Phosphorescent heteroleptic iridium(III) cyclometallates: Improved syntheses of acetylacetonate complexes and quantum chemical studies of their excited state properties

Article abstract of DOI:10.1016/j.poly.2019.114256

We have investigated methods to prepare cyclometallated iridium(III) complexes with efficient photoluminescence spanning a broad color palette. In particular, we find that addition of ancillary ligands to chloro-bridged iridium dimers proceeds cleanly in

Full text of DOI:10.1016/j.poly.2019.114256

Journal Pre-proofs
Phosphorescent Heteroleptic Iridium(III) Cyclometallates: Improved Syntheses
of Acetylacetonate Complexes and Quantum Chemical Studies of Their Excited
State Properties
Robert D. Sanner, Nerine J. Cherepy, Hung Q. Pham, Victor G. Young Jr
PII:
S0277-5387(19)30701-6
DOI:
https://doi.org/10.1016/j.poly.2019.114256
POLY 114256
Reference:
Polyhedron
To appear in:
Received Date:
Revised Date:
Accepted Date:
26 August 2019
21 November 2019
2 December 2019
Please cite this article as: R.D. Sanner, N.J. Cherepy, H.Q. Pham, V.G. Young Jr, Phosphorescent Heteroleptic
Iridium(III) Cyclometallates: Improved Syntheses of Acetylacetonate Complexes and Quantum Chemical Studies
of Their Excited State Properties, Polyhedron (2019), doi: https://doi.org/10.1016/j.poly.2019.114256
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will
undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing
this version to give early visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Elsevier Ltd. All rights reserved.

Phosphorescent Heteroleptic Iridium(III) Cyclometallates: Improved
Syntheses of Acetylacetonate Complexes and Quantum Chemical Studies of
Their Excited State Properties
Robert D. Sanner,*† Nerine J. Cherepy†, Hung Q. Pham,‡ and Victor G. Young, Jr.‡
†Lawrence Livermore National Laboratory, Livermore, CA 94550, United States
‡Department of Chemistry, University of Minnesota, Minneapolis, MN 55455, United States
ABSTRACT: We have investigated methods to prepare cyclometallated iridium(III)
complexes with efficient photoluminescence spanning a broad color palette. In
particular, we find that addition of ancillary ligands to chloro-bridged iridium
dimers proceeds cleanly in refluxing 1,2-dimethoxyethane (DME) without the need
for additional product purification. This represents an improvement over the
conventional use of 2-ethoxyethanol which requires column chromatographic
separation. Our efforts in this work have focused on acetylacetonate complexes
such as (F2ppy)2Ir(acac), where F2ppy = 2-(4’,6’-diflurophenyl)pyridinato. We have
prepared fifteen compounds by the route, eight of which are newly reported; in four
cases we were able to prepare complexes which were inaccessible via the
conventional route. Nine of the complexes were characterized by single crystal x-
ray diffraction and possess the same distorted octahedral geometry around the
iridium with two bidentate phenylpyridine ligands and one bidentate
acetylacetonate ligand. Seven of the complexes exhibited efficient
photoluminescence with colors ranging from yellow to blue and quantum yields of
0.51-0.74. All of the compounds with trifluoromethyl or phenyl substituents on the
acetylacetone displayed emission in the orange with low quantum efficiency. The
use of TD-DFT calculations, along with natural transition orbitals (NTOs), permitted
a detailed interpretation of the electronic structures for the complexes. The nature
of the acceptor orbitals for the low energy triplet state NTOs proved to be an
important predictor for the emission spectra of the compounds.
Keywords:
Photoluminescence
Phosphorescent iridium complex
DFT calculations
OLED
Cyclometallated iridium synthesis
1
1
4
11
8

INTRODUCTION
Phosphorescent cyclometallated iridium(III) compounds have been widely
studied for diverse applications such as solar cells,1-6 sensors,7-10 bioimaging,11-13
and scintillators.14-18 Their electrophosphorescence, in which light emission is
triggered by an electric charge, has also prompted research in the areas of light-
emitting electrochemical cells,19-22 electrogenerated chemiluminescence,23-26 and
organic light-emitting diodes (OLED’s).27-36 The strong spin-orbit coupling
engendered by the iridium atom allows for the mixing of singlet and triplet excited
states through efficient intersystem crossing.37-40 Phosphorescence quantum yields
are thus increased beyond that obtainable with solely fluorescent emitters.41 These
cyclometallated iridium(III) complexes may be classified based on their associated
ligands. Thus, homoleptic compounds such as Ir(ppy)342 (where ppy = 2-
phenylpyridinato) or Ir(F2ppy)343 (where F2ppy = 2-(4’,6’-
difluorophenyl)pyridinato) possess identical ligands. These are differentiated from
heteroleptic compounds such as (F2ppy)2Ir(pic)44 (where pic = picolinate or
pyridine-2-carboxylate) or (F2ppy)(ppy)Ir(acac)45 (where acac = acetylacetonate)
that contain at least two different ligands. The cyclometallated ligands, for instance
phenylpyridine, typically contain a (C^N) core and are distinguished from the
ancillary ligands such as picolinate or acetylacetonate. Although (F2ppy)2Ir(pic),
commonly referred to as FIrpic, has received the most study of the heteroleptic
compounds, 44, 46 the class of complexes containing acetylacetonate as the ancillary
ligand has also been a fertile area of research. Since early reports of (ppy)2Ir(acac)42
and (F2ppy)2Ir(acac), also known as FIracac,47 an assortment of substituted β-
diketonates have appeared. They have been the subject of spectroscopic38 and
theoretical48-51 studies. The synthesis of these complexes typically involves a two-
step process. First, the cyclometallating ligand is reacted with IrCl3 · x H2O following
the example of Nonoyama52 or Sprouse53 to form a chloro-bridged dimeric iridium
compound, e.g. [(ppy)2IrCl]2.53, 54 Next, the ancillary ligand is added by reacting the
iridium dimer with acetylacetone in refluxing 2-ethoxyethanol and base (usually
Na2CO3). 23, 47, 55, 56 In a variation, Baranoff 45 has used tetrabutylammonium
hydroxide as base in refluxing dichloromethane. After the second step it is
necessary to perform additional purification, usually concluding with column
chromatography. As part of our synthesis program for plastic scintillators17, 57 we
have investigated iridium complexes that emit in the deep blue.58 We have now
expanded our spectral range to include longer wavelength emitters, concentrating
on heteroleptic iridium(III) (bis)cyclometallates. This effort has led us to report an
improved synthesis of FIrpic and other picolinates:59 We found that refluxing 1,2-
dimethoxyethane in the presence of sodium carbonate made an excellent reaction
medium for addition of the ancillary picolinate ligand to the chloro-bridged iridium
dimer. Furthermore, we were able to obtain yields equal to or better than those
reported in prior procedures without the need for time-consuming purification
steps (such as column chromatography). We have now broadened the scope of this
reaction by the investigation of acetylacetonates as the ancillary ligand and report
those results herein. The products were characterized by NMR and optical
spectroscopy, x-ray crystallography, and photophysical measurements. Time-
2

dependent density functional theory (TD-DFT) calculations were performed on the
complexes with the goal of interpretation of their absorption and emission spectra.
RESULTS AND DISCUSSION
Synthesis. The reaction to prepare heteroleptic iridium complexes
(X2ppy)2Ir(acac-Ri) using 1,2-dimethoxyethane at reflux with base is shown in
Scheme 1, where X = H or F and Ri are various substituents on the acetylacetonate
ligand, including methyl, trifluoromethyl, cyclopentyl, cyclohexyl, and phenyl.
Reaction times are typically 6-8 hr and workup consists of simple filtration and
washing; yields range from 76-92% except in two cases (2 and 13) which will be
discussed below. An identification guide to compounds 1-15 prepared in this work
is given in Chart 1.
Scheme 1. Synthesis of iridium(III) acetylacetonate complexes.
3

Chart 1. Heteroleptic iridium(III) acetylacetonates prepared in this study.
Compounds 1-7 employed [(ppy)2IrCl]253, 54 as the iridium source while 8-15 used
[(F2ppy)2IrCl]2.47, 60 Compounds 1,42, 45 2,61 4,61 5,61 7,54 8,44, 45, 47, 62 9,63 and 1555, 64
have been previously prepared, usually by mixing the iridium dimer with an
acetylacetone derivative in refluxing 2-ethoxyethanol and base; purification is
accomplished by column chromatography. In our reactions, column
chromatography was only required in the case of compound 13.
4

An NMR comparison of the FIracac obtained for 12 hr reactions by our DME
method and the ethoxyethanol method is shown in Fig. 1 (1H, aromatic region).
Note the higher purity of the crude product obtained in DME as demonstrated in the
bottom trace of the 1H NMR spectra. We believe that the gentler conditions of DME
1
Figure 1. H NMR spectra of crude FIracac prepared in ethoxyethanol (top) and dimethoxyethane (bottom).
reflux (85° C) compared to ethoxyethanol (134°) are at least partly responsible for
this difference. Our method did not result in any loss in yield as all starting material
was consumed and the desired product was cleanly obtained. We have found that
this reaction proceeds smoothly with both (ppy)2Ir and (F2ppy)2Ir moieties and
most acetylacetone derivatives we examined. Although we have only performed
this with a maximum 2 g of iridium starting material, scale-up should not be a
problem. In fact, the ability to omit column chromatography would be most useful
at larger scales.
A few cases deserve special mention. The diketonate ligand that contains
two trifluoromethyl groups, hexafluoroacetylacetone, was unreactive with
[(ppy)2IrCl]2 in either refluxing DME or methoxyethanol; starting materials were
recovered unchanged. The reaction of this ligand with [(F2ppy)2IrCl]2 to form 13
did occur but was very slow: after 24 hr reflux there was still ~20% iridium
starting material left as well as ~10% impurity. The product was successfully
isolated using column chromatography and Soxhlet extraction but with only 40%
yield. Attempts to prepare 13 in refluxing methoxyethanol were unsuccessful:
5

although a product was cleanly formed that contained the (F2ppy)2Ir moiety, 19F
NMR showed there were no CF3 groups present. Both acetylacetone ligands that we
examined containing one trifluoromethyl group, 1,1,1-trifluoro-2,4-pentanedione
and 4,4,4-trifluoro-1-phenyl-1,3-butanedione, reacted swiftly and cleanly with
[(ppy)2IrCl]2 and [(F2ppy)2IrCl]2 to form 3, 6, 10, and 14. It is noteworthy that using
the glycol ether route with refluxing ethoxyethanol gave only non-emissive products
in attempts to prepare 10 and 14, proving the superiority of the DME route. The
ligand 2-acetylcyclohexanone required much longer reaction times with
[(ppy)2IrCl]2 (24 hr) than with [(F2ppy)2IrCl]2 (6 hr) to go to completion, forming 5
and 12, respectively. We also attempted to prepare 5 in refluxing ethoxyethanol,
and an NMR comparison of the crude products obtained for the DME and
ethoxyethanol reactions is illustrated in Fig. 2. Again, as shown for FIracac in Fig. 1,
1
Figure 2. H NMR spectra of crude compound 5 prepared in ethoxyethanol (top) and dimethoxyethane (bottom).
DME provided a much cleaner product than ethoxyethanol. The 2-
acetylcyclopentanone ligand did not show the disparity in reaction rates between
[(ppy)2IrCl]2 and [(F2ppy)2IrCl]2 that was found with its cyclohexanone analog since
4 and 11 both exhibited facile formation in 6 hr at DME reflux. Finally, 3-
methylacetylacetone reacted quickly with [(F2ppy)2IrCl]2 to form 9 but slowly with
[(ppy)2IrCl]2 to form 2: 24 hr was required to consume all iridium starting material
with concomitant formation of the desired product and an unidentified compound
that proved difficult to separate. The attempt to form 2 in ethoxyethanol was much
less successful, as none of the desired product was formed in 24 hr. The previous
6

report61 of compound 2 shows NMR results that are at variance with ours (see
Experimental).
Crystal Structure Determinations. The molecular structure of compounds
3, 5, 6, and 9-14 have been confirmed by X-ray crystallography and ORTEP
diagrams of these complexes are shown in Figs. 3-11; crystal and structure
refinement data are provided in Supporting Information (Tables S1-S9). All the
complexes exhibit the same distorted octahedral geometry with two bidentate
phenylpyridine ligands (coordinated through the pyridine nitrogen atom and a
phenyl carbon atom) and one bidentate -pentanedionate framework ligand
(coordinated through the diketonate oxygen atoms). The iridium-bound nitrogen
atoms of the phenylpyridine ligands are trans to each other while the phenyl carbon
atoms bound to the iridium are cis. The two ligated oxygen atoms in the -
diketonate ligand are then trans to the phenyl carbon atoms of the phenylpyridine
ligand. The Ir-N and Ir-C distances for the phenylpyridine ligand are remarkably
consistent throughout these molecules, averaging 2.036(8) Å and 1.992(11) Å,
respectively. The Ir-O distances for the -diketonate ligand within a molecule are
nearly equal and this distance varies as expected among the various diketonates.
Thus, the ligands with electron-withdrawing CF3 groups show the longest Ir-O bond
lengths (e.g. 2.18 Å for 3) while those without these groups show shorter bonds (e.g.
2.11 Å for 9 and 12). The ligands with both a phenyl and a CF3 group fall midway
between these extremes (2.14 Å for 6 and 14). The Ir-O distance averaged over all
the complexes is 2.14 Å. This is an elongation of the mean value of 2.088 Å reported
in the Cambridge Crystallographic Database65 for an Ir-O bond, illustrating the trans
effect of the phenyl group in the phenylpyridine ligand for these molecules. The C-
Ir-N “bite” angle for the phenylpyridine ligands is consistent, averaging 81.0(3)° for
these complexes. As expected, there is more variation in the O-Ir-O angle for the
collection of diketonate ligands investigated, ranging from to 84.76(7)° for the
acetylcyclohexanone ligand in 5 to 90.0(1)° in the acetylcyclopentanone complex
11; the average for this angle over all the complexes is 87.7°. The phenyl and
pyridine rings in each phenylpyridine ligand are slightly twisted with respect to
each other across the C-C bond linking the two rings. The dihedral angle between
best planes for the two rings is typically in the range of 3°- 6°, although 6 shows a
larger twist angle of 10.1° for one of the ligands. Examination of a packing diagram
for 6 shows close intermolecular contacts between a chlorine atom on the
dichloromethane solvate molecule and two carbon atoms on the pyridine ring.
7

Figure 3. The molecular structure of compound 3a, with atom labeling and displacement ellipsoids
drawn at the 50% probability level. Hydrogen atoms have been removed for clarity. Two molecules
are found in the asymmetric unit; only molecule A is shown in the figure. Only one component of the
disordered trifluoropentanedionate ligand pair is shown. Selected bond distances (Å) and angles (°): Ir1A-
C1A 2.030(9), Ir1A-C12A 1.983(9), Ir1A-N1A 2.048(7), Ir1A-N2A 2.027(8), Ir1A-O1A 2.177(6), Ir1A-
O2A 2.183(6), C1A-Ir1A-N1A 81.0(3), C12A-Ir1A-N2A 80.9 (3), O1A-Ir1A-O2A 88.3(2).
Figure 4. The molecular structure of compound 5, with atom labeling and displacement ellipsoids
drawn at the 50% probability level. Hydrogen atoms and dichloromethane solvent have been
omitted for clarity. Only one component of the disordered acetylcyclohexanoate ligand pair is shown.
Selected bond distances (Å) and angles (°): Ir1-C1 2.001(3), Ir1-C12 1.990(3), Ir1-N1 2.028(2), Ir1-N2
2.041(2), Ir1-O1 2.142(2), Ir1-O2 2.147(2), C1-Ir1-N1 80.74(10), C12-Ir1-N2 80.88(10), O1-Ir1-O2
84.76(7).
8

Figure 5. The molecular structure of compound 6, with atom labeling and displacement ellipsoids
drawn at the 50% probability level. Hydrogen atoms and dichloromethane solvent have been
omitted for clarity. Selected bond distances (Å) and angles (°): Ir1-C1 1.986(8), Ir1-C12 1.986(9), Ir1-
N1 2.054(8), Ir1-N2 2.043(8), Ir1-O1 2.151 (5), Ir1-O2 2.138(7), C1-Ir1-N1 80.2(4), C12-Ir1-N2
80.5(4), O1-Ir1-O2 88.1(3).
Figure 6. The molecular structure of compound 9, with atom labeling and displacement ellipsoids
drawn at the 50% probability level. Hydrogen atoms have been removed for clarity. Selected bond
distances (Å) and angles (°): Ir1-C1 1.987(3), Ir1-C12 1.994(3), Ir1-N1 2.036(2), Ir1-N2 2.035(2), Ir1-
O1 2.108(2), Ir1-O2 2.120(2), C1-Ir1-N1 81.08(10), C12-Ir1-N2 81.05(10), O1-Ir1-O2 86.58(7).
9

Figure 7. The molecular structure of compound 10a, with atom labeling and displacement ellipsoids
drawn at the 50% probability level. Hydrogen atoms have been removed for clarity. Two molecules
are found in the asymmetric unit; only molecule A is shown in the figure. Only one component of the
disordered trifluoropentanedionate ligand pair is shown. Selected bond distances (Å) and angles (°): Ir1A-
C1A 1.988(10), Ir1A-C12A 1.976(10), Ir1A-N1A 2.040(8), Ir1A-N2A 2.038 (7), Ir1A-O1A 2.165(7),
Ir1A-O2A 2.117(7), C1A-Ir1A-N1A 80.4(4), C12A-Ir1A-N2A 81.0 (4), O1A-Ir1A-O2A 87.9(2).
Figure 8. The molecular structure of compound 11, with atom labeling and displacement ellipsoids
drawn at the 50% probability level. Hydrogen atoms have been removed for clarity. Selected bond
distances (Å) and angles (°): Ir1-C1 1.989(3), Ir1-C12 1.991(3), Ir1-N1 2.030(3), Ir1-N2 2.030(3), Ir1-O1
2.130 (3), Ir1-O2 2.134(2), C1-Ir1-N1 81.26(13), C12-Ir1-N2 81.09(13), O1-Ir1-O2 89.97(10).
10

Figure 9. The molecular structure of compound 12, with atom labeling and displacement ellipsoids
drawn at the 50% probability level. Hydrogen atoms have been removed for clarity. Only one
component of the disordered whole molecule pair is shown. Selected bond distances (Å) and angles (°):
Ir1-C1 1.984(6), Ir1-C12 1.993(7), Ir1-N1 2.032(6), Ir1-N2 2.035(6), Ir1-O1 2.105(6), Ir1-O2 2.125(6),
C1-Ir1-N1 81.3(3), C12-Ir1-N2 81.0(4), O1-Ir1-O2 86.1(3).
11

Figure 10. The molecular structure of compound 13, with atom labeling and displacement ellipsoids
drawn at the 50% probability level. Hydrogen atoms have been removed for clarity. Selected bond
distances (Å) and angles (°): Ir1-C1 1.985(2), Ir1-C12 1.983(2), Ir1-N1 2.039(2), Ir1-N2 2.035(2), Ir1-O1
2.156(2), Ir1-O2 2.158(2), C1-Ir1-N1 80.60(8), C12-Ir1-N2 81.01(8), O1-Ir1-O2 87.57(6).
Figure 11. The molecular structure of compound 14, with atom labeling and displacement ellipsoids
drawn at the 50% probability level. Hydrogen atoms and dichloromethane solvent have been
omitted for clarity. Selected bond distances (Å) and angles (°): Ir1-C1 1.985(2), Ir1-C12 1.988(2), Ir1-
N1 2.039(2), Ir1-N2 2.036(2), Ir1-O1 2.151(2), Ir1-O2 2.122(2), C1-Ir1-N1 80.98(9), C12-Ir1-N2
80.91(9), O1-Ir1-O2 87.92(6).
12

The molecular structures of all the complexes were successfully solved,
although there were special features of some, which will now be discussed.
Compound 3. The crystal was a twin by non-merohedry and the masses of the twin
individuals refined to 0.56:0.44. An uncharacteristically long detector distance of 16
cm was required to better separate streaky reflections caused by the twinning. Data
completion was >95% for these conditions while shorter detector distances led to
greater data losses. The twin law corresponds to a rotation of 180° about the (111)
reciprocal axis (see CIF file for the matrix used to convert hkl data). The 1,1,1-
trifluoro-2,4-pentanedionate ligand is disordered across a pseudo-twofold axis that
passes through the iridium atom and the carbon at the 3-position in the
pentanedionate ligand (C25 in Fig. 3); this flips the positions of the methyl and
trifluoromethyl groups in the disordered pair. The asymmetric unit for the
structure has two molecules (Z’=2) that exist as an enantiomeric pair (A and B) and
the disordered ligand exists in ratios of 0.69:0.31 in molecule A and 0.76:0.24 in
molecule B. Reasonable constraints and restraints for the disordered ligand were
applied during refinement (see Supporting Information for details).
Compound 5. The 2-acetylcyclohexanonate ligand is disordered in a 0.62:0.38 ratio
across a pseudo-mirror plane that is perpendicular to the -diketonate plane and
passes through the iridium atom and the carbon in the 2-position in the 2-
acetylcyclohexanone ring (atom C25 in Fig. 4). There is a dichloromethane solvent
molecule near the disordered ligand and this proximity causes the solvent to have
positional disorder as well. Both disorders were linked with a single occupancy free
variable to have the same ratio. Reasonable constraints and restraints for the
disordered ligand were applied during refinement (see Supporting Information for
details). The conformation of the six-membered ring in the 2-acetylcyclohexanonate
ligand is worth noting. Cyclohexene is normally found in the “half-chair”
conformation in which the two allylic carbons and the two unsaturated carbons are
coplanar while the two remaining ring carbon atoms are situated one above and one
below this plane.66, 67 That is the geometry found in this structure. The ring plane is
defined as C25/C26/C29/C30 (or its disordered partner C25’/C26’/C29’/C30’); the
average displacement of C27/C27’ is -0.50 Å and C28/C28’ is +0.28 Å.
Compound 6. The crystal structure is a 1:1 solvate with dichloromethane; the
solvent is disordered in overlapping positions in a 4:1 ratio (Fig. S4). The specimen
was extremely thin (approx. 1m) which necessitated longer than normal 45 second
exposures and higher than expected Rint = 0.064. Examination of the 4,4,4-trifluoro-
1-phenyl-1,3-butanedionate ligand shows the plane of the phenyl group is rotated at
an angle of 25.3(6)° out of the diketonate plane.
Compound 9. The crystal structure of the nonfluorinated ring analog, (ppy)2Ir(3-
methylacetylacetonate) (our compound 2) has been previously reported.68 The
configuration of the diketonate ligand is similar in both 2 and 9.
Compound 10. The crystal was a twin by non-merohedry and the masses of the
twin individuals refined to 0.53:0.47. The twin law corresponds to a rotation of
180° about the crystallographic a* axis (see CIF file for the matrix used to convert
hkl data). There are two molecules in the asymmetric unit, A and B, related to each
other by a pseudo-inversion center at approx. (0.497, 0.248. 0.369). The 1,1,1-
13

trifluoro-2,4-pentanedionate ligand is disordered across a pseudo-twofold axis that
passes through the iridium atom and the carbon at the 3-position in the
pentanedionate ligand (C25 in Fig. 7); this flips the positions of the methyl and
trifluoromethyl groups in the disordered pair. This disordered ligand exists in
ratios of 0.75:0.25 in molecule A and 0.85:0.15 in molecule B. Reasonable
constraints and restraints for the disordered ligand were applied during refinement
(see Supporting Information for details).
Compound 11. The five-carbon ring in the -diketonate ligand shows a puckered
structure. Thus, if we calculate the plane of the two allylic carbons and the two
unsaturated carbons (C26/C28/C25/C29 in Fig. 8), it is found that C27 is displaced
by 0.373(6) Å from the plane. This conformation is typically found for the carbon
atom opposite the double bond in cyclopentene and its derivatives.67, 69 We note
that the crystal structure of the nonfluorinated ring analog, (ppy)2Ir(2-
acetylcyclopentanonate) (our compound 4) has been previously reported.70 This
structure displays a disordered diketonate ligand and dichloromethane solvate
while our structure of the fluorinated ring analog does not show any disorder or
solvated molecules.
Compound 12. This compound was found to be isostructural with 11. While 11 is
well behaved and shows no signs of disorder, 12 exhibits what is best described as
whole-molecule disorder. Early results at 173K suggested that cooling to a lower
temperature could remove the disorder but collection at 100K did not help. Two
complete, partially occupied molecules were refined to a 0.64:0.36 ratio; some
disordered pairs of atoms nearly overlap (Fig.S9). Reasonable constraints and
restraints were applied during refinement (see Supporting Information for details).
Close examination of the six-membered ring in the -diketonate ligand in Figure 9
shows it to exist in the “half-chair” conformation as was found in compound 5. The
ring plane is defined as C25/C26/C29/C30 (or its disordered partner
C25’/C26’/C29’/C30’); the average displacement of C27/C27’ is -0.32 Å and
C28/C28’ is +0.41 Å.
Compound 14. The structure exists as a 1:1 solvate with dichloromethane. It was
found that the plane of the phenyl group is rotated 33.8(1)° out of the plane of the
diketonate ligand.
14

Spectroscopy and Photophysical Properties. The absorption spectra
above 350 nm of complexes 3-6 in the (ppy)2Ir(acac) series are shown in the left
half of Fig. 12 (complete spectra in Fig. S12 in Supporting Information). In addition
1.5
1
4
3
5
4
6
1.0
0.5
0.0
5
6
450 500 550 600 650 700 750 800
wavelength (nm)
350
400
450
500
550
wavelength (nm)
Figure 12. Absorption (left) and emission (right) spectra of (ppy)2Ir(acac) series (0.1 wt% in PMMA discs).
to intense ligand-centered π-π* transitions below 375 nm, the common spectral
features are a band ~400 nm typically assigned as 1MLCT (d  *), a band near 450
nm assigned as mixed singlet and triplet MLCT, and a much weaker absorption at
490 nm usually assigned to 3MLCT between the singlet ground state and the lowest
triplet excited state (S0  T1).42, 51, 71 The emission spectra of 1, 4-6 are given in the
right half of Fig. 12 (compound 3 is omitted due to its extremely weak emission).
Compounds 1, 4, and 5 show a band at approx. 515 nm with a shoulder at longer
wavelength (~540 nm) leading to yellow-green emission. Compound 6 exhibits a
broad, weak band that is centered near 600 nm, red-shifted ~90 nm from the other
compounds giving rise to weak orange emission.
The absorption spectra of complexes 9-14 in the FIracac series are shown in
Fig. 13 (left). As above for the (ppy)2Ir(acac) series, the compounds display three
bands in addition to the intense ligand-centered π-π* transition below 350 nm.
1.5
8
9
9
10
11
12
13
14
10
11
12
13
14
1.0
0.5
0.0
350
400
450
500
550
450 500 550 600 650 700 750 800
wavelength (nm)
wavelength (nm)
Figure 13. Absorption (left) and emission (right) spectra of FIracac series (0.1 wt% in PMMA discs).
15

These occur in the spectral regions of 385 nm, 440 nm, and 465 nm. The weakest
absorption (465 nm) is blue-shifted ~25 nm from the (ppy)2Ir(acac) series. The
emission spectra of 8-14 are given in the right half of Fig. 13 and show more
variation than the absorption spectra for the series. In general, the emission bands
of the FIracac series are blue-shifted about 35 nm compared to the (ppy)2Ir(acac)
series. Compounds 8, 9, 11, and 12 are nearly identical, with the main band at 480
nm and a shoulder at 505 nm leading to greenish blue emission. Compounds 13 and
14 exhibit weak, broad emission that is centered ~610 nm, appearing as an orange
color. This is the same behavior shown by 6 from the (ppy)2Ir(acac) series above.
Compound 10 shares characteristics with both of the aforementioned groups. Thus,
it has a peak at 506 nm with a shoulder ~540 nm giving its spectrum an appearance
similar to the 8, 9, 11, 12 group (although red-shifted by 25 nm). It also exhibits the
weak emission and band broadening of the 13, 14 group. It should be noted that all
the compounds with weak emission (3, 6, 10, 13, 14) contain at least one CF3 group.
Several of the emissive compounds prepared in this work are shown in Fig. 14.
1
4
11
8
Figure 14. Representative emissive compounds (0.1% in PMMA discs) under 400 nm illumination.
The photophysical properties of the complexes are given in Table 1. In
general, the F2ppy complexes exhibit higher emission quantum yields than their ppy
analogs, e.g. FIracac > (ppy)2Ir(acac), although the differences are fairly small. The
2-acetylcyclohexanonate complexes are very emissive; the complex 5 has the
highest quantum efficiency in the (ppy)2Ir(acac) group and compound 12 is nearly
as emissive as FIracac in that group. As discussed above, the compounds with CF3
groups are very weakly emissive, Φ ≤ 0.02. This behavior is typically found in
compounds with one or more aromatic groups on the β-diketonate ligand such as
(F2ppy)2Ir(1-phenyl-1,3-butanedionate),72 (F2ppy)2Ir(avobenzone),23 or
(F2ppy)2Ir(1,3-diphenyl-1,3-propanedionate).55 This was also found in a compound
with one aromatic and one CF3 group: (F2ppy)2Ir(4,4,4-trifluoro-1-naphthyl-1,3-
butanedionate).55 In all cases, the low quantum yields were attributed to emission
occurring from the triplet state of the ancillary ligand (diketonate) instead of the
cyclometallated ligand (F2ppy). Our TD-DFT results are in accord with this
assertion (vide infra). The emission lifetimes (τem) of the molecules in our study are
on the microsecond timescale (ave = 1.4 μsec) typically found for iridium(III)
cyclometallates. The complexes with CF3 groups displayed the fastest nonradiative
rate constants, with values in the range of 5.18 - 9.52 x 105 s-1.
Table 1. Photophysical properties of the iridium complexes.
16

kr (105 s-1)
3.44
-
knr (105 s-1)
2.93
-
a
b
Compound
τem (μs)
1.57
-
λ
em (nm)
Φem
1
3
4
5
6
515, 540
530 (broad)
513, 540
520, 545
605 (broad)
479, 502
484, 506
506, 540
480, 502
485, 508
610
0.54
<0.01
0.51
0.74
0.01
0.74
0.60
0.01
0.55
0.70
0.02
0.01
1.55
1.41
1.91
1.44
1.15
1.04
1.72
1.14
1.13
1.15
3.29
5.25
0.05
5.14
5.22
0.10
3.20
6.14
0.18
0.09
3.16
1.84
5.18
1.81
3.48
9.52
2.62
2.63
8.67
8.61
8
9
10
11
12
13
14
610
a
b
Measured in 0.1% PMMA discs. Relative to FIrpic with Φ = 0.89;38 values are ±20%
TD-DFT Calculations. Time-dependent Kohn-Sham density functional
theory calculations (TD-DFT) were performed on the (ppy)2Iracac series (1-6) and
FIracac series (8-14) in order to gain insight into the electronic structures of the
low-lying excited states. Different choices of the exchange-correlation functional
under the TD-DFT scheme may result in very distinct excitation energies,73
therefore we compared two different approximations, B3LYP and CAM-B3LYP, to
find the most accurate method for our systems. Furthermore, we also benchmarked
different choices of the basis set for Ir since the number of basis functions used for
the metal may affect the metal-to-ligand excitations. Two basis sets were used for
Ir: (i) the Stuttgart/Dresden relativistically corrected effective core potential (ECP)
along with a [6s5p3d] valence basis, (ii) the Los Alamos relativistic ECP along with a
[3s3p2d] valence basis. In both cases, the ECPs were used to replace 60 core
electrons of iridium. In all cases we used the triple-ζ basis set def2-TZVP for the
ligands. Our results (shown in Tables S10 and S11 in the Supporting Information)
revealed that B3LYP gave excellent agreement with the measured values and the
effect of the iridium basis set choice was negligible. The current results are
consistent with our previous TD-DFT calculations on iridium(III) picolinate
complexes59 as well as with the other benchmark on Ir(III) complexes.74 Therefore,
the excited states from B3LYP/LANL2DZ (Ir)/def2-TZVP (ligands) will be used in
the following discussion.
The six lowest energy singlet (S1-S6) and triplet (T1-T6) transitions are given
in Tables S12 (1, 3-6) and S13 (8-14); the spectroscopically significant singlet
excited states are shown in Table 2. Based on oscillator strength calculations, the
band found in the 450-460 nm region in the (ppy)2Iracac series (compounds 1, 3-6)
is assigned to either the S0  S1 (1, 3-5) or S0  S2 (6) transition. The higher energy
band near 400 nm in the series corresponds to the S0  S4 transition, except for
compound 4 which has a better fit to S0  S3. This band cannot be resolved in
compound 5 although the oscillator strength suggests S0  S3 would be the most
likely assignment. As mentioned in the spectroscopy discussion above, the band
near 495 nm in 1, 3-6 is typically assigned to the lowest energy triplet transition S0
17

 T1. Our calculations in Table S12 underestimate the energy of this transition by
0.05-0.08 eV. Now, for compounds 8-14 in the FIracac series, oscillator strengths
suggest the transition in the 425-440 nm region is assigned to either the S0  S1 (8-
12) or S0  S2 (12-14) transition. The higher energy band in the 375-390 nm
region may be either the S0  S3 (8, 9, 11, 12), S0  S4(10, 14), or S0  S5 (13)
transition. The weak band near 465 nm for this series is usually assigned as the
lowest energy triplet transition S0  T1 and our calculations are in very good
agreement (≤0.03 eV) with the observed values except for compounds 13 and 14
which underestimate the experimental energies by 0.24 eV and 0.18 eV,
respectively.
18

Table 2. Comparison of TD-DFT results for low-lying singlet states with experimental data.
a
a,b
a
c
Compound
Excited State
Ecalc(eV)
2.68
3.21
2.74
3.25
2.69
3.08
2.64
3.04
2.74
3.07
2.90
3.30
2.87
3.13
2.96
3.43
2.91
3.16
2.86
3.12
3.02
3.44
2.96
3.17
Osc. Strength (f)
0.033
0.044
0.034
0.010
0.031
0.026
0.034
0.016
0.032
0.017
0.030
0.031
0.030
0.019
0.031
0.034
0.027
0.029
0.030
0.017
0.032
0.053
0.031
0.020
λ
calc (nm)
462
387
453
382
461
402
469
407
453
404
427
376
432
397
419
361
427
392
434
397
411
360
418
391
λ
expt (nm)
460
410
451
402
456
413
462
-
448
400
438
390
440
394
432
380
438
392
438
394
425
373
430
385
1
S1
S4
S1
S4
S1
S3
S1
S3
S2
S4
S1
S3
S1
S3
S1
S4
S1
S3
S1
S3
S2
S5
S2
S4
3
4
5
6
8
9
10
11
12
13
14
a
Values from Table S12-13 using B3LYP functional with def2-TZVP (ligand)/LANL2DZ (Ir) basis set.
c
bOnly singlet states with appreciable oscillator strength are included. Measured in 0.1% PMMA discs.
We have computed natural transition orbitals (NTOs)75 for S1-S6 and T1-T2
states for all compounds in the FIracac series (8-14) and provide full details for
these in Tables S14-S28, including the description of the NTOs in terms of molecular
orbitals. We also present molecular orbital energy level diagrams and contour plots
for complexes 8-14 in Tables S29-36 although we will primarily utilize NTOs in the
following discussion. The use of these NTOs provides a simple qualitative
description of the electronic transitions for these complexes. NTOs for the lowest
energy singlet excited state (S1) for 8 are given in Table 3 and this serves as the
prototype for this transition in all the complexes. We find that the NTOs for this
lowest energy singlet band near 435 nm (S0  S1 for 8-12 and S0  S2 for 13-14)
have a very similar appearance for the entire series. Thus, donation from the Ir d-
orbital to the phenylpyridine ligands occurs (MLppyCT) as well as ligand-ligand
donation between the phenylpyridine groups (LppyLppyCT). There is no involvement
of the acetylacetonate ligand in this transition. As shown in Table S14, this lowest
energy singlet transition is mainly HOMO  LUMO for 8, HOMO  LUMO+1 for 10,
13, and 14 and a combination of HOMO  LUMO and HOMO  LUMO+1 for 9, 11,
and 12. The higher energy singlet transition near 390 nm has more variation in the
NTOs and these are shown for representative compounds in Table 3. For all
19

compounds except 13 the donor orbitals have no contribution from the
phenylpyridine ligand; the electron density is located on the iridium d-orbital and
acetylacetonate ligand. The acceptor orbitals for 8, 9, 11, and 12 involve the
phenylpyridine ligands so we can classify the transition on these compounds as
MLppyCT and LacacLppyCT. These are all S0  S3 transitions. The S3 NTOs for 8 and 11
are nearly identical so only 8 is shown in Table 3. Similarly, 9 and 12 are nearly
identical so only 9 is shown. The acceptor orbital on compound 10 shows
predominant involvement of the phenylpyridine group but some small density
exists on the acetylacetonate ligand as well; this is the S0  S4 transition in Table 3.
Compound 14 differs from 8-12 since it has acceptor orbital electron density solely
on the acetylacetonate ligand (with attached phenyl group) so this transition can be
classified as MLacacCT and LacacLacacCT. There is no involvement of the
phenylpyridine ligand in this S0  S4 transition. Finally, compound 13 (S0  S5
transition) has iridium d-orbital and phenylpyridine ligand involvement in its donor
NTO while the acceptor orbital is localized on the acetylacetonate ligand. This
transition can be classified as MLacacCT and LppyLacacCT. Now, in molecular orbital
terms (see Table S14) this approx. 390 nm singlet transition for the FIracac series
involves HOMO-1  LUMO (compounds 10 and 14), HOMO-1  LUMO+1
(compound 8), and a combination of HOMO-1  LUMO and HOMO-1  LUMO+1
(compounds 9, 11, and 12). Compound 13 is very different, with contributions
from HOMO-4  LUMO and HOMO-2  LUMO.
20

Table 3. Natural transition orbitals (NTOs) for low energy singlet states of 8-10, and 13-14.
Acceptor
Donor
Acceptor
Sn
Sn
Donor
Compound 10
S4
S1
Compound 13
S5
S3
Compound 14
S4
Compound 9
S3
21

The NTOs for selected triplet states calculated at the ground state (S0)
geometry are shown in Table 4. Both T1 and T2 excited states are considered for 8-
12 (T3 is also included for 10, see below) since these states are nearly degenerate
(≤0.03 eV difference) as shown in Table S13. The T1 and T2 NTOs for 8, 9, 11, and
12 are very similar and we provide T1 for compound 8 as a representative example
in Table 4. We find that the low energy triplet transitions for these four compounds
incorporate donation from the Ir d-orbital to the phenylpyridine ligands (MLppyCT)
as well as ligand-ligand donation between the phenylpyridine groups (LppyLppyCT).
There is no involvement of the acetylacetonate ligand in these transitions. For 8, 9,
and 12 this can be described as HOMO  LUMO (T1) and HOMO  LUMO+1 (T2).
For 11 each state is a combination of HOMO  LUMO and HOMO  LUMO+1 which
differ primarily in the phase of the configuration mixing. Compounds 13 and 14 are
also very similar to each other: the T1 state for 13 is given in Table 4 (the much
higher energy T2 state need not be considered). This S0  T1 transition comprises
donation from the iridium d-orbital to the acetylacetonate ligand (MLacacCT) and
ligand-ligand donation within the acetylacetonate (LacacLacacCT). There is no
evidence of involvement of the phenylpyridine ligand in this state. This can be
described primarily as HOMO-1  LUMO although 13 also has contributions from
HOMO-6  LUMO and HOMO-3  LUMO. Finally, compound 10 is set apart in this
FIracac series since the three lowest energy triplet states are nearly degenerate (E
= 2.70, 2.71, and 2.72 eV) with very different NTOs as shown in Table 4. The T1 state
for 10 primarily utilizes phenylpyridine ligands as acceptors so it can be classified
as MLppyCT and LppyLppyCT. The T2 state for 10 is unique among the compounds
studied here since it involves both phenylpyridine and acetylacetonate as acceptor
groups and can be viewed as a combination of MLacacCT and MLppyCT. The T3 state
utilizes only the acetylacetonate ligand as an acceptor and is thus MLacacCT and
LacacLacacCT in nature. The T1 state can be described as HOMO  LUMO+1 while the
T2 and T3 states comprise HOMO  LUMO and HOMO-1  LUMO+2 as the main
contributors.
22

Table 4. Natural transition orbitals (NTOs) for low energy triplet states of 8, 10, and 13.
Acceptor
Tn
Donor
Tn
Acceptor
Donor
Compound 10
Compound 8
T1
T1
Compound 13
T2
T1
T3
23

Phosphorescence from cyclometallated iridium complexes is believed to
originate from the lowest energy triplet state38, 46, 50, 76 so we can use the information
culled from the low energy triplet NTOs presented in Table 4 to better understand
the emission spectra depicted in Fig. 13. The triplet acceptor orbitals for these
complexes can be divided into three types: Type 1 has the acceptor located on the
phenylpyridine ligands (compounds 8, 9, 11, and 12); Type 2 has the acceptor
located on the acetylacetonate ligand (compounds 13 and 14); Type 3 is located on
both the acetylacetonate and phenylpyridine ligands (compound 10, specifically the
T2 state). This treatment has been suggested by Park and coworkers60, 76, 77 utilizing
molecular orbitals (MOs) for explaining Type 1 and 2 emission. We have chosen to
focus on NTOs instead of MOs and also posit Type 3 as a new example.59 The Type 1
emission spectra for 8, 9, 11, and 12 are nearly identical with a fairly narrow band
containing a longer wavelength shoulder; the emission is greenish blue in color (see
Fig. 14 for examples). The Type 2 spectra of 13 and 14 are symmetric and broad
with a bathochromic shift giving rise to weak orange emission. Finally, the Type 3
spectrum of 10 has similarities to the Type 1 compounds with a peak at 506 nm and
a longer wavelength shoulder. However, it is significantly broadened with a tail
extending far into the visible and weak yellow emission.
CONCLUSIONS
We have found that using 1,2-dimethoxyethane as a reaction solvent is
superior to the glycol ethers typically employed for ancillary ligand addition to
dimeric iridium compounds. For example, acetylacetone added cleanly to
[(F2ppy)2IrCl]2 to form FIracac in refluxing dimethoxyethane under basic conditions.
This reaction proceeded quickly in high yield and required no product purification
aside from simple filtration and water wash. This is in contrast to the conventional
synthesis in 2-ethoxyethanol in which column chromatography is needed to obtain
pure material. The ability to omit this purification step becomes even more
important with larger scale preparations. We have prepared fifteen cyclometallated
Ir(III) acetylacetonates via this route, eight of which are newly reported here. In
several instances (compounds 2, 10, 13, and 14), the use of refluxing 2-
ethoxyethanol failed to yield products that were easily obtained with our
dimethoxyethane route. X-ray diffraction studies were carried out on nine of the
compounds to determine their molecular structures. All the complexes exhibit the
same distorted octahedral geometry with two bidentate phenylpyridine ligands and
one bidentate acetylacetonate ligand. The iridium-bound nitrogen atoms of the two
phenylpyridine ligands are trans to each other while the two ligated oxygen atoms
in the acetylacetonate ligand are trans to the iridium-bound phenyl carbon atoms of
the phenylpyridine ligands. The acetylcyclohexanonate complexes 5 and 12 display
the “half-chair” conformation of the six-membered ring typically found in
cyclohexene while the acetylcyclopentanonate complex 11 exists in the puckered
form common in cyclopentene. The phenyl-substituted acetylacetonates 6 and 14
show the phenyl group rotated out of the diketonate plane by 25° and 34°,
respectively.
24

Seven of the compounds were strongly emissive at room temperature with
colors ranging from blue to yellow and quantum yields of 50-75%. All of the
complexes containing a trifluoromethyl group (3, 6, 10, 13, and 14) were very weak
emitters with Φ ≤ 0.02. Measured emission lifetimes for all the compounds were in
the range of 1-2 μs. Time-dependent density functional theory calculations utilizing
the B3LYP functional with LANL2DZ (Ir)/def2-TZVP (ligands) basis sets gave close
agreement with spectroscopic data. These calculations, visualized with natural
transition orbitals for the FIracac series, guided a detailed interpretation of the
absorption and emission spectra for the complexes. We found it useful to study the
low energy triplet NTOs for the FIracac series to understand the phosphorescence in
these complexes. Compounds 8, 9, 11, and 12 have acceptor orbitals located only
on the phenylpyridine ligands and are called Type 1 emitters. These display highly
efficient greenish blue luminescence with fairly narrow spectral bands. Complexes
13 and 14 have acceptor orbitals involving only the ancillary acetylacetonate ligand
and are referred to as Type 2 compounds. These are very inefficient with broad
emission bands in the orange region. Finally, compound 10 features acceptor
orbitals involving both phenylpyridine and acetylacetonate ligands (best illustrated
by its T2 state) and we classify this as Type 3 behavior. This emission spectrum has
the general shape of the Type 1 compounds but is broadened appreciably and red-
shifted.
EXPERIMENTAL SECTION
General Information. All compounds were synthesized in a dry nitrogen
atmosphere, unless otherwise specified. The solvents and commercially available
starting materials were of the highest purity available and used as received. Most
organic compounds were purchased from TCI America and IrCl3 · x H2O was
purchased from Strem Chemicals. Silica column chromatography was done with
silica gel 60 (particle size 40-63 m) from EMD Chemicals. Microanalysis was
performed by Galbraith Laboratories, Knoxville, TN. NMR spectra were recorded on
a Bruker Avance III 600 MHz spectrometer.
Photophysical and optical spectroscopy measurements. Specimens for
absorption and emission measurements were prepared by polymerizing 0.1%
solutions of the compounds in methyl methacrylate. This provided 14 mm diam
cylinders which were cut and polished to yield ~ 1 mm thick x 14 mm diam discs.
Optical absorption spectra were obtained with a Thermo Evolution 220 UV-visible
spectrometer. Emission spectra were measured using 425 nm excitation on a
Thermo Lumina luminescence spectrometer employing backscatter geometry,
excitation and emission slits of 1 mm, a signal averaging time of 0.2 s, and a scan
rate of 300 nm/minute. Quantum yields were estimated by sensitivity correcting
the emission spectra for the PMT and grating efficiency as a function of wavelength,
then normalizing the integrated emission spectra to the optical density of each
sample at the excitation wavelength, 425 nm. To estimate the quantum yields of the
compounds, their integral light yields were standardized relative to FIrpic (obtained
25

from American Dye Source), which has a reported value of 0.89 in PMMA.38
Phosphorescence lifetime measurements were acquired using a flashlamp-pumped
Nd:YAG laser at 266 nm with 20 ns pulses. Luminescence was collected with a
monochromator, set to the emission wavelength maximum for each compound,
coupled to an R928 Hamamatsu PMT and read out by an oscilloscope. For each
sample, 10 traces were acquired and averaged to generate the decay curves, which
were then fit in Igor Pro, for all decays, to single exponentials to obtain the
phosphorescence lifetimes.
X-ray crystallography. Suitable crystals for single crystal diffraction studies on
compounds 3, 5, 6 and 9-14 were obtained by slow evaporation of solvent from
dichloromethane/methanol solutions. Single crystal specimens for 3, 5, 6, and 11-
14 were collected on a Bruker-AXS VENTURE diffractometer with a MoKα
microfocus source (λ = 0.71073 Å) with a PHOTON-II (CPAD) detector. Unit cell
constants were determined initially with APEX3 software.78 Single crystal
specimens for 9 and 10 were collected on a Bruker-AXS APEX-II CCD diffractometer
with a MoKα normal-focus tube source (λ = 0.71073 Å). Unit cell constants were
determined initially with APEX2 software.78 Final unit cell constants for all
structures were calculated on sets of strong reflections selected from the integration
of all data with SAINT.78 Multi-scan scaling and absorption correction were applied
with SADABS or TWINABS based on the Laue class determined for each crystal
structure.78 The crystal structures were solved by direct-methods using SHELXT-
2014.79 Final least-squares refinements were completed with SHELXL-2014.80 All
non-hydrogen atoms were refined with anisotropic displacements. Hydrogen atoms
were placed as riding atoms at calculated positions with isotropic displacements
relative to respective host atoms. Specimens 3 and 10 were determined to be twins
by non-merohedry. Specimen 12 exhibited whole molecule disorder. Various
restraints81 and constraints were applied to 3, 5, 10, and 12 as detailed in the
Supporting Information.
Syntheisis. [(ppy)2IrCl]253, 54 and [(F2ppy)2IrCl]247, 60 were prepared by the
literature methods.
(ppy)2Ir(acetylacetonate) (Compound 1). A 50 mL round-bottom flask was
charged with 1.0 g (9.4 mmol) sodium carbonate and 0.37 g (3.7 mmol) 2,4-
pentanedione in 25 ml 1,2-dimethoxyethane. The mixture was purged with
nitrogen and 1.0 g (0.93 mmol) [(ppy)2IrCl]2 was added. The reaction was brought
to reflux and heated for 8 hr under a nitrogen atmosphere. After cooling to room
temperature, the reaction was filtered through a glass frit. The solids were washed
with three portions of water and three portions of methanol and dried under
1
vacuum to yield 1.0 g of yellow-orange product (89%). H NMR results were
identical with those previously reported.42
(ppy)2Ir(3-methylacetylacetonate) (Compound 2). A 50 mL round-bottom flask
was charged with 1.0 g (9.4 mmol) sodium carbonate and 0.43 g (3.7 mmol) 3-
methyl-2,4-pentanedione in 25 ml 1,2-dimethoxyethane. The mixture was purged
26

with nitrogen and 1.0 g (0.93 mmol) [(ppy)2IrCl]2 was added. The reaction was
brought to reflux and heated for 6 hr under a nitrogen atmosphere. After cooling to
room temperature, the reaction was filtered through a glass frit. The solids were
washed with three portions of water and three portions of methanol and dried
under vacuum to yield 1.0 g of material. NMR analysis revealed equimolar starting
material and product and ~4% of unidentified impurity. The reaction was repeated
with a 24 hr reflux and similar workup to yield 1.1 g of material. NMR analysis
showed no starting material remained, but the impurity seen in the 6 hr reaction
was now present at the 25% level. All attempts to obtain analytically pure material
were unsuccessful. The NMR spectrum we recorded for compound 2 was somewhat
different than that previously reported,61 specifically in the upfield region displaying
the diketonate methyl groups. The prior report found these peaks at 2.17 (3H) and
1.78 (6H) in CDCl3. Our spectrum shows them at 1.90 (6H) and 1.89 (3H) in
excellent agreement with the well-characterized F2ppy analog 9 (see below) which
1
displays the resonances at 1.92 (s, 6H) and 1.91(s, 3H). H NMR (600MHz, CDCl3)
(ppm): 8.55 (d, J = 5.5 Hz, 2H), 7.87 (d, J = 8.1 Hz, 2H), 7.74 (t, J = 7.7 Hz, 2H), 7.57
(d, J = 7.7 Hz, 2H), 7.14 (t, J = 6.6 Hz, 2H), 6.81 (t, J = 7.4 Hz, 2H), 6.69 (t, J = 7.4 Hz,
2H), 6.27 (d, J = 7.6 Hz, 2H), 1.90 (s, 6H), 1.89 (s, 3H).
(ppy)2Ir(1,1,1-trifluoroacetylacetonate) (Compound 3). Procedure as for
compound 1, with 1,1,1-trifluoro-2,4-pentanedione as ligand (76% yield). Anal.
Calcd for IrC27H20N2O2F3: C, 49.61; H, 3.08; N, 4.29. Found: C, 49.54; H, 2.92; N,
1
4.20%. H NMR (600MHz, CDCl3) (ppm): 8.47 (d, J = 5.1 Hz, 1H), 8.41 (d, J = 5.1 Hz,
1H), 7.89 (d, J = 8.3 Hz, 2H), 7.79 (m, 2H), 7.59 (d, J = 7.7 Hz, 1H), 7.56 (d, J = 7.7 Hz,
1H), 7.18 (m, 2H), 6.86 (m, 2H), 6.73 (m, 2H), 6.31 (d, J = 7.0 Hz, 1H), 6.22 (d, J =
7.0Hz, 1H), 5.62 (s, 1H), 1.90 (s, 3H). 19F NMR (565 MHz, CDCl3) (ppm): -74.92 (s,
3F).
(ppy)2Ir(2-acetylcyclopentanonate) (Compound 4). Procedure as for compound
1, with 2-acetylcyclopentanone as ligand (91% yield). A lower resolution NMR
1
spectrum has been reported;61 we now report a higher resolution spectrum. H
NMR (600MHz, CDCl3) (ppm): 8.55 (d, J = 5.3 Hz, 1H), 8.50 (d, J = 5.4 Hz, 1H), 7.86
(d, J = 8.1 Hz, 2H), 7.74 (t, J = 7.1 Hz, 2H), 7.56 (t, J = 6.7 Hz, 2H), 7.14 (t, J = 5.8 Hz,
2H), 6.82 (m, 2H), 6.70 (m, 2H), 6.28 (t, J = 7.1 Hz, 2H), 2.67 (t, J = 6.8 Hz, 2H), 2.23
(m, 1H), 2.13 (m, 1H), 1.86 (s, 3H), 1.79 (m, 1H), 1.73 (m, 1H).
(ppy)2Ir(2-acetylcyclohexanonate) (Compound 5). A 50 mL round-bottom flask
was charged with 1.0 g (9.4 mmol) sodium carbonate and 0.53 g (3.7 mmol) 2-
acetylcyclohexanone in 25 ml 1,2-dimethoxyethane. The mixture was purged with
nitrogen and 1.0 g (0.93 mmol) [(ppy)2IrCl]2 was added. The reaction was brought
to reflux and heated for 6.5 hr under a nitrogen atmosphere. After cooling to room
temperature, the reaction was filtered through a glass frit. The solids were washed
with three portions of water and three portions of methanol and dried under
vacuum to yield 1.0 g of material. NMR analysis revealed a 2:1 ratio of product to
starting material. The reaction was repeated with a 24 hr reflux and similar workup
to yield 1.1 g of orange-yellow product (92%), with no starting material remaining.
27

A lower resolution NMR spectrum has been reported;61 we now report a higher
1
resolution spectrum. H NMR (600MHz, CDCl3) (ppm): 8.58 (d, J = 5.6 Hz, 1H), 8.55
(d, J = 5.6 Hz, 1H), 7.86 (d, J = 8.1 Hz, 2H), 7.73 (t, J = 7.7 Hz, 2H), 7.56 (t, J = 6.3 Hz,
2H), 7.14 (t, J = 5.8 Hz, 2H), 6.81 (q, J = 7.1 Hz, 2H), 6.69 (q, J = 7.6 Hz, 2H), 6.27 (dd, J
= 14.8, 7.6 Hz, 2H), 2.40 (t, J = 5.6 Hz, 2H), 2.09 (m, 1H), 1.95 (m, 1H), 1.86 (s, 3H),
1.62 (m, 2H), 1.60 (m, 1H), 1.57 (m, 1H).
(ppy)2Ir(4,4,4-trifluoro-1-phenyl-1,3-butanedionate) (Compound 6).
Procedure as for compound 1, with 4,4,4-trifluoro-1-phenyl-1,3-butanedione as
ligand (85% yield). Anal. Calcd for IrC32H22N2O2F3: C, 53.70; H, 3.10; N, 3.91.
1
Found: C, 53.42; H, 3.19; N, 3.67%. H NMR (600MHz, CDCl3) (ppm): 8.54 (d, J = 5.7
Hz, 1H), 8.44 (d, J = 5.7 Hz, 1H), 7.89 (dd, J = 12.0, 8.1 Hz, 2H), 7.77 (m, 4H), 7.61 (d, J
= 7.1 Hz, 1H), 7.59 (d, J = 7.1 Hz, 1H), 7.49 (t, J = 7.4 Hz, 1H), 7.33 (m, 2H), 7.17 (t, J =
7.2 Hz, 1H), 7.11 (t, J = 7.2 Hz, 1H), 6.88 (dt, J = 18.5, 7.0 Hz, 2H), 6.75 (m, 2H), 6.34
(d, J = 7.6 Hz, 1H), 6.33 (s, 1H), 6.29 (d, J = 7.6 Hz, 1H). 19F NMR (565 MHz, CDCl3)
(ppm): -74.80 (s, 3F).
(ppy)2Ir(1,3-diphenyl-1,3-propanedionate) (Compound 7). Procedure as for
1
compound 1, with dibenzoylmethane as ligand (86% yield). H NMR results were
identical with those previously reported.54
FIracac (Compound 8). A 100 mL round-bottom flask was charged with 1.75 g
(16.5 mmol) sodium carbonate and 0.66 g (6.6 mmol) 2,4-pentanedione in 55 ml
1,2-dimethoxyethane. The mixture was purged with nitrogen and 2.00 g (1.64
mmol) [(F2ppy)2IrCl]2 was added. The reaction was brought to reflux and heated for
11 hr under a nitrogen atmosphere. The reaction was then chilled in an ice-water
bath prior to filtration through a glass frit. The solids were washed with three
portions of water and two portions of ice-cold methanol and dried under vacuum to
yield 1.60 g of yellow product (73%). A second batch of product may be obtained by
concentrating the original dimethoxyethane filtrate to ~10 mL on a rotovap
followed by filtration and washing as above. The combined yield was 1.85 g (84%)
1
of yellow product. NMR spectra taken of both batches were identical. H and 19F
NMR results were in agreement with those previously reported.45, 47
(F2ppy)2Ir(3-methylacetylacetonate) (Compound 9). A 50 mL round-bottom
flask was charged with 0.87 g (8.2 mmol) sodium carbonate and 0.38 g (3.3 mmol)
3-methyl-2,4-pentanedione in 25 ml 1,2-dimethoxyethane. The mixture was purged
with nitrogen and 1.0 g (0.82 mmol) [(F2ppy)2IrCl]2 was added. The reaction was
brought to reflux and heated for 6.25 hr under a nitrogen atmosphere. The reaction
was then chilled in an ice-water bath prior to filtration through a glass frit. The
solids were washed with three portions of water and three portions of ice-cold
methanol and dried under vacuum to yield 0.95 g of yellow product (84%). Anal.
Calcd for IrC28H21N2O2F4: C, 49.05; H, 3.09; N, 4.09. Found: C, 48.76; H, 3.05; N,
1
4.04%. H NMR (600MHz, CDCl3) (ppm): 8.48 (d, J = 5.6 Hz, 2H), 8.27 (d, J = 8.3 Hz,
2H), 7.80 (t, J = 7.9 Hz, 2H), 7.19 (d, J = 6.6 Hz, 2H), 6.34 (t, J = 10.9 Hz, 2H), 5.66 (d, J
28

= 8.9 Hz, 2H), 1.92 (s, 6H), 1.91(s, 3H). 19F NMR (565 MHz, CDCl3) (ppm): -109.14
(q, J = 9.4 Hz, 2F), -111.25 (t, J = 11.3 Hz, 2F).
(F2ppy)2Ir(1,1,1-trifluoroacetylacetonate) (Compound 10). A 50 mL round-
bottom flask was charged with 0.87 g (8.2 mmol) sodium carbonate and 0.51 g (3.3
mmol) 1,1,1-trifluoro-2,4-pentanedione in 25 ml 1,2-dimethoxyethane. The mixture
was purged with nitrogen and 1.0 g (0.82mmol) [(F2ppy)2IrCl]2 was added. The
reaction was brought to reflux and heated for 12 hr under a nitrogen atmosphere.
After cooling to room temperature, the reaction was filtered through a glass frit.
The solids were washed with two portions of water and two portions of methanol
and dried under vacuum to yield 0.37 g of yellow product. Water was added to the
original dimethoxyethane filtrate to obtain a second crop of product which was
filtered and washed with water and methanol as before. The material was dried
under vacuum to obtain an additional 0.69 g of yellow product (combined yield
88%). NMR spectra taken of both fractions were identical. Anal. Calcd for
1
IrC27H16N2O2F7: C, 44.69; H, 2.22; N, 3.86. Found: C, 44.23; H, 1.88; N, 3.81%. H
NMR (600MHz, CDCl3) (ppm): 8.40 (d, J = 5.7 Hz, 1H), 8.34 (d, J = 5.7 Hz, 1H), 8.29
(d, J = 8.2 Hz, 2H), 7.86 (t, J = 7.8 Hz, 2H), 7.26 (t, J = 7.2 Hz, 1H), 7.23 (t, J = 7.1 Hz,
1H), 6.40 (t, J = 10.1 Hz, 1H), 6.38 (t, J = 10.1 Hz, 1H), 5.70 (d, J = 8.7 Hz, 1H), 5.67 (s,
1H), 5.61 (d, J = 8.8 Hz, 1H), 1.95 (s, 3H). 19F NMR (565 MHz, CDCl3) (ppm): -75.03
(s, 3F), -108.32 (q, J = 9.4 Hz, 1F), -108.49 (q, J = 9.4 Hz, 1F), -110.75 (t, J = 11.4 Hz,
1F), -110.82 (t, J = 11.4 Hz, 1F).
(F2ppy)2Ir(2-acetylcyclopentanonate) (Compound 11). Procedure as for
compound 9, with 2-acetylcyclopentanone as ligand (90% yield). Anal. Calcd for
1
IrC29H21N2O2F4: C, 49.92; H, 3.03; N, 4.01. Found: C, 49.81; H, 2.80; N, 3.95%. H
NMR (600MHz, CDCl3) (ppm): 8.48 (d, J = 5.7 Hz, 1H), 8.42 (d, J = 5.7 Hz, 1H), 8.27
(d, J = 8.3 Hz, 2H), 7.81 (t, J = 7.4 Hz, 2H), 7.20 (m, 2H), 6.35 (m, 2H), 5.68 (t, J = 2.7
Hz, 1H), 5.66 (t, J = 2.7 Hz, 1H), 2.68 (t, J = 7.2 Hz, 2H), 2.26 (m, 1H), 2.16 (m, 1H),
1.70-1.85 (m, 2H), 1.88 (s, 3H). 19F NMR (565 MHz, CDCl3) (ppm): -108.88 (q, J =
9.4 Hz, 1F), -108.90 (q, J = 9.4 Hz, 1F), -111.13 (t, J = 11.5 Hz, 1F), -111.17 (t, J = 11.5
Hz, 1F).
(F2ppy)2Ir(2-acetylcyclohexanonate) (Compound 12). Procedure as for
compound 9, with 2-acetylcyclohexanone as ligand (85% yield). Anal. Calcd for
1
IrC30H23N2O2F4: C, 50.63; H, 3.26; N, 3.94. Found: C, 50.62; H, 3.40; N, 3.90%. H
NMR (600MHz, CDCl3) (ppm): 8.50 (d, J = 5.7 Hz, 1H), 8.47 (d, J = 5.7 Hz, 1H), 8.27
(d, J = 8.1 Hz, 2H), 7.80 (t, J = 7.9 Hz, 2H), 7.19 (m, 2H), 6.34 (m, 2H), 5.67 (dd, J = 8.9,
2.3 Hz, 1H), 5.65 (dd, J = 8.9, 2.3 Hz, 1H), 2.40 (t, J = 5.9 Hz, 2H), 2.12 (m, 1H), 1.97
(m, 1H), 1.89 (s, 3H), 1.60-1.65 (m, 3H), 1.57 (m, 1H). 19F NMR (565 MHz, CDCl3)
(ppm): -109.15 (q, J = 9.5 Hz, 1F), -109.19 (q, J = 9.5 Hz, 1F), -111.25 (t, J = 11.3 Hz,
1F), -111.29 (t, J = 11.3 Hz, 1F).
(F2ppy)2Ir(hexafluoroacetonate) (Compound 13). A 50 mL round-bottom flask
was charged with 0.87 g (8.2 mmol) sodium carbonate and 0.69 g (3.3 mmol)
1,1,1,5,5,5-hexafluoroacetylacetone in 25 ml 1,2-dimethoxyethane. The mixture was
29