kjm3020-lars kristian henriksen
TRANSCRIPT
U��������� �� O���D��������� �� C��������
U������������ R������� II
Deposition of n+ poly-Si thin films
Lars Kristian Henriksen
A paper submitted in fulfilment of the requirementsfor the course
KJM3020.
20 credits
Spring, 2016
AbstractA method for depositing highly phosphorus doped silicon ([P]⇠ 5 ⇥ 1019 cm�3) by sputtering, pro-ducing poly-crystalline and conductive thin films without the use of a substrate bias is explained.The samples were characterized by four point measurements, Hall e�ect measurement, atomic forcemicroscopy, x-ray di�raction and secondary ion mass spectrometry. It is shown that samples etchedby reactive ion etch prior to sputter deposition at 200�C for 30 minutes and subsequent furnaceannealing at 1100�C for 30 minutes produces 270 nm thick films with a sheet resistivity as lowas ⇠ 2.8 ⇥ 10�2 ⌦ cm and a carrier density and mobility of ⇠ 2.0 ⇥ 1019 cm�3 and ⇠ 12 cm2/ Vs,respectively. X-ray di�raction patterns indicated both Si (111) and (220) planes, and atomic forcemicroscopy statistical calculations found the surface to have a root mean square average height of⇠ 1 nm.
i
C�������
1 Introduction 1
2 Theory 22.1 Silicon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.2 Doping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.3 Crystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3 Techniques 53.1 Pre-deposition Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.1.1 Reactive Ion Etch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2 Deposition Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2.1 Electron Beam Assisted Physical Vapor Deposition . . . . . . . . . . . . . . . . 53.2.2 Sputtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.3 Post Deposition Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.3.1 Rapid Thermal Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.3.2 Furnace Annealing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.4 Characterization Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.4.1 4-point probe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.4.2 Hall e�ect measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.4.3 Resistivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93.4.4 Van der Pauw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103.4.5 Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.4.6 X-ray di�raction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.4.7 Secondary Ion Mass Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4 Experimental 144.1 Phase I - Initial investigations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.1.1 Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144.1.2 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2 Phase II - Heat treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154.2.1 Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164.2.2 Post deposition treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164.2.3 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.3 Phase III - Substrate preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184.3.1 Pre Deposition Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184.3.2 Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184.3.3 Post deposition treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184.3.4 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5 Results 215.1 Phase I & II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215.2 Phase III . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.2.1 Four Point Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225.2.2 Hall E�ect Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
ii
5.2.3 Secondary Ion Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . 245.2.4 Atomic force microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275.2.5 X-ray di�raction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6 Discussion 30
7 Conclusions 31
iii
1I�����������
Fall 2015. Paris. Top politicians from 195 countries meet to discuss the future of the planet [1].After Copenhagen Climate Change Conference in 2009 the world doubted that Paris would be agame changer and few believed that a climate deal would be struck. But 12/12/2015 the agreementwas signed, making it the first-ever legally binding climate deal. The agreement spring into actionin 2020 and it is believed that it will help us reach the much talked of two degree goal [2]. Norwayproposed a 40% reduction in greenhouse gases by 2030 [3] and through this indirectly committeditself to become a leading figure in the green technology revolution. To reach the goals set forth inParis, new technology is required, and first and foremost green technology.
An important aspect to consider when finding new ways of harvesting energy is the environmen-tal aspect of both production and use. Solar cells made from potentially toxic materials such as GaAsand CdTe are not uncommon [4] and poses a potential risk to both humans during production of theunits, as well as the environment, during production and the end of life disposal. Making solar cellsof the future greener should therefore also imply making them environmentally friendly and nontoxic.
This project is part of a bigger project with an ultimate goal of creating a tandem Si-Cu2O solarcell capable of capturing a larger portion of the incoming sun radiation, enabling higher e�ciencies,than the single junction silicon solar cells of today, while sticking to low cost materials. As siliconand cuprous oxide are both more enviromentally friendly than other solar cell setups [4] and highlyabundant in the earth crust these are highly attractive materials to investigate.
One of the challenges of this new tandem structure is creating a tunneling junction between thesilicon cell and the cuprous oxide cell to enable electron transport between the two. To achieve thisquantum e�ect a well defined p+-n+ junction is needed and in this paper a method for thin filmdeposition of highly phosphorus doped ([P]⇠ 5 ⇥ 1019 cm�3) silicon targets onto a quartz substrateis explained in an e�ort to facilitate a safe and scalable method for making highly doped silicon thinfilms.
The original project description read
The purpose of this project is to deposit a thin film n+- Si-layer onto a silicon substrateby the use of e-beam evaporation and sputter deposition and to characterize this thin film.The goal is to make a thin film with high charge carrier mobility and concentration. Thethin film should also be a suitable substrate for further Cu2O deposition. The task is partof a bigger project of making a tandem solar cell combining a conventional silicon solarcell with a Cu2O cell.
This description was largely followed but a slight modification was imposed early in the projectmaking sure to keep the time frame set forth; instead of using a silicon substrate quartz would beused to simplify sample characterization. Surface characterization to specifically investigate thepossibility of further growth of Cu2O was not undertaken.
1
2T�����
2.1 SiliconSilicon is a group IV semiconductor and is often referred to as an elemental semiconductor becauseit is composed of one species of atoms. Today silicon is by far the most common semi conductormaterial in rectifiers, transistors and integrated circuits as well as in solar cells due to its highabundance, relatively low cost and superior properties.
(a) Diamond crystal structure unit cell. Source: http:
//hyperphysics.phy-astr.gsu.edu/hbase/solids/sili2.html
0 1
20
1
20 1
2
0 1
20
1
4
3
4
3
4
1
4
(b) Atomic positions in the cubic cell where points at0 and 1
2are on the fcc lattice. The points 1
4and 3
4are on a similar lattice displacement along the bodydiagonal by one-fourth of its length [5].
Figure 2.1: The silicon unit cell crystal in 3D and 2D.
Silicon in a pure and perfect crystalline form takes on a diamond crystal structure, an fcc latticewith a two atoms basis. In this structure all Si-atoms are bound to four other Si-atoms by directionalcovalent bonding as shown in figure 2.1a.
This produces a unit cell consisting of 8 atoms, four of which are completely enclosed in the unitcell, plus four that is on the fcc lattice as shown in figure 2.1b. According to Kittel [5] the latticeconstant for silicon is a = 5.430, where a is the edge of the conventional cubic cell.
Introduction of impurities in the silicon crystal, substituting the Si atoms with group III or Velements such as boron (III,trivalent) or phosphorous (V,pentavalent), introduce a change in charac-teristics and is called doping. This is due to the di�erence in the number of silicon valence electronsand the valence electrons of the introduced impurity.
Boron, an acceptor, has one less valence electron and a mobile positive charge will be introducedto the lattice, often referred to as a hole. Enough impurities introduces a new energy band in thesilicon band gap close to the valence minimum band energy level (VBM). Phosphorous, a donor,will introduce a mobile negative charge, electron, to the lattice and a new energy band close to theconduction band minimum energy level (CBM). By introducing new energy bands close to the VBMor CBM, less energy is needed to excite electrons from the valence band or to the conduction band,enabling conduction.
2
CHAPTER 2. THEORY 2.2. DOPING
Figure 2.2: The pn - junction characteristics. Source:https://upload.wikimedia.org/wikipedia/commons/f/fa/
Pn-junction-equilibrium-graphs.png
Semiconductors withan excess of negativelycharged carriers are calledn-type semiconductors; anexcess of positively chargedcarriers are called p-type.Combining bulk p- and n-type material gives riseto a p-n junction. Asshown in figure 2.2 mo-bile charge carriers dif-fuse over the junction tocancel out charge and bydoing this creates a re-gion around the junctionwith stationary chargeswhich sets up an elec-tric field. At equilibriumthis electric field is greatenough to stop any fur-ther di�usion of free car-riers and a depletion re-gion is formed.
This region acts as abarrier for majority car-riers (electrons on n-sideand holes on p-side), anda low resistance path forminority minority carri-ers. By introducing aforward or reverse biasover the junction the de-pletion region can be ma-nipulated, but it can alsobe manipulated by dop-ing the samples; highlydoped samples will havea smaller depletion zoneas more fixed charge are
concentrated close to the junction. Oppositely, lightly doped samples experience a deep depletionzone as fixed charge is less dense.
With a small enough depletion zone a quantum e�ect called tunneling is more probable. Elec-trons in the p-side valence band tunnel to the n-side conduction band without experiencing thebarrier at the p-n junction since the probability of electron tunneling through an energy barrierincrease exponentially with decreasing tunneling distance.
2.2 DopingDoping semi-conducting materials is today a widely used technology with almost unlimited applica-tions. According to Wikipedia [6], the history of doped semi-conductors may be traced back to 1885when it was first observed that the properties of these materials were due to the impurities withinthe material itself, but it was not until the mid 40’s that a process was formally developed by JohnRobert Woodyard.
One of the first groups to report on incorporating pentavalent (elements with five valence elec-trons) and trivalent impurities (elements with three valence electrons) into a tetrahedral amorphoussemiconductor (a-Si) was Spear and Comber [7] in 1975 by chemically reacting silane gas (SiH4)mixed with phosphine gas (PH3) in a process they called "electrode-less glow discharge", knowntoday as Plasma Enhanced Chemical Vapor Deposition (PECVD). Spear and Comber [7] state that
3
CHAPTER 2. THEORY 2.3. CRYSTALLIZATION
...the e�ect of the pentavalent impurity has been to increase the conductivity of a-Siby seven orders of magnitude...
This process have been perfected over many years and with modern equipment, groups like Ku-mar et al. [8] produce samples with carrier concentrations close to 4⇥ 1020 using PECVD.
Producing highly doped silicon thin film can be accomplished in a number of ways. One partic-ularly promising method is sputtering, this due to the reduced temperature needed for deposition,elimination of toxic gases such as PH3 and ability to pre-dope the Si-targets [9].
Sputtering of highly doped Si is far less discussed in literature than PECVD, but both Jun et al.[10] and Wang et al. [11] has had success with the method producing films with mobility and carrierconcentrations of 1.2 cm2/ V s and 6.5 ⇥ 1017 cm�3 (n-type), respectively. For further discussion onthis topic refer to section 3.2.2.
2.3 CrystallizationDue to the fact that sputtering Si onto a substrate produces amorphous thin-films, a post-depositioncrystallization process is required to achieve the poly-Si films needed for device application since theconductivity and resistivity, are highly dependent on crystallinity. Understanding the crystalliza-tion processes is therefore crucial.
This project utilizes solid phase crystallization due to the negative e�ects liquid phase crystal-lization would have on the p+ - n+ transition layer. Keeping doping levels intact throughout thecrystallization process is important to obtain the desired characteristics.
When crystallizing a Si-film there are mainly two processes a�ecting the outcome: nucleationand growth [12–15]. As Spinella et al. [14] states in their paper, growth rate highly depends onavailable nuclei in the sample after deposition, and that the properties of the polycrystalline Silayers depend on the properties of the as-deposited a-Si layers. Their results indicates that favor-able nucleation sites occur during deposition and in an e�ort to control these nucleation sites a predeposition reactive ion etch is included in this project.
In addition to the nuclei dependence, crystallization rate has a strong dependence on the anneal-ing temperatures as the activation energy for nucleating silicon crystals has a high value of 3.9 eV[16]. Typical crystallization parameters are 600�C for 20h [10], while it is 750�C for 2 minutes [11](rf magnetron sputtered films with and without substrate bias) up to 900�C-1100�C for 3 minutes[17] in the case of rapid thermal processing. This shows that the crystallization rate of sputtereda-Si increases rapidly with increasing temperature up to about 1100�C-1200�C.
4
3T���������
3.1 Pre-deposition Techniques3.1.1 Reactive Ion EtchIn reactive ion etching (RIE), often referred to as dry etching or plasma etching, ionized atoms areaccelerated towards a sample under an applied negative bias. In non-conduction materials such assilicon a radio frequency (rf) potential is applied to maintain the bias [18].
RIE may use chemically reactive plasma to attack the sample, chemically removing the top layerof the sample atom by atom. It is also possible to increase the power and use this as a physical etchby sputtering of the surface (see 3.2.2);it is this physical method that is being used in this project.The steps in RIE are adsorption, reaction and desorption. The amount removed depends on etchingtime, power and reaction gas and is hence controllable as shown by Abe et al. [19].
3.2 Deposition TechniquesThere are a vast number of deposition techniques used in thin film production today. They areusually classified as chemical vapor deposition (CVD) or physical vapor deposition (PVD) and di�erin that CVD utilizes a chemical reaction to obtain film growth while PVD does not. Instead PVDvaporizes a target material by either a thermal (e-beam) or athermal (sputter) process.
CVD and especially plasma-enhanced CVD (PEVD) is a widely used technique in preparation ofsilicon thin films [8, 20–23]. This technique utilize the chemical reaction of multiple gases to growthe film, such as PH3 and SiH4 to produce highly doped silicon films as described by Kumar et al. [8].This technique produces high quality films but does come at a price as the phosphine gas is highlytoxic [24].
With this in mind, the preferred techniques in this project does not involve CVD’s. Both electronbeam assisted PVD (EB-PVD) and sputtering are classified as PVD techniques and the followingdiscussion will mainly concern these.
3.2.1 Electron Beam Assisted Physical Vapor DepositionA quick historical reviewThe predecessor to this technique is thermal evaporation of solids first described by Faraday in the1850’s. Faraday [25] used thermal energy to either evaporate or sublimate the target material. Theroad onward is described in detail by Anders [26] covering the history of cathodic arc coating.
Electron Beam Assisted Physical Vapor Deposition TheoryIn Electron Beam Assisted Physical Vapor Deposition (EB-PVD) a high energy electron beam isproduced and subsequently directed at the target material to induce the vaporization. In this set-upthe source producing the electron and target acts as the cathode and anode, respectively, where theformer is e�ectively a filament exposed to a high voltage (6-40 kV), in turn ejecting a high density
5
CHAPTER 3. TECHNIQUES 3.2. DEPOSITION TECHNIQUES
flow of electrons. These electrons are then deflected and directed at the target material by an appliedelectric and magnetic field.
The chamber where this process occurs should be held at high vacuum as this will increase themean free path of the vaporized atom, reducing the gas-to-gas collisions en route to the substrateand increasing the deposition rate and uniformity.
As Arunkumar et al. [27] further states and concludes, the EB-PVD is a process involving severalcrucial steps and the deposition process depend on the applied voltage, vacuum pressure and electronbeam diameter incident on the target.
3.2.2 SputteringSputtering is an attractive technique for depositing Si films, compared to CVD, due to the re-duced temperature needed for deposition, elimination of toxic gases such as PH3, controlling theH2-concentration in the deposited film and ability to pre-dope the Si-targets [9]. It is also easilyscalable and ’industry friendly’.
A quick historical reviewSputtering was discovered in 1852 and 1858, independently, by William Robert Grove and JuliusPlücker, respectively [28]. The practical use of sputtering may be traced all the way back to 1877[29] when it was used to coat mirrors.
For many years this technique was used, but not well understood. Other less complicated andbetter understood evaporation techniques emerged and research in the sputtering technique halted.It was not until the late 50’s that the demand for higher quality thin films of a variety of materialsagain made sputtering an area of interest for scientists [28].
Sputtering theoryThe sputtering method involves physical vaporization of atoms from a target surface by bombardingit with ionized gas atoms in vacuum. In this project Ar gas is used. The ions are accelerated in anelectric field and hit the target surface, breaking loose atoms from the target which are then free tomove towards, and hit, the substrate.
In the process of breaking o� atoms from the target, the atoms gain momentum. The momentumtransfer theory is based on the work done by Guntherschulze in the 20’s and 30’s and Wehner et al.in the 50’s and 60’s [30] and is the basis for understanding the sputtering process.
In his book Mattox [30] lists the e�ect explained by the momentum transfer theory. Shortenedto some degree, Mattox states that: sputter yield ("ratio of atoms sputtered to the numbers of highincident particles") depends on the energy, mass and angle-of-incident of the bombarding particle.Below a certain energy sputtering does not occurand at high energies the sputtering yield is lowbecause the ions lose much of their energy far below the surface of the target. When sputteringpolycrystalline material, some crystallographic planes sputters faster than others.
Di�erent methods of sputtering exist, the most common being dc (direct current) sputtering andthe rf (radio frequency) sputtering. In the dc system the front side of a cathode is covered with thetarget material to be deposited and the substrate is placed on an anode. The sputtering chamberis filled with sputtering gas, and the glow discharge is maintained by the application of dc voltagebetween the electrodes. The Ar+ ions generated are accelerated at the cathode and sputters thetarget [31]. To maintain the glow discharge a metal target is needed to avoid a immediate buildupof a surface charge on the front side of the target.
Substituting the metal with an insulator would not work with the dc sputtering system. In thecase of sputtering insulating material, the rf sputtering system is used where the glow discharge ismaintained by applying an rf voltage to the insulating target.
The system used for this project is the rf magnetron sputtering system where a cylindrical mag-netic field is applied to the target. The magnetic field traps electrons in the vicinity of the sputteringgas, increasing the collision rate between electrons and the sputtering gas, increasing the plasmadensity and sputtering yield [31].
Sputtering of highly doped siliconGroups first reporting on sputtered doped Si films used impurity chips attached to Si wafers, con-trolling the doping concentration by varying the area ratio of the impurity chip to Si wafer [32, 33].
6
CHAPTER 3. TECHNIQUES 3.3. POST DEPOSITION TECHNIQUES
As highly pre-doped Si wafers became available the preferred technique shifted from co-sputteringto using the highly doped target, giving better control of the majority carrier concentration in thedeposited film.
This technique was used by Fenske and Gorka [34] to prepare phosphorus doped Si films bysputtering of 1-2 m⌦ cm P-doped target producing Si films with carrier concentrations of 7 � 9 ⇥1019 cm�3 and mobilities of 50 cm2/Vs.
Experimental results on films produced by sputtering of highly B and P doped Si-targets, an-nealed by RTA at 1100�C, shows a carrier concentration of 1.6⇥ 1019 cm�3 in the B-doped samples,approximately one order of magnitude higher than the P-doped samples [17]. Wang et al. [17] shownthe doping e�ciency of B and P atoms to be 28% and 0.62%, respectively.
3.3 Post Deposition Techniques3.3.1 Rapid Thermal ProcessingRapid thermal processing (RTP) is a widely used annealing technique in semiconductor technologyever since IBM came up with the technique in the late 60’s [35]. With RTP the wafer is heated quicklyat atmospheric or low pressure [36]. Heat is provided by halogen lamps placed in close proximityto the wafer. The lamps heat the wafer to 1100-1200�C on a timescale of several seconds or less.The wafer is placed in a graphite holder, which in turn is placed on four quartz spikes, keepingthe contact area between the graphite wafer-holder and the surrounding equipment to a minimum.A temperature measurement system is placed in a control loop, measuring the wafer temperaturedirectly.
RTP has many applications including activation of dopants, densification of the deposited filmalong with solid phase crystallization (SPC). Traditionally furnace annealing (3.3.2) was used forSPC but RTA has show to be a well suited substitute and even though higher temperatures areneeded in this process [17, 37] the total thermal budget are orders of magnitude smaller because ofthe short annealing duration.
3.3.2 Furnace AnnealingFurnace annealing is a common technique in semiconductor device fabrication. This type of heattreatment produces much of the same result as with RTP such as dopant activation, film desificationand solid phase crystallization, among other.
Annealing duration in furnaces are usually long such as in Jun et al. [10] paper where annealingtimes of 20 hours are used. This is usually not problematic, but at high temperatures (>1000�C)dopants may start to di�use and a well defined p-n junctions may not be achieved.
A gas flow system is usually part of a furnace annealing set-up enabling oxidation steps, or theopposite; flushing the chamber with gases to suppress oxygen to ensure that oxidation does not takeplace.
3.4 Characterization Techniques3.4.1 4-point probeThe four-point probe is widely used for resistivity measurement in semiconductors. It is an absolutemeasurement with no need of calibrated standards and is therefore often used to provide standardsfor other resistivity measurements [38].
Weimer [39] first proposed the four-point probe in 1916 as a tool to measure the earth resistivity,and in 1954 Valdes [40] adopted the technique for resistivity measurements in semiconductor wafers.
The spacing between the four probes are usually identical as shown in figure 3.1, and it is thisassumption that Schroder [38] uses to derive the following expression for measuring resistivity
⇢ = 4.532U
It (3.1)
where ⇢, U , I and t is resistivity, voltage, current and wafer thickness, respectively. The correctionfactor 4.531 is valid for t s/2 where s is the probe spacing; the film thickness needs to be less thanthe probe spacing, which in this work always applies with film thickness 300 nm.
7
CHAPTER 3. TECHNIQUES 3.4. CHARACTERIZATION TECHNIQUES
s s s
⇠VI
I
Figure 3.1: Four-point probe schematics
The measurement is done by applying a current I over the outside probes, then measuring thevoltage over the two interior probes.
3.4.2 Hall e�ect measurementHall e�ect measurements plays a big part in semiconductor physics because of its ability to measureresistivity, carrier concentration and mobility. All these three factors are discussed in detail laterin this section, but they are all, to some degree, found by utilizing the Hall e�ect.
B
I
VH
1
2
4
3qv
FE
Figure 3.2: Hall e�ect schematics with a I13 � V24 van der Pauw setup on a n-type semiconductor.
The hall e�ect was discovered in 1879 by Hall when he was investigating the forces acting on aconductor carrying a current in a magnetic field [41]
Hall found that a magnetic field applied to a conductor perpendicular to the currentflow direction produces an electric field perpendicular to the magnetic field and the cur-rent
In semiconductor physics this translates to a magnetic field perpendicular to the charge carrierdirection of travel. The force on the carriers is given by
F = q(E + v ⇥ B) (3.2)
and as shown in figure 3.2 this results in a accumulation of carriers, setting up an electric fieldwith magnitude proportional to the magnetic field as well as the carrier velocity by
Ey = Bzvx (3.3)
8
CHAPTER 3. TECHNIQUES 3.4. CHARACTERIZATION TECHNIQUES
often referred to as EH , Hall field. From this induced field it is possible to derive the type ofcharge carrier (n- or p-type), carrier concentration and mobility. The electrical current density maybe expressed as
J = �nqvx (3.4)
where n is the number of charge carriers, q the electron charge and vx the carrier velocity as usedin equation 3.2. The current is then given as
Ix = Jx!t (3.5)
where !t is the cross-sectional area of the semiconductor. The Hall voltage VH (V24 ref fig 3.2),measurable as the the potential di�erence across the sample, is related to the Hall field by
VH = �!Z
0
Ey dy = �Ey` (3.6)
Combining equation 3.3, 3.5 and 3.6
VH = �✓
1
nq
◆IxBz
t(3.7)
where the first term is called the Hall coe�cient
RH =1
nq(3.8)
If this value is negative, the charge carriers are negative, and if positive the charge carriers arepositive. In practice, the polarity of VH determines the sign of the charge carriers [42].
Knowing the Hall voltage at a specific current and magnetic field yields a value for the carrierconcentration through
n =1
qRH=
IxBz
qtVH(3.9)
and as all the quantities on the right hand side of this equation can be measured carrier concentra-tions are easily derived by this method.
3.4.3 ResistivityResistivity in thin films with uniform thicknesses are often referred to as sheet resistivity. If ameasurement of resistance R is made, the sheet resistivity ⇢ can be calculated by [43]
⇢ =R!t
L=
V13/IxL/!t
(3.10)
where L is the length between contacts (1-3 ref fig 3.2).Resistivity is in itself an important property as it contributes to the device series resistance,
capacitance, threshold voltage among others [38].The resistivity depends on the free electron and hole densities n and p and their respective mo-
bilities µn and µp through the relationship
⇢ =1
q (nµn + pµp)(3.11)
9
CHAPTER 3. TECHNIQUES 3.4. CHARACTERIZATION TECHNIQUES
In extrinsic materials where the majority carrier density is many order of magnitude higher thanthe minority carrier density, equation 3.11 can be written
⇢ =1
qnµnfor n-type material (3.12)
⇢ =1
qpµpfor p-type material (3.13)
These equations are used to calculate the mobility in the semiconductor.
3.4.4 Van der PauwThis setup is commonly used in Hall e�ect measurements to measure resistivity, carrier concentra-tion, and carrier mobility (often referred to as Hall mobility). The sample is placed on an isolatingsample holder and set up as shown in figure 3.3 by soldering conducting wires to each corner of asample. Arbitrary shapes may also be used as explained by Schroder [38], but as the calculationsare a lot easier with symmetrical shapes, square samples are most common when using the Van derPauw method.
I V
2 3
1 4
Figure 3.3: Van der Pauw schematics, depicting a I12-V34 set-up.
As shown in figure 3.3 the four contacts are given numbers 1,2,3 and 4. Current is sent betweentwo of the contact points and I12 equals current sent between contact point 1 and 2. The measuredvoltage is denoted in the same way over the contacts not used to pass current, such as V34. Theresulting resistance in this set-up, governed by Ohms law, is denoted as R12�34.
By measuring all contacts, two characteristic resistances RA and RB are found, an average of thefour possible geometric combinations
RA =1
4(R12�34 +R21�34 +R34�12 +R43�12)
RB =1
4(R23�41 +R32�41 +R41�23 +R14�23)
From Schroder [38], equation (1.25) describes the resistivity:
⇢ =⇡
ln(2)tRA +RB
2F (3.14)
where t and F are film thickness and a function of the ratio Rr = RA/RB, respectively. For a sym-metrical sample, such as a perfect square, Rr = 1 and F = 1 allowing equation (3.14) to be simplifiedto
⇢ = 4.532tRA (3.15)
and is the same as used in four point measurements described in section 3.4.1.To measure carrier concentration, mobility and type(n- or p-type) a magnetic field is applied
perpendicular to the sample. Calculations on mobility, carrier concentration, hall coe�cient andtype of charge carrier are described in section 3.4.2.
10
CHAPTER 3. TECHNIQUES 3.4. CHARACTERIZATION TECHNIQUES
3.4.5 Atomic Force MicroscopyAtomic force microscopy was first described in 1986 by Binnig et al. [44] where they state that
...this level of sensitivity clearly penetrates the regime of inter atomic forces betweensingle atoms and opens the door to a variety of applications. The atomic force microscope(AFM) is a new tool designed to exploit this level of sensitivity.
The method enabled scientists to view surface structures in a new and more detailed way. After1986 there has been a lot of improvements on the AFM and today we can achieve high resolution,nano-scale pictures.
Sample
Laser
Cantilever
Photo detector
Figure 3.4: Atomic Force Microscope schematics
To achieve this super detailed viewof surfaces, the weak forces between thesample surface and a very sensitive can-tilever is utilized.
In tapping mode the tip, or probe,is driven to oscillate at or near its res-onance frequency. As the tip closes inon the surface the forces between thetip and the surface causes the amplitudeof the oscillating tip to change. Thischange, detected by the position sensi-tive photo detector, is sent to the piezo-ceramic servo which raises or lowers theprobe to maintain a certain height overthe surface. The movement data of theservo is used to produce a three dimensional topographical mapping of the surface.
3.4.6 X-ray di�ractionX-rays was first discovered in 1895 and in 1912 Max von Laue found that crystalline substances act asthree dimensional gratings for x-rays with wavelengths similar to the crystal lattice parameters [45].
Figure 3.5: Energy transitions resultingin the characteristic CuK ’s. Source: https:
//upload.wikimedia.org/wikipedia/commons/thumb/
d/d8/Copper_K_Rontgen.png/220px-Copper_K_
Rontgen.png
This enabled scientists to probe crystalstructures in a new and revolutionaryway and has become a common tech-nique for the study of crystal structuresand atomic spacing.
The x-rays are produced by bombard-ing a metal with a beam of electrons, ion-izing 1s electrons from the target atom.As electrons from the 2p and 3p levelsdrop down to fill these vacancies, x-raysare released and form characteristic x-ray spectra, the most common being K↵
(2p ! 1s) and K�(3p ! 1s)[46].Copper is a common metal in pro-
ducing x-rays for x-ray di�raction (XRD)characterization and the characteristicwavelength used in XRD is
CuK↵ = 1.5418 Å
K� and other less intense wave-lengths are removed by a monochroma-tor. K↵ consists of K↵1 and K↵2 whereK↵1 has a slightly shorter wavelengthand twice the intensity as K↵2 [45]. IfK↵1 and K↵2 wavelengths are far enough apart in energy to produce a clear reflection it is possibleto filter out K↵2 as well.
11
CHAPTER 3. TECHNIQUES 3.4. CHARACTERIZATION TECHNIQUES
Bragg’s LawBy using Bragg’s law it is possible to determine the crystal lattice dimensions and which crystalplanes are present in the sample. A vast number of reflection patterns are documented and the useof databases are needed when deciding which planes are present in a sample.
Bragg’s law is a universal law concerning the behavior of waves in crystals and other solids.Incoming x-rays with a certain wavelength � penetrate the first atomic layer and interact and scattero� atoms in the second layer as shown in figure 3.6.
Figure 3.6: Bragg reflection schematics with two incoming waves at di�raction. Source:http:
//hyperphysics.phy-astr.gsu.edu/hbase/quantum/imgqua/bragglaw.gif
The distance d denotes the spacing between the first and second layer and the distance 2d sin ✓is the total extra distance traveled by the wave reflected o� the second layer compared to the wavereflected o� the first layer, and if this equals n�, n being an integer, the two waves experience con-structive interference. A peak in intensity is observed and Bragg’s law, n� = 2d sin ✓ is fulfilled.Every part of this equation, except for d, can be found through an experimental setup and solvingfor d produces
d =n�
2 sin ✓(3.16)
Figure 3.7: X-ray di�raction schemat-ics. Source: http://chemwiki.ucdavis.edu/
@api/deki/files/232/=xrd.png?revision=
1&size=bestfit&width=323&height=237
Because of the n� dependency, incoming waveswith � close to the lattice parameter is desirablein solid state characterization. As a result us-ing x-rays produced by copper with a characteris-tic wavelength CuK↵ = 1.5418 Å (see 3.4.6) whencharacterizing silicon with a crystal lattice con-stant of 5.431 Å (in diamond FCC crystal) workswell as these values are within the same order ofmagnitude.
As shown in figure 3.7 an x-ray source sends abeam directly at the sample and a detector is set upto collect the reflected rays. In a typical ✓/2✓-scanthe sample holder rotates at an angle ✓ while thedetector rotates at an angle of 2✓ over a scan inter-val. Intensities are logged at each measurementpoint and yields a plot such as figure 5.11.
3.4.7 Secondary Ion Mass SpectroscopySecondary Ion Mass Spectroscopy (SIMS) is one of the most powerful analytical techniques for semi-conductor characterization. After it was developed in early 60’s it has become a leading characteri-zation method capable of detecting all elements as well as isotopes and molecular species.
The basis of SIMS is the removal of material from the sample surface by sputtering and subse-quently analyzing the ejected material. Only about 1% of the total ejected material are positively
12
CHAPTER 3. TECHNIQUES 3.4. CHARACTERIZATION TECHNIQUES
Figure 3.8: Secondary ion mass spectroscopy schematics.Schroder [47]
or negatively charged and SIMStherefore uses the mass/chargeratio of the ions to analyzethe sample. This is sometimesproblematic as various com-plex molecules form during thesputtering process which maybe confused for natural ele-ments.
Depth profiling is anotherstrength of SIMS. Plotting theintensity of a selected massversus sputtering time andconverting this to density ver-sus depth is possible whenthe primary ion beam current,sputter yield, ionization e�-ciency, atomic fraction to beanalyzed and an instrumentalfactor is known. Some of thesefactors may be hard to find and experimental standards are commonly used Schroder [47].
13
4E�����������
The experimental part of this project was divided into three phases, one initial and two main phases.The final phase builds on the results found in the previous two phases.
General proceduresFused silica substrates were used in all depositions. A 3" quartz wafer was laser cut into 1x1 cm2
pieces and cleaned in acetone, isopropanol and water for 2, 10 and 5 minutes, respectively, dried o�by N2 gas and stored in an airtight plastic bag for later use.
4.1 Phase I - Initial investigationsPhase one was the initial phase where the goal was to find deposition parameters which could pro-duce the desired film quality. A fairly unsystematic approach was used producing a total of five pairsof samples.
4.1.1 DepositionElectron Beam Physical Vapor Deposition300 n of silicon was deposited by electron beam physical vapor deposition (EB-PVD) using an Evo-Vac chamber from Ångstrom Engineering. The target material was highly phosphorus doped ([P]⇠5 ⇥ 1019 cm�3) silicon. No substrate heating was applied during the first deposition; the seconddeposition was done with a substrate temperature of 300�C. Initial chamber pressure at deposi-tion initiation was measured to be 2.0 ⇥ 10�6 Torr in both cases, decreasing during deposition to8.5⇥ 10�7 Torr and 9.2⇥ 10�7 Torr in the low and high temperature run, respectively.
SputteringThree sets of samples were sputter deposited at di�erent temperatures; room temperature, 300�Cand 600�C in a Semicore Triaxis DC/RF Magnetron Sputter with each set containing two samplesto facilitate secondary ion mass spectrometry (SIMS) measurements. A phosphorus-doped targetmaterial was used and all depositions were initiated at a chamber pressure below 2.0 ⇥ 10�6 Torr.An argon gas flow of 50 SCCM was introduced into the chamber and plasma was ignited resultingin a deposition pressure of 7.1 mTorr. A ten minute pre-sputtering period was included to clean thetarget surface.
4.1.2 CharacterizationFour point measurementThe sheet resistance of all samples was measured using a Jandel KM3-AR four point probe unit inhigh resistivity mode and auto ranging was used in order to find applicable current values.
14
CHAPTER 4. EXPERIMENTAL 4.2. PHASE II - HEAT TREATMENT
X-ray di�ractionOne sample from phase I was characterized by X-ray di�raction to extract structural information. ABruker AXS D8 Discover unit was used in ✓/2✓-scan mode from 2✓ = 20� to 2✓ = 90� with incrementsof 0.01� at a scan rate of 0.1 seconds per increment.
4.2 Phase II - Heat treatmentThe idea of phase two was to broaden the angle of view and systematically test the techniques foundto work in literature. Although a lot of the literature was based on plasma enhanced chemicalvapor deposition (PECVD), annealing temperature is usually found to be the determining factor incrystallization of a-Si and this parameter was varied as shown in table 4.1. Annealing temperaturesinvestigated in the literature range from 600�C in a traditional furnace setup [10] to 1100�C in atypical rapid thermal processing (RTP) setup [17] and this was the area of investigations for thepresent study.
High temperature EB-PVD was discarded at this time as the time budget was to great, requiringan extensive cool-down period spanning hours and thus only low temperature depositions were madewith EB-PVD.
Sample ID Technique Rate/Power Deposition Temp RTP-tempERT1 E-beam 1Å/s RT noneERT2 E-beam 1Å/s RT 650�,1minERT3 E-beam 1Å/s RT 650�,2minERT4 E-beam 1Å/s RT 750�,1minERT5 E-beam 1Å/s RT 750�,2minERT6 E-beam 1Å/s RT 850�,1minERT7 E-beam 1Å/s RT 850�,2minERT8 E-beam 1Å/s RT 950�,1minERT9 E-beam 1Å/s RT 950�,2minERT10 E-beam 1Å/s RT 1050�,2minSRT1 Sputtering 50W RT noneSRT2 Sputtering 50W RT 650�,1minSRT3 Sputtering 50W RT 650�,2minSRT4 Sputtering 50W RT 750�,1minSRT5 Sputtering 50W RT 750�,2minSRT6 Sputtering 50W RT 850�,1minSRT7 Sputtering 50W RT 850�,2minSRT8 Sputtering 50W RT 950�,1minSRT9 Sputtering 50W RT 950�,2minSRT10 Sputtering 50W RT 1050�,2minSHT1 Sputtering 50W 600� noneSHT2 Sputtering 50W 600C 650�,1minSHT3 Sputtering 50W 600C 650�,2minSHT4 Sputtering 50W 600C 750�,1minSHT5 Sputtering 50W 600C 750�,2minSHT6 Sputtering 50W 600C 850�,1minSHT7 Sputtering 50W 600C 850�,2minSHT8 Sputtering 50W 600C 950�,1minSHT9 Sputtering 50W 600C 950�,2minSHT10 Sputtering 50W 600C 1050�,2min
Table 4.1: Second phase experimental set-up
15
CHAPTER 4. EXPERIMENTAL 4.2. PHASE II - HEAT TREATMENT
4.2.1 DepositionElectron Beam Physical Vapor Deposition300 nm of Si was deposited using the same recipe as described in section 4.1.1, using an EvoVacchamber from Ångstrom Engineering. The target material was highly phosphorus doped silicon andno substrate heating was applied. Chamber pressure was 2.0 ⇥ 10�6 Torr at initiation, decreasingto ⇡ 8.0⇥ 10�7 Torr during deposition. Substrate rotation was set to 7 rpm.
SputteringTwo sets of samples were deposited, one at room temperature and one at 600�C in a SemicoreTriaxis DC/RF Magnetron Sputter using a highly phosphorus doped target. Pressure was below2.0 ⇥ 10�6 Torr prior to argon gas introduction and the deposition pressure was 7.1 mTorr. Argongas flow was set to 50 SCCM. A 10 minute pre-sputtering period was included to clean the target.
4.2.2 Post deposition treatmentAll samples were annealed by RTP except for one control sample. The annealing temperature werebetween the 650�C to 1050�C range. Initially 1050�C was not part of the plan but was included at alater stage as the lower temperatures did not produce desirable results. All samples were annealedat 1 and 2 minutes at each temperature, except for the 1050�C sample which was only treated for 2minutes. RTP was performed in nitrogen atmosphere in an AnnealSys-Micro system.
4.2.3 CharacterizationFour point measurementsInitial characterization was done by four point probe measurements on all samples at three separatepoints on the sample to see if the resistivity was within the range required to perform Hall e�ectmeasurements. A Jandel KM3-AR unit was used in high resistivity mode and auto ranging wasutilized in order to pinpoint usable current values.
X-ray di�ractionAll samples were characterized by X-ray di�raction. A Bruker AXS D8 Discover unit was used in✓/2✓-scan mode from 2✓ = 20� to 2✓ = 90� with increments of 0.01� at a scan rate of 0.1 seconds perincrement.
Atomic force microscopyAll samples were investigated using a Dimensions 3000 atomic force microscope (AFM) in tappingmode.
16
CHAPTER 4. EXPERIMENTAL 4.2. PHASE II - HEAT TREATMENT
Sam
ple
IDRI
Epo
wer
/tim
eTe
chni
que
Pow
erD
ep.te
mp
RTP-
tem
pRT
P-tim
eFu
rnac
e-te
mp
Furn
ace-
time
Char
acte
riza
tion
SRT1
10
/0Sp
utte
r20
0WRT
1100
�3
m1.
4-po
int,
2.XR
D,3
.AFM
SRT1
20
/0Sp
utte
r20
0WRT
1100
�3
m1.
4-po
int,
2.AF
M,3
.Hal
l,4.
SIM
SSR
T13
0/0
Sput
ter
200W
RT11
00�
30m
1.4-
poin
t,2.
AFM
SRT1
40
/0Sp
utte
r20
0WRT
1100
�30
m1.
4-po
int,
2.AF
M,3
.Hal
l,4.
SIM
SSR
T15
200W
/5m
inSp
utte
r20
0WRT
1100
�3
m1.
4-po
int,
2.AF
M,3
.Hal
l,4.
SIM
SSR
T16
200W
/5m
inSp
utte
r20
0WRT
1100
�3
m1.
4-po
int,
2.AF
MSR
T17
200W
/5m
inSp
utte
r20
0WRT
1100
�30
m1.
4-po
int,
2.AF
MSR
T18
200W
/5m
inSp
utte
r20
0WRT
1100
�30
m1.
4-po
int,
2.AF
M,3
.Hal
l,4.
SIM
SSM
LT1
0/0
Sput
ter
200W
200�
1100
�3
m1.
4-po
int,
2.AF
M,3
.Hal
l,4.
SIM
SSM
LT2
0/0
Sput
ter
200W
200�
1100
�3
m1.
4-po
int,
2.AF
M,3
.Hal
l,4.
SIM
SSM
LT3
0/0
Sput
ter
200W
200�
1100
�30
m1.
4-po
int,
2.AF
MSM
LT4
0/0
Sput
ter
200W
200�
1100
�30
m1.
4-po
int,
2.AF
M,3
.Hal
l,4.
SIM
SSM
LT5
200W
/5m
inSp
utte
r20
0W20
0�11
00�
3m
1.4-
poin
t,2.
AFM
SMLT
620
0W/5
min
Sput
ter
200W
200�
1100
�3
m1.
4-po
int,
2.AF
MSM
LT7
200W
/5m
inSp
utte
r20
0W20
0�11
00�
30m
1.4-
poin
t,2.
AFM
,3.X
RD,
4.H
all,
5.SI
MS
SMLT
820
0W/5
min
Sput
ter
200W
200�
1100
�30
m1.
4-po
int,
2.AF
M,3
.Hal
l,4.
SIM
S
Tabl
e4.
2:Th
ird
phas
eex
peri
men
tals
et-u
p
17
CHAPTER 4. EXPERIMENTAL 4.3. PHASE III - SUBSTRATE PREPARATION
4.3 Phase III - Substrate preparationAs the EB-PVD did not produce desirable results in the first two phases the process was discardedand only sputtering deposition was performed at this stage. A new pre-deposition treatment wasalso included after a discussion with the supervisors where the surface topography became an areaof interest.
As the amorphous films deposited was thought to be dependent on nuclei formation and subse-quent crystallization, it was proposed that an reactive ion etch, resulting in a topographical changeof the quartz substrate, would increase the control over the crystal growth.
As shown in table 4.2 samples were produced in pairs, both for statistical interpretation andfacilitating SIMS measurements (destructive method).
4.3.1 Pre Deposition TreatmentThree samples were etched in argon plasma by an Advanced Vacuum Vision 320 MK II reactive ionetcher by Mikael Sjödin1. The setup as shown in table 4.3. The resulting topography was studied us-ing a Dimension 3000 atomic force microscope and the AFM images were produced in Gwyddion[48].Sample 2 produced the most favorable topography; as shown in figure 4.2 the result was 1-1.5 nmtall hillocks with a lateral width in the 20-50 nm range. Half of the samples were consequentlyetched at 200 W for 5 minutes prior to deposition as shown in table 4.2, while the other half of thesamples were left untreated for reference.
Sample Power TimeID [W] [min]1 100 52 200 53 300 5
Table 4.3: Reactive ion etch experimental setup.
4.3.2 DepositionAs discussed in chapter 5 the room temperature sputtered sample was the only one to produce afilm resistivity low enough to yield sensible data and the high temperature sputter depositions wastherefore discarded.
Sputtering deposition power was increased to 200 W since reports by Wang et al. [17] and Junet al. [10] indicate that significantly higher sputtering powers in addition to keeping a substrate biasis required to produce doped crystalline silicon. Deposition duration was decreased to 30 minutes,aiming for a film thickness of 300 nm.
Two separate runs were initiated at room temperature and 200�C, respecively, with an initialchamber pressure of 2.0 ⇥ 10�6 Torr. After introducing an argon flow of 50 SCCM and subsequentplasma ignition, deposition chamber pressure was measured to be 7.0 mTorr. A pre-sputtering pe-riod of ten minutes was used.
4.3.3 Post deposition treatmentAnnealing temperatures was 1100�C but two di�erent methods were used to see how di�erent dura-tion a�ected the samples. RTP annealing duration was 3 minutes while furnace annealing durationwas 30 min. A AnnealSys-Micro system was used for RTP while a tube furnace was used for furnaceannealing. Both RTP and furnace annealing was performed in nitrogen-atmosphere.
4.3.4 CharacterizationAs available time was beginning to shrink at this point in the project only a selected number ofsamples were fully characterized.
18
CHAPTER 4. EXPERIMENTAL 4.3. PHASE III - SUBSTRATE PREPARATION
(a) Quartz reference sample, untreated. (b) Quartz sample 1 RIE treated at 100W for 5 min
(c) Quartz sample 2 RIE treated at 200W for 5 min (d) Quartz sample 3 RIE treated at 300W for 5 min
Figure 4.1: AFM results following reactive ion etch of quartz samples
Figure 4.2: Profile of etched sample 2
Four point probeAll samples were characterized by a four point probe in the same manner as described in section4.2.3.
19
CHAPTER 4. EXPERIMENTAL 4.3. PHASE III - SUBSTRATE PREPARATION
Hall E�ect MeasurementIt was determined that the resistivity range was within the limits of what was required for Hallmeasurements and Hall measurements were performed on 9 samples as indicated in table 4.2. Halle�ect measurements were performed using a LakeShore EM4 HGA system with a Van der Pauwsetup at room temperature using indium as contacts in the corners of the sample.
IV curve measurement were set up with start/end currents at ±10 µA and ±100 µA with cur-rent steps at 2 µA and 20 µA depending on film resistivity.
For voltage tracking measurement the magnetic field was set to vary from -10 kG to 10 kG2
using an excitation current of 1.0 mA.The variable field measurement was performed at a ±10 kG magnetic field. An excitation
current in the 100 µA to 1 mA range was applied, depending on the film resistivity. The resistivityat zero field was used to calculate Hall mobility.
Atomic force microscopyAll samples were investigated using a Dimension 3000 AFM in tapping mode.
X-ray di�ractionTwo samples were chosen for XRD, SRT11 and SMLT7 which were low and high resistivity perform-ers , respectively. The setup was identical to the ✓/2✓-scan mode described in section 4.2.3.
SIMSSIMS characterization was done on all Hall e�ect measured samples and performed by AlexanderAzarov 3 at MiNaLab. Both 30Si intensity and 31P concentration was measured to find phosphorusdoping concentration as well as film thickness.
2magnetic field given in gauss. 1 G corresponds to 1⇥ 10�4 [email protected]
20
5R������
5.1 Phase I & IIThe first two phases of the project did not produce crystalline Si except for perhaps sample SRT10(sputtered at room temperature, annealed at 1050�C for 2 min) which was the only one with aresistivity low enough to be measured by the four point probe. Results are given in table 5.1 andshows a high resistivity in this sample which is barely measurable with the four point probe unitused. The sample deposited at 600�C and annealed at 1050�C did not exhibit these characteristics.
Sample Current Voltage Film Thickness ResistivityID [nA] [mV] [nm] [⌦ cm]
SRT10 10 90 300 1.2E+4
Table 5.1: Four point measurement results of sample SRT10
X-ray di�raction corroborates the electrical results and as seen in figure 5.1 there are little orno indication of crystal planes in the samples, except in sample SRT10 where a very slight peak isobserved close to 2✓ = 28.443� indicating Si(111)-planes [49].
Figure 5.1: X-ray di�raction patterns of sample ERT1, ERT10, SRT1, SRT10, SRT3, SRT4 and SRT2.
21
CHAPTER 5. RESULTS 5.2. PHASE III
5.2 Phase III5.2.1 Four Point MeasurementsPhase three was a highly successful batch of samples and Si crystallized in allinvestigated samples.As stated in section 4.3.4 initial characterization was done by four point measurements and theresults are shown in figure 5.2. The lowest resistivity was measured in samples subjected to reactiveion etch, sputter deposition at 200�C and 30 min furnace anneal at 1100�C (SMLT7 and SMLT8).Assuming a film thickness of 300 nm, the resistivities of these samples were, according to equation3.1 ⇢ = 2.8⇥10�2 ⌦ cm. The highest resistivity sample pair SRT15/16 had a resistivity ⇢ ⇠ 0.5 ⌦ cm.
Figure 5.2: Phase III four point measurements results at room temperature. Resistivities calculatedassuming 300 nm film thickness. The results are an average of sample pair measurements.
5.2.2 Hall E�ect MeasurementsSince the four point measurements showed relatively low resistivity, Hall e�ect measurements wasfeasible. Sample pair SMLT7/8 exhibited the lowest resistivity at ⇠ 2.8 ⇥ 10�2 ⌦ cm, as well as thehighest carrier density and mobility at ⇠ 2.0⇥1019 cm�3 and ⇠ 12.0 cm2/ Vs, respectively, as shownin figure 5.3, 5.4 and 5.5.
Even the higher resistivity samples exhibited decent characteristics. The lowest carrier concen-tration and mobility was found in sample SRT15 (etched and room temperature sputtered samplewith subsequent RTP treatment) with a resistivity ⇠ 0.42 ⌦ cm, a carrier density and mobility of⇠ 4.9⇥ 1018 cm�3 and ⇠ 2.9 cm2/ Vs, respectively.
22
CHAPTER 5. RESULTS 5.2. PHASE III
Figure 5.3: Phase III Hall e�ect resistivity measurement results. The resistivity calculations usedactual thicknesses found by SIMS measurements.
Figure 5.4: Phase III Hall e�ect carrier density measurement results.
23
CHAPTER 5. RESULTS 5.2. PHASE III
Figure 5.5: Phase III Hall e�ect mobility measurement results.
5.2.3 Secondary Ion Mass SpectrometryTo determine the actual thickness and phosphorous concentration of the samples, SIMS was used.The results are shown in figure 5.6 and 5.7 and show a similar phosphorus concentration in allsamples. Film thicknesses varies from 260 - 280 nm in the samples sputtered at 200�C and 280 -300 nm in the samples sputtered at room temperature.
In figure 5.8 phosphorus concentration in sample SMLT7 is plotted against depth showing asimilar phosphorus concentration throughout the film. Film thickness can be derived from this.
24
CHAPTER 5. RESULTS 5.2. PHASE III
Figure 5.6: Results from thickness measurements by SIMS characterization on Hall measured sam-ples.
Figure 5.7: Results of SIMS characterization on Hall e�ect measured samples. Phosphorous con-centration given is an average of the measured concentrations throughout the film.
25
CHAPTER 5. RESULTS 5.2. PHASE III
Figure 5.8: Phosphorus concentration vs depth in sample SMLT7. The first five points left outbecause of inaccuracy of SIMS measurements near the surface.
26
CHAPTER 5. RESULTS 5.2. PHASE III
5.2.4 Atomic force microscopyAll samples were investigated with AFM to compare the pre-deposition treated samples with theuntreated ones, both in surface height root mean square (RMS) and topography. As figure 5.9 shows,the etched and unetched di�ers significantly; the untreated sample with the lowest RMS have anRMS of almost 1.5 times that of the highest among the treated samples.
AFM images substantiate this as figure 5.10 shows. Figure 5.10e and 5.10f are untreated sampleshaving a visibly higher roughness than the treated samples shown in figure 5.10a to 5.10d.
Figure 5.9: Root Mean Square statistical measurements of the surface height of phase III samplesby AFM characterization.
27
CHAPTER 5. RESULTS 5.2. PHASE III
(a) SMLT8 - RIE (b) SRT18 - RIE
(c) SMLT6 - RIE (d) SRT16 - RIE
(e) SMLT4 - no RIE (f) SRT11 - no RIE
Figure 5.10: Atomic force microscope images of phase III samples.
28
CHAPTER 5. RESULTS 5.2. PHASE III
5.2.5 X-ray di�ractionOnly two samples were measured by XRD at this phase, as described in section 4.3.4. In figure 5.11and 5.12 reflection peaks are observed close to 2✓ = 28.443� and 2✓ = 47.303� indicating the presenceof Si (111) and (220) planes, respectively[49].
In sample SMLT7, a peak is observed close to 2✓ = 32.965� indicating In (101) plane [49]. Thisis due to the indium contacts used in Hall e�ect measurements of this sample. No such peak isobserved in figure 5.12 as no Hall e�ect measurements were made on sample SRT11.
In both XRD patterns there are clear peaks close to 2✓ = 44.74�, 2✓ = 65.135� and 2✓ = 78.229�,indicating Al (200) (220) (311) planes [49], respectively, and these originate from the aluminumsample holder.
Comparing the two samples shows higher intensities and more defined peaks at both the Si(111) reflection and the Si (220) reflection in sample SMLT7, an indication of a higher degree ofcrystallinity in this sample.
Figure 5.11: X-ray di�raction pattern of sample SMLT7. This sample was measured by Hall e�ectmeasurements where indium contacts were used.
Figure 5.12: X-ray di�raction pattern of sample SRT11.
29
6D���������
When looking at why the first two phases did not yield any crystalline silicon, a couple of parametersstands out; in particular deposition technique as well as the lack of post deposition treatment.
As shown in the results section none of the samples deposited by electron beam physical vapordeposition showed any signs of crystallinity, indicating an amorphous film even after post depositionannealing. The reasons for this is unknown; other groups such as Jamil et al. [50] have succeeded indepositing polycrystalline silicon, but further investigation of this was discarded as sputtering waschosen as the favored method.
Sputtered samples investigated from phase I and II were almost exclusively showing signs ofamorphous films, except for one sample, sputtered at room temperature and RTP treated at 1050�Cfor 2 minutes. This sample did show slight signs of crystallinity as well as a resistivity within themeasurable range by four point measurement. This sample (referred to as SRT10 in text) indicatedthat high sputtering temperatures was a potential unfavorable deposition condition and this laidthe foundation for further investigations.
From literature it was found that nuclei formation plays a role in post deposition crystallization[14] and a pre deposition reactive ion etch was therefore included in the experimental setup to tryand manipulate the formation of nuclei. The etch was designed to change the topography to matchtypical crystal grain sizes of about 50 nm. As seen in figure 4.2 this was achieved and by studyingthe resulting topography, such as comparing figure 4.1c and 5.10a, a similarity in topography isobserved, indicating a certain control of crystal growth, although the sample in 5.10a seem to havea slightly rougher surface with broader bumps.
The RIE does not seem to enhance the electrical characteristics of the film. As can be seen infigure 5.3 it seems like the un-etched samples performs better in the room temperature sputteredones; in the 200�C sputtered ones it is opposite as the RIE samples perform slightly better, hencethe RIE process does not seem to enhance electrical properties.
On the other hand there is clear indications that electrical properties largely depend on sputter-ing settings, temperature and power, and annealing duration as figures 5.3, 5.4 and 5.5 clearly shows.These figures show that furnace annealing produces higher quality films than the RTP treatmentdoes, indicating that the increase in duration from 3 to 30 minutes at 1100�C makes a big di�erence;possibly both in crystallizing the a-Si as well as activating the phosphorus dopants present in thedeposited films.
The increase in power from phase I and II to phase III changes the results radically and it seemsthat 200 W sputtering power is far better suited for post deposition solid phase crystallization thansputter deposition at 50 W, possibly due to the densification of the a-Si film [51]. Comparing theresults in figure 5.3 (phase III) with the results in table 5.1 (phase II) the di�erence becomes obvious:
the highest resistivity measured in phase III samples (SRT15): ⇢ ⇠ 0.5 ⌦cm
vs. resistivity measured in phase II sample (SRT10): ⇢ ⇠ 1.2⇥ 104 ⌦cm
As this indicates, the poor performers in phase III performed relatively well, exhibiting decentelectrical properties and even though the furnace annealing resulted in the best films, the rapidthermal processing does produce satisfactory films.
30
7C����������
Depositing a highly doped silicon thin film with desirable electrical properties was achieved by sput-tering a highly doped silicon target, both at room temperatures and 200�C, either untreated or re-active ion etched, with subsequent annealing at 1100�C for 3 and 30 minutes.
It is found that a relatively high sputtering power is needed to deposit films suitable for sub-sequent solid phase crystallization. Longer annealing duration seems to yield a higher degree ofcrystallinity and an increase in the number of activated dopants.
It is also shown that crystallization of silicon in this experimental setup requires relatively highannealing temperatures (> 1000�C) to both crystallize and activate dopants.
31
B�����������
[1] European commission, paris agreement, December 2015. URL http://ec.europa.eu/clima/
policies/international/negotiations/paris/index_en.htm.
[2] Understanding temperature limits, 2010. URL http://www.unep.org/publications/ebooks/
emissionsgapreport/chapter1/Box1a.asp?
[3] Norway’s intended nationally determined contribution, February 2015. URL https://www.
regjeringen.no/en/aktuelt/innsending-av-norges-klimamal-til-fn/id2403782/.
[4] Vasilis M. Fthenakis. Overview of Potential Hazards, pages 1083 – 1096. Academic Press, 2thedition, 2012.
[5] Charles Kittel. Introduction to Solid State Physics, pages 1–22. Wiley, 8th edition, 2005.
[6] Doping (semiconductor), 2016. URL https://en.wikipedia.org/wiki/Doping_(semiconductor).
[7] W.E. Spear and P.G. Le Comber. Substitutional doping of amorphous silicon. Solid State Com-munications, 17(9):1193 – 1196, 1975.
[8] A. Kumar, F. Law, G.K. Dalapati, G.S. Subramanian, P.I. Widenborg, H.R. Tan, and A.G.Aberle. Synthesis and characterization of large-grain solid-phase crystallized polycrystallinesilicon thin films. Journal of Vacuum Science and Technology A, 32(6), 2014.
[9] T. Voutsas, H. Nishiki, M. Atkinson, J. Hartzell, and Y. Nakata. Sputtering technology of sifilms for low-temperature poly-si tfts. SHARP Technical Journal, pages 36 – 42, 2001.
[10] S.-I. Jun, P. D. Rack, T. E. McKnight, A. V. Melechko, and M. L. Simpson. Low-temperaturesolid-phase crystallization of amorphous silicon thin films deposited by rf magnetron sputteringwith substrate bias. Applied Physics Letters, 89(2), 2006.
[11] W. Wang, J. Huang, X. Zhang, Y. Yang, W. Song, and F. Huang. Crystallization of as-depositedamorphous silicon films on glass prepared by magnetron sputtering with di�erent substratebiases and temperatures. Journal of Crystal Growth, 321(1):50 – 54, 2011.
[12] T. Kobayashi, M. Koguchi, S. Iijima, M. Ohkura, and Y. Wada. Nucleation of in situ phosphorus-doped amorphous silicon films deposited by pyrolysis of si2h6 and ph3. Journal of The Electro-chemical Society, 141(5):1365 – 1369, 1994.
[13] T. Matsuyama, N. Terada, T. Baba, T. Sawada, S. Tsuge, K. Wakisaka, and S. Tsuda. High-quality polycrystalline silicon thin film prepared by a solid phase crystallization method. Jour-nal of Non-Crystalline Solids, 198 - 200(2):940 – 944, 1996.
[14] C. Spinella, S. Lombardo, and F. Priolo. Crystal grain nucleation in amorphous silicon. Journalof Applied Physics, 84(10):5383 – 5414, 1998.
[15] W.-E. Hong and J.-S. Ro. Kinetics of solid phase crystallization of amorphous silicon analyzedby raman spectroscopy. Journal of Applied Physics, 114(7), 2013.
[16] I.-W. Wu, A. Chiang, M. Fuse, L. Ovecoglu, and T. Y. Huang. Retardation of nucleation ratefor grain size enhancement by deep silicon ion implantation of low-pressure chemical vapordeposited amorphous silicon films. Journal of Applied Physics, 65(10):4036 – 4039, 1989.
32
BIBLIOGRAPHY BIBLIOGRAPHY
[17] W. Wang, J. Huang, W. Xu, J. Huang, Y. Zeng, and W. Song. Electrical properties of boron-and phosphorous-doped microcrystalline silicon thin films prepared by magnetron sputteringof heavily doped silicon targets. Journal of Materials Science: Materials in Electronics, 24(6):2122 – 2127, 2013.
[18] D. M. Mattox. Handbook of Physical Vapor Deposition (PVD) Processing, pages 475 – 527.William Andrew Publishing, 2nd edition, 2010.
[19] H. Abe, M. Yoneda, and N. Fujiwara. Developments of plasma etching technology for fabricatingsemiconductor devices. Japanese Journal of Applied Physics, 47(3R):1435, 2008.
[20] Y. Masaki, P. G. LeComber, and A. G. Fitzgerald. Solid phase crystallization of thin films ofsi prepared by plasma-enhanced chemical vapor deposition. Journal of Applied Physics, 74(1):129–134, 1993.
[21] J. F. Pierson, K. S. Kim, J. Jolly, and D. Mencaraglia. Crystallization of n-doped amorphoussilicon (pecvd) films: comparison between (spc) and (rta) methods. Journal of Non-CrystallineSolids, 270(1 - 3):91 – 96, 2000.
[22] M. H. Gullanar, Y. H. Zhang, H. Chen, W. S. Wei, G. Y. Xu, T. M. Wang, R. Q. Cui, and W. Z.Shen. E�ect of phosphorus doping on the structural properties in nc-si:h thin films. Journalof Crystal Growth, 256(3 - 4):254 – 260, 2003.
[23] S. P. Ahrenkiel, A. H. Mahan, D. S. Ginley, and Y. Xu. Solid-phase crystallization kineticsand grain structure during thermal annealing of a-si:h grown by chemical vapor deposition.Materials Science and Engineering: B, 176(13):972 – 977, 2011.
[24] Phosphie datasheet, 2016. URL https://pubchem.ncbi.nlm.nih.gov/compound/24404#datasheet=
lcss.
[25] M. Faraday. The bakerian lecture: Experimental relations of gold (and other metals) to light.Philosophical Transactions of the Royal Society of London, 147:145–181, 1857.
[26] A. Anders. Cathodic Arcs: From Fractal Spots to Energetic Condensation, pages 1 – 68. SpringerNew York, 2008.
[27] P. Arunkumar, S. K. Kuanr, and K. S. Babu. Thin Film Structures in Energy Applications,pages 1 – 49. Springer International Publishing, 2015.
[28] E. D. McClanahan and N. Laegreid. Sputtering by Particle Bombardment III: Characteristicsof Sputtered Particles, Technical Applications, pages 339 – 377. Springer Berlin Heidelberg,1991.
[29] A. W. Wright. On the production of transparent metallic film by the electrical discharge inexhausted tubes. American Journal of Science, Series 3 Vol. 13(73):49 – 55, 1877.
[30] D. M. Mattox. Handbook of Physical Vapor Deposition (PVD) Processing, pages 237 – 286.William Andrew Publishing, 2nd edition, 2010.
[31] K. Wasa, M. Kitabatake, and H. Adachi. Thin Film Materials Technology: Sputtering of Com-pound Materials, pages 17–70. William Andrew, 2005.
[32] M. M. de Lima Jr. and F. C. Marques. On the doping mechanism of boron-doped hydrogenatedamorphous silicon deposited by rf-co-sputtering. Journal of Non-Crystalline Solids, 299 - 302,Part 1:605 – 609, 2002.
[33] Y. Ohmura, M. Takahashi, M. Suzuki, N. Sakamoto, and T. Meguro. P-type doping of hydro-genated amorphous silicon films with boron by reactive radio-frequency co-sputtering. 308 -310:257 – 260, 2001.
[34] F. Fenske and B. Gorka. Highly phosphorus-doped crystalline si layers grown by pulse-magnetron sputter deposition. Journal of Applied Physics, 105(7), 2009.
[35] F. Roozeboom. Rapid Thermal Processing, pages 349 – 423. Academic Press, 1993.
33
BIBLIOGRAPHY BIBLIOGRAPHY
[36] R. B. Fair. Rapid Thermal Processing, pages 1 – 11. Academic Press, 1993.
[37] H. Hidayat, A. Kumar, F. Law, C. Ke, P.I. Widenborg, and A.G. Aberle. Impact of rapid thermalannealing temperature on non-metallised polycrystalline silicon thin-film diodes on glass. ThinSolid Films, 534:629 – 635, 2013.
[38] D. Schroder. Semiconductor Material and Device Characterization, pages 1 – 59. Wiley-IEEEPress, 3rd edition, 2006.
[39] F. Weimer. A method of measuring earth resistivity. US Bur. Standards Sci. Paper, (258), 1916.
[40] L. B. Valdes. Resistivity measurements on germanium for transistors. Proceedings of the IRE,42(2):420 – 427, 1954.
[41] D. Schroder. Semiconductor Material and Device Characterization, pages 465 – 522. Wiley-IEEE Press, 3rd edition, 2006.
[42] The hall e�ect, 2015. URL http://courses.washington.edu/phys431/hall_effect/hall_effect.pdf.
[43] Ben Streetman and Sanjay Kumar Banerjee. Solid State Electronic Devices. Pearson, 7thedition, 2015.
[44] G. Binnig, C. F. Quate, and C. Gerber. Atomic force microscope. Phys. Rev. Lett., 56:930 – 933,1986.
[45] X-ray powder di�raction (xrd), 2016. URL http://serc.carleton.edu/research_education/
geochemsheets/techniques/XRD.html.
[46] Characteristic x-rays, 2016. URL http://hyperphysics.phy-astr.gsu.edu/hbase/quantum/xrayc.html.
[47] D. Schroder. Semiconductor Material and Device Characterization, pages 627 – 688. Wiley-IEEE Press, 3rd edition, 2006.
[48] David Ne�as and Petr Klapetek. Gwyddion: an open-source software for SPM data analysis.Central European Journal of Physics, 10:181–188, 2012.
[49] ICDD (2010). PDF-2 2009 (Database), edited by Dr. Soorya Kabekkodu, International Centrefor Di�raction Data, Newtown Square, PA, USA.
[50] S. Jamil, S. K. Gupta, K. Anbalagan, and J. Akhtar. Electron-beam assisted physical vapordeposition of polycrystalline silicon films. Materials Science in Semiconductor Processing, 14(3- 4):287 – 293, 2011.
[51] M. Samuelsson, D. Lundin, J. Jensen, M. A. Raadu, J. T. Gudmundsson, and U. Helmersson.On the film density using high power impulse magnetron sputtering. Surface and CoatingsTechnology, 205(2):591 – 596, 2010.
34