Luminescent platinum(II) complexes containing cyclometallated diaryl ketimine ligands: Synthesis, photophysical and computational properties

Article abstract of DOI:10.1002/ejic.200901159

A new series of platinum(II) complexes containing cyclometallated diaryl ketimme ligands has been synthesised. The route involves reaction of diaryl ketirnmes with K[PtCl₃(dmso)] to obtain frans-[PtCl ₂(imine)(dmso)j species, which underwent cyclometallation upon heating in toluene to give [PtCl(Ar′(ArC=N.H)}(dmso)] complexes 12b-17b. JVHydroxy and. JV-phenyl analogues 18b and 21 were also synthesised. In complexes 19 and 20 the auxiliary chlorido and dmso ligands were replaced by an acetylacetonato ligand. The photophysical properties of the cyclometallated complexes are reported. The emission bands at λ_max ≈ 450 and 550 nm are assigned to mixed-ligand. and MLCT states having significant singlet and triplet character, respectively. By varying the structure of the aromatic ligand the efficiency of phosphorescence can be increased to 4.3% for ISb (Ar′ = 1-naphthyl), Theoretical calculations show that the low-energy transitions in all the cyclometallated systems involve mainly the frontier orbitais, HOMO and LUMO. These are mixed chlorido-metal-ligand to largely π*-C=N transitions. Most of the observed, phosphorescence data can be explained by the geometric change on going from the S_O to the T ₁ states. An organic light-emitting device has been fabricated by using complex 15b as the emissive dopant in a poly(vinylcarbazole) host. Broad electroluminescence spanning the range 500750 nm. was observed.

Full text of DOI:10.1002/ejic.200901159

FULL PAPER
DOI: 10.1002/ejic.200901159
Luminescent Platinum(II) Complexes Containing Cyclometallated Diaryl
Ketimine Ligands: Synthesis, Photophysical and Computational Properties
Shashi U. Pandya,[a] Kathryn C. Moss,[a] Martin R. Bryce,*[a] Andrei S. Batsanov,[a]
Mark A. Fox,[a] Vygintas Jankus,[b] Hameed A. Al Attar,[b] and Andrew P. Monkman[b]
Keywords: Platinum / Complexation / Cyclometallation / Photoluminescence / Density functional calculations / Light-
emitting devices
A new series of platinum(II) complexes containing cyclo-
metallated diaryl ketimine ligands has been synthesised. The
route involves reaction of diaryl ketimines with
K[PtCl3(dmso)] to obtain trans-[PtCl2(imine)(dmso)] species,
which underwent cyclometallation upon heating in toluene
to give [PtCl{ArЈ(ArC=NH)}(dmso)] complexes 12b–17b. N-
Hydroxy and N-phenyl analogues 18b and 21 were also syn-
thesised. In complexes 19 and 20 the auxiliary chlorido and
dmso ligands were replaced by an acetylacetonato ligand.
The photophysical properties of the cyclometallated com-
plexes are reported. The emission bands at λmax ≈ 450 and
550 nm are assigned to mixed-ligand and MLCT states hav-
ing significant singlet and triplet character, respectively. By
varying the structure of the aromatic ligand the efficiency of
phosphorescence can be increased to 4.3% for 15b (ArЈ = 1-
naphthyl). Theoretical calculations show that the low-energy
transitions in all the cyclometallated systems involve mainly
the frontier orbitals, HOMO and LUMO. These are mixed
chlorido–metal–ligand to largely π*-C=N transitions. Most of
the observed phosphorescence data can be explained by the
geometric change on going from the S0 to the T1 states. An
organic light-emitting device has been fabricated by using
complex 15b as the emissive dopant in a poly(vinylcarbazole)
host. Broad electroluminescence spanning the range 500–
750 nm was observed.
[20]
∧ ∧
or [C N C]
coordination mode. The phosphorescence
Introduction
∧ ∧
∧ ∧
quantum yields of [N N C] and [C N C] platinum(II)
∧ ∧
Luminescent transition metal complexes enjoy a diverse
range of applications, e.g. sensors, photocatalysis, nonlinear
optical materials for data storage and organic light-emitting
diodes (OLEDs).[1,2] In this context, platinum complexes
have attracted considerable interest due to their efficient
room-temperature phosphorescence, short emission life-
times and broad emission range.[3] For example, plati-
num(II)–porphyrins are highly luminescent with high quan-
tum yields.[4,5] Room-temperature phosphorescence can be
achieved from cyclometallated PtII complexes due to radia-
tive decay from either excited π–π* or MLCT states. Li-
gands such as 2-phenylpyridine and its analogues have been
extensively studied for this purpose (Figure 1). For specific
systems tend to be much lower than those of [N C N]-
based systems.[11,15] Whereas a number of such cycloplati-
nated ligands have been studied in reference to their lumi-
nescence properties, there have been few reports of other
types of ligands.
Our aim was an integrated, synthetic, photophysical and
computational study of a new family of platinum(II) com-
plexes containing cyclometallated diaryl ketimine ligands
where emission properties might be tuned by varying the
imine ligand structure. It is worth noting that, while several
luminescent platinum complexes with aldimines as ligands
have been reported,[21] complexes of ketimines are very rare
and appear to be restricted to benzophenone imine. At the
outset of our work Scaffidi-Domianello et al. reported
(benzophenone imine)platinum(II) compounds, trans-
[PtCl2(Ph2C=NH)(RRЈSO)] containing a range of different
dialkyl(aryl) sulfoxide ligands.[22] The solution photolumi-
nescence quantum yields of the derived cyclometallated
complexes were very low, the maximium being Φem = 0.019.
Solid-state quantum yields and device fabrication studies
were not reported, and no other imine derivatives were
studied in this work. Our series of complexes described
herein comprise new structural modifications. The diaryl
ketimine ligands are systematically varied, and in two cases
examples, complexes 1,[6] 2,[7] 3,[8] [9] and 5[10] are represen-
4
tative. The emission profile is largely dependent on the sub-
stituents on the cyclometallated ligands and also the type
of coordinating ligand, e.g. [N C N]
[11–15]
[16–19]
∧ ∧
∧ ∧
[N N C]
[a] Department of Chemistry, Durham University,
Durham DH1 3LE, United Kingdom
Fax: +44-0191-384-4737
E-mail: m.r.bryce@durham.ac.uk
[b] Department of Physics, Durham University,
Durham DH1 3LE, United Kingdom
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.200901159.
Eur. J. Inorg. Chem. 2010, 1963–1972
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1963

M. R. Bryce et al.
FULL PAPER
Figure 1. Selection of cyclometallated PtII complexes of (hetero)arylpyridine ligands.
substituents are attached to the imine nitrogen atom. Solu- tion of side products. The purified complexes 12a–17a were
tion and solid-state photophysical properties of the cyclo- then subjected to cyclometallation by heating in toluene for
metallated complexes are reported along with theoretical 15 h. The corresponding complexes 12b–17b were isolated
calculations to support the experimental observations.
in low yields and were observed to be luminescent on TLC
(SiO2-coated plate) under 365 nm irradiation and were no-
tably more soluble in common organic solvents compared
to their precursors 12a–17a. The low yield for the cyclome-
tallation step was due to the instability of the complexes
12b–17b in the reaction mixture; after isolation and purifi-
cation, 12b–17b were stable to storage as solids.
Results and Discussion
Synthesis of Cyclometallated Diaryl Ketimine Platinum(II)
Complexes
In the case of 3,3Ј-difluorobenzophenone oxime, despite
several attempts, vacuum pyrolysis did not yield the corre-
sponding imine. Ryabov et al. have reported orthoplatina-
tions of aryl oximes using dichloridobis(sulfoxide/sulfide)-
platinum(II) complexes in moderate yields, although no
photophysical properties were reported.[25] We therefore
proceeded to obtain compound 18a and the derived cyclo-
metallated product 18b. The structure of cis-PtCl2(oxime)-
(dmso) complex 18a was confirmed by X-ray analysis (see
below). Complex 18b, upon irradiation with 365 nm light,
showed considerably weaker luminescence than complexes
12b–17b.
A series of N-unsubstituted diaryl ketimines 6–11 were
synthesised either by: (i) thermal decomposition of ox-
imes[23] or (ii) by Grignard reaction of a nitrile.[24] Different
substituents were attached to the aryl rings [electron-with-
drawing fluoro (7), electron-releasing methoxy (8), π-ex-
tended aryl (9–11)] to study the effect of these electronic
and steric modifications on the cyclometallation reactivity
and the subsequent photophysical properties of the com-
plexes. The diaryl ketimines were all used as prepared, with-
out purification, due to their instability.
The imines were treated with K[PtCl3(dmso)] obtained in
situ by reaction of K2PtCl4 and dmso (Scheme 1). The
major products obtained were the trans-[PtCl2(imine)-
Replacement of the auxiliary chlorido and dmso ligands
(dmso)] complexes 12a–17a in variable yields. Small with a bidentate acetylacetonato ligand considerably im-
amounts of cis-[PtCl2(imine)(dmso)] (12–15%), cis-[PtCl2- proved the stability of the products. Complexes 19 and 20
(imine)2] (6–8%) and µ-[PtCl(imine)(dmso)] complexes were were prepared in good yields in a one-pot procedure by
also detected in the NMR spectra of the crude product mix- treating cis-[PtCl2(dmso)2][26] with the respective imines (6
ture. The yield for this step of the reaction sequence was and 10) in dry toluene followed by addition of acetyl-
low due to the instability of the imines 6–11 and the forma- acetone and sodium carbonate (Scheme 2). In contrast to
Scheme 1. Synthetic routes to cyclometallated complexes 12b–18b. The structures of 12b–18b are shown in Figure 2.
1964
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 1963–1972

Luminescent Platinum(II) Complexes
Figure 2. Structures of the cyclometallated Pt complexes synthesised in the present study.
12b–18b, complexes 19 and 20 were stable to storage in the Supporting Information). No isomeric product was de-
1
solution (in chloroform, dichloromethane or toluene, no tected in the H NMR analysis of the crude reaction mix-
change was observed by 1H NMR analysis after one week). tures.
One example of an N-phenyl-substituted ketimine complex
21 was prepared (see the Supporting Information). How-
ever, this complex did not show any phosphorescence, so
N-substitution in this series was not explored further.
X-ray Crystal Structure of 18a
The asymmetric unit of 18a comprises two molecules of
similar conformation (Figure 3), related by a pseudo-glide
plane a(x,y,¼) and linked by N–OH···O=S hydrogen bonds
into an infinite chain parallel to the crystallographic x axis.
The Pt atom has a cis-square coordination with a small tet-
rahedral distortion (5.4° twist between the PtCl2 and PtSN
planes). The bond lengths are usual for cis-R2(O)SPt(N-li-
gand)Cl2 complexes,[28] although it is noteworthy that
PtCl2(SOMe2)(HON=CPhMe), close to 18a in composi-
tion, has a trans-square structure.[25b] There is a 7.5° twist
around the N=C(1) double bond of 18a; the bonding plane
of the N atom [N,O(1),Cl(1),Pt] is inclined by ca. 70° to the
Scheme 2. Synthesis of complexes 19 and 20.
The structures of the cyclometallated complexes synthe-
sised in the study are shown in Figure 2. The precursor
complexes 12a–18a and the cyclometallated products 12b–
1
18b, 19–21 were characterised by H and 13C NMR spec-
troscopy, mass spectrometry (ES+ mode), IR spectroscopy
and – where possible – also by CHN elemental analysis. In
the latter series the aromatic proton of the cyclometallated
ring displays characteristic satellites from 3JPt,H coupling to
the 195Pt nucleus. The methyl signal of the dmso ligand is
also shifted downfield in the series 12b–18b (∆δ ≈ 0.1–
3
0.3 ppm) with clearly resolved satellites from JPt,Me with J
≈ 21 Hz. The IR stretching frequency of the C=N bond
generally decreases by ca. 10–20 cm–1 in 12b–18b compared
to 12a–18a, and the peak due to the coordinated dmso is
seen at ca. 1120–1130 cm–1, shifted from that of free dmso
(1037–1124 cm–1) and consistent with coordination to the
platinum atom through the sulfur atom.[27]
It is interesting to note that cyclometallation of the un-
symmetrical diaryl ketimine complexes 15a and 16a oc-
curred exclusively onto the phenyl ring and not the naph-
thyl or phenanthryl rings to yield 15b and 16b, respectively,
Figure 3. X-ray molecular structure of 18a, showing the atomic dis-
placement ellipsoids at 50% probability level. Selected bond lengths
(for two independent molecules) [Å]: Pt–Cl(1) 2.314(1) & 2.315(1),
Pt–Cl(2) 2.286(1) & 2.293(1), Pt–S 2.212(1) & 2.215(1), Pt–N
2.015(3)
& 2.017(3), N=C(1) 1.295(5) & 1.296(5), N–O(1)
as proved by COSY NMR spectroscopy (see Figure S2 in 1.389(4) & 1.384(4).
Eur. J. Inorg. Chem. 2010, 1963–1972
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1965

M. R. Bryce et al.
FULL PAPER
metal coordinating plane (PtCl2SN), whereas the benzene
The bands centred at ca. 450 nm (Figure 4) are not due
rings are inclined to the plane of C(1) by 40–50° in a propel- to Raman scattering (there is a Raman scattering peak in
ler-like fashion.
toluene at ca. 455 nm when excited at 400 nm). Firstly, we
depict the spectra of highly concentrated (10–3 ) solutions
in Figure 4 in order to increase the luminescence emission.
This is valid as we did not observe any significant lumines-
Photophysical Properties
Figures 4 and 5 depict the PL spectra of the complexes cence spectral shape changes with an increase of concentra-
shown in Figure 2. Complexes 12b, 13b and 14b (Figure 4) tion from 10–5 to 10–3 . Furthermore, the band position
have two broad bands each showing vibronic features, one does not change when the excitation wavelength is changed
band centred at ca. 450 nm and the other at ca. 550 nm, the (Figure S6 in Supporting Information). This would not be
latter one blueshifted by ca. 20 nm for complex 13b.
the case for Raman bands. The same argument, based on
We observed that bubbling nitrogen through the solution changing the excitation wavelength, is valid for 17b (Fig-
affects only the 550 nm band, whereas the other band does ures 5 and S7 in Supporting Information). For complexes
not change.
19 and 20 (Figure 5b) peaks at 456 nm are indeed due to
The other notable feature of the broad bands centred at Raman scattering as we observe a shift with the excitation
ca. 550 nm is the very long lifetime even in a non-degassed wavelength. However, there is no singlet character emission
toluene solution. The lifetime data for complex 12b (144 ns) at ca. 450 nm for these complexes. In complex 18b the peak
is shown in Figure S3 in the Supporting Information. This at 455 nm is additionally a Raman peak, but unlike in 19
contrasts with the sub-nanosecond lifetime of the bands and 20 the singlet character emission at ca. 450 nm is pres-
centred at ca. 450 nm. Due to technical limitations, our life- ent. This is clearly demonstrated in Figure S8 in the Sup-
time-measurement system could not measure such short porting Information. We could not conceal the Raman
lifetimes precisely. Thus, there is clear indication that the peaks in 18b, 19 and 20 by increasing the concentration
bands at ca. 450 nm are of singlet character since the life- (as was done with the complexes in Figure 4), because they
times are very short and are not affected by oxygen, exhibited very efficient concentration quenching, and high
whereas the bands at ca. 550 nm are of triplet character concentration solutions were very weakly emissive.
since their lifetimes are much longer than that of conven-
tional fluorophores and the 550 nm bands are quenched by were performed. The isoemissive point, which is clearly
oxygen (see Figure S4 in the Supporting Information). present (Figure 6), shows that both species are interdepen-
In addition, variable-temperature studies of complex 13b
Figure 4. Photoluminescence spectra of complexes 12b, 13b, 14b (a) and 15b, 16b, 21 (b) in 10–3  toluene solution excited at 400 nm.
Figure 5. Photoluminescence spectra of complexes 17b, 18b (a) and 19, 20 (b) in 10–5  toluene solution. Excited at 400 nm (except 17b,
at 355 nm).
1966
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 1963–1972

Luminescent Platinum(II) Complexes
Figure 6. Photoluminescence spectra of complex 13b: (a) in non-degassed 10–3  toluene solution over the temperature range between
163 and 223 K, excited at 385 nm, and (b) expansion to show the isoemissive point at 483 nm.
Table 1. Photophysical data.
Complex
Fluorescence maxima
[nm]
Phosphorescence maxima
[nm]
Phosphorescence quantum yield ΦPL
[%]
12b
13b
14b
15b
16b
17b
18b
19
434, 457
431, 455
433, 455
534, 567
518, 549
531, 553
529, 562
529, 562
525, 556
588
0.007[d]
Ͻ0.007[e]
Ͻ0.007[e]
4.3[c]
[a]
–
–
[a]
3.0[c]
375, 395
Ͻ0.007[e]
Ͻ0.007[e]
Ͻ0.007[e]
Ͻ0.007[e]
0
[b]
–
[a]
–
598
[a]
20
–
591
[a]
21
431, 454
–
[a] No emission observed. [b] Not possible to determine due to Raman peak of toluene at 455 nm. [c] Determined by using 9,10-
diphenylanthracene as a standard[30] for the lower-energy band from 483 to 700 nm; errorsϮ5%. [d] Taken from ref.[22] [e] Emission was
much less than that of compound 12b.
dent on one another, i.e. in that temperature range species is an efficient non-radiative singlet decay channel, which ef-
with triplet character increase at the expense of singlet- fectively quenches all the singlets and/or singlet-to-triplet
character species. Similar isoemissive-point observations harvesting,[31] and only those states survive that cross to the
have been reported in the literature.[29]
triplet manifold.
Furthermore, the absorption band at ca. 400 nm (Fig-
ure S5 in the Supporting Information) can be assigned
mixed-ligand and MLCT character based on theoretical
calculations (see below) and similarities in the energy and
Application in Organic Light-Emitting Devices
Preliminary experiments were performed to demonstrate
intensity of MLCT transitions in other platinum com- the applicability of cyclometallated diaryl ketimine plati-
plexes.[22] The emission bands at λmax ≈ 450 and 550 nm are num(II) complexes as emitters in OLEDs. As complex
assigned to mixed-ligand and MLCT states having signifi- 15b showed the highest PL efficiency it was chosen as an
cant singlet and triplet character, respectively, in agreement emissive dopant in a poly(vinylcarbazole) (PVK) host.
with the theoretical calculations (see below).
The device structure used was ITO/PEDOT-PSS/
Until now the reported quantum efficiencies of cyclo- PVK:40% PBD:12% 15b/Ba/Al. The turn-on voltage is ca.
metallated diaryl ketimine platinum(II) complexes have 8 V (Figure 7), which mostly depends not on the dopant but
been very low.[22] However, by varying the ligand structure on the carrier injection barriers into PVK. The maximum
we have succeeded in increasing the efficiency of phospho- external quantum efficiency of the devices was 0.25% with
rescence almost 500-fold for 15b and 16b (Table 1) in com- a maximum brightness of ca. 18 cd/m2 (Figures S9 and S10
parison to that of the parent complex 12b.[22]
in the Supporting Information). Optimisation of the device
From these observations, we see that for systems with architecture to enhance these values is beyond the scope of
incomplete intersystem crossing (ISC), very poor total emis- this study. It should be noted that the broad electrolumi-
sion quantum yields are measured, whereas in the two com- nescence band spanning the 500–775 nm range (Figure 8)
plexes 15b and 16b, where no obvious fluorescence type makes these complexes attractive candidates for white
emission is observed, there is a concomitant dramatic in- OLEDs.[32] Comparison with the solid-state spectrum of the
crease in total emission, which must all arise from phospho- parent complex 12b and the electroluminescence spectrum
rescence. This implies that in this family of complexes there are shown in Figure S11 in the Supporting Information.
Eur. J. Inorg. Chem. 2010, 1963–1972
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1967

M. R. Bryce et al.
FULL PAPER
pected lowest-energy transition (HOMOǞLUMO) in the
absorption spectrum of 12b would be a mixed chlorido–
metal–ligand to mainly π*-C=N charge transfer. The fron-
tier orbitals (HOMO and LUMO) for 13b–16b, 19–21 are
also of similar characteristics with HOMOs containing 34–
38% Pt-dxz character (Table 2; see Supporting Information
for MO details). Whereas the LUMOs are similar, the
HOMOs for 17b and 18b are different, and these com-
pounds are discussed in detail later.
Figure 7. I/V characteristics for an OLED with complex 15b as the
emissive dopant.
Figure 9. HOMO and LUMO for 12b plotted at 0.04 (e/bohr3)1/2
These orbitals were similarly found for 13b–16b, 19–21.
.
TD-DFT data on the S0 geometries of the cyclometall-
ated complexes 12b–16b, 19–21 show the lowest-energy
transitions to be at 486–523 nm for S0ǞT1 and 401–434 nm
for S0ǞS1 with low oscillation strengths. The latter values
are in good agreement with the weak bands observed at
386–405 nm in the absorption spectra for six of the ten
complexes (Figure S13 in the Supporting Information). We
were not able to observe the expected very weak S0ǞT1
bands at around 500 nm here. The strong bands predicted
Figure 8. Electroluminescence (EL) spectrum of an OLED with
complex 15b as the emissive dopant.
Theoretical Calculations
Given the varied photophysical data obtained for the cy- at around 300 nm for these complexes correspond to the
clometallated systems here, their geometries and electronic aryl-πǞπ*-C=N,aryl-π* transitions and are in agreement
structures were investigated computationally. The optimised with intense bands observed elsewhere[22] in the region of
ground-state (S0) geometry for 12b at the B3LYP// 300 nm for 12b and related complexes.
LANL2DZ/6-311G** level of theory is in reasonable agree-
As the HOMOs and LUMOs of 12b–16b and 19–21 and
ment with the experimentally (X-ray) determined geometry their HOMO–LUMO energy gaps are similar, their emis-
of 12b.[22] (See the Supporting Information for the compu- sion data might have been expected to be similar. However,
tational and geometric details.) This level of theory was ap- the emission data are different with both fluorescence and
plied to calculations on all cyclometallated systems synthe- phosphorescence (dual emission) observed for 12b–14b,
sised here.
strong phosphorescence for the fused ring systems (15b and
The HOMO and LUMO for the S0 geometry of 12b are 16b) and strong fluorescence for the N-phenyl system 21.
depicted in Figure 9. The HOMO is located on the Cl–Pt– Very weak luminescence was found for the acac systems,
C6H4 axis, whereas the LUMO is located mainly on the 19 and 20. The HOMO–1 and LUMO+1 were therefore
C=N bond. The HOMO contains 37% Pt character, examined as they are expected to be involved in these emis-
whereas the LUMO has 8% Pt character. Thus, the ex- sions. Figure 10 shows these orbitals for 12b, 15b, 19, 20
Table 2. Orbital energies in eV and % Pt-MO compositions for frontier orbitals of complexes 12b–18b and 19–21.
12b
13b
14b
15b
16b
17b
18b
19
20
21
LUMO+1
% Pt
LUMO
% Pt
–1.09
16
–2.37
8
–1.38
3
–2.55
8
–0.97
4
–2.04
8
–1.79
2
–2.34
8
–1.78
2
–2.34
8
–1.41
1
–1.97
8
–1.42
37
–2.39
8
–1.06
2
–2.06
8
–1.55
3
–2.06
7
–1.20
33
–2.26
7
HOMO
% Pt
–6.02
37
–6.65
83
–6.28
37
–6.87
63
–5.84
38
–6.13
10
–6.04
37
–6.54
3
–6.09
35
–6.46
1
–5.59
19
–5.63
3
–6.19
27
–6.89
17
–5.73
34
–5.87
41
–5.73
34
–5.85
40
–5.93
37
–6.41
17
HOMO–1
% Pt
HLG[a]
3.65
3.73
3.80
3.70
3.75
3.62
3.80
3.67
3.67
3.67
[a] HOMO–LUMO energy gap.
1968
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 1963–1972

Luminescent Platinum(II) Complexes
Figure 10. HOMO–1 and LUMO+1 plots for 12b, 15b, 19, 20 and 21 at 0.04 (e/bohr3)1/2
.
and 21. Clearly, these orbitals vary with significant metal 1.299 Å in S0 to 1.378 Å in T1 (see Figure S8 in the Sup-
character in the HOMO–1s for 12b, 19 and 20 and in the porting Information for details). Table 3 lists the C=N bond
LUMO+1 for 21. The Pt-dxz orbitals are found in the lengths for both geometries of all systems studied here.
HOMO–1s for the acac complexes close in energies (differ- There are significantly longer C=N bonds in the T1 geome-
ences of only 0.12–0.14 eV) with their HOMOs containing tries for 19–21 compared to those for 12b–16b.
Pt-dyz orbitals. Compound 12b, on the other hand, has the
Pt-dz2 orbital character in HOMO–1. The LUMO+1 of 21 point energy calculations [T1(S0)] indicate how different the
contains substantial Pt*-dxy orbital character. T1 geometries are compared to the S0 geometries. The
The energies of the optimised T1 geometries at S0 single-
Whereas the S0ǞS1 data from TD-DFT computations [T1(S0)] energies for 19–21 are large (0.47–0.54 eV) with re-
on S0 geometries are in agreement with fluorescence data spect to 12b–16b (0.33–0.37 eV). The well-known efficient
(S0ǟS1) for the acac systems 19 and 20, the S0ǞT1 data do emitter Ir(ppy)3 has a [T1(S0)] value of only 0.22 eV[33] re-
not tally with their observed phosphorescence (S0ǟT1) flecting little geometric change between these two states.
data. The S1 geometries may, therefore, be similar to S0 geo- Table 3 also lists the computed energy differences between
metries, but the T1 geometries may be significantly different S0 and T1 geometries (T1 – S0), which agree remarkably well
in 19 and 20. The excited triplet-state (T1) geometries of all with the observed phosphorescence data (Figure 11). The
the systems were thus optimised to aid interpretation of the least successful fits involve 14b and 18b, where hydroxy and
phosphorescence data. The optimised T1 geometry of 12b methoxy groups are present, and thus orientations of these
contains some significant geometrical changes compared to groups are likely to have subtle effects on the energies of
the S0 geometry, notably the C=N bond length from the geometries.
Table 3. Energies and C=N bond lengths for S0 and T1 optimised geometries and comparison of observed and computed photophysical
data for complexes 12b–18b, 19–21.
12b
13b
14b
15b
16b
17b
18b
19
20
21
S0ǞT1 [eV][a]
S0ǞS1 [eV][a]
T1 – S0 [eV][b]
[T1(S0)] [eV][c]
2.47
2.90
2.32
0.37
2.54
3.00
2.40
0.36
2.55
3.09
2.42
0.33
2.51
2.96
2.35
0.35
2.51
2.97
2.36
0.33
2.47
2.92
2.36
0.26
2.30
3.07
2.01
0.65
2.37
2.85
2.13
0.51
2.38
2.88
2.14
0.47
2.47
2.95
2.22
0.54
S0(C=N) [Å]
T1(C=N) [Å]
∆(C=N) [Å][d]
1.299
1.378
0.079
1.300
1.391
0.091
1.303
1.398
0.095
1.297
1.374
0.077
1.297
1.372
0.075
1.304
1.382
0.078
1.298
1.438
0.140
1.305
1.428
0.123
1.304
1.429
0.125
1.307
1.432
0.125
S0ǞT1 [nm][a]
T1 – S0 [nm][b]
501
534
487
512
486
517
494
525
493
525
529 (s)
501
525
540
617
523
582
522
582
503
558
–
[g]
S0ǟT1 phosphorescence[e] 534 (w) 518 (w) 531 (w) 529 (s)
525 (w)[f] 588 (vw) 598 (vw) 591 (vw)
S0ǞS1 [nm][a]
427
413
401
–
419
405 (w)
–
417
425
–
400
–
–
434
–
–
430
421
[i]
[i]
[i]
[i]
S0ǞS1 absorption[e,h]
S0ǟS1 fluorescence[e]
400 (w) 386 (w)
434 (w) 431 (w) 433 (w)
405 (w)
400 (w) 386 (w)
431 (s)
435 (vw) 375 (s)[f]
–
[a] From TD-DFT data on optimised S0 geometries. [b] Difference between energies of optimised T1 and S0 geometries. [c] Difference
between the energy of a single-point S0 calculation on the optimised T1 geometry and the energy of the optimised S0 geometry. [d]
Difference between C=N bond lengths of S0 and T1 geometries. [e] Observed data, the intensities listed (s = strong, w = weak, vw = very
weak) are relative, excitation at 400 nm. [f] Excitation at 355 nm. [g] No emission observed. [h] The expected S0ǞT1 absorption bands
were too weak to be observed. [i] Not recorded.
Eur. J. Inorg. Chem. 2010, 1963–1972
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1969

M. R. Bryce et al.
FULL PAPER
third factor is the relatively small geometric change ob-
served between the S0 and T1 optimised geometries, pro-
moting the radiative decay. Thus, from the phosphorescence
band shapes observed for 15b and 16b, the transitions in-
3
3
volved are likely to be a mixture of LLT and MLCT in
character where the fused-ring π, π*, Pt and π*-C=N orbit-
als are major contributors.
Both weak fluorescence and weak phosphorescence are
observed for the parent dmso complex 12b and the fluorine-
and methoxy-substituted analogues, 13b and 14b, respec-
tively. They have no such conditions that would either result
in strong fluorescence (like 21) or very weak luminescence
(like 19 and 20). The metal character in the HOMOs and
T1 geometries for 12b–14b with the fused-ring systems (15b
and 16b) are similar. It is suggested that the weak dual emis-
sions observed in 12b–14b are due to the larger energy dif-
ferences between the aryl-π* LUMO+1 and HOMO orbit-
als in 12b–14b (4.9 eV) compared to those in 15b and 16b
(4.3 eV). The luminescence-band patterns observed for
these systems indicate Pt,aryl-πǟπ*-C=N,aryl-π* transi-
tions with LLT and MLCT character in both singlet and
Figure 11. Correlation between observed phosphorescence data for
12b–18b, 19 and 20 and the differences between computed total
energies of optimised T1 and S0 geometries.
It is clear that the T1 excited-state geometries of the acac triplet states.
complexes 19 and 20 are very different to the ground-state
The frontier orbitals for 17b and 18b are shown in Fig-
geometries, and thus the phosphorescence S0ǟT1 transi- ure 12. Methyl groups are used instead of octyl groups for
tions are lower in energy than those of the related dmso 17b to reduce the computational effort. The HOMOs have
complexes 12b and 16b, as observed experimentally. The en- lower Pt character at 19% for 17b and 27% for 18b com-
ergy-gap law predicts that the non-radiative decay rate con- pared to other complexes discussed above. The carbazole
stant increases with decreasing energy difference between ligands dominate the HOMO–1 (88% on cyclometallated
the optimised T1 and S0 geometries.[34]
carbazole ligand), HOMO (74% on cyclometallated carb-
The significantly different S1 and T1 geometries (as- azole ligand) and LUMO+1 (96% on second carbazole li-
suming S1 has a similar geometry to S0) for the acac com- gand) in 17b. The much smaller heavy-metal contribution
plexes 19 and 20 are responsible for the low phosphores- in 17b (and thus little ISC promotion) accounts for very
cence observed for these complexes. The presence of un- weak phosphorescence. The strong fluorescence in 17b ef-
favourable acac (26% in HOMO and 26–30% in HOMO– fectively arises from carbazolyl-πǟπ*-C=N,carbazolyl-π*
1) and metal character in the highest occupied orbitals are transitions.
likely to be responsible for the low fluorescence observed
for the acac complexes. Strong fluorescence is often associ-
ated with aryl-πǞπ* transitions.
The N-phenyl system 21 also has a substantially different
T1 geometry. The absence of phosphorescence emission in
21, where the maximum would be predicted to be at
558 nm, is most likely due to significantly different S1 and
T1 geometries (assuming S1 has a similar geometry to S0).
The unfavourable Pt(d)ǞPt(d*) quenching of the radiative
decay with the LUMO+1 as the Pt* orbital may also have
a role in the lack of phosphorescene emission observed for
21.[3] The strong fluorescence observed for 21 (Figure 4)
arises from N-phenyl π-orbital contributions (38% in
HOMO–1) to the Pt,aryl-πǟπ*-C=N,aryl-π* transitions.
The strong phosphorescence observed for the fused-ring
dmso complexes 15b and 16b is due to three highly favour-
able conditions. Firstly, the LUMO+1s and HOMO–1s
consist of essentially the π* (81–83%) and π (90–92%) or-
bitals of the naphthyl and anthracyl groups, respectively,
with the LUMO+1s considerably close to the LUMOs in
energies (0.55–0.56 eV). The second condition is the sub-
stantial heavy-metal character in the HOMOs promoting
Figure 12. Frontier orbitals for 17b and 18b plotted at 0.04 (e/
intersystem crossing (ISC) from singlet to triplet states. The bohr3)1/2
.
1970
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 1963–1972

Luminescent Platinum(II) Complexes
water. The solution was stirred at room temperature for 4–5 h, until
the colour turned from red to yellow indicating that
K[PtCl3(Me2SO)] had been generated. A solution of the diaryl ket-
imine (1 equiv.) in dry dichloromethane was added dropwise to the
yellow solution. The mixture was stirred at room temperature for
10 h. The organic layer was separated and the solvent removed un-
der reduced pressure to give the crude product, which was purified
by column chromatography to afford 12a–17a as yellow solids.
The very weak phosphorescence observed experimentally
for the N-hydroxy system 18b can be explained by the sig-
nificantly different S1 and T1 geometries (assuming S1 has
a similar geometry to S0) increasing the nonradiative decay
rate constant. The undesirable Pt–Pt*-dǞd* quenching of
the radiative decay with the LUMO+1 as the Pt* orbital
may also have a role in the weak phosphorescence emis-
sion.[3] The very weak fluorescence observed experimentally
for the N-hydroxy system 18b may be due to the unfavour-
able substantial Pt character in LUMO+1 (Figure 12) and
the low aryl character of the phenyl group in the frontier
orbitals.
General Procedure for Cyclometallation of the Complexed Imine: A
solution of complex 12a–17a in dry toluene (40 mL) was refluxed
for ca. 15 h. The solvent was removed under reduced pressure to
give the crude product, which was then purified by column
chromatography to give 12b–17b as yellow-orange solids.
Photophysical Measurements: The complexes were dissolved in
toluene. Steady-state absorption spectra and luminescence emission
and excitation spectra of the solutions were recorded by using a
UV/Vis spectrophotometer (Lambda 19 from Perkin–Elmer) and a
commercial spectrofluorimeter (Fluorolog from Jobin Yvon),
respectively. Lifetime measurements were made by using a system
consisting of excitation source, pulsed YAG laser emitting at
355 nm (EKSMA or CryLas GmbH). Samples were excited at 45°
angle to the substrate plane, and the energy of each pulse could be
tuned from µJ up to mJ. With the help of a spectrograph and other
optics, luminescence data were collected by a sensitive iCCD cam-
era (Stanford Computer Optics) with sub-nanosecond resolution.
For low-temperature measurements (down to 77 K) samples were
placed in a cryostat. Quantum efficiencies were determined by
using 9,10-diphenylanthracene as a standard,[30] where samples
were degassed by using at least three pump-thaw cycles before
measurements of the quantum efficiencies.
Conclusions
A series of platinum(II) complexes containing cyclo-
metallated diaryl ketimine ligands has been synthesised.
Their photophysical properties are characterised by emis-
sion bands at λmax ≈ 450 and 550 nm, which are assigned
to singlet and triplet species, respectively. By varying the
structure of the aromatic ligand the efficiency of phospho-
rescence can be increased from Ͻ0.007% for several deriva-
tives to 3.0 and 4.3% for 16b and 15b, respectively. Compre-
hensive theoretical calculations have rationalised the ob-
served photophysical properties of these systems. Computa-
tions have shown that the low-energy transitions in all sys-
tems involve mainly the frontier orbitals, HOMO and
LUMO. These are mixed chlorido–metal–ligand to largely
π*-C=N transitions. Most of the phosphorescence data ob-
served here can be explained by the geometric change on
going from the S0 to the T1 states. The very weak (or zero)
phosphorescence emissions observed for some systems (18b,
19–21) are due to the large changes in geometry and energy
between the S0 and T1 states. The relatively strong phospho-
rescence emissions observed in fused-ring systems (15b and
16b) are due to significant aryl-π or -π* character in their
HOMO–1, HOMO and LUMO+1 and substantial Pt char-
acter (35–37%) in the HOMO promoting favourable ISC
combined with the relatively small geometric change on go-
ing from the S0 to the T1 states. The carbazole system 17b
Supporting Information (see footnote on the first page of this arti-
cle): Details of the synthesis and characterisation of all new com-
pounds; additional photophysical data and spectra; X-ray crystal-
lographic data for 18a; device fabrication and characterisation;
computational data for 12b–18b,19–21.
Acknowledgments
The authors thank One North East, EPSRC, Durham University
and Kodak for funding. We thank Durham University for access
to its High Performance Computing Cluster.
[1] M. S. Lowry, S. Bernhard, Chem. Eur. J. 2006, 12, 7970.
has very weak phosphorescence due to little metal character [2] K. K.-W. Lo, W.-K. Hui, C.-K. Cheng, K. H.-K. Tsang, D. C.-
M. Ng, N. Zhu, K.-K. Cheung, Coord. Chem. Rev. 2005, 249,
(19%) in the HOMO (thus negligible ISC takes place) but
1434.
has strong fluorescence with substantial aryl-π or -π* char-
[3] J. A. G. Williams, S. Develay, D. L. Rochester, L. Murphy,
acter in the frontier orbitals involved in the transitions.
Coord. Chem. Rev. 2008, 252, 2596.
Weak dual emissions observed for the simpler systems (12b–
14b), compared to strong phosphorescene found for fused-
ring systems, are attributed to much less aryl-π* character
in the transitions. This work should stimulate further stud-
ies on luminescent platinum(II) complexes incorporating
[4] Y. Li, A. Rizzo, M. Salerno, M. Mazzeo, C. Huo, Y. Wang, K.
Li, R. Cingolani, G. Gigli, Appl. Phys. Lett. 2006, 89, 061125
and references cited therein.
[5] Q. Huo, Y. Zhang, F. Li, J. Ping, Y. Cao, Organometallics 2005,
24, 4509.
[6] I. R. Laskar, S. F. Hsu, T. M. Chen, Polyhedron 2005, 24, 881.
new ligands, aimed at exploring fundamental photophysical [7] Z. He, W. Y. Wong, X. Yu, H. S. Kwok, Z. Lin, Inorg. Chem.
properties and practical applications.
2006, 45, 10922.
[8] M. Cocchi, J. Kalinowski, D. Virgili, V. Fattori, S. Develay,
J. A. G. Williams, Appl. Phys. Lett. 2007, 90, 163508.
[9] X. Yang, Z. Wang, S. Madakuni, J. Li, G. E. Jabbour, Adv.
Experimental Section
Mater. 2008, 20, 2405.
[10] E. L. Williams, K. Haavisto, J. Li, G. E. Jabbour, Adv. Mater.
General Procedure for the Reaction of Diaryl Ketimines 6–11 with
K[PtCl3(Me2SO)]: A solution of dimethyl sulfoxide (1.04 equiv.) in
water was added dropwise to a solution of K2[PtCl4] (1 equiv.) in
2007, 19, 197.
[11] J. A. G. Willlams, A. Beeby, E. S. Davies, J. A. Weinstein, C.
Wilson, Inorg. Chem. 2003, 42, 8609.
Eur. J. Inorg. Chem. 2010, 1963–1972
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1971

M. R. Bryce et al.
FULL PAPER
[12] W. Sotoyama, T. Satoh, N. Sawatari, H. Inoue, Appl. Phys.
Lett. 2005, 86, 153505.
[13] S. J. Farley, D. L. Rochester, A. L. Thompson, J. A. K.
Howard, J. A. G. Williams, Inorg. Chem. 2005, 44, 9690.
[14] X. Q. Hao, J. F. Gong, C. X. Du, L. Y. Wu, Y. J. Wu, M. P.
Song, Tetrahedron Lett. 2006, 47, 5033.
2603; f) W. Tong, L. Lai, M. C. W. Chan, Dalton Trans. 2008,
1412; g) X. Lü, W. Wong, W. Wong, Eur. J. Inorg. Chem. 2008,
523.
[22] Y. Y. Scaffidi-Domianello, A. A. Nazarov, M. Haukka, M. Ga-
lanski, B. K. Keppler, J. Schneider, P. Du, R. Eisenberg, V. Y.
Kukushkin, Inorg. Chem. 2007, 46, 4469.
[15] C. Baik, W. S. Han, Y. Kang, S. O. Kang, J. Ko, J. Organomet.
Chem. 2006, 691, 5900.
[16] W. Lu, B. X. Mi, M. C. W. Chan, Z. Hui, C. M. Che, N. Zhu,
S. T. Lee, J. Am. Chem. Soc. 2004, 126, 4958.
[17] S. C. F. Kui, I. H. T. Sham, C. C. C. Cheung, C. W. Ma, B. Yan,
N. Zhu, C. M. Che, W. F. Fu, Chem. Eur. J. 2007, 13, 417.
[18] a) W. Lu, M. C. W. Chan, N. Zhu, C. M. Che, C. Li, Z. Hui,
J. Am. Chem. Soc. 2004, 126, 7639; b) T.-C. Cheung, K.-K.
Cheung, S.-M. Peng, C.-M. Che, J. Chem. Soc., Dalton Trans.
1996, 1645; c) C.-K. Koo, K.-L. Wong, C. W.-Y. Man, Y.-w.
Lam, L. K.-Y. So, H.-L. Tam, S.-W. Tsao, K.-W. Cheah, K.-C.
[23] A. Lachman, Org. Synth. Coll. Vol. 1943, 2, 234.
[24] P. L. Pickard, Org. Synth. 1964, 44, 51.
[25] a) A. D. Ryabov, S. Otto, P. V. Samuleev, V. A. Polyakov, L.
Alexandrova, G. M. Kazankov, S. Shova, M. Revenco, J.
Lipkowski, M. H. Johansson, Inorg. Chem. 2002, 41, 4286; b)
S. Otto, A. Chanda, P. V. Samuleev, A. D. Ryabov, Eur. J. Inorg.
Chem. 2006, 2561.
[26] C. Eaborn, K. Kundu, A. Pidcock, J. Chem. Soc., Dalton
Trans. 1981, 933.
[27] L. V. Konovalov, V. G. Pogareva, J. Struct. Chem. 1981, 22,
438.
Lau, Y.-Y. Yang, J.-C. Chen, M. H.-W. Lam, Inorg. Chem. [28] a) N. Nedelec, F. D. Rochon, Inorg. Chem. 2001, 40, 5236; b)
2009, 48, 872.
N. Nedelec, F. D. Rochon, Inorg. Chim. Acta 2001, 319, 95.
[29] Z. R. Grabowski, K. Rotkiewicz, W. Rettig, Chem. Rev. 2003,
103, 3899.
[30] J. R. Lakowitz, Principles of Fluorescence Spectroscopy, Kluwer
Academic/Plenum Publishers, New York, 1999.
[31] H. Yersin, Top. Curr. Chem. 2004, 241, 1.
[32] Review: K. T. Kamtekar, M. R. Bryce, A. P. Monkman, Adv.
Mater. 2010, 22, 572.
[19] W. Sun, H. Zhu, P. M. Barron, Chem. Mater. 2006, 18, 2602.
[20] J. R. Berenguer, E. Lalinde, J. Torroba, Inorg. Chem. 2007, 46,
9919.
[21] a) X. Riera, C. López, A. Caubet, V. Moreno, X. Solens, M.
Font-Bardía, Eur. J. Inorg. Chem. 2001, 2135; b) A. Caubet, C.
Lopez, X. Solans, M. Font-Bardía, J. Organomet. Chem. 2003,
669, 164; c) A. Capapé, M. Crespo, J. Grannel, A. Vizcarro, J.
Zafrilla, M. Font-Bardía, X. Solans, Chem. Commun. 2006, [33] Y. Wu, J. L. Bredas, J. Chem. Phys. 2008, 129, 214305.
4128; d) A. Capapé, M. Crespo, J. Grannel, M. Font-Bardía,
X. Solans, Dalton Trans. 2007, 2030; e) M. Crespo, R. Martín,
T. Calvert, M. Font-Bardía, X. Solans, Polyhedron 2008, 27,
[34] G. S. Tong, C. M. Che, Chem. Eur. J. 2009, 15, 7225.
Received: November 30, 2009
Published Online: March 24, 2010
1972
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 1963–1972