Syntheses, X-ray structures and CVD of titanium(iv) arsine complexes

Article abstract of DOI:10.1039/c001359e

Five titanium arsine compounds have been synthesised via the reaction of TiCl₄ with AsPh₃ (1:1 and 1:2 equivalents), or one equivalent of Ph₂AsCH₂AsPh₂, ^tBuAsH₂ and As(NMe₂)₃. In general, 1:1 and 1:2 adducts of the type [TiCl₄(L)_n] (n = 1, L = AsPh₃, Ph₂AsCH₂AsPh₂, and ^tBuAsH₂; n = 2, L = AsPh₃), were isolated and characterised. However, the reaction of TiCl₄ with As(NMe ₂)₃ resulted in a novel exchange between a Cl and an NMe₂ group, yielding the product [TiCl₃(NMe ₂)(μ-NMe₂)₂AsCl]. The crystal structure of [TiCl₃(NMe₂)(μ-NMe₂)₂AsCl] has been determined and showed that the titanium and arsenic atoms are linked via two bridging NMe₂ groups. Additionally, crystal structures for the 1:1 and 1:2 adducts, [TiCl₄(AsPh₃)] and [TiCl ₄(AsPh₃)₂] have been obtained, with Ti-As bond lengths of 2.7465(13) and 2.7238(7) A observed respectively. The decomposition of the compounds has been investigated using thermogravimetric analysis, aerosol-assisted and low pressure chemical vapour deposition.

Full text of DOI:10.1039/c001359e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Syntheses, X-ray structures and CVD of titanium(IV) arsine complexes†‡
Tegan Thomas, David Pugh, Ivan P. Parkin and Claire J. Carmalt*
Received 21st January 2010, Accepted 14th April 2010
First published as an Advance Article on the web 11th May 2010
DOI: 10.1039/c001359e
Five titanium arsine compounds have been synthesised via the reaction of TiCl4 with AsPh3 (1 : 1 and
1 : 2 equivalents), or one equivalent of Ph2AsCH2AsPh2, tBuAsH2 and As(NMe2)3. In general, 1 : 1 and
1 : 2 adducts of the type [TiCl4(L)n] (n = 1, L = AsPh3, Ph2AsCH2AsPh2, and tBuAsH2; n = 2, L =
AsPh3), were isolated and characterised. However, the reaction of TiCl4 with As(NMe2)3 resulted in a
novel exchange between a Cl and an NMe2 group, yielding the product [TiCl3(NMe2)(m-NMe2)2AsCl].
The crystal structure of [TiCl3(NMe2)(m-NMe2)2AsCl] has been determined and showed that the
titanium and arsenic atoms are linked via two bridging NMe2 groups. Additionally, crystal structures
for the 1 : 1 and 1 : 2 adducts, [TiCl4(AsPh3)] and [TiCl4(AsPh3)2] have been obtained, with Ti–As bond
˚
lengths of 2.7465(13) and 2.7238(7) A observed respectively. The decomposition of the compounds has
been investigated using thermogravimetric analysis, aerosol-assisted and low pressure chemical vapour
deposition.
and [Co(CO)4{AsN(tBu)CH2CH2N(tBu)}].10 Therefore, we were
Introduction
interested in developing simple arsine adducts of titanium(IV)
for potential application as single-source precursors to titanium
arsenide films. Surprisingly, little structural data are available for
complexes of this type. We have been studying a range of group
15 containing metal complexes and assessing their ability to form
metal pnictide thin films via CVD including TaN and NbN,11,12
TiN and TiNxCy,13–15 and TiP.8,9 Here we report the synthesis and
characterisation of five titanium arsine adducts synthesised from
the reaction of TiCl4 with nAsPh3 (n = 1, 2), Ph2AsCH2AsPh2,
tBuAsH2 and As(NMe2)3. Additionally, the use of some of the
complexes in CVD experiments has been explored.
Metal arsenides are known to exhibit a wide range of functional
properties, and display band gap characteristics which are cur-
rently exploited within a range of photonic devices including
lasers and light emitting diodes (LEDs). Apart from key III/V
materials (e.g. GaAs, InAs), CoAs,1 and MnAs,2,3 knowledge
surrounding metal arsenide thin films, particularly via chemical
vapour deposition (CVD), remains limited; typically with only
bulk material formation known. It is well understood that
arsenide compounds are considerably less stable and volatile
than their phosphorus counterparts,4 which has until recently
limited research into arsenic containing thin films. However, recent
advances within CVD technology and the development of single-
source precursors, has now facilitated research into this area.
Single-source precursors circumvent problems associated with
precursor volatility when used with aerosol assisted (AA)CVD,
and additionally enable high levels of stoichiometric control due to
the existence of pre-formed bonds.5 Single-source precursors have
already proven successful in the deposition of III/V materials,6
along with transition metal phosphide thin films. CrP films have
been deposited via single-source pyrolysis routes (for instance
that of [Cr(CO)5(PH3)] to produce CrP films),7 and additionally
within CVD where examples include the low pressure (LP)CVD of
[Co(CO)2{P(tBu)2H}(NO)]1 and [TiCl4(L)n] (n = 2, L = CyhexPH2;
n = 1, L = dppm)8,9 to yield CoP and TiP films respectively.
Single-source precursors available for the deposition of tran-
sition metal arsenide thin films are limited, with CoAs suc-
cessfully being deposited via (MO)CVD of [Co2(CO)6(m-As2)],1
Experimental
General procedures and instrumentation
All manipulations were performed under a dry, oxygen-free
dinitrogen atmosphere using standard Schlenk techniques or in
an Mbraun Unilab glovebox. All solvents used were stored in
alumina columns and dried such that the water concentration was
5–10 ppm. Elemental analysis was conducted using an EA-440
horizontal load analyser supplied by Exeter analytical. 1H NMR
was recorded using a Bru¨ker AMX400 spectrometer. All spectra
were recorded using C6D6, which was dried and degassed over
molecular sieves prior to use; 1H chemical shifts are reported
relative to SiMe4 (d 0.00). Mass spectra were recorded using
a Micromass ZABSE instrument. Thermogravimetric analysis
(TGA) was conducted using a Netzsch STA 449C instrument,
with samples sealed using aluminium sample pans. Due to the
air sensitive nature of the samples, TGA was conducted under
Materials Chemistry Centre, Department of Chemistry, University Col-
lege London, Gordon St., London, WC1H OAJ, England. E-mail:
c.j.carmalt@ucl.ac.uk; Tel: +44(0)2076797528
† Dedicated to Professor David W. H. Rankin on the occasion of his
retirement.
◦
a flow of helium gas; heating rates of 10 C min-1 were used
in all instances. The reagents AsPh3 and Ph2AsCH2AsPh2 were
t
supplied by Sigma Aldrich, and As(NMe2)3 and BuAsH2 by
SAFC Hitech Ltd.; with all reagents being used without further
purification.
‡ CCDC reference numbers 762189–762191. For crystallographic data in
CIF or other electronic format see DOI: 10.1039/c001359e
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 5325–5331 | 5325
©

View Article Online
Synthesis of [TiCl4(AsPh3)] (1)
deep red colour intensified. Once the reaction mixture had reached
room temperature, it was cooled to -70 ◦C for approximately
four weeks, after which a dark brown solid was observed. Excess
reagents and solvents were carefully removed under vacuum,
ensuring maximum product extraction was obtained; a small
A clear colourless solution of AsPh3 (0.83 g, 2.7 mmol) in toluene
(20 cm3) was added with stirring to a 1 M orange-red solution
of TiCl4 in toluene (2.7 cm3, 2.7 mmol). The solution turned
dark red immediately on the addition of the AsPh3 solution. The
reaction mixture was allowed to stir for approximately 30 min,
prior to solvent removal under vacuum to yield a purple-pink
solid (1) (yield 85%, mpt 126 ◦C). Analysis found (calc. for
C18H15AsCl4Ti): C 41.78% (43.60); H 3.24% (3.05). 1H NMR
d (ppm) (C6D6) 7.58 (m, 6H, m-C6H5), 7.13 (m, 9H, o-C6H5
and p-C6H5). Mass Spec. m/z (Cl+, methane): 152 [AsPh]H+,
229 [AsPh2]H+, 306 [AsPh3]H+, 330, 353 [TiAsPh3]H+, 360, 424
[TiCl2AsPh3]H+, 430, 458 [AsPh2][AsPh2]H+, 493 [TiCl4AsPh3]H+,
535 [AsPh2][AsPh3]H+, 612 [AsPh3][AsPh3]H+, 687. Mp 126 ◦C.
Crystals of 1 were isolated via dichloromethane–hexane layering
recrystallisation and characterised by X-ray crystallography.
1
quantity of brown solid remained (4) (yield 16%). H NMR d
(ppm) (C6D6) 3.0 (broad s, 2H, AsH2), 1.3 (s, 9H, tBuAs).
Synthesis of [TiCl3(NMe2)(l-NMe2)2AsCl] (5)
A clear colourless solution of As(NMe2)3 (0.5 cm3, 0.62 g,
3 mmol) in toluene (10 cm3) was added via cannula to a red-
orange 1M solution of TiCl4 in toluene (2.7 cm3, 2.7 mmol).
The solution turned dark green immediately on addition of the
As(NMe2)3 solution. After the addition, the reaction mixture was
allowed to stir for approximately 30 min prior to the removal of
solvent under vacuum to yield a dark-green solid (5) (yield 98%,
mpt 95–97 ◦C). Analysis found (calc. for C6H18AsCl4N3Ti): C
1
18.50% (18.16); H 4.81% (4.57); N 10.70% (10.59). H NMR d
Synthesis of [TiCl4(AsPh3)2] (2)
(ppm) (C6D6) 3.7 (s, 6H, terminal NMe2), 2.7 (s, 12H, m-NMe2).
Mass Spec. m/z (Cl+, methane): 84 [TiCl]H+, 105 [AsN2]H+,
110 [TiClNC]H+, 120 [AsNMe2]H+, 136 [AsNNMe2]H+, 145
[TiCl2NC]H+, 154 [TiCl3]H+, 163 [TiCl2NMe2]H+, 170, 190
[TiCl4]H+, 199 [TiCl3NMe2]H+, 208 [As(NMe2)3]H+, 216, 227,
254, 264, 300, 352, 365, 371, 406. Mp 95 C. Crystals of 5 were
isolated via dichloromethane–hexane layering recrystallisation
and characterised by X-ray crystallography.
A 1M orange-red solution of TiCl4 in toluene (1.3 cm3, 1.3 mmol)
was added to a colourless solution of AsPh3 (0.8 g, 2.61 mmol)
in toluene (20 cm3) with stirring. The solution turned dark red
immediately on addition of the TiCl4 solution. The reaction
mixture was allowed to stir for approximately 30 min, prior
to refluxing under nitrogen for 24 h. After cooling to room
temperature, the solvent was removed under vacuum to yield
a purple-pink solid (2) (yield 85%, mpt 88–90 ◦C). Analysis
found (calc. for C36H30As2Cl4Ti): C 50.60% (53.90); H 3.78%
◦
Crystallography
1
(3.77). H NMR d (ppm) (C6D6) 7.43 (m, 12H, m-C6H5), 7.02
In all instances, single crystals were obtained via dichloromethane–
hexane layering recrystallisation. A summary of the crystal data,
data collection and refinement for compounds 1, 2 and 5 is given
in Table 2. Crystals were mounted on a glass fibre with silicon
(m, 18H, o-C6H5 and p-C6H5). Mass Spec m/z (Cl+, methane):
152 [AsPh]H+, 229 [AsPh2]H+, 306 [AsPh3]H+, 335, 458, 535.
Crystals of 2 were isolated via dichloromethane–hexane layering
recrystallisation and characterised by X-ray crystallography.
R
grease from Fomblinꢀ vacuum oil. Datasets were collected on
a Bru¨ker SMART APEX CCD diffractometer using graphite-
Reaction of TiCl4 and Ph2AsCH2AsPh2 (3)
˚
monochromated Mo-Ka radiation (l1 = 0.71073 A) at 150(2) K.
Data reduction and integration was carried out with SAINT+,16
and absorption corrections applied using SADABS.17 All solutions
and refinements were performed using PLATON,18 the WinGX
package and all software packages within.19 All non-hydrogen
atoms were refined using anisotropic thermal parameters and
hydrogens were added using a riding model. The AsPh3 group
in 2 was disordered over two positions but the occupancy
factors were not fixed during refinement cycles. CCDC reference
numbers 762189 (compound 1), 762190 (compound 2) and 762191
(compound 5).
A pale-yellow solution of Ph2AsCH2AsPh2 (1.28 g, 2.7 mmol)
in toluene (20 cm3) was added with stirring to a red-orange
1M solution of TiCl4 in toluene (2.7 cm3, 2.7 mmol). A dark
red-orange suspension formed immediately on addition of the
Ph2AsCH2AsPh2 solution. The reaction mixture was allowed to
stir for approximately 30 min, prior to the removal of solvent
under vacuum to yield a bright orange solid (3) (yield 86.2%, mpt
116–118 ◦C). Analysis found (calc. for C25H22As2Cl4Ti): C 45.08%
(45.36); H 3.30% (3.35). Mass Spec. m/z/(Cl+, methane): 105,
155, 167 [AsPhCH2]H+, 191 [TiCl4]H+, 229 [AsPh2]H+, 245, 260,
306 [AsPh3]H+, 313 [TiClAsPh2]H+, 330, 395 [Ph2AsCH2AsPh]H+,
472 [Ph2AsCH2AsPh2]H+, 501 [TiCl2As2Ph3]H+, 563, 639, 701,
869. Compound 3 was found to be insoluble in a variety of solvents
(e.g. dichloromethane, hexane, and toluene) and no satisfactory 1H
NMR could be obtained.
Aerosol-assisted (AA)CVD
All AACVD depositions were conducted on 90 mm ¥ 45 mm ¥
4 mm float-glass coated with a 50 nm thick SiCO barrier layer to
stop diffusion of ions from the glass, as supplied by Pilkington.
All substrate was cleaned using petroleum ether (60–80 ◦C) and 2-
propanol and allowed to air dry prior to use. All depositions were
conducted using a horizontal cold wall reactor, with substrates
mounted onto a graphite heating block containing a Whatman
cartridge heater, with temperature control achieved using a Pt–Rh
thermocouple. A Vicks VE5520E Humidifier was used to produce
the aerosol mist. Prior to the conduction of AACVD, equipment
Reaction of TiCl4 and tBuAsH2 (4)
tBuAsH2 (0.36 cm3, 0.36 g, 2.69 mmol) was added to an orange-
red 1M solution of TiCl4 in toluene (2.7 cm3, 2.7 mmol) which was
cooled to -78 ◦C. The solution turned bright red immediately on
t
addition of the BuAsH2 solution. The reaction mixture was al-
lowed to stir while warming to room temperature, during which the
5326 | Dalton Trans., 2010, 39, 5325–5331
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
◦
i
˚
Table 1 Selected bond lengths (A) and angles ( ) for compounds 1, 2 and 5; where = -x + 1,-y + 2,-z + 2
Selected Angles (◦)
˚
Compound
Selected Bond lengths (A)
(1)
Ti(1)–As(1) 2.7465(13)
Ti(1)–Cl(1) 2.244(2)
Ti(1)–Cl(2) 2.185(2)
Ti(1)–Cl(3) 2.220(2)
Ti(1)–Cl(4) 2.193(2)
Cl(1)–Ti(1)–As(1) 175.41(8)
Cl(2)–Ti(1)–As(1) 83.95(6)
Cl(3)–Ti(1)–As(1) 81.02(6)
Cl(4)–Ti(1)–As(1) 82.86(6)
(2)
(5)
Ti(1)–As(1) 2.7238(7)
Ti(1)–Cl(2) 2.2719(11)
Ti(1)–Cl(1) 2.2713(11)
Cl(1)–Ti(1)–As(1) 86.68(4)
Cl(1)–Ti(1)–As(1i) 93.32(4)
As(1)–Ti(1)–As(1i) 180.0
Cl(2)–Ti(1)–As(1) 90.36(4)
Cl(2)–Ti(1)–As(1i) 89.64(4)
N(1)–Ti(1) 2.323(2)
N(2)–Ti(1) 2.417(2)
N(3)–Ti(1) 1.864(2)
Cl(1)–Ti(1) 2.3110(8)
Cl(2)–Ti(1) 2.3814(8)
Cl(3)–Ti(1) 2.3004(8)
Cl(4)–As(1) 2.2075(7)
As(1)–N(1)–Ti(1) 93.89(8)
N(1)–Ti(1)–N(2) 69.62(7)
N(2)–As(1)–N(1) 89.60(9)
As(1)–N(2)–Ti(1) 91.90(8)
N(3)–Ti(1)–N(2) 165.86(9)
N(2)–As(1)–Cl(4) 99.97(7)
Table 2 Crystallographic data for compounds 1, 2 and 5
Film characterisation
X-Ray powder diffraction patterns were obtained using a Bru¨ker
AXS D8 discover machine using monochromatic Cu-Ka radi-
ation. Wavelength dispersive X-ray analysis (WDX) was con-
ducted using a Philips XL30ESEM machine. Scanning electron
microscopy (SEM) was performed using a JSM-6301F scanning
field emission machine.
Compound
(1)
(2)
(5)
Chemical formula
Formula weight
Crystal system
Space group
C18H15AsCl4Ti C36H30As2Cl4Ti C6H18AsCl4N3Ti
495.92
Monoclinic
P21/c
9.5026(19)
11.324(2)
18.683(4)
90
92.501(4)
90
2008.5(7)
4
802.14
Triclinic
¯
P1
396.85
Monoclinic
P21/n
9.5866(16)
15.623(3)
9.8386(16)
90
92.562(3)
90
1472.1(4)
4
1.791
150(2)
3.514
0.71073
12201
3496
˚
a/A
9.686(3)
9.789(3)
10.237(3)
108.824(5)
111.489(5)
92.081(5)
841.7(5)
1
1.582
150(2)
2.548
0.71073
6969
3757
˚
b/A
˚
c/A
a (◦)
b (◦)
g (◦)
Results and discussion
Synthesis and characterisation
3
˚
V/A
Z
Treatment of TiCl4 with one or two equivalents of an arsine, L,
resulted in the immediate formation of a dark-red colour and
after work-up isolation of [TiCl4(L)n] (L = AsPh3, n = 1 (1),
n = 2 (2); L = Ph2AsCH2AsPh2, n = 1 (3)) in high yield
(> 85%). Complexes 1 and 2 have been characterised by 1H
NMR spectroscopy, the results of which indicate that a monomeric
complex [TiCl4(L)n] has been isolated in all reactions (Scheme 1).
Dc/Mg m-3
T/K
1.640
150(2)
2.591
m/mm-1
˚
l/A
0.71073
No. of data collected 16527
No. of unique data 4787
Goodness of fit on F2 1.182
1.053
0.028
0.0461
1.126
0.0258
0.0322
Rint
0.0609
0.0809
1
Final R(|F|) for
F02s(F0)
Unfortunately, H NMR of 3 could not be obtained due to the
insolubility of this compound. However, analytical data were
obtained for compounds 1–3. The carbon and hydrogen analyses
for 3 were satisfactory. However, for compounds 1 and 2, although
Final R(F2) for all
data
0.1496
0.1188
0.0672
and substrate were allowed to heat to the required temperature
under a flow of nitrogen. The compounds were either used as
synthesised or created in situ, with delivery to substrate via N2
carrier gas (1 L min-1). After conduction of the AACVD, the
substrates were allowed to cool under a flow of nitrogen, prior to
their removal and subsequent storage in air.
Low pressure (LP)CVD
All LPCVD depositions were conducted on vertically halved
glass microscope slides with approximate dimensions 1.2 cm ¥
7 cm, which were cleaned using petroleum ether (60–80 ◦C)
and 2-propanol and allowed to air dry prior to use. LPCVD
was conducted within a tube furnace, with approximately 0.5 g
of compound used in all instances. The substrates were placed
horizontally up the length of the tube furnace, with LPCVD
conducted at 600 ◦C under vacuum.
Scheme 1 Schematic illustrating the synthesis of compounds 1–5 via the
reaction of TiCl4 with: (i) AsPh3, (ii) 2AsPh3, (iii) Ph2AsCH2AsPh2, (iv)
tBuAsH2 and (v) As(NMe2)3.
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 5325–5331 | 5327
©

View Article Online
the hydrogen analysis was satisfactory, the carbon analysis was low.
This has been observed previously with some related phosphine
adducts of TiCl4, which was thought to be due to the formation of
either titanium phosphide with carbon impurities or metal carbide
during the thermal decomposition stage of the microanalysis
procedure.9 Therefore, it is possible that the % C for 1 and 2 are
low due to the formation of metal carbide. In order to investigate
whether the compounds are monomeric and if a cis or trans
isomer had formed (for compound 2 when n = 2, L = AsPh3),
the structures of compounds 1 and 2 have been determined.
The crystal structure of 1 was determined by single crystal
X-ray diffraction and the results are shown in Fig. 1; selected bond
lengths and angles are given in Table 1. Compound 1 crystallised
into the monoclinic space group P21/c. It is monomeric in the
solid state with the titanium atom adopting a distorted trigonal
bipyramidal geometry. Three Cl atoms occupy the equatorial
positions while the As atom of the AsPh3 group and one Cl atom
trans-(Ph3As)2TiCl4 geometry, as previously observed for both
the SnCl4 and OsBr4 triphenylarsine adducts.20,21 A related
diphenylphosphine adduct of TiCl4 was also shown to crystallise as
9
˚
the trans isomer. An average Ti–As bond distance of 2.7238(7) A
was observed in compound 2, which is similar to that found in
1. Compounds 1 and 2 both demonstrate Ti–As bond lengths
similar to that previously observed for the titanium bromide
chelating arsine complex [TiBr4{MeC(CH2AsMe2)3}] (approxi-
˚
mately 2.68 A), with the Ti–Cl bond lengths in 2 (approximately
˚
2.27 A) found to be in agreement with that previously reported for
the six coordinate titanium chloride chelating phosphine complex
[TiCl4{o-C6H4(PMe2)2}] (2.28 A approximately).22 All other bond
˚
distances and angles are similar to related complexes and the
octahedral coordination geometry around the titanium metal
atom found in 2 was expected.
Titanium tetrachloride readily forms adducts of the type
[TiCl4(L)n], with donor ligands,23 and compounds 1 and 2 have
been previously described briefly in the literature;24 however, to the
best of our knowledge their structures have not been reported. The
coordination numbers of [TiCl4(L)n] adducts are generally 5 or 6
but in a few cases 8 coordination has been isolated.25 For example,
the reaction of titanium(IV) halides with chelating phosphine
and arsine ligands afforded both 6- and 8-coordinate complexes.
In general, 6-coordinate species were obtained when sterically
demanding ligands, such as Ph2PCH2PPh2 and o-C6H4(PPh3)2,
were used, whereas 8-coordinate complexes were isolated with
phosphines and arsines with relatively small cone angles at the
metal, e.g. o-C2H4(PMe2)2 and o-C6H4(AsMe2)2.25,26 A search of
the Cambridge structural data base revealed that compounds 1
and 2 are the first crystallographically characterised [TiCl4(AsRx)n]
structures without a chelating arsine ligand.
˚
reside in the axial positions. A Ti–As bond length of 2.7465(13) A
was observed, which is approximately equal to the sum of the
˚
independent atomic radii (1.47 and 1.25 A) (Fig. 1) indicating that
the donor ligand is bound relatively weakly.
Unfortunately, X-ray quality crystals of [TiCl4(Ph2-
AsCH2AsPh2)] (3), could not be obtained. It is assumed
that due to the chelating nature of the ligand in 3 that a
6-coordinate geometry would be adopted, similar to that observed
in the analogous phosphine complex [TiCl4(Ph2PCH2PPh2)].22
It was not possible to elucidate the geometry of 3 using IR
spectroscopy due to the highly air sensitive nature of the complex
resulting in partial decomposition during the analysis, or analyse
the compound by 1H NMR due to its insolubility.
Fig. 1 ORTEP representation of [TiCl4AsPh3] (1) with thermal ellipsoids
at the 50% probability level. Hydrogen atoms are omitted for clarity.
The structure of compound 2 was determined by single crystal
X-ray crystallography, the results are shown in Fig. 2; selected
bond lengths and angles are given in Table 1. Compound 2 lies
on an inversion centre and crystallised into the triclinic space
Compounds 1–3 involve the use of tertiary arsines and a
chelating arsine. We were particularly interested in investigating
what effect changing the arsine ligand could have on the structure
of the product obtained. Furthermore, changing the arsine
could affect any film resulting from the use of these complexes,
particularly since the presence of the phenyl group in 1–3 could
lead to carbon contamination. We therefore decided to investigate
the reaction of TiCl4 with a primary arsine, which could result
in a better decomposition route and potentially less carbon
¯
group P1. Compound 2 adopts a centrosymmetric, octahedral
t
incorporation within the film. The primary arsine, BuAsH2,27
has been employed successfully both in dual- and single-source
routes to GaAs and therefore we investigated the use of this
ligand. Treatment of TiCl4 with BuAsH2 in toluene at -78 ◦C
t
resulted in the immediate formation of a dark-red solution. The
resulting complex is extremely volatile and attempts to remove
the solvent in vacuo resulted in the loss of the compound to the
vacuum trap. In order to overcome this problem the solution was
kept at -70 ◦C and after approximately four weeks a dark brown
solid was isolated in low yield. The 1H NMR of the solid suggests
Fig. 2 ORTEP representation of [TiCl4(AsPh3)2] (2) showing one of two
orientations of the disordered AsPh3 group. Thermal ellipsoids at the 50%
probability level and hydrogen atoms are omitted for clarity.
5328 | Dalton Trans., 2010, 39, 5325–5331
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
an adduct, [TiCl4(tBuAsH2)n] (4) has been formed. Thus, peaks
were observed at 1.3 and 3.0 ppm in a 9 : 2 ratio corresponding to
the tBu and AsH2 protons respectively. However, the 1H NMR of
4 clearly shows the formation of an adduct with an upfield shift
in the peaks for the arsine ligand in 4, in comparison to the free
arsine. Unfortunately, the high volatility and air sensitive nature of
compound 4 meant that further sample manipulation was difficult.
Tris(dimethylamino)arsine has also been successfully employed
in the formation of GaAs.27–29 We were therefore interested in
attempting to form a precursor to TiAs using this arsine. In
contrast to the synthesis of 1–4, reaction of one equivalent of
TiCl4 with As(NMe2)3, resulted in the immediate formation of
a dark-green solution. After work-up a dark-green solid (5)
was isolated in high yield (98%), the analytical data for which
suggested the formation of a 1 : 1 adduct, [TiCl4{As(NMe2)3}].
However, the 1H NMR of 5 showed two peaks at 2.7 and 3.7 ppm,
in a 2 : 1 ratio, suggesting the presence of both bridging and
terminal NMe2 groups. In order to fully investigate the nature
of 5 the crystal structure was determined by single crystal X-ray
diffraction, which showed the formation of the novel complex
[TiCl3(NMe2)(m-NMe2)2AsCl] rather than the simple 1 : 1 adduct,
[TiCl4{As(NMe2)3}]. The results are shown in Fig. 3; selected bond
lengths and angles are given in Table 1. Compound 5 crystallised
into the monoclinic space group P21/n. The structure of 5 (Fig. 3),
shows it is monomeric with a 6-coordinate titanium centre which
possesses three terminal Cl ligands and one terminal and two
bridging NMe2 groups in a distorted octahedral geometry. Thus,
two of the NMe2 groups are bridging to the arsenic atom, which
also has a chlorine atom attached. Upon comparing the Ti–N
bond lengths of 5 with that of a similar complex, the terminal and
TGA of compounds (1)–(5)
In order to study the decomposition pathway of compounds 1–5
and to assess their suitability as CVD precursors, thermogravimet-
ric analysis (TGA)/differential scanning calorimetry (DSC) was
carried out (Fig. 4). The TGA/DSC of 1 resulted in a two-step
weight loss profile, with losses occurring at approximately 120 and
220 ◦C. Weight loss completion had occurred by 320 ◦C, where a
residual mass of 11% was reported. The determined residual mass
was not in agreement with that expected for TiAs (25%), however,
it should be noted that sublimation of the precursor could occur
which would result in a higher weight loss than expected. Similarly,
TGA of 2 resulted in a two-step weight ◦loss profile, with losses
occurring at approximately 130 and 360 C (residual mass for 2
6%). Overall the TGA of◦1 and 2 indicates that weight loss is
complete by 320 and 240 C respectively, starting at around the
melting point (120 ◦C). These characteristics indicate the potential
of 1 and 2 as CVD precursors, in LPCVD or AACVD. However, it
is possible that the weight loss could correspond to the dissociation
of the arsine ligand before decomposition due to the weak Ti–As
bonds present in the complexes.
˚
bridging NMe2 Ti–N bond lengths (1.864(2) A and approximately
˚
2.3 A respectively) are found to be consistent with the Ti–N
bond lengths observed for the complex [TiCl2(NMe2)2(HNMe2)]
30
˚
˚
(approximately 1.85 A (C2H6N) and 2.25 A (C2H7N)). The
formation of 5 is the result of an interesting substituent exchange
between one of the dimethylamido groups on As(NMe2)3, and
a chlorine atom from TiCl4 to form the adduct [TiCl3(NMe2)(m-
NMe2)2AsCl], in which the nitrogen atoms of two NMe2 groups
coordinate to the titanium centre in preference to coordination via
the arsenic atom. This is the first time that this type of exchange
reaction has been observed for As. The reason behind the exchange
is not clear, however similar observations have been reported
during solvolysis reactions between TiCl4 and aliphatic primary
and secondary amines.31
Fig. 4 TGA plots for compounds 1–5.
TGA of 3 resulted in a two-step decomposition p◦rofile, with
weight losses occurring at approximately 80 and 200 C. A high
residual mass of 69% was observed which may be attributed to the
formation of titanium carbide or an oligomer as indicated by the
low solubility of 3. Compound 3 was not therefore suitable for use
in CVD.
TGA of 4 resulted in a three step weight loss profile, with
weight losses at approximately 80, 290 and 460 ◦C. A residual
mass of 25% was observed, which is lower than that expected for
TiAs remaining after decomposition (38%), which could be due
to difficulties in the isolation and manipulation of the compound.
Although the formation of 5 is interesting due to the novel
exchange, the compound is thought to be unsuitable as a precursor
to TiAs due to the lack of a pre-formed Ti–As bond. However,
the decomposition of this compound was of interest to investigate
whether a titanium nitride material could be isolated. TGA of
5 resulted in a five-step weight loss profile, with weight losses
occurring at approximately 20, 150, 250, 350and420 ◦C. A residual
mass of 74% was reported, suggesting that this complex is suitable
for use in CVD.
Fig. 3 ORTEP representation of [TiCl3(NMe2)(m-NMe2)2AsCl] (5) with
thermal ellipsoids at the 50% probability level. Hydrogen atoms omitted
for clarity.
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 5325–5331 | 5329
©

View Article Online
AACVD of compounds
(where the arsine and TiCl4 are added into the reaction flask
simultaneously, prior to the conduction of AACVD), whilst the
sequential method involves briefly passing the arsine through the
rig for approximately five minutes, prior to the addition of TiCl4.
The AACVD was conducted from the in situ reaction of TiCl4
and tBuAsH2 in toluene, using substrate temperatures of 400 and
600 ◦C. Strongly adherent rainbow films were deposited in all
instances. X-ray powder diffraction of the resultant films indicated
diffractograms consistent with TiO2 anatase when deposited via
the simultaneous method, with sequential AACVD resulting in
amorphous film deposition. Scanning electron microscopy (SEM)
was conducted on the TiO2 films deposited at 400 ◦C using the
simultaneous method, with island growth agglomerates of TiO2
anatase demonstrated (Fig. 6). Upon comparing the elemental
compositions of the films deposited using the two methods, whilst
no arsenic content was detected via simultaneous AACVD, an
approximate 3 : 1 ratio of Ti : As was observed via the sequential
method. These results suggest some incorporation of arsenic into
the film, which could be the result of the formation of titanium
arsenide, or the deposition of arsenic; however, oxidation largely
results in the formation of TiO2.
Single-source precursors 1, 2, 4 and 5 were used in AACVD in
an attempt to deposit TiAs thin films. Compound 3 was not
investigated due to its poor solubility in a range of solvents.
Compound 1 was generated in situ from the reaction of TiCl4
with one equivalent of AsPh3 in dichloromethane or toluene.
A dark red solution formed immediately and the reaction was
allowed to mix for 2 min before the AACVD reaction was
started. AACVD of 1 was conducted using both toluene and
dichloromethane, at substrate temperatures of 200, 400 and
600 ◦C. In all instances a strongly adherent rainbow film was
deposited, which was typically accompanied by a white non-
adhesive powder, indicative of homogeneous gas phase reactions.
X-ray powder diffraction of the rainbow films resulted in diffrac-
tograms consistent with that of TiO2 anatase, demonstrating peaks
at approximately 25.3, 36.9, 37.8, 38.6, 48.1, 53.9, 55.1, 62.1 and
63.8 2q/◦ (Fig. 5). The deposition of TiO2 was confirmed in
all instances by WDX analysis, with approximate 1 : 2 ratios of
titanium to oxygen being obtained. WDX analysis also indicated
that there was no arsenic present within the deposited films. The
WDX results are in agreement with the TGA results and that
previously reported for [SnCl4(AsPh3)2] where films of SnO2 were
formed rather than tin arsenide.20
Fig. 6 SEM image of TiO2 anatase deposited by the simultaneous
AACVD of TiCl4 and tBuAsH2 at 400 ◦C, at x100,000 magnification.
Fig. 5 Typical TiO2 anatase powder X-ray diffractogram deposited via
the AACVD of compounds 1, 2 and 4.
As previously mentioned 5 is not considered to be an appropri-
ate single-source precursor to TiAs due to lack of pre-formed
Ti–As bonds. The sequential AACVD of 5 resulted in the
deposition of an amorphous film, however due to the minimal
thickness of the deposited film, analysis via WDX was not possible.
In contrast to the in situ formation of 1 for the AACVD
reaction, compound 2 was synthesised (Experimental section)
and isolated prior to use in AACVD. Thus, 2 (0.5 g) was
dissolved in toluene (40 cm3), and AACVD was conducted using
a substrate temperature of 550 ◦C. Similarly to 1, a rainbow
film was deposited, which again upon conduction of powder
X-ray diffraction, produced a diffractogram consistent with TiO2
anatase. WDX analysis was conducted on the resultant rainbow
film, which confirmed TiO2 deposition again with zero arsenic
content.
The AACVD results for both 1 and 2 are consistent with what
was predicted by TGA, with apparent dissociation of arsine during
decomposition, and subsequent oxidation of titanium, resulting in
the deposition of TiO2 anatase. Consequently, precursors 1 and 2
were unsuccessful in depositing TiAs via AACVD.
LPCVD of compounds
From the AACVD results it was evident that the weak Ti–As bonds
(in 1, 2 and 4) were not being maintained during decomposition,
which in addition to the oxophilic nature of titanium, typically
resulted in TiO2 anatase being deposited. As such, LPCVD was
conducted in an attempt to minimise oxygen incorporation into
the system thus eliminating the potential for TiO2 deposition.
LPCVD was conducted on compounds 1, 3 and 5 with typically
amorphous films being deposited. In general, a rainbow/grey and
an orange deposit were observed in all cases, with at least one
of these demonstrating arsenic incorporation (for compounds 1
and 3) and nitrogen incorporation for compound 5, as detected
by WDX. Overall the results of the LPCVD experiments suggest
For precursors 4 and 5, in situ AACVD was conducted
using two different methods; simultaneous and sequential. The
simultaneous method refers to that previously described for 1
5330 | Dalton Trans., 2010, 39, 5325–5331
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
that either TiO2 or arsenic containing films are formed, with no
evidence for the deposition of TiAs via LPCVD.
7 I. M. Watson and J. A. Conner, Thin Solid Films, 1991, 196, L21–
L24.
8 T. S. Lewkebandara, J. W. Proscia and C. H. Winter, Chem. Mater.,
1995, 7, 1053–1054.
9 C. S. Blackman, C. J. Carmalt, I. P. Parkin, L. Apostolico, K. C. Molloy,
A. J. P. White and D. J. Williams, J. Chem. Soc., Dalton Trans., 2002,
2702–2709.
Conclusions
A range of TiCl4 arsine adducts have been synthesised and char-
acterised. The expected five and six-coordination complexes were
obtained for [TiCl4(AsPh3)n] n = 1, 2, [TiCl4Ph2AsCH2AsPh2],
10 F. Klingan, A. Miehr, R. A. Fischer and W. A. Hermann, Appl. Phys.
Lett., 1995, 67, 822–824.
11 A. C. Newport, J. E. Bleau, C. J. Carmalt, I. P. Parkin and S. A. O’Neill,
J. Mater. Chem., 2004, 14, 3333–3336.
t
and [TiCl4 BuAsH2]. However, reaction of TiCl4 and As(NMe2)3
12 J. E. Bleau, C. J. Carmalt, S. A. O’Neill, I. P. Parkin, A. J. P. White and
D. J. Williams, Polyhedron, 2005, 24, 463–468.
resulted in substituent exchange and formation of the novel com-
pound [TiCl3(NMe2)(m-NMe2)2AsCl]. The single-source route to
TiAs was investigated as it uses inexpensive and considerably
less complicated apparatus as well as less toxic precursors (than
for example arsine employed in conventional dual-source CVD
routes). However, the prior synthesis and isolation of very air
sensitive precursors have notable problems with the single-source
route. In addition, the relatively weak Ti–As bonds in these
complexes resulted in oxidation and the formation of titanium
oxide films.
13 C. J. Carmalt, A. C. Newport, I. P. Parkin, A. J. P. White and D. J.
Williams, J. Chem. Soc., Dalton Trans., 2002, 4055–4059.
14 C. J. Carmalt, A. Newport, I. P. Parkin, P. Mountford, A. J. Sealey and
S. R. Dubberley, J. Mater. Chem., 2003, 13, 84–87.
15 C. J. Carmalt, A. C. Newport, S. A. O’Neill, I. P. Parkin, A. J. P. White
and D. J. Williams, Inorg. Chem., 2005, 44, 615–619.
16 V. 6.45a BRUKER AXS Inc., 2003.
17 V. 2.03 BRUKER AXS Inc., 2001.
18 A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7.
19 L. J. Farrugia, J. Appl. Crystallogr., 1999, 32, 837.
20 M. F. Mahon, N. L. Moldovan, K. C. Molloy, A. Muresan, I. Silaghi-
Dumitrescu and L. Silaghi-Dumitrescu, Dalton Trans., 2004, 4017–
4021.
Acknowledgements
21 C. C. Hinckley, M. Matusz and P. D. Robinson, Acta Crystallogr., Sect.
C: Cryst. Struct. Commun., 1988, 44, 371–372.
22 R. Hart, W. Levason, B. Patel and G. Reid, Eur. J. Inorg. Chem., 2001,
2927–2933.
The EPSRC is thanked for the studentship. Pilkington is thanked
for the supply of glass substrate. SAFC Hitech Ltd. is thanked
for the supply of As(NMe2)3 and BuAsH2. Thanks also goes to
Mrs G. A. Maxwell, Dr S. Firth and Mr K. Reeves for help with
microanaysis, TGA/DSC and SEM respectively.
23 W. Levason, B. Patel, G. Reid, V. A. Tolhurst and M. Webster, J. Chem.
t
Soc., Dalton Trans., 2000, 3001–3006.
24 F. Challenger and F. Pritchard, J. Chem. Soc., Trans., 1924, 125, 864–
875.
25 F. A. Cotton, J. H. Matonic, C. A. Murillo and M. A. Petrukhina,
Inorg. Chim. Acta, 1998, 267, 173–176.
26 R. J. H. Clark, J. Lewis and R. S. Nyholm, J. Chem. Soc., 1962, 2460.
27 C. J. Carmalt and S. Basharat, in “Precursors to Semiconducting
Materials” in Comprehensive Organometallic Chemistry, ed. D. O’Hare,
Elsevier Ltd., Oxford, 2007, vol. 12.
References
1 Y. Senzaki and W. L. Gladfelter, Polyhedron, 1994, 13, 1159–1167.
2 P. A. Lane, B. Cockayne, P. J. Wright, P. E. Oliver, M. E. G. Tilsley,
N. A. Smith and I. R. Harris, J. Cryst. Growth, 1994, 143, 237–242.
3 M. E. G. Tilsley, N. A. Smith, B. Cockayne, I. R. Harris, P. A. Lane,
P. E. Oliver and P. J. Wright, J. Alloys Compd., 1997, 248, 125–131.
4 S. M. Godfrey, C. A. McAuliffe, A. G. Mackie and R. G. Pritchard,
Chemistry of arsenic antimony and bismuth, ed. N. C. Norman, 1998.
5 K. L. Choy, Prog. Mater. Sci., 2003, 48, 57–170.
28 M. H. Zimmer, R. Hovel, W. Brysch, A. Brauers and P. Balk, J. Cryst.
Growth, 1991, 107, 348–349.
29 S. Goto, C. Jelen, Y. Nomura, Y. Morshita and Y. Katayama, J. Cryst.
Growth, 1995, 150, 568–573.
30 K. Kirschbaum, O. Conrad and D. M. Giolando, Acta Crystallogr.,
Sect. C: Cryst. Struct. Commun., 2000, 56, ee541–E541.
31 R. T. Cowdell and G. W. A. Fowles, J. Chem. Soc., 1960, 2522–2526.
6 A. H. Cowley and R. A. Jones, Polyhedron, 1994, 13, 1149–1157.
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 5325–5331 | 5331
©