig. 5a). This suggests that the HOMO and
LUMO orbitals in Tb(pobz)3(PO)·(CH3)2CO appear to be loca-
lized to the o-phenoxybenzoate anion, and it plays an impor-
tant role in the transfer of energy to the terbium ion.
The quantum yields of TbCl3(PO)·H2O allow a qualitative
assessment of the sensitization efficiency of ligands, which is
rather high for all the ligands. The highest sensitization
efficiency was observed for PO4. The quantum yields of all the
Tb(pobz)3(PO)·(CH3)2CO complexes reach 87–100% (Table 1),
which makes them prospective luminescent materials for
various applications. Such high quantum yields of
Luminescence properties
The PO1 and PO4 ligands were known earlier to form lumines-
cent complexes with terbium acylpirazolonates or beta-diketo-
nates,16,17 while no data for terbium sensitization ability by
PO2 and PO3 are present in the literature. Besides even in the
case of PO1 and PO4, it was not clear whether the PO-ligand
played a role in the luminescence sensitization, or terbium
luminescence was observed due to the anionic ligand.
Therefore to test the possibility of terbium luminescence sen-
sitization by these ligands, TbCl3(PO)·H2O luminescence was
examined.
All the TbCl3(PO)·H2O (PO = PO1–PO4) complexes exhibit
only terbium ionic luminescence, and the absence of ligand
emission bands in the spectrum and high quantum yields (up
to 70%, Table 1) indicate a high sensitization efficiency
5
7
(Fig. 4b). The different Stark splitting of the D4 → FJ bands
in the luminescence spectra of TbCl3(PO)·H2O indicates a sig-
nificant difference in the coordination environment of the
terbium ion in these complexes, which is also confirmed by
the different lifetimes of the complexes (Table 1).
The excitation spectra of TbCl3(PO)·H2O are strongly
different from each other (Fig. 4a), which is due to the
different energies of the singlet excited states of the PO
Fig. 5 (a) Excitation (λem
= 545 nm) and (b) luminescence (λex =
ligands. The excitation spectrum of TbCl3(PO1)·H2O differs 320 nm) spectra of Tb(pobz)3(PO)·(CH3)2CO complexes.
16352 | Dalton Trans., 2018, 47, 16350–16357
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
Tb(pobz)3(PO)·(CH3)2CO were expected as the PLQY of
Tb(pobz)3(H2O)2 reached 96%.18
Quantum chemical calculations
Quantum chemical modeling was aimed at determining the
energies of the frontier orbitals of the neutral ligands, since
these values are crucial for OLED fabrication as they are
expected to be involved in the electron transport processes. In
addition, the localization of frontier orbitals was determined,
since the knowledge of the orbital localization would allow the
determination of the positions for the substituent introduction
for the directed change of the energies of these orbitals.
For the PO2 ligand, the energies of the frontier orbitals
have already been calculated by Chunmiao Han et al. at the
DFT B3LYP/6-31G* level in ref. 12 and were recalculated here
by the same dual-stage approach as for PO1, PO3, and PO4
(DFT + TDDFT PBE0/6-31G* level).
According to the DFT results, HOMOs are localized on the
linker between the individual phosphine oxides (Fig. 6). This
explains the differences in the energy of the HOMO for
selected PO1–PO4 compounds, since the nature of the linker is
different: aromatic (PO2, PO4), heteroaromatic (PO3), and ali-
phatic (PO1).
It is well known that the DFT method, suitable for ground
state calculations, significantly overestimates the energies of
virtual orbitals containing no electrons, to which the LUMO
belongs. Therefore in our work we used a dual-stage approach
to determine the LUMO energy, based on the calculation of
vertical excitations and then the energy gap (Eg) from TDDFT,
followed by the calculation of the LUMO energy as E(LUMO) =
E(HOMO) + Eg.
To verify the reasonableness of vertical excitation calcu-
lations, the TDDFT absorption spectra were compared with
those experimentally observed. It turned out that they corre-
lated very well both in the position of the bands and in the
relative magnitude of the oscillator strength (Fig. 7). The high
absorption of PO ligands (up to ε = 5 × 104 M−1 cm−1, Fig. 7a)
also makes Tb(pobz)3(PO)·(CH3)2CO complexes promising for
photoluminescence applications.
Fig. 7 Absorption spectra of PO: (a) experimental (10−5 M, acetonitrile)
and (b) calculated.
It was found that only for the PO4 ligand, the lowest energy
band in the TDDFT spectrum corresponds to the HOMO →
LUMO transition. In the remaining cases, all the calculated
absorption bands correspond to the superposition of several elec-
tronic transitions. The most significant contribution is made by
the transitions HOMO−1 → LUMO and HOMO → LUMO
(Table 2). To calculate the energy of the LUMOs, we used the
expression E(LUMO) = E(HOMO−N) + Eg, where E(HOMO−N)
corresponds to either (a) the energy of the orbital with the
highest contribution, or (b) the weighted average energy of the
orbitals from which the transition to the LUMO was observed.
We found that the energy values obtained from the energy
of the main and averaged orbitals (Table 2) practically
coincided. At the same time, as expected, the obtained LUMO
values differ significantly from the data of the DFT method,
particularly the LUMO energy of the PO2 ligand is significantly
different from that from ref. 12.
OLED fabrication
The synthesized complexes were tested in OLEDs as emission
layers within the composite films Tb(L)3 : nPO (L = Cl−, pobz−;
PO = PO1–PO4), where an excess of PO was introduced to
ensure electronic mobility, and the ratio of Tb(L)3 : PO = 1 : n
was varied (Table 3).
The ETL and HTL layers were chosen according to the cal-
culated HOMO and LUMO energy values of the PO. Initially,
the TAZ material with a high LUMO energy (2.8 eV) was
chosen as the electron transport layer due to the high LUMO
level of PO1–PO4. The widely used PVK material was chosen as
the hole transport layer, since its HOMO level (5.8 eV) is well
suited to facilitate the transport of holes from the PEDOT:PSS
hole into the emission layer, as well as PVK is insoluble in
ethanol, from which the emission layer will be deposited.
To determine the best host material, all PO1–PO4 hosts
were tested in OLEDs with the same heterostructures with
TbCl3 : 5PO (PO = PO1–PO4) emission layers made under the
same conditions (A–D devices, Table 3). In the electrolumines-
cence spectra of all A–D diodes, typical emission bands of
terbium luminescence corresponding to the 5D4
sitions were observed.
→
7FJ tran-
The electronic mobility of the PO materials was estimated
by comparing the current values at the same voltage from the
I–V curves (Fig. 8b). The mobility increases in the A–C–B–D
series and the maximum electronic mobility is demon-
strated by the materials PO2 and PO4. The efficiency of the
Fig. 6 Localization of the HOMO in arylphosphine oxides: (a) PO1, (b)
PO2, (c) PO3, and (d) PO4.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16350–16357 | 16353

View Article Online
Paper
Dalton Transactions
Table 2 Quantum chemical calculation data
PO =
PO1
E(HOMO), eV
−6.923
E(LUMO) (DFT), eV
−0.917
Eg, eV
Transition (contributiona)
E(HOMO−N), eV
E(LUMO)b, eV
5.21
HOMO → LUMO (64%)
HOMO−1 → LUMO (15%)
HOMO−1 → LUMO+1 (15%)
HOMO−1 → LUMO+5 (2%)
HOMO → LUMO+1 (1%)
HOMO → LUMO+2 (3%)
HOMO−1 → LUMO (78%)
HOMO → LUMO+2 (14%)
HOMO−3 → LUMO (8%)
HOMO−1 → LUMO (90%)
HOMO → LUMO (10%)
HOMO → LUMO
E(HOMO) = −6.923
Eavg1,2 = −6.933
(a) −1.713
(b) −1.723
PO2
−6.563
−1.295
4.5056
E(HOMO−1) = −6.691
E(HOMO−3) = −6.841
Eavg = −6.705
E(HOMO−1) = −6.876
Eavg = −6.841
E(HOMO) ≡ Eavg = −6.694
(a) −2.185
(b) −2.199
PO3
PO4
−6.84
−1.262
−1.347
4.277
4.659
(a) −2.599,
(b) −2.564
−2.04
−6.694
a Contribution = squared coefficient of Spin Adapted Antisymmetrized Product (SAAP2). b LUMO energy estimated as E(HOMO−N) + Eg, where
E(HOMO−N) corresponds to (a) the energy of the main orbital from which the transition occurs, and (b) the weighted averaged energy of the orbi-
tals from which the transition to the LUMO occurs.
PO* → Tb3+ electroexcitation transfer was evaluated from the
electroluminescence spectra measured at the same current
(Fig. 8a). The electroluminescence intensity increases in the A–
C–B–D series and corresponds to an increase in the sensitizing
ability of the PO ligands which correlates with the values of
the quantum yields of TbCl3(PO)·H2O complexes (Table 1).
Thus, in our case, the mobility and sensitization efficiencies
of the materials of the PO ligands change symbatically in the
Fig. 8 (a) The electroluminescence spectra and (b) the I–V curves of
the devices A–D.
row PO1–PO3–PO2–PO4 and allow us to choose PO4 as the
best material.
To determine the effect of the anionic ligand on the electro-
luminescence intensity, we replaced terbium chloride with
phenoxybenzoate in the composition of the emission layers,
and thus E–H devices were obtained (Table 3). It turned out
that such a replacement resulted in an increase in the
Thus, the best results were observed using Tb(pobz)3(PO4)
as an emitter. The next task was to optimize the structure of
this OLED. Variation of the Tb : PO4 ratio in the composition
electroluminescence intensity by a factor of 1.5–2, and this be- of the emission layer demonstrated that even with a ratio of 1
havior was observed for all the investigated hosts (Fig. S41†).
Thus, it is shown that in spite of the fact that the exciton is
to 1 (device I, Table 3) the electron mobility of PO4 is sufficient
for efficient working of the device, as evidenced by high
generated on the electron transporting PO material, the enough currents (Fig. 9b). The intensity of electrolumines-
excitation is transferred and distributed throughout the
whole organic subsystem of Tb(pobz)3(PO), and the pobz−
anionic ligand participates in the energy transfer process
toward Tb3+.
cence increases (Fig. 9a), and the maximum luminance
reaches 25 Cd m−2 at 12 V.
The introduction of the LiF electron-injecting layer and the
replacement of the electron transport layer TAZ by TPBi, which
Table 3 Device list
Device designation
Heterostructure
Uon a, V
A
B
C
D
E
F
G
H
I
ITO/PEDOT:PSS/PVK/TbCl3:5PO1/TAZ/Al
ITO/PEDOT:PSS/PVK/TbCl3:5PO2/TAZ/Al
ITO/PEDOT:PSS/PVK/TbCl3:5PO3/TAZ/Al
ITO/PEDOT:PSS/PVK/TbCl3:5PO4/TAZ/Al
ITO/PEDOT:PSS/PVK/Tb(pobz)3:5PO1/TAZ/Al
ITO/PEDOT:PSS/PVK/Tb(pobz)3:5PO2/TAZ/Al
ITO/PEDOT:PSS/PVK/Tb(pobz)3:5PO3/TAZ/Al
ITO/PEDOT:PSS/PVK/Tb(pobz)3:5PO4/TAZ/Al
ITO/PEDOT:PSS/PVK/Tb(pobz)3:1PO4/TAZ/Al
ITO/PEDOT:PSS/PVK/Tb(pobz)3:1PO4/TPBi/LiF/Al
14
6.5
8
6
10
6
8
6
6
5
J
a As the Uon value, the voltage, where EL detectable with the naked eye appeared, was considered.
16354 | Dalton Trans., 2018, 47, 16350–16357
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
Photoluminescence quantum yields were determined with a
Fluorolog FL3–22 spectrofluorimeter at room temperature
under excitation into ligand states according to an absolute
method using an integration sphere. The modified de Mello
et al.24 method requires the measurement of (i) La, the inte-
grated intensity of light exiting the sphere when the empty
cuvette is illuminated at the excitation wavelength (Rayleigh
scattering band); (ii) Lc, the same integrated intensity at the
excitation wavelength when the sample is introduced into the
sphere; (iii) Ea the integrated intensity of the entire emission
spectrum of the empty cuvette; and (iv) Ec the integrated inten-
sity of the entire emission spectrum of the cuvette with the
sample. The absolute quantum yield is then given by:
Fig. 9 (a) Electroluminescence spectra of devices H and I at 7 V and (b)
I–V and L–V curves of device J.
has higher electron mobility21 (device J, Table 3) lead to an
increase in currents and in the luminance of the diode, which
reaches 75 Cd m−2 at 14 V (Fig. 9b). Though this value is much
below the state-of-the-art values for the vacuum-deposited
OLEDs, we believe that it is important to publish it to be able
to trace the progress in such a young and prospective area of
research as solution-processed OLED development.
Ec ꢁ Ea
PLQY ¼
ꢂ 100%:
La ꢁ Lc
OLED manufacturing took place in a clean room class
10 000 (Lebedev Physical Institute, Moscow, Russia) in a glove-
box under an argon atmosphere. The substrates were cleaned
by ultrasonication in the following media: NaOH aqueous
solution, distilled water, acetone and 2-propanol for 16 min
each.
Experimental section
Materials and methods
A 40 nm-thick poly(3,4-ethylenedioxythiophene):poly(styre-
nesulfonate) (PEDOT:PSS) hole-injective layer was first de-
All solvents and chemicals were purchased from commercial posited. An aqueous solution of PEDOT:PSS (5 ml) was poured
sources. onto the preheated (70 °C) patterned ITO glass substrate, after
1H NMR spectra were recorded at 25 °C using an Agilent which the substrate was rotated for 60 seconds at 2000 rpm.
400 MR spectrometer with an operating frequency of Finally, the deposited film was annealed in air at 80 °C for
400.130 MHz, 100.613 MHz and 376.498 MHz, respectively. 60 min. As a hole-transport layer PVK solution was spin-coated
Chemical shifts were reported in ppm relative to Me4Si (1H). from toluene (c = 5 g L−1).
Elemental analysis was performed on the microanalytical
The emission layer was spin-coated from the solution (c =
device Heraeus Vario Elementar. Thermal analyses were 5 g L−1) on the heterostucture or on the glass substrate. The
carried out on a thermoanalyzer STA 409 PC Luxx (NETZSCH, ∼30 nm-thick electron-transport/hole-blocking layer (TAZ or
Germany) in the temperature range of 20–1000 °C in air, at a TPBi) was thermally evaporated (Univex-300, LeybordHeraeus)
heating rate of 10° min−1. The composition of the evolved followed by a ∼1 nm-thick LiF layer and a >100 nm-thick
gases was simultaneously monitored during the TA experiment aluminum layer as the cathode under a pressure below 10−6
using a coupled QMS 403C Aeolos quadrupole mass spectro- mm Hg. The thickness of the layers was controlled by using a
meter (NETZSCH, Germany). The mass spectra were recorded quartz indicator. The contacts were attached to the electrodes,
for the species with the following m/z values: 18 (corres- and the device was sealed with epoxy resin (Norland Optical
ponding to H2O), 44 (corresponding to CO2), and 58 (corres- Adhesive).
ponding to (CH3)2CO).
Electroluminescence spectra were recorded on a PicoQuant
Powder X-ray diffraction (PXRD) was performed by using time-correlated single photon counting system used as a con-
Bruker D8 Advance [λ(Cu-Kα) = 1.5418 Å; Ni filter] and Bruker ventional spectrofluorimeter. The spectral resolution was
D8 Advance Vario diffractometers [λ(Cu-Kα1) = 1.54060 Å; 4 nm. All the measurements were carried out in the same con-
Ge(111)-monochromator] with a step size of 0.020°. The pat- figuration, which allows quantitative comparison of the
terns were indexed by using the SVD-Index22 as implemented spectra by dividing by the integration time.
in the TOPAS 4.2 software.23 Then, the powder patterns were
refined by using the Pawley method.
Quantum chemical calculations. The geometry of the
ground state of the PO molecules was optimized using the
Absorption spectra were recorded in the region of DFT method with the functional PBE0 and 6-31G* basis set in
250–800 nm with a PerkinElmer Lambda 650 spectrometer. the NWChem 6.4 package.25 To choose the structural isomer,
Emission and excitation spectra were recorded with
a
the geometry of each possible isomer was optimized and the
Fluorolog 3 spectrofluorometer over excitation with a xenon lowest-energy isomer was chosen. The optimized molecular
lamp. Luminescence lifetime measurements were carried out geometries (Table S4†) were checked for the absence of ima-
and detected on the same system. Lifetimes are averages of at ginary frequencies in the vibrational spectra (Table S5†).
least three independent measurements. All luminescence
TDDFT calculations at the same PBE0/6-31G* level as
decays proved to be perfect single-exponential functions. implemented in the Firefly 8.2 package26 were carried out for
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16350–16357 | 16355

View Article Online
Paper
Dalton Transactions
1
the optimized geometry of PO molecules. 20 singlet single-
excited states were considered.
PO1: Yield = 90%; Bruker H NMR and el. anal. data as in
the literature,28 MS(MALDI): m/z = 416([M + H]+);
The energies of the frontier orbitals were calculated using
PO2: Yield = 90%; Bruker 1H NMR data as in the
the DFT and TDDFT results. The energy of the HOMO was literature.12
obtained as the eigenvalue of DFT calculation for the ground
state (Table S6†). The energy of the LUMO was computed as
described in the Results and discussion section.
MS(MALDI): 569 ([M + H]+);
Elemental analysis: calcd. C 76.05, H 4.61, found C 76.16,
H 4.72.
PO3: Yield = 98%; Bruker 1H NMR (250 MHz, DMSO-d6)
ppm: 8.19–7.47 (m, 13H, Phen-H and, py-H), 0.96 (s, 18H, 2
Synthesis of PO ligands
Phosphine oxides were obtained by the oxidation of the corres- t-Bu);
ponding phosphines.
Bis(diphenylphosphino)methane (for PO1) was purchased
from abcr GmbH.
MS(MALDI): m/z = 440 ([M + H]+);
PO4: Yield = 92%; Bruker H NMR (250 MHz, CDCl3, TMS)
ppm: 7.80–7.43 (m, 28H, Phen-H);
MS(MALDI): m/z = 555 ([M + H]+);
Elemental analysis: calcd. C 77.97, H 5.09 found C 77.93,
H 5.29.
1
1,8-Bis(diphenylphosphino)dibenzofuran (for PO2) was pre-
pared as in ref. 27.
MS(FD): 537.1 (M+), 1072,2 (M++M).
The 1H-NMR spectrum (Bruker, 400 MHz in acetone-d6) was
consistent with ref. 27.
Synthesis of lanthanide complexes
2-Diphenylphosphino-6-di(tert-butyl)phosphinopyridine
(for PO3).
Synthesis of Ln(pobz)3(H2O)2 (Ln = La, Tb). Ln(pobz)3(H2O)2
(Ln = La, Tb) were synthesized by reaction 1 between stoichio-
metric amounts of TbCl3·6H2O or La(NO3)3·6H2O and Kpobz
(in situ from Hpobz and KOH) in water. The products were
dried in air at room temperature.
Synthesis of LnCl3(PO)·H2O (Ln = Gd, Tb; PO = PO1–PO4).
LnCl3(PO)·H2O (Ln = Gd, Tb; PO = PO1–PO4) were synthesized
by the reaction between the corresponding PO and the excess
of LnCl3·6H2O in ethanol. The products were rotor evaporated
to dryness, washed with water from the excess of LnCl3·6H2O
and dried in air at room temperature.
2,6-Dibromopyridine (6.6 g, 28 mmol) was dissolved in dry
THF and stirred for 30 min at −78 °C under an argon blanket.
The first portion of n-butyllithium (5.5 mL, 28 mmol) was
introduced to the solution via a syringe. The reaction mixture
was stirred at −78 °C for one hour and further stirred for
30 minutes at room temperature. After cooling again to
−78 °C, di-tert-butylchlorophosphine (5.4 mL, 28 mmol) was
added dropwise via a syringe, stirring was maintained for 1 h
at this temperature, and warmed slowly to 0 °C. The second
portion of n-butyllithium (5.5 mL of 28 mmol) was introduced
after cooling the reaction mixture again at −78 °C. After stir-
ring the reaction mixture for one hour, chlorodiphenylpho-
sphine (5.1 mL, 28 mmol) was added at once. The reaction
mixture was warmed up to room temperature within 3 h. The
residue obtained after evaporating the solvent was washed
with methanol (2 × 40 mL) to remove inorganics.
TbCl3(PO1)·H2O: MS(MALDI): m/z = 645 ([TbCl2(PO1)]+).
TbCl3(PO2)·H2O: MS(MALDI): m/z = 797 ([TbCl2(PO2)]+).
TbCl3(PO3)·H2O: MS(MALDI): m/z = 668 ([TbCl2(PO3)]+).
TbCl3(PO4)·H2O: MS(MALDI): m/z = 783 ([TbCl2(PO4)]+).
Synthesis of Ln(pobz)3(PO)·(CH3)2CO (Ln = La, Tb; PO =
PO1–PO4). Ln(pobz)3(PO)·(CH3)2CO (Ln = La, Tb) were syn-
thesized by the reaction between stoichiometric amounts of
LnCl3(PO)·H2O and Kpobz (in situ from Hpobz and KOH) in a
water–acetone mixture. The precipitated product was dried in
air at room temperature.
Bis(diphenylphosphino)-4,4′-biphenyl (for PO4).
4,4′-Dibromobiphenyl (9.36 g, 30 mmol) in THF was stirred
at −78 °C for 30 min, while stirring n-butyllithium (25 mL,
62.5 mmol) was added dropwise. The reaction mixture was
allowed to warm up to 0 °C, then chlorodiphenylphosphine
(11.5 mL, 62.5 mmol) was added after cooling again at −78 °C.
The reaction mixture was warmed up to room temperature and
further stirred for 2 h. The residue obtained after evaporating
the solvent was stirred with methanol (2 × 40 mL) to remove
inorganics. The precipitate thus obtained was filtered off and
dried.
Tb(pobz)3(PO1)·(CH3)2CO: MS(MALDI):
m/z = 1001 ([Tb(pobz)2(PO1)]+).
Tb(pobz)3(PO2)·(CH3)2CO: MS(MALDI):
m/z = 1153 ([Tb(pobz)2(PO2)]+).
Tb(pobz)3(PO4)·(CH3)2CO: MS(MALDI):
m/z = 925 ([Tb(pobz)(PO4)-H]+).
General procedure for the oxidation of phosphines: phos-
phine oxides were synthesized by the oxidation of the corres-
Conclusions
ponding phosphines. Three equivalents of H2O2 (30%) per To conclude, in the present work we aimed at the development
dioxide were added slowly to the vigorously stirred THF solu- of a new approach toward host selection for lanthanide-based
tion of the respective phosphine. Stirring was maintained at emitters in OLEDs, earlier proposed by some of us,11 based on
room temperature for 2 h. After the evaporation of THF, the the example of terbium-based emitters. As host materials, aryl-
residue was treated with acetone/ethyl acetate to obtain the phosphine oxides PO = PO1–PO4 were proposed, which were
respective phosphine oxide product as a precipitate. The pre- supposed to both sensitize terbium luminescence and provide
cipitate thus obtained was filtered off and dried in a vacuum.
electron mobility. To assess their sensitization properties,
16356 | Dalton Trans., 2018, 47, 16350–16357
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
TbCl3(PO)·H2O complexes were obtained and thoroughly inves- 10 M. Bredol, U. Kynast and C. Ronda, Adv. Mater., 1991, 3,
tigated. It was found that their quantum yields reach 70%, 361–367.
indicating their high sensitization ability. The high current in 11 V. V. Utochnikova, N. N. Solodukhin, A. N. Aslandukov,
the I–V curve of the OLED based on these materials together
with the pure terbium electroluminescence (EL) spectrum
indicated their sufficient electronic properties and, thus, the
L. Marciniak, I. S. Bushmarinov, A. A. Vashchenko and
N. P. Kuzmina, Org. Electron. Phys. Mater. Appl., 2017, 44,
85–93.
effectiveness of the proposed approach. To select additional 12 C. Han, G. Xie, J. Li, Z. Zhang, H. Xu, Z. Deng,
layers in OLEDs, the frontier orbitals of PO ligands were calcu-
lated based on a complex approach, which allows careful
Y. Zhao, P. Yan and S. Liu, Chem. – Eur. J., 2011, 17, 8947–
8956.
assessment of not only the ground state orbitals, including the 13 P. E. Burrows, A. B. Padmaperuma, L. S. Sapochak,
HOMO, but also the virtual LUMO orbital.
P. Djurovich and M. E. Thompson, Appl. Phys. Lett., 2006,
88, 1–4.
14 A. W. G. Platt, Coord. Chem. Rev., 2017, 340, 62–78.
15 S. Shuvaev, O. Kotova, V. Utochnikova, A. Vaschenko,
L. Puntus, V. Baulin, N. Kuzmina and A. Tzivadze, Inorg.
Chem. Commun., 2012, 20, 73–76.
Conflicts of interest
There are no conflicts to declare.
16 F. Marchetti, C. Pettinari, A. Pizzabiocca, A. A. Drozdov,
S. I. Troyanov, C. O. Zhuravlev, S. N. Semenov,
Y. A. Belousov and I. G. Timokhin, Inorg. Chim. Acta, 2010,
363, 4038–4047.
17 K. Miyata, Y. Konno, T. Nakanishi, A. Kobayashi, M. Kato,
K. Fushimi and Y. Hasegawa, Angew. Chem., Int. Ed., 2013,
52, 6413–6416.
18 A. S. Kalyakina, V. V. Utochnikova, E. Y. Sokolova,
A. A. Vashchenko, L. S. Lepnev, R. Van Deun, A. L. Trigub,
Y. V. Zubavichus, M. Hoffmann, S. Mühl and
N. P. Kuzmina, Org. Electron. Phys. Mater. Appl., 2016, 28,
319–329.
19 V. S. Sastri, J.-C. G. Bunzli, V. R. Rao, G. V. S. Rayudu and
J. R. Perumareddi, Modern aspects of rare earths and their
complexes, Elsevier, Amsterdam, 2003.
20 V. V. Utochnikova, A. S. Kalyakina, I. S. Bushmarinov,
A. A. Vashchenko, L. Marciniak, A. M. Kaczmarek, R. Van
Deun, S. Bräse and N. P. Kuzmina, J. Mater. Chem. C, 2016,
4, 9848–9855.
21 S. Valouch, C. Hönes, S. W. Kettlitz, N. Christ, H. Do,
M. F. G. Klein, H. Kalt, A. Colsmann and U. Lemmer, Org.
Electron. Phys. Mater. Appl., 2012, 13, 2727–2732.
Acknowledgements
This work was supported by the Russian Foundation for Basic
Research (grants #16-53-76018, #17-32-80050). UV thanks
Prezident’s grant MK-2810.2017.3 for the financial support.
MP thanks National Science Centre, Poland, grant no.: OPUS9
2015/17/B/ST5/01038, in part by the Institute of Physical
Chemistry of the Polish Academy of Sciences. The research is
carried out using the equipment of the shared research facili-
ties of HPC computing resources at Lomonosov Moscow State
University.
Notes and references
1 J. Jayabharathi, K. Jayamoorthy and V. Thanikachalam,
J. Organomet. Chem., 2014, 761, 74–83.
2 N. Thejo Kalyani and S. J. Dhoble, Renewable Sustainable
Energy Rev., 2015, 44, 319–347.
3 L. Duan, L. Hou, T.-W. Lee, J. Qiao, D. Zhang, G. Dong,
L. Wang and Y. Qiu, J. Mater. Chem., 2010, 20, 6392.
4 T. W. Lee, T. Noh, H. W. Shin, O. Kwon, J. J. Park, 22 A. A. Coelho, J. Appl. Crystallogr., 2003, 36, 86–95.
B. K. Choi, M. S. Kim, D. W. Shin and Y. R. Kim, Adv. Funct. 23 Bruker, TOPAS 4.2 User Manual, 2009.
Mater., 2009, 19, 1625–1630.
5 L. Deng, W. Li and J. Li, Displays, 2013, 34, 413–417.
24 J. C. De Mello, H. F. Wittmann and R. H. Friend, Adv.
Mater., 1997, 9, 230–232.
6 J. C. G. Bünzli and S. V. Eliseeva, in Comprehensive 25 M. Valiev, E. J. Bylaska, N. Govind, K. Kowalski,
Inorganic Chemistry II: From Elements to Applications, 2nd
edn, 2013, vol. 8, pp. 339–398.
7 H. Xu, Q. Sun, Z. An, Y. Wei and X. Liu, Coord. Chem. Rev.,
2015, 293, 228–249.
T. P. Straatsma, H. J. J. Van Dam, D. Wang, J. Nieplocha,
E. Apra, T. L. Windus and W. A. De Jong, Comput. Phys.
Commun., 2010, 181, 1477–1489.
26 A. A. Granovsky, Firefly version 8, http://classic.chem.msu.
su/gran/firefly/index.html.
8 Z. Chen, F. Ding, F. Hao, Z. Bian, B. Ding, Y. Zhu, F. Chen
and C. Huang, Org. Electron. Phys. Mater. Appl., 2009, 10, 27 M. Kranenburg, Y. E. M. van der Burgt, P. C. J. Kamer,
939–947.
P. W. N. M. van Leeuwen, K. Goubitz and J. Fraanje,
9 O. V. Kotova, V. V. Utochnikova, S. V. Samoylenkov,
Organometallics, 1995, 14, 3081–3089.
A. D. Rusin, L. S. Lepnev and N. P. Kuzmina, J. Mater. 28 S. Kannan, N. Rajalakshmi, K. V. Chetty, V. Venugopal and
Chem., 2012, 22, 4897.
M. G. B. Drew, Polyhedron, 2004, 23, 1527–1533.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans., 2018, 47, 16350–16357 | 16357