ALD Precursors, Precursor Design,
Chemistry and Mechanisms
D. J. H. Emslie
ALD/ALE Tutorial, Denver, Colorado
July 15, 2017
Emslie Group in June 2017, McMaster Univ., Hamilton, ON, Canada
• ALD versus CVD from Chemistry and Film-Deposition Perspectives
• Overview – H2O, H2S, NH3, O2/O3, H2, H2/NH3 etc.
• Precursor design – stability, volatility, m.p., reactivity, evaluation
• Unique Examples of Element ALD:
• Introduction to Cu ALD
• Cu ALD using ZnEt2 as the co-reactant (inc. solution studies)
• Early Transition Metal ALD
• WF6 + Si2H6 (vs TaF5)
• TiCl4 + 1,4-disilyl-substituted 2,5-cyclohexadiene or
1,4-dihydropyrazine co-reagents
• ML2 + BH3(NHMe2) [M = Cr, and possibly Mn]
• MnR2 + H2 (organometallic precursors)
• Summary / Conclusions
OUTLINE 2
CVD
pulsed-
CVD
ALD
Chemistry
Perspective:
Thin Film
Deposition
Perspective:
p-CVD and ALD are similar since they
utilize a precursor and a co-reactant –
this offers more reactivity possibilities
Basic CVD is distinct
in that it involves
just 1 precursor
ALD is distinct in
that it achieves
self-limiting growth
and more
conformal and
uniform deposition
CVD and p-CVD are similar in that they do
not allow for self-limiting growth or highly
conformal and uniform deposition
(not to the extent possible with ALD)
CVD vs ALD from different perspectives
3
ALD OVERVIEW – ACCESS TO DIFFERENT MATERIALS
Co-reactants typically determine the type of thin film deposited (oxide, nitride etc).
Precursors are designed to exhibit the desired reactivity with a particular co-reactant.
Many Common precursors are halide, alkyl, Cp, amido, alkoxide, aminoalkoxide, acac,
b-diketiminate, amidinate complexes.
ALD Precursor
(containing metal
in desired film)
Py(HF)x or TiF4 or TaF5
H2O, O2 or O3
H2S
H2Se or H2Te (toxic!!) or
E(SiMe3)2 (with MClx)
NH3, or
RNH2H2 or
O2 or O3 (noble metals only)
Metal Sulfide
Metal Selenide
Metal Telluride
Elemental Metal
Metal Nitride
Metal Fluoride
Metal Oxide
4
WHY NEW ALD PRECURSORS AND REACTIVITIES ?
Avoidance of CoReactants
which
Alter Chemistry of
Underlying Substrate
New ALD Precursors,
Reactivities, Methods
Materials Currently
Inaccessible by ALD
(in general, or on
required substrate
materials)
Higher GPC, or
More Desirable
Morphology
ALD of Higher
Purity Films
More volatile precursors,
or alternative
reactivities
Larger ALD
Temperature
Window (ideally
with approx.
constant GPC)
ALD at Lower
Temperature
More reactive,
volatile and thermally
robust precursors
Less Potential for
undesirable reactivity
(so long as reaction byproducts
are still
efficiently removed from
the growing film)
Avoidance of
Metal Film
Agglomeration More reactive coreactants
may allow
use of less thermally
robust precursors (with
volatility / reactivity
advantages – e.g.
fluorinated precursors,
or more reactive co-
reactants)
New precursors /
reactivity
New reactivity
(often relying upon
new co-reactants)
Compatible
with a Broader
Range of
Substrates
5
PRECURSOR DESIGN CONSIDERATIONS
Precursor Attributes Additional Comments Comments on Evaluation
Volatile Ideally deliverable at low T (low molecular
weight, alkyl or silyl > aryl, fluorination,
sometimes less symmetrical compounds).
Sublimation/Distillation Temp (or
measurement of vapor pressure
vs temp.).
Thermally Stable Long timescale --- Thermally stable for months
at delivery temperature.
Short timescale – no CVD until high temp.
Heating at delivery temp for 24h.
TGA (thermogravimetric anal.),
ideally at atm. and low pressure.
Reactive Reacts with co-reactant at low T (with wide
temp. window in which desired reactivity is
observed, ideally with fairly constant GPC vs T).
Literature and solution reactivity
can serve as a guide.
In-reactor studies essential.
Reacts with co-reactant to
form volatile and thermally
robust products
Reaction by-products must be readily removed
from the growing thin film.
Expected byproducts can be
prepared. Alternatively, do ALD
and assess film purity.
Low melting point (ideally) Ideally, precursor will be liquid at the delivery
temp (much less important for ALD vs CVD).
Longer alkyl (e.g. n-Bu, i-Bu) groups can help,
but sometimes at the expense of volatility…
---
Reactivity allowing ALD of
multiple materials ?
One precursor for ALD of multiple materials
could be beneficial.
---
No rapid exothermic
decomp. upon heating
--- Exercise caution with certain
classes of compound. TGA.
Scaleable synthesis Synthesis can often be improved if initial ALD
performance is promising.
---
6
Thermal ALD of Copper
CuII precursors with:
• H2 (≥ 150 oC)
• ROH (300 oC)
• HCO2H, then N2H4 (100-170 oC)
• BH3(NHMe2) (130-160 oC)
CuI precursors with:
• Zn metal (>400 oC; impure films)
• H2O, then H2 (375-475 oC).
The above methods require the use of
anhydrous hydrazine or operate at
temperatures ≥ 130 oC, and in the case of very
thin films, this can lead to agglomeration.
At outset of our work in this area, we asked:
Can ZnEt2 be used as a co-reactant for Cu ALD
(by introducing Et groups onto Cu, resulting in
unstable Cu alkyl species) ?
Methods for Thermal Cu ALD: Emslie, D. J. H.; Chadha P.; J. S. Price, Metal ALD and
pulsed-CVD: Fundamental reactions and links with solution
chemistry, Coord. Chem. Rev., 2013, 257, 3282-3296.
Knisley, T. J.; Kalutarage, L. C.; Winter, C. H., Precursors and
chemistry for the ALD of metallic first row transition metal
films, Coord. Chem. Rev., 2013, 257, 3222-3231.
Kalutarage, L. C.; Clendenning, S. B.; Winter, C. H., LowTemperature
Atomic Layer Deposition of Copper Films
Using Borane Dimethylamine as the Reducing Co-reagent,
Chem. Mater. 2014, 26, 3731−3738.
7
PROPOSED REACTION SCHEME
Stepwise
reduction:
Copper(I)
Copper(II)
Copper Metal
The key to each reduction step =
Formation of highly unstable
alkyl copper complexes
8
Temp 5 eq.
25 °C Very pale
yellow soln
50 °C Pale pink soln
75 °C Cu mirror with
reddish pink soln
100 °C Cu mirror with
colorless soln
120 °C Cu mirror with
colorless soln
Screening Reactions – A Typical Reaction
Various simple
Copper(II)
coordination
complexes
An organo-main
group co-reactant
with the potential to
generate copper
alkyl species
Order of Reactivity
from Solution
Screening:
ZnEt2 >
AlMe3 >>
BEt3
9
Temp 5 eq.
25 °C Very pale
yellow soln
50 °C Pale pink soln
75 °C Cu mirror with
reddish pink soln
100 °C Cu mirror with
colorless soln
120 °C Cu mirror with
colorless soln
Screening Reactions – A Typical Reaction
Various simple
Copper(II)
coordination
complexes
An organo-main
group co-reactant
with the potential to
generate copper
alkyl species
Order of Reactivity
from Solution
Screening:
ZnEt2 >
AlMe3 >>
BEt3
No
deposition
N
Cu
N
N N
Et
Et
+ BEt3
10
Pulsed-CVD of Cu Metal using
[Cu(PyrImEt)2] with ZnEt2 at 130 oC
N
Cu
N
N N
Et
Et
+ ZnEt2
Deposition on SiO2
1500 x [6s DEZ / 7s P / 9s CuL2 / 7s P]
Appearance: Metallic Cu
Cu Film thickness ~ 470Å
GPC ~0.31 Å/cycle
Rs: 6.5 Ω/sq; ρ~ 31 Ωcm
B. Vidjayacoumar, D. J. H. Emslie, S. B. Clendenning and J. M. Blackwell et al. Chem. Mater., 2010, 22, 4844-4853.
Not Self Limiting  pulsed-CVD
Dr. Scott B. Clendenning, Intel
11
Pulsed-CVD of Cu Metal using
[Cu(PyrImEt)2] with ZnEt2 at 130 oC
N
Cu
N
N N
Et
Et
+ ZnEt2
Deposition on SiO2
1500 x [6s DEZ / 7s P / 9s CuL2 / 7s P]
Appearance: Metallic Cu
Cu Film thickness ~ 470Å
GPC ~0.31 Å/cycle
Rs: 6.5 Ω/sq; ρ~ 31 Ωcm
B. Vidjayacoumar, D. J. H. Emslie, S. B. Clendenning and J. M. Blackwell et al. Chem. Mater., 2010, 22, 4844-4853.
Not Self Limiting  pulsed-CVD
Dr. Scott B. Clendenning, Intel
• Zinc incorporation results from Zn CVD,
which is significant at > 100 oC
1000 cycles × [1s pulse ZnEt2 / 3s chamber purge] with Cu substrate  Zn CVD
12
Pulsed-CVD of Cu Metal using
[Cu(PyrImEt)2] with ZnEt2 at 130 oC
N
Cu
N
N N
Et
Et
+ ZnEt2
Deposition on SiO2
1500 x [6s DEZ / 7s P / 9s CuL2 / 7s P]
Appearance: Metallic Cu
Cu Film thickness ~ 470Å
GPC ~0.31 Å/cycle
Rs: 6.5 Ω/sq; ρ~ 31 Ωcm
B. Vidjayacoumar, D. J. H. Emslie, S. B. Clendenning and J. M. Blackwell et al. Chem. Mater., 2010, 22, 4844-4853.
Not Self Limiting  pulsed-CVD
Dr. Scott B. Clendenning, Intel
• Zinc incorporation results from Zn CVD,
which is significant at > 100 oC
• Minimum delivery temperature for our
copper precursor was 120 oC.
• With a more volatile copper precursor,
Fischer, Sung et al. demonstrated Cu
metal ALD using ZnEt2 at 100 oC
(Angew. Chem. Int. Ed. 2009, 48, 4536-4539)
1000 cycles × [1s pulse ZnEt2 / 3s chamber purge] with Cu substrate  Zn CVD
13
Pulsed-CVD of Cu Metal using
[Cu(PyrImEt)2] with ZnEt2 at 130 oC
N
Cu
N
N N
Et
Et
+ ZnEt2
Deposition on SiO2
1500 x [6s DEZ / 7s P / 9s CuL2 / 7s P]
Appearance: Metallic Cu
Cu Film thickness ~ 470Å
GPC ~0.31 Å/cycle
Rs: 6.5 Ω/sq; ρ~ 31 Ωcm
B. Vidjayacoumar, D. J. H. Emslie, S. B. Clendenning and J. M. Blackwell et al. Chem. Mater., 2010, 22, 4844-4853.
Not Self Limiting  pulsed-CVD
Dr. Scott B. Clendenning, Intel
• Zinc incorporation results from Zn CVD,
which is significant at > 100 oC
• Minimum delivery temperature for our
copper precursor was 120 oC.
• With a more volatile copper precursor,
Fischer, Sung et al. demonstrated Cu
metal ALD using ZnEt2 at 100 oC
(Angew. Chem. Int. Ed. 2009, 48, 4536-4539)
Reproduced with permission from Angew. Chem.
Int. Ed. 2009, 48, 4536-4539.
SEM
XRD
100-120 oC
SEM TEM
Cu nanotubes formed by deposition into
pores in a polycarbonate membrane
Cu on Si/SiO2
14
What Can we learn from NMR Spectroscopy?
Prepare NMR samples in the glovebox
Record NMR Spectra as Reactions
Proceed (increasing temperature
incrementally if necessary)
Low
Pressure /
Vacuum
NMR tube
For compounds without any unpaired
electrons:
15
x = 0.3
20 oC, 15 min
x = 1.0
20 oC, 15 min
x = 5.0
20 oC, 1 hour
B. Vidjayacoumar, D. J. H. Emslie, J. M. Blackwell and S. B. Clendenning et al. Chem. Mater., 2010, 22, 4854-4866.
NMR Spectrum – CuL2 + x ZnEt2
17
NMR Spectrum – CuL2 + x ZnEt2
x = 0.3
20 oC, 15 min
x = 1.0
20 oC, 15 min
x = 5.0
20 oC, 1 hour
All Stable Intermediates / Byproducts Independently Synthesized.
Selected X-ray Crystal Structures:
B. Vidjayacoumar, D. J. H. Emslie, J. M. Blackwell and S. B. Clendenning et al. Chem. Mater., 2010, 22, 4854-4866.
18
CuL2 + n ZnEt2  Cu metal + byproducts
• Multi step mechanism with
several different available
pathways
• Reaction steps identified by
observation and synthesis of
intermediates and byproducts
• n-Butane is the only gas
formed during reduction from
CuII to CuI.
• Ethylene, ethane and hydrogen
(not n-butane) are formed
during reduction from CuI to
Cu0 (copper metal).
B. Vidjayacoumar, D. J. H. Emslie, J. M. Blackwell and S. B. Clendenning et al. Chem. Mater., 2010, 22, 4854-4866.
19
SOLUTION Reaction Pathway
Stepwise reduction:
Copper(II)
Copper(I)
Copper Metal
via unstable alkyl
copper complexes
Conclusion:
The instability of certain metal
alkyl complexes (generated in
situ) can be exploited for metal
deposition, through the use of
main group alkyl complexes
(e.g. ZnEt2) as co-reactants
B. Vidjayacoumar, D. J. H. Emslie, J. M. Blackwell and S. B. Clendenning et al. Chem. Mater., 2010, 22, 4854-4866.
20
1 18
1
1
H
2.20
2 13 14 15 16 17
2
He
2
3
Li
0.98
4
Be
1.57
5
B
2.04
6
C
2.55
7
N
3.04
8
O
3.44
9
F
3.98
10
Ne
3
11
Na
0.93
12
Mg
1.31
3 4 5 6 7 8 9 10 11 12
13
Al
1.61
14
Si
1.90
15
P
2.19
16
S
2.58
17
Cl
3.16
18
Ar
4
19
K
0.82
20
Ca
1.00
21
Sc
1.36
22
Ti
1.54
23
V
1.63
24
Cr
1.66
25
Mn
1.55
26
Fe
1.83
27
Co
1.88
28
Ni
1.91
29
Cu
1.90
30
Zn
1.65
31
Ga
1.81
32
Ge
2.01
33
As
2.18
34
Se
2.55
35
Br
2.96
36
Kr
3.00
5
37
Rb
0.82
38
Sr
0.95
39
Y
1.22
40
Zr
1.33
41
Nb
1.6
42
Mo
2.16
43
Tc
1.9
44
Ru
2.2
45
Rh
2.28
46
Pd
2.20
47
Ag
1.93
48
Cd
1.69
49
In
1.78
50
Sn
1.96
51
Sb
2.05
52
Te
2.1
53
I
2.66
54
Xe
2.60
6
55
Cs
0.79
56
Ba
0.89
71
Lu
1.27
72
Hf
1.3
73
Ta
1.5
74
W
2.36
75
Re
1.9
76
Os
2.2
77
Ir
2.20
78
Pt
2.28
79
Au
2.54
80
Hg
2.00
81
Tl
2.04
82
Pb
2.33
83
Bi
2.02
84
Po
2.0
85
At
2.2
86
Rn
2.2
7
87
Fr
0.7
88
Ra
0.9
103
Lr
--
104
R
(257)
105
D
(260)
106
S
(263)
107
B
(262)
108
H
(265)
109
M
(266)
110
--
()
111
--
()
112
-
()
114
--
()
116
-
()
118
-
()
Ln
57
La
1.10
58
Ce
1.12
59
Pr
1.13
60
Nd
1.14
61
Pm
--
62
Sm
1.17
63
Eu
--
64
Gd
1.20
65
Tb
--
66
Dy
1.22
67
Ho
1.23
68
Er
1.24
69
Tm
1.25
70
Yb
--
An
89
Ac
1.1
90
Th
1.3
91
Pa
1.5
92
U
1.38
93
Np
1.36
94
Pu
1.28
95
Am
1.3
96
Cm
1.3
97
Bk
1.3
98
Cf
1.3
99
Es
1.3
100
Fm
1.3
101
Md
1.3
102
No
1.3
Thermal ALD of Transition Metals (or pulsed-CVD)
Pauling Electronegativity
2.0 – 2.6
1.7 – 2.0
1.5 – 1.7
1.0 – 1.5
0.8 – 1.0
ALD reported ALD with
conditions
21
THE CHALLENGE OF EARLY TRANSITION METAL ALD
J.W. Klaus, S.J. Ferro, S.M. George, Atomically controlled growth of tungsten and tungsten
nitride using sequential surface reactions, Appl. Surf. Sci., 2000, 162–163 479–491.
Kalutarage, L. C.; Martin, P. D.; Heeg, M. J.; Winter, C. H., Volatile and Thermally Stable Mid to Late Transition Metal
Complexes Containing α-Imino Alkoxide Ligands, a New Strongly Reducing Coreagent, and Thermal Atomic Layer
Deposition of Ni, Co, Fe, and Cr Metal Films, J. Am. Chem. Soc. 2013, 135, 12588.
Klesko, J. P.; Thrush, C. M.; Winter, C. H., Thermal Atomic Layer Deposition of Titanium Films Using Titanium Tetrachloride and
2-Methyl-1,4-bis(trimethylsilyl)-2,5-cyclohexadiene or 1,4-Bis(trimethylsilyl)-1,4-dihydropyrazine, Chem. Mater. 2016, 28, 700.
GPC 0.06 Å/cycle. Rapid oxidation to TiO2 in air, but some Ti metal remains deeper in the film, according to XPS
Analogous reaction with TaF5
generated Ta silicide
Lemonds, A.M.; White, J.M.; Ekerdt, J.G., Surface science investigations of atomic layer
deposition half-reactions using TaF5 and Si2H6, Surf. Sci., 2003, 538, 191.
Mn
Ti
W
Cr
• Ru substrate
• 180-225 oC
• GPC 0.07-0.10 Å/cycle
22
THE CHALLENGE OF EARLY TRANSITION METAL ALD
J.W. Klaus, S.J. Ferro, S.M. George, Atomically controlled growth of tungsten and tungsten
nitride using sequential surface reactions, Appl. Surf. Sci., 2000, 162–163 479–491.
Kalutarage, L. C.; Martin, P. D.; Heeg, M. J.; Winter, C. H., Volatile and Thermally Stable Mid to Late Transition Metal
Complexes Containing α-Imino Alkoxide Ligands, a New Strongly Reducing Coreagent, and Thermal Atomic Layer
Deposition of Ni, Co, Fe, and Cr Metal Films, J. Am. Chem. Soc. 2013, 135, 12588.
Klesko, J. P.; Thrush, C. M.; Winter, C. H., Thermal Atomic Layer Deposition of Titanium Films Using Titanium Tetrachloride and
2-Methyl-1,4-bis(trimethylsilyl)-2,5-cyclohexadiene or 1,4-Bis(trimethylsilyl)-1,4-dihydropyrazine, Chem. Mater. 2016, 28, 700.
GPC 0.06 A/cycle. Rapid oxidation to TiO2 in air, but some Ti metal remains deeper in the film, according to XPS
Analogous reaction with TaF5
generated Ta silicide
Lemonds, A.M.; White, J.M.; Ekerdt, J.G., Surface science investigations of atomic layer
deposition half-reactions using TaF5 and Si2H6, Surf. Sci., 2003, 538, 191.
Mn
Ti
W
Cr
• Ru substrate
• 180-225 oC
• GPC 0.07-0.10 Å/cycle
Disilane strips the fluoride ligands off tungsten.
Similar reactivity observed for SiH4 and B2H6, but less clean.
23
• Ru substrate
• 180-225 oC
• GPC 0.07-0.10 Å/cycle
THE CHALLENGE OF EARLY TRANSITION METAL ALD
J.W. Klaus, S.J. Ferro, S.M. George, Atomically controlled growth of tungsten and tungsten
nitride using sequential surface reactions, Appl. Surf. Sci., 2000, 162–163 479–491.
Kalutarage, L. C.; Martin, P. D.; Heeg, M. J.; Winter, C. H., Volatile and Thermally Stable Mid to Late Transition Metal
Complexes Containing α-Imino Alkoxide Ligands, a New Strongly Reducing Coreagent, and Thermal Atomic Layer
Deposition of Ni, Co, Fe, and Cr Metal Films, J. Am. Chem. Soc. 2013, 135, 12588.
Klesko, J. P.; Thrush, C. M.; Winter, C. H., Thermal Atomic Layer Deposition of Titanium Films Using Titanium Tetrachloride and
2-Methyl-1,4-bis(trimethylsilyl)-2,5-cyclohexadiene or 1,4-Bis(trimethylsilyl)-1,4-dihydropyrazine, Chem. Mater. 2016, 28, 700.
GPC 0.06 A/cycle. Rapid oxidation to TiO2 in air, but some Ti metal remains deeper in the film, according to XPS
Analogous reaction with TaF5
generated Ta silicide
Lemonds, A.M.; White, J.M.; Ekerdt, J.G., Surface science investigations of atomic layer
deposition half-reactions using TaF5 and Si2H6, Surf. Sci., 2003, 538, 191.
Mn
Ti
W
Cr
Conceptually, the two precursors serve as a more reactive
form of a disilane (Me3Si-SiMe3), which strips chloride ligands
from titanium.
Requires a volatile metal chloride, which is likely to limit
translation of methodology to most other early transition metals.
24
THE CHALLENGE OF EARLY TRANSITION METAL ALD
J.W. Klaus, S.J. Ferro, S.M. George, Atomically controlled growth of tungsten and tungsten
nitride using sequential surface reactions, Appl. Surf. Sci., 2000, 162–163 479–491.
Kalutarage, L. C.; Martin, P. D.; Heeg, M. J.; Winter, C. H., Volatile and Thermally Stable Mid to Late Transition Metal
Complexes Containing α-Imino Alkoxide Ligands, a New Strongly Reducing Coreagent, and Thermal Atomic Layer
Deposition of Ni, Co, Fe, and Cr Metal Films, J. Am. Chem. Soc. 2013, 135, 12588.
Klesko, J. P.; Thrush, C. M.; Winter, C. H., Thermal Atomic Layer Deposition of Titanium Films Using Titanium Tetrachloride and
2-Methyl-1,4-bis(trimethylsilyl)-2,5-cyclohexadiene or 1,4-Bis(trimethylsilyl)-1,4-dihydropyrazine, Chem. Mater. 2016, 28, 700.
GPC 0.06 A/cycle. Rapid oxidation to TiO2 in air, but some Ti metal remains deeper in the film, according to XPS
Analogous reaction with TaF5
generated Ta silicide
Lemonds, A.M.; White, J.M.; Ekerdt, J.G., Surface science investigations of atomic layer
deposition half-reactions using TaF5 and Si2H6, Surf. Sci., 2003, 538, 191.
Mn
Ti
W
Cr
Deposition only observed on a ruthenium substrate after a lengthy surface
pre-treatment with BH3(NHMe2).
Deposition ceased at a film thickness of ~ 10 nm.
For Mn, immediate oxidation of the film upon removal from reactor
prevented determination of whether Mn metal or MnO had been deposited.
• Ru substrate
• 180-225 oC
• GPC 0.07-0.10 Å/cycle
25
• Reduction to elemental metal becomes increasingly challenging for more
electropositive metals
• Electropositive metals have a high tendency to form oxides, nitrides and
halides, making many coordination complexes poorly suited for
electropositive metal ALD
An Alternative Approach:
Organometallic Precursors for Electropositive Metal ALD ?
How will the
necessary thermal
stability be achieved
without compromising:
(a) volatility and
(b) reactivity?
Can highly reactive
organometallic precursors
(acyclic hydrocarbyl
complexes, rather than
cyclopentadienyl complexes)
be used for electropositive
metal ALD?
Metal-organic
Complexes
More Reactive Less Reactive
Organometallic
Complexes
Contain a metal
and organic
ligands
Contain direct
M-C (or M-Si, or
M-H) bonds
26
• Reduction to elemental metal becomes increasingly challenging for more
electropositive metals
• Electropositive metals have a high tendency to form oxides, nitrides and
halides, making many coordination complexes poorly suited for
electropositive metal ALD
How will the
necessary thermal
stability be achieved
without compromising:
(a) volatility and
(b) reactivity?
Can highly reactive
organometallic precursors
(acyclic hydrocarbyl
complexes, rather than
cyclopentadienyl complexes)
be used for electropositive
metal ALD?
An Alternative Approach:
Organometallic Precursors for Electropositive Metal ALD ?
27
OVERALL STRATEGY FOR MANGANESE DEPOSITION
The metal precursor must be:
• Thermally robust & volatile
• Reactive towards the desired co-reactant --- for this reason, highly reactive metal
alkyl and allyl complexes are the focus of this work
Precursors:
Alkyl or Allyl Complexes
- Free from anionic
N-donors
- Free from halogens
or O-donors
Highly Reactive
28
OVERALL STRATEGY FOR MANGANESE DEPOSITION
Co-Reactants Selected to form Alkyl / Hydride Intermediate which should
be particularly prone to reductive elimination
Precursors:
Alkyl or Allyl Complexes
- Free from anionic
N-donors
- Free from halogens
or O-donors
Highly Reactive
29
OVERALL STRATEGY FOR MANGANESE DEPOSITION
Co-reactants:
• Class 1 co-reactants (H2, PhSiH3, R’2BH) form the alkyl / hydride complex directly
• Class 2 co-reactants (BEt3, AlEt3, ZnEt2) initially form an unstable ethyl complex
Precursors:
Alkyl or Allyl Complexes
- Free from anionic
N-donors
- Free from halogens
or O-donors
Highly Reactive
30
OVERALL STRATEGY FOR MANGANESE DEPOSITION
Co-reactants:
• Class 1 co-reactants (H2, PhSiH3, R’2BH) form the alkyl / hydride complex directly
• Class 2 co-reactants (BEt3, AlEt3, ZnEt2) initially form an unstable ethyl complex
Precursors:
Alkyl or Allyl Complexes
- Free from anionic
N-donors
- Free from halogens
or O-donors
Highly Reactive
31
OVERALL STRATEGY FOR MANGANESE DEPOSITION
Co-reactants:
• Class 1 co-reactants (H2, PhSiH3, R’2BH) form the alkyl / hydride complex directly
• Class 2 co-reactants (BEt3, AlEt3, ZnEt2) initially form an unstable ethyl complex
Suitable MnR2 precursors required
32
(1) Primary alkyl complexes
- fairly straightforward to prepare
- presence of a-hydrogen atoms may render complexes less thermally stable
(2) Tertiary alkyl complexes [C(SiR3)3 complexes]
- a-hydrogen free, and b-hydrogen free
- Complexes expected to exhibit high thermal stability (resistance to carbide formation)
(3) Bulky allyl complexes
- resistant to common decomposition pathways
- allyl complexes often have higher thermal stability than alkyl complexes
- allyl complexes are much more reactive than cyclopentadienyl complexes
Precursor Design
- Oxygen-free alkyl and allyl complexes
- All complexes will be highly reactive due to the presence of metal-carbon bonds
33
Bulky alkyl group for high thermal stability
Lower symmetry for
increased volatility ?
Allyl group for
increased reactivity
and improved
synthetic
accessibility
3
Tertiary ALKYL and ALLYL Complexes
High thermal stability
due to steric bulk and
absence of a-hydrogens
Synthetically inaccessible in gram
quantities due to extreme steric bulk
1
Decomp.
at ~ 70 oC
2 Insufficient
thermal
stability for
use as an
ALD
precursor
34
MIXED ALKYL (TSI) / ALLYL (allylTMS2) MANGANESE(II) COMPLEXES 35
MIXED ALKYL (TSI) / ALLYL (allylTMS2) MANGANESE(II) COMPLEXES 36
L = X-Li(THF)3; X= 89% Cl, 11% Br L = PEt3 L = dmap
L = quinuclidineL = PMe3L = THF
X-Ray Crystal Structures of [(allylTMS2)Mn(TSI)L] Complexes
Cl
Mn
Mn
Mn
Mn Mn
Mn
Li
O P
P
N
N
N
Mn–O = 2.170(3) Å Mn–P = 2.617(1) Å
Mn–N = 2.163(4) Å
Mn–N = 2.258(4) and 2.270(4) Å
Sublimes 50-70 oC, 10 mTorr
No decomposition after 24h at 100 oC
37
‘MnR2’
dmpe
‘MnR2’
PMe3
dmpedmpm
PRIMARY ALKYL MANGANESE(II) COMPLEXES: R = CH2CMe3
[Mn(CH2CMe3)2]4
not quite thermally
stable / volatile
enough
for ALD
[Mn(CH2SiMe3)2]x
Polymeric
(low volatility)
J. S. Price, P. Chadha, D. J. H. Emslie, Organometallics, 2016, 35, 168-180.
38
‘MnR2’
dmpe
‘MnR2’
PMe3
dmpedmpm
PRIMARY ALKYL MANGANESE(II) COMPLEXES: R = CH2CMe3
[Mn(CH2CMe3)2]4
not quite thermally
stable / volatile
enough
for ALD
[Mn(CH2SiMe3)2]x
Polymeric
(low volatility)
Sublimes 80 oC,
5 mTorr
Decomposes over
5-6h at 120 oC
Melts and sublimes 60 oC, 5 mTorr
Minimal Decomp after 24h at 120 oC
J. S. Price, P. Chadha, D. J. H. Emslie, Organometallics, 2016, 35, 168-180.
0.5 Torr 760 Torr
weight%
100
80
60
40
20
0
0 50 100 150 200 250
Temp (oC)
39
Solution Reactivity of Primary Alkyl Manganese(II) Complexes with H2
Mn PXRD
J. S. Price, P. Chadha, D. J. H. Emslie, Organometallics, 2016, 35, 168-180.
ALD using forming gas
(5% H2 in N2)
• With substrate = 125 oC,
a GPC of 0.2 Å / cycle was
observed on Ru seed.
• The film was nonconductive
after airexposure,
likely due to
complete oxidation…
Ru seed
Thickness ~ 80 Å (ellipsom),
GPC ~ 0.2 Å/cycle
400 x [3s Mn / 5s purge / 2s FG /
15 s purge]
40
Solution Reactivity of Primary Alkyl Manganese(II) Complexes with H2
Mn PXRD
ALD using forming gas
(5% H2 in N2)
• With substrate = 125 oC,
a GPC of 0.2 Å / cycle was
observed on Ru seed.
• The film was nonconductive
after airexposure,
likely due to
complete oxidation…
Ru seed
Thickness ~ 80 Å (ellipsom),
GPC ~ 0.2 Å/cycle
400 x [3s Mn / 5s purge / 2s FG /
15 s purge]
MnS ALD using H2S
(one precursor for
multiple materials)
• With substrate = 110 oC,
a GPC of 0.3 Å / cycle was
observed on Ru seed.
• XPS: ~ 1:1 ratio of Mn:S
with no P and low C.
5-10% O, presumably due
to air exposure…
Ru seed
Thickness ~ 80 Å (ellipsom),
GPC ~ 0.2 Å/cycle
400 x [3s Mn / 5s purge / 2s FG /
15 s purge]
Ru seed
Thk ~ 390 Å (XSEM)
GPC ~ 0.33 Å/cycle
1200x [3s Mn / 8s purge / 3s Mn
/ 4s purge / 0.4s H2S / 8s purge]
41
ALD Reactor Studies
Oven #1
Reactor
Heated
Bubbler #1
RT
bubbler
Foreline
Argon
delivery
manifold
Substrate
heater
Sample
Stage
- Until recently, all ALD studies on our compounds were conducted by collaborators.
- We now have a home-built ALD reactor, so further studies on our Mn cpds are on the
schedule for this summer...
42
- Many reported ALD processes are far from ideal:
- Also, many materials (especially pure elements) simply cannot yet be deposited by thermal ALD.
- Consequently, new precursors, co-reactants, reactivities and methods are required.
- ALD at lower temperature (may require more reactive precursor/co-reactant combinations and/or more
volatile precursors and co-reactants) offers various advantages:
- For metal ALD, most new ALD methods leading to ALD of previously inaccessible materials rely upon new
co-reactants rather than new metal precursors.
- For thermal metal ALD, most precursors are not organometallic (some Pt compounds are exceptions).
- The Cu chemistry highlights the use of a new type of co-reactant (ZnEt2) for metal ALD. The initial focus
was not on Cu precursor design. However, limitations in ZnEt2 thermal stability brought the work full circle,
where precursor volatility now appears to be the limitation in terms of accessing Cu deposition significantly
below 100 oC. --- Solution reactivity studies can provide mechanistic insight.
- The Mn chemistry highlights a new approach for electropositive metal ALD: harnessing the high reactivity
of manganese alkyl and allyl complexes for Mn metal deposition.
- Highly reactive organometallic precursors have the disadvantage of high air-sensitivity.
However, they may allow deposition of multiple materials from a single precursor.
Summary / Conclusions
- High deposition temperature (leading to agglomeration, low film purity, incompatibility with
certain substrates, limited selection of precursors with appropriate thermal stability)
- Limited ALD temperature window
- Very low GPC
- Low film purity
- Undesirable film morphology
- Undesirable substrate selectivity (e.g. no deposition on H-terminated Si)
- Undesirable reactivity between the co-reactant and the underlying substrate.
43
- Many reported ALD processes are far from ideal:
- Also, many materials (especially pure elements) simply cannot yet be deposited by thermal ALD.
- Consequently, new precursors, co-reactants, reactivities and methods are required.
- ALD at lower temperature (may require more reactive precursor/co-reactant combinations and/or more
volatile precursors and co-reactants) offers various advantages:
- For metal ALD, most new ALD methods leading to ALD of previously inaccessible materials rely upon new
co-reactants rather than new metal precursors.
- For thermal metal ALD, most precursors are not organometallic (some Pt compounds are exceptions).
- The Cu chemistry highlights the use of a new type of co-reactant (ZnEt2) for metal ALD. The initial focus
was not on Cu precursor design. However, limitations in ZnEt2 thermal stability brought the work full circle,
where precursor volatility now appears to be the limitation in terms of accessing Cu deposition significantly
below 100 oC. --- Solution reactivity studies can provide mechanistic insight.
- The Mn chemistry highlights a new approach for electropositive metal ALD: harnessing the high reactivity
of manganese alkyl and allyl complexes for Mn metal deposition.
- Highly reactive organometallic precursors have the disadvantage of high air-sensitivity.
However, they may allow deposition of multiple materials from a single precursor.
- Many reported ALD processes are far from ideal:
- Also, many materials (especially pure elements) simply cannot yet be deposited by thermal ALD.
- Consequently, new precursors, co-reactants, reactivities and methods are required.
- ALD at lower temperature (may require more reactive precursor/co-reactant combinations and/or more
volatile precursors and co-reactants) offers various advantages:
- For metal ALD, most new ALD methods leading to ALD of previously inaccessible materials rely upon new
co-reactants rather than new metal precursors.
- For thermal metal ALD, most precursors are not organometallic (some Pt compounds are exceptions).
- The Cu chemistry highlights the use of a new type of co-reactant (ZnEt2) for metal ALD. The initial focus
was not on Cu precursor design. However, limitations in ZnEt2 thermal stability brought the work full circle,
where precursor volatility now appears to be the limitation in terms of accessing Cu deposition significantly
below 100 oC. --- Solution reactivity studies can provide mechanistic insight.
- The Mn chemistry highlights a new approach for electropositive metal ALD: harnessing the high reactivity
of manganese alkyl and allyl complexes for Mn metal deposition.
- Highly reactive organometallic precursors have the disadvantage of high air-sensitivity.
However, they may allow deposition of multiple materials from a single precursor.
- Many reported ALD processes are far from ideal:
- Also, many materials (especially pure elements) simply cannot yet be deposited by thermal ALD.
- Consequently, new precursors, co-reactants, reactivities and methods are required.
- ALD at lower temperature (may require more reactive precursor/co-reactant combinations and/or
more volatile precursors and co-reactants) offers various advantages:
- For metal ALD, most new ALD methods leading to ALD of previously inaccessible materials rely upon new
co-reactants rather than new metal precursors.
- For thermal metal ALD, most precursors are not organometallic (some Pt compounds are exceptions).
- The Cu chemistry highlights the use of a new type of co-reactant (ZnEt2) for metal ALD. The initial focus
was not on Cu precursor design. However, limitations in ZnEt2 thermal stability brought the work full circle,
where precursor volatility now appears to be the limitation in terms of accessing Cu deposition significantly
below 100 oC. --- Solution reactivity studies can provide mechanistic insight.
- The Mn chemistry highlights a new approach for electropositive metal ALD: harnessing the high reactivity
of manganese alkyl and allyl complexes for Mn metal deposition.
- Highly reactive organometallic precursors have the disadvantage of high air-sensitivity.
However, they may allow deposition of multiple materials from a single precursor.
- Avoids agglomeration of thin metal films.
- Compatible with a broader range of substrates, perhaps including those with polymer patterning.
- Leads to more predictable chemistry, potentially generating higher purity films.
- Can allow use of less-thermally robust but higher-volatility (and perhaps higher reactivity)
precursors and co-reactants.
Summary / Conclusions 44
- Many reported ALD processes are far from ideal:
- Also, many materials (especially pure elements) simply cannot yet be deposited by thermal ALD.
- Consequently, new precursors, co-reactants, reactivities and methods are required.
- ALD at lower temperature (may require more reactive precursor/co-reactant combinations and/or more
volatile precursors and co-reactants) offers various advantages:
- For metal ALD, most new ALD methods leading to ALD of previously inaccessible materials rely
upon new co-reactants rather than new metal precursors.
- For thermal metal ALD, most precursors not organometallic (some Pt compounds are exceptions).
- The Cu chemistry highlights the use of a new type of co-reactant (ZnEt2) for metal ALD.
--- The initial focus was not on Cu precursor design. However, limitations in ZnEt2 thermal stability brought
the work full circle, where precursor volatility now appears to be the limitation in terms of accessing Cu
deposition significantly below 100 oC. --- Solution reactivity studies can provide mechanistic insight.
- The Mn chemistry highlights a new approach for electropositive metal ALD: harnessing the high reactivity
of manganese alkyl and allyl complexes for Mn metal deposition.
- Highly reactive organometallic precursors have the disadvantage of high air-sensitivity.
However, they may allow deposition of multiple materials from a single precursor.
- Many reported ALD processes are far from ideal:
- Also, many materials (especially pure elements) simply cannot yet be deposited by thermal ALD.
- Consequently, new precursors, co-reactants, reactivities and methods are required.
- ALD at lower temperature (may require more reactive precursor/co-reactant combinations and/or more
volatile precursors and co-reactants) offers various advantages:
- For metal ALD, most new ALD methods leading to ALD of previously inaccessible materials rely upon new
co-reactants rather than new metal precursors.
- For thermal metal ALD, most precursors not organometallic (some Pt compounds are exceptions).
- The Cu chemistry highlights the use of a new type of co-reactant (ZnEt2) for metal ALD.
--- The initial focus was not on Cu precursor design. However, limitations in ZnEt2 thermal stability brought
the work full circle, where precursor volatility now appears to be the limitation in terms of accessing Cu
deposition significantly below 100 oC. --- Solution reactivity studies can provide mechanistic insight.
- The Mn chemistry highlights a new approach for electropositive metal ALD: harnessing the high reactivity
of manganese alkyl and allyl complexes for Mn metal deposition.
- Highly reactive organometallic precursors have the disadvantage of high air-sensitivity.
However, they may allow deposition of multiple materials from a single precursor.
Summary / Conclusions 45
- Many reported ALD processes are far from ideal:
- Also, many materials (especially pure elements) simply cannot yet be deposited by thermal ALD.
- Consequently, new precursors, co-reactants, reactivities and methods are required.
- ALD at lower temperature (may require more reactive precursor/co-reactant combinations and/or more
volatile precursors and co-reactants) offers various advantages:
- For metal ALD, most new ALD methods leading to ALD of previously inaccessible materials rely upon new
co-reactants rather than new metal precursors.
- For thermal metal ALD, most precursors not organometallic (some Pt compounds are exceptions).
- The Cu chemistry highlights the use of a new type of co-reactant (ZnEt2) for metal ALD.
--- The initial focus was not on Cu precursor design. However, limitations in ZnEt2 thermal stability brought
the work full circle, where precursor volatility now appears to be the limitation in terms of accessing Cu
deposition significantly below 100 oC. --- Solution reactivity studies can provide mechanistic insight.
- The Mn chemistry highlights a new approach for electropositive metal ALD: harnessing the high reactivity
of manganese alkyl and allyl complexes for Mn metal deposition.
- Highly reactive organometallic precursors have the disadvantage of high air-sensitivity.
However, they may allow deposition of multiple materials from a single precursor.
Summary / Conclusions 46
- Many reported ALD processes are far from ideal:
- Also, many materials (especially pure elements) simply cannot yet be deposited by thermal ALD.
- Consequently, new precursors, co-reactants, reactivities and methods are required.
- ALD at lower temperature (may require more reactive precursor/co-reactant combinations and/or more
volatile precursors and co-reactants) offers various advantages:
- For metal ALD, most new ALD methods leading to ALD of previously inaccessible materials rely upon new
co-reactants rather than new metal precursors.
- For thermal metal ALD, most precursors not organometallic (some Pt compounds are exceptions).
- The Cu chemistry highlights the use of a new type of co-reactant (ZnEt2) for metal ALD.
--- The initial focus was not on Cu precursor design. However, limitations in ZnEt2 thermal stability brought
the work full circle, where precursor volatility now appears to be the limitation in terms of accessing Cu
deposition significantly below 100 oC. --- Solution reactivity studies can provide mechanistic insight.
- The Mn chemistry highlights a new approach for electropositive metal ALD: harnessing the high reactivity
of manganese alkyl and allyl complexes for Mn metal deposition.
- Highly reactive organometallic precursors have the disadvantage of high air-sensitivity.
However, they may allow deposition of multiple materials from a single precursor.
Summary / Conclusions 47
Questions ?
Funding for the Cu and Mn work:
through:
ACKNOWLEDGEMENTS
Bala Preeti Jeffrey
Todd
Cu Mn Mn
Nick
Current
Majeda Novan
Current Current
ALD Research in the Group:
48