Synthesis of N-heterocyclic stannylene (Sn(II)) and germylene (Ge(II)) and a Sn(II) amidinate and their application as precursors for atomic layer deposition

Article abstract of DOI:10.1021/cm403901y

Thin films containing germanium or tin have a great variety of current and potential applications, particularly their oxides or chalcogenides. Chemical vapor deposition (CVD) and atomic layer deposition (ALD) are popular ways to make these thin films conformally even on challenging nanostructured substrates. The success of these processes depends on having precursors that are sufficiently stable, volatile, and reactive. In this paper we optimize the syntheses of the following three precursors: 1 and 2 are racemic Ge(II) or Sn(II) cyclic amides made from N²,N³-di-tert-butylbutane- 2,3-diamine, and 3 is bis(N,N'-diisopropylacetamidinato)tin(II). All three compounds are demonstrated to be effective precursors for ALD of their monosulfides, GeS or SnS, by reaction with H₂S. 2 has also been reported previously to make polycrystalline SnO₂ by ALD with oxidizing agents such as H₂O₂. The cyclic amides 1 and 2 are more volatile than the amidinate 3, vaporizing sufficiently for ALD even at precursor temperatures below 40°C, whereas 3 vaporizes at temperatures over 90°C. 1 and 2 can thus be used at lower substrate temperatures than 3. GeS or SnS can be deposited on substrates even at temperatures below 50°C, while ALD of SnS from 3 becomes slow below substrate temperatures of 100 °C because of insufficient vapor pressure. The amount of growth per ALD cycle is higher for the cyclic amide 2 than for the amidinate 3. The GeS films are smooth and amorphous, while the SnS films are polycrystalline and granular. All of these films are uniformly thick inside holes with aspect ratios (depth/diameter) greater than 40:1.

Full text of DOI:10.1021/cm403901y

Article
pubs.acs.org/cm
Synthesis of N‑Heterocyclic Stannylene (Sn(II)) and Germylene (Ge(II))
and a Sn(II) Amidinate and Their Application as Precursors for
Atomic Layer Deposition
Sang Bok Kim,†,⊥ Prasert Sinsermsuksakul,†,⊥ Adam S. Hock,§,¶,⊥ Robert D. Pike,‡ and Roy G. Gordon*,†
†Department of Chemistry & Chemical Biology, Harvard University, Cambridge, Massachusetts 02138, United States
‡Department of Chemistry, The College of William & Mary, Williamsburg, Virginia 23187, United States
§Department of Biological and Chemical Sciences, Illinois Institute of Technology, Chicago, Illinois 60616, United States
¶Chemical Sciences and Engineering Division, Argonne National Laboratory, Argonne, Illinois 60439, United States
S
* Supporting Information
ABSTRACT: Thin films containing germanium or tin have a
great variety of current and potential applications, particularly
their oxides or chalcogenides. Chemical vapor deposition
(CVD) and atomic layer deposition (ALD) are popular ways
to make these thin films conformally even on challenging
nanostructured substrates. The success of these processes
depends on having precursors that are sufficiently stable,
volatile, and reactive. In this paper we optimize the syntheses
of the following three precursors: 1 and 2 are racemic Ge(II)
or Sn(II) cyclic amides made from N2,N3-di-tert-butylbutane-
2,3-diamine, and 3 is bis(N,N′-diisopropylacetamidinato)tin-
(II). All three compounds are demonstrated to be effective
precursors for ALD of their monosulfides, GeS or SnS, by reaction with H2S. 2 has also been reported previously to make
polycrystalline SnO2 by ALD with oxidizing agents such as H2O2. The cyclic amides 1 and 2 are more volatile than the amidinate
3, vaporizing sufficiently for ALD even at precursor temperatures below 40 °C, whereas 3 vaporizes at temperatures over 90 °C. 1
and 2 can thus be used at lower substrate temperatures than 3. GeS or SnS can be deposited on substrates even at temperatures
below 50 °C, while ALD of SnS from 3 becomes slow below substrate temperatures of 100 °C because of insufficient vapor
pressure. The amount of growth per ALD cycle is higher for the cyclic amide 2 than for the amidinate 3. The GeS films are
smooth and amorphous, while the SnS films are polycrystalline and granular. All of these films are uniformly thick inside holes
with aspect ratios (depth/diameter) greater than 40:1.
problems (e.g., toxicity and limited availability) despite their
promising results. In particular, the toxicity of cadmium and
lead and their potential exposure are considered as a threat to
health for CdTe and PbTe.11 Cu2ZnSn(Se,S)422,23 is one of the
most promising Earth abundant photovoltaic absorbers with
11.1% solar conversion efficiency,23 satisfying both toxicity and
scalability problems by using abundant Cu, Zn, and Sn
elements. Metal monochalcogenides such as SnS,11,24−30
SnSe,11,31−33 GeS,11 and GeSe11 are being studied for their
possible application to solar cells due to their low toxicity,
economical large scale production, and a simpler chemistry
compared to Cu2ZnSn(Se,S)4.
1. INTRODUCTION
Metal chalcogenides have attracted increasing interest for their
electrical and optical properties.1 Tin chalcogenides (sulfide,
selenide, telluride) and tin oxide are being used in many
applications. Tin oxide2−5 has attracted great attention due to
its high transparency, infrared reflectivity, and semiconductor
property with an ∼3.62 eV band gap. The unique properties of
tin oxide have been applied to energy-conserving windows for
buildings,6 gas sensors,7 and transparent conductive electrodes
in thin film solar cells.8−10 Germanium chalcogenides such as
GeS and GeSe have attracted attention for their applications
(e.g., optical waveguides and photovoltaic devices).1,11 One of
their possible applications is in phase change nonvolatile
memories (PCRAM). Ge2Sb2Te5 (GST) is the most well-
known and popular phase change material for PCRAM.12−16
Many of the potential PCRAM structures call for deposition of
GST film conformally in deep trenches and pores.17−19
During device fabrication or preparation of functional
surfaces, formation of conformal and uniform films on
substrates in a highly controlled manner is important in
obtaining reproducible desired properties. Sputtering and
Recently, other metal chalcogenides have attracted attention
for solar cell applications. The currently developed solar cells
Received: November 25, 2013
Revised: February 24, 2014
Published: March 12, 2014
made from CdTe,20 PbTe,21 and Cu(In,Ga)(S,Se)2 have
22
© 2014 American Chemical Society
3065
dx.doi.org/10.1021/cm403901y | Chem. Mater. 2014, 26, 3065−3073

Chemistry of Materials
Article
chemical vapor phase deposition methods are currently used in
the formation of metal, metal oxide, metal nitride, and metal
chalcogenide thin films, in the fabrication of many kinds of
devices. Under the umbrella of chemical vapor phase deposition
methods are chemical vapor deposition (CVD) and atomic
layer deposition (ALD). ALD, previously known as atomic
layer epitaxy (ALE), is a modified form of CVD for deposition
of thin films.34,35 ALD differs from CVD insofar as the vapor
phase precursors arrive onto the substrate alternately via a pulse
sequence; the precursor pulses are separated by periods of
purging with inert gas. The film growth is performed by the
surface reaction between species formed on the surface and
incoming precursors by alternating depositions of pairs of
precursors. In each step, once the surface reaction is complete,
any excess precursor can no longer participate in the surface
reaction and is removed with inert gas during purging periods.
As a result, the film growth is self-limiting due to surface
saturation, making homogeneous reactant flux unnecessary.
The self-limiting character of ALD35 allows for easy and
accurate control of film thickness at the atomic layer level and
provides not only excellent large area uniformity but also high
conformality on complex shape surfaces such as lithographically
patterned surfaces and narrow holes. In order to deposit films
by ALD, precursors should be compounds that can perform the
surface saturation reactions. The first requirement for an
optimal precursor compound for ALD is that it should be
highly reactive not only toward the surfaces of substrates but
also toward the surface products of the complementary
precursors such as H2S, H2O2, O3, NH3, or chalcogen metal
complexes14 (e.g., alkylsilyl compounds of tellurium and
selenium). Second, precursors should be volatile at certain
(relatively low) temperatures and thermally stable at the film
growth temperature. In order for precursors to react at the
surface of a substrate, they should not decompose until they
arrive at the surface of the substrate in the ALD chamber
heated at the film growth temperature. Third, any surface
reaction byproducts should be nonreactive toward the surface
and noncorrosive so as not to etch the thin film.
Based on these requirements for ALD precursors, rac-N,N′-
di-tert-butyl-2,3-dimethylenediamine36−38 was chosen as a
suitable ligand for precursors. The bulky tert-butyl group on
each nitrogen limits oligomerization by its steric effect. The
coordination product for metal(II) using this ligand has a lower
molecular weight and higher volatility39 compared to metal(II)
amidinate complexes having two amidinate ligands. Finally, the
bidentate chelating effect is expected to increase the thermal
stability of the cyclic metal amide complexes (N-heterocyclic
stannylene and germylene) compared to that of monodentate
metal amide complexes.
To date, three metal−organic complexes containing rac-
N,N′-di-tert-butyl-2,3-dimethylenediamine have been reported:
1,37 2,5,38,40,41 and rac-1,3-di-tert-butyl-4,5-dimethyl-1,3-diaza-2-
silacyclopentane-2-ylidene.36 The synthesis of 2 was reported
almost simultaneously by Gordon et al.5,40,41 and Russel et al.,38
using slightly different synthetic procedures. Gordon et al. used
2 for ALD of SnO2, and Russel et al. studied the coordination
of 2 to transition metals. Russel et al. used n-butyl lithium to
make the lithium salt of rac-N,N′-di-tert-butyl-2,3-dimethylene-
diamine which is afforded by the protonation of the lithiated
intermediate obtained from the reaction of 1,4-di-tert-butyl-1,4-
diaza-1,3-butadiene and methyl lithium. In contrast, bypassing
isolation of the intermediate product, Gordon et al. used
methyl lithium and 1,4-di-tert-butyl-1,4-diaza-1,3-butadiene to
make the same lithium salt and allowed it to react with SnCl2.
For the other two compounds, intermediate coordination
products are afforded by the reaction of metal chloride
(H2SiCl236or GeCl437) in the presence of 1,4-diazabicyclo-
[2.2.2]octane (DABCO) and rac-N,N′-di-tert-butyl-2,3-
dimethylenediamine, obtained from the same synthetic
procedure reported by Russel et al. The final desired products
were obtained from the reaction with a reducing agent, such as
KC836 or Li.37 The synthetic procedure of 1 used in our group
was simplified compared to that of West et al.37 by using GeCl2·
dioxane and the dilithium salt of rac-N,N′-di-tert-butyl-2,3-
dimethylenediamine which is prepared in situ from 1,4-di-tert-
butyl-1,4-diaza-1,3-butadiene and MeLi. All the synthetic
methods described above for 1 and 2 are summarized in
Table S1. For ALD, the reactions of 2 with H2O2 and 3 with
2−5
H2S were used to deposit SnO2
and SnS24,29,42 films,
respectively, in our group. However, detailed synthetic
procedures and characterizations for both complexes (2 and
3) have not yet been reported.
Here, we report the synthetic procedure and characterization
of 3 and simplified synthetic procedures (in situ synthesis) and
relevant properties of compounds 1 and 2, which may be
advantageous in industrial production. We also demonstrate
their reaction with H2S to form ALD films of GeS and SnS at
relatively low temperature (ca. 50−100 °C). It was found that
the growth rate of SnS film using 2 was higher than that using 3
by ca. 0.3 Å/cycle. In addition, the ALD temperature
window35,43,44 of the SnS film using 2 (50−125 °C) is lower
than that of 3 (100−250 °C)42 and Sn(acac)2 (125−225
°C),30,45 enabling deposition on thermally sensitive substrates
such as transparent flexible polymers.
2. EXPERIMENTAL SECTION
The procedures were performed on a Schlenk line under an
atmosphere of dry nitrogen or argon inside a hood, or in a glovebox.
Tin dichloride (98%, anhydrous), germanium(II) chloride dioxane
complex (1:1), glyoxal solution (40 wt % in H2O), methyllithium
solution (1.6 M in diethyl ether), THF (anhydrous), pentane
(anhydrous), and diethyl ether (anhydrous) were purchased from
Sigma-Aldrich. tert-Butylamine was purchased from Acros. All
chemicals were used as received. All glassware used in the experiments
was dried in an oven at 160 °C. Although an ALD process for SnS
using 3 was previously reported,42 single crystal X-ray structure and
other characterization such as NMR, elemental analysis, and TGA
(Thermal Gravimetric Analysis) of 3 are first reported in this article.
2.1. Synthesis. Synthesis of 1,4-Di-tert-butyl-1,4-diaza-1,3-
butadiene (1,4-Diazadiene).46,47 A glyoxal solution (75 mL (40 w%
in H2O), 651 mmol) in distilled water (180 mL) was added dropwise
via syringe to a solution of tert-butylamine (180 mL, 1698 mmol) in
distilled water (150 mL) placed inside a round-bottom Schlenk flask
(1000 mL) at room temperature. During the reaction, a solidified
white product was observed in the reaction mixture. After 3 h of the
reaction under nitrogen, an additional 50 mL of distilled water were
added to the flask. The reaction mixture was stirred for 10 min and
filtered through a fritted glass filter. The filter cake was washed three
times with distilled water and air-dried with an aspirator. The crude
product was dissolved in diethyl ether, and then the diethyl ether layer
was decanted and evaporated. The lachrymatory white solid was
dissolved in a minimum amount of diethyl ether and recrystallized on
dry ice twice. The recrystallized solid product was dissolved in diethyl
ether, and the solution was dried over magnesium sulfate. Afterward,
magnesium sulfate was filtered out and the filtrate was evaporated
resulting in a white solid. The solid product was additionally purified
by sublimation at 40 °C under vacuum to give a white crystalline solid
(75 g, 68%). 1H NMR (CDCl3, 500 MHz): δ 7.95 (s, 18H), δ 1.27 (s,
2H).
3066
dx.doi.org/10.1021/cm403901y | Chem. Mater. 2014, 26, 3065−3073

Chemistry of Materials
Article
a
Scheme 1. Syntheses of ALD Precursors
a
rac-1,3-Di-tert-butyl-4,5-dimethyl-1,3-diaza-2-germacyclopentane-2-ylidene (1), also named 1,3,2-diazagermolidin-2-ylidene, 1,3-bis(1,1-dimethyl-
ethyl)-4,5-dimethyl- under CAS Registry Number 1167987-04-3; rac-1,3-di-tert-butyl-4,5-dimethyl-1,3-diaza-2-stannacyclopentane-2-ylidene (2),
also named 1,3,2-diazastannolidin-2-ylidene, 1,3-bis(1,1-dimethylethyl)-4,5-dimethyl-, (4R,5R)-rel- under CAS Registry Number: 1268357-44-3; and
bis(N,N′-diisopropylacetamidinato)tin(II) (3), also named tin, bis[N,N′-bis(1-methylethyl)ethanimidamidato-κN,κN′]-, (T-4)- under CAS Registry
number 1421599-46-3.
Synthesis of 1, [Ge(II)(η2-((NBut)CH(CH3)CH(CH3)(NBut))], rac-1,3-
Di-tert-butyl-4,5-dimethyl-1,3-diaza-2-germacyclopentane-2-yli-
dene. MeLi (111 mL of 1.6 M, 177 mmol) was added to anhydrous
diethyl ether (150 mL) in a round-bottom Schlenk flask (500 mL)
under Ar atmosphere. The solution was cooled to 0 °C in an ice bath.
To the solution, solid 1,4-di-tert-butyl-1,4-diaza-1,3-butadiene (14.53 g,
86.3 mmol) was added portionwise while stirring under a strong Ar
flow over a period of 20 min. The ice bath was removed, and the
reaction mixture was stirred at room temperature for 1 h. Anhydrous
tetrahydrofuran was added to the diethyl ether solution of the lithiated
ligand to make 300 mL of the solution. Then, the solution was
transferred via cannula to an addition funnel which was connected to a
round-bottom Schlenk flask (1000 mL) containing GeCl2·dioxane (20
g, 86.3 mmol) and THF (200 mL) (Scheme 1). Then, the round-
bottom Schlenk flask was immersed in a dry ice−acetone bath under a
strong Ar flow. Using the addition funnel, addition of the lithiated
ligand solution to the GeCl2·dioxane in THF was performed over a
period of 4 h after the flask was cooled down. After the addition was
completed, the reaction mixture was kept at −78 °C for 2 h and slowly
warmed up to room temperature. The reaction mixture was kept for
overnight (ca. 12 h). The volatiles were removed on an ice bath under
reduced pressure, and the crude product was extracted with pentane.
The residue was washed two times with fresh anhydrous pentane. The
pentane extract was evaporated on an ice−water bath under reduced
pressure. The obtained crude product was distilled at 40 °C under
vacuum in order to remove a colorless liquid byproduct. The desired
product was distilled (60 °C, 1 Torr) to yield a pale yellow liquid
(13.70 g, 59%). 1H NMR (C6D6, 500 MHz): δ 3.02 (q, J = 6.0, 2H), δ
1.28 (s, 18H) 4H), δ 1.11 (d, J = 6.0, 6H); 13C NMR (C6D6, 100
MHz, δ) 62.98, 55.03, 33.89, 27.46.37
solution of the lithiated ligand was transferred via cannula to an
addition funnel which was connected to the round-bottom Schlenk
flask (1000 mL) containing anhydrous SnCl2 (14.3 g, 75.5 mmol) and
diethyl ether (250 mL) (Scheme 1). Then, the round-bottom Schlenk
flask was immersed in a dry ice-acetone bath under strong Ar flow.
After the flask was cooled down, the rate of Ar flow was reduced.
Using the addition funnel, addition of the lithiated ligand solution to
the SnCl2 in diethyl ether was performed over a period of 4 h. After
the addition was completed, the reaction mixture was kept at −78 °C
for 4 h and then left on the dry ice−acetone bath without adding dry
ice until the bath reached room temperature. The reaction mixture was
additionally stirred overnight at room temperature. The diethyl ether
was removed on an ice bath under reduced pressure, and the crude
product was extracted with pentane. The residue was washed two
times with fresh anhydrous pentane. The pentane extract was
evaporated under reduced pressure resulting in an orange colored
solid. The crude product was dried overnight at room temperature
under vacuum. The solid in the Schlenk flask was transferred to a
sublimator inside a glovebox. Sublimation of the crude product was
performed from an oil bath kept between 45 and 50 °C, and water was
used as the coolant for the coldfinger. A light-orange crystalline solid
1
(17.8 g, 74%) was obtained. H NMR (C6D6, 500 MHz): δ 3.32 (q, J
= 6.0, 2H), δ 1.25 (s. 18H) 4H), δ 1.20 (d, J = 6.0, 6H); 13C NMR
(C6D6, 100 MHz, δ) 64.38, 55.99, 34.13, 28.64. EA Calcd (%): C,
45.46; H, 8.27; N, 8.84. Found (%): C, 45.35; H, 8.17; N, 8.66.
Synthesis of 3, [Sn(II)(η2-MeC(NPri)2)2], Bis(N,N′-diisopropyl-
acetamidinato)tin(II). 1,3-Diisopropylcarbodiimide (38.46 g or
47.72 mL, 305 mmol) was added dropwise via syringe to a
methyllithium solution (200 mL of 1.6 M, 320 mmol) in anhydrous
diethyl ether (250 mL) in a Schlenk flask (500 mL) on an ice bath
under an argon atmosphere. The ice bath was removed 10 min after
the addition, and the solution was stirred for 3 h. The lithium
amidinate solution was transferred via cannula to an addition funnel
connected to a round-bottom Schlenk flask (1000 mL) containing
anhydrous SnCl2 (28.90 g, 152 mmol) and diethyl ether (400 mL)
placed on a dry ice−acetone bath (−78 °C). Using the addition funnel,
the lithium amidinate was added dropwise to SnCl2 in diethyl ether
over a period of 2 h under a strong Ar flow. The dry ice−acetone bath
was removed 30 min after complete addition, and the reaction mixture
Synthesis of 2, [Sn(II)(η2-((NBut)CH(CH3)CH(CH3)(NBut))], rac-1,3-
Di-tert-butyl-4,5-dimethyl-1,3-diaza-2-stannacyclopentane-2-yli-
dene. MeLi (100 mL of 1.6 M, 160 mmol) was added to anhydrous
diethyl ether (150 mL) in a round-bottom Schlenk flask (500 mL)
under Ar atmosphere. The solution was cooled to 0 °C in an ice bath.
To the solution, solid 1,4-di-tert-butyl-1,4-diaza-1,3-butadiene (12.7 g,
75.5 mmol) was added portion-wise while stirring under strong Ar
flow over a period of 20 min. The ice bath was then removed and the
reaction mixture was stirred at room temperature for 1 h. Diethyl ether
3067
dx.doi.org/10.1021/cm403901y | Chem. Mater. 2014, 26, 3065−3073

Chemistry of Materials
Article
Figure 1. ORTEP plots (50% thermal ellipsoids) of the X-ray crystal structures of compounds 2 and 3. (a) Selected bond lengths (Å) and angles
(deg) for 2. Sn(1A)−N(1) = 2.026(7), Sn(1A)−N(2) = 2.055(8), N(1)−C(5) = 1.484(12), N(2)−C(7) = 1.492(12), C(5)−C(7) = 1.520(14),
N(1)−Sn(1A)−N(2) = 81.8(3), C(5)−N(1)−Sn(1A) = 113.6(6), C(1)−N(1)−Sn(1A) = 127.6(6), C(9)−N(2)−Sn(1A) = 129.1(6), C(7)−
N(2)−Sn(1A) = 111.7(6), N(1)−C(5)−C(7) = 108.7(8), N(1)−C(5)−C(6) = 113.6(9), N(2)−C(7)−C(5) = 108.2(8), N(2)−C(7)−C(8) =
112.0(8). (b) Selected bond lengths (Å) and angles (deg) for 3. Sn(1)−N(3) = 2.192(3), Sn(1)−N(2) = 2.195(2), Sn(1)−N(1) = 2.386(3),
Sn(1)−N(4) = 2.388(3), N(1)−C(4) = 1.325(4), N(2)−C(4) = 1.331(4), N(3)−C(12) = 1.333(4), N(4)−C(12) = 1.324(4), N(3)−Sn(1)−N(2)
= 98.12(9), N(3)−Sn(1)−N(1) = 92.88(9), N(2)−Sn(1)−N(1) = 57.82(9), N(1)−Sn(1)−N(4) = 136.40(8), C(4)−N(1)−Sn(1) = 89.80(18),
C(4)−N(2)−Sn(1) = 98.24(18), N(1)−C(4)−N(2) = 113.5(3).
was stirred overnight (ca. 16 h). Volatiles were removed under reduced
pressure, and the yellow solid residue was dissolved in pentane (ca.
450 mL). Insoluble organic and inorganic salts were filtered on Celite
and washed with pentane under Ar. The filtrate was evaporated under
reduced pressure. The resulting light yellow solid product was
transferred to the sublimator inside a glovebox. The sublimation was
performed at 90−95 °C under vacuum without coolant on the
coldfinger (air cooling), yielding a white crystalline solid (42 g, 69%).
A dark red residue was observed in the bottom of the sublimator after
the sublimation was completed. In cases where a red colored or “wet”
solid product was collected on the coldfinger, a second sublimation
could be performed for further purification. 1H NMR (C6D6, 500
MHz): δ 3.67 (sep, J = 6.5, 4H), δ 1.55 (s. 6H), δ 1.28 (d, J = 6.5,
24H); 13C NMR (C6D6, 100 MHz, δ) 165.4, 47.6, 25.5, 12.2. EA
Calcd (%): C, 47.90; H, 8.54; N, 13.97. Found (%): C, 47.72; H, 8.57;
N, 13.99.
2.2. ALD of SnS and GeS Films. GeS and SnS thin films were
deposited by atomic layer deposition (ALD) in a custom-built hot-wall
ALD reactor42,48 from the reaction of 1 or 2 with H2S, respectively.
The growth temperature was between 50 and 200 °C. The
temperatures of 1 and 2 precursor sources were kept at 40 and 45
°C, which give vapor pressures of 0.71 and 0.52 Torr, respectively. A
gas mixture of 4% H2S in N2 (Airgas Inc.) was used as the source of
sulfur. H2S is a toxic, corrosive, and flammable gas (lower flammable
limit of 4%) and should be handled with caution.49 Precursor vapors
assisted by N2 and a gas mixture of 4% H2S in N2 were injected
sequentially into the deposition chamber to allow chemical reactions
to occur successively on the substrate surface for 4 s. The purge time
after precursor vapor or H2S exposure was set to 10 s to completely
remove unreacted and byproduct species and prevent reaction in the
vapor phase. The deposition processes were done using the stop-flow
ALD mode, reported in detailed elsewhere.5,50
were included in the final refinements. Refinement details and
explanations are included in the individual CIF files. Crystallographic
data and final agreement factors are given in Table S2.
Thermal Gravimetric Analysis (TGA). TGA experiment was
performed by the TA Instruments Model Q50 system in the glovebox.
N2 was used as a flow gas. The heating rate and mass of samples used
in this experiment were set at 10 °C/min and ca. 40 mg, respectively.
Thin-Film Characterization. Surface morphology of the films were
examined by using field-emission scanning electron microscopy
(FESEM, Zeiss, Ultra-55). The film thickness was determined using
a combination of cross-sectional SEM and X-ray fluorescence
spectroscopy (XRF, Spectro, Xepos-III); the average thickness of
rough films was determined by XRF utilizing a calibration curve of Sn
Lα1 line intensity (count min−1) versus the film thickness of smooth
films measured by cross-sectional SEM. The elemental compositions
of the films were determined by Rutherford backscattering
spectrometry (RBS, Charles Evans RBS and Ionex 1.7 MV
Tandetron). The crystal structures of SnS and GeS films were
examined by X-ray diffraction (XRD, PANalytical X’Pert Pro) with Cu
Kα radiation (λ = 1.542 Å) using θ−2θ scan.
3. RESULTS AND DISCUSSION
Because 1 is liquid, single crystal X-ray analyses were performed
only for 2 and 3 (Figure 1). Crystallographic data for 2 (Table
S2) exactly matched with those reported by Russel et al.38 2 is
not a square planar type structure. The Sn atom lies out of the
plane made by N1, N2, N3, and N4 (Figure S5).
Figure 2 shows that all three compounds (1, 2, 3) show very
clean single-step vaporization and leave almost no residue.
These TGA results mean that all three compounds vaporize or
sublime without thermal decomposition. As shown in Figure 2,
it was observed that cyclic type complexes (1, 2) start to
vaporize at lower temperature compared to 3, the tin amidinate.
The dependence of the deposition rate of GeS and SnS on
the substrate temperature was studied in the range of 50 to 200
°C to estimate the ALD window; the results are shown in
Figure 3. The growth rate of SnS using 2 and H2S (Figure 3a)
was found to remain nearly constant at ca. 1.2 Å/cycle in the
temperature range 50−125 °C and decreased by more than half
to ca. 0.5 Å/cycle at 150−200 °C. The drop in growth rate
above a certain temperature was also observed in a ALD of
2.3. Characterizations. X-ray Structure Determination. The
structures of compounds 2 and 3 were determined. Diffraction quality
crystals were obtained by sublimation. Crystal mounting and data
collections were performed on a Siemens (Bruker) SMART CCD
diffractometer using Mo Kα radiation. Data reductions were
performed with SAINT, which corrects for Lorentz polarization and
decay. Space groups were assigned by analysis of symmetry. Systematic
absences were determined by XPREP and were further checked by
PLATON. Structures were solved by direct methods and refined
against all data in the 2θ ranges by full-matrix least-squares on F2 using
SHELXS-97 and SHELXL-97. Hydrogen atoms at idealized positions
3068
dx.doi.org/10.1021/cm403901y | Chem. Mater. 2014, 26, 3065−3073

Chemistry of Materials
Article
temperature range below 250 °C. The increase in growth rate
along with carbon contamination detected in SnS films above a
substrate temperature of 250 °C was attributed to the
decomposition of 3.42 Thus, we could conclude that 3 has
thermal stability up to 250 °C as in the case of 2.
Figure 4 shows the growth of SnS using 2 and H2S at
different temperatures. As shown in Figure 3a, the thicknesses
of SnS films are significantly decreased at 150 and 200 °C
(Figure 4c, d) compared to 70 °C. Furthermore, the SnS films
grown at 150 and 200 °C (Figure 4c, d) have less densely
packed crystalline grains, compared to those deposited at 70
°C.
From the examples of the growth rate drop described above,
two conclusions can be made about the drop in the growth rate
of SnS using 2 and H2S above 125 °C. First, the drop in the
growth rate is not produced by decomposition of coprecursor
H2S as in the case of ALD of SnO2 using H2O2. We already
know that H2S is thermally stable at least in the range 100−250
°C from ALD of SnS using 3 and H2S. Second, the drop in the
growth rate is not caused by the formation of denser SnS film as
in the case of ALD of SnS using nitric oxide (NO). As shown in
Figure 4, SnS films become less dense and rougher as the
temperature increases. Therefore, it is more likely that the drop
in the growth rate is produced by desorption of the surface
reaction product of 2 at higher temperature, although the
nature of the desorbed species could not be determined in our
experiments. Additionally, because the surface reaction product
of H2S in the ALD using 2 is the same as that of H2S in the
ALD using 3 (ALD window 100−250 °C),42 it is believed that
there is no desorption of surface reaction product of H2S in the
ALD of SnS using 2 in the temperature range 125−200 °C.
Unlike SnS, a constant growth rate of GeS over a wide
temperature range was not observed (Figure 3b). The growth
rate remained roughly constant at 0.28 Å/cycle between 50 and
75 °C. Above ∼75 °C, the growth rate of GeS decreased
significantly from 0.21 Å/cycle at 100 °C to almost no growth
(0.02 Å/cycle) in the temperature range 150−200 °C,
indicating an increase in precursor desorption with increasing
temperature as in the case of ALD SnS using 2 and H2S. From
the experiments of film growth rate vs substrate temperature for
GeS and SnS (Figure 3), the optimal temperature was chosen
as 70 °C in the ALD window for both compounds.
Figure 2. TGA curves of 1 (solid line), 2 (dash dotted line), and 3
(dashed line).
SnO2 we reported previously using 2.3,5 There are two ways for
ALD of SnO2 using 2 as a Sn source: with hydrogen peroxide
(H2O2) or with nitric oxide (NO). In the case of growing SnO2
film using 2 and H2O2, the growth rate started to drop above
150 °C (1.8 Å/cycle) and reached ca. 1.2 Å/cycle at 250 °C.5
This growth rate drop was explained by thermal decomposition
of H2O2.5 In the growth of SnO2 film using 2 and nitric oxide,
the growth rate started to drop above 125 °C (2.3 Å/cycle) and
reached ca. 1.4 Å/cycle at 200 °C.3 This growth rate, ca.1.4 Å/
cycle, was constant from 200 to 250 °C. The growth rate drop
was reported to be due to the formation of a denser SnO2 film.3
Because growth rate is calculated from dividing the film
thickness by the number of cycles, a denser film has a lower
growth rate because it is less thick compared to a lower density
film. In the ALDs using both combinations of 2 and H2O2 and
2 and NO, an increase in growth rate due to the precursor
decomposition, as in the case of 3 above 250 °C (vide infra),42
was not observed up to 250 °C. This result can be translated to
the thermal stability of 2, at least, up to 250 °C. Meanwhile, the
dependence of growth rate on substrate temperature for SnS
using 3 and H2S studied in our group showed that the growth
rate (0.9 Å/cycle) was constant in the temperature range 100−
250 °C (ALD window) and then started to increase above 250
°C.42 This means that neither 3 nor H2S decomposes in the
For both metal sulfide films studied, reactions between the
precursor and hydrogen sulfide are self-limiting. Saturation of
Figure 3. Film growth rate as a function of substrate temperature: (a) SnS (from 2) and (b) GeS (from 1).
3069
dx.doi.org/10.1021/cm403901y | Chem. Mater. 2014, 26, 3065−3073

Chemistry of Materials
Article
Figure 4. SEM images (top view and side view, scale bar: 250 nm) of SnS films (1000 cycles of 2 and H2S) at temperatures of (a) 70 °C, (b) 120 °C,
(c) 150 °C, and (d) 200 °C.
Figure 5. Data showing that film growth rate at constant temperature approaches a constant value (saturated, self-limiting surface reaction) as doses
of precursor are increased. (a) Growth rate of SnS films with 2 at fixed H2S exposure of 2.2 Torr·s. (b) Growth rate of SnS films with H2S at fixed 2
exposure of 2.4 Torr·s. (c) The thickness of SnS films on SiO2 substrates using 1.8 Torr·s of 2 and 1.1 Torr·s of H2S exposure as a function of the
number of ALD growth cycles at 70 °C: the linear relationship between metal sulfide film thickness and the number of deposition cycles. (d) Growth
rate of GeS films with 1 at fixed H2S exposure of 0.96 Torr·s, e) Growth rate of GeS films with H2S at fixed 1 exposure of 0.90 Torr·s. (f) The
thickness of GeS films on SiO2 substrates using 0.60 Torr·s of 1 and 0.96 Torr·s of H2S as a function of the number of ALD cycles at 70 °C: the
linear relationship between metal sulfide film thickness and the number of deposition cycles.
3070
dx.doi.org/10.1021/cm403901y | Chem. Mater. 2014, 26, 3065−3073

Chemistry of Materials
Article
Figure 6. Proposed mechanisms for the formation of SnS films (a) from 2 and (b) 3.
Figure 7. Cross-sectional SEM images of a hole with aspect ratio of ∼40:1 coated by (a) GeS, (b) SnS at 70 °C and magnified views of the top,
middle, bottom of the coated hole. (c) RBS spectra; (red dotted line) SnS, (black solid line) GeS. (d) XRD spectra of SnS and GeS films.
the growth rate was shown by increasing exposure in Figure 5a,
5b, 5d, 5e. The exposure (product of partial pressure and
exposed time) is the total flux of the precursor to the surface
that causes the surface reaction to go to completion. Figure 5a
show that 2 undergoes self-limiting reactions with the surface
prepared by the H2S. In this experiment (Figure 5a), exposure
3071
dx.doi.org/10.1021/cm403901y | Chem. Mater. 2014, 26, 3065−3073

Chemistry of Materials
Article
of H2S was fixed at 2.2 Torr·s. It was found that the saturation
of growth rate by 2 starts when its exposure is at ca. 1.8 Torr·s.
Figure 5b shows that H2S also undergoes self-limiting reactions
with the surface prepared by 2 whose exposure is fixed at 2.2
Torr·s. The saturation of growth rate by H2S starts when its
exposure is at ca. 1.1 Torr·s. By comparing the saturation points
of each 2 (1.8 Torr·s, Figure 5a) and H2S (1.1 Torr·s, Figure
5b), it reveals that higher exposure of 2 than that of H2S is
needed to saturate the ALD reactions. From this observation, it
seems that the surface reaction between H2S and the surface
prepared by 2 is more facile than that between 2 and the
surface prepared by H2S. At the minimum exposure of 2 (1.8
Torr·s) and H2S (1.1 Torr·s) to attain the surface saturation,
the relation between the thickness and number of cycles was
investigated (Figure 5c). From the slope of thickness versus
number of cycles, the ALD growth per cycle of SnS from 2 and
H2S was determined to be 1.23 Å/cycle.
in situ prepared lithium salt of the ligand. By using an in situ
synthetic method, synthesis of these complexes could be
simplified compared to previously reported methods. GeS films
and SnS films were successfully formed using 1 and 2 with H2S.
Those films were formed even inside the holes having a high
aspect ratio (∼40:1) with uniformity and conformality. By
producing 2 that has a relatively low temperature ALD window
range (50−125 °C), ALD of SnS is possible at temperatures
from 50 to 250 °C in combination with the Sn(II) amidinate
(3) (ALD window: 100−250 °C).
ASSOCIATED CONTENT
■
S
* Supporting Information
Photographic image of complexes (1, 2, 3), 1H, 13C NMR, and
X-ray crystallographic files in CIF format for the structure
determinations of the two compounds in Table S1. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Figure 5d shows that the Ge precursor (1) undergoes self-
limiting reactions with the surface prepared by H2S with fixed
exposure at 0.96 Torr·s. It was found that saturation of the
growth rate by 1 starts when its exposure is at ca. 0.60 Torr·s.
Figure 5e shows that H2S also undergoes self-limiting reactions
with the surface prepared by 1 at a fixed exposure of 0.90 Torr·
s. The saturation of growth rate by H2S starts when its exposure
is at ca. 0.30 Torr·s. From the comparison of the saturation
points of each 1 (0.60 Torr·s, Figure 5d) and H2S (0.30 Torr·s,
Figure 5e), it can be seen that higher exposure of 1 than that of
H2S is needed for surface saturation reaction. From this
observation, it seems, as in the case with 2 and H2S, that the
surface reaction between H2S and the surface prepared by 1 is
more facile than that between the Ge precursor (1) and the
surface prepared by H2S. The growth rate during ALD of GeS
was obtained at 0.60 Torr·s of 1 and 0.96 Torr·s of H2S
exposure and determined to be 0.28 Å/cycle (Figure 5f).
In the case of a tin(II) amidinate complex (3), two nitrogen
atoms of an amidinate ligand still coordinate to tin after 3 reacts
with thiol on the surface (Figure 6). Whereas, after cyclic tin
amide (2) reacts with thiol on the surface, one nitrogen atom is
released from tin and one side of the ligand can rotate freely.
Therefore, after surface reaction, the product of 2 can be
expected to occupy less surface space versus the surface
reaction product of 3. Thus, it is reasonable that the cyclic
complexes 2 (Scheme 1) can react with a higher density of
surface sites compared to 351−53 (Scheme 1). This difference
may be a reason for the difference in the observed ALD growth
rate of SnS film between 2 (1.23 Å/cycle) and 3 (0.90 Å/
cycle).42 However, this steric bulk argument does not explain
why GeS grows at a much slower rate (0.28 Å/cycle) from 1
than SnS does from compound 2, which has a similar structure.
Figure 7a and 7b show cross-sectional SEMs of SnS and GeS
deposited at 70 °C inside a deep hole with an aspect ratio of
∼40:1. The images show a uniform, conformal coating
indicating self-limiting ALD reactions. It was found by
Rutherford backscattering (RBS) that the Sn/S atomic ratio
in SnS film was 1:1 and that the Ge/S atomic ratio in GeS film
was 1:0.85 (Figure 7c). The XRD spectrum of SnS films shown
in Figure 7d exhibits polycrystalline orthorhombic structured
SnS (JCPDS# 39-0354).54,55 No peaks were found in the XRD
of GeS, indicating that it is amorphous.
AUTHOR INFORMATION
Corresponding Author
*E-mail: gordon@chemistry.harvard.edu.
■
Author Contributions
⊥These authors contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the U.S. National Science
Foundation under NSF Award No. CBET-1032955, the U.S.
Department of Energy SunShot Initiative under Contract No.
DEEE0005329. ASH acknowledges the Camille and Henry
Dreyfus Foundation for support as an Environmental
Chemistry Postdoctoral Fellow. This work was performed in
part at Harvard University’s Center for Nanoscale Systems
(CNS), a member of the National Nanotechnology Infra-
structure Network (NNIN), which is supported by the
National Science Foundation under NSF Award No. ECS-
0335765.
REFERENCES
■
(1) Anne, M.-L.; Keirsse, J.; Nazabal, V.; Hyodo, K.; Inoue, S.;
Boussard-Pledel, C.; Lhermite, H.; Charrier, J.; Yanakata, K.; Loreal,
O.; Le Person, J.; Colas, F.; Comper
7398.
̀
e, C.; Bureau, B. Sensors 2009, 9,
(2) Heo, J.; Kim, S. B.; Gordon, R. G. Appl. Phys. Lett. 2012, 101,
113507.
(3) Heo, J.; Kim, S. B.; Gordon, R. G. J. Mater. Chem. 2012, 22, 4599.
(4) Heo, J.; Liu, Y.; Sinsermsuksakul, P.; Li, Z.; Sun, L.; Noh, W.;
Gordon, R. G. J. Phys. Chem. C 2011, 115, 10277.
(5) Heo, J.; Hock, A. S.; Gordon, R. G. Chem. Mater. 2010, 22, 4964.
(6) Gordon, R. J. Non-Cryst. Solids 1997, 218, 81.
(7) Rosental, A.; Tarre, A.; Gerst, A.; Sundqvist, J.; Harsta, A.; Aidla,
A.; Aarik, J.; Sammelselg, V.; Uustare, T. Sens. Actuators, B 2003, 93,
552.
̊
(8) Martinson, A. B. F.; Elam, J. W.; Liu, J.; Pellin, M. J.; Marks, T. J.;
Hupp, J. T. Nano Lett. 2008, 8, 2862.
(9) Lee, Y. S.; Heo, J.; Siah, S. C.; Mailoa, J. P.; Brandt, R. E.; Kim, S.
B.; Gordon, R. G.; Buonassisi, T. Energy Environ. Sci. 2013, 6, 2112.
(10) Lee, Y. S.; Heo, J.; Winkler, M. T.; Siah, S. C.; Kim, S. B.;
Gordon, R. G.; Buonassisi, T. J. Mater. Chem. A 2013, 1, 15416.
(11) Antunez, P. D.; Buckley, J. J.; Brutchey, R. L. Nanoscale 2011, 3,
2399.
4. CONCLUSIONS
Cyclic Ge(II) amide (1) and Sn(II) amide (2) complexes were
synthesized by the reaction of M(II)Cl2 (M = Ge, Sn) and the
(12) Lee, S.-H.; Jung, Y.; Agarwal, R. Nat. Nano 2007, 2, 626.
3072
dx.doi.org/10.1021/cm403901y | Chem. Mater. 2014, 26, 3065−3073

Chemistry of Materials
Article
(13) Lee, S. Y.; Park, Y. S.; Yoon, S. M.; Jung, S. W.; Yu, B. G.
Electron. Lett. 2010, 46, 652.
(48) Hausmann, D. M.; de Rouffignac, P.; Smith, A.; Gordon, R.;
Monsma, D. Thin Solid Films 2003, 443, 1.
(49) Gordon, R. G. J. Chem. Phys. 1962, 37, 2587.
(14) Pore, V.; Hatanpaa, T.; Ritala, M.; Leskela, M. J. Am. Chem. Soc.
̈
̈
̈
(50) Karuturi, S. K.; Liu, L. J.; Su, L. T.; Zhao, Y.; Fan, H. J.; Ge, X.
C.; He, S. L.; Yoong, A. T. I. J. Phys. Chem. C 2010, 114, 14843.
(51) Zydor, A.; Kessler, V. G.; Elliott, S. D. Phys. Chem. Chem. Phys.
2012, 14, 7954.
(52) Puurunen, R. L. Chem. Vap. Deposition 2003, 9, 249.
(53) Halls, M. D.; Raghavachari, K. J. Phys. Chem. A 2004, 108, 2982.
(54) Devika, M.; Reddy, N. K.; Ramesh, K.; Gunasekhar, K. R.;
Gopal, E. S. R.; Reddy, K. T. R. Semicond. Sci. Technol. 2006, 21, 1125.
(55) Devika, M.; Reddy, N. K.; Reddy, D. S.; Reddy, S. V.; Ramesh,
K.; Gopal, E. S. R.; Gunasekhar, K. R.; Ganesan, V.; Hahn, Y. B. J.
Phys.: Condens. Matter 2007, 19, 306003.
2009, 131, 3478.
(15) Pore, V.; Knapas, K.; Hatanpaa, T.; Sarnet, T.; Kemell, M.;
̈
̈
Ritala, M.; Leskela, M.; Mizohata, K. Chem. Mater. 2010, 23, 247.
̈
(16) Knapas, K.; Hatanpaa, T.; Ritala, M.; Leskela, M. Chem. Mater.
̈
̈
̈
2009, 22, 1386.
(17) Zhang, J.; Wang, X.; Wang, X.; Ma, H.; Cheng, K.; Fan, Z.; Li,
Y.; Ji, A.; Yang, F. Appl. Phys. Lett. 2010, 96, 213505.
(18) Kwon, Y.; Kim, J.-H.; Chae, S.; Lee, Y.; Kim, S. G.; Ku, J.; Sohn,
Y.; Park, S. J. Kor. Phys. Soc. 2011, 59, 466.
(19) Yeung, F.; Ahn, S.-J.; Hwang, Y.-N.; Jeong, C.-W.; Song, Y.-J.;
Lee, S.-Y.; Lee, S.-H.; Ryoo, K.-C.; Park, J.-H.; Shin, J.-M.; Jeong, W.-
C.; Kim, Y.-T.; Koh, G.-H.; Jeong, G.-T.; Jeong, H.-S.; Kim, K. Jpn. J.
Appl. Phys. 2005, 44, 2691.
(20) Rajendra, B. V.; Kekuda, D. J. Mater. Sci.: Mater. Electron. 2012,
23, 1805.
(21) Ma, W.; Luther, J. M.; Zheng, H.; Wu, Y.; Alivisatos, A. P. Nano
Lett. 2009, 9, 1699.
(22) Ramasamy, K.; Malik, M. A.; O’Brien, P. Chem. Commun. 2012,
48, 5703.
(23) Barkhouse, D. A. R.; Gunawan, O.; Gokmen, T.; Todorov, T.
K.; Mitzi, D. B. Prog. Photovoltaics 2012, 20, 6.
(24) Sinsermsuksakul, P.; Hartman, K.; Kim, S. B.; Heo, J.; Sun, L.;
Park, H. H.; Chakraborty, R.; Buonassisi, T.; Gordon, R. G. Appl. Phys.
Lett. 2013, 102, 053901.
(25) Sun, L.; Haight, R.; Sinsermsuksakul, P.; Kim, S. B.; Park, H. H.;
Gordon, R. G. Appl. Phys. Lett. 2013, 103, 181904.
(26) Greyson, E. C.; Barton, J. E.; Odom, T. W. Small 2006, 2, 368.
(27) Jiang, T.; Ozin, G. A. J. Mater. Chem. 1998, 8, 1099.
(28) Alexandros, S.; Jason, M. S.; Christopher, A. C.; Andrew, G. C.;
Patrick, S. G.; Andrew, A. R. W. Nanotechnology 2010, 21, 185202.
(29) Sinsermsuksakul, P.; Chakraborty, R.; Kim, S. B.; Heald, S. M.;
Buonassisi, T.; Gordon, R. G. Chem. Mater. 2012, 24, 4556.
(30) Kim, J. Y.; George, S. M. J. Phys. Chem. C 2010, 114, 17597.
(31) Bernardes-Silva, A. C.; Mesquita, A. F.; de Moura Neto, E.;
Porto, A. O.; de Lima, G. M.; Ardisson, J. D.; Lameiras, F. S. Solid State
Commun. 2005, 135, 677.
(32) Liu, S.; Guo, X.; Li, M.; Zhang, W.-H.; Liu, X.; Li, C. Angew.
Chem., Int. Ed. 2011, 50, 12050.
(33) Mathews, N. R. Solar Energy 2012, 86, 1010.
(34) Knez, M.; Nielsch, K.; Niinisto, L. Adv. Mater. 2007, 19, 3425.
̈
(35) George, S. M. Chem. Rev. 2009, 110, 111.
(36) Li, W.; Hill, N. J.; Tomasik, A. C.; Bikzhanova, G.; West, R.
Organometallics 2006, 25, 3802.
(37) Tomasik, A. C.; Hill, N. J.; West, R. J. Organomet. Chem. 2009,
694, 2122.
(38) Mansell, S. M.; Herber, R. H.; Nowik, I.; Ross, D. H.; Russell, C.
A.; Wass, D. F. Inorg. Chem. 2011, 50, 2252.
(39) Li, Z.; Lee, D. K.; Coulter, M.; Rodriguez, L. N. J.; Gordon, R.
G. Dalton Trans. 2008, 0, 2592.
(40) Gordon, R. G.; Hock, A. S.; Heo, J.; Sinsermsuksakul, P.
Provisional application No. 61/320,069, filed on Apr. 1, 2010; US
patent application No. 13/077,241, filed on Mar. 31, 2011; Pub. No.:
US 2012/0027937 A1, Pub. Date: Feb. 2, 2012.
(41) Heo, J.; Hock, A. S.; Gordon, R. G. Proc. AVS Atomic Layer
Deposition Conf. 2010.
(42) Sinsermsuksakul, P.; Heo, J.; Noh, W.; Hock, A. S.; Gordon, R.
G. Adv. Energy Mater. 2011, 1, 1116.
(43) Cassir, M.; Goubin, F.; Bernay, C.; Vernoux, P.; Lincot, D. Appl.
Surf. Sci. 2002, 193, 120.
(44) Putkonen, M.; Sajavaara, T.; Niinisto, L.; Keinonen, J. Anal.
̈
Bioanal. Chem. 2005, 382, 1791.
(45) Bakke, J. R.; Pickrahn, K. L.; Brennan, T. P.; Bent, S. F.
Nanoscale 2011, 3, 3482.
(46) Kliegman, J. M.; Barnes, R. K. Tetrahedron 1970, 26, 2555.
(47) Mistryukov, E. A. Mendeleev Commun. 2006, 16, 258.
3073
dx.doi.org/10.1021/cm403901y | Chem. Mater. 2014, 26, 3065−3073