Homoleptic zirconium amidates: Single source precursors for the aerosol-assisted chemical vapour deposition of ZrO2

Article abstract of DOI:10.1039/c6tc03631g

We report the development of a true single source precursor (i.e. without any need for an exogenous source of oxygen) for the growth of zirconia thin films by aerosol-assisted chemical vapour deposition (AACVD) using an original family of zirconium(iv) amidate derivatives, which are easily prepared by protonolysis of [Zr(NMe₂)₄] with the free amide pro-ligands. In all but one case the reactions resulted in the isolation of the corresponding homoleptic eight-coordinate zirconium(iv)tetrakis(amidato) derivatives. Three of these species along with a tris(amidato)dimethylamido zirconium(iv) derivative have been characterised by single crystal X-ray diffraction analysis. The materials potential of the homoleptic compounds was identified through the application of design criteria derived from consideration of the existing knowledge base relating to the pyrolysis of wholly organic amides. In this manner the thermal decomposition of the homoleptic derivatives benefits from facile, molecularly imposed pyrolysis pathways, which provide for the privileged generation of volatile small molecule by-products and the production of contaminant-free solid oxide material. Thermogravimetric analysis, in conjunction with NMR spectroscopic analysis of the volatile products resulting from their thermal decomposition, indicated the potential of the homoleptic species as exquisite single source precursors to ZrO₂ at moderate temperatures. The compound bearing both N- and C-iso-propyl substituents was, thus, applied as a true single source precursor under ambient pressure AACVD conditions. The resultant films, deposited on either SiO₂-coated glass or quartz substrates, are smooth and comprise small and densely packed crystalline particulates that are shown by XRD to be primarily cubic ZrO₂. Compositional analysis by X-ray photoelectron spectroscopy (XPS) revealed that the oxygen delivered, and the decomposition pathway provided, by the amidate ligand structure yields ZrO₂ films which, though slightly sub-stoichiometric (ZrO_1.8-1.9), contain undetectable levels of carbon incorporation.

Full text of DOI:10.1039/c6tc03631g

Journal of
Materials Chemistry C
View Article Online
View Journal | View Issue
PAPER
Homoleptic zirconium amidates: single source
precursors for the aerosol-assisted chemical
vapour deposition of ZrO2†
Cite this: J. Mater. Chem. C, 2016,
4, 10731
Amanda L. Catherall, Michael S. Hill,* Andrew L. Johnson,* Gabriele Kociok-Kohn
¨
and Mary F. Mahon
We report the development of a true single source precursor (i.e. without any need for an exogenous
source of oxygen) for the growth of zirconia thin films by aerosol-assisted chemical vapour deposition
(AACVD) using an original family of zirconium(IV) amidate derivatives, which are easily prepared by
protonolysis of [Zr(NMe2)4] with the free amide pro-ligands. In all but one case the reactions resulted in
the isolation of the corresponding homoleptic eight-coordinate zirconium(IV)tetrakis(amidato) derivatives.
Three of these species along with a tris(amidato)dimethylamido zirconium(IV) derivative have been
characterised by single crystal X-ray diffraction analysis. The materials potential of the homoleptic
compounds was identified through the application of design criteria derived from consideration of the
existing knowledge base relating to the pyrolysis of wholly organic amides. In this manner the thermal
decomposition of the homoleptic derivatives benefits from facile, molecularly imposed pyrolysis pathways,
which provide for the privileged generation of volatile small molecule by-products and the production of
contaminant-free solid oxide material. Thermogravimetric analysis, in conjunction with NMR spectroscopic
analysis of the volatile products resulting from their thermal decomposition, indicated the potential of
the homoleptic species as exquisite single source precursors to ZrO2 at moderate temperatures. The
compound bearing both N- and C-iso-propyl substituents was, thus, applied as a true single source
precursor under ambient pressure AACVD conditions. The resultant films, deposited on either SiO2-coated
glass or quartz substrates, are smooth and comprise small and densely packed crystalline particulates that
are shown by XRD to be primarily cubic ZrO2. Compositional analysis by X-ray photoelectron
spectroscopy (XPS) revealed that the oxygen delivered, and the decomposition pathway provided, by the
amidate ligand structure yields ZrO2 films which, though slightly sub-stoichiometric (ZrO1.8–1.9), contain
undetectable levels of carbon incorporation.
Received 22nd August 2016,
Accepted 31st October 2016
DOI: 10.1039/c6tc03631g
www.rsc.org/MaterialsC
provided by many of the reported precursors to zirconium
dioxide, which has attracted significant attention as a high
Introduction
The successful deployment of true single-source precursor reagents permittivity (k B 25) dielectric thin film material with low
for the metal organic chemical vapour deposition (MOCVD) leakage currents.1,2 Although the initial evolution of the simple
of metal oxide thin films often necessitates a compromise. b-diketonate species, [Zr(acac)4] (acac = pentane-2,4-dionate),3
For many systems the realisation of the volatility necessary for to more volatile 2,2,6,6-tetramethylheptane-3,5-dionato and the
vapour phase mass transport requires the introduction of fluorinated 1,1,1-trifluoropentane-2,4-dionato analogues facili-
peripheral organic substituents that reduce intermolecular tated enhanced oxide growth rates,4,5 these modifications came
association but which can act as sources of detrimental carbon at the expense of even greater levels of carbon and fluorine
contamination in the deposited films. A case in point is incorporation respectively. While a variety of alkoxide-,6 organo-
metallic-7 and, more recently, amidinate-based8 zirconium
Department of Chemistry, University of Bath, Claverton Down, Bath, BA2 7AY, UK.
E-mail: msh27@bath.ac.uk
systems have also been described, all such precursors require
the use of oxygen as co-reagent and typically high temperatures
(4600 1C) to achieve successful growth. These latter features
are particularly unattractive if deposition onto oxidatively- or
thermally-sensitive substrates is to be achieved. Although use of
† Electronic supplementary information (ESI) available: Experimental procedures,
instrumentation details and characterisation data for compounds 1–6, NMR spectra
relating to the thermal decomposition of compounds 2 and 5, UV-vis, XPS and micro-
XRD characterisation of the zirconia films. CCDC 1498668–1498671. For ESI and
crystallographic data in CIF or other electronic format see DOI: 10.1039/c6tc03631g the carbon-free complex [Zr(NO3)4] has allowed single-source
This journal is ©The Royal Society of Chemistry 2016
J. Mater. Chem. C, 2016, 4, 10731--10739 | 10731

View Article Online
Paper
Journal of Materials Chemistry C
and carbon-free growth at temperatures as low as 300 1C, the
films were found to contain 10–20% excess oxygen while the
potentially explosive nature of the precursor probably limits its
more widespread application.9 More general success has been
achieved by a liquid injection CVD (LICVD) technique utilising
a variety of heteroleptic alkoxides.10 In these cases, however, O2
is again required as a co-reagent resulting in oxygen-rich thin
film stoichiometries.10 Although a variety of volatile amide-
based systems for the atomic layer deposition (ALD) of zirconia
have also been described,11 this approach to thin film growth
necessarily requires a secondary oxygen source (typically water
or ozone) and is potentially limited by its requirements for
highly specialised equipment and high vacuum conditions.
Aerosol-assisted CVD (AACVD) is a modification of the
atmospheric pressure CVD technique that is gradually finding
more widespread application.12 In this case the requirement for
precursor volatility is circumvented through transport of the
chemical precursor to the reaction chamber as a solution in the
form of aerosol droplets using an inert carrier gas. Although
the only limitation on precursor implementation is, thus, its
solubility in an appropriate solvent, for a single source pre-
cursor system the requirement of a well-defined decomposition
pathway with limited competing deleterious processes becomes
even more rigorous. For the specific case of metal oxide AACVD,
the development of true single source precursors is an even
greater priority in order to obviate the hazards inherent in the use
of oxidising co-reagents in the presence of volatilised flammable
solvents. Kin et al. have described the sole use of the AACVD
technique for the deposition of zirconia films using ultrasonic
nebulisation of [Zr(OBu)4], [Zr(OBu)3(acac)], [Zr(OBu)2(acac)2], and
[Zr(OBu)(acac)3] as 0.1 M 1-butanol solutions.13 Although poly-
crystalline thin films were deposited at temperatures between
300 and 550 1C, no compositional data for the films were
provided and, taking account of the behaviour of zirconium
alkoxide and acetylacetonate derivatives under conventional
MOCVD conditions,3–6 it appears likely that these films also
suffered from significant levels of carbon contamination.
With these specific considerations in mind we have implemen-
ted a general programme of research to identify and synthesise
single-source precursors for the AACVD of metal oxides that will
completely fulfil the requirements of not only ultimate materials
integrity but also scalability and cost effectiveness. From this
perspective, metal amidates provide an as yet unexploited
opportunity for precursor design. Whereas it has been under-
Scheme 1 Proposed decomposition pathway for a metal amidate single
source oxide precursor under AACVD conditions.
protonolysis of other amidate ligands whereupon the estab-
lished nitrile elimination step will result in clean formation of
the metal oxide.14 It is notable that the only source of oxygen in
any resultant oxide material arises from the amidate ligands
and that this process should occur with the maintenance of the
metal’s initial oxidation state. Although the materials potential
of several metal amidate derivatives was highlighted some years
ago in the patent literature,16 the embodiments highlighted in
these claims were specifically derived from the volatility of the
metal amidate species and their consequent utility in conventional
direct delivery CVD or atomic layer deposition (ALD) processes.
Most notably, no claims were made for the utility of these species
as single source oxide precursors or as suitable reagents for any
type of solution-based materials deposition process.
In this study we target a series of zirconium amidate deriva-
tives as single source AACVD precursors to zirconium oxide.
Although our general interest in this material is derived from
its technological importance, more specifically, the diamagnetic
nature and stability of the Zr(IV) oxidation state lends itself ideally
to structural and mechanistic analysis. From this perspective
the ZrO2 deposition study described herein provides a suitably
challenging exemplar with which to begin to address the
broader potential of metal amidates as single source precursors
to metal oxides under mild AACVD conditions.
Results and discussion
stood for a century that wholly organic acid amides decompose The zirconium amidate derivatives, compounds 1–6, were
thermally by a selective process of nitrile and water elimination,14 synthesised in a manner similar to that employed by Schafer
no information is available for the products resulting from the and co-workers in their studies of the use of heteroleptic
pyrolysis of metal amidate derivatives. We speculated, therefore, zirconium amidates as pre-catalytic reagents for a variety of
that an archetypical multivalent metal amidate would thermally catalytic transformations.17 Addition of 4 molar equivalents of
decompose under AACVD conditions by the route shown in the relevant amide pro-ligand at 0 1C to a solution of tetra-
Scheme 1. The initial proposed alkene elimination step is kis(dimethylamido)zirconium afforded the homoleptic amidate
reminiscent of the Chugaev elimination established for isolobal complexes, 2–6, in excellent yields (480%) after heating at 60 1C
and isoelectronic metal xanthate-based sulphide precursor for 18 hours (Scheme 2). In contrast, and presumably for steric
chemistry.15 Although there are no directly relevant literature reasons, the corresponding reaction of [Zr(NMe2)4] with the
species to provide comparison, we postulate that the acidity of N-tert-butyl-2,2-dimethylpropanamide pro-ligand bearing both
the resultant metal hemiamilate will be sufficient to effect the N-t-Bu (R1) and C-t-Bu (R2) substitution was observed to be
10732 | J. Mater. Chem. C, 2016, 4, 10731--10739
This journal is ©The Royal Society of Chemistry 2016

View Article Online
Journal of Materials Chemistry C
Paper
Scheme 2 Synthesis of the homoleptic zirconium derivatives, compounds 2–6.
limited to protonolysis of only three dimethylamide residues to compound 1 displayed only a single signal associated with its
provide the heteroleptic species, tris(N-tert-butyl-2,2-dimethyl- sp2 amidato carbon centres in its 13C{1H} NMR spectrum. This
propanamidato)dimethylamido-zirconium, compound 1. All six signal was observed to resonate at d 165.3 ppm, significantly
compounds were readily isolated as colourless crystalline solids less deshielded than the comparable signals of the homoleptic
after removal of the reaction solvent.
compounds (2–6, ca. d 186–189 ppm).
Compounds 1–6 were characterised in d8-toluene solution by
The likely origin of these solution-based observations was
1H NMR and 13C NMR spectroscopy and by elemental analysis. resolved by X-ray diffraction analysis performed on single crystals
The 1H NMR spectra of compounds 2–6 confirmed the absence of of compounds 1, 2, 5 and 6 after crystallisation from toluene at
the dimethylamido singlet resonance at ca. d 3.10 ppm and clearly room temperature. The results of these experiments are illustrated
evidenced their constitution as homoleptic tetrakis(amidato) in Fig. 1 while details of the analyses and selected bond length and
species with, in each case, all four amidate ligands in equivalent angle data are presented in Tables 1 and 2 respectively.
environments. In contrast the 1H NMR spectrum of compound 1
Fig. 1(a) illustrates that compound 1 is a distorted octa-
displayed a resonance at d 3.06 ppm characteristic of the reten- hedral tris-amidato dimethylamido zirconium species. Two of
tion of a single dimethylamide ligand along with two sets of tert- the amidate ligands are bound in a chelating k2 fashion while
butyl methyl signals for each of the C- and N-bonded organic the third coordinates to the zirconium centre via a k1-binding
residues, which were observed in a 2 : 1 ratio by integration. mode through its basic oxygen centre, presumably to alleviate
Although a 1H EXSY NMR experiment demonstrated that these unfavourable steric congestion as a result of the multiplicity of
latter signals were not exchanging on the NMR timescale, bulky tertiary-butyl substituents. Although the Zr–O bond length
Fig. 1 ORTEP representations (25% probability ellipsoids) of (a) compound 1; (b) compound 2; (c) compound 5; (d) the Zr(1) containing molecule of
compound 6. Hydrogen atoms are removed for clarity.
This journal is ©The Royal Society of Chemistry 2016
J. Mater. Chem. C, 2016, 4, 10731--10739 | 10733

View Article Online
Paper
Journal of Materials Chemistry C
Table 1 Single crystal X-ray diffraction analysis of compounds 1, 2, 5 and 6
Compound
1
2
5
6
Empirical formula
Formula weight
Temperature/K
Crystal system
C29H60N4O3Zr
604.03
150.00(10)
Monoclinic
P21/n
C32H64N4O4Zr
660.09
150(2)
Monoclinic
C2
C32H64N4O4Zr
660.09
150(2)
Monoclinic
Cc
C28H56N4O4Zr
603.98
150.0(3)
Triclinic
%
P1
Space group
a/Å
b/Å
c/Å
a/1
b/1
g/1
10.8077(5)
17.3778(8)
18.1390(9)
90
90.090(4)
90
3406.7(3)
4
1.178
18.3482(7)
11.9920(5)
8.9002(4)
90
112.527(2)
90
1808.90(13)
2
1.212
20.1357(5)
15.0949(4)
12.2443(3)
90
98.836(2)
90
3677.44(16)
4
1.192
11.7212(6)
16.1788(9)
19.0915(10)
68.599(5)
79.594(4)
81.912(5)
3304.3(3)
4
U/Å3
Z
r
calc/g cmÀ3
1.214
0.367
m/mmÀ1
0.354
0.341
0.336
F(000)
1304.0
712
1424
1296.0
Crystal size/mm3
2y range for data collection/1
Index ranges
0.214 Â 0.197 Â 0.132
6.738 to 52.136
À13 r h r 13,
À21 r k r 20,
À22 r l r 22
25 934
0.250 Â 0.200 Â 0.150
3.398 to 27.090
À23 r h r 23,
À15 r k r 15,
À11 r l r 11
13 994
0.256 Â 0.222 Â 0.046
3.36 to 27.48
À26 r h r 26;
À19 r k r 19;
À15 r l r 10
6983
0.25 Â 0.171 Â 0.1
6.602 to 52.744
À12 r h r 14,
À18 r k r 20,
À23 r l r 22
28 277
Reflections collected
Independent reflections, Rint
Data/restraints/parameters
Goodness-of-fit on F2
6663, 0.0506
6663/195/385
1.026
0.0521, 0.1245
0.0710, 0.1360
1.41/À0.45
3951, 0.0518
3951/80/342
1.061
0.0318, 0.0631
0.0342, 0.0643
0.248/À0.701
6983, 0.0000 – hklf5 data
6983/2/391
1.024
0.0517, 0.1460
0.0649, 0.1553
1.570/À1.785
13 381, 0.0587
13 381/0/699
0.810
0.0464, 0.0701
0.0924, 0.0756
0.56/À0.51
Final R1, wR2[I Z 2s(I)]
Final R1, wR2[all data]
Largest diff. peak/hole/e ÅÀ3
Table 2 Selected bond lengths and angles for compounds 1, 2, 5 and 6
an alternative localised alkoxo-imino tautomer. Furthermore, we
suggest that the close correspondence of these measurements is,
thus, reflective of the facility for rapid exchange between k1- and
k2-binding modes highlighted in the solution NMR spectra of
this compound (vide supra).
In contrast to the heteroleptic constitution of compound 1
the tetrakis(amidato) derivatives, compounds 2, 5 and 6 adopt a
common 8-coordinate dodecahedral geometry in the solid state
1
2
5
6a
Zr(1)–N(1)
Zr(1)–N(2)
Zr(1)–N(3)
Zr(1)–N(4)
Zr(1)–O(1)
Zr(1)–O(2)
Zr(1)–O(3)
Zr(1)–O(4)
2.368(4)
2.296(3)
2.027(4)
2.307(11)
2.334(11)
2.298(8)
2.378(9)
2.221(12)
2.280(12)
2.089(3)
2.154(3)
2.315(5)
2.289(5)
2.301(5)
2.313(5)
2.194(4)
2.202(4)
2.200(3)
2.198(4)
2.280(3)
2.272(2)
2.302(3)
2.295(2)
2.207(2)
2.197(2)
2.194(2)
—
—
2.125(3)
2.164(3)
1.978(3)
2.1893(19) with pseudo-D2d symmetry (Fig. 1b–d). Although there are some
minor variations in the relevant Zr–O and Zr–N interactions and
O(1)–Zr(1)–O(2) 156.97(13)
77.87(13)
128.91(16) 125.92(8)
across the various C–O and C–N bonds within the amidate
ligands, these measurements are closely comparable for all three
compounds and lie within the ranges observed in the only
previous directly comparable homoleptic zirconium(IV) amidate,
the tetrakis(N-(2,6-dimethylphenyl)C-tert-butylamidate) zirconium
described by Schafer and co-workers.17b Although this previously
described species displayed a similar 8-coordinate geometry, it
was also noted to adopt a sterically disfavoured N-cis configuration
of the amidate ligands, which was ascribed to the additional
stability arising from p-stacking interactions of the N-ortho-xylyl
substituents. Although compounds 2, 5 and 6 lack the potential
for similar stabilising interactions, the coordination geometry
about zirconium in all three compounds is closely comparable
to this earlier structure and any minor variations may be simply
attributed to the differing steric profiles of the N- and C-bonded
amidate alkyl residues.17b
N(1)–Zr(1)–N(2)
O(1)–Zr(1)–N(1)
O(2)–Zr(1)–N(2)
O(1)–Zr(1)–N(3) 107.62(15)
O(2)–Zr(1)–N(3) 83.43(12)
N(1)–Zr(1)–O(3) 146.47(13)
N(2)–Zr(1)–O(3)
O(1)–Zr(1)–O(3)
O(2)–Zr(1)–O(3) 108.31(13) 126.5(5)
92.36(15) 167.03(16)
89.68(19)
57.69(15)
58.05(16)
134.34(16) 134.79(9)
85.42(17) 83.32(9)
133.86(17) 135.89(9)
84.63(16)
76.24(15)
127.91(16) 131.96(7)
91.28(17) 87.40(9)
166.96(18) 164.64(9)
91.31(9)
58.42(8)
58.67(8)
57.16(13)
57.92(10)
55.3(5)
58.9(4)
84.1(6)
82.8(6)
89.4(7)
84.4(6)
94.15(13)
88.67(8)
77.78(8)
89.59(13) 128.0(6)
N(1)–Zr(1)–N(4)
N(2)–Zr(1)–N(4)
—
—
86.5(8)
90.8(8)
a
Zr(1)-containing molecule only. The corresponding measurements for
the Zr(2)-containing molecule are closely comparable.
to the unique amidate [Zr(1)–O(3) 1.978(3) Å] is unsurprisingly
shorter than either of the Zr–O contacts of the bidentate ligands
[Zr(1)–O(1) 2.125(3); Zr(1)–O(2) 2.164(3) Å], it is notable that the
C–N bonds lengths across all three amidate moieties are closely
comparable [C(1)–N(1) 1.271(6); C(10)–N(2) 1.297(5); C(20)–N(4)
1.280(7) Å]. On this basis, we suggest that, despite this signifi-
Thermal analysis
cant adjustment in binding mode, the unidentate ligand is also Thermogravimetric analysis (TGA) was performed on com-
best represented as a delocalised amidate anion, rather than as pounds 1–6 (shown for compounds 2, 5 and 6 in Fig. 2).
10734 | J. Mater. Chem. C, 2016, 4, 10731--10739
This journal is ©The Royal Society of Chemistry 2016

View Article Online
Journal of Materials Chemistry C
Paper
The thermal decomposition of the heteroleptic species, com- the doublet at d 1.56 ppm were readily assigned as the respec-
pound 1, was essentially complete at temperatures less than tive methine, methylene and methyl resonances of propylene
400 1C, but provided a residual mass significantly in excess of while the heptet and doublet signals at d 1.75 and 0.63 ppm
that expected for ZrO2 (318 (1) versus 123.2 g molÀ1 for ZrO2). respectively confirmed the production of isobutyronitrile as a
Although the homoleptic amidate derivatives 2–6 displayed further volatile product of thermal decomposition. In accord
similar two stage thermal profiles, with an onset of decomposi- with the generalised decomposition pathway depicted in
tion ca. 150 1C, the differing N- and C-substitution patterns Scheme 1, the generation of these small molecule species was
provided significant variability in both the apparent mode of also accompanied by the evolution of significant quantities of
decomposition and the nature of the ultimate involatile resi- the protonated N-isopropylisobutyramide ligand precursor,
due. While the residual mass of compound 2 (106.74 g molÀ1
was indicative of some whole molecule volatility, the residues tities of intact compound 6 onto the cold walls of the heated
resulting from the decomposition of compounds 3 (167 g molÀ1
silica tube projecting from the tube furnace (see Fig. 3b).
)
which was observed to co-condense along with sublimed quan-
)
and 4 (199.22 g molÀ1) were much higher than that expected for Analogous decomposition of compounds 2 and 5 similarly
ZrO2 (123.2 g molÀ1). In contrast, the breakdown of compounds resulted in the evolution of the free amide precursors alongside
5 (124 g molÀ1) and 6 (129 g molÀ1) provided stable residual gaseous 2-methylpropene and isobutyronitrile (2) and propene
masses, which were effectively commensurate with the for- and pivalonitrile (5) respectively. On the basis of these empiri-
mation of zirconium dioxide. For both of these latter compounds cal observations and the TGA data described above we, thus,
decomposition was essentially complete at 350 1C and occurred propose that the thermolysis of compounds 2, 5 and 6 occurs by
as apparent two-step processes. While compound 5 provided an the common mechanism illustrated in schematic form as
initial mass loss event corresponding to ca. 7% mass between Scheme 1 and, for the specific case of compound 6, as shown
100 1C and 290 1C, a similar but more clearly defined separation in Scheme 3. Under this regime, the initial mass loss processes
of decomposition steps was observed for compound 6. In this highlighted in the thermogravimetric analyses may be rationa-
case an initial mass loss of ca. 16% had been observed by 200 1C lised as a consequence of b-hydride elimination and alkene
prior to more rapid weight loss across an approximate 25 1C generation from the N-alkyl residue of the relevant amidate
window, which we suggest represents the operation of several ligand. Zirconium oxide formation may then result from a
synchronous processes (vide infra).
sequence of amidate protonation and nitrile elimination steps
Further insight into the nature of these thermal processes reminiscent of the long established thermal behaviour of
was provided by a series of decomposition reactions performed wholly organic amides.14 The observed variations in thermal
on compounds 2, 5 and 6. Guided by the decomposition stability of the homoleptic zirconium amidate species are, thus,
temperatures indicated by the TGA data, samples (ca. 0.1 g) a reflection of the compounds’ relative facility toward N-alkyl
of each compound were heated (2, 5, 400 1C; 6 350 1C) under a b-hydride elimination and subsequent nitrile extrusion corres-
static vacuum in an apparatus which allowed the trapping of ponding to the C-alkyl amidate residue.
any volatile products into the d8-toluene solvent contained
Thin film deposition
within a J. Youngs NMR tube cooled to À78 1C. Subsequent
appraisal of the resultant solutions by 1H NMR spectroscopy Based on its sharper and slightly lower decomposition tempera-
highlighted a number of features common to all three decom- ture (Fig. 2) and its apparent rational decomposition pathway,
position reactions. The 1H NMR spectrum resulting from the compound 6 was identified as the most promising precursor to
gaseous by-products of the decomposition of compound 6 is carry forward for AACVD studies. Precursor solutions (0.05 M)
illustrated in Fig. 3a. The multiplets at d 5.71 and 4.98 ppm and were prepared by dissolving the chosen compound in toluene
and deposition was carried out under hot wall conditions onto
both silica-coated glass and quartz substrates in the temperature
range 350–600 1C for 60 minutes using nitrogen as a carrier gas.
The as-deposited films were optically clear but presented a very
pale brown colouration (see Fig. S20, ESI†). Post deposition
annealing by heating of the samples to 600 1C in air for a further
60 minutes, however, provided completely transparent and
colourless films which were well adhered to the substrate and
could not be removed by Scotch tape. The maximum thicknesses
of the annealed films deposited on quartz were determined with
ellipsometry using Complete Ease software (J.A Woollam Co.,
Inc.). These results (Fig. 4) highlight that the film thickness
increases with increasing temperature between 400 and 550 1C,
suggesting that film growth occurs under kinetic control
between these temperatures and with surface-based reactions
incorporating the rate determining step of the growth process.
Fig. 2 Thermogravimetric analysis of compounds 2, 5 and 6.
A resultant Arrhenius plot for this regime allowed an estimation
This journal is ©The Royal Society of Chemistry 2016
J. Mater. Chem. C, 2016, 4, 10731--10739 | 10735

View Article Online
Paper
Journal of Materials Chemistry C
Fig. 3 1H NMR spectra (d8-toluene, 298 K) of (a) the volatile products resulting from the in vacuo thermal decomposition of compound 6 highlighting
the formation of propene and isobutyronitrile; (b) the solid condensate products resulting from the thermal decomposition of compound 6, primarily a
mixture of sublimed compound 6 and the N-isopropylisobutyramide pro-ligand; (c) the N-isopropylisobutyramide pro-ligand; (d) compound 6.
samples aside from the film prepared at 400 1C displayed
maxima associated with the presence of the cubic phase of
zirconia (ICDD 04-016-6178). The sample prepared at 550 1C
also produced reflections consistent with the production of
monoclinic ZrO2 (Baddeleyite, ICDD 01-083-0943), while the
samples deposited at temperatures of 500 1C or lower displayed
an additional unidentified peak at a 2y value of 461. Notably
this latter peak, which we tentatively suggest is a consequence
of the incorporation of a metalorganic by-product of the
precursor decomposition, was observed to persist even after
post deposition annealing at 600 1C. The identification of the
baddeleyite phase, however, was further confirmed by analysis
Scheme 3 Postulated mechanism for the thermal decomposition of
compound 6.
of the film deposited at 550 1C by Raman spectroscopy
(Fig. S24, ESI†) while the films, irrespective of annealing, were
found to be non-conducting (four-point probe analysis). Optical
analysis of the annealed films deposited on quartz substrates
of the activation barrier toward this process as 42.60 Æ confirmed the high optical transparency of the films across all
10.23 kJ molÀ1. In contrast, at temperatures above 550 1C the visible wavelengths (Fig. S22, ESI†) and Tauc plots (Fig. S23,
film thickness is observed to decline in an effectively linear ESI†) derived from these data indicated band gaps of 5.7–5.8 eV
fashion suggesting that, at these higher temperatures, film consistent with the presence of ZrO2.
growth by AACVD is likely to be mass transport limited as
result of competitive gas phase reactions.
The morphology of the as-deposited films on SiO2-coated
glass was analysed by scanning electron microscopy (SEM) and
ZrO2 may exist in one of four basic crystalline polymorphs atomic force microscopy (AFM) (Fig. 5a and b). Both micro-
(monoclinic, orthorhombic, cubic, or tetragonal) or as a mix- scopic techniques highlighted a smooth and largely featureless
ture of polymorphs. For example, a mixture of monoclinic and coverage of the substrate by the films with a few small lumps
either orthorhombic or tetragonal phases is generally found in included across the surface. Cross sectional SEM images corro-
CVD ZrO2 thin films.8b Analysis of the non-annealed films borated the estimated film thicknesses illustrated in Fig. 4
deposited from compound 6 at 400, 450, 500, 550 and 600 1C and revealed that the films comprised of densely packed
by micro-X-ray diffraction (Micro-XRD), indicated that all the microscopic particulates with a sharp interface to the silica
10736 | J. Mater. Chem. C, 2016, 4, 10731--10739
This journal is ©The Royal Society of Chemistry 2016

View Article Online
Journal of Materials Chemistry C
Paper
Fig. 4 Film thickness (nm) determined by ellipsometry versus temperature (1C) for annealed zirconia thin films grown on quartz substrates by AACVD
from compound 6.
Fig. 5 (a) AFM image for film grown at 500 1C from complex 6; (b) cross section SEM image (801 specimen tilt, 100 000Â magnification) of film grown at
500 1C from complex 6; plan view SEM images of sample annealed at 600 1C; (c) 1300Â magnification; (d) 6000Â magnification.
substrate surface. Although the films were observed to remain the films, however, is notably consistent through the entire
well adhered to the substrate after annealing at 600 1C, micro- depth profile and indicates that AACVD utilising compound 6
scopic analysis revealed the formation of multidirectional as a single source precursor provides zirconia films which are
cracks which emanated from the larger particulates embedded remarkably uncontaminated by carbon within the detection
in the film (Fig. 5c and d). We suggest this latter observation is limits of the instrumentation (40.1 at%, Fig. S34, ESI†). In this
a consequence of differential thermal expansion between the regard, it is further notable that no elements other than
zirconia films and the glass substrate.
zirconium and oxygen could be detected throughout the bulk
Examination of the film composition was undertaken by of the coatings to the interface with the silica-coated glass
X-ray photoelectron spectroscopy (XPS) which was carried out on substrate. Comparison of the oxygen to zirconium ratios high-
both as-deposited films grown at 500 1C and those which were lighted that the annealed samples had a stoichiometry which
subsequently annealed in air at 600 1C. All the samples studied was effectively ZrO2, whereas the as-deposited samples were
provided very similar spectra and Fig. 6 illustrates typical XPS slightly sub-stoichiometric with average compositions between
depth profiles of (a) an as-deposited and (b) an annealed zirconia ZrO1.8 and ZrO1.9. Although the presence of lower oxidation
film. These data provided evidence for adventitious carbon state zirconium centres could not be discriminated in the XPS
contamination on the film surfaces. The bulk composition of data, the undetectable levels of carbon incorporation and this
This journal is ©The Royal Society of Chemistry 2016
J. Mater. Chem. C, 2016, 4, 10731--10739 | 10737

View Article Online
Paper
Journal of Materials Chemistry C
Acknowledgements
We thank the University of Bath for funding a PhD scholarship
(ALC) in support of the EPSRC-funded Supersolar Hub.
Pilkington-NSG are thanked for assistance with micro-XRD,
SEM and XPS measurements.
References
1 (a) A. C. Jones, H. C. Aspinall and P. R. Chalker, Chemical
Vapour Deposition of Metal Oxides for Microelectronics
Applications, in Chemical Vapour Deposition: Precursors, Processes
and Applications, ed. A. C. Jones and M. L. Hitchman, Royal
Society of Chemistry, 2009, 357; (b) A. Devi, Coord. Chem.
Rev., 2013, 257, 3332.
2 Y. Kim, J. Koo, J. W. Han, S. Choi, H. Jeon and C. G. Park,
J. Appl. Phys., 2002, 92, 5443.
3 (a) M. Pulver and G. Wahl, Electrochem. Soc., Proc., 1977,
97–25, 960; (b) R. C. Smith, T. Z. Ma, N. Hoilien, L. Y. Tsung,
M. J. Bevan, L. Colombo, J. Roberts, S. A. Campbell and
W. L. Gladfelter, Adv. Mater. Opt. Electron., 2000, 10, 105.
4 A. C. Jones, T. J. Leedham, P. J. Wright, M. J. Crosbie,
D. J. Wiliams, P. A. Lane and P. O’Brien, Mater. Res. Soc.
Symp. Proc., 1998, 495, 11.
5 (a) M. Balog, M. Schieber, M. Michman and S. Patai,
J. Cryst. Growth, 1972, 17, 298; see also: (b) S. V. Pasko,
L. G. Hubert-Pfalzgraf, A. Abrutis, P. Richard, A. Bartasyte
and V. Kazlauskiene, J. Mater. Chem., 2004, 14, 1245.
6 (a) A. Sartori, N. El Habra, C. De Zorzi, S. Sitran, M. Casarin,
G. Cavinato, C. Sada, R. Gerbasi and G. Rossetto, Chem. Vap.
Deposition, 2012, 18, 151; (b) J. S. Na, D. H. Kim, K. Yong and
S. W. Rhee, J. Electrochem. Soc., 2002, 149, C23; (c) J. J.
Gallegos, T. L. Ward, T. J. Boyle, M. A. Rodriguez and L. P.
Francisco, Chem. Vap. Deposition, 2000, 6, 21; (d) H. J. Frenck,
E. Oesterschulze, R. Beckmann, W. Kulisch and R. Kassing,
Mater. Sci. Eng., A, 1991, 139, 394; (e) B. J. Gould, I. M. Povey,
M. E. Pemble and W. R. Flavell, J. Mater. Chem., 1994, 4, 1815;
( f ) Y. Takahashi, T. Kawae and M. Nasu, J. Cryst. Growth,
1986, 74, 409.
7 (a) S. Codato, G. Carta, G. Rossetto, G. A. Rizzi, P. Zanella,
P. Scardi and M. Leoni, Chem. Vap. Deposition, 1999, 5, 159;
(b) J. M. Gaskell, A. C. Jones, K. Black, P. R. Chalker,
T. Leese, A. Kingsley, R. Odedra and P. N. Heyes, Surf. Coat.
Technol., 2007, 201, 9095.
8 (a) V.-S. Dang, M. Banerjee, H. Zhu, N. B. Srinivasan,
H. Parala, J. Pfetzing-Micklich, A. D. Wieck and A. Devi,
Chem. Vap. Deposition, 2014, 20, 320; (b) R. Thomas,
R. Bhakta, A. Milanov, A. Devi and P. Ehrhart, Chem. Vap.
Deposition, 2007, 13, 98.
Fig. 6 XPS depth profiles of (a) ZrO2 film deposited from compound 6 at
500 1C; (b) ZrO2 film deposited from compound 6 at 500 1C and
subsequently annealed at 600 1C in air.
slight sub-stoichiometry in oxygen, lead us to postulate that the
light colouration associated with the as-deposited films arises
from the presence of zirconium colour centres rather than any
contaminating impurity.
Conclusion
In summary, we present a series of homoleptic zirconium
amidate derivatives that display notably specific and predict-
able mechanisms of thermal decomposition at temperatures
lower than 400 1C. One of these derivatives, tetrakis(N-iso-
propyl-isobutyramidato)zirconium (6) has been assessed as a
single source precursor to ZrO2 thin films by an aerosol assisted
CVD process. Although the resultant films are crystalline for all
deposition temperatures above 400 1C, the films display a
minor sub-stoichiometry in oxygen. Compositional analysis,
however, within the detection limits of XPS, indicated that
the as-deposited films comprised zirconium and oxygen as
the sole constituent atoms. We suggest that his latter observa-
tion, together with the specificity of the decomposition path-
ways, demonstrate not only the potential of zirconium amidate
derivatives to act as true single source precursor reagents to
zirconium oxide but also that metal amidate species in general
hold much broader and, as yet, untapped potential as a class of
easily synthesised and readily manipulable oxide precursor
9 (a) R. C. Smith, T. Z. Ma, N. Hoilien, L. Y. Tsung, M. J. Bevan,
L. Colombo, J. Roberts, S. A. Campbell and W. L. Gladfelter,
Adv. Mater. Opt. Electron., 2000, 10, 105; (b) D. G. Colombo,
D. C. Gilmer, V. G. Young, S. A. Campbell and W. L. Gladfelter,
Chem. Vap. Deposition, 1998, 4, 220.
reagents. We are continuing to examine these possibilities 10 (a) A. C. Jones, T. J. Leedham, P. J. Wright, M. J. Crosbie,
and we will describe our observations in future publications.
D. J. Wiliams, K. A. Fleeting, H. O. Davies, D. J. Otway and
10738 | J. Mater. Chem. C, 2016, 4, 10731--10739
This journal is ©The Royal Society of Chemistry 2016

View Article Online
Journal of Materials Chemistry C
Paper
P. O’Brien, Chem. Vap. Deposition, 1998, 4, 197;
(b) P. A. Williams, J. L. Roberts, A. C. Jones, P. R. Chalker,
J. F. Bickley, A. Steiner, H. O. Davies and T. J. Leedham,
J. Mater. Chem., 2002, 12, 165; (c) P. A. Williams, J. L. Roberts,
A. C. Jones, P. R. Chalker, N. L. Tobin, J. F. Bickley,
H. O. Davies, L. M. Smith and T. J. Leedham, Chem. Vap.
Deposition, 2002, 8, 163; (d) Y. F. Loo, R. O’Kane, A. C. Jones,
H. C. Aspinall, R. J. Potter, P. R. Chalker, J. F. Bickley, S. Taylor
and L. M. Smith, Chem. Vap. Deposition, 2005, 11, 299;
(e) Y. F. Loo, R. O’Kane, A. C. Jones, H. C. Aspinall,
R. J. Potter, P. R. Chalker, J. F. Bickley, S. Taylor and
L. M. Smith, J. Mater. Chem., 2005, 15, 1896.
M. S. Hill, R. Hill, K. C. Molloy, H. Saponia, D. Stanton and
A. Sudlow, Dalton Trans., 2011, 40, 6893; (c) H. C. Leventis,
S. P. King, A. Sudlow, M. S. Hill, K. C. Molloy and S. A.
Haque, Nano Lett., 2010, 10, 1253; (d) A. J. MacLachlan,
F. T. F. O’Mahony, A. L. Sudlow, M. S. Hill, K. C. Molloy,
J. Nelson and S. A. Haque, ChemPhysChem, 2014, 15, 1019;
(e) J. Akhtar, M. Afzaal, M. A. Vincent, N. A. Burton,
I. H. Hillier and P. O’Brien, Chem. Commun., 2011,
47, 1991; ( f ) E. A. Lewis, P. D. McNaughter, Z. J. Yin,
Y. Q. Chen, J. R. Brent, S. A. Saah, J. Raftery,
J. A. M. Awudza, M. A. Malik, P. O’Brien and S. J. Haigh,
Chem. Mater., 2015, 27, 2127.
11 See: T. Blanquart, J. Niinisto, M. Ritala and M. Leskela, 16 (a) C. H. Winter, T. J. Knisley and P. K. D. M. C. Karunarathne,
Chem. Vap. Deposition, 2014, 20, 189, and references therein.
12 (a) X. Hou and K.-L. Choy, Chem. Vap. Deposition, 2006,
12, 583; (b) C. E. Knapp and C. J. Carmalt, Chem. Soc. Rev.,
2016, 45, 1036; (c) T. Wildsmith, M. S. Hill, A. L. Johnson,
Thermally stable volatile film precursors, WO 2010/132871 A1,
2010; (b) S. V. Ivanov, W. H. Bailey III, X. Lei and D. P. Spence,
Amidate precursors for depositing metal containing films, US
8642797 B2, 2014.
A. J. Kingsley and K. C. Molloy, Chem. Commun., 2013, 17 (a) C. Li, R. K. Thomson, B. Gillon, B. O. Patrick and
49, 8773; (d) I. Y. Ahmet, M. S. Hill, A. L. Johnson and
L. M. Peter, Chem. Mater., 2015, 27, 7680.
13 D.-Y. Kim, C.-H. Lee and S. J. Park, J. Mater. Res., 1996,
11, 2583.
14 (a) R. S. Borener and A. L. Ward, J. Am. Chem. Soc., 1916,
38, 2503; (b) R. S. Borener and A. L. Ward, J. Am. Chem. Soc.,
1916, 38, 2503.
L. L. Schafer, Chem. Commun., 2003, 2462; (b) P. R. Payne,
R. K. Thomson, D. M. Medeiros, G. Wan and L. L. Schafer,
Dalton Trans., 2013, 42, 15670; (c) R. K. Thomson,
F. E. Zahariev, Z. Zhang, B. O. Patrick, Y. A. Wang and
L. L. Schafer, Inorg. Chem., 2005, 44, 8680; (d) R. K. Thomson
and L. L. Schafer, Organometallics, 2010, 29, 3546;
(e) M. Haehnel, J. C.-H. Yim, L. L. Schafer and U. Rosenthal,
Angew. Chem., Int. Ed., 2013, 52, 11415; ( f ) R. K. Thomson,
F. E. Zahariev, Z. Zhang, B. O. Patrick, Y. A. Wang and
L. L. Schafer, Inorg. Chem., 2005, 44, 8680.
¨
15 See, for example: (a) N. Alam, M. S. Hill, G. Kociok-Kohn,
M. Zeller, M. Mazhar and K. C. Molloy, Chem. Mater., 2008,
¨
20, 6157; (b) J. M. Clark, G. Kociok-Kohn, N. J. Harnett,
This journal is ©The Royal Society of Chemistry 2016
J. Mater. Chem. C, 2016, 4, 10731--10739 | 10739