Precursors As Enablers of ALD Technology: Contributions From University of Helsinki

Coordination Chemistry Reviews 257 (2013) 32973322
Contents lists available at ScienceDirect
Coordination Chemistry Reviews

journal homepage: www.elsevier.com/locate/ccr
Review
Precursors as enablers of ALD technology: Contributions from

University of Helsinki
Timo Hatanp , Mikko Ritala, Markku Leskel
Laboratory of Inorganic Chemistry, Department of Chemistry, University of Helsinki, P.O.Box 55, 00014 University of Helsinki, Finland
Contents
1.
2.
3.
4.
5.
6.
7.
8.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3298
Alkaline earth metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3298
2.1.
-Diketonates of alkaline earth metals as precursors for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3299
2.2.
Cyclopentadienyl compounds of alkaline earth metals as precursors for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3300
Group 4 elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3305
3.1.
Group 4 metal halides as precursors for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3305
3.2.
Group 4 alkoxides as precursors for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3306
3.3.
Group 4 alkylamides as precursors for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3307
3.4.
Group 4 Cyclopentadienyl compounds for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3307
3.4.1.
Bis-cyclopentadienyl compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3307
3.4.2.
Mono-cyclopentadienyl compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3307
3.5.
Other group 4 precursors for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3308
Group 15 and 16 elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3309
4.1.
Alkylsilyl compounds of group 15 and 16 elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3309
Bismuth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3311
5.1.
Bismuth chloride as precursor for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3312
5.2.
Organometallic compounds of bismuth as precursors for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3312
5.3.
Tris(Bis(Trimethylsilyl)amido)Bismuth(III) as a precursor for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3312
5.4.
Bismuth compounds with oxygen based ligands as precursors for ALD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3313
5.4.1.
-Diketonates of bismuth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3313
5.4.2.
Bismuth alkoxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3313
5.4.3.
Bismuth carboxylates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3315
Pt group metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3315
Silver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3315
Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3319
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3319
Abbreviations: acac, acetylacetonate; ALD, atomic layer deposition; CET, capacitance equivalent thickness; CHT, cycloheptatrienyl; CHD, cyclohexadiene; CMOS, complementary metal oxide semiconductor; CN, coordination number; Cp, cyclopentadienide; COD, cyclooctadiene; CVD, chemical vapour deposition; dien, diethylenetriamine;
diglyme, bis(2-methoxyethyl) ether; DMAE, dimethylaminoethoxide; DRAM, Dynamic random access memory; EDS, energy dispersive spectrometry; EOT, Equivalent oxide
thickness; FESEM, eld emission scanning electron microscope; fod, 1,1,1,2,2,3,3-heptauro-7,7-dimethyl-4,6-octanedione; FRAM, ferroelectric random access memory;
GST, Germanium antimony telluride; hfac, hexauoroacetylacetonate; HRTEM, high resolution transmission electron microscopy; me, methoxyethoxy; mmp, 1-methoxy2-methyl-2-propanolate; MOCVD, metal organic chemical vapour deposition; MS, mass spectroscopy; MTHD, methoxy-2,2,6,6-tetramethylheptane-3,5-heptanedione, me;
METHD, methoxyethoxy-2,2,6,6-tetramethylheptane-3,5-heptanedione; MOSFET, metal oxide semiconductor eld effect transistor; NVRAM, non-volatile random access
memory; PCRAM, Phase change random access memory; PEALD, plasma enhanced atomic layer deposition; Piv, pivalate; REALD, radical enhanced atomic layer deposition; SBT, strontium bismuth tantalate; SDTA, single differential thermal analysis; SEM, scanning electron microscope; SIMS, Secondary ion mass spectrometry; SOFC,
solid oxide fuel cell; TG, thermo gravimetry; TGA, thermo gravimetric analysis; THF, tetrahydrofurane; tmhd, tetramethylheptanedione; Tp, tris(pyrazolyl)borate; TpEt2 ,
tris(3,5-diethylpyrazolyl)borate; trien, triethylenetetraamine; triglyme, triethylene glycol dimethyl ether.
Corresponding author. Tel.: +358 919150222, fax: +358 919150198.
E-mail address: timo.hatanpaa@helsinki. (T. Hatanp).
0010-8545/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ccr.2013.07.002
3298
a r t i c l e
T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322
i n f o
Article history:
Received 4 March 2013
Received in revised form 3 July 2013
Accepted 3 July 2013
Available online 18 July 2013
Keywords:
Atomic layer deposition
ALD precursors
High-K materials
Phase change memory
Alkaline earth metals
Group 4 metals
Bismuth
Iridium
Silver
a b s t r a c t
The review focuses on ALD precursors of selected elements such as alkaline earth (Mg, Ca, Sr, Ba), group
4 metals, bismuth, silver, iridium, selenium, tellurium and antimony. These elements are needed in different high tech applications but are challenging for ALD. Their precursor design needs careful balancing
between volatility, thermal stability and reactivitythe key properties of ALD precursors. The extensive
studies showed that cyclopentadienyl based precursors of alkaline earth metals are versatile ALD precursors which react with both water and ozone forming oxide. Group 4 ALD chemistry has been studied
very widely and many good precursors have been found for the oxide ALD. From a bunch of different
compound types studied the most promising ALD precursor for bismuth is Bi(OCMe2 i Pr)3 which shows
stable ALD process with water at 150250 C. The success in depositing noble metal lms by ALD can
be attributed more to the reactant rather than the metal precursor. Ru, Pt, Ir, Rh and Os lms can be
deposited from various organometallic and metal organic precursors using O2 as the other precursor.
Typically temperatures above 225 C are needed. Using O3 as a reactant lms can be deposited at lower
temperatures. Noble metal oxides are obtained below approx. 200 C and metallic lms above that. By
supplying both O3 and H2 as consecutive pulses, noble metal lms can be deposited well below 200 C.
For silver phosphine stabilized carboxylato and -diketonato complexes are thermally stable enough
enabling hydrogen plasma enhanced ALD of silver metal lms. Alkylsilyl compounds of selenium and
tellurium are versatile ALD precursors for metal selenide and telluride lms when combined with metal
chloride precursors. The use of alkylsilyl compounds is not limited to group 16 elements but can also
been used for group 15.
2013 Elsevier B.V. All rights reserved.
1. Introduction
The success of ALD is built on chemistry. The great benets of
ALD (conformality, uniformity, atomic level thickness control, etc.)
can be enjoyed only when proper precursors have been identied
for the material of an interest. The precursors need to meet a number of requirements (Table 1) among which the combination of
high reactivity and good thermal stability often sets major challenges. Volatility is important too, and though moderately volatile
solids are used even industrially, liquids and gases are preferred for
convenience.
The ALD chemistry has been approached from two directions:
scientic curiosity and from needs of industry. Over the years the
emphasis has been shifted toward the latter as the ALD has become
better known and more widely used thin lm deposition technique.
For example, when we in 1993 studied HfCl4 as a precursor for
ALD of HfO2 [1] this was done for comparison to the earlier studied TiCl4 [2], and neither we nor anyone else had a clue that ALD
HfO2 would become a high-k gate oxide in the leading MOSFETs
15 years later. By contrast, our recent breakthrough in ALD of tellurides and selenides with their alkylsilyl compounds as precursors
[3], was motivated by a request addressed to us from semiconductor companies via an ALD tool manufacturer that we collaborate
with. On the other hand, also in 1990s we did application motivated
precursor research but that focused on electroluminescent display materials rather than the current mainstream semiconductors.
Soon after the results and knowledge gained in that research were
successfully beneted in developing an ALD process for SrTiO3 [4]
that had gained increasing interest in IC applications because of its
very high-k value.
ALD metal precursors have traditionally been grouped based
on their ligands (Fig. 1): halides, -diketonates, alkoxides, alkyl
amides, alkyls, cyclopentadienyls, amidinates, guanidinates, etc.
and even elements [5]. Inside these groups the precursor properties
have been tuned to meet the ALD requirements by tailoring the ligands, particularly by changing the size of the ligands and by adding
donor functionalities to the hydrocarbon groups of the ligands.
Another option has been to add adduct ligand to the coordination
sphere of the metals. A newer trend has involved heteroleptic metal
complexes where two or more different ligands are used together
for further tuning of the precursor properties.
In this paper we review some of our own contributions to the
ALD precursor chemistry over the past 20 years. The aim of the
review is not to present a balanced view of the eld but to focus

on studies and discoveries done at the University of Helsinki. Thus
references to other studies are kept in minimum. A number of
different kinds of precursor groups will be covered. Each chapter
begins with a brief summary of applications of ALD lms containing
the given elements, thereby demonstrating how in ALD precursor chemistry basic and applied research are connected seamlessly.
Properties (structure, volatility, thermal stability, reactivity) of precursors studied and main results from the lm growth experiments
are reviewed. About the ALD reaction mechanisms only some illustrative examples will be given since in situ reaction mechanism
studies on ALD processes were reviewed in detail very recently by
Knapas and Ritala [6].
2. Alkaline earth metals
Alkaline-earth metals are constituents of many technologically
important materials. SrTiO3 and Ba1x Srx TiO3 are high permittivity dielectric materials for future DRAMs [7]. BaTiO3 is not
only a dielectric but also piezoelectric and ferroelectric material
which may be used in e.g. ceramic multilayer capacitors and thermistor elements [8,9]. CaS, SrS and BaS are host materials for
luminescent materials [10,11]. SrBi2 Ta2 O9 is a ferroelectric material which may be used in FRAMs [12]. Strontium and barium
are also constituents of the well-known superconducting cuprates
La2x Srx CuO2 , YBa2 Cu3 O7x and Bi2 Sr2 CaCu2 O8+x [13]. Fluorides of
magnesium and calcium are materials suited for optical coatings
[14].
Especially the larger alkaline earth metals are challenging from
the CVD/ALD precursor chemistry point of view: their compounds
have a tendency to form oligomeric species which have low
volatility. Thus sufciently volatile, thermally stable and reactive
compounds of alkaline earth metals are not that common.
Before the interest in ALD, a large variety of -diketonates and
also many other compounds had been used in CVD as precursors for alkaline earth metals. In addition a variety of compounds
had been suggested as possible precursors due to their volatility [15]. These compounds include silylamides and phosphides,
poly(pyrazaolyl)borates, alkoxides, -ketoiminates, pyrazolates,
pyrrolates and cyclopentadienyl compounds with and without
neutral ancillary Lewis base ligands. Precursors used for ALD of
alkaline-earth metal containing lms are quite few and they are collected into Table 2. Basically the precursors have been restricted to
3299
Fig. 1. General classes of metal precursors. Backbones of the ligands are shown and in addition the ligands may contain donorfuntionalities in the variable Rx parts.
-diketonates and cyclopentadienyl compounds which have been

extensively studied at the University of Helsinki too.
2.1. -Diketonates of alkaline earth metals as precursors for ALD
The rst alkaline-earth metal compounds introduced as precursors for ALD were -diketonates. These are the most well-known
volatile compounds of alkaline earth metals. Their volatilization
properties have been known for decades [116] but the real boom
came in connection to high-temperature superconductors. The
CVD of oxide superconductors required volatile compounds for
Ca, Sr and Ba as well as for rare earths [117]. The number of different -diketonato ligands is high but by far the most used is
the thd-ligand (thd = 2,2,6,6-tetramethyl-3,5-heptanedione). The
-diketonates have several advantages such as sufciently easy
synthesis and reasonable thermal stability but also several disadvantages like limited volatility because of oligomerization and low
reactivity. In thin lm growth experiments there have also been
problems which have been associated with some kind of instability
of the -diketonate precursors at their evaporation temperatures
[44]. This behavior may have been partly related to the quality of
the thd-compounds as the actual precursors used may have been
different O and OH groups containing oligomeric species formed
when different synthesis methods without exclusion of water and
air have been used, or the compounds have been handled or stored
incorrectly.
The large size and low charge of alkaline earth metal ions
and only bidentate character of the -diketonato ligands causes
coordinative unsaturation in M(-diketonato)2 complexes. The
unsaturation is decreased by oligomerization and oligomers with
four metal centers have been reported for Ba(thd)2 [118] and
if additional aqua, OH or O ligands are coordinated the size of
oligomers increases further [119]. By using neutral adduct ligands
such as polyethers and polyamines it is possible to decrease the
Table 1
Requirements for ALD precursors.
Requirement
Comment
Volatility
Volatility is needed for efcient

transportation. For stable ux liquids
and gases are preferable.
Decomposition of precursors is not
allowed. Instability would destroy the
self-limiting lm growth mechanism
of ALD.
Aggressive and complete reactions
with co-precursors to ensure fast
completion of surface reactions and
thus short cycle times, and for high
lm purity.
Competing reactions may prevent lm
growth.
To avoid corrosion. Readsorption of the
byproducts may slow the lm growth.
Would destroy the self-limiting growth
behavior.
Sufciently pure precursors to meet
the requirements specic to each
application.
Easy synthesis with good yield, simple
compounds with cheap ligands.
Stable at ambient air, if possible.
If possible.
Thermal stability
Reactivity
No etching of the lm or
substrate material
Unreactive and volatile
byproducts
No dissolution into the lm or
substrate
Purity
Inexpensive
Easy to handle
Non-toxic and
environmentally friendly
3300
molecule size to monomolecular one [120]. The volatility of the

alkaline earth thd complexes increases in series Ba < Sr < Ca < Mg
and this is in line with the oligomerization and ion size. Sr(thd)2
and Ca(thd)2 form trimers when crystallized in dry organic
solvents and also need an adduct ligand such as polyamines
or polyethers to be monomolecular [41,121]. The mass spectrometric studies show that also in gas phase the alkaline earth
thd-complexes exist in oligomeric forms with the dimeric species
being the dominant in the mass spectra [116,122]. Magnesium as
the smallest member in the series forms a dimeric thd-complex
[21] which again can be turned to monomeric by ancillary neutral ligands [123]. The mass spectrometric and sublimation studies
of adducted thd-complexes of the larger alkaline earth metals Sr
and Ba suggest that the neutral ligand dissociates upon sublimation or in the gas phasein this sense adducts with uorinated
-diketonato ligands are better because many of them evaporate as
intact molecules. The thd-complexes of Mg and Ca make an exception: diamine and polyamine adducts stay intact in the gas phase
below a certain dissociation temperature [41,123].
Alkaline earth metal thd-complexes have been used in ALD to
deposit the corresponding suldes and uorides using H2 S and HF
as the co-precursors, respectively [10,35,3741,4447,110]. CaS,
SrS and BaS lms doped with different rare earth metal ions
have extensively been studied as phosphor materials for electroluminescent at panel displays [10,11]. Fluorides have also been
deposited using reactions between the thd-complexes and TiF4 and
TaF5 .[1820,36] This method avoids the usage of highly corrosive
and toxic HF and also much higher growth rates were obtained,
e.g. with the Ca(thd)2 + TiF4 process a four times higher growth rate
(1.6 A/cycle)
compared to the Ca(thd)2 + HF process was achieved.
In contrast to H2 S and HF, water does not react effectively
with the -diketonates of larger alkaline earth metals at temperatures below decomposition [44]. This insufcient reactivity
of -diketonates is the main limitation for their usage in ALD of
oxide lms. However, growth of an oxide by the Cathd2 + H2 O process has been reported [32] and in two studies SrO also grew,
though with a low rate of 0.1 A/cycle,

from Sr(thd)2 and H2 O
at 250 C [50,55]. Also when Sr(thd)2 + H2 O cycles were mixed
with Ti(Oi Pr)4 + H2 O or Ti(Oi Pr)2 (thd)2 cycles SrTiO3 lms could be
grown [50,55].
Because water has low reactivity with -diketonates more reactive oxygen sources have been introduced. Ozone is the common
second choice for oxygen precursor in ALD of oxide lms. With thdcomplexes of the larger alkaline earth metals ozone however forms
carbonate lms [48]. It is possible to get rid of the carbonate by post
deposition annealing but it makes an additional process step and
involves a risk of defect formation. In contrast with the larger alkaline earth metals, Mg(thd)2 forms quite pure MgO with O3 [22] and
a constant growth rate of 0.27 A/cycle

is observed at 225250 C.
A third common oxygen source in ALD is oxygen plasma. SrO and
SrTiO3 have been deposited using Sr(METHD)2 , Ti(Oi Pr)4 and oxygen plasma as precursors and in contrast to the ozone processes
no carbonate was formed. The ALD window was reported also very
wide (150275 C) [63]. In one study MgO lm was grown from
[Mg2 (thd)4 ] and H2 O2 but with a low rate (0.100.15 A/cycle)

and
at relatively high temperatures of 325425 C [21].
2.2. Cyclopentadienyl compounds of alkaline earth metals as
precursors for ALD
In Cp compounds the metal is usually coordinated to all ve carbon atoms in the Cp ring (Fig. 2) and this gives shielding and stability
while still keeping the reactivity. The Cp rings can be substituted to
increase the bulkiness of the ligand and it is also possible to bridge
two rings together to an ansa-compound. Cp compounds are known
for most of the metals and many of them are volatile. Also a large
number of cyclopentadienyl compounds of strontium and barium

have been synthesized and characterized quite well [124,125]. The
Cp compounds of alkaline earth metals (Mg, Ca, Sr, Ba) are very
important in ALD because they enable the growth of oxide lms at
reasonable temperatures and growth rates using water as oxygen
source which is very difcult with the alternative thd-precursors.
Cyclopentadienyl compounds have been widely studied starting from the mid-eighties [124] with one of the main motivations
mentioned, in addition to general curiosity, being the possibility
to use these compounds as CVD precursors. However, before 1999
there were only two actual articles reporting usage of cyclopentadienyl compounds of heavier alkaline earth metals as precursors for
CVD [126,127]. In addition to these articles the usage of cyclopentadienyl compounds of strontium and barium in CVD has been
mentioned in several mostly Japanese patents starting as early as
1990 [128]. In contrast, cyclopentadienyl compounds of magnesium had been used in CVD more widely starting from at least 1989
[129]. Most obvious reason for the quite low number of reports
is that cyclopentadienyl compounds of alkaline earth metals are
extremely sensitive to air (oxygen) and moisture which makes it
quite difcult to use these compounds in CVD where all the precursors are introduced into the reactor simultaneously. Also for CVD
of thin lms containing alkaline earth metals there are many far
more suitable compounds than the cyclopentadienyls, like different -diketonates and -ketoiminates which have lower reactivity
and lower thermal stability, often desired in conventional CVD.
In our work we synthesized cyclopentadienyl compounds of
strontium and barium with different cyclopentadienyl ligands and
tested different well known strategies for making them more
volatile and thermally stable i.e. lled the coordination sphere of
the metal centers with multidentate Lewis base molecules, utilized large and sterically demanding cyclopentadienyl ligands to
cover the metal centers, and utilized donor functionalized ligands
(Fig. 2) [65,113,130,131]. Table 3 lists the compounds prepared and
studied.
Most of the as-prepared compounds contained coordinated THF
used as a solvent in the synthesis. After crystallization a xed
amount (x = 1 or 2) of THF was attached to each molecule. THF
could be removed by sublimation, heating under vacuum or by
using the solvent reux method. Thus the loosely bonded THF is
lost before or during the evaporation of the compounds and has no
relevant effect on the thermal properties of the compounds. The
isolated compounds were pure as characterized by elemental analysis, NMR and mass spectroscopy. All MS showed [ML2 ]+ species.
Ions where the neutral solvent or added Lewis base molecule
would have been coordinated to barium were not seen. For the
adducts Ba(Me5 C5 )2 A (A = dien, trien, diglyme, triglyme) this is
in contrast with the results of vacuum sublimation experiments
which resulted in sublimates with composition exactly the same as
before sublimation. Crystal structures were determined for compounds 1, 5, 7, 10, 11, 13 (Fig. 2), 14, 16, 17 (Fig. 2), 18 (Fig. 2)
and 19 [65,113,130132]. For the others, except the propyl bridged
bis-cyclopentadienyl compounds 8 and 9, and Ba(Me5 C5 )2 (trien)
(12), from which single crystals could not be grown, crystal structures were already known. All compounds with coordinated Lewis
bases are strictly monomeric. Ba(Me5 C5 )2 is known to form loose
chains where some methyl groups of neighbouring molecules are in
distance of interaction from the barium atom [133]. Similar interactions are found in the structure of Ba(t Bu3 C5 H2 )2 [130]. In the
compounds Ba(Me2 NC2 H4 C5 Me4 )2 (18) and Ba(EtOC2 H4 C5 Me4 )2
(19) the donor atoms O and N are coordinated to the metal center.
Melting point temperatures of the cyclopentadienyl compounds
prepared in general are quite high [113,125]. The lowest melting temperatures are observed for the compounds with bulkier
ligands i Pr3 C5 H2 and t Bu3 C5 H2 - , the lowest being 115 C for
Ba(i Pr3 C5 H2 )2 (THF)2 [132,134]. Most likely the i Pr groups in the
3301
Table 2
Group II precursors used in ALD.
Metal precursor
Other precursor
Material deposited
Reference
Mg
Te
MgTe
[16,17]
Mg(thd)2
TiF4 TaF5
H2 O2
O3
MgF2
MgF2
MgO
MgO
[18,19]
[20]
[21]
[22]
Mg(C5 H5 )2
H2 O
AlMe3 , O3
MgO
MgAl2 O4
[2330]
[30]
Mg(EtC5 H4 )2
Ca(thd)2
H2 O
H2 O
O3
O3 , (MeO)3 PO, H2 O
HF
TiF4 , TaF5
H2 S
MgO
CaO
CaO
Ca10 (PO4 )6 (OH)2
CaF2
[31]
[32]
[33,34]
[34]
[35]
[18,36]
[10,3741]
Ca(i Pr3 C5 H2 )2
H2 O
HfCl4 , H2 O
CaO
CaO-HfO2
[42]
Ca(Tp)2
H2 O
CaB2 O4
[43]
Sr(thd)2
HF
H2 S
O3
O3 , Ti(Oi Pr)4 , H2 O
H2 O, Ti(Oi Pr)4
Ti(Oi Pr)4 , O2 plasma
H2 O
H2 O, Ti(Oi Pr)2 (thd)2
La(thd)3 , Fe(thd)3 , O3
Ti(Oi Pr)4 , H2 O plasma
H2 S, Se
SrF2
SrS
SrO(SrCO3 )
SrTiO3
SrTiO3
SrTiO3
SrO
SrTiO3
La1x Srx FeO3
SrTiO3
SrS1x Sex
[35]
[10,38,4447]
[48]
[48,49]
[50,51]
[52]
[5254]
[55,58]
[59]
[60,61]
[62]
Sr(METHD)2
Sr(MTHD)2
Sr(Me5 C5 )2
O2 plasma
O3
H2 S
SrO, SrTiO3
SrO
SrS
[63]
[64]
[65]
Sr(n PrMe4 C5 )2
H2 O
La(Me4 C5 H)3 , Mn(Me4 C5 H)2 , H2 O
SrO
La2 O3 SrOMnO
[66]
[67]
Sr(i Pr3 C5 H2 )2
H2 O
O2
O2 , RuO4
O3
O2 plasma
Ti(Oi Pr)4 , H2 O
H2 O, Ti(Oi Pr)2 (thd)2 , O3
Ti(NEtMe)4 , O3
(Me5 C5 )Ti(OMe)3 , O3
(MeC5 H4 )2 Hf(OMe)Me, O2 plasma
H2 S
SrO
SrO
SrRuO3
SrO
SrO
SrTiO3
SrTiO3
SrTiO3
SrTiO3
SrHfO3
SrS
[6872]
[73]
[73]
[71,74]
[72]
[4,70,7577]
[78]
[79,80]
[74,81]
[72]
[65,82]
Sr(i Pr3 C5 H2 )2 (DME)
(Me5 C5 )Ti(OMe)3 , O2 plasma

H2 O
O3
(MeCp)Ti(OMe)3 , O3
SrTiO3
SrO
SrO
SrTiO3
SrTiO3
[83,85]
[71]
[71]
[71]
[71]
Sr(t Bu3 C5 H2 )2
H2 O
Ti(OMe)4 , H2 O
Ti(OMe)4 , O3
Ti(Oi Pr)4 , O2 plasma
(MeC5 H4 )2 Hf(OMe)Me, H2 O
SrO
SrTiO3
SrTiO3
SrTiO3
SrTiO3
SrO-HfO2
[86,87]
[86,99]
[100,101]
[102]
[87]
[72]
Sr(Tp)2
Sr(t But Am2 imidazolato)2
H2 O
O3
SrB2 O4
SrO
[103]
[104]
Sr[Ta(OEt)5 (me)]2
O2 plasma
H2 O
SrTa2 O6
[105]
[106,107]
CaS
3302
Table 2 (Continued )
Metal precursor
Other precursor
Material deposited
Reference
Sr[Ta(OEt)5 (dmae)]2
H2 O
O2 plasma
Bi(N(SiMe3 )2 )3 , H2 O
SrTa2 O6
SrBiTaO
[106,107]
[108]
[109]
Ba(thd)2
H2 S
BaS
[10,110]
Ba(Me5 C5 )2
Ti(Oi Pr)4 , H2 O
H2 S
BaTiO3
BaS
[4,75]
[65,82]
Ba(n PrMe4 C5 )2
Zr(NMe2 )4 , Y(MeC5 H4 )3 , H2 O
BaOZrO2 Y2 O3
[111]
Ba(t Bu3 C5 H2 )2
H2 O
Ti(Oi Pr)4 , H2 O
Ti(OMe)4 , H2 O
BaO
BaTiO3
BaTiO3
[75,112]
[113]
[112,114]
Ba(TpE t 2 )2
Ba(t But Am2 imidazolato)2
H2 O
O3
BaB2 O4
BaO
[115]
[104]
Fig. 2. Molecular structures of Ba(Me5 C5 )2 (diglyme), Ba(t Bu3 C5 H2 )2 and Ba(Me4 C5 C2 H4 NMe)2 representing the three different ways to pursue volatility and (thermal)
stability [113,130].
i Pr C H
3 5 2
ligand have freedom to bend and rotate so that the melting temperature is lowered. Adducts with multidentate Lewis bases
and compounds with donor functionalized Cp ligands had remarkably high melting temperatures well above 200 C.
In order to study the differences in stability and volatility
between the compounds thermogravimetric analyses (TGA) were
carried out under nitrogen atmosphere (Figs. 35) as is the common practice while evaluating ALD and CVD precursors. Typically
a dynamic TGA measurement is done i.e. the temperature of the
sample is increased with a constant heating rate while the weight
of the sample is followed as a function of time and temperature.
A TG curve showing a weight loss in a single step and residue
Table 3
Cyclopentadienyl compounds studied at Laboratory of inorganic Chemistry, University of Helsinki.
1
2
3
4
5
6
7
8
9
10
1114
15
16
17
18
19
20
Compound
Reference
Sr(C5 Me5 )2 (THF)2

Sr(i Pr3 C5 H2 )2 (THF)
Sr(i Pr3 C5 H2 )2
Sr(t Bu3 C5 H2 )2 (THF)
Sr(t Bu3 C5 H2 )2
Sr(C9 H7 )2 (THF)x
Sr(Me7 C9 )2 (THF)2
SrC3 H6 (i PrC5 H3 )2
SrC3 H6 (t BuC5 H3 )2
Ba(Me5 C5 H)2 (THF)x
Ba(C5 Me5 )2 (A) A = dien, trien, diglyme, triglyme
Ba(i Pr3 C5 H2 )2 (THF)2
Ba(t Bu3 C5 H2 )2 (THF)
Ba(t Bu3 C5 H2 )2
Ba(Me2 NC2 H4 C5 Me4 )2
Ba(EtOC2 H4 C5 Me4 )2
Ca(i Pr3 C5 H2 )2 (THF)2
[65]
[65]
[132]
[130]
[130]
[131,132]
[131,132]
[131,132]
[131,132]
[65,113]
[113]
[132]
[113]
[130]
[113]
[113]
[42]
Fig. 3. TG curves of cyclopentadienyl compounds of strontium. Flowing 1 atm N2

atmosphere, 10 C/min, 10 mg samples.
Fig. 4. TG curves of cyclopentadienyl compounds of barium with donor functionalized ligands and Me5 C5 ligands + ancillary Lewis base ligands. Flowing 1 atm N2
atmosphere, 10 C/min, 10 mg samples.
close to zero indicates a good volatility and thermal stability. TGA

gives information of thermal stability of a compound studied only
in the condensed state, not in the gas phase. Therefore when the
compound studied is highly volatile and there is no or very little
residue after the measurement it can only be said that the compound is thermally stable at least up to the temperature where the
evaporation process was complete. Larger residues indicate decomposition of the samples. Multiple steps indicate dissociation or
decomposition of the compounds studied, like dissociation of neutral adduct ligands at lower temperature followed by evaporation
of the main compound at higher temperature, or decomposition of
the precursor compound into a stable and non-volatile solid. Those
complexes which have protecting, sterically demanding ligands,
like i Pr3 C5 H2 and t Bu3 C5 H2 (25, 1517) were found to evaporate and leave residues below 8%. Compounds Sr(Me5 C5 )2 (THF)x
and Sr(Me5 C5 )2 (THF)x (x = 02) left 18 and 37% residues, respectively, indicating some decomposition. Under vacuum the above
mentioned compounds 15, 10 and 1517 sublimed at 110220 C
with a loss of possibly coordinated THF solvent. All the other complexes, i.e. those having pentamethyl substituted Cp + multidentate
Lewis base ligand (1114), indenyl (6, 7), N and O functionalized Cp
(18, 19) or ansa ligands (8, 9) seemed mainly to decompose in the
Fig.
5. TGA
curves
of
Ba(Me5 C5 )2 (THF)2
Ba(i Pr3 C5 H2 )2 (THF)2
and
Ba(t Bu3 C5 H2 )2 (THF). Inset shows TGA curves measured for Lewis base free
Sr(t Bu3 C5 H2 )2 and Ba(t Bu3 C5 H2 )2 . Flowing 1 atm N2 atmosphere, 10 C/min,
10 mg samples.
3303
TG measurements leaving 3070% residues. To make any conclusions what are the decomposition paths is quite impossible. Clearly
adding multidentate Lewis bases to Ba(Me5 C5 )2 (1114) did not
enhance but lowered the volatility as compared to the Lewis base
free Ba(Me5 C5 )2 or the THF adduct (Fig. 4). Also the thermal stability
was not enhanced. This was also conrmed by vacuum sublimation
experiments: vacuum sublimation of the adducts 1114 required
higher temperatures (180260 C) than pure Ba(Me5 C5 )2 and also
the sublimation yields were low (for 11 63% and for 1214 3239%)
though it seems that the compounds sublimed intact as the sublimates were found to be identical with the unsublimed compounds.
Noticeable is that TGA showed Ba(Me2 NC2 H4 C5 Me4 )2 (18) to be
stable up to ca 290 C above which it decomposes rapidly. Under
vacuum Ba(Me2 NC2 H4 C5 Me4 )2 (18), however, showed very good
sublimation behavior at 200260 C.
In conclusion the best approach for making volatile and thermally stable cyclopentadienyl compounds of Sr and Ba was
denitely the usage of large and sterically demanding cyclopentadienyl ligands. The i Pr3 C5 H2 and t Bu3 C5 H2 ligands were found to
make a good match with Sr and Ba. With similar ligands the thermal
properties of Sr and Ba compounds are very similar. Inset in Fig. 5
shows TG curves measured for Sr(t Bu3 C5 H2 )2 and Ba(t Bu3 C5 H2 )2 .
It seems likely that i Pr4 C5 H2 and (Me3 Si)3 C5 H2 ligands could
be equally good [135,136] but extremely crowded ligand i Pr5 C5
[137] may be even too bulky for Sr and Ba in the sense that
the reactivity with H2 O is lowered. However, making the tetra
and pentaisopropyl substituted cyclopentadienyl ligands is more
complicated compared to the trisubstituted ligands [135,137].
Adducting the cyclopentadienyl compounds or using donorfunctionalized cyclopentadienyl ligands was found not to be benecial.
Volatilities or thermal stabilities of adduct compounds were not
enhanced if compared to the Lewis base free compounds.
In general cyclopentadienyl compounds of alkaline earth metals
have been used especially in ALD of oxides of Mg, Ca, Sr and Ba
but also suldes of strontium and barium have been deposited
(Table 2). The advantages of the cyclopentadienyl compounds over
the -diketonates are high reactivity against oxygen precursors and
higher thermal stability. Oxygen precursors used with cyclopentadienyls are H2 O, O3 and oxygen plasma. Unlike the -diketonates of
Sr and Ba, the cyclopentadienyl compounds form sufciently pure
Sr and Ba containing lms with all three oxygen precursors.
SrS and BaS were deposited using Sr(Me2 C5 )2 , Ba(Me5 C5 )2 and
Sr(i Pr3 C5 H2 )2 as metal precursors and H2 S as a sulfur precursor
[65,82]. Growth rates of the processes at different temperatures
are shown in Fig. 6. Typical ALD window was observed for all
Fig. 6. Growth rate of SrS and BaS thin lms as a function of the deposition temperature. At 120 C, a pulse length of 1.5 s was used with complex 1 to ensure its uniform
distribution across the substrate [65]. Reprinted with permission from Chem. Mater.
14 (2002) 1937. Copyright 2002 American Chemical Society.
3304
three processes. As can be seen from the gure, the temperature

where the growth rate starts to increase is lower when Sr(Me5 C5 )2
(>350 C) and Ba(Me5 C5 )2 (>300 C) are used as precursors as
compared to Sr(i Pr3 C5 H2 )2 (>380 C) [82]. The increase of the
growth rate at high temperatures is associated to a decomposition of the precursors. The temperature independent growth rates
under the self-limiting conditions observed for the three precursors follow the order Ba(Me5 C5 )2 > Sr(Me5 C5 )2 > Sr(i Pr3 C5 H2 )2 as
expected: Ba is a larger cation than Sr, and i Pr3 C5 H2 is a larger
ligand than Me5 C5 requiring larger surface area in the saturatively
formed monolayer of the precursor on the surface of the growing
lm. The sulde lms deposited were all polycrystalline. Impurity levels in the lms deposited at 300 C were low (C 0.1 at.%, H
0.30.5 at.%, O 0.30.6 at.%). When compared with the previous SrS
and BaS processes using the thd compounds and H2 S, the usage
of the Cp compounds allowed lower deposition temperatures as
lower evaporation temperatures could be used. The Cp compounds
also resulted in higher or similar growth rates, depending on the
temperature, as the thd compounds. Roughnesses of the SrS lms
grown using the Cp compounds of strontium were also lower.
We have studied cyclopentadienyl compounds for both binary
and ternary oxides of the larger alkaline earth metals. Oxide
materials deposited have been CaO, SrO, BaO, SrTiO3 , BaTiO3 and
Ba1x Srx TiO3 . In all the processes H2 O was used as the oxygen precursor. Water is a difcult precursor for the binary oxides of the
larger alkaline earth metals because of the easy formation of metal
hydroxide [112]. The hydroxide formation and the reverse reaction
of dehydration back to the oxide are strongly dependent on the
process temperature and can greatly complicate the ALD growth
characteristics of the binary oxides. Fortunately in the far more
important ternary oxides, such as SrTiO3 and BaTiO3 , the hydration/dehydration processes are effectively stopped by the presence
of the other metal oxide [70,112].
The binary oxides of Ca, Sr and Ba are not important materials
due to their high reactivity in ambient atmosphere. However, SrO
and BaO growth experiments were carried out in order to study the
absorption of H2 O in the growing lms [70,112]. The precursors
used were Ba(t Bu3 C5 H2 )2 , Sr(i Pr3 C5 H2 )2 and Sr(t Bu3 C5 H2 )2 . It is
known that especially BaO is very keen in absorbing large amounts
H2 O and forming Ba(OH)2 (H2 O)n which could interfere the growth
of Ba oxide materials. It was found that at 250340 C the growth
chemistry of BaO is signicantly inuenced by the H2 O absorption. The amount of H2 O absorbed decreases with increasing H2 O
purge times and growth temperature. Up to 290 C the lms were
Ba(OH)2 . At 340 C there was still some Ba(OH)2 in the lms. SrO
has a slightly lower reactivity towards H2 O than BaO and it proved
to be much easier to grow uniform SrO lms. At 300 C the growth
rates were 0.42 and 0.28 A/cycle

for processes using Sr(i Pr3 C5 H2 )2
t
and Sr( Bu3 C5 H2 )2 , respectively. At this temperature the lms were
crystalline (1 1 1) oriented SrO. With calcium there is much less
problems with hydration. CaO was deposited using Ca(i Pr3 C5 H2 )2
and H2 O. At 300 C the growth rate was around 1 A/cycle

[42].
Finally, MgCp2 is a good precursor for ALD of MgO lms. Constant
growth rate of 1.16 A/cycle

could be achieved with the MgCp2 H2 O
process at 200300 C as reported by Putkonen et al. [26].
For SrTiO3 two sets of precursors were used: Sr(i Pr3 C5 H2 )2 ,
Ti(Oi Pr)4 , H2 O [4,76], and Sr(t Bu3 C5 H2 )2 , Ti(OMe)4 and H2 O [70].
With the rst set temperature range of 250325 C was studied.
The growth was the most ideal at 250 C giving highly uniform but
mostly amorphous lms with a growth rate of 0.5 A/cycle.

At 325 C
which is already above the decomposition limit of Ti(Oi Pr)4 the
SrTiO3 lms began to crystallize during the growth when exceeding approx. 50 nm thickness, rst with random orientation and
then in preferred (1 0 0) orientation. High water doses resulting
in Sr(OH)2 formation enhanced the crystallization and also the
substrates had an effect. On crystalline platinum the lms were
crystalline whereas on amorphous Al2 O3 they were amorphous

[70]. When SrTiO3 was deposited on an MBE-grown seed layer
at 325 C the lm grew epitaxially and no interface between the
MBE and ALD layers could be seen with HRTEM (Fig. 7). To deposit
stoichiometric SrTiO3 a cycling ratio of 5:6 (Sr/Ti) was found to be
appropriate.
With the second set of thermally more stable precursors
Sr(t Bu3 C5 H2 )2 and Ti(OMe)4 , SrTiO3 was grown at 300 C [70]. With
the 1:1 Sr:Ti cycling ratio a self-limiting growth rate of 0.4 A/cycle
and a Sr:Ti ratio of 3:7 in the lm were achieved. The 2:1 cycling
ratio was found to result in slightly Sr rich lms but reaching the
saturation was found to be more complicated and an increase in the
growth rate was observed. The average surface composition during
the SrO cycle clearly inuences the growth behaviour [70]. Surface
composition effecting the growth behaviour seems to be a general
feature in ALD of ternary compounds.
Dielectric properties of the SrTiO3 lms prepared were tested in
MIM capacitors. A 490 nm lm grown at 325 C using Sr(i Pr3 C5 H2 )2 ,
Ti(Oi Pr)4 and H2 O had a permittivity of 130 after annealing in air at
500 C. A 49 nm lm grown at 300 C using Sr(t Bu3 C5 H2 )2 , Ti(OMe)4
and H2 O and annealed at 750 C had a permittivity of 90 at 1 V,
resulting in EOT of 2.2 nm. Leakage currents of <50 nm lms were
fairly high, above 105 A/cm2 at 1 V [70].
Deposition of BaTiO3 was tested with three sets of precursors:
Ba(Me3 C5 )2 , Ti(Oi Pr)4 and H2 O [4], Ba(t Bu3 C5 H2 )2 , Ti(Oi Pr)4 and
H2 O [113], and Ba(t Bu3 C5 H2 )2 , Ti(OMe)4 and H2 O [112]. The rst
set resulted in reasonably uniform lms up to deposition temperatures of 275 C. With the second set crystalline BaTiO3 lms were
growth rate. The high growth

grown at 350 C with 1.0 A/cycle
rate observed in this preliminary study was associated to Ti(Oi Pr)4
decomposition and formation of barium hydroxide as an intermediate. At 250 C the process was difcult to control due to the
enhanced hydroxide formation at this temperature. Purge times
after water and barium precursor pulses strongly affected the
growth rates but still uniform lms could be deposited with xed
pulse lengths and long purge times after water. With the third
precursor set using Ti(OMe)4 instead of Ti(Oi Pr)4 higher process
temperatures could be used and self-limiting growth conditions
were found [112]. Films grown from Ba(t Bu3 C5 H2 )2 and Ti(OMe)4
at 340 C with a 1:1 pulsing ratio showed a Ba/Ti ratio of 0.56 in
the lm. With a pulsing ratio 2:1 the stoichiometry of the lm
improved to a Ba/Ti ratio of 1.05. Growth rates ranged from 0.37 and
MIM capacitors annealed at 800 C with platinum

to 0.50 A/cycle.
electrodes and 140 nm thick BaTiO3 lms showed permittivity of
280 and EOT below 2 nm.
ALD also allows the deposition of Ba1-x Srx TiO3 (BST) lms of different composition simply by changing the Ba/Sr pulsing ratio. BST
lms were deposited at 300 C using Sr(t Bu3 C5 H2 )2 , Ba(t Bu3 C5 H2 )2 ,
Ti(OMe)4 and H2 O as the precursors [70]. With a 1:1:1 Ba:Sr:Ti pulsing ratio the (Ba + Sr):Ti ratio was close to 1 but the Ba:Sr ratio varied
from 1.30:1 to 1.52:1 depending on the Ba precursor dose. In MIM
structures with Pt electrodes, after 2 min annealing at 750 C, 35 nm
BST layers showed a permittivity of 135, an EOT of 1.08 nm, and a
leakage current density of 1 105 A/cm2 at 1 V.
After the initial research on ALD of SrTiO3 and BaTiO3 with
small scale research reactor the processes were scaled-up to
larger reactors in co-operation with ASM Microchemistry [70,114].
While depositing SrTiO3 on larger substrates using Sr(i Pr3 C5 H2 )2 ,
Ti(Oi Pr)4 and H2 O the process temperature was limited to 250 C
due to the limited thermal stability of Ti(Oi Pr)4 that is more
readily observed on large substrates. Anyway lms of uniform
composition and thickness could be grown on 30 cm 40 cm
glass substrates [70]. Soon after the process with the improved
Ti and Sr precursors Ti(OMe)4 and Sr(t Bu3 C5 H2 )2 was adapted
by ASM Microchemistry for ALD of SrTiO3 on up to 300 mm
wafers. The BaTiO3 process using Ba(t Bu3 C5 H2 )2 , Ti(OMe)4 and
3305
Fig. 7. HRTEM images showing the inuence of an epitaxial MBE seed layer on ALD growth. A SiO2 layer which forms during the processing is seen between Si and the
epitaxial SrTiO3 [70].
H2 O was rst scaled up on 200 mm silicon wafers [114]. Later

these processes have been extensively studied at IMEC for device
integration [8699].
The reactivity of the Sr and Ba Cp compounds with water is
high and from that point of view there is no need for other oxygen
precursors [4,86]. But as noted, with water the hydroxide formation may cause complications. In addition, slow purging of water
especially from large batch reactors with high surface areas has
made industry to prefer ozone as an alternative oxygen precursor. STO lms can be grown with ozone but the state-of-the-art
Sr-rich ALD STO lms which show EOT < 0.5 nm with low leakage (107 A/cm2 ) at 1 V have been made with water using the
Sr(t Bu3 C5 H2 )2 H2 OTi(OMe)4 H2 O process at 250 C followed by
annealing at 600 C [88,92]. The structure of the MIM capacitor was
Ru/RuOx /Sr rich STO/TiN. The RuOx /STO interface may also contain
rutile TiO2 . Both RuOx and rutile TiO2 are benecial surfaces for
ALD of STO. They help the crystallization of the STO lm thereby
minimizing the dead-layer in STO. In addition TiO2 may increase
the Ti content in STO lms and RuOx can heal traps associated to
oxygen vacancies by releasing oxygen during annealing with the
both processes decreasing the leakage current density.
3. Group 4 elements
Oxides of the group 4 metals form one of the most studied materials group in ALD. The interest towards ZrO2 and HfO2 stems from
their good dielectric properties. ALD HfO2 is used industrially as a
high-k gate oxide in MOSFET transistors [138] and ALD ZrO2 as a
dielectric material in DRAM capacitors [139]. The high aspect ratios
in DRAMs and also the need to decrease the leakage current by making ZrO2 Al2 O3 ZrO2 nanolaminates call for the use of ALD in the
thin lm processing. TiO2 is a versatile material which also has a
high dielectric constant especially in rutile form but is not used
as a dielectric material because of the problems of high leakage
currents. TiO2 lms nd applications as photocatalysts, protective
coatings, and as a component in titanates and mixed oxide and
nanolaminates such as Alx Tiy Oz (ATO) that is used commercially in
the electroluminescent displays. Group 4 metal oxides have several
crystal structures with properties of their own. Therefore in thin
lm deposition besides the composition and purity also the formation of the desired crystal structure is important. In case of TiO2
the anatase form is preferred in photocatalysts, and the rutile form
when a high-k value is needed. ZrO2 and HfO2 have monoclinic,
tetragonal and cubic forms from which the last ones are desired in
high-k applications.
The recent development in Ti precursors for ALD aims at the
desired properties mentioned above: high reactivity, good thermal
stability, preferably liquid at room temperature, and good compatibility with Sr(R3 C5 H2 )2 precursors for deposition of SrTiO3 . The
latest complexes are heteroleptic following the idea that one of the
ligands gives reactivity and the other thermal stability.
3.1. Group 4 metal halides as precursors for ALD
Chlorides of group 4 metals are very important precursors since
they are volatile, stable, cheap and highly reactive with water. The
chloride ALD process for all these oxides has been known for over
20 years [1,2,140]. Chlorides have certain drawbacks such as ZrCl4
and HfCl4 are solids, the reaction by-product is corrosive HCl, and
incompatibility with some other lm constituents like strontium in
SrTiO3 that readily forms SrCl2 . The problem of being solid and possibly causing particle problems has been overcome and HfCl4 , for
example is used in industrial scale [141]. TiCl4 is extensively used
ALD precursor when cheap, easily delivered precursor is needed [5].
From the other halides iodides have been in some extent studied
as possible precursors for the TiO2 , ZrO2 and HfO2 lms [142,143].
The chloridewater process works at a wide temperature range:
from room temperature (with the highly volatile TiCl4 ) to at least
750 C. As mentioned above, in the growth of ZrO2 lms for the
high-k applications the crystal structure formed is important. The
most stable phase is the monoclinic one but its k-value is only
around 20 while the tetragonal and cubic phases have values of
3040. The ZrCl4 H2 O process produces amorphous lms below
210 C, tetragonal at 210300 C, and a mixture of tetragonal and
monoclinic at higher temperatures. Interestingly the k-values of
the lms remained rather low (around 20) despite the crystal
structure [144]. HfCl4 H2 O process on the other hand, produces
mainly monoclinic phase which also shows k-value close to 20
[145]. With decreasing lm thickness the fraction of the tetragonal
phase increases, however, as observed in for example HfO2 Ta2 O5
nanolaminates [145].
Metal chlorides react also with metal alkoxides producing either
a binary oxide or mixed oxide depending on whether the metals in
the two precursors are the same or different, respectively [146]. The
byproduct of this reaction is alkylchloride [147]. Because no separate oxygen precursor is used, and oxygen is bound to a metal
3306
Fig. 8. Thermal decomposition of Ti(OMe)4 , Ti(OEt)4 , Ti(Oi Pr)4 and Ti(Ot Bu)4 . ALD
equipment with in situ QMS was used to produce the data. The amount of decomposition was determined by dividing the main fragment of the ligand by a precursor
fragment [149].
with high oxygen afnity, the process oxidises the substrate or

underlying lm less than the normal oxide ALD processes.
3.2. Group 4 alkoxides as precursors for ALD
Metal alkoxides are common precursors in CVD of oxide lms.
In ALD simple titanium alkoxides (OMe, OEt, i OPr, t OBu) have been
widely studied and used with water to deposit TiO2 . Alkoxides of
Zr and Hf, by contrast, have been much less used in ALD because
of their lower thermal stability compared to the Ti alkoxides. The
smaller alkoxides tend to oligomerize and the oligomerization also
depends on the metal increasing in the order Ti < Zr Hf. On the
other hand, the stability of the alkoxides decreases in a series
OMe > OEt > Oi Pr > Ot Bu which has been best conrmed with titanium (Fig. 8) [148,149]. Monomolecular complexes are formed only
with iso-propoxy and t-butoxy ligands.
TiO2 with anatase structure can be grown from titanium alkoxides and water between 225 and 350 C. Because of thermal
decomposition of the alkoxides, the maximum temperatures with
the titanium ethoxide and i-propoxide are about 325 C while the
methoxide allows the use of 350 C [148,149]. Close to the decomposition onset temperature however the decomposition is so slow
that the lms are still reasonably uniform and a linear dependence of lm thickness on the number of deposition cycles is valid
[150,151]. The growth rate per cycle follows inversely the size of
the precursor being the highest with the smallest methoxide ligand
that gives a rate of 0.65 A/cycle

at 350 C.
Polycrystalline ZrO2 lms can be grown from Zr(Ot Bu)4 and
water at 200300 C. The growth rate decreased strongly when
the temperature was increased from 200 to 300 C. On the other
hand, the growth rate increased with higher precursor doses
and longer pulse times indicating decomposition of Zr(Ot Bu)4 .
However, the polycrystalline lms showed a high-k value of 32
indicating tetragonal or cubic structure [152]. Hf(Ot Bu)4 is also an
extremely air and moisture sensitive compound and as difcult to
use in ALD [153].
The poor thermal stability of zirconium and hafnium alkoxides
has been tried to be improved by using ligands that have been donor
functionalized by amine or ether groups. These have been used both
in homoleptic and heteroleptic complexes but with a general conclusion that thermal stability and high reactivity with water cannot
be achieved at the same time because coordinative saturation that
improves stability also decreases reactivity. The use of a dimeric
heteroleptic precursor [Zr(Ot Bu)2 (dmae)2 ]2 with primary alkoxy
Fig. 9. Crystal structure of [Zr(Ot Bu)2 (dmae)2 ]2 dimer [154]. Reproduced with a
permission from J. Non-Cryst. Solids 303 (2002) 24. Copyright 2002, Elsevier.
ligand dmae (dmae = dimethylaminoethoxide) (Fig. 9) allowed to

increase slightly the deposition temperature but thermal decomposition of the metal precursor was anyway seen in the experiments
[154]. The situation was similar when Ot Bu ligand was changed to
Oi Pr or a complete homoleptic dimethylaminoethoxide complex,
Zr(dmae)4 , was used [155]. The monomeric heteroleptic alkoxide
Hf(Ot Bu)2 (mmp)2 with tertiary mmp ligand (mmp = 1-methoxy-2methyl-2-propanolate) (Fig. 10) [156] showed rather good thermal
stability up to 275 C but at these temperatures the reactions with
water remain incomplete resulting in low growth rate and high
residue contents [157]. Fairly homogeneous lms could be grown
at 240 C but at 340 C some decomposition occurred which could
be seen in the increase of the growth rate from 0.58 A at 240 C
at 340 C. The homoleptic Hf(mmp)4 behaved quite

to 1.48 A/cycle
similarly [158]. A homoleptic hydroxylamide Hf(ONEt2 )4 , when
used together with water, showed a narrow growth temperature
range with the optimum being 300 C where reasonably good sat
was achieved [159]. The lms were quite
uration at 0.5 A/cycle
severely contaminated with hydrogen (11 at.%) and carbon (6 at.%),
however, while the nitrogen content was below 0.1 at.%. At 250 C
the growth was weak, showing that the bidentate hydroxylamide
Fig. 10. Crystal structure of M(Ot Bu)2 (mmp)2 (M = Zr, Hf) [156]. Reproduced with a
permission from Chem. Vap. Deposition 8 (2002) 163. Copyright 2002, John Wiley
and Sons.
3307
ligands shield hafnium more effectively than the related monodentate NR2 and OR ligands.
3.3. Group 4 alkylamides as precursors for ALD
Alkylamides form another important group of precursors for the
group 4 metals. The homoleptic tetrakis complexes used in ALD
usually contain dimethyl-, ethylmethyl- or diethyl-substituents at
the nitrogen. Especially for zirconium the ethylmethylamide is
widely used in oxide depositions. Alkylamides have several advantages: easy synthesis, high reactivity with water, relatively small
Besides oxide
molecules resulting in growth rates about 1 A/cycle.
lms alkylamides have been studied as precursors for nitride combined in thermal processes with ammonia or in plasma processes
with ammonia, nitrogen or hydrogen plasma [5].
Thermal stability of alkylamides of group 4 metals shows a
reverse order compared to alkoxides: titanium alkylamides are less
stable than the corresponding zirconium and hafnium complexes.
Thermal decomposition of Ti(NMe2 )4 has been reported to start at
180 C [160]. TiO2 lms have been grown from this precursor and
the growth rate is high at low temperatures (<100 C) and decreases
strongly towards 325 C [161]. Hausmann et al. [162] reported
much higher temperatures for zirconium and hafnium, the maximum deposition temperatures being for Zr(NMe2 )4 , Zr(NMeEt)4 ,
and Zr(NEt2 )4 250, 300 and 350 C, and for Hf(NMe2 )4 , Hf(NMeEt)4 ,
and Hf(NEt2 )4 350, 400 and 450 C, respectively. It seems that the
decomposition temperature depends on the reactor design since in
our reactors decomposition is seen already at the above mentioned
temperatures. For example, with Zr(NMeEt)4 there is a growth rate
difference between 275 and 300 C [163]. Low decomposition temperatures have also been observed for Hf(NMeEt)4 in batch reactors
[164].
Good results have been reported for ZrO2 lms grown from
Zr(NMeEt)4 and water or ozone. It is possible to reach permittivity
values close to 40 which is a sign of tetragonal or cubic structure
[165]. This can lead to CET values as low as 0.8 nm [166]. Hf(NMe2 )4
and Hf(NMeEt)4 precursors behave differently. The oxide lms
deposited from them with water show predominantly monoclinic
structure and permittivity values less than 20 [167,168]. This seems
to be more the property of hafnium oxide. High-k value forms of
HfO2 can be achieved by doping the lms for example with rare
earth ions. Good results have been reported for HfO2 doped with
7% of yttrium [169]. Although with ZrO2 it is easier to the tetragonal or cubic form than with HfO2 doping ZrO2 with for example
strontium assists in formation of these high permittivity phases
[170].
3.4. Group 4 Cyclopentadienyl compounds for ALD
3.4.1. Bis-cyclopentadienyl compounds
Bis-cyclopentadienyl (Cp) compounds represent true
organometallic compounds used as precursors for group 4
metal oxides. Two Cp rings coordinate in 5 fashion to the metal
and the other two ligands usually are either chlorides or methyl
groups. The bis-cyclopentadienyl compounds react rapidly with
both water and ozone and resulting oxide material is very pure.
The thermal stability of the bis-Cp compounds is higher than that
of alkoxides and alkylamides and deposition temperatures > 300 C
are achievable for ZrO2 and HfO2 [170]. Cp compounds of titanium
have only sparsely been studied in ALD.
The ALD of both ZrO2 and HfO2 has been studied using different bis-Cp compounds and both water and ozone as oxygen
precursors [171173]. The growth rate is typically 0.50.6 A/cycle

(Fig. 11). The in situ reaction mechanism studies showed that in
the Cp2 Zr(CH3 )2 water process 90% of methyl groups and 40%
of Cp groups were released during the metal precursor pulse in
Fig. 11. Growth rate of HfO2 lms as a function of deposition temperature. Precursors Cp2 HfMe2 and water. The inset shows the growth rate at 350 as a function of
the Cp2 HfMe2 pulse length [173]. Reproduced by permission of The Royal Society
of Chemistry.
reactions with the surface OH groups (Fig. 12) [174]. This is in

line with the expected bonding strength of the ligands. The polycrystalline lms of HfO2 show monoclinic structure while ZrO2
forms mixed phases. Thus, the presence of the low permittivity
monoclinic phase in ZrO2 has an effect on CET values which were
>1 nm.
Cp compounds of the group 4 metals are usually solids. Industry prefers liquid source materials and the tuning of the precursor
molecules to liquid at room temperature is one of the driving forces
in the precursor development. In bis-Cp compounds such a small
variation in structure as a change of one methyl ligand to a methoxy
ligand lowered the melting point below room temperature. The
studied (CpMe)2 Zr(OMe)Me compound showed similar thermal
stability as the bis-methyl compounds [175]. The methoxy substituted compound improved slightly also the other weak property of
Cp compounds, viz. low growth rate which increased by 0.1 A/cycle.

The reaction mechanism studies showed that methyl and methoxy
groups are mainly removed during the metal precursor pulse and
the CpMe ligand during water/ozone pulse [176]. XRD measurements showed that the ZrO2 lms formed contained mixed phases
and permittivity value remained quite low, however.
3.4.2. Mono-cyclopentadienyl compounds
As described above, the Cp-precursors offer good thermal stability but modest growth rate and tendency to form lms with a
mixture of oxide phases. Zirconium alkylamides, on the other hand
show higher volatility and growth rate but suffer from thermal
instability. Films grown from alkylamides show in addition better
phase outcome where the high permittivity tetragonal and cubic
phases are dominating. In mixed ligand complexes (CpR)Zr(NMe2 )3
(R = H, Me or Et) the benets of the both precursor types can be
combined [177]. Pure, conformal lms (Fig. 13) were grown by an
ozone process with a growth rate close to 1 A/cycle.

No signs of precursor decomposition were seen at 300 C. A CET value of 0.7 nm
with targeted low leakage current density was measured from a
MIM capacitor with 6 nm ZrO2 lm. When the similar cyclopentadienyl and alkylamido ligands were applied to hafnium the results
were not as good as with zirconium. HfO2 lms grew in the
(CpR)Hf(NMe2 )3 O3 process with a rate of 0.8 A/cycle

but the lms
contained mixed phases with the monoclinic phase being dominant
[178]. The as-deposited thin lms (<10 nm) were amorphous but
after annealing they contained mixed phases. Only yttrium doped
lms showed high permittivity structures.
3308
Fig. 12. Surface reactions in the Cp2 ZrMe2 /D2 O ALD process at 350 C as proposed
based on in situ studies with QMS and QCM. Upon adsorption of Cp2 ZrMe2 more
methyl than Cp ligands are released in exchange reactions with the surface OD
groups (b). The following D2 O pulse releases the rest of the ligands as MeD and CpD
releasedconverts the surface back to OD-terminated [174]. Reprinted with permission from Langmuir 21 (2005) 7321. Copyright 2005 American Chemical Society.
Mixed ligand complex (Me5 C5 )Ti(OMe)3 has been found to be

thermally highly stable[179181] but unfortunately they lack reactivity against water and more powerful oxidizers like O3 have to be
used. Initial ALD investigations revealed a very wide ALD process
window with deposition temperatures up to 375400 C [179,180].
Rose et al. [181] found that at 260330 C the growth rate increased
slowly from 0.22 to 0.29 A/cycle

and then raised up to 0.50 A/cycle
at 350 C. The density of lms was the highest at 330 C and dropped
signicantly at 350 C. However, the growth was still linear at
350 C. SrTiO3 was deposited at 250 and 300 C using Sr(i Pr3 C5 H2 )2 ,
(Me5 C5 )Ti(OMe)3 and O3 as precursors [81]. Films grown at 300 C
using SrO:TiO2 cycle sequences of 10:10 and 5:5 both possessed a
Sr/Ti atomic ratio of 0.90. The lms grown at 300 C and crystallized by annealing had high leakage while lms grown at 250 C
had much lower leakage.
3.5. Other group 4 precursors for ALD
Low thermal stability of titanium complexes has been a concern especially in deposition of SrTiO3 lms where it is desired
that the lms grow directly into the crystalline form. Compounds aimed to solve this have been e.g. the above mentioned
Fig. 13. Cross-sectional SEM images of a ZrO2 lm grown by the CpZr(NMe2 )3 /O3
process at 275 C into 60:1 aspect ratio trenches. The upper image shows the top
part of the trench, those below the middle and bottom parts. The trench opening is 115 nm and depth 6.75 mm. Labels denote the ZrO2 lm thicknesses [177].
Reproduced by permission of The Royal Society of Chemistry.
(Me5 C5 )Ti(OMe)3 and Ti(i OPr)2 (thd)2 [5557,71,79,81,8385,87].

Interestingly the recent approach to use heteroleptic alkylamidoalkoxo and acetamidinato-alkoxo complexes also improved the
thermal stability although the homoleptic parent compounds
show low thermal stability [182]. TiO2 lms can be grown from
Ti(NMe2 )2 (Oi Pr)2 and Ti(i PrNCHNi Pr)(Oi Pr)3 together with water
in a controlled way up to 350 C with a rates of 0.75 and 0.45 A/cycle.

Yet another approach to use heteroleptic complexes involves
a combination of cyclopentadienyl and cycloheptatrienyl
(CHT = C7 H7 ) ligands (Fig. 14) [183]. These complexes have
been synthesized for Ti and Zr. The compounds are thermally very
stable and growth temperatures exceeding 400 C are possible.
The oxide growth requires the use of ozone and growth rates of
have been recorded. In ZrO2 lms the high permit0.70.8 A/cycle

tivity phases are dominating and accordingly low CET values were
3309
Fig. 15. Molecular structure of (t BuMe2 Si)2 Te [3]. Reproduced with permission from
J. Am. Chem. Soc. 131 (2009) 3478. Copyright 2009 American Chemical Society.
harmful level. By contrast, the hydrides of selenium and tellurium

are much more dangerous and would require extensive safety systems and exhaust gas treatment. Alkyl compounds of selenium and
tellurium have been quite widely used in CVD but in ALD they do
not provide efcient exchange reactions with the common metal
precursors. In group 15 there is similar trend: NH3 can be used with
appropriate precautions while PH3 , AsH3 and SbH3 are much more
dangerous gases.
4.1. Alkylsilyl compounds of group 15 and 16 elements
Fig. 14. Molecular structure of (MeCp)Zr(C7 H7 ) (above) and TG curves measured

for CpTi(C7 H7 ) and (MeCp)Zr(C7 H7 ) [183]. Reprinted with permission from Chem.
Mater. 24 (2012) 2002. Copyright 2012 American Chemical Society.
reported. TiO2 lms on the other hand showed anatase structure

at low temperature (250 C), rutile structure at high temperatures
(400 C), and mixed phases in between.
4. Group 15 and 16 elements
Metal selenide and telluride thin lms have been gaining continuously increasing interest. Two well-known examples are CuInSe2
solar cell absorber and Ge2 Sb2 Te5 phase change material. Phasechange random-access memory (PCRAM) is a promising technology
for future non-volatile data storage [184186]. Reaching high integration densities together with low power consumption in PCRAMs
requires that memory cells are scaled into dimensions of a few
nanometers only. Connement of the phase change material into
small areas and volumes is most efciently done by depositing the
material into tiny vias. This requires excellent conformality from
the thin lm deposition method and makes ALD an attractive technique for replacing the so far dominant sputtering methods. From
the ALD chemistry point of view the common phase change material Ge2 Sb2 Te5 (GST) was seen a major challenge because it is made
of three elements that had not been studied much earlier and were
all considered difcult to ALD.
For non-metals hydrides are often an ideal choice as they allow
clean ligand elimination by protonation. However, hydrides of
many nonmetals are highly toxic. In group 16 water of course is
without this concern, and H2 S can be used with appropriate precautions because its odor threshold limit is much lower than the
A breakthrough in ALD of GST and related materials, including

selenides, was made when alkylsilyl compounds (R3 Si)2 M (M = Te,
Se) were identied as efcient ALD precursors with good volatility,
thermal stability and high reactivity [3]. Soon after, it was also realized that alkylsilyl compounds of group 15 elements have similar
properties [187,188]. Fig. 15 shows the structure of (t BuMe2 Si)2 Te
and Fig. 16 TG curves measured for a set of group 15 and 16 alkylsilyl compounds. TGA measurements reveal that all compounds of
the group 16 elements are volatile and they exhibit pure evaporation behavior. Their thermal stability also seems to be good.
Volatility of the compounds depends strongly on their molecular
weight. Compounds of group 15 elements lose weight at higher
temperatures compared to the group 16 Se and Te compounds.
Residues seen at the end of the measurements get larger in the order
As < Sb < Bi. DTA (not shown) revealed that the step of weight loss
is accompanied by heat release (exotherm) which indicates some
decomposition. However (Et3 Si)3 As leaves practically no residue
apparently because the decomposition products are volatile too. All
Fig. 16. TG curves measured for different alkylsilyl compounds of As, Sb, Bi, Se and
Te [188].
3310
Fig. 17. Temperature dependencies of ALD growth rates of various materials

deposited with metal chlorides and alkylsilyl precursors of tellurium and antimony
[3,187,190].
compounds studied are sensitive to air and moisture. Compounds

with larger alkylsilyl groups are less sensitive. Most of the compounds studied i.e. those with Me3 Si and Et3 Si ligands are liquids
while those with t BuMe2 Si are low melting solids.
Compounds of tellurium and selenium react efciently with various metal halides forming the corresponding metal tellurides and
selenides. For example, Sb2 Te3 , GeTe and GST lms were deposited
by ALD at remarkably low temperature of 90 C using (Et3 Si)2 Te,
SbCl3 and GeCl2 C4 H8 O2 as precursors, where C4 H8 O2 is a neutral
dioxane adduct ligand.
The alkylsilyl compounds of tellurium and selenium offer a
straightforward exchange reaction route to the deposition of the
metal telluride and efcient elimination of the chloride ligands of
the metal precursor, the overall reaction being:
(R 3 Si)2 Te(g) + MCl2 (g) MTe(s) + 2R 3 SiCl(g)
(1)
In situ reaction mechanism studies with a quartz crystal

microbalance and quadrupole mass spectrometer veried the
dechlorosilylation reaction mechanism and showed that the ligand
exchange and formation of R3 SiCl as a byproduct occur stepwise
during both metal chloride and alkylsilyl tellurium pulses [189].
The favorability of this reaction is supported by the Lewis hardsoft acidbase concept: in (R3 Si)2 Se and (R3 Si)2 Te silicon, a hard
Lewis acid, is bonded to the heavy group 16 elements which are soft
Lewis bases, thus forming an unfavorable hard-soft Lewis acidbase
pair. Upon exchange reaction with metal chlorides, silicon becomes
bonded to a harder base. Further, if the metal of the metal chloride
is a soft Lewis acid (large cation with a low charge), the other product of the exchange reaction will form a favorable softsoft Lewis
acidbase pair (MSe or MTe).
Both Sb2 Te3 and GeTe binary processes showed the self-limiting
behavior as expected in ALD, i.e. the growth rates saturated with
long enough precursor pulses. Despite the low deposition temperatures the saturation level was reached with reasonably short
pulse times of only 1 s. Temperature-wise the processes showed
peculiar features compared to the ALD processes in general. The
growth of Sb2 Te3 occurred with a good rate of about 0.6 A/cycle
at
as low temperature as 60 C but above that the growth rate dropped
rapidly (Fig. 17). Also with GeTe the highest rate was achieved at
the lowest possible deposition temperature of 90 C, the limitation coming from GeCl2 C4 H8 O2 that needs a source temperature
of 70 C [3,190]. Above 90 C the GeTe deposition rate decreased
rapidly and dropped to zero already at 150 C. Consequently the
largest concern related to these processes arises from the modest
volatility of GeCl2 C4 H8 O2 combined with the strong temperature
Fig. 18. GeSbTe phase diagram showing compositions reached so far with the
metal chloridealkylsilyl tellurium/antimony based ALD processes.
dependence of the growth rate. Various alternatives have been

looked for GeCl2 C4 H8 O2 but no better precursor has been found
yet. Reasons for these temperature dependences are still under discussion but they seem to be related to these specic lm materials
rather than this chemistry in general because zinc telluride could
be grown at a high temperature of 400 C using the same tellurium
precursor and ZnCl2 [3].
Ternary GST lms were deposited by mixing the binary ALD
was
cycles of GeTe and Sb2 Te3 . A growth rate of about 0.30 A/cycle
obtained where the cycle refers to a sum of binary cycles of GeTe
and SbTe [3,191]. A common way for controlling the ternary composition in ALD is to vary the cycle ratio of the binary constituents.
Ideally, the lm should adopt the stoichiometry corresponding to
the composition of the stable ternary compound but because GST
and related ternary materials are more alloy-like and many stoichiometric compositions exist along the pseudo-binary line between
GeTe and Sb2 Te3 (Ge1 Sb2 Te4 , Ge2 Sb2 Te5 , Ge3 Sb2 Te6 , etc.) there is
no strong driving force towards a particular stoichiometry. Indeed,
in the ternary process the antimony content increased and the germanium and tellurium contents decreased with increasing SbCl3
exposure times without approaching any specic composition.
Anyhow, the GST composition could still be adjusted close to the
stoichiometric Ge2 Sb2 Te5 by tuning the ratio of the binary cycles
GeTe and SbTe. On the other hand, if so desired, the composition
can be adjusted to other stoichiometries on the GeTeSb2 Te3 tie
line as well by just changing the binary cycle ratio.
The dechlorosilylation reaction was later extended to also elemental antimony (Fig. 17) and various metal antimonides, also
those of germanium [187]. Elemental Sb lms were deposited from
SbCl3 and (Et3 Si)3 Sb, while GeCl2 C4 H8 O2 and (Et3 Si)3 Sb deposited
a germaniumantimony alloy with a germanium rich composition of 82 at.% Ge and 18 at.% Sb. The germanium rich composition
points to a reaction mechanism more complicated than the simple
ligand exchange; most likely some antimony to germanium substitution is involved, too. Contrary to the GeSbTe system, GeSb
has no stoichiometric compound and tends to elemental segregation. Anyhow, by mixing the ALD cycles of elemental Sb with those
of Ge82 Sb18 , the lm composition could be tuned in between these
two extremes. Furthermore, when also GeTe and/or Sb2 Te3 ALD
cycles were added, the GeSbTe composition could be tuned over
a wide range of compositions in the ternary phase diagram (Fig. 18).
Despite the low deposition temperature, the GST lms deposited
from the germanium and antimony chlorides contained reasonably
low impurity contents of about 2.4 at.% oxygen, 1.0 at.% hydrogen,
0.7 at.% carbon, and 0.6 at.% chlorine as analyzed by Time-of-Flight

Energy Recoil Detection Analysis (TOF-ERDA) [3,191]. In elemental Sb lms no chlorine could be detected with ERDA [187]. The
silicon residues were more difcult to analyze because the lms
were on Si substrate, but the silicon content appeared to be below
1 at. %. Low levels of silicon and carbon in the GST lms have
also been veried with Secondary Ion Mass Spectrometry (SIMS)
[191].
The binary Sb2 Te3 and elemental Sb lms were found to grow
in polycrystalline form whereas with GeTe both amorphous and
polycrystalline lms have been obtained. The ternary GST has
usually been amorphous in the as-deposited state which can be
considered as a fortunate consequence of the low deposition temperature because strong crystallization during the growth could
prevent complete lling of the three-dimensional PCRAM memory
cell structures. With the amorphous GST and the GeTe lms complete lling of sub-100 nm holes in actual PCRAM device structures
was successfully demonstrated (Fig. 19).
Phase change properties of the ALD GST lms have been studied
with high temperature X-ray diffraction and resistivity measurements and compared to the sputtered GST lms. In general, the
phase change properties exhibited by the ALD GST lms have been
similar to their sputtered counterparts though in the ALD lms
the phase change temperatures and resistivities were somewhat
higher. The small amounts of impurities, oxygen in particular, in the
ALD GST lms likely explain these differences. Laser pulse experiments have veried that the ALD GST lms can be repeatedly
switched between amorphous and crystalline states.
While the early research on ALD of GST reviewed above was
done on lms made in small research reactor with 50 mm 50 mm
substrates, later the GST ALD process was scaled up to 200 mm
wafers with less than 3% 1 thickness nonuniformity [192].
This also allowed fabrication of integrated PCRAM devices which
demonstrated that the memory cells with ALD GST could be
switched between the low and high resistivity states up to 106
cycles with a very stable resistance for both states.
5. Bismuth
Many applications have been suggested in literature for thin
lms containing bismuth. Binary bismuth oxide can be used in
sensors and solid electrolytes (ionic conductors in solid state oxide
3311
Fig. 19. Transmission electron microscopy image of a GeTe lm ling a

31 nm 84 nm via indicated by the arrow [190]. Reproduced with permission from
J. Electrochem. Soc. 158 (2011) D694. Copyright 2011, The Electrochemical Society.
fuel cells) [193196]. Ternary and quaternary bismuth containing

metal oxides have a wide spectrum of potential applications. These
include various data storage devices (DRAM, FRAM, magnetic,
optical, holographic), electro-optical modulators, oxidation catalysts, gas sensors, solid oxide fuel cells, photocatalysts, lithium ion
batteries, lasers, optical amplier, superconductors, and oxygen
conducting solid electrolytes [193]. Bismuth suldes, selenides
and tellurides are other materials of interest. They nd use for
example in thermoelectrics [197].
Unlike for example in the case of the larger alkaline earth metals
strontium and barium even the simplest compounds of bismuth
, like BiCl3 , BiMe3 and Bi(Ot Bu)3 are quite volatile. This is because
bismuth tends to form compounds that are monomeric or have
a low oligomerization degree which in turn may be explained by
the sufciently high electronegativity (2.02) and charge to ionic
Table 4
Compounds of bismuth used as precursors in ALD.
Precursor
Other precursors
Material
References
BiCl3
(Et3 Si)2 Se (Et3 Si)2 Te

H2 S
Bi2 Se3
Bi2 Te3
Bi2 S3
[3]
[3]
[200]
Bi(Ph)3
Ti(Oi Pr)3 + H2 O
O3 + Ti(Oi Pr)4 + H2 O
Bi(Ph)3 and SrTa2 (OEt)10 me2 in n-butylacetate + O2 plasma
BiTiO pyrochlore
Bi4 Ti3 O12
SrBi2 Ta2 O9
[201]
[202]
[203]
Bi(CH2 SiMe3 )3
O3
BiSiO
[204]
Bi(N(SiMe3 )2 )3
Ta(OEt)5 + H2 O
Ti(OMe)4 + H2 O
Ti(OMe)4 + H2 O
SrTa2 (OEt)10 (dmae)2 + H2 O
BiTaO
BiTiO pyrochlore
Bi4 Ti3 O12
SrBi2 Ta2 O9
[109]
[205]
[205]
[109]
Bi(thd)3
Bi(OCMe2 i Pr)3
H2 O
H2 O
Bi2 O3
Bi2 O3
[206,208]
[206]
Bi(mmp)3
Bi(mmp)3 and Ti(mmp)4 in ethylcyclohexane + O3 /O2 Ti(Oi Pr)5 + H2 O

Ti(Oi Pr)4 + Si(OEt)4 + H2 O
Bi(mmp)3 , Ti(mmp)4 and Si(OEt)4 in ethylcyclohexane + O3
Bi(mmp)3 and Ti(mmp)4 in ethylcyclohexane, AlMe3 + H2 O
Bi4 Ti3 O12

BiTiO pyrochlore
BiTiSiO
BiTiAlO
[209]
[210]
[211,212]
[213]
3312
radius ratio (Bi3+ , 2.91(CN = 6)) of bismuth [198,199]. So volatility

is not the main issue while trying to nd suitable ALD precursors
for bismuth. The issues are more like insufcient reactivity and low
thermal stability.
ALD processes have been reported for several bismuth oxide and
chalcogenide materials. These processes are collected into Table 4.
Successful ALD has been reported for eight bismuth precursors:
BiCl3 , BiPh3 , Bi(CH2 SiMe3 )3 , Bi(N(SiMe3 )2 )3 , Bi(thd)3 , Bi(mmp)2 ,
Bi(Ot Bu)3 and Bi(OCMe2 i Pr)3 .
5.1. Bismuth chloride as precursor for ALD

BiCl3 is a volatile and thermally stable compound with a vapor
pressure of 0.1 mbar at 173 C [214]. Reported source temperatures in ALD have been 140 C [3]. In the reactions with (Et3 Si)2 Se
and (Et3 Si)2 Te selenides and tellurides of bismuth are easily
formed [3]. Also Bi2 S3 can be deposited using H2 S as the sulfur precursor [200]. With these reactants formation of bismuth
chalcogenides is favorable and this may be partially explained by
the soft Lewis acid Bi3+ coming together with the soft Lewis base
(S2 , Se2 , Te2 ). The problem with BiCl3 is its inability to form
pure oxide lms. With water, O2 and O3 the lms formed are
oxychloride BiOCl or otherwise heavily contaminated by chlorine
[215].
5.2. Organometallic compounds of bismuth as precursors for ALD

Bi(Ph)3 [201203] and Bi(CH2 SiMe3 )3 [204] are the only true
organometallic bismuth compounds that have been used for ALD.
Trialkyl and triaryl compounds of bismuth are well volatile and
they have been used largely in MOCVD. Down sides of these
compounds is their lack of reactivity against H2 O, their potential
explosion hazard and harmfulness [216]. Though Bi(Ph)3 does not
form any lm with H2 O, BiTiO lms with Bi/Ti ratios up to 0.6
were achieved when Bi(Ph)3 H2 O reaction cycles were mixed with
titanium tetraisopropoxideH2 O cycles. Further Bi incorporation
was difcult, however [201]. In addition, bismuth was incorporated into the lms in a metallic form rather than as an oxide, and
post-deposition annealing was needed to oxidize bismuth. With
O3 Bi(Ph)3 forms dark non-uniform lms [202]. When Bi(Ph)3 + O3
cycles were mixed with Ti(Oi Pr)4 + H2 O cycles BiTiO could be
grown [202]. In liquid injection plasma enhanced ALD of strontium
bismuth tantalate (SBT) Bi(Ph)3 has been used as a component in
the solution [203]. Simultaneous injection of two or more metal
precursors can, however, lead to composition nonuniformity while
coating deep trench structures because the precursors are competing for the same chemisorption sites. Nevertheless, uniform
SBT composition over a 5-in. wafer was reported by Shin et al.
[203].
Bismuth silicates, Bix Siy Oz have been deposited from
Bi(CH2 SiMe3 )3 and O3 at temperatures between 200 and 450 C
[204]. Fig. 20 shows the structure and TGA curve measured for
Bi(CH2 SiMe3 )3 . TG curve for Bi(i Pr)3 is included for volatility comparison. In the Bi(CH2 SiMe3 )3 O3 process Bi(CH2 SiMe3 )3 acted as
a precursor for both bismuth and silicon. Constant growth rate
of 0.4 A/cycle
independent of the Bi(CH2 SiMe3 )3 precursor pulse
time above 1 s was observed at 250350 C. Films deposited above
350 C showed a clear thickness prole indicating decomposition
of the precursor. The Si/Bi ratio increased with the increasing
deposition temperature but in the lms deposited at 250350 C
the ratio was around 2. The as-deposited lms were amorphous
but annealing either in N2 or O2 atmosphere resulted in crystalline
lms showing the presence of orthorhombic Bi2 SiO5 phase and
some other bismuth silicates.
Fig. 20. Above: Molecular structure of Bi(CH2 SiMe3 )3 . Below: TG curves of Bi(i Pr)3
and Bi(CH2 SiMe3 )3 [204].
5.3. Tris(Bis(Trimethylsilyl)amido)Bismuth(III) as a precursor for

ALD
Even the simplest alkylamides of bismuth like Bi(NMe2 )3 are
known to be monomeric and quite volatile [217], and they have
been mentioned as possible MOCVD precursors though actual
usage has not been reported. Down sides of these compounds
considering ALD usage are their light sensitivity, extremely high
sensitivity to air and moisture, and thermal instability. Compared
to simple alkylamides the silylamide Bi(N(SiMe3 )2 )3 is more stable
and less sensitive [109]. TGA under normal pressure N2 atmosphere
shows a single step weight loss of 98% at 220270 C (Fig. 21). This
result suggests that Bi(N(SiMe3 )2 )3 is stable even up to 270 C, but
ALD growth experiments implied that decomposition takes place
already at temperatures exceeding 200 C. The compound melts at
133137 C and was a solid at 120 C ALD source temperature used.
ALD Growth of BiOx was achieved at very narrow temperature window between 190 and 200 C using Bi(N(SiMe3 )2 )3 and H2 O as the
precursors. Uniform amorphous BiOx thin lms could be deposited
with growth rates between 0.08 and 0.23 A/cycle,

but the reproducibility of the process was poor [109]. The irreproducibility in
the binary oxide growth was thought to be due to partial reduction
of bismuth to metallic form, leading to termination of the growth
[205]. Other peculiarities in the BiOx growth were also observed.
FESEM revealed that the grown amorphous BiOx lms were granular while amorphous lms are usually smooth and featureless. The
3313
Fig. 22. Sawyer-Tower QV loop of a Pt/SBT/Pt capacitor structure [70].
70 and remanent polarization of 0.75 C/cm2 were measured for

190 nm SBT lm (Fig. 22) [70].
Mixing Bi(N(SiMe3 )2 )3 H2 O cycles with Ti(OMe)4 H2 O cycles
in 1:1 ratio resulted in smooth amorphous lms with a Bi/Ti ratio of
0.33. 7:1 cycle ratio was needed for obtaining phase-pure Bi4 Ti3 O12
lms with sufcient Bi content. Crystallization of the BiTiO lms
was carried out at 500750 C. Ferroelectric lms with remanent
polarizations up to 0.5 C/cm2 (51 nm lm) were prepared on
Pt/SiO2 /Si-substrates by annealing rst at 500 C and then at 600 C
[70].
Main limitations of the Bi(N(SiMe3 )2 )3 precursor are its low
thermal stability and the high evaporation temperature needed
that together limit the growth temperature to very narrow range
of 190200 C. The tendency of the precursor to deposit binary bismuth oxide unreliably is a further limitation. In addition, it is also
likely that with O3 there would be silicon incorporation in the lms.
5.4. Bismuth compounds with oxygen based ligands as precursors
for ALD
Fig. 21. Above: Crystal structure of Bi(N(SiMe3 )2 )3 . Below: TG curve of
Bi(N(SiMe3 )2 )3 (N2 , 1 atm, 10 C/min, 10 mg sample) [109].
granular morphology was thought to originate from a reduction of

bismuth to an elemental form at some point of the growth, followed
by its reoxidation during the further deposition cycles as there was
no metallic bismuth in the lms [205].
Deposition of tantalum and titanium containing multicomponent lms was found more reproducible than BiOx
deposition. It is likely that mixing the BiOx growth cycles with the
other metal oxide cycles stabilizes the growth. Multicomponent
BiTaO, SrBiTaO [109] and Bi4 Ti3 O12 [205] lms were
made by mixing Bi(N(SiMe3 )2 )3 H2 O cycles with Ta(OEt)5 H2 O,
SrTa2 (OEt)10 (dmae)2 H2 O and Ti(OMe)4 H2 O cycles. The lm
growth was similarly found to be limited to 190200 C and below
by the thermal stability of Bi(N(SiMe3 )2 )3 . In the case of all the multicomponent oxides the metal component ratios could be varied by
varying the binary oxide cycle ratios.
When BiTaO was deposited at 190 C using a xed BiOx /Ta2 O5
cycle ratio of 2:1 BiTaO the growth rate was 0.33 A/cycle
and
the Bi/Ta ratio was around 1.1. Amorphous SrBiTaO (SBT) lms
were deposited by mixing two Bi(N(SiMe3 )2 )3 H2 O cycles with one
SrTa2 (OEt)10 (dmae)2 H2 O cycle. The lms deposited were found
strontium decient with a stoichiometry of Sr0.8 Bi2.1 Ta2 Ox though
the Sr/Ta ratio in the precursor was xed to 1:2. The desired ferroelectric layered perovskite, i.e. the so-called Aurivillius phase SBT
was realized after careful annealing treatments but still there was
always some pyrochlore phase present in the lms. Permittivity of
Several bismuth compounds with oxygen based ligands are

known and their thermal properties have been studied. Volatile
compounds can be found in all the common compound groups i.e.
alkoxides, -diketonates and carboxylates.
5.4.1. -Diketonates of bismuth
Several -diketonates of bismuth are known [218]. They are
volatile, sufciently thermally stable compounds and they react
with H2 O. In ALD only Bi(thd)3 has been used [206208]. Bi(thd)3
forms loose dimers in the solid state (Fig. 23) and is most likely
monomeric in the gas phase. Bi(thd)3 melts at 130134 C and TGA
shows 97% weight loss in a single step at 160300 C indicating
evaporation (Fig. 24). Bi2 O3 was deposited from Bi(thd)3 and H2 O
at temperatures between 200 and 350 C. The ALD temperature
window was found at 270300 C but the growth rate was low,
and believed to be due to the large steric hindrance of

0.1 A/cycle,
the ligands.
5.4.2. Bismuth alkoxides
Bismuth alkoxides form an interesting group of bismuth precursors. The simplest alkoxides Bi(OMe)3 , Bi(OEt)3 , Bi(Oi Pr)3
have very poor volatility but even the smallest tertiary alkoxide
Bi(Ot Bu)3 may be sublimed cleanly already at 80100 C/0.05 mbar
[206,219,220]. It seems that in general bismuth alkoxides with tertiary alkoxy ligands have good volatility while those with primary
or secondary alkoxy ligands do not. There are several examples
which show that the structures of bismuth alkoxides with primary or secondary alkoxy ligands are oligomeric while the ones
3314
Fig. 23. Molecular structure of Bi(thd)3 . Loose dimers are formed in solid state.
Fig. 24. TG curves of different bismuth alkoxides, -diketonate Bi(thd)3 and carboxylates Bi(O2 Ct Bu)3 and Bi(O2 CMe)3 [206]. Reproduced by permission of The Royal
Society of Chemistry.
Fig. 25. Molcular structures of Bi(Ot Bu)3 (above) and Bi(OCi Pr3 )3 (hydrogen atoms
omitted for clarity) [206].
with tertiary alkoxy ligands are monomeric [219,221]. Structures

of Bi(Ot Bu)3 and Bi(Oi Pr3 )3 are shown in Fig. 25. Among the bismuth alkoxides with tertiary alkoxy ligands the volatility seems to
decrease with increasing molecular weight i.e. Bi(Ot Bu)3 is the most
volatile as can be seen from TG curves shown in Fig. 24. Thermal
stability of Bi(Ot Bu)3 however is not good and it was found that
a compound with a bit larger alkoxy ligandOCMe2 i Pr seems to
be more stable. Using even larger tertiary alkoxy ligands seems to
lower the thermal stability. Bismuth alkoxides are in general well
reactive with water though the reaction is not vigorous.
Bi(Ot Bu) has been only tested in a preliminary manner for ALD
of Bi2 O3 using H2 O as the oxygen precursor [206]. Bi2 O3 lm grew
but the precursor, solid at the evaporation temperature of 70 C
(mp. 139 C), seemed to suffer from ageing making repeatable ALD
impossible.
Bi(OCMe2 i Pr)3 is more stable alkoxide and was studied in more
detail [206]. Bi(OCMe2 i Pr)3 has the advantage of being a liquid (mp.
40 C) at the temperature used for its evaporation (85 C). At 150 C
the Bi(OCMe2 i Pr)3 H2 O process was found to produce Bi2 O3 with
a growth rate of 0.38 A/cycle.

ALD type growth was veried fully
at this temperature but it was also conrmed that no signicant
decomposition took place even at temperatures of 200250 C. At
240 C the growth rate was found to be 0.30 A/cycle

[200,222]. Films
grown at 150 C were amorphous as-deposited while at growth
temperatures above 200 C a thus far unidentied crystalline bismuth oxide phase was observed. TOF-ERDA results revealed that
the lms were oxygen rich, probably due to OH residues, and had
otherwise very low impurity levels.
Bi(OCMe2 i Pr)3 has also been studied in ALD of Bi4 Ti3 O12 and
SrBi2 Ta2 O9 at 240 C [200,222]. For Bi4 Ti3 O12 the other precursors
were Ti(OMe)4 and H2 O. Mixing Bi(OCMe2 i Pr)3 H2 O cycles with
Ti(OMe)4 H2 O cycles in a 3:1 ratio gave a Bi/Ti ratio of 1.47 in the
lm. Bi4 Ti3 O12 , an orthorhombic ferroelectric phase was observed
after annealing the lms in air at 700 C. For SrBi2 Ta2 O9 the other
precursors used were Ta(OEt)5 , Sr(i PrC5 H2 )2 or Sr(t Bu3 C5 H2 )2 , and
H2 O.
Bi(mmp)3 i.e. Bi(OCMe2 CH2 OMe)3 is an alkoxide with a tertiary
donor functionalized ligand. This compound has a very good volatility and can be sublimed at 60 C/0.01 mbar. It has been used to
deposit Bix Tiy Oz , Bix Tiy Siz Ow and Bix Tiy Alz Ow (Table 4) [209213].
In three studies Bi(mmp)3 was delivered by liquid injection as a part
of a solvent mixture [209,212,213]. In two studies Bi(mmp)3 H2 O
cycles were combined with Ti(Oi Pr)4 H2 O and Si(OEt)4 H2 O cycles
and the ALD type of growth of BiTiO and BiTiSiO were
reported at 225 C [30,211]. Our own studies on the Bi(mmp)3 H2 O
process revealed that the precursor was decomposing quite rapidly
at 250 C [200]. Bi(mmp)3 was also seen to suffer from ageing in the
heated source.
3315
5.4.3. Bismuth carboxylates

Carboxylates form another group of bismuth compounds that
may be potential ALD precursors. Bismuth formate is completely
unvolatile but acetate and other primary and secondary carboxylates may be sublimed at 200230 C under vacuum with 7080%
yields [223]. Bismuth pivalate, Bi(O2 Ct Bu)3 , which has tetrameric
structure [224], can be sublimed already at 160 C/0.05 mbar and is
stable up to 400 C [206,223]. TG measured under normal pressure
N2 atmosphere also shows almost complete evaporation before
310 C (Fig. 24). Clear advantages of Bi(O2 Ct Bu)3 are high thermal
stability, easy handling and synthesis which does not require any
exclusion of air and moisture. Disadvantage is the lack of reactivity against H2 O and O2 . Bi(O2 Ct Bu)3 has been tested with H2 O for
ALD only preliminarily and it was found that at 250 C there was no
With more
growth and at 300 C the growth rate was 0.06 A/cycle.
powerful oxidizing reagents like O3 or oxygen plasma it may be
possible to deposit bismuth oxide more efciently.
6. Pt group metals
Thin lms of noble metals and their oxides nd use in a wide
variety of applications from catalysis to microelectronics. In integrated circuits there main use is as electrode materials in MOSFETs
and various memory capacitors. The main benets of noble metals
are high work function and resistance against oxidation, and even
if oxidized under oxygen plasma or ozone, the resulting noble
metal oxide layers are still conducting. On the other hand, noble
metal oxides are sometimes preferred as the bottom electrodes
because their structural compatibility with high-k oxides promotes
the crystallization of the latter.
ALD of noble metals was done for the rst time in 2003 when Ru
was deposited from RuCp2 and O2 [225,226]. In this case the success
can be attributed more to the reactant rather than the metal precursor. Before that molecular oxygen had been considered mostly
unreactive in ALD as only in a few special cases it could have been
used for ALD of oxides. Therefore it was surprising to observe O2
serving as an efcient reactant for ALD of noble metals where reducing agent in the rst place was looked for. The high reactivity in the
noble metal ALD process was concluded to be due to the activation
of O2 by the noble metal lm itself via dissociative adsorption into
atomic oxygen. Indeed, in situ reaction mechanism studies revealed
that the RuCp2 O2 ALD Ru process involved combustion reactions
producing CO2 and H2 O as the main byproducts [225,227]. Substantial amount of these were released already during the RuCp2
pulse which was interpreted in terms of atomic oxygen remaining
adsorbed on, and to some extent also in, the Ru lm surface after
the O2 pulse. Later a similar process was demonstrated using a
-diketonate Ru(thd)3 [228], thereby showing that the metal precursor does not need to be an organometallic compound, and the O2
based noble metal ALD chemistry was expanded to Rh [229231],
Pd [231,232], Os [233], Ir [231,234236], Pt [231,232,234,235].
The O2 based ALD metal processes suffer, however, from poor
nucleation. This is thought to be due to a lack of catalytic activation
of O2 before the rst nuclei of the noble metal form, a mechanism which is still to be explained. Also the growth temperatures
needed with O2 are typically 225 C and above. Therefore ozone
as a more efcient oxidizing agent was tested in place of O2 and
was found to enable lm deposition at much lower temperatures
than O2 [237240]. Typically lms deposited below about 200 C
were noble metal oxides and those deposited above 200 C were
metallic even when ozone was used. On the other hand, when H2
was added in the pulsing sequence (metal-precursorO3 H2 ) noble
metals could be deposited at as low temperatures as 165 C [238].
Taking Ir and IrO2 as an example, the rst process examined
was Ir(acac)3 O2 that deposited metallic Ir at 225 C and above
Fig. 26. Above: Molecular structure of (MeC5 H4 )Ir(CHD). Below: TG and SDTA
curves of the compound [236]. Reproduced by permission of The Royal Society of
Chemistry.
[236]. When O2 was replaced with ozone, the Ir(acac)3 O3 process deposited IrO2 at 165200 C and metallic Ir above 200 C.
The low temperature limit of 165 C was in this case set by the
low volatility of Ir(acac)3 that requires a source temperature of
155 C [239]. In other words, with the introduction of O3 H2 [238]
in place of O2 , the metal precursor Ir(acac)3 became the limiting
factor for the low temperature deposition of Ir and created a
need for new iridium precursor. To ll this gap, (MeCp)Ir(CHD)
(MeCp = methylcyclopentadienyl, CHD = cyclohexadiene) was synthesized and tested with both O2 and O3 [236,241]. (MeCp)Ir(CHD)
has a mp of 3639 C and could be delivered from a source temperature of 4550 C. Structure and TG/SDTA curves are shown
in Fig. 26. The (MeCp)Ir(CHD)O2 process was effective only at
225 C and above depositing metallic Ir and was thus similar to the
Ir(acac)3 O2 process. By contrast, the (MeCp)Ir(CHD)O3 process
deposited iridium oxide lms from as low as 100 C up to 180 C
while at 200 C metallic iridium was obtained. Upon adding hydrogen the (MeCp)Ir(CHD)O3 H2 process deposited metallic iridium
lms down to 120 C.
7. Silver
Silver has the lowest resistivity of all known elements,
1.59 cm, and is highly reective and antibacterial material.
Thus silver is used in electrical conductors, highly reective mirrors decorative coatings and antimicrobial coatings [242]. In future
silver could be used in microelectronics interconnects [243]. Compared to many other conducting elements like gold, silver and its
3316
compounds are cheap. Recently, silver nanostructures have

achieved much attention as substrate materials for surface
enhanced Raman scattering [244].
Silver is a group 11 element and resides between Cu and Au. The
most common oxidation state of silver is +I. Paulings electronegativity of silver is 1.9 [198] which cause less polar bonds to form
with ligands if compared for example alkaline earth metals. Ionic
radius of Ag+ is 0.67 A (CN = 2) [199] and the charge to radius ratio
1.49 which means that AgL compounds should be keen to increase
their coordination number through oligomerization or adduct formation. Ag+ is a soft Lewis acid and thus forms stronger bond with
soft Lewis bases like phosphorus and sulfur.
Considering ALD of silver lms the main issues are nding a
silver precursor with sufcient thermal stability and choosing a
suitable co-precursor which converts the silver precursor to a silver lm. Most of the known silver compounds which show volatility
are thermally quite unstable. In the case of the platinum group
metals (Ru, Os, Rh, Ir, Pd, Pt) the usage of O2 , O3 and O3 + H2 as
co-precursors has been found to result in good quality metal lms
from many kinds of noble metal precursors (see above). In the case
of silver this approach apparently has not been found effective and
to date the ALD of silver lms has relied on using a reducing agent
[245247].
While looking for silver precursors for CVD the main objective has been to make a volatile compound which decomposes at
relatively low temperature in the CVD apparatus [248]. This is in
contrast with the demands set for good ALD precursor which should
not show any decomposition. The most common silver precursors
used for CVD are -diketonates and carboxylates with and without adduct forming ligands [248]. Many of the compounds with no
adduct forming ligands are oligomeric or polymeric with low or no
volatility. In several studies on CVD precursors with oxygen based
ligands it has been noticed that the volatility of the complexes is
improved when ligands with appropriate bulkiness and/or uorination are used. It has also been noticed that adduct forming ligands
and their type have crucial effect on the thermal stability and
volatility of the complexes. Electronic properties i.e. the -acceptor
strength of the adduct forming ligands seem to be especially important. Neutral ligands introduced silver complexes include alkene
[249], alkyne [250], methylisocyanide [251], polyamines [252],
polyethers [253,254] and phosphines [248]. Phosphines seem to
form the strongest bonds with silver. The AgO bond stability of
compounds with tertiary phosphines as adduct forming ligands
increases in order: PMe3 < PEt3 < PPh3 , following the -acceptor
strength of the ligands. However, PPh3 is not a good ligand considering the volatility.
ALD of silver is quite scarce and there seem to be only
three different processes reported [245247]. Chalker et al. used
Ag(hfac)(COD) and propanol and the resulting deposits were silver nanoparticles rather than thin lms [247]. In our own work
we employed plasma activated hydrogen to deposit silver from
Ag(Piv)(PEt3 ) and Ag(fod)(PEt3 ) [245,246].
While looking for the most promising precursors for ALD a
set of silver compounds was synthesized and studied. The compounds studied were amidinates, carboxylates and -diketonates:
Ag(i PrNCMeCCi Pr) (Ag1), [Ag(iPrNCMeCNiPr)]2 (PEt3 ) (Ag2),
Ag(O2 CMe) (Ag3), Ag(O2 CCF3 ) (Ag4), Ag(Piv) (Ag5), Ag(Piv)(PEt3 )
(Ag6), Ag(Piv)(PBu3 ) (Ag7), Ag(fod) (Ag8), Ag(hfac)(PMe3 ) (Ag9),
Ag(thd)(PEt3 ) (Ag10), Ag(fod)(PEt3 )(Ag11). Structures of compounds Ag2, Ag6, Ag10 and Ag11 were solved while the structures
of Ag1, Ag5, Ag7 and Ag9 were already known. Amidinate Ag1
forms dimers and trimers in the same cystal structure [258].
Different N atoms in the ligand bond to different Ag atoms. By
adding PEt3 to this compound it was hoped that a monomer would
form. However this did not happen and instead a dimer with one
PEt3 bonded to one Ag atom (Ag2) was seen (Fig. 27) [259], though
Fig. 27. Molecular structure of [Ag(i PrNCMeNi Pr)]2 (PEt3 ) [259].
PEt3 was added in 1:1 proportion. Bi(Piv)3 (Ag5) is polymeric in the

solid state [260] and Ag3 and Ag4 are also believed to be polymeric.
Ag6 forms a trimeric structure shown in Fig. 28 [245]. Ag7 with the
larger phosphine ligand PBu3 forms a dimer [261]. Ag8 is believed
to be polymeric and Ag9 is a monomeric [262]. In Ag10 there is
a weak interaction between Ag atoms of two Ag(thd)(PEt3 ) units
i.e. the compound is dimeric (Fig. 29) [246]. Ag(fod)(PEt3 ) (Ag11)
shows a dimeric structure where the -diketonato ligands which
are chelated to one Ag atom also share one oxygen atom with the
second Ag atom(Fig. 30) [246]. The bridging O atom is on the side
of the C3 F7 -chain of the fod ligand. Ag atoms are not interacting
but there is an interaction between a F atom and Ag (AgF distance
3.111 A).
To study the volatility and thermal stability of the compounds
TGA/SDTA measurements were done (Fig. 31). All the prepared
compounds except Ag(fod)(PEt3 ) (Ag11) left residues which were
Fig. 28. Molecular structure of [Ag(Piv)(PEt3 )]3 [245]. Trimeric structure.
3317
Fig. 29. Molecular structure of Ag(thd)(PEt3 ) [246]. Dimeric structure.
Fig. 30. Molecular structure of Ag(fod)(PEt3 ) [246]. Dimeric structure.
Fig. 31. TG curves of different silver compounds prepared.
larger or approximately equal with the theoretical amount of

silver in the compounds i.e. the compounds decomposed. In contrast, Ag(fod)(PEt3 ) left only a 6% residue while the Ag content
is 22.7%, which proves that most of the compound evaporated.
All the compounds studied except Ag(fod)(PEt3 ), which showed
an endotherm due to evaporation, also showed heat release during the TG/SDTA measurements indicating decomposition. TGA
lets us to suggest that Ag(fod)(PEt3 ) could be thermally stable
at least up to 230 C, but ALD experiments showed that already
at temperatures exceeding 140 C there was thermal decomposition [246]. Considering the temperature where the weight
change takes place in TGA being an indication of thermal stability of the compounds in the condensed state (solid, liquid),
the thermal stability increases in order: [Ag(i PrNCMeNi Pr)]2 (PEt3 )
(Ag2) < Ag(i PrNCMeNi Pr) (Ag1) < Ag(fod) (Ag8) < Ag(hfac)(PMe3 )
(Ag9) Ag(thd)(PEt3 ) (Ag10) < Ag(Piv)(PEt3 ) (Ag6) Ag(Piv)(PBu3 )
(Ag7) < Ag(O2 CMe) (Ag3) Ag(Piv) (Ag5) < Ag(O2 CCF3 ) (Ag4), the
most stable being the unadducted carboxylates. For Ag(fod)(PEt3 )
(Ag11) it is more complicated to evaluate the temperature of
decomposition as the compound evaporates but roughly it could
3318
Fig. 32. SEM pictures of Ag lms deposited using the Ag(Piv)(PEt3 )H radicals process [245]. Reproduced with a permission from Chem. Vap. Dep. 13 (2007) 408. Copyright
2007, John Wiley and Sons.
be similar with that of Ag(Piv)(PEt3 ). Clearly, for all the other silver compounds studied except Ag(fod)(PEt3 ) the temperature of
efcient evaporation at 1 atm is higher than the temperature of
decomposition. Indeed, the volatility of the carboxylates is so low
that sublimation even under vacuum requires temperatures so
high that partial decomposition, indicated by the low sublimation
yields, takes place [255]. It seems obvious that in Ag(fod)(PEt3 )
the volatility and thermal stability are in optimal balance. In
addition to Ag(fod)(PEt3 ) literature reports some other silver compounds which show evaporation before decomposition in TGA.
These include Ag(thd)(Pn Bu3 ) [256], Ag(fod)(PMe3 ), Ag(fod)(PEt3 )
[257] and Ag(hfac)(CNMe) [251].
It does not seem to be possible to make any straight conclusions concerning the thermal stabilities of the compounds on
Fig. 33. (a) AFM image of a 17 nm thick Ag lm deposited at 120 C. (b) SEM image
of the same lm [246]. Reproduced with permission from Chem. Mater. 23 (2011)
2901. Copyright 2011 American Chemical Society.
basis of the crystal structures. Concerning the volatility, polymeric compounds are less volatile than the monomeric or dimeric
compounds.
In an early work Ag(Piv)(PEt3 ) was employed as the silver
precursor and hydrogen radicals produced with remote plasma
(REALD) were used as the reducing agent [245]. In addition
Ag(Piv)(PBu3 ) was tested. It was found that the Ag(Piv)(PEt3 ) precursor starts to decompose at temperatures above 160 C. As the
precursor needed to be heated to 125 C to volatilize it, the ALD
window was limited to 140160 C. It was conrmed that the
usage of molecular H2 did not yield any lms while H radicals
did work. Growth of Ag lms saturated with 3 s Ag(Piv)(PEt3 ) and
7 s hydrogen radical pulses and the growth rate was 1.2 A/cycle.
The lms grew conformally as can be seen from the SEM picture
(Fig. 32). The lms were quite rough but appeared mirror-like.
Impurity levels of the lms were quite high on P (4.0 at.%), O
(10 at.%), H (5.0 at.%) and Si (4.0 at.%). Carbon content was only
1.0 at.%. Resistivity of a 40 nm lm was measured to be 6 cm.
With Ag(Piv)(PBu3 ) lms grew slowly, the resistivities of the lms
were high and their appearance was dark. The worse results were
assumed to be possibly due to slower reaction kinetics or the PBu3
ligands leaving more impurities in the lms. Later on upscaling
the Ag(Piv)(PEt3 )H plasma process was tried on 200 mm Si wafers
but no uniform lms could be deposited over the whole substrate.
To overcome this limitation another precursor Ag(fod)(PEt3 ) was
introduced.
Ag(fod)(PEt3 ) was used with plasma-activated hydrogen
(PEALD) [246]. Again, mirror-like continuous lms were deposited
on silicon and glass substrates with a diameter of 200 mm. Films
thinner than 10 nm were non-continuous as the growth was found
to start with a formation of nanoclusters. The rms-roughness
of a 17 nm thick lm was 4.8 nm and a 50 nm lm had roughness of 5.4 nm (Fig. 33). ALD-window for the process was found
at 120140 C. At 120 C the growth rate was 0.3 A/cycle

and at
140 C, where some signs of decomposition were already seen,
0.38 A/cycle.
The observed growth rate is much lower compared
to that observed for the Ag(Piv)(PEt3 )H radicals process and this
may be due to the larger fod ligand. The deposited lms were
polycrystalline and the resistivity of a 22 nm lm was 68 cm.
Elemental contents of a 17 nm lm deposited at 120 C were Ag
(85 at.%), H (3 at.%), C (3 at.%), O (0.9 at.%), P (0.9 at.%), F (0.5 at.%)
and N (0.7 at.%) as measured with TOF-ERDA. It is notable that the F
contents of the lms were low though there was even a AgF interaction in the solid state structure of the precursor. Also the lms
contained considerably less impurities than the lms deposited
with the Ag(Piv)(PEt3 )H radicals process. These differences may,
however, be partly affected by the reactor congurations too. Anyhow, a clear advantage of Ag(fod)(PEt3 ) over Ag(Piv)(PEt3 ), as tested
in the same reactor, is the ability to deposit uniform lms over large
area substrates.
8. Summary and conclusions

Precursors form the basis upon which the successful exploitation of ALD is built on. Each thin lm material of an interest can
be deposited with all the advantageous features of ALD only after
nding proper precursors. While ALD precursor development has
mostly been focused on metal precursors, advances in non-metal
precursors and co-reactants have brought the ALD technology forward as well. Also, innovative combinations of the precursors have
opened new avenues in the ALD chemistry. In University of Helsinki
the ALD precursor chemistry research has been a balanced combination of basic and applied research, i.e. it has been driven by both
scientic curiosity and needs of industry.
In 1990s -diketonates of alkaline earth metals were intensively
studied as precursors for ALD of sulde lms which act as phosphors in thin lm electroluminescent displays. The studies were
especially aiming to blue light emitting lms. As alternatives to
the -diketonates, alkaline earth cyclopentadienyls were developed and tested. While the cyclopentadienyls brought only minor
improvements to the sulde ALD, they turned highly valuable in
ALD of alkaline earth oxides containing lms, in particular SrTiO3 ,
BaTiO3 and (Ba,Sr)TiO3 , that became of interest at the end of 1990s
as possible high permittivity dielectric materials for DRAM capacitors but could not deposit from the -diketonates with water.
The best cyclopentadienyl compounds for this purpose were those
with sufciently bulky cyclopentadienyl ligands i.e. Ca(i Pr3 C5 H2 )2 ,
Sr(t Bu3 C5 H2 )2 and Ba(t Bu3 c5 H2 )2 . To date cyclopentadienyl compounds are the state-of-the-art alkaline earth metal precursors for
ALD.
ALD of group 4 oxide lms has been studied for over 20 years
at the University of Helsinki. Metal chlorideH2 O processes were
studied in detail and also processes where metal chloride reacts
with metal alkoxide to form binary or mixed metal oxide were
invented. While the simple alkoxides are relatively good precursors for titanium, those of zirconium and hafnium suffer from poor
thermal stability donor functionalized alkoxides of Zr and Hf were
found thermally more stable but suffered from low reactivity. Alkylamides form an important precursor group for group 4 metals,
especially for Zr and Hf. They are very reactive with water but
the thermal stability is limited. Films grown from the alkylamides
show in addition advantageous phase outcome where the high permittivity tetragonal and cubic phases are dominating. Yet another
group of group 4 precursors studied at University of Helsinki has
been cyclopentadienyl compounds. The bis-cyclopentadienyl compounds react rapidly with both water and ozone and the resulting
oxide materials are very pure. The thermal stability of the bis-Cp
compounds is higher than that of alkoxides and alkylamides and
deposition temperatures > 300 C are achievable for ZrO2 and HfO2 .
The Cp-precursors offer good thermal stability but modest growth
rate and tendency to form lms with a mixture of oxide phases.
In heteroleptic complexes (CpR)Zr(NMe2 )3 (R = H, Me or Et) the
benets of the both precursor types can be combined.
Our studies on group 15 and 16 elements were initiated by
a need to deposit germanium antimony telluride lms for phase
change memory applications. The key was to nd suitable tellurium precursor and among the many candidates studied the
alkylsilyl compounds were most promising. They react efciently
with metal chlorides like GeCl2 dioxane and SbCl3 forming the
desired tellurides. Immediately it was also realized that the chemistry may be expanded to other group 16 elements and also group
15 elements.
Studies on bismuth precursors were motivated by the bismuth containing ternary oxides. Different bismuth compounds
were synthesized and the most suitable for the oxide processes
has so far been a bismuth alkoxide Bi(OCMe2 i Pr)3 with tertiary
alkoxy ligands. With this low melting, volatile and thermally stable
3319
compound and H2 O good quality lms could be deposited reliably

in a wide temperature range.
Noble metal lm deposition by ALD has been a subject at University of Helsinki for over 10 years. The rst reported ALD process
for noble metals was a ruthenium process using RuCp2 and O2 as
the precursors. Afterwards also all the other platinum group metals
(Rh, Pd, Os, Ir, Pt) were deposited using the O2 based chemistry
and various kinds of noble metal precursors. By changing O2 to
a more efcient oxidizing agent O3 oxide lms of these metals
could be deposited at temperatures below 200 C while metals were
obtained above 200 C. Adding H2 to the pulsing sequence (metal
precursorO3 H2 ) metal lms could be deposited at low temperatures with the evaporation temperature of the metal precursor
often being the limit. For example, Ir was deposited at 120 C using
a highly volatile Ir precursor (MeCp)Ir(CHD).
ALD of silver has remained quite scarce so far. We have studied a set of silver precursors and found two promising compounds
for ALD of silver lms both containing anionic bidentate oxygen
donor ligand and a neutral phosphine adduct ligand. The silver lm
ALD requires the use of hydrogen radicals or plasma as a reducing
reactant. Uniform silver thin lms have been deposited on 200 mm
wafers.
The role of precursor chemistry has been and will also in the
future be crucial in developing new ALD processes and thereby
bringing the whole ALD technology forward. In the future the challenges for ALD precursor chemistry are still to nd precursors
and precursor combinations for making new materials and materials whose deposition need improvement. New materials could
be for example halides other than uorides, borides and sulfates.
Processes needing serious improvements are many and as an example one could mention electropositive metals (Al, Ti), nitrides and
ternary compounds. There will always be also need to ne tune the
properties of the precursor compounds to fulll the requirements
of lm deposition for different applications. For designing and synthesizing metal precursors there is a set of known ligands available.
A recent trend has been to combine different ligands in heteroleptic
compounds in order to combine the best properties of each ligand
type. This approach has already been successful for example in the
case of group 4 metals and will most likely be explored for also other
metals in the future. On the other hand, while looking for new precursors for ALD, it should be kept in mind that besides completely
new exotic compounds there is a large variety of known compounds
synthesized earlier for some other purpose or just for curiosity that
could be possible precursors for ALD too; they just have not been
found for the purpose yet. Also it should be remembered that the
non-metal precursors, reducing agents and other co-reactants may
be as crucial for the success in ALD as the metal precursors.
References
[1] M. Ritala, M. Leskel, E. Nyknen, P. Soininen, L. Niinist, Thin Solid Films 225
(1993) 288.
[2] M. Ritala, M. Leskel, L. Niinist, T. Prohaska, G. Friedbacher, M. Grasserbauer,
Thin Solid Films 250 (1994) 72.
[3] V. Pore, T. Hatanp, M. Ritala, M. Leskel, J. Am. Chem. Soc. 131 (2009) 3478.
[4] M. Vehkamki, T. Hatanp, T. Hnninen, M. Ritala, M. Leskel, Electrochem.
Solid State Lett. 2 (1999) 504.
[5] V. Miikkulainen, M. Ritala, M. Leskel, R.L. Puurunen, J. Appl. Phys. 113 (2013)
021301.
[6] K. Knapas, M. Ritala, Crit. Rev. Solid State Mater. Sci. 38 (2013) 167.
[7] C.S. Hwang, Mater. Sci. Eng. B56 (1998) 178.
[8] C. Pithan, D. Hennings, R. Waser, Int. J. Appl. Ceram. Technol. 2 (2005) 1.
[9] B. Heinen, R. Waser, J. Mater. Sci. 33 (1998) 4603.
[10] M. Tammenmaa, H. Antson, M. Asplund, L. Hiltunen, M. Leskel, L. Niinist, E.
Ristolainen, J. Cryst. Growth 84 (1987) 151.
[11] W.-M. Li, M. Ritala, M. Leskel, R. Lappalainen, J. Jokinen, E. Soininen, B. Huttl,
E. Nyknen, L. Niinist, J. Appl. Phys. 84 (1998) 1029.
[12] C.A. Paz de Araujo, J.D. Cuchiaro, L.D. Mcmillan, M.C. Scott, J.F. Scott, Nature
374 (1995) 627.
[13] M. Leskel, H. Mls, L. Niinist, Supercond. Sci. Technol. 6 (1993) 627.
3320
[14] M. Friz, F. Waibel, Optical Coatings, in: N. Kaiser, H.K. Pulker (Eds.), Springer
Series in Optical Sciences, vol. 88, Springer, Berlin, 2003, pp. 105130.
[15] W.A. Wojtczak, P.F. Fleig, M.J. Hampden-Smith, Adv. Organomet. Chem. 40
(1996) 215.
[16] J.M. Hartmann, J. Cibert, F. Kany, H. Mariette, M. Charleux, P. Alleysson, R.
Langer, G. Feuillet, J. Appl. Phys. 80 (1996) 6257.
[17] J.M. Hartmann, M. Charleux, H. Mariette, J.L. Rouviere, Appl. Surf. Sci. 112
(1997) 142.
[18] T. Pilvi, M. Ritala, M. Leskel, M. Bischoff, U. Kaiser, N. Kaiser, Appl. Opt. 47
(2008) C271.
[19] T. Pilvi, T. Hatanp, E. Puukilainen, K. Arstila, M. Bischoff, U. Kaiser, N. Kaiser,
M. Leskel, M. Ritala, J. Mater. Chem. 17 (2007) 5077.
[20] T. Pilvi, E. Puukilainen, U. Kreissig, M. Leskel, M. Ritala, Chem. Mater. 20
(2008) 5023.
[21] T. Hatanp, J. Ihanus, J. Kansikas, I. Mutikainen, M. Ritala, M. Leskel, Chem.
Mater. 11 (1999) 1846.
[22] M. Putkonen, L.-S. Johansson, E. Rauhala, L. Niinist, J. Mater. Chem. 9 (1999)
2449.
[23] R. Huang, A.H. Kitai, J. Electron. Mater. 22 (1993) 215.
[24] R. Huang, A.H. Kitai, Appl. Phys. Lett. 61 (1992) 1450.
[25] R. Huang, A.H. Kitai, J. Mater. Sci. Lett. 12 (1993) 1444.
[26] M. Putkonen, T. Sajavaara, L. Niinist, J. Mater. Chem. 10 (2000) 1857.
[27] T. Trndahl, C. Platzer-Bjrkman, J. Kessler, M. Edoff, Prog. Photovolt. 15
(2007) 225.
[28] R. Mantovan, M. Georcieva, M. Perego, H.L. Lu, A. Zenkevich, G. Scarel, M.
Fanciulli, Acta Phys. Pol. A 112 (2007) 1271.
[29] H.-L. Lu, S.-J. Ding, D.W. Zhang, Electrochem. Solid-State Lett. 13 (2010) G25.
[30] M. Putkonen, M. Nieminen, L. Niinist, Thin Solid Films 466 (2004) 103.
[31] B.B. Burton, D.N. Goldstein, S.M. George, J. Phys. Chem. C 113 (2009)
1939.
[32] J. Aarik, A. Aidla, A. Jaek, M. Leskel, L. Niinist, Appl. Surf. Sci. 75 (1994) 33.
[33] O. Nilsen, H. Fjellvg, A. Kjekshus, Thin Solid Films 450 (2004) 240.
[34] M. Putkonen, T. Sajavaara, P. Rahkila, L. Xu, S. Cheng, L. Niinist, H.J. Whitlow,
Thin Solid Films 517 (2009) 5819.
[35] M. Ylilammi, T. Ranta-Aho, J. Electrochem. Soc. 141 (1994) 1278.
[36] T. Pilvi, K. Arstila, M. Leskel, M. Ritala, Chem. Mater. 19 (2007) 3387.
[37] S. Dey, S.J. Jun, Appl. Surf. Sci. 143 (1999) 191.
[38] P. Soininen, L. Niinist, E. Nyknen, M. Leskel, Appl. Surf. Sci. 75 (1994) 99.
[39] M. Leskel, M. Mkel, L. Niinist, E. Nyknen, M. Tammenmaa, Chemtronics
3 (1988) 113.
[40] J. Rautanen, L. Niinist, E. Nyknen, P. Soininen, M. Leskel, Appl. Surf. Sci.
82/83 (1994) 553.
[41] T. Hnninen, I. Mutikainen, V. Saanila, M. Ritala, M. Leskel, J.C. Hanson, Chem.
Mater. 9 (1997) 1234.
[42] K. Kukli, M. Ritala, T. Sajavaara, T. Hnninen, M. Leskel, Thin Solid Films 500
(2006) 322.
[43] M.J. Saly, F. Munnik, C.H. Winter, J. Mater. Chem. 20 (2010) 9995.
[44] J. Aarik, A. Aidla, A. Jaek, M. Leskel, L. Niinist, J. Mater. Chem. 4 (1994),
1239.=34=mm324.
[45] S. Moehnke, M. Bowen, S.S. Sun, R. Tuenge, 4th Intl. Conf. on the Science
and Technol. of Display Phosphors, Bend, Oregon, 1998, p. 203, Extended
Abstracts.
[46] M. Leskel, L. Niinist, E. Nyknen, P. Soininen, M. Tiitta, Mater. Res. Soc. Symp.
Proc. 222 (1991) 315.
[47] P. Soininen, E. Nyknen, L. Niinist, M. Leskel, Chem. Vap. Deposition 2 (1996)
69.
[48] A. Kosola, M. Putkonen, L.-S. Johansson, L. Niinist, Appl. Surf. Sci. 211 (2003)
102.
[49] T. Hara, T. Ishiguro, Sens. Actuators BB136 (2009) 489.
[50] O.S. Kwon, S.W. Lee, J.H. Han, C.S. Hwang, J. Electrochem. Soc. 154 (2007)
G127.
[51] S.W. Lee, J.H. Han, O.S. Kwon, C.S. Hwang, J. Electrochem. Soc. 155 (2008)
G253.
[52] J.H. Lee, Y.J. Cho, Y.S. Min, D. Kim, S.W. Rhee, J. Vac. Sci. Techol. A 20 (2002)
1828.
[53] J.-H. Ahn, J.-Y. Kim, J.-H. Kim, J.-S. Roh, S.-W. Kang, ECS Trans. 16 (2008) 335.
[54] J.-H. Ahn, J.-Y. Kim, J.-H. Kim, J.-S. Roh, S.-W. Kang, Electrochem. Solid State
Lett. 12 (2009) G5.
[55] S.W. Lee, O.S. Kwon, J.H. Han, C.S. Hwang, Appl. Phys. Lett. 92 (2008) 222903.
[56] S.W. Lee, J.H. Han, C.S. Hwang, ECS Trans. 19 (2009) 685.
[57] S.W. Lee, O.S. Kwon, J.H. Han, C.S. Hwang, ECS Trans. 6 (2007) 87.
[58] S.W. Lee, J.H. Han, C.S. Hwang, Electrochem. Solid State Lett. 12 (2009) G69.
[59] M. Lie, O. Nilsen, H. Fjellvag, A. Kjekshus, Dalton Trans. (2009) 481.
[60] S.W. Lee, O.S. Kwon, C.S. Hwang, Microelectron. Eng. 80 (2005) 158.
[61] O.S. Kwon, S.K. Kim, M. Cho, C.S. Hwang, J. Jeong, J. Electrochem. Soc. 152
(2005) C229.
[62] J. Ihanus, M. Ritala, M. Leskel, Electrochem. Soc. Proc. 9725 (1997)
1423.
[63] D.S. Kil, J.M. Lee, J.S. Roh, Chem. Vap. Deposition 8 (2002) 195.
[64] K.-C. Kim, K. Cho, K. Lee, Y. Kim, J.H. Choi, J.-S. Lim, J.Y. Kim, W.-D. Kim, O.S.
Kwon, Y.S. Tak, J.-H. Chung, Y.-S. Kim, S.-T. Kim, W. Han, ECS Trans. 11 (2007)
131.
[65] J. Ihanus, T. Hnninen, T. Hatanp, T. Aaltonen, I. Mutikainen, T. Sajavaara, J.
Keinonen, M. Ritala, M. Leskel, Chem. Mater. 14 (2002) 1937.
[66] T.P. Holme, F.B. Prinz, J. Phys. Chem. A 111 (2007) 8147.
[67] T.P. Holme, C. Lee, F.B. Prinz, Solid State Ionics 179 (2008) 1540.
[68] C.B. Zhang, L. Wielunski, B.G. Willis, Appl. Surf. Sci. 257 (2011) 4826.
[69] H. Wang, X. Jiang, B.G. Willis, J. Vac. Sci. Techol. A 30 (2012), 01A133/1.
[70] M. Vehkamki, Ph.D. Thesis, University of Helsinki, 2007, http://urn.
/URN:ISBN:978-952-10-4452-6
[71] R. Katamreddy, Z. Wang, V. Omarjee, P.V. Rao, C. Dussarrat, N. Blasco, ECS
Trans. 25 (2009) 217.
[72] K. Black, M. Werner, R. Rowlands-Jones, P.R. Chalker, M.J. Rosseinsky, Chem.
Mater. 23 (2011) 2518.
[73] J.H. Han, W. Lee, W. Jeon, S.W. Lee, C.S. Hwang, C. Ko, J. Gatineau, Chem. Mater.
24 (2012) 4686.
[74] S. Riedel, J. Neidhardt, S. Jansen, L. Wilde, J. Sundqvist, E. Erben, S. Teichert, A.
Michaelis, J. Appl. Phys. 109 (2011), 094101/1.
[75] M. Ritala, K. Kukli, M. Vehkamki, T. Hnninen, T. Hatanp, P.I. Risnen, M.
Leskel, Proc. Electrochem. Soc 2000-13 (2000) 597.
[76] M. Vehkamki, T. Hnninen, M. Ritala, M. Leskel, T. Sajavaara, E. Rauhala, J.
Keinonen, Chem. Vap. Deposition 7 (2001) 75.
[77] A. Rahtu, T. Hnninen, M. Ritala, J. Phys. IV: Proc. 11 (2001), Pr3/923.
[78] M.D. McDaniel, A. Posadas, T.Q. Ngo, A. Dhamdhere, D.J. Smith, A.A. Demkov,
J.G. Ekerdt, J. Vac. Sci. Techol. A 31 (2013) 01A136.
[79] S.W. Lee, J.H. Han, S. Han, W. Lee, J.H. Jang, M. Seo, S.K. Kim, C. Dussarat, J.
Gatineau, Y.-S. Min, C.S. Hwang, Chem. Mater. 23 (2011) 2227.
[80] W. Lee, J.H. Han, S.W. Lee, S. Han, W.J. Jeon, C.S. Hwang, J. Mater. Chem. 22
(2012) 15037.
[81] H. Garcia, H. Castan, A. Gomez, S. Duenas, L. Bailon, K. Kukli, M. Kariniemi,
M. Kemell, J. Niinist, M. Ritala, M. Leskel, J. Vac. Sci. Techol. B 29 (2011),
01AC04/1.
[82] J. Ihanus, T. Hnninen, T. Hatanp, M. Ritala, M. Leskel, J. Electrochem. Soc.
151 (2004) H221.
[83] E. Langereis, R. Roijmans, F. Roozeboom, M.C.M. van de Sanden, W.M.M.
Kessels, J. Electrochem. Soc. 158 (2011) G34.
[84] V. Longo, N. Leick, F. Roozeboom, W.M.M. Kessels, ECS Trans. 41 (2011)
63.
[85] V. Longo, N. Leick, F. Roozeboom, W.M.M. Kessels, ECS J. Solid State Sci. Technol. 2 (2013) N15.
[86] M. Popovici, S. Van Elshocht, N. Menou, J. Swerts, D. Pierreux, A. Delabie, B.
Brijs, T. Conard, K. Opsomer, J.W. Maes, D.J. Wouters, J.A. Kittl, J. Electrochem.
Soc. 157 (2010) G1.
[87] M. Popovici, K. Tomida, J. Swerts, P. Favia, A. Delabie, H. Bender, C. Adelmann,
H. Tielens, B. Brijs, B. Kaczer, M.A. Pawlak, M.-S. Kim, L. Altimime, S. Van
Elshocht, J.A. Kittl, Phys. Status Solidi A 208 (2011) 1920.
[88] M.A. Pawlak, B. Kaczer, M.-S. Kim, M. Popovici, J. Swerts, W.-C. Wang, K.
Opsomer, P. Favia, K. Tomida, A. Belmonte, B. Govoreanu, C. Vrancken, C.
Demeurisse, H. Bender, V.V. Afanasev, I. Debusschere, L. Altimime, J.A. Kittl,
Appl. Phys. Lett. 98 (2011) 182902.
[89] M.A. Pawlak, B. Kaczer, M.-S. Kim, M. Popovici, K. Tomida, J. Swerts, K.
Opsomer, W. Polspoel, P. Favia, C. Vrancken, C. Demeurisse, W.-C. Wang, V.V.
Afanasev, W. Vandervorst, H. Bender, I. Debusschere, L. Altimime, J.A. Kittl,
Appl. Phys. Lett. 97 (2010), 162906/1.
[90] C. Wenger, M. Lukosius, T. Blomberg, A. Abrutis, P. Baumann, G. Ruhl, ECS
Trans. 41 (2011) 53.
[91] D. Pierreux, V. Machkaoutsan, E. Tois, J. Swerts, T. Schram, C. Adelmann, S.
Van Elshocht, M. Popovic, T. Conard, J. Tseng, L-. Ragnarsson, J.W. Maes, ECS
Trans. 25 (2009) 263.
[92] M.A. Pawlak, J. Swerts, M. Popovici, B. Kaczer, M.-S. Kim, W.-C. Wang, K.
Tomida, B. Govoreanu, V.V. Afanasev, M. Schaekers, W. Vandervorst, J.A. Kittl,
Appl. Phys. Lett. 101 (2012) 042901.
[93] N. Menou, M. Popovici, S. Clima, K. Opsomer, W. Polspoel, B. Kaczer, G. Rampelberg, K. Tomida, M.A. Pawlak, C. Detavernier, D. Pierreux, J. Swerts, J.W.
Maes, D. Manger, M. Badylevich, V.V. Afanasev, T. Conard, T.P. Favia, H.
Bender, B. Brijs, W. Vandervorst, S. Van Elshocht, G. Pourtois, D.J. Wouters,
S. Biesemans, J.A. Kittl, J. Appl. Phys. 106 (2009) 094101.
[94] N. Menou, M. Popovici, K. Opsomer, B. Kaczer, M.A. Pawlak, C. Adelmann, A.
Franquet, P. Favia, H. Bender, C. Detavernier, S. Van Elshocht, D.J. Wouters, S.
Biesemans, J.A. Kittl, Jpn. J. Appl. Phys. 49 (2010) 04DD01.
[95] M. Popovici, S. Van Elshocht, N. Menou, P. Favia, H. Bender, E. Rosseel, J. Swerts,
C. Adelmann, C. Vrancken, A. Moussa, H. Tielens, K. Tomida, M. Pawlak, B.
Kaczer, G. Schoofs, W. Vandervorst, D.J. Wouters, J.A. Kittl, J. Vac. Sci. Technol.
B 29 (2011) 01A304.
[96] M.A. Pawlak, B. Kaczer, W. Wang, M. Kim, M. Popovici, J. Swerts, K. Tomida,
K. Opsomer, M. Schaekers, C. Vrancken, B. Govoreanu, A. Belmonte, C.
Demeurisse, I. Debusschere, L. Altimime, V.V. Afanasev, J.A. Kittl, VLSI Technology (VLSIT), 2011 Symposium on Topic(s): Communication, Networking
& Broadcasting; Components, Circuits, Devices & Systems; Computing &
Processing (Hardware/Software); Engineered Materials, Dielectrics & Plasmas
(2011) 168.
[97] M.A. Pawlak, M. Popovici, J. Swerts, K. Tomida, K. Min-Soo, B. Kaczer, K.
Opsomer, M. Schaekers, P. Favia, H. Bender, C. Vrancken, B. Govoreanu,
C. Demeurisse, W. Wan-Chih, V.V. Afanasev, I. Debusschere, L. Altimime,
J.A. Kittl, Electron Devices Meeting, 2010, IEDM 2010, IEEE International
Topic(s): Components, ircuits, Devices & Systems; Engineered Materials,
Dielectrics & Plasmas; Fields, Waves & Electromagnetics (2010) 11.7.1,
http://dx.doi.org/10.1109/IEDM.2010.5703344
[98] N. Menou, X.P. Wang, B. Kaczer, W. Polspoel, M. Popovici, K. Opsomer, M.A.
Pawlak, W. Knaepen, C. Detavernier, T. Blomberg, D. Pierreux, J. Swerts,
J.W. Maes, P. Favia, H. Bender, B. Brijs, W. Vandervorst, S. Van Elshocht, D.J.
Wouters, S. Biesemans, J.A. Kittl, Electron Devices Meeting, 2008, IEDM 2008,
[99]
[100]
[101]
[102]
[103]
[104]
[105]
[106]
[107]
[108]
[109]
[110]
[111]
[112]
[113]
[114]
[115]
[116]
[117]
[118]
[119]
[120]
[121]
[122]
[123]
[124]
[125]
[126]
[127]
[128]
[129]
[130]
[131]
[132]
[133]
[134]
[135]
[136]
[137]
[138]
[139]
[140]
[141]
[142]
[143]
[144]
[145]
[146]
[147]
IEEE International Topic(s): Components, Circuits, Devices & Systems; Engineered Materials, Dielectrics & Plasmas; Photonics & Electro-Optics (2008)
pp. 14, http://dx.doi.org/10.1109/IEDM.2008.4796852
J.A. Kittl, K. Opsomer, M. Popovici, N. Menou, B. Kaczer, X.P. Wang, C. Adelmann, M.A. Pawlak, K. Tomida, A. Rothschild, B. Govoreanu, R. Degraeve, M.
Schaekers, M. Zahid, A. Delabie, J. Meersschaut, W. Polspoel, S. Clima, G. Pourtois, W. Knaepen, C. Detavernier, V.V. Afanasev, T. Blomberg, D. Pierreux, J.
Swerts, P. Fischer, J.W. Maes, D. Manger, W. Vandervorst, T. Conard, A. Franquet, P. Favia, H. Bender, B. Brijs, S. Van Elshocht, M. Jurczak, J. Van Houdt, D.J.
Wouters, ECS Trans. 19 (2009) 29.
T. Blomberg, J. Anttila, S. Haukka, M. Tuominen, M. Lukosius, Ch. Wenger, T.
Saukkonen, Thin Solid Films 520 (2012) 6535.
C. Baristiran-Kaynak, M. Lukosius, B. Tillack, Ch Wenger, T. Blomberg, G. Ruhl,
Microelectron. Eng. 88 (2011) 1521.
C.J. Yim, S.U. Kim, Y.S. Kang, M.-H. Cho, D.-H. Ko, Electrochem. Solid-State Lett.
14 (2011) G45.
M.J. Saly, F. Munnik, C.H. Winter, Chem. Vap. Deposition 17 (2011) 128.
J.A.T. Norman, M. Perez, M.S. Kim, X. Lei, S. Ivanov, A. Derecskei-Kovacs, L.
Matz, I. Buchanan, A.L. Rheingold, Inorg. Chem. 50 (2011) 12396.
W.-C. Shin, S.-O. Ryu, I.-K. You, B.-G. Yu, W.-J. Lee, K.-J. Choi, S.-G. Yoon, J.
Electrochem. Soc. 151 (2004) C292.
R.J. Potter, P.A. Marshall, J.L. Roberts, A.C. Jones, P.R. Chalker, M. Vehkamki,
M. Ritala, M. Leskel, P.A. Williams, H.O. Davies, N.L. Tobin, L.M. Smith, Mater.
Res. Soc. Symp. Proc. 784 (2004) 97.
M. Vehkamki, M. Ritala, M. Leskel, A.C. Jones, H.O. Davies, T. Sajavaara, E.
Rauhala, J. Electrochem. Soc. 151 (2004) F69.
W.-J. Lee, I.-K. You, S.-O. Ryu, B.-G. Yu, K.I. Cho, S.-G. Yoon, C.-S. Lee, Jpn. J.
Appl. Phys. Part 1 40 (2001) 6941.
M. Vehkamki, T. Hatanp, M. Ritala, M. Leskel, J. Mater. Chem. 14 (2004)
3191.
V. Saanila, J. Ihanus, M. Ritala, M. Leskel, Chem. Vap. Deposition 4 (1998) 227.
J.H. Shim, J.S. Park, J. An, T.M. Guer, S. Kang, F.B. Prinz, Chem. Mater. 21 (2009)
3290.
M. Vehkamki, T. Hatanp, M. Ritala, M. Leskel, S. Vyrynen, E. Rauhala,
Chem. Vap. Deposition 13 (2007) 239.
T. Hatanp, M. Vehkamki, I. Mutikainen, J. Kansikas, M. Ritala, M. Leskel,
Dalton Trans. (2004) 1181.
R. Matero, A. Rahtu, S. Haukka, M. Tuominen, M. Vehkamki, T. Hatanp, M.
Ritala, M. Leskel, ECS Trans. 1 (2006) 137.
M.J. Saly, F. Munnik, R.J. Baird, C.H. Winter, Chem. Mater. 21 (2009) 3742.
J.E. Schwarberg, R.E. Sievers, R.W. Moshier, Anal. Chem. 42 (1970) 1828.
M. Leskel, H. Mls, L. Niinist, Supercond. Sci. Technol. 6 (1993) 627.
A. Gleizes, S. Sans-Lenain, D. Medus, C.R. Acad. Sci. Paris Ser. II 313 (1991) 761.
G. Rossetto, A. Polo, F. Benetollo, M. Porchia, P. Zanella, Polyhedron 11 (1992)
979.
S.R. Drake, S.S.S. Miller, M.B. Hursthouse, K.M.A. Malik, Polyhedron 12 (1993)
1621.
T. Hnninen, I. Mutikainen, M. Leskel, unpublished results; J. Brooks, H.O.
Davies, T.J. Leedham, A.C. Jones, A. Steiner, Chem. Vap. Deposition 6 (2000)
66.
M. Leskel, L. Niinist, E. Nyknen, P. Soininen, M. Tiitta, Thermochim. Acta
175 (1991) 91.
T. Hatanp, J. Kansikas, I. Mutikainen, M. Leskel, Inorg. Chem. 40 (2001) 788.
T.P. Hanusa, Polyhedron 9 (1990) 1345.
T.P. Hanusa, Chem. Rev. 93 (1993) 1023.
M.J. Benac, A.H. Cowley, R.A. Jones, A.F. Tasch Jr., Chem. Mater. 1 (1989) 289.
Y. Tasaki, R. Sakamoto, Y. Ogawa, S. Yoshizawa, J. Ishiai, S. Akase, Jpn. J. Appl.
Phys. Part 1 33 (1994) 5400.
Y. Yamada, T. Morimoto, Jpn. Patent, JP02225317, 1990.
C. Blaauw, R.A. Bruce, C.J. Miner, A.J. Howard, B. Emmerstorfer, A.J. Springthorpe, J. El. Mater. 18 (1989) 567.
T. Hatanp, M. Ritala, M. Leskel, J. Organomet. Chem. 692 (2007) 5256.
T. Hatanp, T. Hnninen, J. Ihanus, J. Kansikas, I. Mutikainen, M. Vehkamki,
M. Ritala, M. Leskel, Electrochem. Soc. Proc. 2003-08 (2003) 516.
T. Hnninen, T. Hatanp, M. Ritala, M. Leskel, unpublished results.
R.A. Williams, T.P. Hanusa, J.C. Huffman, J. Chem. Soc., Chem. Commun. (1988)
1045.
D.J. Burkey, R.A. Williams, T.P. Hanusa, Organometallics 12 (1993) 1331.
R.A. Williams, K.F. Tesh, T.P. Hanusa, J. Am. Chem. Soc. 113 (1991) 4843.
M.J. Harvey, K.T. Quisenberry, T.P. Hanusa, V.G. Young Jr., Eur. J. Inorg. Chem.
(2003) 3383.
H. Sitzmann, T. Dezember, M. Ruck, Angew. Chem. Int. Ed. 37 (1998) 3113.
M.T. Bohr, R.S. Chau, T. Ghani, K. Mistry, IEEE Spectrum 44 (10) (2007) 30.
S.K. Kim, S.W. Lee, J.H. Han, B. Lee, S. Han, C.S. Hwang, Adv. Funct. Mater. XX
(2010) 1.
M. Ritala, M. Leskel, Appl. Surf. Sci. 75 (1994) 333.
M. Verghese, S. Battle, J. Roeder, Semicond. Int., April 1 (2008).
K. Kukli, M. Ritala, M. Schuisky, M. Leskel, T. Sajavaara, J. Keinonen, T. Uustare,
A. Hrsta, Chem. Vap. Deposition 6 (2000) 303.
K. Kukli, K. Forsgren, J. Aarik, T. Uustare, A. Aidla, A. Niskanen, M. Ritala, M.
Leskel, A. Hrsta, J. Cryst. Growth 231 (2001) 262.
K. Kukli, M. Ritala, J. Aarik, T. Uustare, M. Leskel, J. Appl. Phys. 92 (2002) 1833.
K. Kukli, J. Ihanus, M. Ritala, M. Leskel, Appl. Phys. Lett. 68 (1996) 3737.
M. Ritala, K. Kukli, A. Rahtu, P.I. Risnen, M. Leskel, T. Sajavaara, J. Keinonen,
Science 288 (2000) 319.
A. Rahtu, M. Ritala, M. Leskel, Chem. Mater. 13 (2001) 1528.
3321
[148] V. Pore, A. Rahtu, M. Leskel, M. Ritala, Chem. Vap. Deposition 10 (2004) 143.
[149] A. Rahtu, PhD Thesis, University of Helsinki, 2002, http://urn./URN:
ISBN:952-10-0646-3
[150] M. Ritala, M. Leskel, L. Niinist, P. Haussalo, Chem. Mater. 5 (1993) 1174.
[151] M. Ritala, M. Leskel, E. Rauhala, Chem. Mater. 6 (1994) 556.
[152] K. Kukli, M. Ritala, M. Leskel, Chem. Vap. Deposition 6 (2000) 297.
[153] A.C. Jones, H.C. Aspinall, R.P. Chalker, R.J. Potter, K. Kukli, A. Rahtu, M. Ritala,
M. Leskel, J. Mater. Chem. 14 (2004) 3101.
[154] R. Matero, M. Ritala, M. Leskel, A.C. Jones, P.A. Wlliams, F.J. Bickley, A. Steiner,
T.J. Leedham, H.O. Davies, J. Non-Cryst. Solids 303 (2002) 24.
[155] R. Matero, M. Ritala, M. Leskel, T. Sajavaara, A.C. Jones, J.L. Roberts, Chem.
Mater. 16 (2004) 5630.
[156] P.A. Williams, J.L. Roberts, A.C. Jones, P.R. Chalker, N.L. Tobin, J.F. Bickley, H.O.
Davies, L.M. Smith, T.J. Leedham, Chem. Vap. Deposition 8 (2002) 163.
[157] K. Kukli, M. Ritala, M. Leskel, T. Sajavaara, J. Keinonen, A.C. Jones, J.L. Roberts,
[158] K. Kukli, M. Ritala, M. Leskel, T. Sajavaara, J. Keinonen, A.C. Jones, J.L. Roberts,
Chem. Mater. 15 (2003) 1722.
[159] K. Kukli, M. Ritala, M. Leskel, T. Sajavaara, J. Keinonen, A.C. Jones, N.L. Tobin,
[160] J.W. Elam, M. Schuisky, J.D. Ferguson, S.M. George, Thin Solid Films 436 (2003)
145.
[161] Q. Xie, Y.-L. Jiang, C. Detavernier, D. Deduytsche, R.L. van Meirhaeghe, G.-P.
Ru, B.-Z. Li, X.-P. Qu, J. Appl. Phys. 102 (2007) 083521.
[162] D.M. Hausmann, E. Kim, J. Becker, R.G. Gordon, Chem. Mater. 14 (2002) 4350.
[163] K. Kukli, M. Ritala, T. Sajavaara, J. Keinonen, M. Leskel, Chem. Vap. Deposition
8 (2002) 199.
[164] E. Granneman, P. Fischer, D. Pierreux, H. Terhorst, P. Zagwijn, Surf.
Coat.Technol. 201 (2007) 8899.
[165] S.K. Kim, C.S. Hwang, Electrochem. Solid State Lett. 11 (2008) G9.
[166] J.-H. Kim, V.A. Ignatova, J. Heitman, L. Oberbeck, J. Phys. D: Appl. Phys. 41
(2008) 172005.
[167] K. Kukli, M. Ritala, J. Lu, A. Hrsta, M. Leskel, J. Electrochem. Soc. 151 (2004)
F189.
[168] K. Kukli, T. Pilvi, M. Ritala, T. Sajavaara, J. Lu, M. Leskel, Thin Solid Films 491
(2005) 328.
[169] J. Niinist, K. Kukli, T. Sajavaara, M. Ritala, M. Leskel, L. Oberbeck, J. Sundqvist,
U. Schrder, Electrochem. Solid State Lett. 12 (2009) G1.
[170] J. Niinist, K. Kukli, M.J. Heikkil, M. Ritala, M. Leskel, Adv. Eng. Mater. 11
(2009) 222.
[171] M. Putkonen, J. Niinist, K. Kukli, T. Sajavaara, M. Karppinen, H. Yamauchi, L.
Niinist, Chem. Vap. Deposition 9 (2003) 207.
[172] J. Niinist, M. Putkonen, L. Niinist, K. Kukli, M. Ritala, M. Leskel, J. Appl. Phys.
95 (2004) 84.
[173] J. Niinist, M. Putkonen, L. Niinist, S.L. Stoll, K. Kukli, T. Sajavaara, M. Ritala,
M. Leskel, J. Mater. Chem. 15 (2005) 2271.
[174] J. Niinist, A. Rahtu, M. Putkonen, M. Ritala, M. Leskel, L. Niinist, Langmuir
21 (2005) 7321.
[175] J. Niinist, K. Kukli, A. Tamm, M. Putkonen, C.L. Dezelah, L. Niinist, J. Lu, F.
Song, P. Williams, P.N. Heyns, M. Ritala, M. Leskel, J. Mater. Chem. 18 (2008)
3385.
[176] K. Knapas, M. Ritala, Chem. Mater. 20 (2008) 5698.
[177] J. Niinist, K. Kukli, M. Kariniemi, M. Ritala, M. Leskel, N. Blasco, A. Pinchart, C.
Lachaud, N. Laaroussi, Z. Wang, C. Dussarrat, J. Mater. Chem. 18 (2008) 5243.
[178] J. Niinist, M. Mntymki, K. Kukli, L. Costelle, E. Puukilainen, M. Ritala, M.
Leskel, J. Cryst. Growth 312 (2010) 245.
[179] N. Blasco, A. Pinchart, C. Lachaud, A. Zauner, J. Gatineau, S. Ogava, B. Feist, R.
Katamreddy, N. Stafford, Z. Wang, V. Omarjee, C. Dussarrat, AVS 8th International Conference on Atomic Layer Deposition, Bruges, Belgium, AVS, New
York (2008) (Presentation available on DVD).
[180] R. Katamreddy, V. Omarjee, B. Feist, C. Dussarrat, ECS Trans. 16 (2008) 113.
[181] M. Rose, J. Niinist, P. Michalowski, L. Gerlich, L. Wilde, I. Endler, J.W. Bartha,
J. Phys. Chem. C 113 (2009) 21825.
[182] T. Blanquart, J. Niinist, V. Longo, V.R. Pallem, C. Dussarrat, M. Ritala, M.
Leskel, Chem. Mater. 24 (2012) 3420.
[183] J. Niinist, T. Hatanp, M. Mntymki, M. Kariniemi, L. Costelle, K. Mizohata,
K. Kukli, M. Ritala, M. Leskel, Chem. Mater. 24 (2012) 2002.
[184] S.R. Ovshinsky, Phys. Rev. Lett. 21 (1968) 1450.
[185] A.L. Lacaita, Solid-State Electron. 50 (2006) 24.
[186] S. Raoux, G.W. Burr, M.J. Breitwisch, C.T. Rettner, Y.-C. Chen, R.M. Shelby, M.
Salinga, D. Krebs, S.-H. Chen, H.-L. Lung, C.H. Lam, IBM J. Res. Dev. 52 (2008)
465.
[187] Pore, K. Knapas, T. Hatanp, T. Sarnet, M. Kemell, M. Ritala, M. Leskel, K.
Mizohata, Chem. Mater. 23 (2011) 247.
[188] T. Hatanp, V. Pore, T. Sarnet, M. Ritala, M. Leskel, ALD2010, 10th
international Conference on Atomic Layer Deposition, Proceedings, Poster
presentation M8, Seoul, Korea, June 2023, 2010.
[189] K. Knapas, T. Hatanp, M. Ritala, M. Leskel, Chem. Mater. 22 (2010) 1386.
[190] T. Sarnet, V. Pore, T. Hatanp, M. Ritala, M. Leskel, A. Schrott, Y. Zhu, S. Raoux,
H.-Y. Cheng, J. Electrochem. Soc. 158 (2011) D694.
[191] M. Ritala, V. Pore, T. Hatanp, M. Heikkil, M. Leskel, K. Mizohata, A. Schrott,
S. Raoux, S.M. Rossnagel, Microel. Eng. 86 (2009) 1946.
[192] S. Maitrejean, S. Lhostis, S. Haukka, C. Jahan, E. Gourvest, R. Matero, T.
Blomberg, A. Toffoli, A. Persico, C. Jayet, M. Veillerot, J.P. Barnes, F. Pierre,
F. Fillot, L. Perniola, V. Sousa, H. Sprey, F. Boulanger, B. de Salvo, T.
Billon, 2011 IEEE International Interconnect Technology Conference and
3322
[193]
[194]
[195]
[196]
[197]
[198]
[199]
[200]
[201]
[202]
[203]
[204]
[205]
[206]
[207]
[208]
[209]
[210]
[211]
[212]
[213]
[214]
[215]
[216]
[217]
[218]
[219]
[220]
[221]
[222]
[223]
[224]
[225]

2011 Materials for Advanced Metallization (IITC/MAM), Dresden (2011),
http://dx.doi.org/10.1109/IITC.2011.5940298
M. Mehring, Coord. Chem. Rev. 251 (2007) 974.
P. Shuk, H.D. Wiemhfer, U. Guth, W. Gopel, M. Greenblatt, Solid State Ionics
89 (1996) 179.
N.M. Sammes, G.A. Tompsett, H. Nfe, F. Aldinger, J. Eur. Ceram. Soc. 1801
(1999).
A. Cabot, A. Marsal, J. Arbiol, J.R. Morante, Sens. Actuator B Chem. 99 (2004)
74.
R. Venkatasubramanian, E. Siivola, T. Colpitts, B. OQuinn, Nature 413 (2001)
597.
A.L. Allred, J. Inorg. Nucl. Chem. 17 (1961) 215.
R.D. Shannon, Acta Crystallogr. Sect. A 32 (1976) 751.
M. Vehkamki, unpublished results.
M. Schuisky, K. Kukli, M. Ritala, A. Hrsta, M. Leskel, Chem. Vap. Deposition
6 (2000) 139.
J. Harjuoja, S. Vyrynen, M. Putkonen, L. Niinist, E. Rauhala, J. Cryst. Growth
285 (2006) 376.
W.C. Shin, S.O. Ryu, I.K. You, S.M. Yoon, S.M. Cho, N.Y. Lee, K.D. Kim, B.G. Yu,
W.J. Lee, K.J. Choi, s.G. Yoon, Electrochem. Solid State Lett. 7 (2004) F31.
J. Harjuoja, T. Hatanp, M. Vehkamki, S. vyrynen, M. Putkonen, L. Niinist,
M. Ritala, M. Leskel, E. Rauhala, Chem. Vap. Deposition 11 (2005) 362.
M. Vehkamki, T. Hatanp, M. Kemell, M. Ritala, M. Leskel, Chem. Mater. 18
(2006) 3883.
T. Hatanp, M. Vehkamki, M. Ritala, M. Leskel, Dalton Trans. 39 (2010)
3219.
Y.D. Shen, Y.W. Li, W.M. Li, J.Z. Zhang, Z.G. Hu, J.H. Chu, J. Phys. Chem. C 116
(2012) 3449.
E. stereng, K.B. Gandrud, O. Nilsen, H. Fjellvag, BALD2009, Book of abstracts,
Uppsala, Sweden (2009).
Y.J. Cho, Y.-S. Min, J.-H. Lee, B.-S. Seo, J.K. Lee, Y.S. Park, Integr. Ferroelectr. 59
(2003) 1483.
G.W. Hwang, W.D. Kim, Y.-S. Min, Y.J. Cho, C.S. Hwang, J. Electrochem. Soc.
153 (2006) F20.
G.W. Hwang, W.D. Kim, C.S. Hwang, Y.-S. Min, Y.J. Cho, J. Electrochem. Soc.
154 (2007) H915.
Y.-S. Min, Y.J. Cho, I.P. Asanov, J.H. Han, W.D. Kim, C.S. Hwang, Chem. Vap.
Deposition 11 (2005) 38.
Y.-S. Min, I.P. Asanov, C.S. Hwang, Electrochem. Solid State Lett. 9 (2006) G231.
Outokumpu HSC Chemistry for Windows Chemical Reaction and Equilibrium Software with Extensive Thermochemical Database, Antti Roine, 2002.
M. Schuisky, Acta. Univ. Uppsa. Abstr. Uppsala Diss. Sci. 594 (2000).
H. Suzuki, Y. Matano, in: H. Suzuki, Y. Matano (Eds.), Organobismuth Chemistry, Elsevier, Amsterdam, 2001.
W. Clegg, N.A. Compton, R.J. Errington, G.A. Fisher, M.E. Green, D.C.R. Hockless,
N.C. Norman, Inorg. Chem. 30 (1991) 4680.
A.P. Pisarevskii, L.I. Martynenko, N.G. Dzyubenko, Russ. J. Inorg. Chem. 37
(1992) 38.
M.-C. Massiani, R. Papiernik, L.G. Hubert-Pfalzgraf, J.-C. Daran, Polyhedron 10
(1991) 437.
W.J. Evans, J.H. Hain Jr., J.W. Ziller, J. Chem. Soc., Chem. Commun. (1989) 1628.
P.A. Williams, A.C. Jones, M.J. Crosbie, P.J. Wright, J.J. Bickley, A. Steiner, H.O.
Davies, T.J. Leedham, G.W. Critchlow, Chem. Vap. Deposition 7 (2001) 205.
M. Vehkamki, T. Hatanp, M. Ritala, M. Leskel, US Patent, US7713584 (B2),
May 11, 2010.
A.P. Pisarevskii, L.I. Martynenko, N.G. Dzyubenko, Russ. J. Inorg. Chem. 35
(1990) 843.
S.I. Troyanov, A.P. Pisarevsky, J. Chem. Soc., Chem. Commun. (1993) 335.
T. Aaltonen, A. Rahtu, M. Ritala, M. Leskel, Electrochem. Solid State Lett. 6
(2003) C130.
[226] T. Aaltonen, P. Alen, M. Ritala, M. Leskel, Chem. Vap. Deposition 9 (2003) 45.
[227] T. Aaltonen, A. Rahtu, M. Ritala, M. Leskel, Electrochem. Soc. Proc 2003-8
(2003) 946.
[228] T. Aaltonen, M. Ritala, Y.L. Tung, Y. Chi, K. Arstila, K. Meinander, M. Leskel, J.
Mater. Res. 19 (2004) 3353.
[229] T. Aaltonen, M. Ritala, M. Leskel, Electrochem. Solid State Lett. 8 (2005) C99.
[230] K.J. Park, J.M. Doub, T. Gougousi, G.N. Parsons, Appl. Phys. Lett. 86 (2005)
051903.
[231] T. Aaltonen, M. Ritala, M. Leskel, in: D. Erb, P., Ramm, K., Masu, A. Osaki (Eds.),
Advanced Metallization Conference 2004 (AMC 2004) 663.
[232] T. Aaltonen, M. Ritala, Yl. Tung, Y. Chi, K. Arstila, K. Meinander, M. Leskel, J.
Mater. Res. 19 (2004) 3353.
[233] J. Hmlinen, T. Sajavaara, E. Puukilainen, M. Ritala, M. Leskel, Chem. Mater.
24 (2012) 55.
[234] E. Frm, M. Kemell, M. Ritala, M. Leskel, J. Phys. Chem. C 112 (2008) 15791.
[235] E. Frm, M. Kemell, E. Santala, M. Ritala, M. Leskel, J. Electrochem. Soc. 157
(2010) K10.
[236] J. Hmlinen, T. Hatanp, E. Puukilainen, L. Costelle, T. Pilvi, M. Ritala, M.
Leskel, J. Mater. Chem. 20 (2010) 7669.
[237] J. Hmlinen, F. Munnik, M. Ritala, M. Leskel, J. Electrochem. Soc. 156 (2009)
D418.
[238] J. Hmlinen, E. Puukilainen, M. Kemell, L. Costelle, M. Ritala, M. Leskel,
Chem. Mater. 21 (2009) 4868.
[239] J. Hmlinen, M. Kemell, F. Munnik, U. Kreissig, M. Ritala, M. Leskel, Chem.
Mater. 20 (2008) 2903.
[240] J. Hmlinen, F. Munnik, M. Ritala, M. Leskel, Chem. Mater. 20 (2008) 6840.
[241] J. Hmlinen, T. Hatanp, E. Puukilainen, T. Sajavaara, M. Ritala, M. Leskel,
J. Mater. Chem. 21 (2011) 16488.
[242] A.P. Piedade, M.T. Vieira, A. Martins, F. Silva, Nanotechnology 18 (2007)
105103.
[243] International Technology Roadmap for Semiconductors, [Online] (2009),
http://www.itrs.net/Links/2009ITRS/Home2009.htm
[244] M. Moskovits, Rev. Modern Phys. 57 (1985) 783.
[245] A. Niskanen, T. Hatanp, K. Arstila, M. Leskel, M. Ritala, Chem. Vap. Deposition 13 (2007) 408.
[246] M. Kariniemi, J. Niinist, T. Hatanp, M. Kemell, T. Sajavaara, M. Ritala, M.
Leskel, Chem. Mater. 23 (2011) 2901.
[247] P.R. Chalker, S. Romani, P.A. Marshall, M.J. Rosseinsky, S. Rushworth, P.A.
Williams, Nanotechnology 21 (2010) 405602.
[248] A. Grodzicki, I. Lakomska, P. Piszczek, I. Szymanska, E. Szlyk, Coord. Chem.
Rev. 249 (2005) 2232.
[249] K.-M. Chi, C.-T. Lin, S.-M. Peng, G.-H. Lee, Organometallics 15 (1996) 2660.
[250] C.-M. Chi, Y.-H. Lu, Chem. Vap. Deposition 7 (2001) 117.
[251] Z. Yuan, N.H. Dryden, X. Li, J.J. Vittal, R.J. Puddephatt, J. Mater. Chem. 5 (1995)
303.
[252] J.A. Darr, M. Poliakoff, W.-S. Li, A.J. Blake, J. Chem. Soc., Dalton Trans. (1997)
2869.
[253] M.E. Fragala, G. Malandrino, O. Puglisi, C. Benelli, Chem. Mater. 12 (2000) 290.
[254] J.A. Darr, M. Poliakoff, A.J. Blake, W.-S. Li, Inorg. Chem. 37 (1998) 5491.
[255] S.E. Paramonov, N.P. Kuzmina, S.I. Troyanov, Polyhedron 22 (2003) 837.
[256] C.N. Silvana, K.K. Banger, P.J. Toscano, J.T. Welch, Polyhedron 21 (2002)
1289.
[257] Z. Yuan, N.H. Dryden, J.J. Vittal, R.J. Puddephatt, Chem. Mater. 7 (1995) 1696.
[258] B.S. Lim, A. Rahtu, J.-S. Park, R.G. Gordon, Inorg. Chem. 42 (2003) 7951.
[259] T. Hatanp, unpublished results.
[260] S.E. Paramonov, E.V. Mychio, S.I. Troyanov, N.P. Kuzmina, Russ. J. Inorg. Chem.
45 (2000) 2003.
[261] S.E. Paramonov, N.P. Kuzmina, S.I. Troyanov, Polyhedron 22 (2003) 837.
[262] Z. Yuan, N.H. Dryden, J.J. Vittal, R.J. Puddephat, Chem. Mater. 7 (1995)
1696.

Precursors As Enablers of ALD Technology: Contributions From University of Helsinki

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Precursors As Enablers of ALD Technology: Contributions From University of Helsinki

Uploaded by

Copyright:

Available Formats

Coordination Chemistry Reviews 257 (2013) 32973322

Contents lists available at ScienceDirect

Coordination Chemistry Reviews

Precursors as enablers of ALD technology: Contributions from

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

review is not to present a balanced view of the eld but to focus

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

-diketonates and cyclopentadienyl compounds which have been

Volatility is needed for efcient

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

molecule size to monomolecular one [120]. The volatility of the

though with a low rate of 0.1 A/cycle,

a constant growth rate of 0.27 A/cycle

[Mg2 (thd)4 ] and H2 O2 but with a low rate (0.100.15 A/cycle)

number of cyclopentadienyl compounds of strontium and barium

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

Sr(i Pr3 C5 H2 )2 (DME)

(Me5 C5 )Ti(OMe)3 , O2 plasma

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

Sr(C5 Me5 )2 (THF)2

Fig. 3. TG curves of cyclopentadienyl compounds of strontium. Flowing 1 atm N2

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

close to zero indicates a good volatility and thermal stability. TGA

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

three processes. As can be seen from the gure, the temperature

rates were 0.42 and 0.28 A/cycle

and H2 O. At 300 C the growth rate was around 1 A/cycle

growth rate of 1.16 A/cycle

mostly amorphous lms with a growth rate of 0.5 A/cycle.

crystalline whereas on amorphous Al2 O3 they were amorphous

growth rate. The high growth

MIM capacitors annealed at 800 C with platinum

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

H2 O was rst scaled up on 200 mm silicon wafers [114]. Later

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

with high oxygen afnity, the process oxidises the substrate or

that gives a rate of 0.65 A/cycle

ligand dmae (dmae = dimethylaminoethoxide) (Fig. 9) allowed to

at 340 C. The homoleptic Hf(mmp)4 behaved quite

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

precursors [171173]. The growth rate is typically 0.50.6 A/cycle

reactions with the surface OH groups (Fig. 12) [174]. This is in

Cp compounds, viz. low growth rate which increased by 0.1 A/cycle.

ozone process with a growth rate close to 1 A/cycle.

(CpR)Hf(NMe2 )3 O3 process with a rate of 0.8 A/cycle

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

Mixed ligand complex (Me5 C5 )Ti(OMe)3 has been found to be

slowly from 0.22 to 0.29 A/cycle

(Me5 C5 )Ti(OMe)3 and Ti(i OPr)2 (thd)2 [5557,71,79,81,8385,87].

in a controlled way up to 350 C with a rates of 0.75 and 0.45 A/cycle.

have been recorded. In ZrO2 lms the high permit0.70.8 A/cycle

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

harmful level. By contrast, the hydrides of selenium and tellurium

Fig. 14. Molecular structure of (MeCp)Zr(C7 H7 ) (above) and TG curves measured

reported. TiO2 lms on the other hand showed anatase structure

A breakthrough in ALD of GST and related materials, including

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

Fig. 17. Temperature dependencies of ALD growth rates of various materials

compounds studied are sensitive to air and moisture. Compounds

In situ reaction mechanism studies with a quartz crystal

dependence of the growth rate. Various alternatives have been

T. Hatanp et al. / Coordination Chemistry Reviews 257 (2013) 32973322

0.7 at.% carbon, and 0.6 at.% chlorine as analyzed by Time-of-Flight