Metallic thin film on sige si substrates 2

2.2 Sample handling and treatment 2.2.1 Growth method of Si0.8Ge0.2 virtual substrates The Si0.8Ge0.2001 virtual substrates used in this project are grown by chemical vapor deposition CV

Chapter 2 Experimental

2.1 Introduction

In this chapter, we will first explain how to clean substrates, deposit Ni thin films and anneal the samples in an ultra-high-vacuum (UHV) environment Then we will introduce the main characterization facilities used in this thesis as well as methods

of data analysis

2.2 Sample handling and treatment

2.2.1 Growth method of Si0.8Ge0.2 virtual substrates

The Si0.8Ge0.2(001) virtual substrates used in this project are grown by chemical vapor deposition (CVD) on Si(001) substrates at 800oC in Imperial College, UK Starting from a clean Si(001) surface, a buffer Si1-xGex layer of 3µm thick is first grown with the Ge content (x) linearly increasing from 0% to 20% After that, a

Si0.8Ge0.2 layer of 1µm thick with constant Ge content (20%) is grown on the top with boron doping level of 1017/cm3 A schematic drawing of the structure is shown in Fig 2.1 (a) There is no chemical mechanical polishing (CMP) treatment after the growth Therefore, a cross-hatch pattern is present on the surfaces (Fig 2.1 (b))

The “cross-hatch” formation on Si1-xGex VS is believed to be associated with the strain relaxation process during the growth of VS However, its formation mechanism is still under investigation Gao et al attributed the mechanism to the stress-

induced morphological instability221 During growth, the strain caused by the lattice mismatch drives diffusional atomic flux along the film surface such that an initial flat film gradually evolves into an undulating profile, i.e., cross-hatch pattern However, using channeling-contrast-microcopy (CCM), Seng et al found evidence showingthat the cross hatch structure is present in both compositionally graded structures and constant composition SiGe layer222 Seng believed that the pattern is associated with the presence of slight lattice tilt

Fig 2.1 (a) Schematic drawing of Si0.8Ge0.2 virtual substrate structure and (b) the typical 10 µm × 10 µm AFM image of Si0.8Ge0.2 virtual substrate after growth

2.2.2 Cleaning methods for Si, Ge and Si0.8Ge0.2 virtual substrates

Hydrogen terminated Si and Si0.8Ge0.2 samples are prepared by using a modified RCA wet chemical cleaning method, which is described in details in Table 2.1 and Fig 2.2 It is important to note that only Teflon© tweezers are used to handle samples throughout the cleaning process

The cut Si and Si0.8Ge0.2 substrates are firstly immersed in Solution A to remove the Si and Ge native oxides Following that, they are rinsed in de-ionized (DI) water to wash away the remaining hydrofluoric acid (HF) Note that DI water with less than 4 parts per billion or with electrical resistivity not less than 18.2 MΩ•cm should

be used for rinsing Next, the substrates are introduced into Solution B, where the Si and Si0.8Ge0.2 surfaces are re-oxidized, before the substrates are re-immersed into Solution A again The “oxide strip-and-grow” process is repeated three times to enhance the removal of carbon contaminations Further removal of organic contaminants can be achieved by transferring the substrates into Solution C The solvating action of NH4OH and the powerful oxidizing action of H2O2 will eliminate stubborn and residual carbon species Finally, the Si and Si0.8Ge0.2 substrates are exposed to fresh Solution A, followed by a dip in DI water before mounting on the sample holder and introducing into the UHV chamber as quickly as possible to minimize dwell time of the sample in air

To clean the as-received Ge substrates, a simpler cleaning sequence is adopted The Ge wafers are introduced in Solution A for 1 minute followed by rinsing in DI water for 1 minute This process is repeated for three times and thereafter the Ge samples are quickly loaded into the UHV chamber Both Solutions B and C significantly degrade Ge’s morphology and therefore are avoided

Solution A Solution B Solution C

Fig 2.2 Flow chart summarizing the modified RCA cleaning procedures

The final immersion of Si, Ge and Si0.8Ge0.2 substrates in HF solutions causes H-termination at the surfaces by bond formation between Si & Ge dangling bonds at the surfaces and hydrogen (H), thus passivating the dangling bonds of surface Si and

Ge atoms with H and leaving surfaces with low defect density155 This results in the Si,

Ge and Si0.8Ge0.2 surfaces being less susceptible to (1) airborne contaminations prior to introduction into the UHV chamber; (2) hydrocarbons from turbo-molecular pumps while substrate resides in the load-lock chamber; (3) residual oxygen and hydrocarbons

in the UHV chamber156 The samples are thus stable in the UHV for extended periods

Solution B at 60°C (5 min)

Rinse in Deionized Water (1 min) Repeat 3 times

Solution A (1 min)

Rinse in Deionized Water (1 min)

Solution C (10 min)

Prior to Ni deposition, the cleanliness of the sample’s surface is examined by XPS Figure 2.3 represents typical survey spectra taken from the as-received, H-terminated and clean Si(001) surfaces, respectively Compared to the as-received Si surface, the XPS peaks belonging to C 1s, O 1s and O KLL are clearly not detected on the H-terminated and clean Si surfaces, which indicate that this modified RCA cleaning method is effective in removing the native oxide and carbonaceous contaminants Similar cleaning effects have also been observed on Ge and Si0.8Ge0.2

surfaces and hence are not shown here

Alternatively, clean Si, Ge and Si0.8Ge0.2 surfaces can be prepared by heating the as-received Si, Ge and Si0.8Ge0.2 surfaces to 650oC, 400oC and 650oC inside the UHV chamber, respectively The corresponding native oxides (SiO2 or GeO2) on the surfaces can be thermally decomposed and then desorbed from surface However, this method is not adopted to prepare the clean Si, Ge and Si0.8Ge0.2 surfaces because it roughens the surface morphology and leads to carbide formation on the surfaces

Fig 2.3 XPS survey spectra of as-received, H-terminated and clean Si (001) surfaces

1100 1000 900 800 700 600 500 400 300 200 100 0 Clean Si(001) Surface

H-terminated Si(001) surface

As-received Si(001) surface

2.2.3 Ni deposition method

In this dissertation work, Ni deposition on Si, Ge and Si0.8Ge0.2 surface is achieved by using an OMICRON EFM 3 UHV electron-beam evaporator (Fig 2.4(a)) Evaporant materials, i.e., Ni, are in the form of rods The bombarding electron beam, from a heated tungsten (W) filament, induces a temperature rise of the evaporant, causing Ni evaporation The evaporator is equipped with a flux monitor (Fig 2.4(b)), which is actually an ion collector and is located in the evaporant exit column At a given electron emission current, IE, and e-beam energy, the ion flux measured is directly proportional to the flux of evaporated atoms Once calibrated, the flux monitor can replace a quartz thickness monitor by continuously monitoring the evaporation rate The interior components in the front part of the evaporator can be seen more clearly in Fig 2.4(b)

(a)

(b) Fig 2.4 shows (a) the entire OMICRON EFM 3 structure and (b) the cross-section view of front part of e-beam evaporator (Source: OMICRON EFM3 user manual)

Before evaporation, the sample is positioned on the same axis as the evaporant exit column Filament current, If, will be increased to about 1.95 A to supply the bombardment electrons A positive bias ranging from 850 V to 950 V is applied to the evaporant rod This is to attract and to focus the electrons towards the end of the rod

At this stage, IE would have increased and the end of the evaporant rod should be heated up The parameters mentioned can be adjusted until a suitable and stable flux is achieved The Ni deposition can then be started by opening the cell shutter To

terminate the deposition process, the cell shutter is closed and the control parameters decrease slowly to cool the evaporant rods

2.2.4 Annealing method

In-situ sample annealing inside XPS system was achieved through resistive heating in the UHV Analysis chamber During annealing the pressure was kept at ~10-9mbar region The sample temperature was monitored by a thermocouple in direct contact with the sample surface A picture of two annealing sample holders is shown in Fig 2.5

Fig 2.5 Pictures of two annealing sample holders

Thermal Couples

2.3 Ultrahigh Vacuum (UHV) System

2.3.1 Why UHV is needed

The second reason is related to the requirement of photoelectron spectroscopy (PES) analysis PES techniques, i.e., X-ray photoelectron spectroscopy (XPS), Ultraviolet photoelectron spectroscopy (UPS), Auger electron spectroscopy (AES), typically analyze only the few top atomic layers (<10nm) with a sensitivity of a few percent of an atomic layer Hence, these techniques are extremely sensitive to minute amounts of surface contamination, which requires that the time for a clean surface to

be covered by a layer of contamination inside the UHV chamber must be less than the time to perform the analysis experiment From the consideration of the kinetic theory

of gases, the rate of the surface bombardment by gas molecules is given by:

The UHV environment is also essential in prolonging the lifetime of various PES components, i.e., X-ray gun, channeltrons, etc Therefore, a UHV environment is clearly indispensable in order to in-situ probe the reactions on sample surfaces by XPS

2.3.2 UHV system layout

The UHV system used in this work is a VG ESCALAB 220i-XL Imaging XPS and is shown in Fig 2.6 It consists of three chambers - fast entry air lock (FEAL), preparation chamber and analysis chamber The FEAL and preparation chamber are

pumped down by turbo molecular pumps with rotary pumps, while the analysis chamber is pumped down by two ion pumps and one titanium sublimation pump (TSP) The pressures in preparation chamber and analysis chamber are monitored separately by two ion gauges

Fig 2.6 Layout of VG ESCALAB 220i-XL Imaging XPS system

There is one pneumatic gate valve in between FEAL and preparation chamber and a second pneumatic gate valve in between preparation chamber and analysis chamber in order to provide isolation for each chamber The FEAL chamber is used to introduce the samples from the ambient into the system quickly and then pump down

to mid-vacuum level (~10-8mbar range) Next the samples are transferred from FEAL

to preparation chamber via an extendable transfer arm after the pneumatic gate valve between them is open A nickel e-beam evaporator is installed in the preparation chamber and is used to deposit Ni onto the samples’ surfaces, during which the

Preparation Chamber

FEAL Analysis Chamber

Ni e-beam

evaporator

Analyzer

Transferring arm

pressure is maintained at ~2-3×10-9 mbar Thereafter the samples are transferred from preparation chamber to analysis chamber through another extendable transfer arm, and the XPS analysis is carried out there A 5-axis manipulator is installed in the analysis chamber, which is able to heat as well as to cool the annealing sample holder shown in Fig 2.5 After the experiments, the samples are taken out by the reverse sequences of operation

2.3.3 Bake-out procedures

Once the UHV chambers are vented to air for maintenance, a bake-out is required to restore the system back to UHV condition quickly After the components which cannot withstand high temperatures (i.e., electrical cables, plastic covers, etc) are removed from the system, the entire system is covered and sealed by baking panels and is heated up to 110oC, where the system stays for ~15 hours By doing so, the water and other gas molecules absorbed on the inner chamber walls will have larger kinetic energies to desorb from the wall and to be pumped away quickly Once the system is cooled down and all the components containing filaments are degassed, a UHV environment can be achieved

2.4 In-situ X-ray Photoelectron Spectroscopy (XPS)

2.4.1 The basic principles

X-ray photoelectron spectroscopy (XPS), also known as Electron Spectroscopy for Chemical Analysis (ESCA), has been widely used to investigate the chemical composition at the sample surface, after it was introduced in the mid 1960s by K Siegbahn and his research group at the University of Uppsala, Sweden XPS is able to detect most of the elements except hydrogen and helium The principle of the technique is based on the photoelectric effect outlined by Einstein in 1905 (Fig 2.7)223 The photoemission principle states that if an atom absorbs a photon, an electron from the core shell of the atom will be ejected out from the atom, provided that the photon energy is greater than the sum of that electron binding energy (EB) and the work function (Φ) between Fermi level (EF) and Vacuum level (EV) The Fermi level (EF) is defined to be at zero binding energy position Hence, the resultant kinetic energy, EK, possesses by the photoelectron will be given in Eqn (2.4) by applying the principle of the energy conservation:

EK = hν - EB - Φ (2.4) Although it is the kinetic energies of the outgoing photoelectrons that are measured experimentally, spectra are usually displayed on a binding energy scale to allow easy elemental identification This is achieved by changing eqn (2.4) into

EB = hν - EK - Φ (2.5) Here, if hν and Φ are known, the measured EK would allow us to obtain the EB

Fig 2.7 Schematic diagram of the photoemission process The vacuum level (EV) is the energy of an electron at rest (zero kinetic energy) in a vacuum far away from neighboring particles such that it has no interaction with them The work function (Ф)

of a solid is defined as the minimum energy to remove an electron from the highest occupied energy level in the solid to the ‘vacuum level’

2.4.2 Instrumentation

The VG ESCALAB 220i-XL Imaging XPS is equipped with a monochromatic

Al Kα (1486.7 eV) and two un-monochromated Mg Kα (1253.6 eV) and Al Kα X-ray source, a concentric hemispherical electrostatic energy analyzer (CHA) and a magnetic immersion lens (XL lens) to increase the sensitivity of the instrument for small area XPS These components would be briefly described in the following sections

2.4.2.1 X-ray source

The anode material for an X-ray source has to meet the following two requirements Firstly, the line width must not limit the energy resolution required in the

technique and, secondly, the characteristic X-ray energy must be high enough that a sufficient range of core-shell electrons can be exited for most elements in the periodic table

Fig 2.8 Schematic cross section of the X-ray source

The monochromated Al Kα X-ray source meets the above requirement and has been used in our study As shown in Fig 2.8, after a focused electron beam (15KeV) is bombarded on the anode material, i.e., Al, a characteristic Al X-ray is produced This X-ray is then diffracted by a pair of quartz crystals, which are arranged in an angle such that only Al Kα wavelength meets the Bragg diffraction requirement and hence is focused into a well defined spot on the sample surface with an energy resolution of around 0.45eV In addition there is a thin aluminum window at the exit of Al Kα which can allow the majority of the X-ray to pass through but prevent unwanted low-energy electrons from reaching the sample surface

2.4.2.2 Electron energy analyzer

The core of the XPS technique is to measure the photoelectrons’ kinetic energies accurately, which is achieved via an electron energy analyzer or spectrometer

Filament

Cathode

Wehnelt Einzel Lens

Anode X-ray Beam Electron Beam

(marked in Fig 2.6) From the schematic cross-section of the spectrometer (Fig 2.9), it can be seen that the electron energy analyzer consists of two electrically isolated concentric hemispheres with a potential difference between them The electrostatic field in between the two hemispheres filters the photoelectrons by only allowing the passage of electrons with a chosen kinetic energy (the ‘pass energy’) to the detector Electrons having kinetic energy less or more than the chosen ‘pass energy’ collide with either the inner or outer hemisphere and neutralize To collect photoelectrons with higher kinetic energies than the pass energy, they are slowed down to the pass energy

at the entrance of the hemisphere by applying a negative voltage at the retard plate (ALPHA-PLATES) Through varying the negative voltage applied on the retard plate, electrons with different kinetic energies are allowed to pass the analyzer and reach the detector

The analyzer can be operated in either constant analyzer energy (CAE) mode or constant retard ratio (CRR) mode In CAE mode, the pass energy, Ep, is constant as the kinetic energy is scanned by the lens to build up a spectrum In the CRR mode, the voltage difference between the hemispheres is proportional to the retarding voltage of the analyzer

Each mode has its advantages and disadvantages CAE mode has the advantage

of fixed resolution at all kinetic energies However, at low kinetic energies the sensitivity increases in CAE but decreases in CRR In general, the number of electrons ejected from a sample is greatest at low kinetic energies While CAE tends to exaggerate this effect, CRR tends to counter it and produces a flatter spectrum at low kinetic energies Because of the properties of the analyzer, historical reasons and the