Metallic thin film on sige si substrates 1

1.2.3 Growth mode and interfacial reaction of Ni on Si, Ge and Si0.8Ge0.2 substrates 1.2.3.1 Determining the growth modes by AES or XPS Investigations of the first stages of thin film gr

Chapter 1 Introduction

1.1 Introduction

Silicon has been the dominant substrate material in the semiconductor industry for over 30 years This is because silicon is abundant in nature, non-toxic, strong, and a good conductor of heat In addition, very high purity wafers with very large diameter (300mm) can be grown at low cost compared to other substrates Another important advantage of Si is the superior thermal stabilities of silicon dioxide and silicon nitride, both of which act as necessary insulating films in the semiconductor device structures

The advance of microelectronic technology has pushed the search for new materials and novel device structures Some III-V semiconductors, e.g GaAs, InP, demonstrate superior electronic properties compared to silicon However, they still cannot replace Si due to their high cost and lack of high quality oxides In addition, defect density in III-V crystal increases with wafer diameter, which makes fabrication

of large diameter III-V wafers with high purity difficult Thus, it would be easier and cost-effective to integrate new materials or device structures with existing Si technology rather than to develop a new production line Among several interesting materials, Si1-xGex is a promising candidate due to its high hole mobility and its compatibility with the current Si technology Details on the structure of Si1-xGex and concerns with its applications in devices will be addressed in the next section

Fig 1.1 The arrangement of atoms in a single crystal substrate for both Si and Ge

There had been considerable interest by many groups to grow bulk, unstrained

Si1-xGex in the 1960s and early 1970s However, they failed due to the difficulty to

produce bulk Si1-xGex crystal with acceptable radial and axial homogeneity1,2 On the other hand, the development of low-temperature growth techniques, such as molecular beam epitaxy (MBE) and chemical vapor deposition (CVD), attracted a rapid increase

of different groups dealing with thin Si1-xGex films since the 1980s2 Epitaxial Si1-xGex

layers were first grown by MBE, but are now grown commercially by chemical vapor deposition (CVD) in the Si industry due to its lower cost

From the studies on Si1-xGex thin films deposited on Si substrates, it was soon realized that Si1-xGex films are tetragonally distorted when the film is thinner than a critical thickness, due to the inherent lattice mismatch of ~4.2% between Si and Ge Such Si1-xGex films are strained and have the same lattice parameter as that of Si Strain is recognized later to be an important material property, because many parameters, such as band gaps, band offsets, effective mass, etc, depend largely on strain However, upon annealing at high temperatures, the strains in the Si1-xGex/Si heterostructures can be relieved by restoring to the intrinsic cubic lattice constant of

Si1-xGex Such undesired strain relaxation degrades the performance of devices built on such substrates and poses a challenge to device stability

When growing beyond the critical thickness, the Si1-xGex films relax and results

in the appearance of misfit dislocations and significant degradation of the substrate quality To resolve this problem, a buffer layer in between the Si substrate and the desired Si1-xGex film has been proposed Ge concentration in the buffer layer gradually increases from 0% to the target value (x%) within a few microns The buffer layer serves to accommodate the misfit between Si substrate and the Si1-xGex layer On top

of this buffer layer, a Si1-xGex layer with constant Ge% is grown This layer is relaxed and can be treated as a virtual “bulk” substrates (VS) for further device

strain-fabrication In this work, we will mainly focus on the strain-relaxed Si1-xGex virtual substrates (VS) with a constant Ge content of 20%

1.2.1.2 Ni

Nickel is a transition metal with a face-centered cubic (FCC) crystal structure

at room temperature (RT) The lattice parameter of the unit cell as shown in Fig 2.2 is

occupy the “F” positions in the CaF2 unit cell2,3 NiSi2’s lattice constant is 5.406 Å, only 0.4% smaller than that of Si Hence, epitaxial growth of NiSi2 on Si is not unexpected and has been observed on both Si(111) and Si(001) surfaces by using various growth methods, e.g solid phase reaction, molecular beam epitaxy and template methods4

Ni

Si

Fig 1.3 Crystal structures of (a) NiSi and (b) NiSi2

Ni silicides are usually grown through a two-step process Firstly, Ni thin films are deposited on Si surfaces at room temperature (RT) Thereafter, the thin films are annealed to high temperatures to promote the reaction between Ni and Si With a thick

Ni film (~ 1000Å) on Si, annealing at 250oC first leads to the growth of Ni2Si phase NiSi phase starts to form when the entire Ni film has been transferred into Ni2Si5 Eventually the NiSi2 is formed in between 700oC and 800oC

Among three types of Ni silicides, NiSi is the one with most technological importance because it is able to replace CoSi2 as a contact material Compared to CoSi2, NiSi is produced at lower temperature by one-step annealing and its low sheet resistance (14-20 µΩ cm) remains unchanged even for linewidths below 0.1 µm6. In

addition, Si consumption is also relatively low during NiSi formation This is very crucial for the use of thin silicon on insulator (SOI) substrates and shallow junctions2

1.2.1.4 Ni germanides

There are mainly 2 common Ni germanides: Ni2Ge and NiGe They share similar crystallographic structures and have close lattice parameters with their counterparts of silicides: Ni2Si and NiSi, respectively7,8 Contrary to the appearance of NiSi2 at high temperature (>700oC), NiGe2 is not believed to exist in any system even after annealing to 700oC (either thin film or bulk)9,10, which makes NiGe the terminal phase in Ni-Ge system

Among the different phases of Ni germanides, NiGe has the most technological importance because it has the lowest resistivity (14-20 µΩ cm)10 and is stable over widest temperature window11 Similar to NiSi, NiGe also possesses an orthorhombic crystal structure as shown in Fig 1.3 (a) with a=5.381 Å, b=3.428 Å & c=5.811 Å12

For thick Ni layers (100-150nm) grown on hydrogen-terminated Ge(001) Ge) surface, it is generally agreed that Ni2Ge is the first phase to form after annealing between 150oC and 250oC13, although there is a controversy in the literature on the first phase obtained by reacting Ni thin film with Ge For example, others observed that

(H-Ni5Ge3 is the first formation phase149 Between 250oC and 600oC, NiGe is formed with

a flat NiGe-Ge interface as seen by XTEM The best epitaxy of NiGe on Ge(001) is observed at 500oC9 Above 600oC, NiGe starts to agglomerate and form islands9 Further increasing the annealing temperature leads to a bigger grain size and a rough surface

1.2.1.5 Ni germanosilicides

There have been intensive studies on the reactions between thick Ni films and

Si1-xGex substrates in the context of interfacial phase formation sequence, thermal and morphological stability For Ni layers (>10nm) deposited on Si1-xGex substrates, it has been found from XRD results that a Ni-rich phase (predominantly Ni2(Si1-xGex) with NiSi1-xGex) is formed when annealed below 325oC14,15 When the annealing temperature increases further (<500oC), Ni layer is fully transformed into Ni mono-germanosilicide (NiSi1-xGex) with very low sheet resistance (Rsh) values14-20 The interface of the NiSi1-xGex/Si1-xGex substrate is smooth and the thickness of NiSi1-xGex

is uniform as observed by the crossed-section TEM21 Seger et al reported that the Ni germanosilicide (NiSi1-xGex) has an orthorhombic crystal structure and the corresponding lattice parameters as a function of Ge content x can be expressed as a=5.24+0.19x Å, b=3.25+0.16x Å and c=5.68+0.15x Å19 When the temperature is elevated further, Ni germanosilicide film starts to break up and then agglomerates, causing an increase in Rsh value At even higher temperatures, Ge is expelled from Ni mono-germanosilicide (NiSi1-xGex) and segregates to the regions between NiSi1-xGex

grains As a result of Ge segregation, two phases are formed simultaneously While it

is generally agreed the first phase is the Ge-rich Si1-zGez (z>x), there is a controversy

in the literature about the second phase Some researchers observed the second phase

to be a Ge-deficient Ni mono-germanosilicide (NiSi1-yGey,y<x)15,16,22 They agreed that

it is not possible to completely deplete Ge from NiSi1-yGey and argue that Ge presence helps to suppress NiSi2 and Ni(Si1-yGey)2 formation16-17,19,154 As a result, Ni(Si1-xGex)2

does not exist as a stable phase, except for small values of x17,153 On the other hand, K

Do152, J S Luo219 and N Rath220 observed the formation of Ge-deficient Ni germanosilicide (Ni(Si1-yGey)2,y<x) by XRD, TEM, etc

di-1.2.2 Thermal stability of relaxed Si1-xGex virtual substrates

Si1-xGex virtual substrates (VS) have attracted considerable interest because of their compatibility with current Si technology together with their potential application

in high-speed electronic23 and optoelectronic devices24 Similar to Si-based device applications, it is essential to prepare clean and defect free Si1-xGex VS surface However, surfaces of Si1-xGex VS are much more complicated than Si substrates since

it has two components with different material properties, e.g thermal stability, chemical reactivity and bond strength Preferential surface segregation of Ge14,20,22,25-

28

, smaller Ge-Ge bond strength29 and desorption barrier compared to Si for example may result in the degradation of the Si1-xGex VS surface, especially at high temperatures These factors will inevitably change the desired Si1-xGex VS composition and lead to instability at the metal/Si1-xGex interface Hence, device reliability will be compromised

In this respect, there are surprisingly few reported studies which probe the effect

of temperature on the stability and morphology of Si1-xGex VS surfaces A lack of such information for instance has led to conflicting explanations and confusion about formation of blobs and holes22,25,30 often seen after annealing Si1-xGex VS deposited with thin metal films, e.g., Co or Ni It is not clear whether these defects are a direct result of material consumption due to phase formation and reaction at the Ni or Co/Si1-

xGex VS interface or due to the degradation of the Si1-xGex VS substrate itself at these

temperatures Clearly, knowledge of Si1-xGex thermal stability and hence the determination of a temperature window for metal deposition and annealing without substrate influence will be a subject of great interest, since defects arising from substrate will compromise the device performance directly Therefore, Si0.8Ge0.2 will

be used as an example to investigate the thermal stability of Si1-xGex virtual substrates

at high temperatures The results are presented in the Chapter 3 of the thesis

1.2.3 Growth mode and interfacial reaction of Ni on Si, Ge and Si0.8Ge0.2 substrates

1.2.3.1 Determining the growth modes by AES or XPS

Investigations of the first stages of thin film growth by Auger electron spectroscopy (AES) or X-ray photoelectron spectroscopy (XPS) began in the late 1960’s31-38 The first study is Cs and K monolayer (ML) adsorption on clean Ge and Si surfaces by Weber and Peria31 In both material systems, AES studies were performed

by plotting the time-variations of the Auger signals for both condensate and substrate The variations with time are found to be close to exponential Coupled with knowledge

of the escape depths of the Auger or photoelectrons, the data gives information on film thickness and modes of growth Gallon et al33-34 provided an analysis which shows that

in the case of film growth in a monolayer-by-monolayer fashion, the plots would form

a series of straight lines to which the exponential form would be an envelope Changes

in the slope (breaks or kinks) would occur at the completion of each monolayer Bauer and Poppa35 found clear evidence of such breaks for a number of condensate/substrate

n s

I

systems Monitoring growth mode this way (i.e measuring the adsorbate and the substrate signals (S) while increasing the exposure of the overlayer as a function of time (t)39) is now regarded as a standard method in electron spectroscopy, mostly AES

or XPS The S-t plots (electron spectroscopy signal versus evaporation time) should have a characteristic shape depending on the growth mode of the adsorbate metal36

If diffusion is not the limiting factor, Fig 1.4 illustrates the three main types of growth mechanisms and their accompanying XPS signal versus deposition time

Fig 1.4 Utilization of variations in XPS signal as a function of layer thickness to elucidate thin film growth (a) layer-by-layer mode (F-M); (b) island growth mode (V-W); (c) layer plus island growth mode (S-K) The dot line represents the completion of one monolayer of adsorbate39

The first film growth mechanism is layer by layer mode (Frank-van der Merwe growth) A new atomic layer does not start until the preceding layer is completed The interaction between adsorbate and substrate is preferred over itself The second growth mode occurs in the form of three-dimensional crystallites (Volmer-Weber) The adsorbate prefers to stick to itself rather than to the substrate Finally, the third

mechanism is an intermediate situation, between the extremes of the mechanism one and two The initial growth is layer by layer before three-dimensional islands form (Stranski-Krastanov mechanism) The layer-by-layer (F-M) growth is characterized by

a series of linear segments of different gradients Each “break point” (where linear segments intersect) corresponds to the completion of each monolayer before the next layer commences For Volmer-Weber (island) growth, large areas of the substrate remain clear of adsorbed material Therefore the substrate and adsorbate signal intensities show slow rates of decay and increase respectively The Stranski-Krastanov growth falls midway between these two cases (Fig 1.4(c)) Depending on the critical coverage, the XPS intensity vs coverage signals will consist of one or more linear segments followed by a slower decay/increase related to the formation and growth of islands

However, it is not easy to experimentally determine the actual growth mode based only on the characteristics of the S-t curves While the S-t curves for FM growth mode are evidenced by “breaks”, they are less obvious when growth shows no breaks

in the signal and when the data can be fitted with a single exponent For example, Cu deposition on TiO2 at RT showed no breaks40 The exponential rise and decay appear

to indicate a FM growth mode but the absence of breaks in the S-t curve makes the assignment of growth mode difficult Thus, in examining the growth modes of thin films, we will have to combine AES/XPS with other one or more techniques to avoid mistakes in interpreting the data In this thesis, both XPS and AFM will be employed

in order to probe the growth mode of Ni on clean and H-terminated Si, Ge and

Si0.8Ge0.2 substrates at RT The results are presented in Chapter 4 of the thesis

1.2.3.2 Ni growth on Si substrates at RT

Self-aligned silicidation (salicidation) process is widely used to grow silicides, which can lower gate resistance (and therefore the RC delay), reduce the source and drain parasitic resistance and increase the drive current of the MOS transistors41 As device dimensions continue to scale down to sub-half micron, the once commonly used TiSi2 displays linewidth-dependent sheet resistance caused by incomplete C49 to C54 phase transformation42 Great efforts have been put in to search for new metals which can replace Ti to form silicides and two possible metals are Co and Ni However, CoSi2 also encounters problems in ultra shallow junction application owing to its high

Si consumption and high leakage current43 NiSi, as a result, has attracted intensive attention because of its low sheet resistivity, low Si consumption, low temperature one-stop annealing and no linewidth dependence42 Nevertheless, NiSi has been reported to agglomerate at temperatures as low as 600oC and undergo phase transformation to high resistivity NiSi2 at 750oC41 The roughness of the interface between NiSi and Si substrate can influence the resistance of NiSi films, therefore it is critical to probe the initial growth of Ni on Si substrates in terms of understanding Ni growth mode and chemical interaction at RT before proceeding to high temperature anneal studies

Previous studies of Ni deposition on clean Si(111) and Si(001) surfaces at RT have revealed the following results It has been widely accepted that Ni reacts immediately with the clean Si(111) surface upon deposition at RT44-48, initially forming a thin NiSi2-like layer at the interface44,46-48 Using high-resolution medium energy ion scattering (MEIS) and reflection high energy electron diffraction (RHEED),

Nishimura et al47 found that for Ni coverage of 1ML, the RHEED pattern shows a weak 7×7 structure, which disappears completely for Ni coverages above 3ML During evaporation of Ni on clean Si(111), Murano et al observed a LEED pattern of 7×7 structure until 2ML but a halo pattern above 2ML49 With increasing Ni-coverage the surface is covered by Ni-rich silicides A metallic Ni-layer thus starts to stack above the silicide layer for Ni coverages more than 10ML Based on the Auger peak-to-peak height (APPH) of Ni (NVV,61eV) and Si (LVV,92eV) versus Ni coverages, Murano concluded that Ni growth mode on clean Si(111) is Stranski-Krastanov (SK) type

While the results on clean Si(111) surfaces are much consistent, the findings on

Ni behavior on clean Si(001)-2×1 surface are inconsistent in the literature Below 6-7

ML of Ni, Kilper et al48 found, by photoelectron spectroscopy, that a homogeneous amorphous Ni silicide layer forms Further growth of Ni leads to formation of metallic

Ni films above the silicide layer By using synchrotron ultraviolet photoemission, Wen

et al50 concluded that at 1ML Ni coverage, a large number of Ni atoms penetrate into the tetrahedral sites in the subsurface region to form an amorphous NiSi2 (a-NiSi2) bonding environment Other Ni atoms occupy two different sites above the Si dimers

to form an amorphous Ni2Si (a-Ni2Si) and amorphous NiSi (a-NiSi) At Ni coverages higher than 2 ML, the penetration of a-NiSi2 is completed and the formation of a-Ni2Si and a-NiSi becomes dominant The growth mode is layer-by-layer-like for a-Ni2Si and a-NiSi Only when the Ni coverage is larger than a critical coverage of about 4 ML does the growth of amorphous a-Ni2Si become dominant However, Hoummada et al51did not observe the formation of NiSi2 at the interface by atom probe tomography Instead, they demonstrated the formation of a uniform NiSi layer with Ni2Si particles

The Ni source they used to deposit Ni thin film contains 5at% of Pt It is unclear if the 5at% of Pt could have any influence on the interfacial reaction

Recently, hydrogen-terminated Si (H-Si) surfaces have attracted great interest since hydrogen can act as a passivating layer or surfactant by terminating the reactive

Si dangling bonds present on clean surfaces52 Termination of Si dangling bonds by hydrogen is described to alter surface energy, adatom diffusion, segregation and nucleation behaviour, resulting in the modification of the growth mode of thin films Therefore, studies of growth and structure of thin films deposited on H-Si surfaces have been investigated for a large variety of metals These metals include 3d-transition metals such as Ti49,53, Fe54, Co54-57 and Ni49,58-59, noble metals such as Cu60, Ag 53,61and Au62, and trivalent simple metals such as Al63-64 and In65 In some of these studies, hydrogen does behave as a ‘‘surfactant’’ by suppressing not only the formation of three dimensional islands10 but also interfacial metal silicides63-65

Among the aforementioned list of metals, Ni in particular, diffuses very rapidly into the bulk of Si66 Therefore, it is difficult to form a sharp and flat interface between

Ni and Si just by deposition and annealing of Ni on Si Hence, hydrogen has been adopted as a surfactant to modify the growth kinetics of Ni on Si substrates While NiSi2-like layer is formed when depositing Ni on clean Si(111) surface, both Murano

et al49 and Hirose et al58-59 did not observe spontaneous formation of NiSi2 during growth of Ni on H-Si(111) at RT, although Hirose58 found that Ni atoms diffuse beneath the Si surface without breaking the surface Si-Si bonds Both of them agreed that hydrogen atoms terminating the Si surface dangling bonds suppress the silicide reaction at RT Murano et al49 reported that 7×7 LEED pattern is observed until 2ML

of Ni, and the 1×1 structure of Si(111) is also observed until 3ML of Ni The LEED

spots disappear above 3 ML while no ordered structure is observed Based on Auger peak-to-peak height (APPH) of Ni and Si as a function of Ni coverage, Murano et al inferred that the growth mode of Ni on H-Si(111) is layer-by-layer (FM) type instead

of layer-plus-island (SK) type in the case of clean Si(111)

Similar to the work on H-Si(111) surfaces, the findings about Ni interaction with H-Si(001) surfaces appear also to be inconsistent in the literature By using scanning tunneling microscopy (STM), Yoshimura67 believed that Ni atoms deposited

on H-terminated Si(001) surface are prevented from reacting with Si atoms As a result, clusters form and they are bigger on H-Si(001) surface than that on clean Si(001) surface due to enhanced surface migration However, by repeating a similar STM experiment, Yoshimura66 reported an opposite phenomena, namely, the Ni clusters are smaller on H-Si(001) surface than that on clean Si(001) To the best of our knowledge, there is lack of information about Ni growth mode on H-Si(001) surface

In addition, it is still a controversy whether there is interfacial reactions when Ni is grown on H-Si(001) and H-Si(111) surfaces at RT

1.2.3.3 Ni growth on Ge substrates at RT

Although Ge is the semiconductor used to build the first bipolar transistor by Bardeen and Brattain68, it is soon replaced by Si due to the inferior thermal stability and water solubility of germanium oxides Since then, Si has been dominating as the primary substrate in VLSI owing much to the stability, physical properties and processability of its oxide As device dimensions keep shrinking, SiO2 is expected to

be replaced by other high-dielectric constant (high-k) materials69-70, e.g., ZrO2 and

HfO2 The replacement of SiO2 will considerably reduce the technological advantages

of Si and provide room for the integration of other semiconductor materials with better performance for device fabrication There is great interest in using Ge as a channel material for high-performance logic circuit application because of Ge’s high intrinsic mobilities of both electrons and holes1 and its compatibility with Si processing technology For instance, Ge metal-oxide-semiconductor field-effect transistors (MOSFET) with germanium oxynitride dielectrics have been successfully demonstrated71 The promising results suggest that by adopting Ge as the channel material, complementary-metal-oxide-semiconductor (CMOS) technology will be able

to extend beyond the limits imposed by Si

To maximize the superior transport properties of Ge and reduce the high series resistance observed in these Ge MOSFETs, metal germanides natually appear as candidates for contacting Ge, in the same way that self-aligned metal silicides are used

in standard CMOS process today to contact the source, drain, and gate regions of the transistors A systematic investigation of the thermally induced reactions of 20 transition metals with Ge has been carried out11 It was discovered that two monogermanides, NiGe and PdGe, are the most promising candidates as they can be formed at low temperatures, stay stable over a wide range of temperatures, and exhibit low resistivities (22 µΩ cm for NiGe and 30 µΩ cm for PdGe on amorphous Ge (a-Ge)) Due to the wide application of Ni in current VLSI technology, interest has been focused on further understanding the Ni-Ge materials system

Compared to the extensive work about Ni growth on Si substrates at RT3-6,44-67, there is a paucity of information pertaining to Ni deposition on both clean and H-

Ge(001) substrates at RT Therefore, we will adopt the existing work about Ni growth

on clean and H-Ge(111) surfaces as a reference

By exploiting LEED and AES, Girardeaux et al72 noticed a large increase in Ni signal upon initial deposition of Ni on clean Ge(111) at ambient temperature The Ni signal then increased slowly and the Ge signal was always detected even after 2 hours

of deposition Judging from the Auger signal versus deposition time plot (“AS-t”), they concluded that the growth mode of Ni on Ge(111) is SK type: after a continuous Ni is formed initially, 3D nickel crystallites grow subsequently They also inferred that surface compound can form during growth of germanium on the nickel layer Nevertheless, they did not provide any information about the possible Ni germanide phase at the interface By using RBS, Smith et al73 also showed that Ni reacts with Ge substrates and it effectively displaces 3ML of Ge atoms from their original lattice positions, forming an interfacial layer of NiGe for Ni coverage below 3ML However, further deposition of Ni leads to a continuous, uniform metal film, similar to layer-by-layer growth Although both of them used repeated cycles of sputtering and annealing method to prepare the clean Ge(111) surfaces, it is not clear whether the sputtering methods adopted by different groups will modify the original Ge(111) surfaces in different ways and therefore result in different Ni growth modes

To probe the effects of surfactant, i.e., hydrogen, on the surface reaction, Thundat et al74-75 studied the growth of Ni by electrodeposition on H-Ge(111) with the aid of STM The resulting topography shows a smooth surface with an average roughness of 5Å for Ni coverages below 5ML With the help of X-ray standing wave (XSW) technique, they concluded that Ni reacts with Ge substrate at the interface, forming a few layers of nickel germanide At a coverage of approximately 50ML Ni,

the STM topography shows a surface decorated with Ni islands Thundat did not point out, however, that the growth mode of Ni on H-Ge(111) by electrodeposition seems to remain as SK mode based on their STM and XSW results Contrary to Thundat’s observation, Hsieh et al9 found by TEM that no interfacial interaction is detected in the as-deposited samples, after using electron gun deposition of 30nm Ni on H-Ge(111) This conflict may be because the much thicker Ni film deposited by Hsieh hinders the observation of the interface reaction, which only takes place during the initial deposition of very thin Ni films

Although a lot of research have been focused on the formation and stability of

Ni germanides on Ge(001) and Ge(111) substrates at high temperature11-13,76-81,current studies lack a clear understanding on the growth mode of Ni on clean & H-Ge(001) substrates at RT The controversy about whether Ni reacts with H-Ge(001) at RT needs to be resolved as well

1.2.3.4 Ni growth on Si1-xGex substrates at RT

With the emergence of Si1-xGex based materials for high performance devices,

it is increasingly important to understand the interaction between metals and Si1-xGex

substrates in order to develop suitable germanosilicides for interconnects and contact applications Among various germanosilicides, Ni germanosilicide (NiSi1-xGex) is very promising due to its low resistance and the compatibility of Ni with current Si technology Although there are many papers concerning Ni germanosilicide formation

at high temperatures14-22, there seems to be a lack of information regarding the growth

mode of Ni on both clean and H-Si1-xGex(001) substrates at RT as well as the reports regarding the possible interaction between Ni and Si1-xGex upon deposition at RT

1.2.4 Oxidation of Ni silicides, Ni germanides and Ni germanosilicides at RT

Si is destroyed followed by the formation of an amorphous SiO2 overlayer The oxidation continues to proceed further inwards only after the current Si layer is completely terminated by oxygen Therefore layer-by-layer growth of native oxide has