Reductive elimination of hypersilyl halides from zinc(II) complexes. implications for electropositive metal thin film growth

Article abstract of DOI:10.1021/ic502184f

Treatment of Zn(Si(SiMe₃)₃)₂ with ZnX₂ (X = Cl, Br, I) in tetrahydrofuran (THF) at 23 °C afforded [Zn(Si(SiMe₃)₃)X(THF)]₂ in 83-99% yield. X-ray crystal structures revealed dimeric structures with Zn₂X₂ cores. Thermogravimetric analyses of [Zn(Si(SiMe₃)₃)X(THF)]₂ demonstrated a loss of coordinated THF between 50 and 155°C and then single-step weight losses between 200 and 275°C. The nonvolatile residue was zinc metal in all cases. Bulk thermolyses of [Zn(Si(SiMe₃)₃)X(THF)]₂ between 210 and 250°C afforded zinc metal in 97-99% yield, Si(SiMe₃)₃X in 91-94% yield, and THF in 81-98% yield. Density functional theory calculations confirmed that zinc formation becomes energetically favorable upon THF loss. Similar reactions are likely to be general for M(SiR₃)_n/MX_n pairs and may lead to new metal-film-growth processes for chemical vapor deposition and atomic layer deposition.

Full text of DOI:10.1021/ic502184f

Communication
pubs.acs.org/IC
Reductive Elimination of Hypersilyl Halides from Zinc(II) Complexes.
Implications for Electropositive Metal Thin Film Growth
Chatu T. Sirimanne,† Marissa M. Kerrigan,† Philip D. Martin,† Ravindra K. Kanjolia,‡ Simon D. Elliott,§
and Charles H. Winter*,†
†Department of Chemistry, Wayne State University, Detroit, Michigan 48202, United States
‡SAFC Hitech, 1429 Hilldale Avenue, Haverhill, Massachusetts 01832, United States
§Tyndall National Institute, University College Cork, Lee Maltings, Cork, Ireland
S
* Supporting Information
ruthenium substrate to decompose BH3(NHMe2) to more a
ABSTRACT: Treatment of Zn(Si(SiMe3)3)2 with ZnX2
(X = Cl, Br, I) in tetrahydrofuran (THF) at 23 °C afforded
[Zn(Si(SiMe3)3)X(THF)]2 in 83−99% yield. X-ray crystal
structures revealed dimeric structures with Zn2X2 cores.
Thermogravimetric analyses of [Zn(Si(SiMe3)3)X(THF)]2
demonstrated a loss of coordinated THF between 50 and
155 °C and then single-step weight losses between 200
and 275 °C. The nonvolatile residue was zinc metal in all
cases. Bulk thermolyses of [Zn(Si(SiMe3)3)X(THF)]2
between 210 and 250 °C afforded zinc metal in 97−99%
yield, Si(SiMe3)3X in 91−94% yield, and THF in 81−98%
yield. Density functional theory calculations confirmed
that zinc formation becomes energetically favorable upon
THF loss. Similar reactions are likely to be general for
M(SiR3)n/MXn pairs and may lead to new metal-film-
growth processes for chemical vapor deposition and
atomic layer deposition.
reactive reducing species, and the growth stopped once the film
covered the ruthenium surface. Elemental zinc films have been
grown by chemical vapor deposition (CVD) at ≥150 °C by
thermal decomposition of bis(allyl)zinc.6 Thermal decomposi-
tion of ZnEt2 to zinc metal occurs below 127 °C on many
surfaces.7 Additionally, the use of ZnEt2 as a reducing agent in
ALD growth with Cu(OCHMeCH2NMe2)2 affords Cu−Zn
alloys at ≥120 °C through parasitic thermal decomposition of
ZnEt2 to zinc metal.8 However, no ALD processes for zinc metal
have been reported because of a lack of appropriate chemical
precursors.
Recently, ALD growth of antimony thin films was
demonstrated using SbCl3 and Sb(SiEt3)3.9 This process
proceeds by elimination of SiEt3Cl and is a totally new approach
for ALD of element films, in this case a nonmetal.9 Related
reactions could represent powerful methodologies for ALD
growth of metallic first-row transition- and electropositive-metal
films and could avoid potential problems associated with H2 and
highly reactive hydride reagents.1,5 Within this context, we
describe a series of reactions that occur upon treatment of
bis(hypersilyl)zinc, Zn(Si(SiMe3)3)2,11 with ZnX2 (X = Cl, Br, I)
ultimately to afford zinc metal and Si(SiMe3)3X. These results
demonstrate that reductive eliminations of silyl halides from
zinc(II) centers are facile, thus suggesting new, potentially
general chemistry for the growth of metal films. Significantly, the
electrochemical potential of the Zn2+ ion (E° = −0.7618 V10) is
significantly more negative than that of the Sb3+ ion (E° = 0.152
V10), implying that silyl halide reductive eliminations may
provide general access to many electropositive metals.
hin films of metals have many important applications in
1
Tmicroelectronics and magnetic devices.
Atomic layer
deposition (ALD) is a film-growth technique that provides
conformal coverage of nanoscale features and subnanometer
control over film thicknesses because of its self-limited growth
mechanism.2 In many future devices, transition metal and other
metal films will need to be grown by ALD to meet performance
requirements.1−3 Such ALD growth requires a rapid, comple-
mentary reaction between a metal precursor and a reducing
coreagent to afford the metal. Considerable progress has been
made on the ALD growth of metallic copper and noble metal
films1−3 largely because of the positive electrochemical potentials
(E°) of ions of these metals and attendant ease of reduction to
the metals with a range of reagents.1−3 By contrast, most other
metal ions have negative E° values and are much more difficult to
reduce to metals.1 As such, thermal ALD growth of these metals
is difficult and remains poorly developed.1 H2 has been the most
commonly used as the reducing coreagent to date in ALD,1,4 but
many metal ions have low reactivities toward H2 at desired ALD
growth temperatures of ≤200 °C.1 We recently reported that
BH3(NHMe2) serves as a powerful reducing coreagent for
nickel(II), cobalt(II), iron(II), chromium(II), and possibly
manganese(II) α-imino alkoxide precursors in the ALD growth
of these metals at ≤200 °C.5 However, these processes required a
Zn(Si(SiMe3)3)2 was prepared according to a literature
procedure11a and treated with 1 equiv of ZnX2 (X = Cl, Br, I)
in tetrahydrofuran (THF; Scheme 1). No color changes or
precipitation occurred upon mixing at ambient temperature or
refluxing for 18 h. Analysis of these reaction mixtures after
stirring at 23 °C for 3 h indicated the formation of products 1−3,
as outlined in Scheme 1. Complexes 1−3 were isolated by
crystallization and were characterized by spectral and analytical
data and by X-ray crystallography.12 The 1H and 13C{1H} NMR
spectra revealed the expected resonances for the Si(SiMe3)3 and
THF ligands, and the former resonances were shifted from those
Received: September 8, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ic502184f | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Communication
benzene-d6, decanting of the extracts, and vacuum drying of the
gray powders. Powder X-ray diffraction revealed the gray
powders to be zinc metal, with isolated yields of 96−99%.12 1H
and 13C{1H} NMR spectra of the extracts showed that the
soluble products were Si(Si(CH3)3)3X (91−94%) and THF
(81−98%), based on a comparison with the NMR data of
authentic samples.12
Scheme 1. Synthesis and Thermolysis of 1−3
Density functional theory (DFT) calculations using VASP14
and TURBOMOLE15 were used to predict the structures of
intermediates and quantify the changes in bonding, as detailed in
the Supporting Information.12 The gas-phase compounds
[Zn(Si(SiMe3)3)Cl]n increase in stability upon oligomerization,
with n = 4 predicted to be the most stable [ΔE = −49.1 kJ/mol of
zinc relative to ZnCl2 and Zn(Si(SiMe3)3)2]. The tetramer is
predicted to contain a cubic Zn4Cl4 core. The X-ray crystal
structure of [Zn(Si(tBu)3)Br]4 has been reported16 and exists
with a cubic Zn4Br4 core, thus supporting the DFT calculations of
[Zn(Si(SiMe3)3)Cl]4. However, it is not clear if the tetramer is
accessible after THF loss in the current experiments.
Thermolysis of solid 2 at 111 °C and 0.05 Torr for 24 h in a
drying tube afforded a THF-free complex of the apparent
formula [Zn(Si(SiMe3)3)Br]x;12 however, X-ray-quality crystals
could not be obtained despite multiple attempts, and the
compound was not pursued further.
of Zn(Si(SiMe3)3)2. Complex 1 was previously reported using a
different synthetic method.13 Complexes 1−3 have similar
dimeric structures, with one Si(SiMe3)3 and one THF ligand per
zinc ion and two halide ions that bridge between the zinc ions.
The Si(SiMe3)3 groups are anti to each other within the dimers,
to avoid steric crowding. Figure 1 shows a perspective view of 2.
The Zn−Br, Zn−Si, and Zn−O bond lengths are 2.5030(6) and
2.5382(6), 2.352(1), and 2.101(3) Å, respectively.
The DFT calculations revealed that the formation of THF
adducts from [Zn(Si(SiMe3)3)Cl]n is thermodynamically
favorable for n = 1−3. For formation from ZnCl2, Zn(Si-
(SiMe3)3)2, and THF, the dimer [Zn(Si(SiMe3)3)Cl(THF)]2 is
computed to be the most stable (ΔE = −68.4 kJ/mol of zinc
neglecting entropy), consistent with the isolation of 1−3.
However, the trimer [Zn(Si(SiMe3)3)Cl(THF)]3 is predicted to
be only <2 kJ/mol less stable than the dimer. The calculated Zn−
Cl (2.40−2.43 Å) and Zn−Si (2.39 Å) distances in the dimer are
within 2% of those observed in the X-ray crystal structure of 1
[2.391(2) and 2.347(2) Å], which is within the accuracy of DFT.
We next used DFT to predict the thermodynamic stabilities of
various zinc complexes toward elimination of Si(SiMe3)3X.
Consistent with the high stabilities of 1−3 documented above,
the THF-coordinated dimer 1 is thermodynamically stable by
+22 kJ/mol of zinc at 0 K with respect to the reductive
elimination of Si(SiMe3)3Cl and the production of zinc metal.
However, a loss of THF ligands causes the resulting THF-free
[Zn(Si(SiMe3)3)Cl]2 to become metastable by −12.5 kJ/mol of
zinc toward the reductive elimination of Si(SiMe3)3Cl and the
formation of zinc metal. Because entropy has been neglected, this
energy difference (+22 → −12.5 kJ/mol) reflects metastable
bonding within the complex. Similar switches to metastability
upon a loss of THF ligands are computed for the monomer
Zn(Si(SiMe3)3)Cl(THF) (+14.4 → −45.9 kJ/mol of zinc) and
trimer [Zn(Si(SiMe3)3)Cl(THF)]3 (+20.9 → −12.5 kJ/mol of
zinc). The metastability toward reductive elimination is likely
due to lower coordination numbers at zinc upon THF loss.
These observations are consistent with experiment because the
TGA data reveal that THF loss from 1−3 occurs prior to
decomposition. By contrast, reductive elimination from the
tetramer [Zn(Si(SiMe3)3)Cl]4 is not favored, apparently because
each zinc ion is four-coordinate and thus coordinatively
saturated, although the small energy cost (ΔE = +2.9 kJ/mol
of zinc) can probably be overcome at elevated temperatures.
The overall reaction Zn(Si(SiMe3)3)2 + ZnCl2 → 2Zn +
2Si(SiMe3)3Cl is computed by DFT to show ΔE = −46 kJ/mol
of zinc and ΔG = −52 kJ/mol of zinc at T = 100 °C, indicating
that it is a thermodynamically viable route to the ALD of zinc.
Figure 1. Perspective view of 2 with thermal ellipsoids at the 50% level.
Selected bond lengths (Å) and angles (deg): Zn−Br 2.5030(6), Zn−Br′
2.5382(6), Zn−Si2 2.352(1), Zn−O1 2.101(3); Zn−Br−Zn′ 81.16(2),
Br−Zn−Br′ 93.84(2), Br−Zn−Si2 124.52(3), Br−Zn′−Br′ 121.70(3),
Br−Zn−O 96.88(9), Br′−Zn−O 94.56(9).
Thermogravimetric analyses (TGA) of 1−3 were carried out
to understand their solid-state decompositions. The THF ligands
are labile, and their loss was observed starting at 50 °C and was
complete by 155 °C. Then, single-step weight losses occurred
between 200 and 275 °C to afford 11.67, 6.02, and 3.84% weight
residues for 1−3, respectively. The percentages of zinc in 1−3 are
15.55, 14.06, and 12.77%, respectively. The lower observed
residues, compared to those predicted by the percentage of zinc
in 1−3, may originate from partial sublimation of 1−3 or the
THF-free products thereof.
Encouraged by the TGA data, we next explored solid-state
thermolyses of 1−3.12 Pressure tubes fitted with Teflon stoppers
were charged with ∼0.4 mmol each of 1−3 and hexamethyl-
benzene as an internal standard. These tubes were then heated at
210 (1), 225 (2), and 250 °C (3) for 4 h. During the thermolyses,
the white solids gradually transformed to granular gray solids and
colorless liquids condensed at the cool area near the Teflon
stoppers. Workup entailed extraction of the flask contents with
B
dx.doi.org/10.1021/ic502184f | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Communication
This prediction is consistent with the observed decompositions
of 1−3 upon thermolysis. For reference, analogous DFT
calculations of the ALD reaction for the deposition of antimony
ACKNOWLEDGMENTS
■
We are grateful to the U.S. National Science Foundation (Grant
CHE-1212574), Science Foundation Ireland (Grant 09/IN.1/
I2628), and SAFC Hitech for support of this research.
9
from SbCl3 and Sb(SiEt3)3 predict a favorable ΔE value of −140
kJ/mol of antimony.12 The highest occupied molecular orbital of
Zn(Si(SiMe3)3)2 is found to be Si−Zn−Si σ bonding, and this is
the ultimate source of electrons that reduce the zinc ion to
metallic form. The [Zn(Si(SiMe3)3)Cl]n clusters also show 2n
electrons in high-lying σ(Zn−Si) orbitals that become available
for the reduction of zinc when the Zn−Si bond is broken. At the
same time, the ligating silicon atoms of the hypersilyl groups
become oxidized. The DFT calculations reveal a similar role for
σ(Sb−Si) bonding orbitals in Sb(SiEt3)3.
REFERENCES
■
(1) (a) Knisley, T. J.; Kalutarage, L. C.; Winter, C. H. Coord. Chem. Rev.
2013, 257, 3222−3231. (b) Ramos, K. B.; Saly, M. J.; Chabal, Y. J. Coord.
Chem. Rev. 2013, 257, 3271−3281. (c) Emslie, D. J. H.; Chadha, P.;
Price, J. S. Coord. Chem. Rev. 2013, 257, 3282−3296. (d) Leskela, M.;
̈
Ritala, M.; Nilsen, O. MRS Bull. 2011, 36, 877−884.
(2) (a) George, S. M. Chem. Rev. 2010, 110, 111−131. (b) Leskela, M.;
̈
Ritala, M. Angew. Chem., Int. Ed. 2003, 42, 5548−5554. (c) Putkonen,
M.; Niinisto, L. Top. Organomet. Chem. 2005, 9, 125−145.
̈
This work has several implications for the growth of metal
films by CVD and ALD using silyl halide elimination reactions.
Most importantly, reductive elimination of Si(SiMe3)3X is
predicted to be energetically favorable for the zinc(II) ion,
although no prediction about kinetics can be made without
knowledge of a reaction pathway in solution or during ALD. The
formation of zinc metal from 1−3 demonstrates favorable
reactions at 210−250 °C. The metastable unsolvated adducts
[ZnSi(SiMe3)3Cl]n are representative of structures that might
form on the growing surface in a potential film growth process
using ZnCl2 and Zn(Si(SiMe3)3)2, which implies that the present
work may lead to a zinc metal ALD process. In this vein,
Zn(Si(SiMe3)3)2 sublimes at 110 °C and 0.05 Torr, decomposes
thermally at about 350 °C, and thus has excellent ALD precursor
properties.12 Virtually all ALD processes for zinc-containing
films use ZnEt2 as a precursor,17 and our calculations predict ΔE
= −103 kJ/mol of zinc for the reaction Zn(Si(SiMe3)3)2 + ZnEt2
→ 2Zn + 2Si(SiMe3)3Et.12 This is even more exothermic than
the formation of zinc metal from ZnCl2 and Zn(Si(SiMe3)3)2 and
may be a promising ALD approach if the depositions proceed at
or below the thermal decomposition temperature of ZnEt2
(∼130 °C).7,8 The favorable energetics with ZnEt2 suggest that
precursors other than metal halides can likely be used. However,
more volatile and thermally stable zinc precursors need to be
developed. Finally, zinc is an excellent model for first-row
transition metals because its metal radius is about the same as
those of vanadium and chromium, the E° value of the zinc(II) ion
(E° = −0.74 V10) lies between those of iron(II) (E° = −0.44 V10)
and chromium(II) (E° = −0.94 V10), and the coordination
chemistry of zinc(II) is similar to first-row transition-metal(II)
(3) Hamalainen, J.; Ritala, M.; Leskela, M. Chem. Mater. 2014, 26,
786−801.
̈
̈
̈
̈
(4) Lim, B. S.; Rahtu, A.; Gordon, R. G. Nat. Mater. 2003, 2, 748−754.
(5) Kalutarage, L. C.; Martin, P. D.; Heeg, M. J.; Winter, C. H. J. Am.
Chem. Soc. 2013, 135, 12588−12591.
(6) Cheon, J.; Dubois, L. H.; Girolami, G. S. Chem. Mater. 1994, 6,
2279−2287.
(7) (a) Rueter, M. A.; Vohs, J. M. Surf. Sci. 1992, 262, 42−50.
(b) Kovacs, I.; Iost, N.; Solymosi, F. J. Chem. Phys. 1994, 101, 4236−
4247.
(8) (a) Vidjayacoumar, B.; Emslie, D. J. H.; Clendenning, S. B.;
Blackwell, J. M.; Britten, J. F.; Rheingold, A. Chem. Mater. 2010, 22,
4844−4853. (b) Lee, B. H.; Hwang, J. K.; Nam, J. W.; Lee, S. U.; Kim, J.
T.; Koo, S.-M.; Baunemann, A.; Fischer, R. A.; Sung, M. M. Angew.
Chem., Int. Ed. 2009, 48, 4536−4539.
(9) Pore, V.; Knapas, K.; Hatanpaa, T.; Sarnet, T.; Kemell, M.; Ritala,
̈
̈
M.; Leskela, M.; Mizohata, K. Chem. Mater. 2011, 23, 247−254.
̈
(10) Handbook of Chemistry and Physics, 92nd ed.; CRC Press: Boca
Raton, FL, 2011−2012; pp 5-80−5-89, http://www.hbcpnetbase.com/.
(11) (a) Arnold, J.; Tilley, T. D.; Rheingold, A. L.; Geib, S. J. Inorg.
Chem. 1987, 26, 2106−2109. (b) Gaderbauer, W.; Balatoni, I.; Wagner,
H.; Baumgartner, J.; Marschner, C. Dalton Trans. 2010, 39, 1598−1603.
(12) See the Supporting Information.
(13) Nanjo, M.; Oda, T.; Mochida, K. J. Organomet. Chem. 2003, 672,
100−108.
(14) Kresse, G.; Furthmuller, J. Phys. Rev. B 1996, 54, 11169.
̈
(15) Ahlrichs, R.; Bar, M.; Haser, M.; Horn, H.; Kolmel, C. Chem. Phys.
Lett. 1989, 162, 165.
̈
̈
̈
(16) Wiberg, N.; Amelunxen, K.; Lerner, H.-W.; Noth, H.; Appel, A.;
Knizek, J.; Polborn, K. Z. Anorg. Allg. Chem. 1997, 623, 1861−1870.
(17) (a) Malm, J.; Sahramo, E.; Perala, J.; Sajavaara, T.; Karppinen, M.
̈
̈
̈
Thin Solid Films 2011, 519, 5319−5322. (b) Makino, H.; Miyake, A.;
Yamada, T.; Yamamoto, N.; Yamamoto, T. Thin Solid Films 2009, 517,
3138−3142.
(18) (a) Nakamoto, M.; Yamasaki, T.; Sekiguchi, A. J. Am. Chem. Soc.
2005, 127, 6965−6955. (b) Heyn, R. H.; Tilley, T. D. Inorg. Chim. Acta
2002, 341, 91−98. (c) Farwell, J. D.; Lappert, M. F.; Marschner, C.;
Strissel, C.; Tilley, T. D. J. Organomet. Chem. 2000, 603, 185−188.
(d) Klett, J.; Klinkhammer, K. W.; Niemeyer, M. Chem.Eur. J. 1999, 5,
2531−2536.
18
ions. Analogous reactions of M(SiR3)n and MXn are likely to
afford metals and should be similarly exothermic, which may lead
to new growth processes for metal films upon appropriate
precursor development.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthetic procedures and analytical and spectroscopic data for
1−3, X-ray crystallographic data for 1−3 in CIF format, and
details of the DFT calculations. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: chw@chem.wayne.edu.
Notes
The authors declare no competing financial interest.
C
dx.doi.org/10.1021/ic502184f | Inorg. Chem. XXXX, XXX, XXX−XXX