Ce(IV) complexes with donor-functionalized alkoxide ligands: Improved precursors for chemical vapor deposition of CeO2

Article abstract of DOI:10.1021/ic201593s

Thin films of ceria (CeO₂) have many applications, and their synthesis by liquid-injection MOCVD (metal-organic chemical vapor deposition) or ALD (atomic layer deposition) requires volatile precursor compounds. Here we report the synthesis of a

Full text of DOI:10.1021/ic201593s

ARTICLE
pubs.acs.org/IC
Ce(IV) Complexes with Donor-Functionalized Alkoxide Ligands:
Improved Precursors for Chemical Vapor Deposition of CeO2
Helen C. Aspinall,* John Bacsa, Anthony C. Jones, and Jacqueline S. Wrench
Department of Chemistry, Donnan and Robert Robinson Laboratories, University of Liverpool, Crown Street, Liverpool, L69 7ZD, U.K.
Kate Black, Paul R. Chalker, Peter J. King, Paul Marshall, and Matthew Werner
Department of Engineering, George Holt Building, University of Liverpool, Liverpool, L69 3GH, U.K.
Hywel O. Davies and Rajesh Odedra
SAFC Hitech, Power Road, Bromborough, Wirral, CH62 3QF, U.K.
S Supporting Information
b
ABSTRACT: Thin films of ceria (CeO2) have many applica-
tions, and their synthesis by liquid-injection MOCVD (metalꢀ
organic chemical vapor deposition) or ALD (atomic layer de-
position) requires volatile precursor compounds. Here we
report the synthesis of a series of homoleptic and heteroleptic
Ce(IV) complexes with donor-functionalized alkoxide li-
gands mmp (1-methoxy-2-methylpropan-2-olate), dmap
(1-(dimethylamino)propan-2-olate), and dmop (2-(4,4-di-
methyl-4,5-dihydrooxazol-2-yl)propan-2-olate) and their
potential as precursors for MOCVD and ALD of CeO2.
New complexes were synthesized by alcohol exchange reac-
tions with [Ce(OBut)4]. [Ce(mmp)4] and [Ce(dmap)4] were
both found to be excellent precursors for liquid-injection MOCVD of CeO2, depositing high purity thin films with very low carbon
contamination, and both have a large temperature window for diffusion controlled growth (350ꢀ600 °C for [Ce(mmp)4];
300ꢀ600 °C for [Ce(dmap)4]). [Ce(mmp)4] is also an excellent precursor for liquid-injection ALD of CeO2 using H2O as oxygen
source and demonstrates self-limiting growth from 150 to 350 °C. [Ce(dmap)4] has lower thermal stability than [Ce(mmp)4] and
does not show self-limiting growth in ALD. Heteroleptic complexes show a tendency to undergo ligand redistribution reactions to
form mixtures in solution and are unsuitable as precursors for liquid-injection CVD.
’ INTRODUCTION
with H2O, which is a real advantage in ALD. Facile reaction of a
precursor with H2O means that H2O can be used as an oxygen
source for ALD of oxide materials: this is experimentally preferable to
use of more aggressive oxygen sources such as ozone. The Ce(IV)
alkoxide [Ce(OCMe2Pri)4] has been investigated as a precursor for
MOCVD of CeO2,8 but there have been no subsequent reports of its
use, and we found that it had a very narrow temperature window for
diffusion-controlled growth (see Results and Discussion). We there-
fore identified a need for more effective precursors for CeO2, and in
this paper, we report on a series of Ce(IV) alkoxides and their use for
liquid-injection MOCVD and ALD of CeO2.
Thin films of CeO2 have many applications: e.g. as a dielectric
in complementary metal oxide semiconductor (CMOS) and
memory applications;1 as a dopant to stabilize high-k phases of
HfO2;2 in magneto-optical devices;3 in sensors;4 as an anode
material in solid oxide fuel cells.5 MOCVD (metalꢀorganic
chemical vapor deposition) and ALD (atomic layer deposition)
are important techniques for the preparation of oxide thin
films,6,7 and both of these techniques require volatile precursor
compounds. [Ce(thd)4] (thd =2,2,6,6-tetramethylheptane-3,5-
dionate) is probably the most widely used precursor for CVD of
CeO2. It is air- and moisture-stable, reasonably volatile, easy to
synthesize, and has high thermal stability. Although these properties
mean that it is quite easy to handle, they also mean that it is a less than
ideal precursor for CVD. Alkoxide complexes are more attractive as
CVD precursors: they have lower thermal stability than β-diketo-
nates and hence lower deposition temperatures in MOCVD, and
they are powerful Brønsted bases, meaning that they react readily
It is apparent that Ce complexes with monodentate alkoxide
ligands are less than ideal as CVD precursors, and so we chose to
work with bidentate donor-functionalized alkoxide ligands de-
rived from the alcohols shown in Chart 1 below.
Received: July 25, 2011
Published: October 21, 2011
r
2011 American Chemical Society
11644
dx.doi.org/10.1021/ic201593s Inorg. Chem. 2011, 50, 11644–11652
|

Inorganic Chemistry
ARTICLE
Chart 1. Structures of Donor Functionalized Alcohols
Calcd for CeC16H40O4N4: C 39.01, H 8.18, N 11.37; found C, 38.88,
H 8.08, N 11.19%. A 1H NMR spectrum is shown in Figure S2 of the
Supporting Information.
[Ce(dmap)4] 3 was synthesized following the route used for
[Ce(mmp)4] using CAN (7.00 g, 12.8 mmol), NaOBut (7.36g, 76.6mmol),
and dmapH (6.30 mL, 51.1 mmol). Recrystallization from toluene gave
the complex as yellow crystals. Yield: 3.10 g (44%).
1H NMR (400 MHz, CDCl3): δ(ppm) 1.12 (3H, d, J = 6.13 Hz,
OCH(CH3)CH2N(CH3)2), 2.13 (2H, m, OCH(CH3)CH2N(CH3)2),
2.27 (6H, s, OCH(CH3)CH2N(CH3)2), 3.78 (1H, br m, OCH-
(CH3)CH2N(CH3)2). 13C NMR (100 MHz, CDCl3): δ(ppm) 24.6
(OCH(CH3)CH2N(CH3)2), 46.1 (OCH(CH3)CH2N(CH3)2), 70.3
(OCH(CH3)CH2N(CH3)2). Calcd for C20H48CeN4O4: C 43.78, H
8.82, N 10.21; found C 43.96, H 8.76, N 10.07%.
[Ce(dmop)4] 4 A solution of [NH4]2[Ce(NO3)6] (CAN) (1.00 g,
1.82 mmol) in dimethoxyethane (DME) (20 mL) was added at room
temperature to a stirred solution of [NaOBut] (1.05 g, 10.9 mmol) in
DME. After removal of the NaNO3 precipitate, the filtrate was stirred at
0 °C, and a solution of dmopH (1.13 g, 7.30 mmol) in DME (10 mL) was
added. The reaction mixture was allowed to warm to room temperature
and stirred overnight. Volatiles were removed in vacuo, and the yellow
solid residue was extracted with toluene (3 ꢁ 50 mL), which was then
concentrated to give the product as yellow crystals on cooling to ꢀ18 °C.
Yield: 437 mg (31.3%).
These ligands (apart from dmomp) have all been used
successfully with other metals as precursors for CVD of oxides,9,10
and we have already published a preliminary communication on
the use of [Ce(mmp)4] for MOCVD of CeO2.11
1H NMR (400 MHz, C7D8): δ(ppm) 1.52 (6H, s, NC(CH3)2CH2),
1.54 (6H, s, OC(CH3)2C), 3.82 (2H, s, NC(CH3)2CH2). 13C NMR (100
MHz, C7D8): δ(ppm) 27.9 (NC(CH3)2CH2), 28.0 (OC(CH3)2C), 77.0
(NC(CH3)2CH2), 78.4 (OC(CH3)2C), 80.6 (NC(CH3)2CH2). Calcd for
CeC32H56N4O8: C 50.25, H 7.38, N 7.32; found C 50.07, H 7.49, N 7.19%.
Characterization of [Ce2(dmop)6O] 5. [Ce2(dmop)6O] was found
as the residue after attempts to sublime [Ce(dmop)4] in vacuo.
1H NMR (400 MHz, C7D8): δ(ppm) 1.02 (s), 1.16 (d, J = 3.3 Hz),
1.44 (br s), 1.52 (br m), 1.53 (br s), 1.65 (s), 1.74 (s), 1.78 (s), 1.84 (s),
2.08 (br m), 3.47 (s), 3.57 (s), 3.64 (d, J = 3.3 Hz), 3.79 (br s), 3.82 (br
s), 3.90 (br s), 3.92 (br s). 13CNMR (100 MHz, C7D8): δ(ppm) 27.6,
30.0, 28.2, 28.4, 28.6, 28.7, 28.8 (NC(CH3)2CH2), 30.1, 30.4, 30.8, 31.2,
32.1, 32.2 (OC(CH3)2C), 66.7, 66.9, 67.3 (NC(CH3)2CH2), 77.1, 78.4,
79.4 (OC(CH3)2C), 80.7, 81.8, 82.5, 82.7 (NC(CH3)2CH2), 171.4
(OC(CH3)2C). Calcd for Ce2C48H84N6O13: C 46.74, H 6.86, N 6.81;
found C 46.89, H 6.92, N 6.85%.
[Ce(dmomp)4] 6 was synthesized in the same way as [Ce(mmp)4]
using [NH4]2[Ce(NO3)6] (532 mg, 0.970 mmol) and NaOBut (562
mg, 5.85 mmol). Following the preparation of [Ce(OBut)4], the
intermediate Hdmomp (800 mg, 3.90 mmol) in DME (10 mL) was
added to the yellow solution with stirring at room temperature. A color
change to deep red was seen and after 2 h the volatiles were removed in
vacuo to give a powdery amorphous red solid. Yield: 328 mg (35%).
1H NMR (400 MHz, CDCl3): δ(ppm) 0.95 (3H, s, NC(CH3)2CH2),
1.21 (3H, s, NC(CH3)2CH2), 2.04 (3H, s, CH3(C6H3O)), 3.87 (1H, d, J =
7.9 Hz, NC(CH3)2CH2), 3.97 [1H, d, J = 7.8 Hz, NC(CH3)2CH2), 6.54
[1H, t, C6H3), 7.23 (1H, s, C6H3), 7.71 (1H, d of d, C6H3). 13C NMR (100
MHz, CDCl3): δ(ppm) 15.2 (CH3(C6H3O)), 24.6 (NC(CH3)2CH2),
27.5 (NC(CH3)2CH2), 68.7 (NC(CH3)2CH2), 77.6 (NC(CH3)2CH2),
116.0 (aromatic), 116.4 (aromatic), 123.9 (aromatic), 126.6 (aromatic),
133.0 (aromatic), 161.8 (CH2(C5H3CO)), 165.5 (CH3(C6H3O)CN). MS
(EI): m/z = 956.3 (M+); calcd. for [C48H56CeN4O8]+, 956.3.
’ EXPERIMENTAL SECTION
General. Hmmp, [NH4]2[Ce(NO3)6] (CAN) and [Ce(thd)4] were
supplied by SAFC Hitech. Hdmop12 and Hdmomp13 were synthesized
by the published routes. Other starting materials were purchased from
Aldrich.
The preparation of all cerium complexes was performed under strictly
anaerobic and anhydrous conditions using standard Schlenk techniques.
Nondeuterated solvents were distilled from sodium/benzophenone
ketyl and stored under N2 over activated 4 Å molecular sieves prior to
use. Deuterated solvents were distilled from CaH2 prior to use. NMR
spectra were recorded on a Bruker Avance 400 spectrometer. NMR
samples of cerium complexes were sealed under vacuum. Mass spectra
were recorded on a Finnigan MAT 95 XP instrument (EI and CI) by the
EPSRC National Mass Spectrometry Service at Swansea University.
Elemental analyses were performed by Mr. Stephen Boyer of the School
of Health Sciences, London Metropolitan University.
Synthesis of Complexes. [Ce(mmp)4] 1. A solution of [NH4]2-
[Ce(NO3)6] (10.00 g, 18.24 mmol) in dimethoxyethane (DME)
(77 mL) was added at room temperature to a stirred solution of NaOBut
(10.52 g, 109.4 mmol) in DME. After removal of the NaNO3 precipitate,
the filtrate was stirred at 0 °C and Hmmp (8.52 mL, 73.0 mmol) was
added. The reaction mixture was heated under reflux for 1 h and then left
to cool to room temperature. Volatiles were removed in vacuo, and the
resulting yellow oil was extracted with hexane (3 ꢁ 100 mL), which
was then concentrated to give the product as yellow crystals on
cooling to ꢀ18 °C. (5.24 g, 52% yield).
M.p.: 105ꢀ106 °C (0.8 Torr). 1H NMR (400 MHz, C7D8): δ(ppm):
1.27 (6H, s OC(CH3)2CH2OCH3), 3.60 (2H, s, OC(CH3)2CH2OCH3),
3.49 [3H, s, OC(CH3)2CH2OCH3). 13C NMR (100 MHz, C7D8):
δ(ppm): 26.9 (OC(CH3)2CH2OCH3), 58.2 (OC(CH3)2CH2OCH3),
75.0 (OC(CH3)2CH2OCH3), 84.1 (OC(CH3)2CH2OCH3). Calcd. for
CeC20H44O8: C, 43.46 ; H, 8.02; found: C, 42.58; H, 7.88%.
[Ce(mmp)2(dbm)2] 7. A solution of Hdbm (dibenzoylmethane)
(800 mg, 3.60 mmol) in toluene (20 mL) was added dropwise via
canula to a stirred solution of [Ce(mmp)4] (1.00 g, 1.80 mmol) in
toluene (20 mL) at 0 °C. After full addition the reaction was stirred for
1 h at 0 °C and allowed to warm to room temperature for a further 1 h.
The volatiles from the deep orange solution were removed in vacuo, and
the crude oil recrystallized from toluene to give a low melting point solid
at ꢀ18 °C. Yield: 447 mg (31.2%).
[Ce(dmae)4] 2 was synthesized following the route used for
[Ce(mmp)4] using CAN (1.00 g, 1.82 mmol), NaOBut (1.05 g, 96.1
mmol), and dmaeH (0.75 mL, 7.3 mmol). The reaction mixture was
refluxed for 2 h and volatiles removed in vacuo to give a yellow oil. Yield:
350 mg (39%).
11645
dx.doi.org/10.1021/ic201593s |Inorg. Chem. 2011, 50, 11644–11652

Inorganic Chemistry
ARTICLE
Table 1. Conditions for Thin Film Growth
(2H, s, NC(CH3)2CH2), 6.56 (1H, t, J = 7.3 Hz, CH3(C6H3O)), 7.22
(1H, d, J = 7.2 Hz, CH3(C6H3O)), 7.86 (1H, d, J = 8.0 Hz, CH3(C6H3O)).
13C NMR (100 MHz, C7D8): δ(ppm) 17.0 (CH3(C6H3O)), 28.3
(OC(CH3)2CH2OCH3), 29.4 (NC(CH3)2CH2), 59.2 (OC(CH3)2-
CH2OCH3), 68.6 (OC(CH3)2CH2OCH3), 78.9 (NC(CH3)2CH2), 84.5
(OC(CH3)2CH2OCH3), 115.6 (NC(CH3)2CH2), 117.0 (CH3(C6H3O)),
127.0 (CH3(C6H3O)), 128.5 (CH3(C6H3O)), 129.3 (CH3(C6H3O)),
134.9 (CH3(C6H3O)), 166.0 (CH3(C6H3O)), 167.1 ((C6H3O)CNC-
(CH3)2). Calcd for CeC34H50N2O8: C 54.10, H 6.68, N 3.7; found C
53.86, H 6.45, N 3.56%. MS (EI): m/z = 752.3 ([M ꢀ 2H]+), calcd for
[CeC34H50N2O8 ꢀ 2H]+ 752.2; MS (CI, NH3): m/z = 753.3 ([M ꢀ H]+),
calcd for [CeC34H50N2O8 ꢀ H] 753.2.
General Growth Conditions
solvent
toluene
precursor conc
evaporator temperature
reactor pressure
argon flow rate
0.05 M
100 °C
1 mbar
200 cm3/min
MOCVD Growth Conditions
substrate temps
200ꢀ600 °C
100 cm3/minꢀ1
Reaction of [Ce(mmp)4] with 2 equiv of Hdmop. A solution of Hdmop
(499 mg, 3.17 mmol) in toluene (10 mL) was added dropwise via canula to a
stirred solution of [Ce(mmp)4] (877 mg, 1.59 mmol) in toluene (20 mL) at
0 °C, and the reaction was stirred for 2 h. The volatiles were removed in
vacuo, and the crude materials recrystallized in toluene to give [Ce2(mmp)2-
(dmop)4O] 10 as yellow crystals (253 mg, 19.3% yield).
oxygen flow rate
ALD Growth Conditions
substrate temps
oxygen source
100ꢀ350 °C
H2O: ([Ce(mmp)4], [Ce(dmap)4],
[Ce(OC(CH3)2Pri])
O3: ([Ce(thd)4])
300
1H NMR (400 MHz, C7D8): δ(ppm) 1.42 (6H, s), 1.48 (24H, m),
1.57 (6H, s), 1.65 (6H s), 1.74 (6H, s), 1.79 (6H, s), 1.96 (6H, s), 3.39
(6H s), 3.34 (2H, d, J = 2.2 Hz), 3.38 (2H, d, J = 2.2 Hz), 3.74 (2H, d, J =
8.1 Hz), 3.76 (2H, d, J = 8.1 Hz), 3.78 (2H, d, J = 8.1 Hz), 3.91 (2H, d, J =
8.1 Hz). Calcd for Ce2C42H78O13N4: C 44.75, H 6.97, N 4.97; found C
44.60, H 6.84, N 4.90%.
number of cycles
pulse sequence (s)
2/2/0.5/3.5
precursor/purge/oxygen source/purge
Attempted Synthesis of [Ce(ONep)4]. [Ce(OBut)4] was prepared
from [NH4]2[Ce(NO3)6] (1.00 g, 1.82 mmol) in dimethoxyethane
(DME) (20 mL) and NaOBut (1.05 g, 10.9 mmol) in DME. After
removal of the NaNO3 precipitate, the filtrate was stirred at room
temperature and a solution of neopentyl alcohol (642 mg, 7.28 mmol) in
toluene (4 mL) was added. The reaction mixture was heated under reflux
for 1 h, and then left to cool to room temperature. Volatiles were
removed in vacuo and recrystallization of the oil attempted in toluene at
ꢀ18 °C. Yield of crystalline material = 340 mg. Single crystal X-ray
diffraction showed this material to be [Ce3(OBut)(ONep)9O] 11.
Calcd for Ce3C49H108O11: C 45.49, H 8.41%; found C 40.89, H
7.63%. 1H NMR (400 MHz, C7D8): δ(ppm) 0.81(13 H, broad s), 3.04
(2H, broad s). 13C NMR (C7D8): δ(ppm) 26.7, 27.1.
Crystal Structure Determinations. Crystals were mounted on
glass fibers and placed in a cold stream at 100 K. Single crystal X-ray data
were collected on a Bruker D8 diffractometer with an APEX CCD
detector, and 1.5 kW graphite monochromated Mo radiation. The
detector to crystal distance was 5.98 cm. Exposure times of 10 s per
frame and scan widths of 0.3° were used throughout the data collections.
The frames were integrated with SAINT v7.45a.14 The structures were
solved and refined with X-SEED,15 a graphical interface to SHELX.16
Crystal data are summarized in Table 2, and a cif file is available in the
Supporting Information.
1H NMR (400 MHz, d8-thf): δ(ppm) 1.06 (6H, s, C(CH3)2), 3.08
(2H, s, CH2), 3.28 (3H, s, OCH3), 6.85 (1H, s, (C6H5CO)2CH), 7.27
(m, aromatic), 7.38 (m, aromatic), 8.10 (m, aromatic). 13C NMR (100
MHz, d8-thf): δ(ppm) 26.8 (OC(CH3)2CH2OCH3), 59.2 (OC(CH3)2-
CH2OCH3), 69.6 (OC(CH3)2CH2OCH3), 82.7 (OC(CH3)2CH2OCH3),
98.3 ((C6H5CO)2CH), 128.8 (aromatic), 129.6 (aromatic), 130.5
(aromatic), 130.6 (aromatic), 131.2 (aromatic), 131.9 (aromatic), 139.8
(aromatic), 185.0 ((C6H5CO)2CH). Calcd for CeC40H44O8: C 60.59, H
5.59; found C 60.68, H 5.40%. MS (ASAP): m/z = 793.22 ([M + H]+);
calcd for [CeC40H44O8 + H] 793.22.
[Ce(dmap)2(dbm)2] 8. A solution of Hdbm (dibenzoylmethane)
(433 mg, 1.93 mmol) in toluene (4 mL) was added dropwise via canula
to a stirred solution of [Ce(mmp)4] (530 mg, 0.970 mmol) in toluene
(20 mL) at 0 °C, and the reaction was stirred for 2 h. Volatiles were
removed in vacuo to give the compound as an orange/red oil. Yield:
613 mg (79.9%).
1H NMR (400 MHz, C7D8): δ(ppm) 1.08 (3H, d, J = 5.8 Hz,
OCH(CH3)CH2N(CH3)2), 2.25 [6H, s, OCH(CH3)CH2N(CH3)2),
2.44 (1H, d, J = 9.8 Hz, OCH(CH3)CH2N(CH3)2), 3.41 [1H, t, J = 11.4
Hz, OCH(CH3)CH2N(CH3)2), 5.23 [1H, br m, OCH(CH3)CH2N-
(CH3)2), 6.43 (1H, s, (C6H5CO)2CH), 7.07 [6H, t, J = 7.5 Hz,
aromatic), 7.17 [5H, d, J = 7.1 Hz, aromatic], 7.17 (2H, d, J = 8.0 Hz,
aromatic), 7.26 [3H, t, J = 7.1 Hz, aromatic), 7.78 (3H, d, J = 7.5 Hz,
aromatic). 13C NMR (100 MHz, CDCl3): δ(ppm): 20.3 (OCH(CH3)-
CH2N(CH3)2), 46.2 (OCH(CH3)CH2N(CH3)2), 71.3 (OCH(CH3)-
CH2N(CH3)2), 95.7 ((C6H5)2(CO)2CH), 124.2 (aromatic), 126.5
(aromatic), 126.8 (aromatic), 127.1 (aromatic), 127.9 (aromatic),
129.3 (aromatic), 136.6 (aromatic), 139.1 (aromatic), 182.9 ((C6H5)2-
(CO)2CH). Calcd forC40H46CeN2O6: C 60.74, H 5.86, N 3.54;foundC
60.64, H 6.03, N 3.46%
[Ce(mmp)2(dmomp)2] 9. A solution of dmompH (179 mg, 0.870
mmol) in toluene (10 mL) was added dropwise via canula to a stirred
solution of [Ce(mmp)4] (241 mg, 0.870 mmol) in toluene (10 mL) at
ꢀ83 °C (liquid N2/ethylacetate). On addition of the ligand a color
change from yellow to orange was seen. After stirring for 1 h the volatiles
were removed in vacuo to give an orange oil. Yield: 185 mg (56.2%).
1H NMR (400 MHz, C7D8): δ(ppm) 1.14 (3H, br, s, NC-
(CH3)2CH2), 1.31 (6H, s, OC(CH3)2CH2OCH3), 1.66 (3H, br s,
NC(CH3)2CH2), 2.49 (3H, s, CH3(C6H3O)), 3.19 (2H, s, OC-
(CH3)2CH2OCH3), 3.22 (3H, s, OC(CH3)2CH2OCH3), 3.56
Thin Film Growth and Characterization. The CeO2 films were
deposited on n-Si(100) wafers. Deposition of the CeO2 films was
performed by liquid-injection ALD or liquid-injection MOCVD using
an Aixtron AIX 200FE AVD reactor fitted with a modified liquid-injector
system.17 The delivery of cerium precursor was achieved using a 0.05 M
solution of the precursor compound in toluene. Deionized water or
ozone (for [Ce(thd)4]) was used as a source of oxygen for ALD growth.
Details of the MOCVD and ALD growth conditions are summarized in
Table 1.
X-ray diffraction patterns were measured using a Rigaku Miniflex
diffractometer using Cu Kα radiation (0.154051 nm, 40 kV, 50 mA)
spanning a 2θ range of 20ꢀ50° at a scan rate of 0.01°/min.
Raman spectra were acquired using a Jobin-Yvon LabRam HR
consisting of a confocal microscope coupled to a single grating spectro-
meter equipped with a notch filter and a CCD camera detector. The
Raman measurements were taken using a confocal aperture of 300 μm to
limit the light taken to the center of the microscope lens and hence
11646
dx.doi.org/10.1021/ic201593s |Inorg. Chem. 2011, 50, 11644–11652

Inorganic Chemistry
ARTICLE
Table 2. Crystal Structure and Refinement Data
[Ce(dmap)4] 3
[Ce(dmop)4] C7H84
[Ce2(mmp)2(dmop)4O] C7H810
[Ce3(OBut)(ONep)9O] 11
3
3
habit
prism
prism
prism
prism
formula
formula weight
space group
a, Å
C20H48CeN4O4
548.74
P6122
C39H64CeN4O8
857.06
P21/c
C49H86Ce2N4O13
1219.46
P21
C73H152Ce3O11
1626.31
P3c1
9.8993(3)
9.8993(3)
46.8411(16)
90.0
11.514(5)
12.546(6)
29.583(13)
90.0
10.791(3)
18.425(5)
14.533(4)
90.0
19.0918(17)
19.0918(17)
23.342(2)
90.0
b, Å
c, Å
α, deg
β, deg
90.0
96.245(8)
90.0
99.530(4)
90.0
90.0
γ, deg
120.0
120.0
volume, Å3
3975.3(2)
6
4248(3)
4
2849.6(14)
2
7368.2(11)
4
Z
Fcalc, g/cm
1.375
1.340
1.421
1.466
F000
1716
1792
1256
3408
temp, K
100(2)
2.4
100(2)
2.15
100(2)
3.24
100(2)
4.74
R1 (I > 2σ(I)), %
goodness of fit
1.335
1.069
1.057
0.969
improve the quality of signal achieved. The measurements spanned
150ꢀ1500 cmꢀ1 with four accumulations and an exposure time of 30 s.
All of the spectra were recorded in a backscattering geometry using an
incident wavelength of 325 nm from a HeꢀCd laser.
Scheme 1. β-Hydride Elimination Reaction of a tert-Butox-
ide Complex
’ RESULTS AND DISCUSSION
Choice of Potential Precursor Complexes. Some Ce(III)
complexes, for example β-ketoiminates,18ꢀ20 have been used as
precursors for CVD of CeO2. However the synthesis of volatile
Ce(III) precursors poses a real challenge: it is difficult to fill the
coordination sphere of the large Ce3+ ion. Therefore, many
Ce(III) complexes with monodentate ligands form dimers or
oligomers, and their volatility is consequently reduced. A further
potential problem is that many Ce(III) complexes are very prone
to at least partial oxidation, making them difficult to obtain and
keep in a pure state. We therefore chose to work with Ce(IV): the
increased charge and decreased radius of Ce4+ compared with
Ce3+ means that, providing the +4 oxidation state can be
stabilized, it should be less difficult to obtain monomeric (and
hence potentially more volatile) complexes. There are several
well-known Ce(IV) complexes with monodentate alkoxide li-
gands e.g. [Ce(OBut)4], [Ce2(OiPr)8], and [Ce(OCMe2Pri)4].
[Ce(OBut)4] and [Ce2(OiPr)8] are unsuitable as CVD precur-
sors: they are thermally very unstable and readily decompose to
form O-bridged clusters with much reduced volatility.21,22
[Ce(OBut)4] is also reported to undergo partial reduction in
solution, resulting in the formation of the mixed-valence cluster
[Ce3(OBut)11] which contains two Ce4+ and one Ce3+.23
[Ce(OCMe2Pri)4] has been used for MOCVD of CeO2;8
however, we found that it has an extremely narrow temperature
window for MOCVD (see Figure 9) which seriously limits its use
in our reactor system. The thermal instability of [Ce(OBut)4]
and [Ce2(OiPr)8] can be attributed to the facile elimination
reaction shown in Scheme 1 below (generally referred to by CVD
chemists as a “β-hydride” elimination reaction)24 and formation
of a reactive CeꢀOH moiety which can then react further to
form oxo-bridged species.
We therefore initially decided to investigate [Ce(ONep)4]
(ONep = neopentoxide, OCH2CMe3) which cannot undergo
the β-hydride elimination reaction and so should show enhanced
stability. However, synthesis of this complex via a salt metathesis
reaction of [NH4]2[Ce(NO3)6] with NaONep was unsuccessful,
and the alcohol exchange reaction of [Ce(OBut)4] with 4 equiv
of HONep gave an oily product which crystallized to give [Ce3(μ3-O)
(μ3-OBut)(μ2-ONep)3(ONep)6] 11 (see Figure 7). The reten-
tion of one OBut ligand in this complex suggests that the
condensation to an oxo-bridged trimer had occurred before the
alcohol exchange reaction was complete.
Synthesis of Complexes. Homoleptic Complexes. We initi-
ally attempted to synthesize new complexes by a salt metathesis
reaction between NaOR and ceric ammonium nitrate ([NH4]2-
[Ce(NO3)6], CAN);25 however, this method never gave satis-
factory yields and frequently gave none of the desired product at
all. We therefore used alcohol exchange reactions of
[Ce(OBut)4], which is readily synthesized from NaOBut and
[NH4]2[Ce(NO3)6]. [Ce(OBut)4] is somewhat fragile in solu-
tion, readily decomposing to form [Ce3(μ3-O)(OBut)10],21 and
so, we used a simplified synthesis, avoiding the isolation of
[Ce(OBut)4], as outlined in Scheme 2.
The synthesis of homoleptic complexes by this route was
straightforward for all the donor-functionalized alkoxide ligands
that we investigated. For example, to synthesize [Ce(mmp)4] 1,
4 equiv of Hmmp was added at room temperature to the
[Ce(OBut)4] solution, and after heating to reflux for 2 h, pure
product was obtained as yellow crystals in up to 80% yield.
11647
dx.doi.org/10.1021/ic201593s |Inorg. Chem. 2011, 50, 11644–11652

Inorganic Chemistry
ARTICLE
Scheme 2. Synthesis of Cerium(IV) Alkoxides via Alcohol
Exchange
[Ce(mmp)2L2] and [Ce(dmap)2L2] by careful addition at low
temperature of 2 equiv of HL to [Ce(mmp)4] or [Ce(dmap)4],
and the complexes were characterized by a combination of
elemental analysis, NMR spectroscopy, and mass spectrometry.
In the case of L = dmop, we were unable to isolate [Ce(mmp)2-
(dmop)2], but instead the oxo-bridged dimer [Ce2(mmp)2-
(dmop)4(μ2-O)] 10 was formed and characterized by single
crystal X-ray diffraction. Heteroleptic β-diketonate complexes
with L = dibenzoylmethanate (dbm) were also prepared by careful
reaction of [Ce(mmp)4] or [Ce(dmap)4] with 2 equiv of Hdbm
(dibenzoylmethane); however, all of our attempts to prepare
heteroleptic complexes with thd resulted only in formation of
[Ce(thd)4]. The heteroleptic complexes were generally found to
be labile in solution: ligand redistribution reactions occurred
readily to form mixtures, and so, these complexes were not
investigated further as CVD precursors.
Description of Structures. The static structure of an 8-co-
ordinate complex with four bidentate chelating ligands must be
chiral unless it has exactly square-prismatic geometry and the
ligands contain a plane of symmetry. Although chirality of
6-coordinate tris-chelate complexes has been the subject of many
studies, the corresponding phenomenon in 8-coordinate com-
plexes has been relatively unexplored. Here we have an oppor-
tunity to examine the structures of three related tetrakis-chelate
complexes, including one containing chiral ligands.
[Ce(mmp)4] 1. A preliminary report of this complex has
already been published;11 an ORTEP plot of the complex is
given in Figure 1 for comparison with other related structures.
The Ce atom lies on a crystallographic C2 axis (see Figure 1b),
and it has approximate D2 symmetry. The room temperature 1H
NMR spectrum of this complex (Figure S1 of the Supporting
Information) shows that in solution all four mmp ligands are
equivalent, and all contain a time-averaged plane of symmetry,
which is consistent with D2d symmetry. Rapid inversion of the
chelate rings would account for the high-symmetry solution
structure; we have been unable to freeze out this process by
low temperature 1H NMR spectroscopy.
Similar reaction conditions were used for [Ce(dmap)4] 3 and
[Ce(dmomp)4] 6, and these complexes were isolated as crystal-
line solids. Unfortunately 6 had poor solubility in toluene and
decomposed in more polar solvents such as THF or CH2Cl2 and
was not investigated further. Synthesis of [Ce(dmop)4] 4
required addition of Hdmop to be performed at 0 °C, followed
by reaction at room temperature; otherwise, the product was
contaminated with what we believe from analytical data to be
[Ce2(dmop)6O] 5 (see the Experimental Section).
The reaction of [Ce(OBut)4] with 4 equiv of Hdmae, which is
less sterically demanding than Hdmap, was not straightforward:
addition of Hdmae at 0 °C followed by stirring at room tem-
perature resulted in incomplete reaction. When the addition of
Hdmae was performed at room temperature followed by heating
to reflux for 2 h, a yellow oil analyzing as [Ce(dmae)4] 2 was
isolated, but its 1H NMR spectrum was complex (see Figure S2),
indicating several different ligand environments. The reaction
was not always reproducible, and attempts to sublime 2 in vacuo
resulted in decomposition, and elimination of dmaeH. 2 also
displayed instability in solution, and we therefore abandoned
further investigations of this complex.
Except for [Ce(dmae)4] 2 and [Ce(dmomp)4] 6, all of the
homoleptic donor-functionalized alkoxide complexes that we
prepared were found by NMR spectroscopy to be indefinitely
stable in solution at room temperature: unlike [Ce(OBut)4] and
[Ce2(OPri)8(HOPri)2], there was no evidence of condensation
reactions to form oxo-bridged clusters nor was there any
evidence of reduction to Ce(III).
Heteroleptic Complexes. We synthesized several heteroleptic
complexes with the aim of modifying reactivity and volatility to
optimizeperformancefor CVD. Initially weinvestigated formation
of heteroleptic complexes by addition of a mixture of HL pro-
ligands to [Ce(OBut)4]; however, this method was never success-
ful, generally leading to a mixture of homoleptic complexes.
We therefore prepared a series of heteroleptic complexes
[Ce(dmap)4] 3. The dmap ligand is chiral by virtue of the α-
Me substituent, and as we used racemic Hdmap in the prepara-
tion of 3, the formation of diastereomeric mixtures would be
possible. In the solid state the complex crystallizes a single
Figure 1. (a) ORTEP plot of [Ce(mmp)4] 1. Selected distances (Å) and angles (deg): Ce(1)ꢀO(1), 2.155(5); Ce(1)ꢀO(2), 2.667(4) ; Ce(1)ꢀO(3),
2.155(4); Ce(1)ꢀO(4), 2.684(4); O(2)ꢀCe(1)ꢀO(1), 63.99(15); O(4)ꢀCe(1)ꢀO(3), 63.81(16); O(1)ꢀCe(1)ꢀO2*, 76.54(14); O(3)ꢀCe-
(1)ꢀO(4)*, 76.29(13); O(2)ꢀCe(1)ꢀO(3), 153.13(16); O(1)ꢀCe(1)ꢀO(3), 96.09(17); O(1)ꢀCe(1)ꢀO(4), 153.34(13). O(1), O(1)*, etc. are
related by a C2 rotation. (b) View along crystallographic C2 axis.
11648
dx.doi.org/10.1021/ic201593s |Inorg. Chem. 2011, 50, 11644–11652

Inorganic Chemistry
ARTICLE
Figure 3. [Ce(dmap)4] 3 viewed along the crystallographic C2 axis.
(C atoms are omitted for clarity).
Figure 2. ORTEP plot of [Ce(dmap)4] 3. Selected distances (Å) and
angles (deg) are the following: Ce1ꢀO1, 2.1661(19); Ce1ꢀO2,
2.171(2); Ce1ꢀN1, 2.784(2); Ce1ꢀN2, 2.851(3); O1ꢀCe1ꢀN1,
64.51(8); O2ꢀCe1ꢀN2, 64.00(7); O1ꢀCe1ꢀO1, 94.06(12); N1ꢀ
Ce1ꢀN1, 122.16(11); O2ꢀCe1ꢀO2, 97.97(12); N2ꢀCe1ꢀN2,
120.20(11). O1, O1*; N1, N1*; etc. are related by a C2 rotation.
diastereomer (Figure 2) containing two R-dmap ligands (O2N2
and O2*N2*) and two S-dmap ligands (O1N1 and O1*N1*).
Similar disastereoselectivity has been observed in [Cu(R-dmap)
(S-dmap)]26 and [Zr2(OiPr)5(μ2-OiPr)(μ2-R-dmap)(μ2-S-dmap)].9
However, despite the racemic nature of the ligand combination in
3, the complex itself is chiral, and it undergoes a spontaneous
resolution to crystallize in the rare chiral space group P6122 (143
examples out of a total of 537 899 in the Cambridge Crystal-
lographic Database; January 2011). The donor atoms of the two
S-ligands (N1, O1, N1*, O1*) are coplanar, as are those of the
two R-ligands (N2, O2, N2*, O2*).
A simplified view of the complex along the crystallographic C2
axis (Figure 3) shows how the two parallelograms defined by
N1O1N1*O1* and N2O2N2*O2* are twisted by ca. 84° with
respect to each other, thus rendering the complex chiral at the Ce
atom. The average bite angle of the dmap ligands is 64.25°; the
average CeꢀO distance is 2.168 Å, and the average CeꢀN
distance is 2.817 Å.
Figure 4. ORTEP plot of [Ce(dmop)4] 4. Selected distances (Å) and
angles (deg): Ce1ꢀO1, 2.1811(11); Ce1ꢀO3, 2.1973(11); Ce1ꢀO5,
2.1758(11); Ce1ꢀO7, 2.1836(11); Ce1ꢀN1, 2.7124(14); Ce1ꢀN2,
2.7246(14); Ce1ꢀN3, 2.6587(13); Ce1ꢀN4, 2.7499(14); O1ꢀCe1ꢀ
N1, 63.94(4); O3ꢀCe1ꢀN2, 64.16(4); O5ꢀCe1ꢀN3, 64.70(4);
O7ꢀCe1ꢀN4, 64.74(4); N3ꢀCe1ꢀN1, 122.32(4); N2ꢀCe1ꢀN4,
106.04(4); O5ꢀCe1ꢀO1, 99.99(4) ; O7ꢀCe1ꢀO3, 113.63(4).
[Ce(mmp)4] adopts a similarly chiral structure in the solid
state, but both enantiomers are present in the unit cell.
The simplicity of the H NMR spectrum of [Ce(dmap)4]
O5O1Ce1 and O3O7Ce1), making the complex chiral, but in
this case the unit cell has a center of symmetry and so both
enantiomeric forms of the complex are present. The average bite
angle of the dmop ligands in 4 is 64.38°, very similar to that of the
dmap ligands in [Ce(dmap)4]. The average CeꢀO distance
(2.184 Å) is slightly longer than in [Ce(dmap)4] whereas the
average CeꢀN distance (2.711 Å) is slightly shorter.
1
(Figure S3 of the Supporting Information) indicates that the
solution structure consists of a single highly symmetrical diaster-
eomer with all four ligands equivalent, which is a higher symmetry
than observed in the solid state. A rotation about the crystal-
lographic C2 axis of the plane defined by N2O2N2*O2* to give an
eclipsed conformation rather than the staggered conformation
shown in Figure 3 would introduce a plane of symmetry into the
complex. All four ligands are equivalent in the resulting C2h
structure. We have not been able to freeze out this process by low
temperature NMR spectroscopy.
[Ce(dmop)4] 4. An ORTEP plot of 4 is shown in Figure 4. The
structure of this 8-coordinate complex in the solid state is very
similar to that of [Ce(dmap)4]: the donor atoms (N1, O1, N3, O5)
of two of the chelating dmop ligands are close to coplanar, as are
those of the other two ligands (N2, O3, N4, O7). As shown in
Figure 5, the two planes are twisted by ca. 51° with respect to
each other (this is the angle between the two planes defined by
1
The H NMR spectrum of 4 (Figure S5 of the Supporting
Information), like those of [Ce(mmp)4] and [Ce(dmap)4],
shows a highly symmetrical structure in solution at room tem-
perature: all of thedmop ligands are equivalent, and all have a time-
averaged plane of symmetry. A time-averaged plane of symmetry
through the dmop ligands can only be achieved through a facile
fluxional process at room temperature, resulting in isomerization/
racemization of the complex. We have been unable to freeze out
this fluxional process by low temperature NMR spectroscopy.
[Ce2(μ2-O)(dmop)4(mmp)2] 10. This oxo-bridged dimer is
chiral with approximate C2 symmetry, and it crystallizes in the
chiral space group P21. An ORTEP plot of the complex is shown
11649
dx.doi.org/10.1021/ic201593s |Inorg. Chem. 2011, 50, 11644–11652

Inorganic Chemistry
ARTICLE
Figure 6. ORTEP plot of [Ce2(dmop)4(mmp)2(O)] 10. Selected dis-
tances (Å) and angles (deg): Ce1ꢀO12, 2.078(5); Ce1ꢀO5, 2.106(5);
Ce1ꢀO10, 2.239(5); Ce1ꢀO3, 2.395(5); Ce1ꢀO1, 2.407(5); Ce1ꢀN1,
2.603(5); Ce1ꢀN4, 2.619(6); Ce2ꢀO5, 2.106(4); Ce2ꢀO8, 2.143(5);
Ce2ꢀO6, 2.190(5); Ce2ꢀO1, 2.379(5); Ce2ꢀO3, 2.459(5); Ce2ꢀN3,
2.556(6); Ce2ꢀN2, 2.653(5); Ce2ꢀO1ꢀCe1 89.85(15); Ce1ꢀO3ꢀCe2
88.25(15); Ce1ꢀO5ꢀCe2, 106.72(17); O6ꢀCe2ꢀN3, 65.99(18); O1ꢀ
Ce1ꢀN1, 63.71(19); O10ꢀCe1ꢀN4, 64.72(18); O3ꢀCe2ꢀN2,
62.29(18).
Figure 5. Coordination sphere of one enantiomeric form of [Ce(dmop)4]
(C atoms omitted for clarity).
in Figure 6 below, along with selected distances and angles. The
two Ce atoms are symmetrically bridged by an oxo ligand and the
alkoxide groups of two dmop ligands, and there is an approximate
C2 axis passing through O5 and the midpoint between Ce1 and
Ce2. The mmp ligands are both “dangling”, as in several other
rare earth complexes with this ligand,27,28 demonstrating again
that the OMe group is not a powerful donor. The CeꢀO and
CeꢀN distances for the nonbridging dmop ligands are very
similar to those in [Ce(dmop)4]; as expected, the corresponding
distances for the bridging dmop ligands are somewhat longer.
The CeꢀO (alkoxide) distances for the dangling mmp ligands
(average 2.110 Å) are very similar to the corresponding distances
in the chelating mmp ligands of [Ce(mmp)4] (average 2.152 Å).
1
The H NMR spectrum of [Ce2(μ2-O)(dmop)4(mmp)2]
(Figure S6 of the Supporting Information) shows eight unique
Me groups, indicating that the solid state structure is retained in
solution.
[Ce3(μ3-O)(μ3-OBut)(μ2-ONep)3(ONep)6] 11. An ORTEP
plot of this complex is shown in Figure 7. The [M3(μ3-O)2]
moiety is a familiar motif in f-element alkoxide chemistry, and
examples are known where both μ3-O atoms are alkoxides29 and
where one is an alkoxide and one an oxide30 or hydroxide.28 11 is
the first fully characterized example with three Ce(IV) atoms and
the structure is analogous to that proposed for [Ce3O(OBut)10]21
by analogy with [U3O(OBut)10].30 The μ3-O and μ3-OBut
ligands are on a crystallographic C3 axis and so are bonded
symmetrically to all three Ce atoms, whereas the μ2-OBut bridges
are slightly unsymmetrical with CeꢀO distances of 2.331(3) and
2.305(5) Å.
Figure 7. ORTEP plot of [Ce3(μ3-O)(μ3-OBut)(μ2-ONep)3-
(ONep)6] 11. O3, O30, O3#, etc. are related by a C3 rotation. Selected
distances (Å) and angles (deg): Ce1ꢀO1, 2.530(4); Ce1ꢀO2, 2.250(2);
Ce1ꢀO3, 2.331(3); Ce1ꢀO3#, 2.305(5); Ce1ꢀO4, 2.054(4); Ce1ꢀO5,
2.081(4); Ce1ꢀO1ꢀCe1#, 87.52(18); Ce1ꢀO2ꢀCe1#, 102.13(13);
Ce1ꢀO3ꢀCe1#, 98.06(15).
Applications to MOCVD and ALD of CeO2. TGA Studies.
Precursors for MOCVD or ALD must generally have some
volatility, and TGA can be a useful technique for assessing this.
Figure 8 shows TGA traces for [Ce(mmp)4], [Ce(dmap)4],
[Ce(dmop)4], and [Ce(thd)4]. [Ce(mmp)4] sublimed cleanly at
atmospheric pressure: sublimation was complete at 250 °C (50%
weight loss at 226 °C), leaving a residue of <1.5%. This can be
compared with [Ce(thd)4], probably the most widely used pre-
cursor for CVD of CeO2, (50% weight loss at 275 °C and no
residue). None of [Ce(dmap)4], [Ce(dmop)4], or [Ce(dmomp)4]
displayed any volatility at atmospheric pressure: all left substan-
tial residues (30% and 26%, respectively), consistent with
decomposition to CeO2. However, we routinely use liquid-
injection MOCVD and ALD at reduced pressures, and so, lack
of volatility at atmospheric pressure is not necessarily a serious
problem. Both [Ce(mmp)4] and [Ce(dmap)4] sublimed in
vacuo; however, attempts to sublime [Ce(dmop)4] in vacuo
resulted only in sublimation of Hdmop, and a material analyzing
as [Ce2(dmop)6(O)] was isolated from the residue. [Ce-
(dmomp)4] was also involatile in vacuo, so we did not pro-
ceed with CVD studies using complexes of dmop or dmomp
ligands.
MOCVD. In order to assess the potential of our new precursors,
we compared them with the known precursors [Ce(thd)4] and
[Ce(OCMe2Pri)4].8 MOCVD growth curves for [Ce(mmp)4],
[Ce(dmap)4], [Ce(thd)4], and [Ce(OCMe2Pri)4] with O2 as
coreactant are shown in Figure 9 below. [Ce(mmp)4] 1 and
[Ce(dmap)4] 3 both have low temperatures for onset of growth,
and although these temperatures were slightly higher than we
observed for [Ce(OCMe2Pri)4], the temperature windows for
11650
dx.doi.org/10.1021/ic201593s |Inorg. Chem. 2011, 50, 11644–11652

Inorganic Chemistry
ARTICLE
Figure 8. TGA traces for Ce complexes (N2 atmosphere; heating rate
10 °C/min).
Figure 10. Liquid-injection ALD growth curves for [Ce(thd)4],
[Ce(OCMe2Pri)4], [Ce(mmp)4], and [Ce(dmap)4]. The oxygen
source was H2O for [Ce(OCMe2Pri)4], [Ce(mmp)4], and [Ce(dmap)4];
it was O3 for [Ce(thd)4].
Figure 9. Liquid-injection MOCVD growth curves for [Ce(thd)4],
[Ce(OCMe2Pri)4], [Ce(mmp)4], and [Ce(dmap)4].
MOCVD growth were dramatically wider. We also investigated
the possibility of using 1 and 3 as single-source precursors for
CeO2, i.e. in the absence of O2. Under these conditions, 1
showed growth rates at 350 and 500 °C that were only ca. 10%
of those observed in the presence of O2 whereas 3 showed a
growth rate at 400 °C that was ca. 55% of that in the presence of
O2, and therefore, significantly higher than that achieved with
[Ce(thd)4] in the presence of O2. 3 therefore has potential as a
single source precursor for CeO2. In addition to the obvious
growth rate advantages of 1 and 3 when compared with
[Ce(thd)4], we also found that our new alkoxide precursors
are much more soluble in toluene than [Ce(thd)4]. We have
found that the poor solubility of [Ce(thd)4] frequently results in
experimentally inconvenient injector blockages.
ALD. ALD growth curves for [Ce(mmp)4] 1 and [Ce(dmap)4]
3 with H2O as coreactant are shown in Figure 10, along with data
for [Ce(thd)4] and [Ce(OCMe2Pri)4] for comparison. It is
important to note that due to the weak Brønsted basicity of
[Ce(thd)4], no growth of CeO2 is observed using H2O as
coreactant, and ozone must be used as the oxygen source. 1
has excellent ALD characteristics: much higher growth rates than
for [Ce(OCMe2Pri)4]/H2O, or [Ce(thd)4]/O3, and a large
window for self-limiting ALD growth from 150 to 350 °C, which
is important in order to achieve high reproducibility of film
growth and precise control of film thickness by the number of
Figure 11. Raman spectra (325 and 514 nm) and XRD (inset) of CeO2
thin film from [Ce(dmap)4] (MOCVD, 400 °C, as-grown).
deposition cycles. 3 is less thermally stable than 1 and does not
show self-limiting behavior in ALD: there is a large increase in
growth rate between 100 and 200 °C indicating thermal decom-
position, which is consistent with the onset of MOCVD growth
at 200 °C for this precursor.
Characterization of Thin Films. The CeO2 thin films were
characterized by XRD and Raman spectroscopy. Figure 11
shows the 325 and 514 nm Raman spectra for as-grown thin films
deposited by MOCVD from [Ce(dmap)4] at 400 °C. Ceria has
one triply degenerate Raman mode at ∼465 cmꢀ1 with F2g
symmetry which is present in traces from both laser sources.
The A1g modes seen at ∼600 and ∼1180 cmꢀ1 are second-order
modes only picked up by the 325 nm source. Importantly, neither
trace shows evidence of the characteristic D and G carbon bands at
∼1350 and ∼1600 cmꢀ1, respectively, demonstrating minimal
carbon inclusion in the films during deposition. These measure-
ments and the XRD plot in the inset show that the films were cubic
as grown, and this was found for all the films grown in the
300ꢀ600 °C temperature window. Films grown at 350 °C by
MOCVD were analyzed by XPS giving the following composi-
tions: [Ce(mmp)4] precursor Ce, 30.6; O, 62.0; C, 7.4 at %;
11651
dx.doi.org/10.1021/ic201593s |Inorg. Chem. 2011, 50, 11644–11652

Inorganic Chemistry
ARTICLE
[Ce(dmap)4] precursor Ce, 31.8; O, 65.9; C, 2.3 at %. The
compositions of the films were derived from the peak areas of
the Ce(5d(3/2, 5/2)), O(1s), and C(1s) signals after a nominal
ion bombardment to remove adventitious surface contamination.
The measurement of carbon content is complicated by the
presence of an overlapping Ce XPS peak, and results are are to
some extent notional. The XPS data show that the Ce:O ratio is
close to 1:2 for films grown from [Ce(mmp)4] and from
[Ce(dmap)4]. Combined with Raman and XRD data this con-
firms the bulk composition of the films to be CeO2. A detailed
study of materials properties of CeO2 grown by ALD using
[Ce(mmp)4] is published elsewhere.31
(4) Jasinski, P.; Suzuki, T.; Anderson, H. U. Sens. Actuators B-Chem.
2003, 95, 73–77.
(5) Atkinson, A.; Barnett, S.; Gorte, R. J.; Irvine, J. T. S.; McEvoy,
A. J.; Mogensen, M.; Singhal, S. C.; Vohs, J. Nat. Mater. 2004, 3, 17–27.
(6) Jones, A. C.; Aspinall, H. C.; Chalker, P. R.; Potter, R. J.;
Manning, T. D.; Loo, Y. F.; O’Kane, R.; Gaskell, J. M.; Smith, L. M.
Chem. Vap. Deposition 2006, 12, 83–98.
(7) Leskela, M.; Ritala, M. Angew. Chem., Int. Ed. 2003, 42, 5548–5554.
(8) Suh, S.; Guan, J.; Miinea, L. A.; Lehn, J.-S. M.; Hoffman, D. M.
Chem. Mater. 2004, 16, 1667–1673.
(9) Fleeting, K. A.; O’Brien, P.; Jones, A. C.; Otway, D. J.; Andrew,
A. J. P.; Williams, D. J. J. Chem. Soc.-Dalton Trans. 1999, 2853–2859.
(10) Loo, Y. F.; O’Kane, R.; Jones, A. C.; Aspinall, H. C.; Potter, R. J.;
Chalker, P. R.; Bickley, J. F.; Taylor, S.; Smith, L. M. Chem. Vap.
Deposition 2005, 11, 299–305.
(11) Wrench, J. S.; Black, K.; Aspinall, H. C.; Jones, A. C.; Bacsa, J.;
Chalker, P. R.; King, P. J.; Werner, M.; Davies, H. O.; Heys, P. N. Chem.
Vap. Deposition 2009, 15, 259–261.
(12) Pridgen, L. N.; Miller, G. J. Heterocycl. Chem. 1983, 20, 1223–1230.
(13) Aspinall, H. C.; Beckingham, O.; Farrar, M. D.; Greeves, N.;
Thomas, C. D. Tetrahedron Lett. 2011, 52, 5120–5123.
(14) SAINT V7.68a; BRUKER AXS Inc.: WI, USA, 2009.
(15) Barbour, L. J. J. Supramolec. Chem. 2001, 1, 189–191.
(16) Sheldrake, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr.
2008, 64, 112–122.
(17) Potter, R. J.; Chalker, P. R.; Manning, T. D.; Aspinall, H. C.;
Loo, Y. F.; Jones, A. C.; Smith, L. M.; Critchlow, G. W.; Schumacher, M.
Chem. Vap. Deposition 2005, 11, 159–169.
(18) Belot, J. A.; Wang, A.; McNeely, R. J.; Liable-Sands, L.;
Rheingold, A. L.; Marks, T. J. Chem. Vap. Deposition 1999, 5, 65–69.
(19) Edleman, N. L.; Wang, A.; Belot, J. A.; Metz, A. W.; Babcock,
J. R.; Kawaoka, A. M.; Ni, J.; Metz, M. V.; Flaschenriem, C. J.; Stern,
C. L.; Liable-Sands, L. M.; Rheingold, A. L.; Markworth, P. R.; Chang,
R. P. H.; Chudzik, M. P.; Kannewurf, C. R.; Marks, T. J. Inorg. Chem.
2002, 41, 5005–5023.
(20) Wang, A.; Belot, J. A.; Marks, T. J.; Markworth, P. R.; Chang,
R. P. H.; Chudzik, M. P.; Kannewurf, C. R. Physica C (Amsterdam) 1999,
320, 154–160.
’ CONCLUSIONS
We have prepared a series of homoleptic and heteroleptic
complexes of cerium(IV) with donor-functionalized alkoxide
ligands and investigated their potential as precursors for the
MOCVD and ALD of CeO2. The homoleptic complexes with
mmp, dmap, or dmop ligands show good solution stability with
no tendency to decompose either to oxo-bridged structures or to
Ce(III) species. The heteroleptic complexes have a tendency to
undergo ligand redistribution reactions in solution. The
eight-coordinate homoleptic complexes [Ce(mmp)4] 1 and
[Ce(dmap)4] 3 are both excellent precursors for liquid-
injection MOCVD of CeO2: they exhibit high growth rates
and large growth windows, and the thin films show very low
carbon contamination, particularly those grown from 3. 1 is
also an excellent precursor for liquid-injection ALD of CeO2
using H2O as coreactant, demonstrating a window for self-
limiting growth between 150 and 350 °C. Both of these new
precursors have significant advantages when compared with
the widely used [Ce(thd)4]: higher growth rates in both
MOCVD (approximately double) and ALD (approximately
6 times higher), ability to use H2O instead of ozone as
coreactant in ALD, and much better solubility, eliminating
problems of injector blockages in the reactor.
(21) Evans, W. J.; Deming, T. J.; Olofson, J. M.; Ziller, J. W. Inorg.
Chem. 1989, 28, 4027–4034.
(22) Sirio, C.; HubertPfalzgraf, L. G.; Bois, C. Polyhedron 1997,
16, 1129–1136.
(23) Arnold, P. L.; Casely, I. J.; Zlatogorsky, S.; Wilson, C. Helv.
Chim. Acta 2009, 92, 2291–2303.
(24) Cameron, M. A.; George, S. M. Thin Solid Films 1999, 348, 90–98.
(25) Gradeff, P. S.; Schreiber, F. G.; Mauermann, H. J. Less-Common
Met. 1986, 126, 335–338.
’ ASSOCIATED CONTENT
S
Supporting Information. NMR spectra of complexes
b
and crystallographic data as a CIF file. This material is available
free of charge via the Internet at http://pubs.acs.org.
(26) Goel, S. C.; Kramer, K. S.; Chiang, M. Y.; Buhro, W. E.
Polyhedron 1990, 9, 611–613.
’ AUTHOR INFORMATION
(27) Anwander, R.; Munck, F. C.; Priermeier, T.; Scherer, W.;
Runte, O.; Herrmann, W. A. Inorg. Chem. 1997, 36, 3545–3552.
(28) Aspinall, H. C.; Bickley, J. F.; Gaskell, J. M.; Jones, A. C.; Labat,
G.; Chalker, P. R.; Williams, P. A. Inorg. Chem. 2007, 46, 5852–5860.
(29) Boyle, T. J.; Ottley, L. A. M. Chem. Rev. 2008, 108, 1896–1917.
(30) Cotton, F. A.; Marler, D. O.; Schwotzer, W. Inorg. Chim. Acta
1984, 85, L31–L32.
(31) King, P. J.; Werner, M.; Chalker, P. R.; Jones, A. C.; Aspinall,
H. C.; Basca, J.; Wrench, J. S.; Black, K.; Davies, H. O.; Heys, P. N. Thin
Solid Films 2011, 519, 4192–4195.
Corresponding Author
*E-mail: hca@liv.ac.uk.
’ ACKNOWLEDGMENT
We thank EPSRC for studentships to J.S.W. and P.J.K., and the
EPSRC National Mass Spectrometry Service Centre Swansea for
mass spectra.
’ REFERENCES
(1) Sharma, R. K.; Kumar, A.; Anthony, J. M. JOM-J. Minerals Met.
Mater. Soc. 2001, 53, 53–55.
(2) Chalker, P. R.; Werner, M.; Romani, S.; Potter, R. J.; Black, K.;
Aspinall, H. C.; Jones, A. C.; Zhao, C. Z.; Taylor, S.; Heys, P. N. Appl.
Phys. Lett. 2008, 93, 182911.
(3) Bi, L.; Kim, H. S.; Dionne, G. F.; Speakman, S. A.; Bono, D.;
Ross, C. A. J. Appl. Phys. 2008, 103, 07D138.
11652
dx.doi.org/10.1021/ic201593s |Inorg. Chem. 2011, 50, 11644–11652