Tungsten nitrido complexes as precursors for low temperature chemical vapor deposition of WNxCy films as diffusion barriers for Cu metallization

Article abstract of DOI:10.1021/ja4117582

Tungsten nitrido complexes of the form WN(NR₂)₃ [R = combinations of Me, Et, ⁱPr, ⁿPr] have been synthesized as precursors for the chemical vapor deposition of WN_xC_y, a material of interest for diffusion barriers in Cu-metallized integrated circuits. These precursors bear a fully nitrogen coordinated ligand environment and a nitrido moiety (Wi - N) designed to minimize the temperature required for film deposition. Mass spectrometry and solid state thermolysis of the precursors generated common fragments by loss of free dialkylamines from monomeric and dimeric tungsten species. DFT calculations on WN(NMe ₂)₃ indicated the lowest gas phase energy pathway for loss of HNMe₂ to be β-H transfer following formation of a nitrido bridged dimer. Amorphous films of WN_xC_y were grown from WN(NMe₂)₃ as a single source precursor at temperatures ranging from 125 to 650 C using aerosol-assisted chemical vapor deposition (AACVD) with pyridine as the solvent. Films with stoichiometry approaching W₂NC were grown between 150 and 450 C, and films grown at 150 C were highly smooth, with a RMS roughness of 0.5 nm. In diffusion barrier tests, 30 nm of film withstood Cu penetration when annealed at 500 C for 30 min.

Full text of DOI:10.1021/ja4117582

Article
pubs.acs.org/JACS
Tungsten Nitrido Complexes as Precursors for Low Temperature
Chemical Vapor Deposition of WN C Films as Diffusion Barriers for
x y
Cu Metallization
†
‡
†
†
†
K. Randall McClain, Christopher O’Donohue, Arijit Koley, Richard O. Bonsu, Khalil A. Abboud,
‡
‡
,†
Joseph C. Revelli, Timothy J. Anderson, and Lisa McElwee-White*
†
Department of Chemistry, University of Florida, Gainesville, Florida, 32611-7200, United States
‡
Department of Chemical Engineering, University of Florida, Gainesville, Florida, 32611-6005, United States
*
S Supporting Information
ABSTRACT: Tungsten nitrido complexes of the form
i
n
WN(NR ) [R = combinations of Me, Et, Pr, Pr] have been
2
3
synthesized as precursors for the chemical vapor deposition of
WN C , a material of interest for diffusion barriers in Cu-
x
y
metallized integrated circuits. These precursors bear a fully
nitrogen coordinated ligand environment and a nitrido moiety
(
WN) designed to minimize the temperature required for film
deposition. Mass spectrometry and solid state thermolysis of the
precursors generated common fragments by loss of free dialkylamines from monomeric and dimeric tungsten species. DFT calculations
on WN(NMe ) indicated the lowest gas phase energy pathway for loss of HNMe to be β-H transfer following formation of a nitrido
2
3
2
bridged dimer. Amorphous films of WN C were grown from WN(NMe ) as a single source precursor at temperatures ranging from
x
y
2 3
1
25 to 650 °C using aerosol-assisted chemical vapor deposition (AACVD) with pyridine as the solvent. Films with stoichiometry
approaching W NC were grown between 150 and 450 °C, and films grown at 150 °C were highly smooth, with a RMS roughness of
2
0.5 nm. In diffusion barrier tests, 30 nm of film withstood Cu penetration when annealed at 500 °C for 30 min.
1
2,34−37
INTRODUCTION
phase.
Titanium nitride is an effective diffusion barrier
■
and adhesion promoter for Al−Cu and tungsten interconnects
but may prove ineffective as a Cu barrier due to its polycrystalline
nature that allows diffusion along grain boundaries.
addition to metal nitrides, “self-forming” MnSi O barriers have
The continuing decrease in microelectronic circuit dimensions
has necessitated a shift in interconnect materials to minimize RC
resistance × capacitance) time delay, which has predominantly
38−40
In
(
x
y
limited overall device speed. To reduce this time delay,
metallization schemes based on aluminum, tungsten, and SiO2
are being replaced with copper and low-κ (low-dielectric-
41−45
received recent attention.
To form these barriers, a Cu−Mn
alloy is deposited by PVD or Mn metal is deposited onto the
Cu layer via PVD or CVD. Mn then diffuses through Cu to the
Cu/dielectric interface during a subsequent annealing step
above 350 °C). The resulting MnSi O barrier, formed from
reaction of Mn with the SiO -based dielectric, has a negligible
impact on the resistivity of the Cu interconnect because it forms
at the surface of the dielectric layer. These highly resistive
materials, however, cannot be used as barriers between Cu and
semiconducting layers.
1
constant) materials. This transition has generated a need for
efficient diffusion barrier materials to prevent destructive
interlayer diffusion of copper into adjacent semiconducting and
(
x y
2
−4
2
dielectric layers.
Present Cu interconnect barriers employ
a bilayer system composed of a TaN diffusion barrier capped1
with Ta to promote the adhesion and nucleation of textured Cu.
This TaN/Ta bilayer is deposited using PVD (physical vapor
deposition) or ionized-PVD, the latter offering advantages in
5
WN and the ternary material WN C offer several advantages
x
x y
terms of conformality. While advances in this legacy system will
over other transition metal nitrides as Cu diffusion barriers
in terms of resistivity, barrier performance, and processing
simplicity. For WN , stoichiometry ranging from W N to WN
has been investigated.
likely extend its use, the “line-of-sight” deposition nature of PVD
will become prohibitive in increasingly aggressive high aspect
ratio features and an eventual transition to a more conformal
deposition process, such as CVD (chemical vapor deposition) or
ALD (atomic layer deposition), will occur.
Much research has focused on the development of CVD/
46
x
2
47,48
Tungsten-rich WN barriers closer
x
6
to W N have proven to be most desirable due to their lower
2
resistivity and Cu migration rates. Similarly, growth of WN C
films with tungsten-rich stoichiometry approaching W NC has
resulted in good Cu barriers with low resistivity.
x y
ALD systems for diffusion barriers based on transition metal
2
33,49,50
7
8−16
17−23
nitrides, especially the nitrides of tantalum,
titanium,
2
4−33
and tungsten.
While an effective Cu barrier when deposited
via PVD, TaN deposition using CVD/ALD generally results in
Received: November 28, 2013
Published: January 2, 2014
x
nitrogen rich stoichiometry closer to the high resistivity Ta N5
3
©
2014 American Chemical Society
1650
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
i
Early coreactant systems for deposition of WN employed
that of Cl (CH CN)W(N Pr) (Table 1). The next logical step
4 3
x
volatile homoleptic tungsten compounds such as WF or
was to eliminate the N-substituent bond by incorporation of
a terminal nitrido moiety in 1. Replacement of the chloride
6
25,27
63
W(CO) with ammonia as the nitrogen source.
While cost-
6
effective, these systems generally require high temperature, result
in the production of corrosive byproducts, and lead to the
incorporation of undesirable impurities in the films. In an effort to
alleviate these limitations, much research has focused on tungsten
complexes bearing N-coordinated ligands as WN /WN C
ligands of the earlier imido and hydrazido complexes with
dimethylamides in 1 circumvents the formation of the corrosive
byproduct HCl, which was observed by mass spectrometry in the
29
reactor effluent from depositions using Cl (CH CN)W(NPh).
4
3
By varying the dialkylamido ligands in nitrido complexes
2−9 and 11, the effect of ligand size and branching on pertinent
properties, such as volatility and thermal stability, was
ascertained.
x
x
y
2
6,28,32
precursors. These complexes generally contain amide,
2
6,28,30,31,36
33,51
52
imido,
nate
hydrazido,
ligands. CVD or ALD using these “metal−organic”
WN /WN C precursors, either in single-source mode or in the
amidinate, and/or guanidi-
5
3,54
Precursor Synthesis. After 1 was synthesized in high
x
x y
t
63
presence of an appropriate coreactant (usually H or NH ), has
yield from WN(O Bu) and Zr(NMe ) , synthesis of other
2
3
3 2 4
t
resulted in films virtually free of unsuitable impurities and without
the formation of corrosive byproducts.
WN(NR ) complexes from WN(O Bu) was explored. As
2 3 3
expected, the reaction of WN(O Bu) with Zr(NEt ) in toluene
t
3
2 4
While several ALD chemistries are reactive enough to allow
deposition of WN /WN C at lower temperatures (180−
resulted in complete amide/alkoxide exchange, as observed
in the H NMR spectra of the product mixture. However, the
1
x
x
y
3
2,55−58
350 °C),
higher temperatures (>350 °C) are generally
product, WN(NEt ) (2), could not be isolated by selective
2
3
still required for dissociation of critical bonds during film growth
crystallization or sublimation.
under single source CVD conditions (Table 1). Temperatures
Transamination from 1 using dialkylamines was chosen as a
strategy due to the near thermoneutrality of the exchange and the
relative volatility of the dimethylamine product. Indeed, reaction
of 1 with excess diethylamine (60 equiv) at reflux, either neat
or in pentane solution, for 3 h resulted in complete substitu-
tion to produce 2. The crude product was purified easily by
recrystallization from hexamethyldisiloxane, resulting in a high
yield (85%) of analytically pure 2.
Table 1. Minimum Temperature for CVD of WN C Films
x
y
from Single Source Precursors
precursor
T (°C)
ref
t
Cl W(N Bu) (py)
550
450
450
400
300
125
61
25
30
48
62
a
2
2
2
t
t
W(N Bu) (NH Bu)
2
2
i
Cl (CH CN)W(N Pr)
4
3
t
i
W(N Bu) (NMe ){( Pr N) CNMe }
2
2
2
2
2
Scheme 1. Synthesis of 2
Cl (CH CN)W(NNMe )
4
3
2
WN(NMe2)3
a
This work.
above 350 °C can compromise some low-κ materials due to their
low thermal stability and can encourage crystallization of barrier
materials resulting in low-energy diffusion pathways for Cu
i
The homoleptic amides WN(N(Me)Et) (3), WN(N(Et)Pr)
3
3
n
(
4), and WN(N Pr ) (5) were synthesized from 1 by refluxing
2 3
2
migration by grain boundary formation. Design and synthesis of
with a large excess of the appropriate dialkylamine and purified
by recrystallization from hexamethyldisiloxane. The reactions
were sensitive to the steric demand of the incoming amine, with
4 obtained in refluxing pentane (36 °C), while formation of 5
required refluxing in neat N-ethylisopropylamine (71 °C).
precursors for the deposition of metal nitride films at low
5
9,60
temperature (below 350 °C) remain of great interest.
We previously reported the synthesis and CVD of WNx
nanospheres from WN(NMe ) (1) at temperatures as low as
2
3
6
3
7
5 °C. We now report the synthesis and characterization
of several potential WN /WN C precursors of the general
x
x
y
Scheme 2. Synthesis of 3−5
formula WN(NR ) (2−9 and 11) and describe the deposition
2
3
of continuous WN C thin films from 1 at temperatures as low
x
y
as 125 °C.
RESULTS AND DISCUSSION
Precursor Design. Three of the critical issues in CVD
precursor design are minimization of the deposition temperature,
the avoidance of deleterious side reactions, and optimization
of certain physical properties such as solubility and volatility.
■
60
In an earlier study involving the kinetics of film growth from the
i
tungsten imido complexes Cl (CH CN)W(NR) [R = Ph, Pr,
4
3
allyl], correlation of the apparent activation energy for film
growth with the bond dissociation energy (BDE) of the
N(imido)−C bond implicated cleavage of that bond as the rate
i
Formation of WN(N Pr ) could not be achieved by trans-
2
3
31
determining step for film growth. Modification of the precursor
design to replace the critical N(imido)−C bond with a weaker
N−N bond resulted in preparation of the hydrazido complexes
amination of 1, even under forcing conditions in neat refluxing
i
diisopropylamine (85 °C). However, WN(NMe ) (N Pr ) (6)
2
2
2
i
and WN(NMe )(N Pr ) (7) were obtained by manipulation
2
2 2
6
2
Cl (CH CN)W(NNR ) (R = Me , Ph , −(CH ) −). As
of the reaction conditions. Transamination of 1 with excess
diisopropylamine in refluxing pentane solution resulted in
selective formation of 6, while conducting the transamination
4
3
2
2
2
2
2 5
predicted from the BDEs, the deposition temperature for
Cl (CH CN)W(NNMe ) was 300 °C, a drop of 150 °C from
4
3
2
1
651
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
1
13
Scheme 3. Synthesis of 6−9
Spectroscopic Characterization of 2−11. The H and C
NMR spectra of complexes 2−6 and 8 displayed sharp first-order
signals indicative of unrestricted rotation about all W−N
amide
bonds, as found in 1. The NMR spectra of complexes 7 and 9
exhibited diastereotopic methyl groups on the diisopropylamide
ligands. This was evidenced by the appearance of two doublets in
1
the H NMR spectra and two closely spaced singlets for the
methyl groups in the 13C spectra. The variable temperature H
1
NMR spectra of 7 (Figure S-1, Supporting Information) and 9
(
Figure S-2, Supporting Information) indicated restricted
bond rotation at lower temperatures, evidenced by
Overall yield from 1.
W−N
amide
coalescence of the diastereotopic methyl doublets into a broad
of 1 in refluxing diisopropylamine allowed for the selective
synthesis of 7. In both cases, the product could be purified by
recrystallization.
The remaining dimethylamide groups of 6 and 7 were cleanly
replaced by transamination with excess diethylamine in refluxing
singlet. The methyl doublets coalesce at −25 and −5 °C for 7
and 9, respectively. In contrast, the H and C NMR spectra
of compounds 10 and 11 indicated restricted W−N
rotation at room temperature with broad signals for the methyl
groups of the diisopropylamide ligands. The behavior of
compound 11 is similar to its Cr(VI) analogue CrN(N Pr ) ,
in which restricted Cr−N
the basis of NMR results. Variable temperature H NMR
spectra of 10 (Figure S-3, Supporting Information) and 11
(
of the methyl doublets at 25 °C for both 10 and 11. Compound 10
displayed four doublets for the methyl protons at low temperature
(−45 °C in toluene-d
high temperature (75 °C in benzene-d
the presence of diastereotopic methyl groups. It should be noted
that the proton signal for the tert-butyl group of 10 remained
a sharp singlet at all temperatures, indicating unrestricted W−O
bond rotation and weaker π-donor character for this ligand.
Compound 11 displayed two doublets for the methyl protons at
1
13
bond
amide
i
pentane or benzene solution to produce WN(NEt ) (N Pr ) (8)
2
2
2
i
i
and WN(NEt )(N Pr ) (9), respectively. Analytically pure 8 was
2 3
2
2 2
bond rotation was postulated on
obtained by recrystallization from hexamethyldisiloxane, while
pure 9 was more efficiently obtained by sublimation owing to its
very high solubility. Attempts to generate 8 and 9 by trans-
amination of 2 with diisopropylamine did not result in any signifi-
cant substitution, even upon reflux in neat diisopropylamine.
amide
6
5
1
Figure S-4, Supporting Information) confirmed coalescence
i
t
WN(NPr ) (11) waseventuallysynthesized from WN(O Bu)
2
3
3
t
using an alternative two-step procedure. Reaction of WN(O Bu)
8
) and two doublets for the methyl protons at
), which is consistent with
3
i
64
with La(NPr ) ·2THF (0.70 equiv) resulted in amide/alkoxide
6
2
3
t
i
scrambling to generate WN(O Bu)(NPr ) (10), which was
2
2
t
isolated by sublimation after workup. Reaction of WN(O Bu) or
3
i
1
1
0 with additional La(NPr ) ·2THF did not result in formation of
2 3
1. However, 11 could be obtained by reaction of 10 with excess
lithium diisopropylamide in Et O at 65 °C.
2
low temperature (−40 °C in toluene-d ) and one doublet for the
methyl protons at high temperature (75 °C in benzene-d ).
8
Scheme 4. Synthesis of 10 and 11
6
X-ray Crystal Structure of 7. Complex 7 was chosen for
X-ray structure determination because of its heterosubstituted
sterically crowded ligand environment and temperature depend-
ent NMR spectra. Crystals suitable for crystallography were
grown from concentrated hexamethyldisiloxane solution at low
temperature. The asymmetric unit of 7 consists of the three
chemically equivalent but crystallographically independent
complexes: A, B, and C (Table 3). Accordingly, the following
discussion will use bond lengths and angles from complex A
unless otherwise specified. As seen in Figure 1, 7 adopts a
Compounds 2−11 were obtained as mostly colorless, air
and moisture sensitive crystalline solids after crystallization or
sublimation. Samples of the purified 2−9 and 11 were sublimed
to assay their volatility and stability during gas phase transport
(
Table 2). Interestingly, compounds 3 and 5, which contain
63
distorted tetrahedral geometry similar to 1 and 2 (Figure S-5,
Supporting Information). The average (A, B, and C) W1N1
Table 2. Sublimation Data for 2, 4, 6−9, and 11
(
WN) bond length of 1.683(3) in 7 is very similar to that in 1
[
(
WN 1.680(2)] and slightly longer than that in WN(N-
compound
pressure (mTorr)
T (°C)
% recovery
i
66
63
Ar) Pr) (Ar = 3,5-C H Me ) [WN 1.669(5)]. As in 1,
3 6 3 2
2
4
6
7
8
9
150−250
150−250
350−450
350−450
250−350
250−350
100−200
75−80
80−90
65−70
65−70
62−68
58−68
115−125
55
40
35
40
30
30
90
the WN
bond lengths of 7 are shorter [WN
amide
avg
i
1
.953(4)] than those in WN(N(Ar) Pr) (Ar = 3,5-C H Me )
3
6
3
2
i
[
WN(Ar) Pr 1.972(3)], a result consistent with the stronger
i
67
donor (σ, π) character of NR (R = Me, Pr) versus
2
i

N(Ar) Pr (Ar = 3,5-C H Me ). The total bond angles around
6
3
2
2
N2/N3/N4 in 7 are consistent with sp hybridization [359.9(3),
11
3
60.0(3), and 360.1(4)°, respectively], and the amide planes
of 7 are parallel to the WN axis, allowing π donation into the
2
2
unsymmetrically substituted amide ligands, did not sublime well.
Compounds 2, 4, and 6−9 all sublimed with moderate levels
of decomposition (recovery = 30−55%). Compounds 7 and 9
sublimed with higher yield from the crude reaction mixtures
than as pure substances. Compound 11 proved to be unique,
subliming at higher temperature with excellent thermal stability
W based d
and d orbitals. A distortion toward trigonal
x −y
xy
pyramidal geometry is observed in 7, evidenced by the contrac-
tion [{NWN} , 105.15(16)°] and expansion [{NW
avg.
N} , 113.42(15)°] of bond angles around W. As in 1, this is
avg.
caused by steric repulsion between the more bulky amide groups
and a tendency for low overlap between fully occupied bonding
68
(
recovery = 90%).
orbitals.
1
652
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
Table 3. Selected Bond Lengths (Å) and Angles (deg) for 7
W1A−N1A
W1B−N1B
W1C−N1C
W1A−N2A
W1B−N2B
W1C−N2C
W1A−N3A
W1B−N3B
W1C−N3C
W1A−N4A
W1B−N4B
1.688(3)
1.679(3)
1.681(3)
1.949(3)
1.952(4)
1.957(4)
1.950(3)
1.951(4)
1.955(3)
1.961(4)
1.948(3)
1.945(3)
105.47(15)
106.04(15)
104.95(16)
N1A−W1A−N3A
N1B−W1B−N3B
N1C−W1C−N3C
N1A−W1A−N4A
N1B−W1B−N4B
N1C−W1C−N4C
N2A−W1A−N3A
N2B−W1B−N3B
N2C−W1C−N3C
N2A−W1A−N4A
N2B−W1B−N4B
N2C−W1C−N4C
N3A−W1A−N4A
N3B−W1B−N4B
N3C−W1C−N4C
104.52(12)
104.50(16)
104.84(16)
105.45(16)
106.20(13)
106.17(12)
113.33(15)
110.87(14)
111.62(15)
113.65(14)
115.00(14)
115.55(14)
113.29(15)
113.27(14)
112.64(15)
W1C−N4C
N1A−W1A−N2A
N1B−W1B−N2B
N1C−W1C−N2C
+
monomeric [(M-HNR ) + H] and dimeric [(M -HNR ) +
2
2
2
+
+
+
H] species. Both [(M-HNR ) + H] and [(M -HNR ) + H]
2
2
2
ions were clearly evident in the mass spectra of compounds 1−4.
Heterosubstituted compounds 6−8 generated fragments corre-
sponding to loss of both types of dialkylamines from monomeric
and dimeric species, but in most cases (7−9), the relative
abundance of some of the fragments was very low (<1%).
However, the mass spectrum of compound 6 clearly displayed all
+
i
four possible fragments {[(M -HNMe ) + H] , [(M -HN Pr ) +
2
2
2
2
+
+
i
+
H] , [(M-HNMe ) + H] , [(M-HN Pr ) + H] } with higher
relative abundance. Only [M + H] and the fragment cor-
responding to loss of diisopropylamine from a monomeric species
2
2
+
i
+
[
(M-HNPr ) + H] were clearly observed for compound 11.
2
Thermolysis of Neat 1 and 2. Thermolyses of 1 and 2 were
conducted to identify the volatile byproducts of their thermal
decomposition. Neat 1 was heated to 145 °C under 1 atm
of Ar, and the volatile products were condensed at −78 °C.
1
H NMR and GC-EI-MS analysis of the condensate confirmed
the presence of both dimethylamine and trimethylamine. The
material remaining after thermolysis was a black powder. Neat
2 was heated to 205 °C under 1 atm of argon, and the volatile
Figure 1. Displacement ellipsoids drawing of 7 (molecule A is shown);
ellipsoids are drawn at 50% probability.
1
13
products were condensed at −78 °C. H NMR, C NMR, and
Mass Spectrometry. Since the fragmentation patterns
GC-EI-MS analysis of the condensate confirmed the presence of
1
13
observed during mass spectrometry often correlate well with
diethylamine. The H NMR and C NMR spectra indicate that
a small amount of N-ethylethanimine was also present. Again, the
material remaining after thermolysis was a black powder.
Computational Study of the Decomposition of 1. Since
CI-MS (compounds 1−9 and 11) and thermolysis (compounds
1 and 2) both indicate formation of dialkylamines during
decomposition of WN(NR ) species, a computational study of
3
1,69,70
observed CVD decomposition pathways,
the positive ion
CI (chemical ionization) mass spectra of compounds 1−9 and 11
were obtained (Table 4). For all compounds, the appropriate
+
[
M + H] ion was observed in high relative abundance, indicating
the moderate stability of these thermally sensitive compounds
under the ionization conditions. For compounds 1−9 and 11,
additional tungsten-containing ions were observed. Some of
these fragments correspond to loss of dialkylamine from
2
3
dialkylamine formation was performed using 1 as a model com-
pound. To investigate the pathway for production of NHMe2
Table 4. Relative Abundances for Positive Ion CI Mass Spectra of Compounds 1−9 and 11
compound (MW)
selected ions (m/z)(relative abundance) [formula]
+
+
+
1 (330)
2 (414)
3 (372)
4 (456)
5 (498)
6 (386)
7 (442)
8 (442)
9 (470)
11 (498)
616(31) [(M -HNMe ) + H] ; 331(95) [M + H] ; 286(7) [(M-HNMe ) + H]
2 2 2
+
+
+
756(8) [(M -HNEt ) + H] ; 415(72) [M + H] ; 342(5) [(M-HNEt ) + H]
2
2
2
+
+
+
686(2) [(M -HN(Me)Et) + H] ; 373(52) [M + H] ; 314(4) [(M-HN(Me)Et) + H]
2
i
+
+
i
+
826(1) [(M -HN(Et) Pr) + H] ; 457(95) [M + H] ; 370(4) [(M-HN(Et) Pr) + H]
2
n
+
+
896(1) [(M -HN Pr ) + H] ; 499(12) [M + H]
2
2
+
i
+
+
+
i
+
728(26) [(M -HNMe ) + H] ; 672(11) [(M -HN Pr ) + H] ; 387(96) [M + H] ; 342(6) [(M-HNMe ) + H] ; 286(2) [(M-HN Pr ) + H]
2
2
2
2
2
2
443(23.4) [M + H]+
+
+
812(1) [(M -HNEt ) + H] ; 443(15) [M + H]
2
2
+
+
i
+
471(83) [M + H] ; 398(1) [(M-HNEt ) + H] ; 370(2) [(M-HN Pr ) + H]
2
2
+
i
499(52) [M + H] ; 398(1) [(M-HN Pr ) + H]
2
1
653
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
3
1,71−73
Scheme 5. Thermolysis of 1 and 2
horizontal reactor).
The temperature dependence of film
1
.32
growth rate in the low-temperature regime was estimated as T
.
Moreover, there was no growth rate dependence on the substrate
orientation (i.e., p-type boron-doped Si(100) and n-type arsenic-
doped Si(111)), which further supports mass transfer limited
growth.
The maximum growth rate observed in the low-temperature
regime was 4.9 Å/min at a deposition temperature of 400 °C,
whereas the minimum growth rate was 2.3 Å/min at 125 °C.
These growth rates are less than that of typical CVD growth
74
rates, which supports the possibility that the rate determining
step is bimolecular, consistent with the location of lower energy
transition states for amine loss by DFT (Figure 2).
upon thermal decomposition of 1, two possible decomposition
routes were modeled by DFT calculations. In the first (pathway
A, Figure 2), the amine arises from an intramolecular β-proton
abstraction by one of the amido ligands (via TS1-A) to generate a
dimethylamine ligand in INT1-A, which is displaced upon
coordination of a second molecule of 1 (via TS2-A) to afford
INT2-A. In the second pathway (pathway B, Figure 2),
dimerization of 1 through a bridging nitrido ligand precedes
the proton transfer to generate the amine. The activation energy
At higher temperature, >400 °C, the growth rate decreases
with increasing temperature with an apparent activation energy
of −9.37 ± 0.22 kcal/mol (−0.386 ± 0.008 eV). This regime
exhibits an abrupt onset, occurring at 450 °C, and falls to
0.3 Å/min at 650 °C. The decreasing-growth-rate regime is likely
dominated by premature and aggressive gas-phase reactions
upstream of the substrate, desorption of adatoms from the
surface, or a combination of the two. Parasitic homogeneous
reaction of 1 is evidenced by the observation of wall deposition
upstream of the heated susceptor, no measurable dependence of
growth rate on the Si substrate orientation, as well as a significant
increase in carbon content in the films, which is most likely from
incorporation of byproducts from upstream reactions.
⧧
for the rate determining step in pathway A (ΔG = 49.4 kcal/mol;
⧧
ΔH = 47.6 kcal/mol via TS1-A) is calculated to be significantly
⧧
⧧
higher than that for pathway B (ΔG = 31.0 kcal/mol; ΔH =
1
4.0 kcal/mol via TS-B). These results suggest that decom-
position of 1 to release dimethylamine could involve
intermolecular interactions in the gas phase above the substrate.
Another possibility under deposition conditions is a surface
reaction in which coordination of a nitride moiety to an incoming
precursor molecule precedes the loss of NH(CH ) . The latter is
WNxCy Film Microstructure. For barrier films, an
amorphous microstructure is more desirable than a polycrystal-
3
2
line one because amorphous materials offer no low-energy4
consistent with the formation of dimethylamine during the solid
state thermolysis of 1.
Formation of dimethylamine according to the proposed
routes is overall endothermic by approximately 8.2 kcal/mol
diffusion pathways, such as grain boundaries, for Cu migration.
It should be noted that depositions using 1 as a single-source
precursor yield highly amorphous thin films with no detectable
crystallinity (Figure 4A) within the relatively large temperature
range 125−650 °C. Figure 4A depicts a 2θ scan for the lowest
(
ΔG = 24.4 kcal/mol) according to pathway A, or 5.8 kcal/mol
(ΔG = 7.7 kcal/mol) according to pathway B. These results are
(
125 °C), highest (650 °C), and intermediate (350 °C) deposi-
not unreasonable for a reaction occurring under film deposition
conditions.
tion temperatures, and clearly shows the amorphous nature of
the films per the absence of any crystallographic peaks (as
compared to the dominant reflections for powders indicated by
the stick pattern in Figure 4B). On the basis of previous work
using other precursors, the primary β-WN C reflections are
WN C Film Growth. Pyridine was chosen as the solvent for
x
y
AACVD because of its stability, appropriate volatility, lack of
undesirable elements, and ability to solubilize 1. To investigate
the possibility that pyridine might coordinate to the precursor,
x
y
expected to result from a solid solution of β-W N and β-WC
2
1−x
pyridine was added to a sample of 1 in benzene-d . No evidence
6
1
and that their respective reflections are very approximate to each
other, hence illustrated as a single reflection. The appropriate
crystallographic planes are labeled for these reflections, and they
of pyridine coordination to 1 was detected in the H NMR
spectra. Typically, films grown using a solution of 1 in pyridine
(
0.051 M) under the conditions described in the Experimental
correspond to the planes of both β-W N and β-WC solutions.
2
1−x
Section resulted in film thicknesses less than 100 nm at a
substrate temperature as low as 125 °C. Films grown using 1 as a
single-source precursor under N ambient adhered to one of two
Similarly, the primary WO pattern is an amalgamation of
3
reflections for three common WO phases: cubic, tetragonal, and
3
2
monoclinic.
growth regimes, as seen in the Arrhenius plot of growth rate (G)
vs reciprocal temperature in Figure 3. The error bars shown in
the figure are representative of the film thickness standard
deviation near the center of the substrate, for multiple samples
and measurements. A small apparent activation energy of
WN C Film Composition. X-ray photoelectron spectros-
x
y
copy (XPS) was used to quantify atomic composition in the film
and to establish the valence states of each element in the film.
Figure 5 shows the atomic concentration in the films versus
deposition temperature. Only four elements were present in the
film: W, C, N, and O. The error bars represent a sensitivity of one
atomic percent for the instrument used.
In the deposition temperature range 150−450 °C, the
film composition was relatively constant with average atomic
percent values of 41, 13.6, 13.6, and 31.5% for W, N, C, and O,
1
.39 ± 0.12 kcal/mol (0.056 ± 0.005 eV) is estimated in the
low temperature region, 125−400 °C. A similar low activation
energy was observed during film growth using the allylimido
3
1
complex Cl (CH CN)W(NCH CHCH ) (0.15 eV).
4
3
2
2
The allylimido complex is expected to have a higher activation
energy for film growth than that of 1, due to the presence of
the imido substituent (see the Precursor Design section). This
low apparent activation energy is consistent with mass transfer
respectively. This results in a film consisting of WO
carbonitride phase WN0.45 0.45. Typically, WN films where
and the
3
C
C
x y
n
limited growth, which shows a T temperature dependence,
x and y ≈ 0.5 provide the lowest resistivity and highest Cu
migration resistance.
where n is dependent on the reactor design (e.g., n = 1.5 for a
1
654
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
−1
Figure 2. Energy profiles (ΔG and ΔH, in kcal mol ) and bond lengths (Å) for pathway A involving initial unimolecular β-proton abstraction by amide
and pathway B involving initial bimolecular β-proton abstraction by amide. Calculations performed at the M06L/LanL2DZ/6-31G* level in the gas
phase at room temperature.
There is a noticeable change in composition for films grown
above 550 °C and below 150 °C. Above 550 °C, the change is
most likely caused by upstream gas phase decomposition of both
precursor and solvent. A dramatic increase in C and a sharp
decrease in W are consistent with solvent decomposition and/or
precursor disappearance by parasitic reaction. Incorporation of
oxygen in the film is reduced by the stuffing of carbon into the
film. Below 150 °C, it is likely that the precursor does not
decompose fully and ligands are incorporated into the film.
At these low temperatures, the sticking probability of gaseous
molecules is high and any background oxygen or water vapor is
likely integrated into the film.
75
76
1
655
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
Figure 3. Arrhenius plot of growth rate (G) for films deposited with 1
depicting the two well-defined growth regimes.
Figure 6. Peak evolution with deposition temperature for primary peaks
(
A) W 4f, (B) N 1s, (C) C 1s, and (D) O 1s. The superimposed stick
77
patterns identify typical BEs for relevant valence states associated with
each respective element.
W 4f5/2 peaks, respectively. This implies a reduction of W, most
likely to an oxide solid solution of WO and WO , as WO . The
Figure 4. (A) Grazing incidence X-ray diffraction (GIXRD) patterns
taken for films deposited using compound 1. (B) Typical crystallo-
graphic planes for relevant materials with respect to their preferred
structures and according to their relative intensities.
3
2
x
percentage of the W signal in the oxide state, however, is reduced
to about 25% with the balance consistent with a WN C solid
x
y
solution. The W 4f peak for the sample grown at the highest
temperature (650 °C) indicates 85% of the W resides in a WN C
x
y
energy state, with only about 15% of the W signal existing in a
slightly more reduced (0.8 eV negative shift) WO form.
x
The N 1s peak position is consistently at a BE of 397.5 eV
for all deposition temperatures except 125 °C, implying it is
W bound. CN can invoke multiple BEs for the N 1s peak
x
depending on the nature of the film, but these typically range
from 398.5 to 400.6 eV. Deconvolution of the peak for the film
deposited at 125 °C generated two distinct peaks, one at 400.1
eV (∼30% of the N signal), corresponding to CN , and the other
x
at 397.5 eV, corresponding to WN .
x
The C 1s peak is nominally located at 283.5 eV in WN C ,
x
y
Figure 5. Atomic composition of films as a function of deposition
temperature as determined by XPS.
2
84.5 eV as amorphous C, and 286 eV in CN . Only films
x
deposited at 125 °C yielded a C 1s peak located at 286 eV
285.9 eV) (∼30% of the C signal), which is consistent with the
(
Figure 6 illustrates the evolution of the primary elemental
peaks in the XPS spectra of films grown at various deposition
temperatures. With increasing deposition temperature, the W 4f
peak shifted from higher binding energies (BEs) to lower ones,
suggesting a transition from a higher oxidation state to a lower
N 1s data. Growth temperatures from 150 to 275 °C produced
films almost entirely (85%) in a carbidic state. At intermediate
temperatures (300 °C < T < 500 °C), the carbon predominantly
resided in a carbidic (∼90%) state, with the balance as
amorphous carbon. Above 550 °C, the amorphous carbon
content increased linearly until 650 °C to a maximum 45%,
balanced by 55% carbidic C. An increase in free C is expected
with increasing deposition temperature, since solvent decom-
position is more likely. It appears that growing at 125 °C resulted
in incomplete pyrolysis of 1, which is evident in both the C 1s and
N 1s peaks.
one. Typical BEs for tungsten oxide, WO , reside around 35.7
3
and 37.7 eV for the W 4f7/2 and W 4f5/2 peaks, respectively.
Typical BEs for WN C reside around 32.1 and 34.1 eV for the
x
y
77
W 4f7/2 and W 4f5/2 peaks, respectively. W 4f peaks for WC1−x
and WN are separated by 0.2 eV and are generally considered
x
indistinguishable, so a single peak for WN C was assumed. It can
x
y
be seen that as deposition temperature increases W becomes
predominately bound as a WN C solid solution.
Oxygen is an impurity in the film and is likely incorporated by
post-growth exposure to air and/or residual oxygen and water
vapor in the reactor during film growth. Due to the amorphous
nature and low density of the films, significant in-diffusion of
oxygen is possible post-growth. The background O-source partial
pressure was estimated to be ∼35 mTorr (∼0.0001 mol fraction).
x
y
Deconvolution of the W 4f peak for the sample grown at
25 °C suggests that approximately 60% of the W is bound to O.
For the film at the 250 °C growth temperature, the oxide peaks
1
shift 0.5 to 35.2 and 37.2 eV (i.e., lower BE) for the W 4f7/2 and
1
656
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
O levels in the films reduce at higher growth temperature because
of sufficient thermal budget to replace O with C in the crystal
lattice, as well as the inhibition of post-growth incorporation due
to C stuffing in the lattice. The O 1s peak for adsorbed oxygen is
positioned at 531.6 eV, whereas those for WO and WO are
interfacial resistance due to electron scattering, which is
8
3,84
especially important to low resistance Cu contacts
in devices.
WN C Film Density Measurements. Films deposited
x
y
using 1 were analyzed with X-ray reflectivity (XRR) to determine
film densities as a function of deposition temperature. Figure 8
3
2
approximately 530.5 and 530.4 eV, respectively, so for simplicity,
they are considered as a single peak. At temperatures less than
3
00 °C, the O is bound nearly equally, 55 and 45%, as adventitious
oxide and WO , respectively. At temperatures above 300 °C, the
O signal is increasingly consistent with the BE of WO (58%),
x
x
whereas the adventitious oxide contribution is reduced to 42%.
It has been reported that using H or a H /NH as a carrier
2
2
3
51,61,78,79
gas reduces O incorporation.
Unfortunately, film growth
proved to be difficult under those conditions due to stable
precursor transport issues.
WN C Film Surface Roughness. Films deposited using 1
x
y
resulted in smooth films with root-mean-squared (RMS)
roughness values on average below 2.5 nm. At lower temper-
atures, the roughness is near that of the underlying Si substrate
Figure 8. Correlation between deposition temperature and measured
film density.
(
1.8 Å), as can be seen in Figure 7A. The sharp increase in
depicts this dependency and clearly shows film densification
with increasing deposition temperature, except at 650 °C.
This is typical of CVD type growth as transport rates in the solid
8
0,81
increase,
which facilitates increased crystallinity. Density
decreases at 650 °C when large amounts of amorphous C are
85
incorporated into the film. Amorphous C has a density between
3
2
and 3 g/cm , consistent with a decrease in film density upon
mixing with WN C . Typical bulk values for crystalline WN C
x y
x
y
8
6
3
have been reported at approximately 15.6 g/cm , whereas bulk
values for amorphous WN C have been reported between
and 13.1 g/cm . The relatively deflated density values can be
87−89
x
y
3
3
attributed to oxygen incorporation, which enhances the disorder
in the film. Despite the relatively low density values, it has been
shown that film microstructure is of more importance than film
90
density for diffusion barrier applications.
Diffusion Barrier Tests. The integrity of the films as
diffusion barriers was evaluated by depositing approximately
3
3
00 nm of Cu (according to the Experimental Section) onto
0 nm (per CS-SEM) of WN C film and annealing the Cu/
x
y
WN C /Si stack at 500 °C for 30 min. To detect large-scale
x
y
diffusion of Cu through the barriers, XRD and four point
probe (4PP) measurements were performed. Figure 9 illustrates
the changes in crystallinity between pre- and post-annealing of
Figure 7. (A) Average surface roughness for films deposited from 1 at
different deposition temperatures (black). The dotted horizontal line is
the RMS roughness of the underlying bare Si. (B) Atomic force
microscopy (AFM) micrograph for films deposited from 1 at 150 °C.
(
C) AFM micrograph for films deposited from 1 at 550 °C.
roughness at 175 °C could be indicative of a Stranski−Krastanov
growth mechanism, upon transitioning from a two-dimensional
to a three-dimensional growth. It has been shown that WN films
x
undergo compressive stresses, which facilitates island formation.
The surface roughness displayed an approximately linear
correlation to growth rate from 125 to 350 °C. This most likely
is the result of insufficient energy for adsorbed species (adatoms)
80
to overcome the surface diffusion barrier, so the increasing
81
growth rate propagates underlying features more quickly.
Between 350 and 450 °C, the surface roughness becomes nearly
80
constant, which is the result of adatom mobility saturation. At
higher temperatures, the surface mobility barrier is nonexistent,
82
which allows for the smoothing of films, as can be seen from
50 to 650 °C. This relatively smooth surface promotes
strong adhesion to neighboring layers and limits the amount of
Figure 9. XRD patterns depicting the crystallinity for pre- and post-
4
annealed Cu/WN C /Si material stacks with an annealed Cu/Si stack as
x
y
reference.
1
657
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
the samples. Peaks for Cu(111), Cu(200), and Cu(220) are
evident at 43.47, 50.65, and 74.28 2θ°, respectively. Cu(111) is
the preferred orientation for barrier applications, since it offers
the lowest electrical resistivity and highest electromigration
micrometer-sized inverted pyramidal features (appearing as dark
rectangles in Figure 10C) would be visible on the surfaces in the
plan-view images. The lack of inverted pyramidal features in
Figure 10A and B is more sensitive evidence for successful barrier
performance during annealing.
4
,91
resistance.
The ratio of the two most common peaks,
Cu(111) to Cu(200), is approximately 5:3, which increases to
5
:2 with annealing. These ratios are consistent with other reports
CONCLUSION
■
9
2,93
for deposition of Cu by MOCVD.
The Cu O peak that
2
In conclusion, we have prepared tungsten nitrido complexes of
appears during the annealing process results from background
oxygen and water vapor in the deposition chamber. The critical
peaks are the Cu Si (110) and Cu Si (103) peaks that appear
at 44.75 and 45.30 2θ°, respectively, in the annealed Cu/Si
stack but do not appear in either the pre- or post-annealed
Cu/WN C /Si stacks. These results indicate that barrier integrity
the type WN(NR ) [1−9 and 11; R = combinations of Me, Et,
2
3
i
n
Pr, Pr] as single source precursors for the deposition of WN C
x
y
3
3
films. Inclusion of the terminal nitrido ligand in place of the
imido or hydrazido ligands in previously reported precursors
dramatically lowered the film deposition temperature to 125 °C.
WN(NMe ) (1) proved to be a viable precursor for the deposi-
x
y
2
3
was maintained.
tion of robust WN C thin films for Cu diffusion barrier pur-
x
y
The onset of Cu Si precipitates can drastically increase the
3
poses at very low growth temperatures. Further investigation of
barrier materials grown from complexes 2−9 and 11 is currently
underway.
resistivity of the material stacks, as observed in the Cu/Si stack
upon annealing. The as-deposited and annealed Cu/Si stack
had resistivity measurements of 5.4 ± 0.7 and 384 ± 5.7 μΩ-cm,
respectively. In contrast, the resistivity values of the as-deposited
and annealed Cu/WN C /Si stacks were 6.8 ± 0.3 and 6.3 ±
EXPERIMENTAL SECTION
General Procedures. Unless otherwise specified, all manipulations
were performed under an inert atmosphere (N , Ar) using standard
Schlenk and glovebox techniques. Toluene, benzene, and hexane were
purified using an MBraun MB-SP solvent purification system and
stored over activated 3 Å molecular sieves prior to use. Diethyl ether and
tetrahydrofuran were dried using sodium/benzophenone, distilled, and
stored over activated 3 Å molecular sieves prior to use. Benzene-d6,
■
x
y
0
.4 μΩ-cm, respectively, which is consistent with no Cu diffusion
through the barrier film. The slight difference in resistivity
between the as-deposited Cu/WN C /Si and Cu/Si stacks can
2
x
y
be attributed to the more resistive barrier film (vs Cu) as well as
interfacial resistance.
An etch-pit test was also performed to provide insight into
small-scale Cu diffusion through the barrier. This method is
generally considered more sensitive than either resistivity or
XRD measurements. Figure 10 displays SEM plan-view images of
toluene-d , and THF-d8 were purchased from Cambridge Isotope
8
Laboratories, Inc., and stored over activated 3 Å molecular sieves
(15% w/v) for several days in an inert atmosphere glovebox prior to use.
Anhydrous tert-butanol, pentane, and redistilled diisopropylamine were
purchased from Sigma-Aldrich and used as received. Anhydrous 1,2-
dimethoxyethane, diglyme, and pyridine were purchased from Sigma-
Aldrich and stored over activated 3 Å molecular sieves for several days in
an inert atmosphere glovebox prior to use. Hexamethyldisiloxane,
diethylamine, dipropylamine, N-ethylmethylamine, and N-ethylisopro-
pylamine were purchased from Sigma-Aldrich or Alfa Aesar and stored
over activated 3 Å molecular sieves (15% w/v) for several days in an inert
atmosphere glovebox prior to use. Solid lithium diisopropylamide was
prepared from the reaction of butyllithium (1.6 M in hexanes) and
excess diisopropylamine (1.25 equiv.) and used shortly after
i
64
preparation. La(N Pr ) ·2THF was prepared using published
2
3
t
methods. WN(O Bu) and 1 were prepared by modified procedures
3
63
(
Supporting Information) adapted from a previous publication. All
1
13
other chemicals were purchased and used as received. H and C NMR
spectra were obtained using Gemini, Mercury, or VXR 300 MHz
spectrometers (Varian shims/probes; Oxford magnets) using the
residual protons of the deuterated solvents as reference peaks. Mass
spectra were obtained with a Thermo Scientific Trace GC DSQ mass
spectrometer using the DIP-CI mode of operation. Elemental analysis
results were attained from Complete Analysis Laboratories, Inc.
WN(NEt ) (2). In a 100 mL Schlenk flask, 2.22 g (6.72 mmol) of 1
2
3
and 40 mL of pentane were combined to create a white slurry. To the
slurry, 21 mL of diethylamine (15 g, 200 mmol) was added all at once to
give an orange/red suspension. A reflux condenser was attached, and the
reaction mixture was brought to a gentle reflux under argon. After 3 h,
the resulting light amber solution was returned to room temperature and
all volatiles were removed under vacuum. The remaining rusty/tan
residue was extracted by stirring with hexamethyldisiloxane (10 mL) for
15 min and filtered through Celite. The filter pad was washed with
additional hexamethyldisiloxane (2 × 5 mL). The resulting amber
filtrate was concentrated to saturation under vacuum and placed in the
glovebox refrigerator (−15 °C) overnight. The mother liquor was
removed, and the pure product was dried under vacuum to yield 2.37 g
Figure 10. (A) Plan-view of etch-pit test for pre-annealed Cu/WN C /
x
y
Si stack, (B) plan-view of etch-pit test for post-annealed Cu/WN C /Si
x
y
stack, and (C) plan-view of etch-pit test of a failed barrier for comparison
purposes. (D) Cross-sectional view of post-annealed Cu/WN C /Si stack.
x
y
etched films on both the pre- (Figure 10A) and post-annealed
Cu/WN C /Si stacks (Figure 10B), in addition to a cross-
1
x
y
(5.72 mmol) of pure 2 as white needle-like crystals (85%). H NMR
sectional image of a post-annealed Cu/WN C /Si stack showing
(C D , 300 MHz): δ = 1.16 [t, J = 7 Hz, 18 H, −N(CH CH ) ]; 3.58 [q,
x
y
6
6
2
3 2
13
the intact barrier (Figure 10D). In the event of barrier failure,
J = 7 Hz, 12 H, −N(CH CH ) ]. C NMR (C D , 300 MHz): δ = 17.8
2
3
2
6
6
1
658
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
i
[
[
3
−N(CH CH ) ]; 52.6 [−N(CH CH ) ]. DIP-CI-MS: calcd for
WN(NMe ) (NPr ) (6). In a 100 mL Schlenk flask, 0.25 g (0.76 mmol)
2 2 2
2
3
2
2
3
2
+
M + H] 415.2060; found 415.2062. C H N W (414.24): calcd C
of 1 and 60 mL of pentane were combined to create a white slurry.
To the slurry, 3.0 mL of diisopropylamine (2.2 g, 21 mmol) was added
to give an orange/red suspension. A reflux condenser was attached and
the mixture was brought to a gentle reflux under argon. After 12 h, the
resulting amber/brown solution was returned to room temperature and
solvent was removed under vacuum. The product was extracted by
stirring with hexamethyldisiloxane (30 mL) for 15 min and filtered
through Celite. The filter pad was washed with additional hexa-
methyldisiloxane (2 × 5 mL). The resulting brown filtrate was con-
centrated to saturation and left in the glovebox refrigerator (−15 °C)
overnight. The mother liquor was removed, and the product was dried
under vacuum to yield 0.19 g (0.49 mmol) of pure 6 as colorless
12
30
4
4.79, H 7.30, N 13.53; found C 34.76, H 7.19, N 13.60.
WN(N(Me)Et) (3). In a 10 mL Schlenk flask, 0.25 g (0.76 mmol) of
3
WN(NMe ) and 3.0 mL of N-ethylmethylamine (2.1 g, 35 mmol) were
2
3
combined to create an orange/red suspension. A reflux condenser was
attached, and the reaction mixture was brought to a gentle reflux under
argon. After 2.5 h, the resulting orange/amber solution was returned to
room temperature and all volatiles were removed under vacuum. The
rusty/tan residue was extracted by stirring with excess pentane (7 mL)
for 15 min and filtered through Celite. The filter pad was washed with
additional pentane (2 × 3 mL). All volatiles were removed from the
resulting amber filtrate under vacuum. The crude product was dissolved
in minimal hexamethyldisiloxane, concentrated to saturation, and placed
in the glovebox refrigerator (−15 °C) overnight. The mother liquor
1
microcrystals (65%). H NMR (C D , 300 MHz): δ = 1.31 [d, J = 6 Hz,
6
6
12 H, −N(CH(CH ) ) ]; 3.35 [s, 12 H, −N(CH ) ]; 3.41 [m, J = 6 Hz,
3
2
2
3 2
13
was removed, and the product was dried under vacuum to yield 0.15 g
2 H, −N(CH(CH ) ) ]. C NMR (C D , 300 MHz): δ = 26.4
3
2
2
6
6
1
(
0.40 mmol) of pure 3 as a tan microcrystalline powder (53%). H NMR
[−N(CH(CH ) ) ]; 51.3 [−N(CH ) ]; 51.9 [−N(CH(CH ) ) ]. DIP-
3
2
2
3
2
3 2 2
+
(
3
C D , 300 MHz): δ = 1.11 [t, J = 7 Hz, 9 H, −NCH (CH CH )];
CI-MS: calcd for [M + H] 387.1747; found 387.1750. C H N W
6
6
3
2
3
10 26 4
.35 [s, 9 H, −NCH (CH CH )]; 3.58 [q, J = 7 Hz, 6 H,
(386.19): calcd C 31.10, H 6.79, N 14.51; found C 30.93, H 6.83,
N 14.43.
WN(NMe )(NPr ) (7). In a 100 mL Schlenk flask, 0.50 g (1.5 mmol)
3
2
3
1
3
−
NCH (CH CH )]. C NMR (C D , 300 MHz): δ = 17.0
3
2
3
6
6
i
[
[
−NCH (CH CH )]; 47.3 [−NCH (CH CH )]; 57.4
3
2
3
3
2
3
2
2 2
+
−NCH (CH CH )]. DIP-CI-MS: calcd for [M + H] 373.1584;
of 1 and 10 mL of diisopropylamine (7.2 g, 71 mmol) were combined to
give an orange/red suspension. A reflux condenser was attached, and the
mixture was brought to a gentle reflux under argon. After 12 h, solvent
was removed from the dark brown solution under vacuum at room
temperature. The crude product was extracted by stirring with
hexamethyldisiloxane (50 mL) for 15 min and filtered through Celite.
The filter pad was washed with additional hexamethyldisiloxane
(2 × 5 mL). The resulting brown filtrate was concentrated to saturation
and left in the glovebox refrigerator (−15 °C) overnight. The mother
liquor was removed, and the product was dried under vacuum to yield
3
2
3
found 373.1599. C H N W (372.16): calcd C 29.05, H 6.50, N 15.05;
9
24
4
found C 28.91, H 6.43, N 15.01.
i
WN(N(Et) Pr) (4). In a 25 mL Schlenk flask, 0.21 g (0.64 mmol) of 1
3
and 4.0 mL (2.9 g, 33 mmol) of N-ethylisopropylamine were combined
to give an orange/red suspension (some white precipitate). A reflux
condenser was attached, and the reaction mixture was brought to a
gentle reflux under argon. After 4.5 h, the resulting dark brown solution
was returned to room temperature and all volatiles were removed under
vacuum to leave a tan/brown solid residue. The residue was extracted
by stirring with hexamethyldisiloxane (5 mL) for 15 min and filtered
through Celite. The filter pad was washed with additional hexa-
methyldisiloxane (2 × 3 mL). The resulting dark amber filtrate was
concentrated to saturation under vacuum and placed in the glovebox
refrigerator (−15 °C) overnight. The mother liquor was removed, and
the product was dried under vacuum to yield crude 4 as a light brown
powder (mostly pure by H NMR). The product was recrystallized with
minimal hexamethyldisiloxane to yield 0.10 g (0.22 mmol) of pure 4
as a light brown microcrystalline powder (34%). H NMR (C D ,
3
1
6
1
0.40 g (0.90 mmol) of pure 7 as colorless microcrystals (60%). H NMR
(C D , 300 MHz): δ = 1.29 [d, J = 6 Hz, 12 H, −N(CH-
6
6
(CH ) (CH ) ) ]; 1.37 [d, J = 6 Hz, 12 H, −N(CH(CH ) (CH ) ) ];
3
a
3
b
2
3
a
3 b 2
3.37 [s, 6 H, −N(CH ) ]; 3.45 [m, J = 6.4 Hz, 4 H, −N-
3
2
1
3
(CH(CH ) (CH ) ) ]. C NMR (C D , 300 MHz): δ = 26.2
3
a
3
b
2
6
6
[−N(CH(CH ) (CH ) ) ]; 27.2 [−N(CH(CH ) (CH ) ) ]; 51.4
3
a
3
b
2
3
a
3 b 2
1
[−N(CH ) ]; 52.3 [−N(CH(CH ) −(CH ) ) ]. DIP-CI-MS: calcd
3
2
3
a
3 b 2
+
for [M + H] 443.2368; found 443.2364. C H N W (442.29): calcd
14 34 4
1
C 38.02, H 7.75, N 12.67; found C 37.89, H 7.68, N 12.63
WN(NEt ) (NPr ) (8). In a 200 mL Schlenk flask, 0.50 g (1.5 mmol) of
6
6
i
00 MHz): δ = 1.05 [t, J = 7 Hz, 9 H, −NCH CH (CH(CH ) )];
2
3
3
2
2 2
2
.45 [d, J = 7 Hz, 18 H, −NCH CH (CH(CH ) )]; 3.20 [q, J = 7 Hz,
1 and 80 mL of pentane were combined to form a white slurry. To the
slurry, 6.0 mL of diisopropylamine (4.3 g, 43 mmol) was added to give
an orange/red suspension. A reflux condenser was attached, and the
mixture was brought to a gentle reflux under argon. After 12 h, solvent
was removed from the amber/brown solution under vacuum at room
temperature. The residue was extracted by stirring with hexamethyldi-
siloxane (50 mL) for 15 min and filtered through Celite. The filter
pad was washed with additional hexamethyldisiloxane (2 × 5 mL).
Solvent was removed from the resulting brown filtrate under vacuum
to give crude 6. This was dissolved in pentane (60 mL), and 3.0 mL of
diethylamine (2.1 g, 29 mmol) was added. A reflux condenser was
attached, and the mixture was brought to a gentle reflux under argon.
After 12 h, solvent was removed from the dark brown solution
under vacuum at room temperature. Sublimation of the crude residue
(60−65 °C/200−300 mtorr) afforded 0.36 g (0.83 mmol) of pure 8
2
3
3 2
H, −NCH CH −(CH(CH ) )]; 3.32 [m, J = 7 Hz, 3 H,
2
3
3 2
13
−
NCH CH (CH(CH ) )]. C NMR (C D , 300 MHz): δ = 19.0
2 3 3 2 6 6
[
−NCH CH (CH(CH ) )]; 27.2 [−NCH CH (CH(CH ) )]; 49.1
2
3
3
2
2
3
3 2
[
−NCH CH (CH(CH ) )]; 56.8 [−NCH CH (CH(CH ) )]. DIP-
2
3
3
2
2
3
3 2
+
CI-MS: calcd for [M + H] 457.2524; found 457.2516. C H N W
15
36
4
(
456.32): calcd C 39.48, H 7.95, N 12.28; found C 39.42, H 7.86,
N 12.19.n
WN(N Pr ) (5). In a 25 mL Schlenk flask, 0.25 g (0.76 mmol) of 1 and
2
3
1
3
2 mL of pentane were combined to create a white slurry. To the slurry,
.0 mL of dipropylamine (2.2 g, 22 mmol) was added all at once to give
an orange/red suspension. A reflux condenser was attached, and the
reaction mixture was brought to a gentle reflux under argon. After 18 h,
the resulting orange/red solution was returned to room temperature and
all volatiles were removed under vacuum to leave a tan solid residue.
The residue was extracted by stirring with hexamethyldisiloxane (7 mL)
for 15 min and filtered through Celite. The filter pad was washed with
additional hexamethyldisiloxane (2 × 3 mL). The resulting amber
filtrate was concentrated to saturation under vacuum and placed in the
glovebox refrigerator (−15 °C) overnight. The mother liquor was
1
as a light amber microcrystalline powder (55%). H NMR (C D ,
6
6
300 MHz): δ = 1.15 [t, J = 7 Hz, 12 H, −N(CH CH ) ]; 1.34 [d,
2
3 2
J = 6 Hz, 12 H, −N(CH(CH ) ) ]; 3.42 [m, J = 6 Hz, 2 H,
3
2 2
1
3
−N(CH(CH ) ) ]; 3.51−3.72 [m, J = 7 Hz, 8 H, N(CH CH ) ].
C
3
2
2
2
3 2
NMR (C D , 300 MHz): δ = 17.6 [−N(CH CH )]; 26.6 [−N(CH-
6
6
2
3
removed, and the pure product was dried under vacuum to yield 0.25 g
(CH ) ) ]; 51.2 [−N(CH CH ) ]; 52.7 [−N(CH(CH ) ) ]. DIP-CI-
3
2
2
2
3
2
3 2 2
1
+
(
0.50 mmol) of pure 5 as off-white needles (66%). H NMR (C D ,
MS: calcd for [M + H] 443.2373; found 443.2366. C H N W
6
6
14 34 4
3
00 MHz): δ = 0.90 [t, J = 7 Hz, 18 H, −N(CH CH CH ) ]; 1.64
(442.29): calcd C 38.02, H 7.75, N 12.67; found C 37.79, H 7.71,
N 12.58.
2
2
3 2
[
−
[
m, J = 7 Hz, 12 H, −N(CH CH CH ) ]; 3.62 [t, J = 7 Hz, 12 H,
2
2
3 2
1
3
i
N(CH CH CH ) ]. C NMR (C D , 300 MHz): δ = 12.0
WN(NEt )(NPr ) (9). In a 100 mL Schlenk flask, 0.50 g (1.5 mmol) of
2
2
3
2
6
6
2
2 2
−N(CH CH CH ) ]; 25.1 [−N(CH CH CH ) ]; 61.1 [−N-
CH CH CH ) ]. DIP-CI-MS: calcd for [M + H] 499.3000; found
1 and 10 mL of diisopropylamine (7.2 g, 71 mmol) were combined
to create an orange/red suspension. A reflux condenser was attached,
and the mixture was brought to a gentle reflux under argon. After 12 h,
solvent was removed from the dark brown solution under vacuum at
2
2
3
2
2
2
3 2
+
(
2
2
3 2
499.2991. C H N W (498.40): calcd C 43.38, H 8.49, N 11.24; found
18
42
4
C 43.26, H 8.39, N 11.13.
1
659
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
room temperature. The residue was extracted by stirring with
hexamethyldisiloxane (50 mL) for 15 min and filtered through Celite.
The filter pad was washed with additional hexamethyldisiloxane
Substrate Preparation. The p-type boron-doped Si(100) sub-
strates were acquired from Cemat Silicon and the n-type arsenic-doped
Si(111) substrates from Crysteco. The wafers were cut to 1 cm × 1 cm,
treated for 3 min each in boiling trichloroethylene, boiling acetone, and
boiling methanol to remove any organic contamination on the surface,
and subsequently rinsed for 30 s in boiling deionized water. Afterward,
(
2 × 5 mL). Solvent was removed from the resulting brown filtrate
under vacuum to give crude 7. This was dissolved in benzene (60 mL),
and 3.0 mL of diethylamine (2.1 g, 29 mmol) was added. A reflux
condenser was attached, and the mixture was brought to a gentle reflux
under argon. After 12 h, solvent was removed from the dark brown
solution under vacuum at room temperature. Sublimation of the crude
product (60−65 °C/200−300 mtorr) afforded 0.35 g (0.75 mmol) of
the substrates were placed in a buffered-oxide-etch (6:1 NH F:HF in
4
water) for 2 min to remove the native surface oxide and leave a relatively
air-stable hydrogen terminated surface. The samples were blown dry
using 99.999% N , wiped with an acetone soaked lint-free alpha swab,
1
2
pure 9 as a light amber microcrystalline powder (50%). H NMR (C D ,
6
6
and then blown dry again.
3
1
00 MHz): δ = 1.15 [t, J = 7 Hz, 6 H, −N(CH CH ) ]; 1.26 [d, J = 6 Hz,
2
3 2
Film Growth. A custom built aerosol-assisted chemical vapor
deposition reactor was used for the deposition studies, and is described
2 H, −N(CH(CH ) (CH ) ) ]; 1.33 [br. d, 12 H, −N(CH-
3
a
3 b 2
(
CH ) (CH ) ) ]; 3.46 [m, J = 6 Hz, 4 H, −N(CH(CH ) (CH ) ) ];
30
3
a
3
b
2
3
a
3 b 2
13
elsewhere. Precursor 1 was dissolved in pyridine at a concentration of
3
.62 [q, J = 7 Hz, 4 H, −N(CH CH ) ]. C NMR (C D , 300 MHz):
2
3
2
6
6
0
.051 M and delivered at a rate of 4 mL/h and transported with 99.999%
δ = 17.2 [−N(CH CH )]; 26.8 [−N(CH(CH ) (CH ) ) ]; 51.4
2
3
3
a
3 b 2
purity N to the reaction zone through heated lines at 55 °C. This
2
[
−N(CH CH ) ]; 52.7 [−N(CH(CH ) −(CH ) ) ]. DIP-CI-MS:
2
3
2
3
a
3
b
2
−6
+
corresponds to a 3.40 × 10 mol/min precursor delivery rate, or
calcd for [M + H] 471.2687; found 471.2705. C H N W (470.35):
16
38
4
−5
approximately a 7.47 × 10 gas-phase mol fraction of precursor. The
reactor operating pressure was maintained at 350 Torr, for bulk
materials studies, with deposition temperatures ranging from 125 to
650 °C. The deposition time was 150 min.
calcd C 40.86, H 8.14, N 11.91; found C 40.78, H 7.99, N 11.76.
t
i
WN(O Bu)(NPr ) (10). In a 200 mL Schlenk flask, 1.71 g (2.94
2
2
i
mmol) of La(N Pr ) ·2THF and 100 mL of THF were combined
to form a clear solution. To this solution, 1.75 g (4.19 mmol) of
WN(O Bu) was added all at once to create a golden-amber solution.
2
3
t
Materials Characterization. The materials characterization was
performed at the Major Analytical Instrumentation Center at the
University of Florida. To ensure a high signal-to-noise ratio, GIXRD was
employed in place of the traditional powder XRD. A Philips X’Pert
MRD system was used to perform the X-ray diffraction scans using Cu
Kα radiation with a scan rate of 0.03° 2θ per second. XRR measurements
were also performed using this instrument with density information
3
The reaction flask was heated to 60 °C under Ar (initial darkening of
the reaction mixture was observed, followed by an eventual return to
golden-amber). After 24 h, solvent was removed under vacuum at room
temperature to leave an amber residue. Pentane (50 mL) and diglyme
(
0.32 mL, 0.30 g, 2.2 mmol) were added, and the resulting slurry
(
off-white precipitate) was stirred for 1 h. The slurry was filtered
through Celite, and the filter pad was washed with additional pentane
2 × 10 mL). Solvent was removed from the resulting amber filtrate
under vacuum, and the solid amber residue was heated under vacuum
60−65 °C/200−300 mtorr) overnight. Sublimation (90−105 °C/
00−300 mtorr) of the crude material overnight afforded 0.94 g
derived from the critical angle, θ , in the θ−2θ scans. XPS was performed
c
(
using a Perkin-Elmer PHI 5100 ESCA System with an Al anode (hν =
1
486.3 eV) at 15 kV and 300 mA. The system had a nominal sputter rate
(
2
of approximately 4 Å/min, and collected data were deconvoluted using
the RBD Analysis suite of software. Pretreatment of samples involved
sputtering for 20 min to remove surface contamination. This sputtering
time was verified by collecting data after each minute of sputtering
until compositional profiles became steady. A VEECO Dimension 3100
Atomic Force Microscope (AFM) was used to perform root-mean-
square (RMS) surface roughness measurements. Typical measurements
included a scan rate of 2 Hz with a resolution of 512 × 512 lines for a
1
(
2.0 mmol) of pure 10 as a white microcrystalline powder (48%). H
NMR (25 °C, C D , 300 MHz): δ = 1.26 [br. s, 24 H, −N(CH-
6
6
(
−
6
CH ) ) ]; 1.52 [s, 9 H, −OC(CH ) ]; 3.54 [m, J = 6 Hz, 4 H,
3
2
2
3 3
1
N(CH(CH ) ) ]. H NMR (75 °C, C D , 300 MHz): δ = 1.27 [d, J =
3 2 2 6 6
Hz, 12 H, −N(CH(CH ) (CH ) ) ]; 1.34 [d, J = 6 Hz, 12 H,
3
a
3 b 2
−
N(CH(CH ) (CH ) ) ]; 1.51 [s, 9 H, -OC(CH ) ]; 3.61 [m, J = 6 Hz,
3 a 3 b 2 3 3
1
3
4
2
3
H, −N(CH(CH ) (CH ) ) ]. C NMR (75 °C, C D , 300 MHz): δ =
3
a
3
b
2
6
6
2
μm × 6 μm area. An FEI XL-40 FESEM was used for imaging purposes
5.7 [−N(CH(CH ) (CH ) ) ]; 26.3 [−N(CH(CH ) −(CH ) ) ];
3
a
3
b
2
3
a
3 b 2
with 20 kV as the accelerating voltage and 10 mm working distance.
Diffusion Barrier Testing. Complex 1 was used to deposit films for
Cu diffusion barrier applications. 300 nm of MOCVD Cu was deposited
2.6 [−OC(CH ) ]; 52.3 [−N(CH(CH ) (CH ) ) ]; 77.7
3
3
3
a
3 b 2
+
[−OC(CH ) ]. DIP-CI-MS: calcd for [M + H] 472.2527; found
3
3
4
72.2514. C H N OW (471.33): calcd C 40.77, H 7.91, N 8.92; found
16 37 3
onto freshly prepared (approximately 5 min of post-growth exposure
C 40.69, H 8.04, N 8.97.
i
I,II
WN(NPr ) (11). In a 100 mL pressure flask, 1.23 g (11.5 mmol) of
to air) WN C /Si using (hfac)Cu (TMVS). The Cu/WN C /Si stack
x
y
x
y
2
3
freshly prepared lithium diisopropylamide and 10 mL of diethyl ether
were combined to form a clear solution. To this solution, 0.83 g
was subsequently annealed at a temperature of 500 °C for 30 min under
a flow of 99.999% N . The annealed Cu/WN /Si stacks were then
C
x y
2
(
1.8 mmol) of 10 was added all at once to create a light amber solution.
analyzed by XRD, SEM, and 4PP. Etch-pit testing was performed by
removing the Cu layer with a dilute nitric acid solution, then removing
the barrier with a solution of NH OH, H O , and H O in a ratio of 1:1:4
The pressure flask was sealed and immediately heated to 65 °C (the
color of the reaction solution darkened initially and eventually became
dark red/amber with tan precipitate). After 24 h, the pressure flask was
returned to room temperature and solvent was removed under vacuum.
The crude residue was extracted by stirring with pentane (100 mL) for
4
2
2
2
for 10 min, followed by the etching of Cu Si for 5 s with Secco etchant
3
[
(K Cr O :H O, 1:1):HF, 1:1].
2 2 7 2
3
0 min and filtered through Celite. The filter pad was washed with
ASSOCIATED CONTENT
additional pentane (2 × 10 mL). The resulting dark red/amber filtrate
was concentrated to approximately 5 mL and placed in the glovebox
freezer (−15 °C) overnight to yield 0.62 g of crude 11 as tan powder
after removal of the mother liquor and drying under vacuum. The
product was further purified by sublimation (115−125 °C/300−400
mtorr) overnight to yield 0.50 g (1.0 mmol) of pure 11 as a white
■
*
S
Supporting Information
General experimental procedures. Modified procedures for
t
synthesis of WN(O Bu) and 1. Procedures for thermolysis of
3
1
1 and 2. Variable temperature H NMR spectra for 7, 9, 10, and
1
microcrystalline powder (57%). H NMR (25 °C, C D , 300 MHz):
11. X-ray experimental procedure, tables of crystal data and
structural refinement, tables of bond distances, bond angles,
positional parameters, and anisotropic displacement parameters
for 2 and 7. Experimental procedure for DFT calculations and
tables of optimized Cartesian coordinates for decomposition
pathways A and B. This material is available free of charge via the
Internet at http://pubs.acs.org.
6
6
δ = 1.33 [br. s, 36 H, −N(CH(CH ) ) ]; 3.50 [m, J = 6 Hz, 6 H,
3
2 2
1
−
N(CH(CH ) ) ]. H NMR (75 °C, C D , 300 MHz): δ = 1.34 [d,
3 2 2 6 6
J = 6 Hz, 36 H, −N(CH(CH ) ) ]; 3.56 [m, J = 6 Hz, 6 H,
3
2 2
13
−
N(CH(CH ) ) ]. C NMR (75 °C, C D , 300 MHz): δ = 26.4
3 2 2 6 6
[
−N(CH(CH ) ) ]; 51.7 [−N(CH(CH ) ) ]. DIP-CI-MS: calcd for
3
2
2
3 2 2
+
[
M + H] 499.3000; found 499.3027. C H N W (498.40): calcd
18
42
4
C 43.38, H 8.49, N 11.24; found C 43.43, H 8.56, N 11.28.
1
660
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
(
29) Bchir, O. J.; Green, K. M.; Hlad, M. S.; Anderson, T. J.; Brooks, B.
AUTHOR INFORMATION
Corresponding Author
lmwhite@chem.ufl.edu
Notes
■
C.; Wilder, C. B.; Powell, D. H.; McElwee-White, L. J. Organomet. Chem.
003, 684, 338.
30) Bchir, O. J.; Johnston, S. W.; Cuadra, A. C.; Anderson, T. J.; Ortiz,
2
(
C. G.; Brooks, B. C.; Powell, D. H.; McElwee-White, L. J. Cryst. Growth
2003, 249, 262.
The authors declare no competing financial interest.
(31) Bchir, O. J.; Green, K. M.; Ajmera, H. M.; Zapp, E. A.; Anderson,
ACKNOWLEDGMENTS
T. J.; Brooks, B. C.; Reitfort, L. L.; Powell, D. H.; Abboud, K. A.;
McElwee-White, L. J. Am. Chem. Soc. 2005, 127, 7825.
■
We would like to thank the National Science Foundation for
support under NSF Grant CHE-0911640. K.A.A. wishes to
acknowledge the National Science Foundation and the
University of Florida for funding of the purchase of the X-ray
equipment. We would also like to thank the Major Analytical
Instrumentation Center at the University of Florida for providing
and maintaining materials characterization equipment.
(
32) Dezelah, C. L.; El-Kadri, O. M.; Kukli, K.; Arstila, K.; Baird, R. J.;
Lu, J.; Niinisto, L.; Winter, C. H. J. Mater. Chem. 2007, 17, 1109.
33) Ajmera, H. M.; Anderson, T. J.; Koller, J.; McElwee-White, L.;
Norton, D. P. Thin Solid Films 2009, 517, 6038.
34) Kim, H. J. Vac. Sci. Technol., B 2003, 21, 2231.
(
(
(35) Ritala, M.; Kalsi, P.; Riihelae, D.; Kukli, K.; Leskelae, M.; Jokinen,
J. Chem. Mater. 1999, 11, 1712.
(36) Juppo, M.; Ritala, M.; Leskela, M. J. Electrochem. Soc. 2000, 147,
REFERENCES
3377.
■
(37) Cho, S. L.; Kim, K. B.; Min, S. H.; Shin, H. K.; Kim, S. D. J.
(
2
1) ITRS International Technology Roadmap for Semiconductors:
011 ed., International SEMATECH, 2011.
2) Lee, C.; Kuo, Y. L. JOM 2007, 59, 44.
3) Singer, P. Semicond. Int. 2006, 29, 32.
Electrochem. Soc. 1999, 146, 3724.
(38) Park, K. C.; Kim, K. B. J. Electrochem. Soc. 1995, 142, 3109.
(39) Park, K. C.; Kim, K. B. Mater. Res. Soc. Symp. Proc. 1995, 391, 211.
(40) Park, K. C.; Kim, K. B.; Raaijmakers, I.; Ngan, K. J. Appl. Phys.
(
(
(
4) Kaloyeros, A. E.; Eisenbraun, E. Annu. Rev. Mater. Sci. 2000, 30,
1
(
996, 80, 5674.
41) Casey, P.; Bogan, J.; McCoy, A.; Lozano, J. G.; Nellist, P. D.;
Hughes, G. J. Appl. Phys. 2012, 112.
363.
(
(
5) Rossnagel, S. M.; Hopwood, J. Appl. Phys. Lett. 1993, 63, 3285.
6) ITRS. Emerging Research Materials Difficult Challenges. [Online
(42) Sun, H. Q.; Qin, X. D.; Zaera, F. J. Phys. Chem. Lett. 2011, 2, 2525.
Early Access]. Published Online: 2011. http://www.itrs.net/Links/
011ITRS/2011Chapters/2011ERM.pdf.
7) Kafizas, A.; Carmalt, C. J.; Parkin, I. P. Coord. Chem. Rev. 2013, 257,
073.
8) Holloway, K.; Fryer, P. M.; Cabral, C., Jr.; Harper, J. M. E.; Bailey,
P. J.; Kelleher, K. H. J. Appl. Phys. 1992, 71, 5433.
(43) Matsumoto, K.; Neishi, K.; Itoh, H.; Sato, H.; Hosaka, S.; Koike, J.
2
(
2
(
Appl. Phys. Express 2009, 2.
44) Moon, D. Y.; Han, D. S.; Park, J. H.; Shin, S. Y.; Park, J. W.; Kim, B.
M.; Cho, J. Y. Thin Solid Films 2012, 521, 146.
45) Au, Y.; Lin, Y. B.; Kim, H.; Beh, E.; Liu, Y. Q.; Gordon, R. G. J.
Electrochem. Soc. 2010, 157, D341.
46) Lee, B. H.; Yong, K. J. Vac. Sci. Technol., B: Microelectron.
Nanometer Struct.–Process., Meas., Phenom. 2004, 22, 2375.
47) Suh, B. S.; Lee, Y. J.; Hwang, J. S.; Park, C. O. Thin Solid Films
999, 348, 299.
48) Uekubo, M.; Oku, T.; Nii, K.; Murakami, M.; Takahiro, K.;
Yamaguchi, S.; Nakano, T.; Ohta, T. Thin Solid Films 1996, 286, 170.
49) Koh, W.; Kumar, D.; Li, W. M.; Sprey, H.; Raaijmakers, I. J. Solid
State Technol. 2005, 48, 54.
50) Kim, S.-H.; Oh, S. S.; Kim, H.-M.; Kang, D.-H.; Kim, K.-B.; Li, W.-
M.; Haukka, S.; Tuominen, M. J. Electrochem. Soc. 2004, 151, C272.
51) Kim, D.; Kim, O. H.; Ajmera, H. M.; Anderson, T.; Koller, J.;
Abboud, K. A.; McElwee-White, L. J. Electrochem. Soc. 2011, 158, H618.
52) Gwildies, V.; Thiede, T. B.; Amirjalayer, S.; Alsamman, L.; Devi,
A.; Fischer, R. A. Inorg. Chem. 2010, 49, 8487.
53) Ajmera, H. M.; Heitsch, A. T.; Anderson, T. J.; Wilder, C. B.;
(
(
(
(
9) Chiu, H. T.; Chang, W. P. J. Mater. Sci. Lett. 1992, 11, 96.
10) Tsai, M. H.; Sun, S. C.; Chiu, H. T.; Tsai, C. E.; Chuang, S. H.
(
Appl. Phys. Lett. 1995, 67, 1128.
11) Min, K.-H.; Chun, K.-C.; Kim, K.-B. J. Vac. Sci. Technol., B 1996,
4, 3263.
12) Tsai, M. H.; Sun, S. C.; Tsai, C. E.; Chuang, S. H.; Chiu, H. T. J.
Appl. Phys. 1996, 79, 6932.
13) Cho, S.-L.; Kim, K.-B.; Min, S.-H.; Shin, H.-K. Mater. Res. Soc.
Symp. Proc. 1998, 514, 531.
14) Chen, X.; Peterson, G. G.; Goldberg, C.; Nuesca, G.; Frisch, H. L.;
Kaloyeros, A. E.; Arkles, B.; Sullivan, J. J. Mater. Res. 1999, 14, 2043.
15) Kaloyeros, A. E.; Chen, X. M.; Stark, T.; Kumar, K.; Seo, S.;
Peterson, G. G.; Frisch, H. L.; Arkles, B.; Sullivan, J. J. Electrochem. Soc.
999, 146, 170.
16) Chen, T.; Xu, C.; Baum, T. H.; Stauf, G. T.; Roeder, J. F.;
DiPasquale, A. G.; Rheingold, A. L. Chem. Mater. 2010, 22, 27.
17) Yun, J. Y.; Park, M. Y.; Rhee, S. W. J. Electrochem. Soc. 1999, 146,
804.
18) Cheung, N.; Vonseefeld, H.; Nicolet, M. A. J. Electrochem. Soc.
979, 126, C346.
(
1
(
(
1
(
(
(
(
(
(
(
(
1
(
(
Reitfort, L. L.; McElwee-White, L.; Norton, D. P. J. Vac. Sci. Technol., B
2008, 26, 1800.
(54) Rische, D.; Parala, H.; Gemel, E.; Winter, M.; Fischer, R. A. Chem.
Mater. 2006, 18, 6075.
(
1
(
1
(
19) Kurtz, S. R.; Gordon, R. G. Thin Solid Films 1986, 140, 277.
(55) Becker, J. S.; Gordon, R. G. Appl. Phys. Lett. 2003, 82, 2239.
(56) Anacleto, A. C.; Blasco, N.; Pinchart, A.; Marot, Y.; Lachaud, C.
Surf. Coat. Technol. 2007, 201, 9120.
(
(
20) Sherman, A. J. Electrochem. Soc. 1990, 137, 1892.
21) Winter, C. H.; Sheridan, P. H.; Lewkebandara, T. S.; Heeg, M. J.;
Proscia, J. W. J. Am. Chem. Soc. 1992, 114, 1095.
(57) Li, W.-M. Chem. Vap. Deposition 2013, 19, 82.
(58) Zeng, W.; Wang, X.; Kumar, S.; Peters, D. W.; Eisenbraun, E. T. J.
Mater. Res. 2007, 22, 703.
(59) McElwee-White, L.; Koller, J.; Kim, D.; Anderson, T. J. ECS Trans.
2009, 25, 161.
(60) McElwee-White, L. Dalton Trans. 2006, 5327.
(61) Potts, S. E.; Carmalt, C. J.; Blackman, C. S.; Leese, T.; Davies, H.
O. Dalton Trans. 2008, 5730.
(62) Koller, J.; Ajmera, H. M.; Abboud, K. A.; Anderson, T. J.;
McElwee-White, L. Inorg. Chem. 2008, 47, 4457.
(63) McClain, K. R.; O’Donohue, C.; Shi, Z.; Walker, A. V.; Abboud, K.
A.; Anderson, T.; McElwee-White, L. Eur. J. Inorg. Chem. 2012, 2012,
4579.
(
(
22) Musher, J. N.; Gordon, R. G. J. Electrochem. Soc. 1996, 143, 736.
23) Min, J.-S.; Son, Y.-W.; Kang, W.-G.; Chun, S.-S.; Kang, S.-W. Jpn.
J. Appl. Phys., Part 1 1998, 37, 4999.
24) Kim, S. H.; Kim, J. K.; Lee, J. H.; Kwak, N.; Kim, J.; Jung, S. H.;
Hong, M. R.; Lee, S. H.; Collins, J.; Sohn, H. J. Electrochem. Soc. 2007,
54, D435.
25) Nakajima, T.; Watanabe, K.; Watanabe, N. J. Electrochem. Soc.
987, 134, 3175.
(
1
(
1
(
26) Chiu, H. T.; Chuang, S. H. J. Mater. Res. 1993, 8, 1353.
(
27) Kelsey, J. E.; Goldberg, C.; Nuesca, G.; Peterson, G.; Kaloyeros, A.
E.; Arkles, B. J. Vac. Sci. Technol., B 1999, 17, 1101.
28) Becker, J. S.; Suh, S.; Wang, S.; Gordon, R. G. Chem. Mater. 2003,
5, 2969.
(
1
(64) Aspinall, H. C.; Tillotson, M. R. Polyhedron 1994, 13, 3229.
1
661
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662

Journal of the American Chemical Society
Article
(
65) Kumar, N.; Yang, Y.; Noh, W.; Girolami, G. S.; Abelson, J. R.
Chem. Mater. 2007, 19, 3802.
66) Clough, C. R.; Greco, J. B.; Figueroa, J. S.; Diaconescu, P. L.;
Davis, W. M.; Cummins, C. C. J. Am. Chem. Soc. 2004, 126, 7742.
67) DiFranco, S. A.; Maciulis, N. A.; Staples, R. J.; Batrice, R. J.; Odom,
A. L. Inorg. Chem. 2011, 51, 1187.
(
(
(
68) Firman, T. K.; Landis, C. R. J. Am. Chem. Soc. 2001, 123, 11728.
(
69) Lewkebandara, T. S.; Sheridan, P. H.; Heeg, M. J.; Rheingold, A.
L.; Winter, C. H. Inorg. Chem. 1994, 33, 5879.
70) Amato, C. C.; Hudson, J. B.; Interrante, L. V. Mater. Res. Soc.
Symp. Proc. 1990, 168, 119.
71) Chae, Y.; Houf, W. G.; McDaniel, A. H.; Troup, J.; Allendorf, M.
D. J. Electrochem. Soc. 2004, 151, C527.
72) Chae, Y.; Houf, W. G.; McDaniel, A. H.; Allendorf, M. D. J.
Electrochem. Soc. 2006, 153, C309.
73) Sun, S. C. In Solid-State and Integrated Circuit Technology
(
(
(
(
Proceedings; IEEE, 1998; pp 243−246.
(
74) Choy, K. L. Prog. Mater. Sci. 2003, 48, 57.
(
75) Chen, H.; Yang, Y.; Hu, Z.; Huo, K. F.; Ma, Y. W.; Chen, Y.; Wang,
X. S.; Lu, Y. N. J. Phys. Chem. B 2006, 110, 16422.
(
76) Kreuzer, H. J. J. Chem. Phys. 1996, 104, 9593.
(
77) NIST X-ray Photoelectron Spectroscopy Database, Version 4.1;
National Institute of Standards and Technology: Gaithersburg, MD,
012; http://srdata.nist.gov/xps/.
78) Ajmera, H. M.; Heitsch, A. T.; Bchir, O. J.; Anderson, T. J.;
Reitfort, L. L.; McElwee-White, L. J. Electrochem. Soc. 2008, 155, H829.
79) Kim, D.; Kim, O. H.; Anderson, T.; Koller, J.; McElwee-White, L.;
Leu, L. C.; Tsai, J. M.; Norton, D. P. J. Vac. Sci. Technol., A 2009, 27, 943.
80) Karpenko, O. P.; Yalisove, S. M.; Eaglesham, D. J. J. Appl. Phys.
997, 82, 1157.
81) Chevrier, J.; Cruz, A.; Pinto, N.; Berbezier, I.; Derrien, J. J. Phys. I
994, 4, 1309.
82) Takahashi, K.; Nakayama, M.; Yokoyama, S.; Kimura, T.;
Tokumitsu, E.; Funakubo, H. Appl. Surf. Sci. 2003, 216, 296.
83) Abom, A. E.; Haasch, R. T.; Hellgren, N.; Finnegan, N.; Hultman,
L.; Eriksson, M. J. Appl. Phys. 2003, 93, 9760.
84) Israelachvili, J.; Maeda, N.; Rosenberg, K. J.; Akbulut, M. J. Mater.
Res. 2005, 20, 1952.
85) Ferrari, A. C.; Libassi, A.; Tanner, B. K.; Stolojan, V.; Yuan, J.;
Brown, L. M.; Rodil, S. E.; Kleinsorge, B.; Robertson, J. Phys. Rev. B
000, 62, 11089.
86) Suetin, D. V.; Shein, I. R.; Ivanovskii, A. L. J. Struct. Chem. 2009,
0, 1.
87) Hoyas, A. M.; Travaly, Y.; Schuhmacher, J.; Sajavaara, T.; Whelan,
2
(
(
(
1
(
1
(
(
(
(
2
(
5
(
C. M.; Eyckens, B.; Richard, O.; Giangrandi, S.; Brijs, B.; Vandervorst,
W.; Maex, K.; Celis, J. P.; Jonas, A. M.; Vantomme, A. J. Appl. Phys. 2006,
99.
(
(
88) Marcus, S. D.; Foster, R. F. Thin Solid Films 1993, 236, 330.
89) Volders, H.; Tokei, Z.; Bender, H.; Brijs, B.; Caluwaerts, R.;
Carbonell, L.; Conard, T.; Drijbooms, C.; Franquet, A.; Garaud, S.;
Hoflijk, I.; Moussa, A.; Sinapi, F.; Travaly, Y.; Vanhaeren, D.; Vereecke,
G.; Zhao, C.; Li, W. M.; Sprey, H.; Jonas, A. M. Microelectron. Eng. 2007,
8
(
4, 2460.
90) Kim, S. H.; Chung, D. S.; Park, K. C.; Kim, K. B.; Min, S. H. J.
Electrochem. Soc. 1999, 146, 1455.
(
91) Shacham-Diamand, Y. J. Electron. Mater. 2001, 30, 336.
(
92) Yang, D.; Hong, J.; Richards, D.; Cale, T. J. Vac. Sci. Technol., B
2
(
002, 20, 495.
93) Kroger, R.; Eizenberg, M.; Cong, D.; Yoshida, N.; Chen, L.;
Ramaswami, S.; Carl, D. J. Electrochem. Soc. 1999, 146, 3248.
1
662
dx.doi.org/10.1021/ja4117582 | J. Am. Chem. Soc. 2014, 136, 1650−1662