INVESTIGATION OF SINGLE CRYSTAL DIAMOND FOR SWIFT HEAVY ION BEAM DETECTORS AND STUDY OF DIAMOND GROWTH FOR IMPROVEMENT OF DIAMOND MATERIAL FOR DETECTORS By A yan Bhattacharya A DISSERTATION Submitted to Michigan State University in partial fu lfillment of the requirements for the degree of Electrical E ngineering Doctor of Philosophy 2018 ABSTRACT INVESTIGATION OF SINGLE CRYSTAL DIAMOND FOR SWIFT HEAVY ION BEAM DETECTORS AND S TUDY OF DIAMOND GROWTH FOR IMPROVEMENT OF D IAMOND MATERIALS FOR DET ECTORS By A yan B hattacharya Diamond is a material with outstanding mechanical, electrical and optical properties. Single crystal diamond has a large bandgap (~ 5.47 eV), large atomic displacement energy (~ 43eV) , high electric breakdown field (10 7 V cm - 1 ) and high thermal conductivity at room temperature. All these superior properties make diamond inherently a superb radiation tolerant material for harsh radiation environments. In this study, single crystal diamo nd (scd) s ubstrates grown at Michigan State University (MSU) by microwave plasma - assisted chemical vapor deposition (MPACVD) are tested to develop detectors for swift heavy ion beam. Prior to beam irradiation, the material properties of the diamond were ch aracterized by UV - VIS spectroscopy and FTIR (Fourier Transform Infrared Spectroscopy). After fabrication of detectors, their performance was tested by irradiating with swift heavy ion (SHI) beams in the range of 100 - 150 MeV/u at the National Superconducting Cyclotron Laboratory (NSCL) at MSU. In ad dition to MSU grown samples, commercial electronic grade samples (Microwave Enterprises Ltd.) were also investigated under the same radiation envi ronment. Post irradiation, samples ware characterized by the transient current technique (TCT) to understand how charge transport properties get affected by the beam irradiation. The charge collection efficiency (CCE) and the lifetime of the holes and ele ctrons created were studied. A completely non - irradiated commercially available electronic grade diamond was a lso tested in the same testing configuration to generate a reference. Beam irradiated samples were also characterized by X - ray diffraction and Ram an spectroscopy to characterize for any structural damage. The overall characterizations mostly confirmed a de terioration of charge transport properties. However, any evidence of substantial structural damage by the irradiation could not be found. One rele vant observation was that the MSU lab grown diamond had a shorter carrier lifetime primarily due to more nitro gen impurities present in the grown diamond. Next, to improve charge transport performance of MSU lab grown diamond substrates for el ectronic applications, single crystal diamond was deposited in a low nitrogen environment to grow thick layers ( In general, epitaxial growth on surface close to the (001) crystallographic plane at a low nitrogen environment often suffers from de fects forming on the surface. Such defects arise from dislocations (already present in the substrate), twining during the growth process . A possible solution to this issue is to grow samples on misoriented substrates (i.e. on surfaces that are slightly til ted from the (001) plane). The resulting surface profile largely depends on the growth condition (i.e. temperature, pressu re, methane concentration), as well as the substrate misorientation angle. The deposition on a misoriented surface happens with a step flow growth process . It is found in research literature that impurities present in the deposition gas generally tend to c oncentrate at steps more than on terrace. Hence the spatial size of the step height and terrace width distribution produced as a resul t of misorientation angle variation is important for the quality of the deposited diamond. A detailed study of the distrib ution of step height and step height and terrace width distributions can help decide the selection of offcut angle, doping concentration and growth layer thic kness, which otherwise may create localized and non - uniform distribution of impurities and non - uniform breakdown electric field s . iv ACK NOWLEDGEMENTS I would like to begin with express ing my sincere appreciation and gratitude to my major advisor Dr. Timothy A. Grotjohn for providing me with this opportunity and his mentorship and support in this j ourney of my PhD. His patience and thoughtful suggestions always helped me to keep things rolling when everything appeared to hit almost a dead end. Next, I wou ld like to e xtend my thanks to Dr. Andreas Stolz , one of my committee members, for his guidance and help to conduct my most important experiments . Without his st rong support , this work could not have been accomplished. I would also like to thank my other committee members , Dr. Timothy P. Hogan and Dr. Qi Hua Fan , for their valuable advices and suggestions . I woul d also lik e to thank National Superconducting Cyclotron Laboratory (NSCL) at Michigan State University (MSU) for providing me their outstanding facility to conduct some experiments, which are the integral part of my PhD dissertation. I would also like to extend m y sincere thanks to Dr. Jes Asmussen for his valuable s uggestions and guidance during my exploration of the interesting and exciting work of single crystal diamond deposition . I am also thankful to Dr. Donnie Reinhard for being in my committee for few years b efore his retirement. My sincere appreciation also goes to Dr. Per Askeland of Composite Material and structure center (CMSC), MSU and Dr. Baokang Bi of W. M. Keck Microfabrication Facility at MSU for their valuable guidance in operating diffe rent instrum ent for characterization. My sincere thanks are also extended to Mr. Brian Wright and Mr. Karl Dersch for their help in troubleshooting and fixing complex technical issues in the CVD reactors. Last but not the least, I a m also thankful to all the fello w graduate s tudents and colleagues at Fraunhofer CCD, USA. v Finally, this work could not have been completed without the unconditional support of my family members, my in - laws, especially my wife Dr. Aparajita Banerjee for her support to fight and handle a ll the extre me stressful moments during this journey. I do not have any words to express my gratitude to my late mother, who had been my strong pillar of support, however, could not witness the end of this long journey. I would like to thank all my friends within and outside MSU for the important role they played in this journey. vi TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ....................... viii LIST OF FIGURES ................................ ................................ ................................ ....................... ix KEY TO ABBREVIATIONS ................................ ................................ ................................ ...... xiv Chapter 1. Introduction ................................ ................................ ................................ ................... 1 1.1 Research motivation ................................ ................................ ................................ .............. 1 1.2. Objectives ................................ ................................ ................................ ............................. 4 1.3. Dissertation out line ................................ ................................ ................................ .............. 6 Chapter 2. Introduction to radiation detectors ................................ ................................ ................ 9 2.1. Properties of diamond material ................................ ................................ ............................ 9 2.2. Radiation detection basics ................................ ................................ ................................ .. 11 2.3. Types of radiation detectors ................................ ................................ ............................... 16 2.4. Interaction of radiation beams with material ................................ ................................ ...... 20 2.5. Current state of art of diamond - based radiation detectors ................................ ................. 28 Chapter 3. Performance test of radiation detectors ................................ ................................ ....... 33 3.1. Fabrication of radiation detectors ................................ ................................ ....................... 33 3.2. Irradiation of diamond detectors by Swift Heavy Ion (SHI) beams ................................ ... 35 Chapter 4. Post irradiation characterization of diamond detectors ................................ ............... 41 4.1. Electrical characterization ................................ ................................ ................................ .. 42 4.1.1. Transient current technique ................................ ................................ ......................... 42 4.1.2. Leakage current test ................................ ................................ ................................ ..... 51 4.2. Optical characterization ................................ ................................ ................................ ...... 57 4.2.1. UV - Vis spectroscopy ................................ ................................ ................................ ... 57 4.2.2. Fourier Transform Infra red spectroscopy (FTIR) ................................ ....................... 61 4.3. Structural c haracterization ................................ ................................ ................................ .. 63 4.3.1. High resolution X - ray diffraction (HRXRD) ................................ .............................. 64 4.3.2. Raman spectroscopy ................................ ................................ ................................ .... 69 Chapter 5. Single crystal diamond deposition by chemical vapor deposition (CVD) .................. 7 4 5.1. Brief overview of histo rical events ................................ ................................ .................... 7 4 5.2. Hot filament CVD ................................ ................................ ................................ .............. 7 6 5.3 . Plasma Enhanced CVD (PECVD) ................................ ................................ ..................... 7 8 5.4. Microwave Plasma - Assisted CVD (MPACVD). ................................ ............................... 79 5.5. Deposition of single crystal diamond ................................ ................................ ................. 82 5.5.1. Chemical reactions for CVD diamond growth ................................ ............................ 85 5.5.2. Microwave plasma assisted C VD reactor ................................ ................................ .... 90 Chapter 6. Single crystal dia mond deposition on misoriented/offcut substrates ........................ 109 vii 6.1. Overview of scd deposition on misoriented substrates ................................ .................... 109 6.2. Aim of current study ................................ ................................ ................................ ........ 120 6.3. Experimental procedure ................................ ................................ ................................ ... 121 6.3.1. SCD deposition process ................................ ................................ ............................. 121 6.4. Optical characterization of offcut grown samples ................................ ............................ 125 6.4.1. Operation of optical microscope ................................ ................................ ............... 125 6.4.2. Observations and discussions ................................ ................................ .................... 128 6.5. Study of distribution of step and terrace growth ................................ .............................. 137 6.5.1. Basic operation of DEKTAK profilometer ................................ ............................... 141 6.5.2. Observations and discussions ................................ ................................ .................... 144 Chapter 7. Conclusions and future directions ................................ ................................ ............. 153 7.1. Conclusions ................................ ................................ ................................ ...................... 153 7.2. Future investigations ................................ ................................ ................................ ........ 157 BIBLIOGRAPHY ................................ ................................ ................................ ....................... 159 viii LIST OF TABLES Table 1.1. Some important material properties of diamond ................................ ......................... 10 Table 4.1 . Full width half maximum (fwhm) measured for peak (004) a cross the sample 68 Table 5.1. Differences in important design parameters between Reactor A, B and C ................. 9 7 Table 5.2. Description of the different sections/attachments of reactor - B ................................ ... 99 Table 6.1. Lateral and v ertical growths on di fferent offcut angle substrates .............................. 1 3 5 Table 6.2 . Average step height and terrace width distribut ion for different offcut growth ........ 15 2 ix LIST OF FIGURE S Figure 2.1. Bond of carbon atoms in diamond structure ................................ ................................ . 9 Figure 2. 2 . Penetration dept hs of different ionizing sources [9] ................................ ................... 1 1 Figure 2.3. A schematic of detector`s circuitry ................................ ................................ ............. 1 2 Figure 2.4. Current measurement mode of a detector [10] ................................ ........................... 13 Figure 2.5. Current in a detector due to steady state irradiation ................................ ................... 14 Figure 2.6. Operation of a detector i n pulse mode [10] ................................ ................................ 14 Figure 2.7. The output signal for a detector circuit with small time constant (b) and for a circuit with large time constant (c) given a particle interaction (a) [10] ................................ .................. 1 5 Figure 2.8 . Charge collected by different gas filled detectors [5] ................................ ................ 1 7 Figu re 2.9. A p - n junction acting as a radiation detector [12] ................................ ...................... 19 Figure 2 .10 . Energy loss diagram of different particles [5] ................................ .......................... 2 3 Figure 2.11. Mean energy loss rates for particles in different medium [19] ................................ 2 4 Figure 2.12. Stopping power vs Energy [16] ................................ ................................ ................ 25 Figure 2. 13. Energy distribution of a monoenergetic beam varying over penetration depth [ 19] ................................ ................................ ................................ ................................ .............. 2 6 Figure 3.1. Top surface of the single crystal CVD diamond plate with electrodes ...................... 3 4 Figure 3.2. Signal captured from lab grown SCD detector (GYJ - 148C) irradiated w ith 96 Zr beam. (The upper signal represents the response of heavily irradi ated segment and the lower signal shows response of lightly irradiated segment) ................................ ................................ .............. 3 6 Figure 3.3. The output voltage from lightly irradiated (dark blue) and heavily irradiated (cyan) a fter a certain particle fluence (2.77 10 12 particles/cm 2 ) for different bias voltages ................. 3 7 Figure 3.4. The output voltage measured at different time (with a range of bias voltages) during irradiation from heavily irradiated segment (red curve) vs lightly irradiated segment (black curve) ................................ ................................ ................................ ................................ ............ 38 Figure 3.5. Relative signal drop of detectors over particle fluence ................................ .............. 39 x Figure 4.1. Transient current signal shape for different quality of diamond [ 36] ........................ 4 3 Figure 4.2. Transient current signal for electron and holes fro m a single crystal diamond [ 36] .. 4 4 Fig ure 4. 3. Transient current signals from commercial diamond - 1 ................................ .............. 4 5 Figure 4.4. Transient current signal from lightly irradiated segment of commercial diamond - 2 ................................ ................................ ................................ ................................ ................... 4 7 Figure 4.5. Transient current signal from heavily irradi ated segment of commercial diamond - 2 ................................ ................................ ................................ ................................ ...................... . 4 7 Figure 4.6. Charge collection versus applied field ................................ ................................ ....... 4 8 Figure 4.7. Transient signal generated from a MSU lab grown diamond ................................ ..... 5 0 Figure 4.8. Transient current signal generated from a non - irradiated commercial electro nic grade diamond ................................ ................................ ................................ ................................ ......... 5 0 Figur e 4.9. Leakage current in a polycrystalline pixel detector [ 44] ................................ ............ 5 2 Figure 4.10. Leakage current measurement on commercial diamond - 2 ................................ ....... 5 3 Figure 4.11. Repeat run of leakage current measurement (a & b) with segments measured in diffe rent orders ................................ ................................ ................................ .............................. 5 3 Figure 4.12. Leakage current measurement with sample illuminated with red LED ................... 5 5 Figure 4.13. Transient response of heavily irradiated segments of commercial diamond 2 ........ 5 6 Figure 4.14. Transient response of lightly irradiated segments of co mmercial diamond - 2 ........ 5 6 Figure 4.15. % Transmission and absorbance of different quality diamond ................................ 59 Figure 4.16. Effect of SHI beam irradiation on samples GYJ 148A and 148C ............................ 6 0 Figure 4.17. % Transmission on SHI beam irradiated samples ................................ .................... 6 0 Figure 4.18. Infrare d spectrum of commercial and lab grown diamonds ................................ ..... 6 2 Figure 4.19. Post irradiation infrared spectrum of lab grown diamonds ................................ ...... 6 3 Figure 4.20. Rocking curve scan of GaN sample [ 54] ................................ ................................ .. 6 5 Figure 4.21. A schematic of different positions for rocking curve scan ................................ ....... 6 6 xi Figu re 4.22. The distribution of rocking curve scan at different spots of sample GYJ - 148C ................................ ................................ ................................ ................................ ....................... 6 7 Figure 4.23. A schematic of Raman spectrometer [ 54] ................................ ................................ 7 0 Figure 4.24. Raman shift of an irradiated diamond, (a) damaged region , (b) undamaged region . ................................ ................................ ................................ ................................ .................... ... 7 0 Figure 4.2 5. Raman spectrum collected from the instrument used for this work (a) electronic grade commercial diamond, (b) HPHT diamond and (c) a l ab - grown non - irradiated diamond ................................ ................................ ................................ ................................ ....................... 7 2 Figure 4.26. Raman spectrum collected from SHI irradiated diamonds . ................................ ...... 7 3 Figure 5. 1. A schematic of HFCVD operation [72]. ................................ ................................ ..... 7 7 Figure 5.2. Reaction steps in PECVD process [73]. ................................ ................................ ..... 79 Figure 5.3. A cross sectional configuration of a MPACVD reactor [74]. ................................ .... 8 1 Figure 5.4. A schematic showing the process of diamond deposition during CVD process [20 ] , [ 84]. ................................ ................................ ................................ ................................ ............... 8 3 Figure 5.5. Bachman C - H - O triangle describing the gas composition required for dia mond growth [ ... . 85 Figure 5. 6. A cross - sectional view of a typical MSU designed reactor [97]. ............................... 9 2 Figure 5. 7 . (a) Hybrid microwave cavity applicator (cross sect ional view) with cylindrical and coaxial intersecting at z=0 plane (b) cross sectional view of standing wave pattern in Reactor B (smaller substrate holder) [104]. ................................ ................................ ................................ ... 9 4 Figure 5.8 . Cross section of generalized design of Reactor C [106]. ................................ ............ 9 6 Figure 5. 9 . A sch ematic of the entire diamond deposition system based upon reactor B design . ................................ ................................ ................................ ................................ ....................... 9 8 Figure 5.10. Design of the cooling stage for reactor B [ 107,108]. ................................ ............. 10 4 Figure 5.11. Dime nsions of a 1.8 mm deep pocket holder [ 108]. ................................ ............... 10 5 Figure 6.1. Appearance of defects on sc d surfaces with very low percentage of CH 4 addition [ 114]. ................................ ................................ ................................ ................................ ........... 110 Figure 6.2. Appearance of an etch - pit, as viewed in AFM [ 120]. ................................ .............. 11 3 xii Figure 6.3. Growth rate vs misorientation angle toward [110] at (a) 6% CH 4 and 800°C, (b) 1% CH 4 and 800°C, and (c) 1% CH 4 a nd 1000°C as reported in reference [12 4 ] ............................ 11 6 Figure 6.4. DICM images of the surface morphologies of (001) homoepitaxial diamond films grown for 5hr at 875°C w ith CH 4 concentrations of (a) 1% (b) 2%, and (c) 6% in H 2 [125 ] . .... 11 7 Figure 6.5. Orientations of offcut substrates along different directions from the (100) plane . .. 12 2 Figure 6.6. Schematic of a diamond deposition process showing the pocket recess t hat holds the diamond substrate . ................................ ................................ ................................ ...................... 12 5 Figure 6.7. ( a) S chematic of differential interference contrast imaging (DICM) mode (b) Dark field imaging mode [ 130]. ................................ ................................ ................................ ........... 12 7 Figure 6.8. An image (25x) of a polished HPHT substrate and corresponding birefringence (top row) and image (25x) of commercial CVD plate and corresp on ding birefringence (bottom row) . ................................ ................................ ................................ ................................ ..................... 12 9 Figure 6.9. An optical micrograph (25 X ) of a n epilayer deposited on a commercial CVD plate (left) and the same surface looked at higher magnification (100 X ) . ................................ .......... 130 Figure 6.10. Top surface of SCD deposited layers for different offcut angles ([100] direction bevel) & different growth durations. The thinner edge of each beveled sample i s facing the bottom of each micrograph . ................................ ................................ ................................ ........ 1 3 1 Figure 6.11 . Top surface of deposited layer on [110] beve l and [100] parallel substrates . ........ 13 2 Figure 6.12. The top surface and side view ( a & b) of three different offcut angles (2.5°, 5° and 10° along [100)] bevel) showing the lateral growth of the deposited layers. All pi ctures are for 24 hours growth . ................................ ................................ ................................ .............................. 13 3 Figure 6.13. Vertical growth rate of deposited layer on different offcut samples me asured at different locati ons for 24 hours deposition. ................................ ................................ ................ 13 6 Figure 6.14. High magnification (500x) pictures of step - terrace formation (a) 5° bevel (c) 10° with the bevel along the [100] direction and the corresponding [110] direction bevel at (b) 5° and (d ) 10° . ................................ ................................ ................................ ................................ ........ 13 8 Figure 6.15. Step and terrace propagation along (a) 5° and (b) 10° offcut parallel substrate growth . ................................ ................................ ................................ ................................ ........ 13 9 Figure 6.16. Surface roughness profile of a selected area (f lat region) of 10° bevel growth ..... 1 4 1 Figure 6.17 . A schematic of the operation of stylus - based surface profilometer [134 ]. ............ . 14 2 Figure 6.18. Image of the measurement stage of Veeco 6M profilometer [135]. ....................... 14 3 xiii Figure 6.19 . Schematic showing leveling (from (a) to (b)) and measurement of step height and terrace width on a 10° offcut b evel sample . ................................ ................................ ................ 14 5 Figure 6.20 . Step - terrace distributions (bar diagrams) at different sections of the 10° offcut [100] bevel direction grown layer . ................................ ................................ ................................ ....... 14 5 Figure 6.21 . Step - terrace distributions (bar diagrams) at different sections of a for a 5° offcut [100] bevel grown layer . ................................ ................................ ................................ ............. 14 7 Figure 6.22 . Step - terrace structures and their distributions (bar diagrams) for 5° offcut (a) (100) bevel (b) (110) bevel and (c) (100) parallel substrate thick growth . ................................ .......... 14 9 Figure 6.23 . Step - terrace structures and their distributi ons (bar diagrams) for 10° offcut (a) (100) bevel (b) (110) bevel and (c) (100) parallel substrate thick growth . ................................ .......... 1 5 1 xiv KEY TO ABBREVIATIONS AFM Atomic force microscopy C CD C harge collection distance CERN European Organization for Nuclear Research CMS Comp act muon solenoid CVD C hemical vapor deposition DF Dark field DICM Differential interference contrast imaging FH Flat head hillocks FTIR Fourier - transform infrared FWHM Full width half maximum HFCVD Hot filament chemical vapor deposition HL - LHC High Lumino sity Large Hadron Collider HPHT H igh pressure high temperature HPMS High Pressure Microwave Source HRXRD H igh resolution X - ray diffraction IBIC I on beam induced current LED Light emitting diaode LHC Large Hadron Collider MCPR Microwave cavity plasma reacto r MFC Mass flow controllers MIP M inimum ionizing particles MPACVD M icrowave plasma assisted chemical vapor deposition xv NIRIM Nat ional Institute for Research in Inorganic Materials NSCL National Superconducting Cyclotron Laboratory PACVD Plasma Ass isted Chem ical Vapor Deposition pcd P olycrystalline diamond pCVD P olycrystalline chemical vapor deposition PECVD Plasma enhanced chemical vapor deposition PH Pyramidal hillocks scCVD S ingle crystal chemical vapor deposition scd Single crystal diamonds SHI S wift heav y ion TCT T ransient current technique TE Transverse electric TEM T ransm ission electron microscopy TM T ransverse magnetic UC unepitaxial crystals UV - VIS ultra violet - visible 1 1.1 Research m otivation Carbon has many different allot ropes out of which diamond stands as the most cherished one among all due to its forever high demand and popularity. The sp 3 bonding of carbon atoms in diamond a ttributes to its several extreme properties in terms of electrical, mechanical or thermal behavior. It is one of the hardest and most chemically inert materials. Dia mond has become a true engineering material with many of its electronic properties superior in comparison to some other contemporary electronic materials. It is already being steadily explor ed in the domain of power electronics and heat sinks, as diamond - b ased devices c an be operated with minimal cooling or without additional cooling. However even with many potential advantages, the cost of production and a reliable quality control of diamond substrates still limits diamond in comparison with the well - estab lished silicon - based technology. Natural diamond already carries a great cost with it due to the complexity of excavation and related costs and limited worldwide resources. Hence their appli cation hardly expanded outside the jewelry business. It was only a fter many decades of trials and failures, that the first reliable report from General Electric Company from Schenectady, New York in 1954 showed a pathway towards producing lab grown diamond s and opened a platform towards engi neering applications of diamon d [1] . In the last couple of decades, diamond has emerged as a semiconductor material. At present, industrial application s of diamond ranges beyond its traditional application as abrasives to the field s of optics, RF MEMS technology, heat sinks, power electronics and high energy particle detector s [2] . 2 The first diamond synthesis process mimicked the natu ral process of diamond gr owth, i.e. carbon was crystallized into diamond at a pressure of 70 Kbar/inch 2 and a temperature close to 2500 °C by using a belt press. This technique evolved as the "High P ressure H igh T emperature (HPHT)" method of diamond growth . Diamond growth techn olo gy has gone through many transformations and advancements. Currently , the microwave plasma assisted chemical vapor deposition (MPACVD) technique stands as a very efficient and popular method of diamond growth for the highest qualit y diamond . Although advan cement of diamond growth still has many obstacles to overcome to make diamond based electronic devices scalable for mass production, there are a few areas where diamond has already shown superiority and radiation detectors for high energy particles is one such case. This application of diamond is based upon the inherent strength of sp 3 bonds in the diamond crystal structure, and the related high radiation hardness of diamond, something which silicon - based technology does not off er. As silicon is not as radi ation hard as diamond, beyond a certain harsh radiation environment threshold the well - established Si based detection systems need to be upgraded by relatively new diamond - based detection systems. Some studies on diamond based detection system s at CERN (Eu ropean Nuclear Research Center, Geneva, Switzerland) have already shown promise that diamond - based detectors will survive in harsh environments for years [3] . This has generated many collaborations between industry and pa rticle physics groups to come 100% charge collection at the initial stage of irradiation. It is already an accepted fact that diamond`s extreme radiation hardness will be a great advantage to push the limit of higher particle energy and fluence. Hence in the near future at CERN the target is to push the integrated luminosity by two orders of magnitud e higher than its current state [4] . This higher luminosity is a challenge 3 for detectors w hich need to survive fluences of 1 MeV neutron equivalent on the order 10 16 cm - 2 [5] . Diamond with its higher radiation hardness is ex pected to meet such challenge s. As it is important to understand the performance of diamond detectors in radiation harsh environments over a long - time frame , it is also important to know the lifespan and process of degradation of detectors in different ty pes of high energy beams, esp ecially the physical state of the material diamond after being exposed to heavy ion irradiation for a prolong time. The RD42 collaboration at CERN has made some significant studies over the last two decades to understand such p henomena. Some of their exten sive study shows the change in detector properties under different energies of proton and neutron beams. However, a global parametric distribution is still hard to establish as the material degradation at different energy regim e s requires different physica l models for explanation. In this proposed study one of the main goals is to look at how single crystal diamond degrades in swift heavy ion (SHI) beams [6] generat ed at the National Superconducting Cyclotron Laboratory (NSCL) at Michigan State University. In this type of beam, it is the mass of the particle toge ther wit h the kinetic energy and finally the charged state of the particles that determines the interactio n of the particles with the diamond detector material. The major goal of this work has been dedicated in understanding how the quality of diamond material af fects the radiation induced degradation process. Although higher quality of diamond (with very low nitrogen impurities) is always desired for any electronic applications , the cost of such diamond plates are very high. Therefore, comparisons of performance have been made between commercially available electronic grade diamond and lab grown diamonds at MSU to establish a more cost - effective research pathway on detectors. However due to presence of much higher nitrogen impurities in some of our own lab grown diamonds , the lifetime of the hole and 4 electron carriers generated by high energy particles was sh orter than in low er nitrogen impurities diamond substrates. As a next step to upgrade the MSU lab grown diamonds work was performed to establis h diamond grow th in a low nitrogen environment. However, growing thick diamond plates using a homoepitaxial proc ess in a low nitrogen environment with a top growth surface close to [001] crystallographic plane is a challenge. Several research literature s sh ow that controlling defects generat ed from dislocations and associated twinning is an extensive c hallenge when growth is done with a low nitrogen impurity level . As a method to overcome this problem misoriented substrates (with misorientation angle from t he (100) plane along either the [100] or [110] directions ) are used to deposit single crystal diam ond , which allows step flow growth across the surface. However, a selection of the growth condition (feed gas ratio, temperature , etc. ) combined with selectio n of the misoriented angle has a strong influence in the appearance of the top growth surface duri ng deposition and the final surface structure, i.e. the surface has steps and terraces in the order of nanometers to micro meters dimens ions . Therefore, it is important to have an idea of step height and terrace width distribution with respect to misorienta tion angle because the incorporation of impurities is higher in the step regions. This can also play a crucial role when selective impurity/dopant are added t o feed - gas to create p - type or n - type diamond. Hence a second major part of this thesis explore s s ystematically the step height and terrace width distribution with varied misorientation angle for a set of growth condition in MPACVD diamond deposition . 1.2. Objective s Th e purpose of this study was to investigate and improve the performance of single crystal diamond detectors in swift heavy ion (SHI) beams. A focus of this study is understanding the 5 material degradation of the detectors after a high flux of SHI particle irradiation in the energy range of 120 MeV/u. The key objectives are : Quantify and understand the performance of diamond detectors made from both MSU lab grown and commercial electronic grade diamonds. This study serves a purpose of unders tanding the use of CVD grown diamond of various qualities for SHI beam detection. Quantify the lif etime of diamond detectors for SHI beam detection and to understand their eventual degradation. An important aspect of this work is to predict the typical s pan of time that these detectors can produce a reasonable quality of detection signal. A study of th e detector`s output at different fluence s of high energy particles is used as an estimation for life span prediction. Measure the effect of SHI beam irradi ation on the electrical, optical and structural properties of the diamonds. The effect of irradiatio n on diamond material is determined by the energy and type of the beam and the type of interaction that goes on inside the material. This work looks to unde rstand the type of the degradation that happens in the diamond detectors. Explore pathways of impro vement in lab grown diamond for detector applications. It is always desired to have a diamond with very low impurities and minimal defects for detectors. Sp ecifically, both low impurity levels and minimal defects enhance s the charge collection distance of the diamond detector. However, for lab grown diamonds, it is still a great challenge to get both impurities and defects well under control, as there is a hi gh probability of defects propagating from the seeds used to grow single crystal diamond plates. Research literature s show that some control in defect propagation can be achieved by off axis growth of the plates from the crystallographic axis (001). Th is proposed work will further investigate off - axis 6 growth of thick diamond plates and then character ize them to establish better growth techniques of lab grown diamonds applicable for detector fabrication. 1.3. Dissertation outline The plan mentioned in previous section of this thesis work has been conducted, divided and explained in the next six chap ters. The chapter 2 begins with listing important properties of diamond which makes it an interesting material, especially for radia tion detectors. The chapter next covers the working principles of radiation detectors and their types. Then the discussio n moves to the kind of interaction s that takes place when energetic particles impinge and penetrate a host material. Here a simplified scenario is described for the sake of the general aim of this thesis. However, this area of interaction of heavy ion beam s with materials (especially with diamond) is still an evolving field. This chapter finally covers the state of worldwide research that has been going towards exploration of single crystal and polycrystalline diamond as a material for radiation detector. The basic structure of radiation detector made from single crystal diamond plates is discussed in the first section of chapter 3. T h e detector fabrication re lies on a parallel plate capacitance structure where a dielectric material (diamond plate) is grow n with electrodes in two flat surfaces. The next section discusses how these detectors are irradiated with a swift heavy ion beam of en ergy 120 - 150 MeV/nucleon at the National Superconducting Cyclotron Laboratory (NSCL) at MSU. Further this section describ es how the response of the detector is measured (output voltage) and eventually translated into analyzing/quantification of degradation and the basic parametric nature that the degradation curve follows. 7 Chapter 4 investigates the effect of beam irradiat ion on the electrical, optical and structural properties of diamond. The first broad section covers electrical characterization. The basis of this work is built upon the transient current technique, where a small - scale irradiation on one side of the sampl e generates charges, which are swept across the sample with an externally applied electric field. The first section discusses how transient current signals are measured with an external circuit and eventually how charge collection and lifetime of charges a re calculat ed from the signal curves. The second section of electrical characterization investigates if any significant change takes place in the leakage current of the detectors, which is expected to stay in the picoampere ranges within a range of applied electric f ield for a good quality of diamond . In the next broad section, the optical properties of beam irradiated diamonds were studied. The optical transmission of diamond and its absorbance were measured in the UV - Vis and infrared region. This study he lped to com pare between pre - irradiated and post irradiated optical properties of diamond plates. The last section in this chapter looks for evidence of any structural damages that might have happened due to the beam irradiation. A rocking curve study of th e (004) pla ne in the diamond was measured by a high - resolution x - ray diffraction system and Raman spectroscopy of the irradiated diamond plates was also investigated. An important area directly related to this work is the growth of single crystal diamond . Chapter 5 comprises a general discussion of single crystal diamond deposition by the chemical vapor deposition technique. The chapter begins with chronological events through which diamond deposition by chemical vapor deposition technique has reached its current st ate. The next section describes a general overview of processes associated with chemical vapor deposition of diamond. Finally, the chapter discusses the microwave plasma assisted chemical vapor deposition (MPACVD) technique. In particular , diffe rent types of MPACVD reactors available worldwide 8 are described with an expanded emphasis on reactors designed at Michigan State University (MSU) and their operation. The importance of diamond deposition on misoriented substrates (i.e. substrates with th e top growt h surface prepared at a miscut angle from the (001) crystallographic plane) is discussed at chapter 6. The chapter begins with a detailed literature review on the aspects of diamond deposition on misoriented substrates. The next section discusse s the relev ance of this growth strategy with this current work. The specific reason is to reduce the nitrogen content in the diamond which will give a better performance of diamond detectors. This section is followed by a detailed explanation of specific s teps associ ated with growth on misoriented substrates, especially the preparation of offcut angled substrates. The next section in this chapter covers the characterization of the growth samples examined using an optical microscope, i.e. discussion on the d ifferent fe atures present on the surface of the grown layer. The final broad section describes in detail how step and terraces appear on misoriented substrates. Their dependence on growth time and the offcut angle is carefully explored. This section ends w ith buildin g up an overall distribution of step heights and terrace widths for grown layers on different misoriented substrates. Finally, chapter 7 concludes and summarizes the overall work performed in this thesis. It also provides future direction towa rds growth of thick single crystal plates in a low nitrogen impurity environment by using misoriented substrates for future applications of lab grown diamonds for radiation detectors and other electronic devices. 9 2 .1 . Properties of diamond material Diamond is comprised of carbon atoms forming a tetrahedral structure, connected by strong sp 3 hybrid bonds (Figure 2 . 1) [7] . The strong sp 3 bond results in diamond s many exceptional properties. Single crystal diamond has a very high atomic density, highest bulk modulus and highest thermal conductivity. It is also highly transparent in far infra - red to visible and partially in UV range. Diamond has been t raditionally used as an abrasive material and it still enjoys the status of being one of the hardest materials. Diamond is a wide bandgap semiconductor in the semiconductor materials group with high carrier mobility. Table 1.1 [8] shows some of the very outstanding properties of diamond as a semiconducting material. Figure 2.1. Bond of carbon atoms in diamond structure 10 Table 1.1 . Some important material properties of diamond [ 8] An application of single crystal diamond is radiation beam detection in harsh radiation environments. The idea behind this application is to develop detectors for high radiation environments where silicon - based detector technology fails quickly. Single cry stal diamond is considered to be a radiation hard material due to its high atomic displacement energy of 40 - 50 eV (compared to 13 - 20 eV for Si) and its high breakdown electric field. In addition, the higher carrier mobility helps in the better realization of fast detectors. Diamond is considered as a wide bandgap semiconductor which behaves almost like an insulator in undoped condition. Therefore, while used as a detector it doesn`t require the formation of p - n junctions by doping. Instead single crystal un doped diamond material can perform quite well as solid state detector . Mechanical hardness ~ 100 GPa High est bulk modulus (1.2 10 12 N m - 2 ) Highest thermal conductivity @ room temperature (2 10 3 W m - 1 K - 1 ) Thermal expansion coefficient @ room temperature (0.8 10 - 6 K - 1 ) Wide bandgap (~5.47 eV) High breakdown fiel d (~ 10 7 V cm - 1 ) Displacement ener gy 40 - 50 eV/atom Carrier mobility ( 2 400 cm 2 V - 1 s - 1 for electrons and 2100 cm 2 V - 1 s - 1 for holes Energy required to create e - h pair 13 eV 11 2.2. R adiation detection basic s The term "radiation" expresses the phenomenon of transportation of energy and mass through space. This energy originates from a source and travels through space till it gets partially or fully absorbed by any material. Radiation can be broadly classifie d i nto two forms including electromagnetic radiation and particle radiation. Radiation can be of either ionizing or non - ionizing. Ionizing radiation can create unbound charges in a material by liberating electrons from the atom. The amount of charge genera ted depends on many factors like the type of radiation, the energy of the radiation with which it strikes the material, the target material itself and the depth it - - - rays and X - rays. Th ere - - - rays. Also bombarding an y stable nuclei with high energy particles produced in accelerators can yield these kinds of radiations. A schem atic in figure 2.2 . shows the penetration depths of different ionizing radiations [9] . Figure 2.2. Penetration depths of different ionizing sources [9] 12 The basis of detection of radiation relies on the interaction of the radiat ion beam with a suitable material build into a detector device. Thus, material selection plays an important role in the fabrication of radiation detectors. The interaction of the be am with the detector material is converted to electrical signals sensed thr ough external circuit elements. Currently, most of the radiation detectors depend on ionization as the principle of operation. Charges are generated through ionization and any movem ent of the charges produced by an applied electric field before they recomb ine is sensed through the external circuit (as shown in a schematic in figure 2 .3). The number of quanta of radiation that interacts with the material within a particular time span generally determines the external signal that is generated. Figure 2.3. A sc hematic of detector`s circuitry The modes of detection can be broadly classified into two main mechanisms, (i) current mode detectio n and (ii) pulse mode detection [10] . In the (i) current measurement mode of detection a current meter or more precisely a nano/pico ammeter is attached to the output terminals of the radiation detector. A schematic of the mechanism is shown below in figure 2 .4, 13 Figure 2.4. Current measurement mode o f a detector [ 10 ] The final output from this measuring device can be expressed as, where the device has a response time T, where i(t') is the current generated by a single particle interacting with the detector. The actual output I(t) is cal culated by averagi ng many of the events that happen within the time frame, T. A steady - state irradiation of a detector can be expressed as combination of a constant current I o i (t), arising from the random nature of radiation as shown in the figure 2 .5. 14 Figure 2.5. Current in a detector due to steady state irradiation [10] Figure 2.6. Operation of a detector in pulse mode [10] The (ii) pulse mode detection provides information about the amplitude and timing of an individual radiation event. The nature of the response is guided by the circuit (preamplifier) the detector is connected to. The equivalent circuit diagram is shown in figure 2 .6. The effective inpu t resistance is R and the equivalent detector and circuit capacitance is represented by C. It is the time - the measuring circuit is gi o different regimes the pulse mode operation can work. c , where t c is the charge collection time), the current flowing through the external load follows the exact shape of the instantaneous current in the d etector. Hence the voltage v(t) across the resistor also takes nearly the same shape of the time 15 dependent current, as shown in figure 2.7 (b). When timing information or high event rate becomes an important issue then radiation detectors are operated in s uch conditions. detector`s charge collection time, i.e ( c ). In this case the signal pulse current i(t) charges the capacitor and increases load voltage v(t) . Then the capaci tor will slowly discharge through the resistor and v(t) will gradually reach back to zero. The voltage variation is shown in figure 2 .7 (c). Figure 2 .7 . The output signal for a detector circuit with small time constant (b) and for a circuit with large time constant (c) given a particle interaction (a) [ 10 ] It is observed that the rise time for the signal to reach the maximum value is dependent on the dete response (i.e. its charge collection time) , and not influenced by the load circuit. In 16 contra st, the decay time is only determined by the time constant of the external circuit. Also, the ratio of the total charge generated in the detector Q and the capacitance of the load circuit C decides the maximum value of the external voltage, i.e. V max . Ther efore, if the capacitance C remains constant for each pulse, the voltage maintains a linear dependence on the generated charge . In such cases, a charge sens itive preamplifier circuit is used which can eliminate the variation of capacitance on output voltag e. Pulse mode operation is the most widely used detec tor configuration for radiation as it holds more information through the detection of individual pulses . 2. 3 . Types o f radiation d etectors Radiation detectors are classified based upon the type of me dium used for the radiation beam to interact with and the way the interaction is captured to convert it into a measurable physical quantity [11] . On some occasion, some devices may only coun t the rate of impinged particles and hence known as counters, however the major goal of detectors is to quan tify the energy or study the beam profile both in terms of spatial and temporal distribution. In a gas - filled detector the ionization happens in a gaseous medium, i.e. when a radiation beam passes through, a small volume of gas gets ionized and produce fre e electrons and ions [12] . An applied electric field dr ives the electrons and ions to respective electrodes to generate a current signal in the output. The generated current amplitude depends on amount of generated charges, strength of the field, geometry of electrodes and recombination of charges. There are f ew different types of gas filled detectors, namely; ionization chambers, proportional counters and Geiger counters as their mode of operation is d ecided based upon the range of applied voltage, as seen in figure 2.8 . In an ionization chamber a weak electr ic field is applied between parallel plates to collect most of the ions produced by the incident radiation. Ionization chambers are used to detect mostly x - rays and 17 - rays. Ionization chambers work in either current detection mode or pulse counting mode. I n proportional and Geiger counters a wire along the axis of the hollow cylinder acts as the positive electrode and the hollow cylinder surrounding the wire is used for the negative electrode. When the gas is ionized the electrons and the ions are swept away from the sensing volume to respective electrodes. In proportional counter the amount of collected charges is larger than that by ionization cham ber. Proportional counters use stronger electric drift fields and primary charges are accelerated to high enough energies to cause secondary ionization. This gas amplification process generates a larger amount of total charge which is still proportional to the amount of primarily generated ions. Figure 2.8 . Charge c ollected by different gas filled detectors [5] 18 In Geiger counters the gas amplification process takes place over a larger volume through photons generated through secondary collisions over the entire vol ume. The gas amplification is eventually limited when the field due to the secondary electrons are strong enough to interact with the applied field and the effective field is barely strong enough to support any gas amplification. In this scenario the outpu t current no longer is proportional to the number of primarily generated ions and it gets saturated to a value depending on the detector geometry and the applied field. Generally, Geiger counter works in a pulse mode and the pulse rate is counted through a n external circuit. Fig ure 2 .8 depicts the respective ranges of operation of different gas filled detectors. A Scintillation detector is another type of radiation detector which converts interaction of ionizing radiation into luminescence. Through ionizati on, an electron can get excited from valence band to conduction band or to any mid - gap states created due to impurity. An exciton is formed when the hole and the electron form an electrostatically bonded pair. The excited electron decays to the ground stat e and releases the energy in the visible to UV range. This released energy spectra is directed to a photocathode in a photomultiplier where electrons are emitted due to photoelectric effect. The electrons generate a current in the external circuit by a mul tiplication process. Scintillation detectors are characterized by a fast response and significant intensity of the generated signal. Scintillation detectors are comprise d of either organic or inorganic material a s detection material . Generally inorganic s cintillators are slightly slower in response as compared to organic one s , however inorganic scintillators are more efficient due to having denser structure and higher atomic number. Scintillation detectors are qui te widely used for medical purposes, high - e nergy physics experiments and security purposes. Semiconductor detectors [13] are another widely used detector type that work almost like ionization chambers. These are solid state detectors where the radiation creates electron and hole 19 pairs that are swept away unde r the influence of an external electric fie ld to generate a pulsed signal through external electronics. The major advantage of semiconductor detectors is that these devices offer excellent energy resolution as it requires less energy to create an e - h pair for semiconductors (~3.6 eV for Si, 2.6 eV for Ge and 13.4 eV for diamond) which is generally less as compared to energy required to create an electron - ion separation. Si and Ge are the most used material s for sem iconductor detectors. In general, for a sem iconductor radiation detector a p - n junctio n is used to generate a charge depleted region. For this detection purpose, the p - n junction is operated in a reverse bias mode, which results to an increase of depletion region width as well as a drop for juncti on capacitance. When a radiation beam impin ges in the junction region it creates electron hole pairs which can get swept across the junction by an external applied field and reach respective electrode to generate a current in the external circuit. A schema tic of the basic operation is shown in figu re 2. 9. Figure 2 .9 . A p - n junction acting as a radiation detector [12] The probability of thermal generation electron - hole pair P(T) is expressed as 20 w here T is absolute temperature, E g is the band gap of an intrinsic semiconductor, k is the Boltzmann constant and C is a proportionality constant characteristic of the material, k T= 0.025 eV at room temperature. Comparing between two popular semicond uctors, Si (E g = 1.1 eV) and Ge (E g = 0.67 eV), it is quite essential that Ge based semicond uctor detectors require cooling to minimize the dark current`s effect due to its small band gap which is not necessarily required for Si. W ide bandgap semiconductor s like diamond with E g = 5.45 eV not only need no cooling , it has a very low intrinsically c arrier concentration which offers low dark current without forming a pn junction. For Si based detector, a junction formation is necessary to create a large resisti ve path. So s ingle crystal diamonds for radiation beam detectors act s like a solid - state ion ization chamber. For diamond semiconductors, a simple parallel plate capacitance structures can be used for radiation beam detectors. 2 .4 . Interaction o f r adiatio n b eams with m aterial The detection mechanism in any type of detector relies on the interaction (and interaction volume) of the radiation beam with the material used in the detector [14] . When a beam of radiation penetrates into a material either part of it or the full beam might pass through the material, with some possible scattering. Therefo re, in all practical events the interaction of the beam with the material results into partial translation of energy into the electrons or the nuclei of the atoms. Generally, charge particles interact with host/absorber material through electromagnetic int eraction, whereas for neutral particles the nuclear interaction predominates which fur ther generates charge particles. The electromagnetic interaction between a charged particle beams and host atoms comprises of two steps (i) excitation and ionization o f atoms and (ii) bremsstrahlung [15] , [16] . 21 The bremsstrahlung process is mostly associated with radiation loss es for electrons, as the interaction is inversely proportional to squared mass of incident particle. Each charged particle goes through several events of scattering and d eflection at each interaction before it comes to rest. The electromagnetic interaction causes transfer of energy even without any direct collision, i.e. even the charged particles can interact with the atoms while passing through its vicinity. The result o f the transfer of energy would lead to either raising the atom to higher energy level, i.e. excitation or removal of an electron from its parent atom by ionization. For interaction of heavy ions, beside excitation and ionization, other phenomena like redis tribution of charges, creation of secondary electrons and occasionally some structural modifications of the host material can occur . The secondary electrons are only generated if the primary electrons (generated from ionization) have enough energy to furth er knock out electrons from atom. The secondary electrons are also referred as - ray. An important parameter, which helps to characterize the interaction of charge particle beam , is energy loss per unit length or stopping power [17] of the material. The stopping power depends on both the incident particle and the target material. The major contribution to stopping power arises due to electronic and nuclear interactio ns. Hence the total stopping power (S) is the sum of the electron stopping power and nuclear stopping power. The negative sign indicates the drop of energy of the particle with increase in penetration length in the material. A detailed expression of the energy (E) loss over a distance (x) for the charged particle passing through a material is expressed by Bethe - Bloch distribution [18] , [19] given as; (a) 22 where the terms are defined as r e : classical electron radius = 2.817 10 - 13 cm. m e : electron mass, N a: Avogadro`s number = 6.022 10 - 23 mol - 1 I: mean excitation potential., Z: atomic number of absorbing materials A: atomic speed of particle with respect to c and W max : maximum energy transfer in a single collision. For si mplicity of discussion the Bethe - Block distribution can be qualitatively expressed as; Hence the particle velocity is a prime parameter for the energy loss distribution, i.e. for non - relativistic regime, dE/dx varies inversely with kinetic energy. This c an be interpreted as, that at low particle velocity the cha rge particles have a higher chance to stay near the electrons and therefore higher is the transfer of energy. On the other hand, for different particles moving at the same velocity the charge state z will be the factor for energy loss. The energy loss diag ram for different charged particle versus kinetic energy with Bethe - B loch distribution is shown in figure 2 - particles loose energy at much higher rate than proton s at the same kinetic energy level due to its charge state. The distribution also shows that for some of the particles above (c) ( b ) 23 several hundreds of MeV energy range the stoppin g power reaches a near minimum value. At this con dition these particles act as " minimum ioni zing particles" (MIP). The electrons also act the same way at an energy level close to or lower than 1 MeV. Figure 2 .10 . Energy loss diagram of different partic les [5] Figure 2.11 shows that for a given particle with specific charge, the rate of energy loss has a similar value for different materials except hydrogen. However, there is a marginal drop happening in the rate of energy loss with increase in the host m aterials atomic number Z. Other than hydrogen, the minimum ionizing energy has minimal variation, ranging between 1 - 2 MeV/g.cm 2 with change in Z from 7 to 100. Beyond minimum ionizing point, with increase in ener gy the energy loss rises slowly [20] . 24 Figure 2 .11. Mean e nergy loss rates fo r particles in different medium [19] As a heavy charged particle moving with low energy (~MeV) goes deeper inside a material it further loses energy and stopping power changes accordingly. As stopping power is a measurement of ionization capability of the particles inside the material, any ch ange in stopping power is equivalent to alteration of ionization capability. The following simplified expression of Bethe - - particles inside a material. where = 0.30548 is a constant for a given material and I is considered as 10 - 4 MeV, a typical number for low Z materials. The corresponding plot is shown in figure 2 .12. ( e ) 25 Figure 2.12. Stopping power vs Energy [16] As seen, with the gradual increase in stopping power, the ionization increases until it reaches to a maximum point known as Bragg peak [21] . Beyond this point the particle sharply loses all the energy and becomes neutralized. This overall phenomena (plot) is known as Bragg curve. For heavy charged particles, there appears a point where the particles do not have suffici ent energy for ionization and the loss of energy mainly happens due to nuclear collisions. In an overall situation the electronic stopping dominates the particles interaction with material as long as the particle velocity is in the rela tivistic range. In c ase of electronic stopping the particle beam nearly follows a straight trajectory with minimum deflection. Based upon its initial energy the particles gradually loose energy and go through an intermediate stopping. In this range the cha rged particles pick up electrons. The capture of higher orbital electrons continues until the ions become neutral. 26 The beam trajectory goes through some directional change and slows down further. In the final range of stopping, the particles reach low ioni zation state or almo st a neutral state. One of the important phenomena associated with the interaction of particles is that the energy loss process is statistical in nature. An expansion of the energy curve always happens after charged particles pass thro ugh a particular thi ckness of the absorber. The width of the energy distribution curve is a measurement of energy straggling which varies along the distance travelled by the particle. Figure 2 .13 shows, how with the penetrating distance the energy profile of a monoenergetic p articles gets skewed as an effect of energy straggling. Figure 2. 13. Energy distribution of a monoenergetic beam varying over penetration depth [19] As an e ffect of energy stragglin g the Bragg curve (in figure 2 .13) will slightly get deformed in the end with a small tail. 27 Another important parameter related to particle interaction is the range or the average distance a particle beam can travel in an absorber . This could be computed by integrating the average power over the entire energy spectrum of the incident particle, The range R(T) shows some statistical fluctuation around a mean value as it goes throug h multiple coulomb scattering. The fluctuation associated with the range is known as range straggling. The end result of diff erent types of interaction of radiation with matter can be radiation damage. Some radiation damage can be partially healed but f or some damages the material is permanently degraded. So, the different scenarios can be described by different types of defects [22] . (i) Impurities are caused by radiation w hen a nuclei transmutes to another radioactive nuclei. Neutron and ion irradiation can generate new radioactive species. Neutron capture by a nucleus can produce an isotope different than the original detector material and the new isotope may decay. This c an often dis tort the chemical composition of the detector material. Ion irradiation causes an impurity to be added to the detector material and it can also cause new radioactive species to be created by interacting with the detector material atomic nucleus . These impu rities can slowly and gradually affect the electrical and material properties of the materials. (ii) Displacement damage takes place due to transfer of kinetic energy from radiation to atomic displacement in the detector material. An elastic co llision can displace an atom from its lattice position which is a primary knock on and triggers multiple secondary atomic displacements. The displaced atoms are referred as interstitials and the void position is known as vacancies. Together, a Frenkel pai r is formed. Displacement damage is a clear evidence of nuclear interaction. Vacancies and interstitials can ( f ) 28 become mobile at sufficiently high temperature and therefore occasionally a high temperature annealing process can initiate recombination of both a nd hence hel p to reduce the number of defects (radiation damage) accordingly. 2 .5 . Current s tate o f a rt o f diamond - based radiation detectors Diamonds inherent properties like chemical inertness, stability and operational capability at high temperatures and extreme radiation resilience make it a very useful material at extreme environments. Historically, the high cost of mining and processing of natural diamonds restricted tion of synt hetic diamond was reported. The very first kind of synthetic diamonds were grown by a technique of processing graphite at high pressure and high temperature and eventually converting it to diamond, better known as HPHT diamonds. The next couple of decades had seen some other different approaches to grow synthetic diamond which finally resulted to the invention of chemical vapor deposition (CVD) technique which has gone through further improvement and refinement. Currently microwave plasma assist ed chemical vapor deposition (MPACVD) [23] stands as a very efficient process to grow high quality diamond. Industrial applications of diamond range beyond its traditional use as an abrasive to include the fields of optics, RF and microwave transmission tec hnology, hea t sinks, and electronics. [ 24] An impo rtant material property of diamond is the high displacement energy (~ 45 eV/atom), much higher than of other contemporary semiconductors. (e.g. 20 eV for Ge, 2 5 eV for Si and 9.9 eV for GaAs ) [ 22, 25] . This property makes diamond an attractive material for radiation detectors in the fields of nuclear and high energy physics. Initially, diamond - based detectors had to face steep competition with silicon - based particle detectors due to the optimized product quality of Si, but 29 with steady improvement in the quality of synthetic diamond the current scenario appears to be very promising for diamond. R ecently, clinical radiotherapy applied diamond for base dosimeters [26] because diamond offers fast response wi th a very low volume of deposited dose and because to human tissue equivalence (atomic nu mber Z=6 is close to Z eff of biological tissues) it requires less correction to the signal read. Ciancaglioni et al . discussed the performance o f a metal - semiconductor - metal structure made of polycrystalline diamond (PCD) [27] , which was applied for x - ray dosimeter. The comparison between diamond and silicon P CD based dosimeter s show diamond to be more suitable compared to silicon - based devices. Single crystal diamonds are also used for high energy X - ray detection [28] in synch rotron radiation sources. These detectors show good linearity with respect to X - ray photon flux. Single crystal diamond has also been tried for successful application as a photodiode for X - ray radiometry [29] . The response of the photodiode improves by applying a pulsed bias to minimize the effect of tr apping and de - trapping of charges. In another field of application, photo current and spectral response of diamond - based photoconductors, built of high quality single crystal diamond , have been studied in the deep UV range [30] . A m ultilayer diamond detection system was used for monitoring space radiation to ich measures cosmic flux to decide safe operating conditions of spacecrafts. In the field of high energy physic s, diamond - based radiation detectors are making a breakthrough by upgrading silicon - based detection technology because diamond offers higher rad iation hardness. This however took some close collaboration between particle physicist and manufacturers of synt hetic diamond to bring the state of art to a very advanced and promising level. It is expected that in the high - energy regime of hundreds of MeV /nucleon to few 30 GeV/nucleon range diamond will be more radiation tolerant and will survive longer than existing silicon - based detectors. In a comparative study between diamond and silicon irradiated with protons and neutrons the diamond material ranged fro m slightly more radiation hard at lower energies to significantly more radiation hard at higher energies above 0 .1 GeV [31] . The RD42 collabor ation program of CERN has conducted a comprehensive study on the performance of diamond detectors in radiation harsh environments. The major mot ivation was to establish detectors which can survive for many years for ATLAS and CMS experiments at the LHC (La rge Hadron Collider) [ 32] . The current planned upgrades in LHC to High Luminosity (HL - LHC) will increase the fluence to the inner mos t tracking detector to 2 x 10 16 /cm 2 . A report from this group a few years back, discussed the response of scCVD (single crystal CVD) and pCVD (p olycrystalline CVD) based diamond detectors up to 1.8 10 16 protons/cm 2 , which showed great potential for diamond base detectors to survive this type of radiation harshness. Two decades ago, poly - crystalline diamonds (PCD) were used with pion and proton b eams in order to study radiation tolerance of diamond while in use as tracking detectors by the RD42 group. Polycrystalline diamond substrates suffer from non - uniform charge collection due to the presence of grain boundaries. As a result, the quest for an improvement in charge collection distance continued and a study on polycrystalline diamond detector perform ance was reported from the RD42 collaboration group [33] . Proton irradiation on strip trackers made of polycrystalline diamond were shown to retain 25% of their initial signal stren gth after exposed to a flux of 20 10 15 p/cm 2 [34,35] . However, with the advent in microwave plasma - assisted chemical vapor deposition (MPACVD), single crystal diamonds became more readily available. Single crystal diamonds ( scd ) offer higher charge collection dist ances, and nearly 100% charge collection efficiency for samples with very low impurities, and superior spec troscopic 31 properties and time resolution [36] . In one of the pioneering studies, the RD42 group compared the radiation hardness of both PCD and SCD with proton irradiation . The SCD di amond strip detector responded with a higher pulse height as compared to PCD diamond. However, the most int eresting observation was the degradation of the charge collection distance (CCD) with particle fluence. With a required shift in the SCD data point i t was found that both types of diamond followed the same nature of damage (i.e. assigned same damage consta nt on the fitting curve). The study of radiation hardness in both types of diamond continued wit h different proton energy beams [37] . As an int eresting outcome, the damage constant was found to marginally drop with proton energy increased from 25 MeV proton to 24 GeV proton beams. In this project we are studying the radiation tolerance of single crystal diamond in swift heavy ion (SHI) beams. SH I beams are ion beams from elements with heavy nuclei accelerated to relativistic energies. Future rare iso tope facilities will provide particle beams of heavy ions with unprecedented intensities with particle rates of up to 10 8 particles per second. Diamo nd is one of the few materials for tracking and timing detectors that is expected to withstand such beams w ith reasonably useful lifetimes. A few reports have shown promising outcomes of the interaction of diamond with SHI beams [38,39] . The radiation damage for relativistic Au beams was studied by Pietrasz ko et al. 35 for fluences of ~10 14 cm - 2 . Gracia et al. looked into possible amorphization of diamond samples irradiated up to 40 MeV beam energy [40] . Even at a very high electronic stopping power range, the authors have found that nuclear stopp ing power created the major effect in the diamond. In their study the authors showed that the surface damag e increases proportionally with nuclear stopping power. In another report, a single crystal diamond - based detector (for time of flight study) had bee n irradiated by 197 Au beam. After collecting nearly 10 14 /cm 2 Au ions the 32 pulse height measurement by ion b eam induced current (IBIC) [41] showed a drop of almost 5.1 times in the area of radiation damage. In this work SHI beams of Sn and Zr were used to irradiate diamond detectors and study the degradation at a similar fluence range. Both lab orator y grown and co mmercially available single crystal diamond grown by the MPACVD process were used to construct detectors for SHI beams measurements and their performance are compared. Samples were irradiated at the National Superconducting Cyclotron Laborator y (NSCL) at M SU. 33 For our experiments, lab grown single crystal diamond substrates produced in MSU MPA CVD reactors were cut and polished to fabricate detectors. In addition, electronic grade co mmercial diam ond s ubstrates were also used to build detectors and subsequently irradiate in the same kind of beam. The transient current technique (TCT) [42 44] was measured to quantif y the charge transport mechanism, charge collection efficiency and carrier lifetime in the irradiated part of the SCD crystal and compared to a lightly irradiated and non - irradiated SCD diamond substrate . In addition, UV - VIS spectroscopy, FTIR, X - ray diffr action, Raman spectroscopy was also used to understand the post irradiation effe ct on the diamond substrates. 3.1. Fabrication of radiation detectors In this study two different types of diamond plates were utilized to fabricate detectors. The first batch of diamond samples were grown and processed at Michigan State University (MSU) by the MPACVD technique. Single crystal diamonds were grown on commercially available high pressure high temperature (HPHT) seeds, at 180 Torr pressure for 48 hours with f eed gases of 400 sccm H 2 and 20 sccm CH 4 in a deposition reactor operating at 2.45 GHz microwave power of 2.5 kWatts [45] . After growth, the samples were laser cut fr om the seed, mechanically polished and cleaned in boiling acid (at 350 o C in H 2 SO 4 /HNO 3 (1:1) for 30 minutes followed by HCl for 20 minutes) to prepare 3.5 mm 3.5 mm samples with thickness of 0.6 ~ 0.7 mm. The other batch of samples, 3.0 mm x 3.0 mm with 0.5 mm thickness substrates of electronic grade , were purchased from a commercial vendor and cleaned using the same procedure. 34 Any further residues were removed by cleaning the samples using acetone and methanol in an ultrason ic bath. The surfaces (both t op and back) of the detectors were next exposed to oxygen plasma for 2 minutes to terminate with oxygen in order to reduce any surface conduction. Figure 3.1. Top surface of the single crystal CV D diamond plate with electrodes After cleaning and oxygen termination, 150 Å of Ti and 1500 Å of Au were deposited on both sides of the diamond plates by thermal evaporation at room temperature. The top surface had two rectangular electrodes (2 mm × 1.1 mm) and the bottom surf ace had one square (2 mm × 2 mm) ele ctrode. Post evaporation, the samples were annealed in vacuum at 600 o C for 2 hours. This process helps Ti to form a carbide, which results in good adhesion and formation of ohmic contacts. Finally, gold - plated aluminum wires were attached to the electrode s using silver paint, connecting them through transmission lines on a printed circuit board to a current amp lifier with coaxial cables. Figure 3.1 shows the top surface (surface which faces the beam) of the diamond detec tor, wired to external cables. 35 The d iamond plates with electrodes on both sides, were placed into custom made mounts. The detector mount was placed inside a n ultra - high vacuum chamber in the beam`s trajectory. The diamond electrodes were connected to a curr ent - sensitive low - noise amplifier [46] with 2.3 GHz bandwidth and +38 dB gain, through coaxial cables. The output of the amplifier was recorded at a digital oscilloscope. A schematic of the basic operation of the detector was shown earlier in figure 2.3 of chapter 2. A third diamond sample was used in this experiment for comparison as a reference sample. This electronic grade 3.0 × 3.0 × 0.5 mm 3 diamon d sample was purchased from a differ ent vendor. Square electrodes (2 mm × 2 mm) were deposited on both sides of this diamond for transient current measurements. This particular sample was never exposed to the radiation beam; hence it retained its pristine nature. 3.2 . Irradiation o f d iamon d d etectors by Swift Heavy Ion (SHI) b eams As mentioned earlier, the detectors were made of two different kind of diamond sources, one batch from diamond grown at MSU (samples are named as GYJ 148A, GYJ 148C) with nitro gen impurity level above few 100 ppb , and the other batch was from commercially available electronic grade diamond (commercial diamond - 2) with low nitrogen impurity. To understand the detection process and gradual degradation, the detectors were exposed to beams of 124 Sn and 36 Figure 3.2. Signal capture d from lab grown SCD detector (GYJ - 148C) irradiated with 96 Zr beam. (The upper signal represents the response of heavily irradiated segment and the lower signal shows response of lightly irradiated segment ) 96 Zr ions with an energy of 120 MeV/ u accelerated at the Coupled Cyclotron Facility at NSCL, MSU. The intensity of the beam was varied from 10 2 particles/s/mm 2 to 10 6 particles/s/mm 2 by suitably changing attenuator foils in the injection line of the cyclotron. The range of 124 Sn ion beam is 1.95 mm for diamond and for 96 Zr ions, it is 2.25 mm. The approximate thickness of the lab grown samples are 0.7 mm. The energy loss is 3802 MeV for 124 Sn beam with 89.31 MeV/u remaining energy and 2395 MeV for 96 Zr beam with energy left as 95.02 MeV/u. T he signal from a detector made of MSU - grown diamond (GYJ - 148C) irradiated by the 96 Zr beam, captured at the oscilloscope, is shown in figure 3.2 . The maximum applied bias field for lab grown diamonds (GYJ 148A & G However, for the commercial diamond, to avoid any saturation by the preamplifier, the maximum 3.2 , the signal from the heavily irradiated segment is show n in channel - 1 and the lightly irradiated p art at channel - 2. This shows an ensemble of multiple traces (where each trace is generated from a single particle) with a pulse width of ~ 2 37 ns. The amplitude in the channel - 1 (with most probable charge generation , in the dark band) dropped partially as co mpared to channel - 2 due to the onset of degradation from the damage created by the beam in the heavily irradiated segment. The collection of screenshots of the oscilloscope at a different time (i.e. after a certa in fluence) is shown in figure 3.3 . The gra ph measures the signal amplitude at diffe rent applied bias voltages to estimate whether any change in the amplitude has happened. Figure 3.3. The output voltage from lightly irradiated (dark blue) and heavily irradiated (cyan) after a certain particle fluence (2.77 10 12 particles/cm 2 ) for different bias voltages The maximum signal amplitude for both lightly irradiated (dark blue) and heavily irradiated (cyan) and their ratio are used as a corresponding figure of merit to estimate degradation, which w ill be discussed in the next section. 38 3.3. Results of beam i rradiation on diamond detectors Figure 3.4 shows signal variation versus applied detector bias at four different fluence levels for lightly irradiated and heavily irradiated segments. As seen i n the bias scan - 4, the red curve representing the signal from heavily irradiated segment, falls off by a large margin from the signal of the lightly irradiated curve. Figure 3.4. The output voltage measured at different time (with a range of bias voltag es) during irradiation from heavily irradiated segment (r ed curve ) vs lightly irradiated segment (black curve) A ratio of the signal str engths (i.e. output voltage) , a s calculated at the highest bias voltage to define the relative signal, is used to create the degradation curve shown in figure 3.5 . 39 Figure 3.5. Relative signal drop of detectors over particle fluence The relative drop irradiated segment as compared to the lightly irradiated s egment is plotted as a function of particle fluence in order to understand the degradation of the detectors. This plot is built on th e assumption that the amplitude of the output signal from the detectors (made from diamond plates of nearly same thickness but different quality) reflects the charge generated in the material, and therefore the drop of signal is directly caused by the rise of defects/trapping inside the volume exposed to the energetic beam. Figure 3.5 shows the relative signal (%) drop in the heavily irradiated segment for the two MSU - grown diamond samples (GYJ - 148A & GYJ - 148C) and the commercial diamond - 2. The detector mad e of the MSU - grown diamond (GYJ - 148A) was irradiated by a 124 Sn beam, whereas the other two detectors made from MSU - grown diamond (GYJ - 148C) and commercial diamond - 2 were irradiated by 96 Zr. For the detectors fabricated from MSU - grown diamonds the signal dropped to near 55% after the same amount of fluence. A parametric equation [47] has been used to describe the scenario (f) 40 as; where A 0 ents the particle fluence the detectors are impinged with, parameter k is the damage constant and c is the offset. The hyperbolic fit (shown as dotted line in figure 3.5 ) appears as a good fit on the experimental points with the damage constant k as 0.88 10 - 14 V - 1 cm 2 . In the study of radiation hardness of diamond detectors by RD42 collaboration group, the distribution of charge collection distance (CCD) was measured with particle fluence for different proton beam energy. With a parametric equation of the same kind, the damage constant was reported as k (by RD42 for protons) ~10 - 18 - 1 cm 2 for both polycrystalline and scCVD. 41 As discussed earlier, the swift heavy ion beams initially loose energy due to ionization loss. However, bey ond a certain point the beams gradually reach an energy regime where most of the energy goes into non - ionizing energy loss and eventually end up of w ithin a narrow range of depth causing atomic displacement. A single incident particle can initiate a chain of primary and secondary interactions and end up creating defects (vacancies, interstitials, Frenkel pairs) or otherwise adding strain to the materia l. In many cases the diamond sample itself has some traps already present at its pristine stage and the hea vy ion beam irradiation triggers formation of further cascade of damages inside. Inside the atomic structure, some of the sp 3 bonds may reshape to a stable form of sp 2 bonds. Therefore, this change in the structure (both physically and chemically) disturb s the electronic properties of the diamond. Impurities can also form inside a material when the incident ion picks up the necessary number of electro ns to become neutralized and therefore settling as a foreign species inside the host structure. The effect of radiation induced damage can be investigated by using the transient current technique ( TCT ) , which measures the charge transport properties of se miconducting material. It is also worth looking at the extent of damage in the material and at the interfac e of the electrodes by measuring the leakage current. In general diamond shows extraordinarily low leakage current compared to silicon based devices. One of the major indication s of radiation damage in silicon based detectors is the increase in leakage cur rent. Diamond can maintain very low leakage current even after significant radiation induced damage. 42 4.1. Electrical c haracterization This sectio n presents the electrical characterization techniques to determine the effect of the beam irradiation on ch arge transport properties of the single crystal diamond. 4 .1.1. Transient c urrent t echnique The transient current technique (TCT) is used to measure the charge transport properties of pristine and radiation beam irradiated semiconducting samples. A meta l - semiconductor - metal capacitive structure is configured (the same used for the detector) to ca rry out this experiment. The TCT technique is a measurement of the current generated due to drift of free carriers under the influence of an external electric fi eld. In this experiment, ionized particle generated free charges travel from one side of the sa mple to the other side under the influence of an external electric field. The shape of the current pulse depends solely on the properties of the material under t est. Important parameters for particle detectors, like drift velocity, carrier mobility, carrie r lifetime and field distribution can be analyzed from this measurement. Once hole and electron charges are created near one side (electrode) of the sample eithe r an electron or hole charge cloud , based upon the polarity of the applied field , will drift du e to the electric field to the other electrode. A rectangular signal versus time is generated during the time it takes for the charge cloud to reach the far thes t away electrode when minimal trapping occurs. The nature of the current signal can be expresse d as [43] The first term in the exponent relates to the contribution from the effective space charge and the (g) 43 second term represents the exponential decrease of current due to charge trapping. In a diamond sample, with significa nt charge trapping, the current signal is described as where is the initial generated charge, is the sample thickness, is the drift velocity and is the effective trapping time o f respective carriers. The trapping centers inside causes the amp litude to drop exponentially. Any non - uniform space charge distribution will lead to a non - uniform field distribution within the sample. Figure 4.1 shows the expected form of transient curren t shapes of different qualities of diamond [36] . As seen in figure 4.1 the generated charge reduces exponentially for a pcd sample or samples where there is already substantial traps present. The flat top signal represents drift of charges in very hi gh - quality diamond with uniform electric field. 6 Figure 4.1. Transient current signal shape for different quality of diamond [36] ( g ) 44 Figure 4.2. Transient current signal for electron and hole s from a single crystal diamond [36] The shape of the signal v ersus time partly depends on the type hole - electron pair excitation source used to ionized charges at one side of the sample, e.g . for laser induced charges the field will get modified by the high density of charges leading to a space - charge - limited curren t flow regi - particle, the low density of charges ensures that TCT signals are generated in space - charge - free regime. However, the quality of the diamond (defects, impurities) is more crucial in guiding the shapes of TCT signals. Figure 4.2 shows the transient current signal at different electric field for both carrier types i n a high - quality single c rystal diamond [36] . In the current work, a 232 U - particle source of 5.4 MeV energy was collimated through a 1 mm diameter hole on top of selected sections of the irradiated and pristine detectors. There was a 13 mm air gap between the source and the sample, which amounted to an energy loss of 1.28 MeV in the air. The - particles penetrate 8 much smaller distance compared to the sample thicknesses (530 of 6.3 10 2 /s/cm 2 and the responses from the detector were collected in a short duration o f 5~7 minutes time to minimize polarization effect. The e lectrodes of the detector were connected to a 45 broadband current amplifier (Cividec Instrumentation: C2HV0081, 2GHz/40dB) and single traces were stored with a Tektronix (TDS 3054B) oscilloscope. To u nderstand the transient current signal in the current set up, a commercially available electronic grade single crystal diamond was tested at first. This particular diamond was designated as commercial SCD - 1 and held at its pristine stage in th e entire study , i.e. this sample was never irradiated with any ion beam. This sample was used as a reference sample for the entire study. The measured voltage V m (t) was converted to a transient current output i m (t) by the following equation ( i ) The major uncertainty in th e measurement appears from the two parameters A amp and R in , which were calibrated with the output of a charge - sensitive amplifier (Cividec Cx - L, with sensitivity 12 mV/fc). The calibration values have been used to convert for all the samples currently under study. Figure 4. 3. Transient current signals from commercial diamond - 1 ( i ) 46 As observed in figure 4.3 , for both type of charges the amplitude drops gradually during the dr ift motion. This could either happen as the diamond sample may have im purities causing charge trapping or there could be strain present in the sample causing polarization of the field. A birefringence study of this pristine sample has shown reasonable str ain present in the sample. Holes get trapped more in the sample than e lectron s as shown by the amplitude of the hole current decreasing faster versus time. The initial current amplitude (at t=0+) depend s on the amount of energy that was deposited, and the equivalent charge generated in the sample, the applied external field and the mobility of the charge carriers. TCT studies were also done on an ir radiated commercial substrate ( 96 Zr beam of energy 120MeV/u) of slightly different type (higher nitr ogen impu rity than commercial diamond - 1 but lower nitrogen than MSU grown samples ). As mentioned in the substrate preparation section above, the substrate was prepared by depositing two rectangular electrodes on top of the substrate, one electrode section was wher e the swift heavy ion beam was directly focused on and named as heavily irradiated segment, and the other section is only irradiated by the less intense halo of the approximately Gaussian shaped beam profile and been termed as lightly irradiated s egment. T he direct beam is expected to induce more defects in the area of the heavily irradiated segment and therefore disturbing the charge transport more in that segment. The next two figures give examples of the transient current signal of the two segme nts of co mmercial SCD - 2. 47 Figure 4.4. Transient current signal from lightly irradiated segment of commercial diamond - 2 Figure 4.5. Transient current signal from heavily irradiated segment of commercial diamond - 2 As observed, comparing figures 4.4 and 4.5 the transient current signals of the more heavily irradiated diamond region are narrow er in nature and have a sharper exponential decreas e in amplitude versus time . Specifically, in figure 4.5 , which shows the eff ect of direct beam damage (heavily irradiated), the irradiation causes more trapping centers or defects to be produced yielding a narrower pulse signal, i.e. the drop of approximately 50% of the amplitude happens more rapidly 48 in the heavily irradiated segm ent as compared to the lightly irradiated one. This indicate s that in the heavily irradiated segment the damage is d ue to the beam irradiation. The beam induced damage also shows slightly smaller amplitude for both electron and hole generated currents at time t=0 as compared to the lightly irradiated s egment. The effect of the beam induced damage c an be further quantified by calculating the collected charges (Q col ) of the different segments of the commercial diamond - 2. The collected charge is f ound by defining a suitable time (by using figure 4.3, 4.4 a nd 4.5 ) and integrating the current pulses over time as g iven in equation (j) where t start and t end are defined as the time the current signal rises above and drops below a threshold level r espectively. The calculated charge by the above integral is plotted with respect to the applied field as shown in figure 4.6. Figure 4.6. Charge collection versus applied field ( j ) 49 The overall charge collection was calibrated by using a charge sensitive amp lifier (Cividec Cx - L, CxL0051). For the co mmercial diamond - 1, charges saturated beyond an applied field ± 0.5 collection was at lower values due to the presence of traps. For calibration, the charge sensiti ve amplifier was used at higher fields for commercial diamond - 1. The collected charge was recorded at 0.60 a number of measurements. The gain of the amplifier was separately measured to estimate the A amp. Further with the collec ted charge value, the product (A amp. R in ) w as adjusted and applied to convert the measured voltage to current i m. for the generated signal from all samples . The degradation created by the swift heavy ion beam is quite evident as the charge collection in the heavily irradiated segment was less than 20 fC (within the applied field range) as compared to lightly irradiated segment where the collected charge is close to 30 ~ 35 fC. The ion beam deposits its energy in the material, which will primarily ionize and create electron - hole pairs. However, some the SHI deposited energy will initiate dislocation of atoms from the lattice. The dislocated atoms (both carbon and especially any i mpurities) can result in traps being created. Finally, a comparison of transient signal generated from non - irradiated segme nts of MSU and commercially available electronic grade is shown in figure s 4.7 and 4.8. As it is evident that the charge drift time s are extremely short for MSU lab grown diamonds (less than 2 nano second).in comparison to the commercially available elect ronic grade diamonds. This clearly establishes the facts that the lab grown diamonds have more charge traps arising from excess nitr ogen - related traps or other defects. 50 Figure 4.7. Transient signal generat ed from a MSU lab grown diamond Figure 4.8. Transient current signal generated from a non - irradiated comm ercial electronic grade diamond 51 4 .1.2. Leakage c urrent t est As a wide bandgap material, intrinsic diamond also shows very high room temperature resistivity. In addition, the energy required to c reate free carrier is also very high, hence altogether diamond acts nearly like an insulator in its undoped condition. Hen ce in general the leakage current associated with a good quality of single crystal diamond material is very low ( < 10 - 12 A). The leakag e current characterization of a particular diamond sample becomes useful in finding a safe biasing range where noise level stays within a certain lower limit. The leakage current can be generated both due to the bulk properties as well as the surface` cond ition. The leakage current from the bulk arises due to thermally activated carriers. These carriers can generate from trap levels inside or occasionally from defects that penetrates deep into the substrate. Hence a low leakage current, which is signature o f a good quality of diamond is desirable for its usage for any electronic applications. Figure 4.9 shows an I - V characteri stic of a pixel polycrystalline detector [48] . Polycrystalline diamonds are used as beam loss monitors in accelerators, hence a relatively high leaka ge current still makes it quite applicable. However, in some cases polycrystalline diamonds show erratic distribution or r ise in leakage current, eventually settling to a very high leakage current. Occasionally an applied magnetic field can alter this beha vior as studied by Muller et al [49] . 52 Figure 4.9. Leakage current in a polycryst alline pixel detector [ 44] In this experiment a Keithley 6485 picoammeter is used to measure the leakage current for both the lightly and heavily irradiated segments of the commercial diamond - 2. The externally ap plied voltage was varied, and the maximum was limited to 550 V. The leakage current test was conducted in the commercial diamond - 2, in both the lightly irradiated and heavily irradiated segment. In both cases the same contacts (Au and Ti ohmic) used durin g beam radiation test, were used to carry out the test . The follo wing figures show the nature of leakage current`s distribution collected at different stages from the two segments. 53 Figure 4.10. Leakage current measurement on commercial diamond - 2 Figure 4.11. Repeat run of leakage current measu rement (a & b) with segments measured in differe nt orders In the first set (figures 4.10 & 4.11 (a)) the leakage current of heavily irradiated segment (noted in figure as Irr) was measured, followed by measur ement of the lightly irradiated segme nt ( or non - irradiated segment, noted as NIrr) where as in the other set ( figure 4.11 (b), the order was reversed, i.e. the lightly irradiated segment was looked into before the other segment. In all cases (a) ( b ) 54 the heavily i rradiated segment showed marginally hig her leakage current after the increase in current for the positive polarity, which only appeared in one polarity of the field. Also, in the second set of data the current rises to an order less than it happened in the first set, which might have happened d ue to light in the background. A next set of experiment (fig 4.12 ) , where the leakage current test in the heavily irradiated segment was carried out first shows almost no difference (other than the current in lightly irradiated segment between 300v - 350 v i s higher compared to its other part) between the distribution of the two segments. In all cases the I - V curve shows very similar nature to leakage current distribution of a polycrystalline diamond (as shown in figure 4.9 ). The commercial diamond - 2 has high impurities and trapping centers, which could release free charges over a certain high field, which eventually shows a sharp rise of current. A separate measurement has been done to verify whether the rise in the curr ent is due to any transient effect . In multiple trials of test the leakage current show similar distribution i.e. a polarity dependent signature of a rise. Also, the leakage current in all the tests stay ed within the range of 10 - 12 A, showing the quality o f the commercial diamond - 2 sample after irradiation. There was no degradation of the contacts observed under optical microscope. To have a better understanding in the rise of the current, the leakage currents were measured in the presence of strong light on the sample. The background was turne d dark and a red LED was used to illuminate the samples for a short time. Figure 4.12 shows the leakage current measured while the sample is illuminated. 55 Figure 4.12. Leakage current measurement with sample illuminated with red LED A comparison of the current axis among the two figures ( 4. 1 1 & 4. 1 2 ) clearly shows that some traps get activated by the red LED. These are shallow level traps which are located within 1.5 eV of the conduction band and get activated with the applied field in presence of the illumination. As it is found that the field required for the activation and the total rise is nearly the same for both lightly irradiated and heavily irradiated segments of the diamond. In an extension of this study on the transient cur rent response, the rise of current at three different high fields was studied to understand whether the spike of the current drops back after some time or not. To realize this the applied vol tage was increased to three different high values above 300 V (as it is seen that the current rapidly increases after 300 V). The following figures (figures 4.13 & 4.14 ) show the rise and effective time taken by the leakage current component to drop at thr ee different voltages at 365 V, 425 V and 465 V respectively , aft er the light (LED) was turned off. 56 Figure 4.13. Transient response of heavily irradiated s egments of commercial diamond 2 Figure 4.14. Transient response of lightly irradiated segments of commercial diamond - 2 57 Figures 4.13 and 4.14 reveal that the transient component of the leakage current slowly fades to almost zero after some time. However, for heavily irradiated segment it takes slightly shorter time for the trans ient component to decay as compared to the lightly irradiated segment, which might b e due to presence of additional defects in the heavily irradiated segment caused by the beam. Therefore, this experiment shows how heavily irradiated segment responses to v isible light activated traps and the typical time required to normalize. 4 .2. Optic al c haracterization A pristine quality single crystal diamond is an optically transparent material throughout the visible range and throughout large portions of the infrar ed region. However, any impurities both deliberately or unintentionally added affect the optical properties, which therefore makes it a very interesting technique to study the quality of diamond samples. Nitrogen is the most familiar impurity present in di amond and based upon the quantity of nitrogen the diamond can appear as pale yellow or brown in color. Hence different forms of nitrogen impurities could be studied by looking at absorption peaks in the visible and infrared range. 4. 2.1. UV - Vis s pectroscopy In UV - VIS spectroscopy optical properties of materials (e.g. absorbance, transmi ttance) are measured w ithin the range of 200 - 400 nm ( uv) and 400 - 800 nm (visible) light. This method helps both in qualitative and quantitative analysis of solids and liquids (or compounds present in liquid). The basis of this method relies on Beer Lambert `s law [50] for analyzing the absorbance of a material, which is a unique characteristic for any material under test. One can derive the 58 concentration of a molecule in a solution from the absorbance, however this has limit ation for higher concentrations. Diamond is transparent in the near UV to visible region at waveleng ths longer than its strong absorption peak at 225 nm (due to its indirect band gap). Diamond also has absorption peaks in infrared region 2500 - 6000 nm. [2,51] . For any solid material a rough surface adds up as a source of scattering and therefore needs to be accommodated for as a correction to figure out the transmittance or absorbance. Therefore, an improper surface finishing can cause a significant signal loss and therefore end up showing a lower transmittance. This is also an issue related to transmi ttance of polycrystalline diamond, where light scatters at grain boundaries. Internal strain can als o cause higher absorbance, which could be partially reduced by high temperature annealing. Otherwise a high - quality type - IIa diamond with a smooth polished surface shows high transmittance beyond the characteristic cut - off wavelength of 22 5 nm. In this stu dy a Perkin - Elmer UV - VIS spectrometer (LAMBDA - 900) is used for measuring the optical properties of different diamond samples. A slit width of 3 - 5 mm has bee n used (based upon the noise present) with variable integration time at different segments of the wa velength. Figure 4.15 shows the absorbance and transmission spectra of a high - quality commercial diamonds along with the two lab grown diamond (GYJ 148 A and GYJ 148 C). 59 Figure 4.15. % Transmission and absorbance of different quality dia mond The commercial diamonds show almost i deal diamond characteristics, i.e. a high transmission (68 - 71%) throughout the uv - visible range and a sharp absorption near the band edge around 220 nm. T he %T value partly varies between the two commercial diamonds (from two different vendors) due to main ly marginal difference in thickness and also the background spectra (as these data were collected o n two different days), otherwise their response in gene ral are very similar. The other two diamond samples ( GYJ - 148A and 148C) were grown in MSU in diamond sy stem - IV (Reactor C configuration). These two samples have higher nitrogen content which shows up as high in the absorbance as compared to the other sampl es. These two samples also had a signal from silicon that may be present due to erosion of the quartz dome during growth . The signature of Si absorption appears around 732 nm in the transmission spectra. The marginal rise in absorbance for commercial diamo nd - 2 is mostly from the instrumental error (slight noisy background collection). Next, we look into t he effect of irradiation through optical characterization. Figure 4.16 shows the effect of the SHI beam irradiation after removal and cleaning of the elec trodes. 60 Figure 4.16. Effect of SHI beam irradiation on samples GYJ 148A and 148C As seen in the figure 4.16 both beams (Zr and Sn) leave a similar kind of signature after a similar dose of particles. The next set of diagrams in figure 4.17 show the transmission spectra of the lightly irradiated and heavily irradiated segments separately through a small aperture (2mm in diamet er). In the same scanned range there are some distinct differences seen between the two regions. For both diamonds strong absorption takes place around 350 nm, which happens due to irradiation produced defects, and the signatu re of the damage, i.e. a subst antial drop in transmission around 650 nm and 325 nm. However, the lightly irradiated segment appears very different for Figure 4.17. % Transmission on SHI beam irradiated samples Heavily irradiated segment 61 both diamonds. This could be due to GYJ 148C has slightly higher nitro gen impurity content than GYJ 148A as seen in figure 4.16 . Hence the effect of beam halo is more prominent in GYJ - 148A. 4 .2.2. Fourier T ransform I nfrared s pectroscopy (FTIR) Fourier Transform Infrared spectroscopy is a techn ique where infrared (IR) radiation is passed through a sample to study the molecular absorption or transmission characteristics [52] over a selected range of photon energ ies (wavelength s ). This method is helpful in identifying different kinds of bonds present in a n organic compound. When a material absorbs IR radiation, the molec ules get excited and goes to a higher vibrational state. The particular wavelength or range absorbed by the molecule is a function of the energy difference between the ground sate and the excited vibrational state, the wavelength is tagged as the character istics wavelength of the molecule. Mostly it is the mid infrared range span in the wave number range 400 cm - 1 to 4000 cm - 1 that is used for this application. The advantage of the mathematical technique " Fourier Transform" is that it allows to analyze the w hole range of spectrum at one shot, a n interferometer is used to achieve this very fast operation. In this experiment a sample is scanned by in Perkin Elmer spectrum one FTIR spectrometer, which can scan between wave number 400 cm - 1 to 4500 cm - 1 . F igure 4 .18 in this section shows the primary features of diamond infrared spectrum . 62 Figure 4.18. Infrared spectrum of commercial and lab grown diamonds In figure 4.18 , the infrared responses of commercial diamond - 1 and the two lab grown diamonds are depicte d. The strong absorption peaks present between 1900 to 2300 cm - 1 are assoc iated with C - C bonds in diamond [53] . The lab - grown diamonds show additional absorption around 1100 cm - 1 which is a signature of the presence of nitrogen impurities. Th ere are minor absorption signals at around 2900 cm - 1 (for the lab grown diamonds) and at 3100 cm - 1 for all the diamonds. These absorptions come from C - H x groups. A careful selection of low methane flow concentration ensures minimal absorption in this regio n. Figure 4.19 shows the infrared spectrum of the lab grown samples after irradiation. 63 Figure 4.19. Post irradiation infrared spectrum of lab grown diamonds It is quite evident from figure 4.19 , that the beam induced damages, or any added impurities rem ain non - respon sive in the mid infrared range, hence there is no additional absorption that appears in the spectrum. As a conclusion from the optical studies, it is found that SHI beam over the fluence of 10 13 particles - cm - 2 creates some color centers in th e single crystal diamond plates. Therefore, the absorption in the visible range gets affected the most, and in addition a part of the UV absorption also suffers. However, in the infra - red region the diamond retains all its c haracteristics. 4 .3. Structura l c haracterization Swift heavy ions that either pass through material or stop within material often produce displacement of atoms. This displacement of atoms, if large enough, can be seen by using structural analysis of th e material. In this section the analysis of beam irradiated samples by X - ray diffraction and Raman spectroscopy is performed to look for any damage that results in crystal structure deformation. In general, for any ion beam reacting with material, the effe ct is determined by the 64 ion ener gy, the fluence and the stopping range of the beam inside the material (or whether the beam completely passes through the material). Even though the low energy regime is more prone to cause structural damages, there is evide nce that in the high energy regi me where electronic stopping power is a dominating process, structural damages can happen for dielectric and semiconductor materials. At high energies the damage occurs along the path of the high energy ion , where the charge d ions produce cylindrical shape of damage along the trajectory . At any time, the electrostatic energy transfers outwards to the neighboring atoms and columnar defects are formed due to the shock waves. This phenomenon is known as track formation. However, up until now there has not been any reports showing any track formation happening in diamonds . Damage that is observed is mostly at low energy ranges due to dominance of nuclear stopping power. In a recent report Garcia et al. showed detailed experimental evidence of amorphization in op tical grade diamond material by irradiating with beam energies of a few MeV to a maximum of 40 MeV, which includes electronic stopping power to a maximum range of 14KeV/nm [40] . Therefore, to look for any major s tructural changes in the irradia ted samples in this work, a high - resolution X - ray diffraction method is applied. In high resolution X - ray diffraction (HRXRD) any feature related to crystallographic planes can be measured with more precision than normal x - r ay diffraction. 4 .3.1. High r e solution X - ray d iffraction (HRXRD) The XRD characterization for this study has been conducted with a Bruker AXS D8 system. Some detailed information about the system could be found in the attached reference [54] . The two major methods which are commonly used in HRXRD system are - curve scan. In - are present in the sample 65 and the corres ponding interplanar spacing. At each step the angle between the incident beam and the sample surface ( law gets satisfied. This method is also h elpful in studying for any stress/strain present in the material. On the other hand, rocking curve scan is more related to single crystal material analysis. It is mainly carried out to analyze the deviation of the orientation of a set of planes from its ex pected position. Figure 4.20. Rocking curve scan of GaN sample [54] As shown in figure 4.20 , for a rocking curve scan the sample is positioned with the beam [55] . Next the sam ple fixed. For substantial deviation if the planes are partially distorted d ue to strain . In this work one of the SHI beam irradiated diamond samples , GYJ 148C was selected for a rocking cu rve scan at the (004) orientation. A 50 micron slit width is used towards the source and a normal nickel filter is attached at the detector end . The rocking curve scan is carried at several spots on the sample travelling from the heavily irradiated segment to the lightly irradiated segment 66 as shown in figure 4.21 ° 119.5101 °, and in the next step it is rotated within ± 5.0 ° . Figure 4.21 shows a schematic of the strategy followed for rocking curve scans at different (approximate) spots on the sample GYJ - 148C. C ommercial diamond sample - 2 is also tested as a reference for thi s set up. Before any rocking curve scan a y - scan of the sample is done to get an approximate range to cover by the beam spot. Figure 4.21. A schematic of different p ositions for rocking curve scan The next set of plots in f igure 4.22 shows the rocking curve scans at the different spots of the sample, as well as the full width at half maximum (fwhm) distribution . A Gaussian curve fitting tool is applied here to get the fwhm value. The figure begins with a scan at he avily irradiated segment (scan 1) and goes all the way to lightly irradiated segment (scan 12 & 13). As observed 67 in each of the scan there exists a small peak along a tail with the (004) peak. The appearance of 2 line f rom the source 1 line. Figure 4.22. The distribution of rocking curve scan at different spots of sample GYJ - 148C The following table lists the distribution of fwhm as measured from the lightly irradiated to heavily irradiated regio n. The fwhm is calculated by using a b est fit method of a Gaussian curve over the points by the O rigin software ( https://www.originlab.com/ ). 68 Table 4.1 . Full width half maximum (fwhm) measured for peak (004) across the sample GYJ 148C. Sample name f whm (Arc sec) GYJ - 148C scan1 205.27 GY J - 148C scan 2 216.50 GYJ - 148C scan 3 218.41 GYJ - 148C scan 4 216.75 GYJ - 148C scan 5 208.87 GYJ - 148C scan 6 193.17 GYJ - 148C scan 7 199.69 GYJ - 148C scan 8 222.62 GYJ - 148C scan 9 194.25 GYJ - 148C scan 10 236.95 GYJ - 148C scan 11 240.5 6 GYJ - 148C scan 12 298.40 GYJ - 148C scan 13 357.0 5 Commercial diamond - 2 148. 35 It is evident from the table 4.1 that within the resolution limit of the rocking curve measurement set up, the fwhm show slightly higher fwhm around heavily irradiated segment (scan 2,3,4) com pared to lightly irradiated segment (scan 5,6,7). This needed to be verified with further repeti tion of experiments and conducting the same study on the other sample. (GYJ - 148A). 69 However higher fwhm values are observed towards scan - 12 and 13 which is almo st towards the edge of lightly irradiated segment. These increases could happen as the sample GY J - 148C has some visible non - diamond phases appearing in the figure 4.22 . 4 .3.2. Raman s pectroscopy Raman spectroscopy relies on a change in energy of a monochromatic light due to molecular vibration of atoms it is incident upon. The energy shift is realized as the characteristic of a particular molecule or a specific molecular bond existing in a test material. The Raman effect is a very weak interaction. In general incident photons get scattered elastically while interacting with a molecule. However, a small fraction of the total photon s goes through inelas tic scattering and hence emerge with a lower frequency (energy) , this effect i s known as Stokes scattering. In the other scenario where the molecule loses energy and therefore the photons end up shifting to higher energy side is called anti - stokes scatteri ng. The chance of anti - stoke scattering is low er . An experimental set up a Raman spectrometer is comprised of (a) a laser source (UV or visible laser), (b) a combination of different lenses to illuminate the sample, and (c) a notch filter and spectrometer to collect and analyze the scattered light and fin ally a charge coupled device to store the spectrum. Figure 4.23 shows a schematic of a Raman spectrometer set up. 70 . Figure 4. 23. A schemati c of Raman spectrometer [54] Raman spectroscopy [56] is a useful technique to characterize radiation damage in a material. Based on the shift of the Raman peak position or its widening one can characterize the strain and crystalline quality of the material. There have been numerous reports on qualitative F igure 4.24. Raman shift of an irradiated diamond, (a) damaged region , (b) undamaged region analysis of single crystal diamonds by Raman spectroscopy [57,58] . However there have been f ew reports on Raman spectra on radia tion damage o f single crystal diamond. Kalish et al . showed that for an ion implanted diamond there exist a critical damage threshold for the concentration of 71 damage produced vacancies above which diamonds get graphitize d and below that level it can be re s tored through proper annealing [59] . This has been supported by Raman spectra analysis on regions selectively implanted. A carefully selected annealing range for sampl es implanted below the critical dama ge limit results to a gradual decay of the Raman peak at 1630 cm - 1 , which is a Raman signal representing a specific vacancy. In another report, Yang et al . looked at Raman spectra o f a single crystal diamond irradiated with an 855 MeV electron beam over a 1mm 2 area done over several years to reach a total fluence o n the order of 10 19 . The Raman spectra in the damaged diamond showed significant change with additional peaks appearing due to amorphous carbon, as shown in th e figure 4.24 . The Raman spectra ind icates the such a dose produced substantial defects [60] . In this work, the Raman spectra at both heavily and lightly irradiat ed segments of the samples GYJ - 148A and 148C were measured. A Renishaw inVia Raman microscope was used to conduct the study. The instrument has a visible (532 nm green laser) and a UV laser. For this study the samples were illuminated by the green laser. T he power of the laser was adjusted to 1% to avoid any saturation in the ccd camera. The spot sized measured of the sample could be adjusted by using different objective lenses to get suitable laser illumina tion spot size . In our experiment the laser spot s spectrometer was used, which can vary between 10 - frame were used. Additional ap plication of the confocal lens in the optical path helps to reduce the beam divergence. As a first test of the measurement technique the electronic grade commercial sample - 2 and a HPHT sample is checked in the spectrometer. Both of the samples show the stro ng diamond characteristics peak at 1332 cm - 1 as t he only signal and hence ensures good quality as seen in figure 4.25. 72 Figure 4.25. Raman spectrum collected from the instrument used for this work (a) electronic grade commercial diamond, (b) HPHT diamond and (c) a l ab - grown non - irra diated diamond However, the lab grown diamond shows some additional feature peak around 1430 cm - 1 , which arises due to photoluminescence from nitrogen vacancies. There is als o visible signal for the sample from 1500 cm - 1 to 2150 cm - 1 w hich is due to the pr esence of trace amount s of amorphous carbon in the substrate. Figures 4.26 show the Raman spectra of the irradiated samples. ( c ) ( b ) 73 Figure 4.26. Raman spectrum collect ed from SHI irradiated diamonds As figure 4.26 reveals with comparison to 4.25 (c) that all these labs grown diamonds have nitrogen produced photoluminescence signals . The different curves (represented in different colors) appe ar from different spots of the measurement conducted. This confirms that these samples have higher fluorescence as the ba ckground due to presence of higher concentration of nitrogen (compared to commercial samples). Also , since no additional peak s w e re obse rved in the heavily irradiated segment , it is indicated that the type and fluence of SHI beams used for irradiat ion did n ot produce enough damage to change the Raman spectra. 74 5 .1. Brief overview of historical events The basic idea of growing diamond in the laboratory lies fundamentally on the understanding that first diamond is solely comprised of carbon atoms and secondly that it requires intense pressure and temperature in the upper mantle of earth to form natural diamond. In the early sixteenth century Averani and Targioni [61] studied the combustion of diamond, but it was the significant work cond ucted by British Chemist Smithson Tennant in 1797 that proved for the first time th e combustion of diamond produces CO 2 i.e. diamond is made of pure carbon. This idea opened a new series of attempts to convert diamond from the easily available carbon allot rope graphite. However, graphite is more stable at room temperature and pressure , which therefore made the idea very difficult to achie ve. In the late eighteenth century out of many attempts some premature success stories were reported by J. B.Hannay in 188 0, Henry Moissan in 1890 and later by Charles P arsons in the early nineteenth century, but some of the results could not be repeated an d therefore left some doubts about the outcomes [62] . In 1939 O. I. Leipunsky from Soviet Academy of Sciences, Moscow w as able to create a phase diagram to identify the right kind of temperature and pressure to convert graphite into diamond [63] . In 1941, P.W. Bridgman jointly with General Electric, Carbondum and Norton initiated a new venture to investigate diamond synthesis. Further in 1951, a new group was formed at General Electric with a focus on the role of high pressure in the diamond synthe sis process. The group worked on modifying the design of Bridgman`s anvil to elevate the pressure to 10 6 psi . This was fo llowed by some novel modifications by H. T. Hall to create a belt apparatus. Finally, in 1955 the 75 first success in the history of lab g rown diamond was reported by H.P. Bundy, H. T. Hall and their group . This work initiated a new generation of diamond synt hesis at high pressure a nd high temperature (HPHT). However due to government secrecy the details about the ultra - high pressure high temperature belt apparatus and the corresponding diamonds grown were not published until 1959 by H. P. Bovenkerk [64] . On the o ther hand, there was two competing projects between Union Carbide and General Electric going on around the same time to take a different diamond synthesis route, i.e. growing diamond at low pressure. After several years of attempt s it was Union Carbide Sci entist William G. Eversole who had the first success in the United States to produce diamond at low pressure on a hot diamond surface. Eversole`s study for this alternative approach started in 1949 by using CO as the source gas an d subsequently mov ed to me thane and other gases. The only document ation of his work c an be found in a patent [65] , [66] . However , at almost the s ame time Soviet scientist Boris Derjyagin and Borys Spitsyn applied for a patent in 1956 to claim the success of growing diamond in carbon using a gaseous environment [67] , although the patent was not published until later in 1980. The work of Eversole was revisited and picked up with great interest by J. C. Angus of Case Western Reserve University, USA . He gr e w diamond on diamond seed crystals at s ub - atm ospheric pressure , which validated Eversole`s work [68] . Further around 1976 Deryagin`s group reve a led the idea of growing diamond on non - diamond substrates. Nonetheless , all these early attempts suffered f rom very low growth rate, but their continued efforts led to the discovery of the critical role played by atomic hydrogen towards etching of graphite while promoting diamond growth. In the early 1980s, a breakthrough was made in Japan. Matsumoto et al. at the National Institute for Research in Inorganic Materials (NIRIM) developed a hot filament reactor [69] , which 76 unleash ed an outpour ing of worldwide intere st in low pressure diamond growth. The major outcome of this process was its capability to grow diamond on non - diamond su rfaces at almost 1µm/hr growth rate. This was followed by another landmark innovation in the deposition of diamond on non - diamond subst rates but using a microwave plasma discharge. This work opened a new direction of diamond deposition which is still consi dered as one of the most efficient technique s . The Hot filament chemical vapor deposition ( HFC V D ) systems are quite regularly used in i ndustry for large area deposition. Some details of these methods will be discussed in the next section. 5 .2. Hot filame nt CVD The h ot filament chemical vapor deposition (HFCVD) technique was the first of its kind, i.e. a technique to grow diamond at low pressure [70] . Also this technique gained immense popularity for offering nucleation and growth of diamond on non - diamond subs trates [71] . In this method, a gas mixture of hydrogen and small p ercentage (0.1~2%) of methane gas is flow ed into a vacuum chamber to raise the pressure in a range of 1 - 100 Torr. The gas passes through a wire mesh or array made of W (most widely used), Ta, or Mo. The wire array/ mesh is electrically heated to a temperatu re close to 2000 °C so the hydrogen gas mol ecules go through thermal dissociation and create atomic hydrogen. This atomic hydrogen further converts methane (CH 4 ) to methyl radicals (CH 3 ) and other hydrocarbon species. During the diamond deposition a Si or Mo substrate is placed a few mm to a cm bel ow the mesh at some elevated temperature (700 °C ~ 1000 °C) to deposit the diamond. The typical growth rate in HFCVD system s can vary within 0.1 - 20 µm/hr. Low er growth rates typically produce higher quality films . Figure 5. 1 shows a basic schematic of the HFCVD principle . 77 Although HFCVD is widely used to deposit polycrystalline diamond of reasonable quality, a limitation is the wire mesh act s as a source of contaminati on as well. Figure 5. 1. A schematic of H FCVD operation [72] During the deposition process the wire mesh /filament , especially for tungsten wires, is carburize d . This results to an additional incubation time at the start of the deposition. This causes slow and non - uniform growth rates during initial stages of growth. Also, at higher temp erature the filaments can oxidize, and they occasionally break as well. Since the gas activation process is a thermal process there is a low density of ions in the gas phase, hence biasing the substrate doesn`t work very effectively. 78 5 .3. Plasma Enhanced CVD (PECVD) Another method of diamond deposition is energizing the gaseous environment with a plasma to create the deposition process species . This process is called plasma enhanced chemical vapor deposition (PECVD). This technique c an overcome some of the drawbacks of HFCVD, especially in terms of contaminants. PECVD is a family of techniques of using a plasma created with different type s of dis charges such as DC - plasma, RF - plasma, microwave plasma, or electron cyclotron resonance mi crowave plasma. Other than DC plasma, in all other cases a plasma discharge is generated by applying a high frequency electric field. The gas is partially ionized and within a certain volume converts into a collection of electron, ions and neutral gas mole cules, which is overall electrically neutral. The highly energized electrons further impact the hydrogen molecules to dissociate them into atomic hydrogens. The elect rons temperature reaches over 10,000 K. However due to the neutral molecules the overall g as temperature stays lower. In low pressure plasmas, the electrons have a longer mean free path, fewer collision s with neutral molecules and therefore less energy tra nsfer, so the overall neutral species temperature in the plasma is low. The exact opposite scenario happens at high er pressure s and therefore the concentration of reactive species and overall temperature of the plasma is high er . The two important factors guiding the PECVD process are the initial chemical reactions in the non - equilibrium temper ature plasma volume and the flux of species to the surface where diamond deposition occurs. A general schematic description of the PECVD process is shown in figure 5. 2. 79 Figure 5. 2. Reaction steps in PECVD process [73] Within all the different types of CVD techniques, the microwave plasma assisted chemical vapor deposition (MPACVD) technique is an extensively adopted technique for high quality diamond deposition. Therefore, the MPACVD technique will be discussed in more detail in the following section. 5 .4. Microwave Plasma - Assisted CVD (MPACVD) . The m icrowave plasma assisted CVD technique is currently the most applied method worldwide to grow high quality diamond. In the MPACVD technique, microwave power is coupled to the pro cess gas volume through a d ielectric window to excite the discharge. The process chamber is the fundamental part of the electromagnetic cavity, where only a particular mode of electromagnetic wave can oscillate. The mode is selected based upon the cavities shape and dimension s . The microwave electric field heats up the electron s and the oscillating electrons transfer their energy to neutral gas molecules through subsequent collisions. The MPACVD process is very efficient from low pressure (10 mtorr) to ver y high pressure (~400 To rr). The 80 behavior of the discharge changes quite significantly with the variation of the pressure. As the pressure is increased in the process chamber, the plasma discharge shrinks in volume and creates a high er - power density regime of operation. At higher pressure s , the discharge also becomes more spatially inhomogeneous (i.e. the core is at a higher temperature surrounded by a sharp transition to lower temperature s at the boundary of the plasma ). The chemical reactions among the ga s species start once the discharge has reached the required temperature, and the growth of diamond initiates when the right condition s are present on the substrate. Generally, the substrate is either penetrating the discharge or kept in very close proximit y at the edge of the pla sma . Some of the major advantages of MPACVD process is the stability and easy reproducibility of the microwave discharge, the high plasma density and low sheath potential, and the possibility of mixing different feed gases with less chance of contamination [10]. In the MPACVD process, the microwave power is generated from a magnetron operating at 2450 MHz (as it is more available) which is passed through a load matched waveguide. Another commonly used frequency is 915 MHz, which is used to deposit on larg er substrates. A general configuration of a MPACVD is shown in the figure 5. 3. 81 Figure 5. 3. A cross sectional configuration of a MPACVD reactor [74] Although a heater beneath the substrate is optional , at higher pressures a cooled stage is used to efficiently control the top surfa ce temperatu re of the substrate on which the diamond is deposited . Controlling the deposition temperature is important for getting high quality diamond deposition. The power to ignite and maintain the plasma depends on absorbed power P abs (where P abs = P so urce P refl ected ) where P source is the microwave power from the power supply and P reflected is the microwave power reflected from the reactor . The MPACVD process still draws lot of attention from both industry and academia in terms of different reactor d esign s [75] , [76] , [77] , [78] , [45] as well as optimization of the process to make efficient and faster growth of diamond. 82 5. 5 . Deposition of s ingle c rystal d iamond The nature of diamond growth in the CVD p rocess is an ensemble of multiple complex chemical and physical process that deposit hydrocarbon species on the substrate [79] , [80] , [81] . Some of these complex reactions get altered based upon plasma growth condition s and substrate surface conditions . Before going in to the actual details of the chemical process, a sim plified view of diamond growth process is shown in figure 5.4. The process gases are fed to the deposition chamber of the reactor. The activation of the reactant gases takes place next by energy supplied from a microwave, RF or h ot - filament source. This pr oceed s by the active radicals and molecules moving closer to the growth surface and finally chemical reactions on the surface that grow diamond [82] , [8] , [83] . 83 Figure 5. 4 . A schematic showing the process of diamon d dep osition during CVD process [ 20], [84] The diamond deposition process by CVD technique is an e xample of turning gaseous reactants in to a solid - state material . Therefore, the fundamental aspect of this process relies on the chemical reaction s of the species at the surface, the orientation and morphology of the surface and the temperature of the surf ace . As already s tated, atomic hydrogen plays the most crucial role for the diamond CVD process . T he diamond growth conditions are preceded by the substrate surface being almost entirely saturated by hydrogen atoms. Otherwise dangling bonds created by surf ace abstraction o r thermal desorption may lead to graphitization. Also, atomic hydrogen etches graphite faster, leaving sp 3 bonded carbon at the surface . Then a small percentage of carbon containing precursor gas is introduced in the chamber. Out of variet y of choices, methane has become the most used carbon precursor gas over the years. Occasionally hydrogen atoms de - 84 absorb from the surface to recreate a d angling bond and forming H 2 . Then, due to the presence of a hydrocarbon species (CH 4 ) in the plasma, a t a small percentage the hydrogen abstraction sites methyl radicals (CH 3 ) can react with the open surface sites and carbon atoms get attached to the bond s. The combination of this process (hydrogen abstraction and methyl addition) continues to the next ad jacent sites and diamond grows . The chemical reaction s for diamond growth rely on the relationship of external control parameters to the deposition CVD c hemistry within the reactor operating region of diamond growth . A detailed framework of diamond growth based on the input gas compositions was proposed and f urther defined by Bachmann et a l. [85] in 1994. Their initial work was built on the experimental condition of nearly seventy research works that used a variety of different CVD techniques. The processing input gas mixtures for diamond growth is plotted on a triangular plot with the carbon , hydr ogen and oxygen (C - H - O) composition indicated on the diagram as shown in figure 5. 5. One this diagram the region of diamond growth is indicated. 85 Figure 5. 5 . Bachman C - H - O triangle describing the gas composition required for dia mond growth [11] The impor tant prediction from Bachmann was that, irrespective of the nature of growth, diamond will only get deposited when the gas composition is above the CO to hydrogen tie - line [11]. Outside this narrow triangle of diamond growth, the result either ends up into no growth (i.e. if gas composition moves towards the O vertex) or nondiamond carbon phase depositio n if the carbon concentration is too large. 5 . 5 . 1 . Chemical reactions for CVD diamond growth The basic process of diamond deposition is primarily built u pon two important sequences, i.e. (a) the creation of atomic hydrogen in the discharge and (b) the breaking the carbon containing gas species into carbon containing radicals. T he chemistry of diamond deposition is a complex 86 process as it involves (i) compe titive deposition process es between sp 2 and sp 3 structure growth and (ii) the possibility of other chemical reactions taking place due to presence of other carbon containing compounds. An early study of Frenklach and Wang [86] , [87] covered some of the se processes like the growth rates of sp 2 and sp 3 carbons, and their respective etching processes to have a better understanding of the diamond deposition process. In a plasma assisted CVD system a simplified set of chemical reactions (a - k) can be formulate d to describe the diamond deposition process. The electrons are energized in the plasma by the power coupled to system and the ene rgetic electrons dissociate the molecular hydrogen into atomic hydrogen. At lower energy density, atomic hydrogen is produced through electron impact dissociation [88] , [89] (a) This above process is balanced by diffusion of some hydrogen atoms to the wall or recombination of the atomic hydrogen in the gas phase. Although the recombination of hydro gen atoms in the gas phase is a slower process , especially at lower pressures . The recombination process can be writt en as; (b) which is a pressure dependent process, where a third body (M) is involved to carry away the exces s h eat. In presence of any hydrocarbon species, a combination of reactions can also take place in the plasma discharge that use the atomic hydrogen. Some example reactions include reactions that convert stable molecules (such as CH 4 ) to reactive radicals (s uch as CH 3 ) and vice versa. 87 (c) and (d ) This homogeneous recombination process (c) can be neglected at low pressure s, but for higher pressures with a considerable amount of hydrocarbon the hydrogen atom distribution in the gas phase is affected. However, hydrogen atom loss primarily happens at reactor wall s and at the diamond growth surface. D iamond deposition requires the formation of active diamond growth sites on the growth surface where a sp 3 radical or other carbon con taining radical addition to the surface can happen [23]. Under typical growth condition s, the number of carbon surface radicals per unit area ( depends upon the rate of abstraction of hydrogen from the surface (e) and the loss of surface carbon radical sites caused by atomic hydrogen adsorption on the surface, (f) where C d H and C d* represent the diamond surface site bonded to hydrogen and carbon surface radical sites, respectively. The corresponding reactions that fil l a carbon surface radical via the dissociation of hydr ogen , (g) is much slower in comparison with reaction (f ). also depends on other growth species; 88 (h) and decomposition of d iamond structure on the surface to graphite (and it`s reverse reaction), (i) and (j) where, denotes a sp 2 structure site. At st eady state condition s for the surface radicals, and at high atomic H concentrations, the terms k 2 [H] and k 1 [H] become the controlling factor of the surface density, hence, The overall surface site density is which is observed in any standard CVD growth processes. At low hydrogen atom concentration s, dominates the denominator however has some effect which makes, 89 where K 1 is the equilibrium cons tant. From the equation above ( n), it is evident that diamond growth depends on the equilibrium condition of H atoms concentration over H 2 gas molecules concentration. Low concentration of H atoms slows down etching of non - diamond surface carbon sites and total number of diamond growth sites. This is one of the key points of defect generation of diamond (although this is not the most comprehensive explanation for all kinds of defects that can occur during diamond growth). A generic idea of defect growth r elies on the interaction of H atoms and sp 2 incorporation. The basic assumption in the generic model of defect growth is that an adsorbate reacts with another nearby adsorbate prior to forming the lattice [90] , for example two methyl groups rea ct and over grow to a sp 2 defect. Therefore, at a fixed substrate temperature, the defect fraction in the film follows an empirical model as, where G is the growth ra te. The order n is determined experimentally. The growth rate G is proportional to [CH 3 ] [H] and in partial equilibrium, CH 3 is approximately proportional to [CH 4 ] [H]/[H 2 ] which translates into, In g eneral, n=2 is selected and 90 H ence , from a qualitative viewpoint , higher atomic hydrog en improves film quality (low er sp 2 defects formation) however it also needs a low concentration of carbon rich gas species in the growth process. Some def ects with sp 3 characteristics also form if the kinetics of bond formation between adjacent carbons is slower than the additional layer of carbon deposition due to a high gas flux of carbon rich gas [91] . 5. 5 . 2 . Microwave plasma ass isted CVD re actor The very early approaches of low pressure CVD processes (Spitsyn, Derjaguin, 1980, Eversole 1962, Angus 1968) [92] , [93] had low grow th rate s of a few nanometers/hr and hence they could not be applied to any industrial level growth. The major hurdle was that the gas tempera tures w ere very low (600°C ~1200°C). The breakthrough appeared when the gas phase temperature and substrate temperature could be separately controlled . This idea allowed the creation of higher atomic hydrogen concentrations in the gas phase as well as an i ncrease in the carbon growth species that allowed growth rate increases. The formation of a hot zone in the gas phase with a hot filament by Matsumoto et a l. [9] from NIRIM led to a substantial high er growth rate and this process became a popul ar one in in dustrial applications. In the next year, 1983, after the s uccess of HFCVD process Kamo et al . from the same institute, reported a methane/hydrogen plasma discharge formed with a 2.45 GHz microwave source that successfully deposited diamond fi lms [94] . Later their reactor design became popular as the NIRIM type reactor, which used a silica tube that is attached to a rectangular waveguide to ignite the plasma. The typical opera ting conditions for these reactors are 80 - 100 Torr with microwave power range of 100W to 1.5 kW. 91 In 1987, P. K. Bachmann in jo int work with Applied Science and Technology (ASTeX), designed a new type of 2.45 GHz CVD reactor by using a bell jar (dome shap ed) with an inner diameter of 10 cm. The microwave power is guided through an antenna to the wave guide which contains the bell jar. The plasma is generated at the center of the bell jar where the field is maximum in strength. The design also had a scope o f moving the substrate holder based upon a thermally floating stage or with an option of externally heat ed stage. Typical operat ing condition used for this reactor are 40 - 70 Torr with 1 kW power. In 1992, an improved design of the bell jar reactor was an nounced by AsTeX, further known as High Pressure Microwave Source (HPMS) reactor . It was upgraded with a higher power 5kW magn etron power supply to increase the growth rate. Also, another major change was the bell jar was replaced by a silica microwave w indow. This advanced system could operate at a wider range of pressure including 10 - 120 Torr. Later AsTeX designed a 915 MHz sys tem to enlarge the deposition area up to 30 cm diameter. At present some newer AsTeX 2.45 GHz reactors with power capability of 8kW and capable of covering an area of 50 mm diameter [95] have been developed. In 1998, Funer et a l. developed a microwave cavity, ellipsoidal in shape with microwave power fed from the top [78] via an antenna. The majo r outcome of this design was that the concentration of the field just above the surface of the substrate could be maximized with a very stable discharge. Both 2.45 GHz and 915 MHz systems were configured using this design. The systems were operated in a pr essure range of 30 - 150 Torr with 2 - 6 kW power for 2.45 GHz and 20 - 60 kW for 915 MHz and were able to deposit uniformly on wafers with 2 - 6 - inch diameter. From 1986 - 1995, Michigan State University developed the microwave cavity plasma reactor (MCPR) and it was licensed by Wavemat/Norton [96] . A 2.45 GHz m icrowave generator 92 transmits the microwave energy to the cylindrical cavity through a rectangular waveguide via a coaxial wavegu ide transition and finally ending at an excitation probe. The cavity consists of cylindrical metal walls and a movable short pla te defining the effective length of the cavity. At the end of the cavity a quartz dome encloses the discharge, which is hemisphe rical in shape. Since the entire deposition work of this thesis is carried out on an MSU reactor, the parameters related to this reactor will be discussed with more detail. A general design configuration of the MSU MCPR reactor is shown in the figure 5. 6 . Figure 5. 6 . A cross - sectional view of a typical MSU designed reactor [ 97] T he physical parameters are defined as Z=0: the plane that defines the boundary between cylindrical waveguide section (Z>0) and coaxial waveguide section (Z<0). 93 R1 & R2: cavity applicator radii, R3: Cooling stage radius, R4: substrate holder radius Lp: probe length and Ls: short length L1 - L2: This defines the substrate position. The substrate holder and cooling stage radii were modified at different stages of the reactor development and the design evolved to three different type s of MCPR reactor. The initial design of the reactor evolved fro m a microwave plasma disk ion source in the 1980s and gradually through several years of work it was modified to MCPR reactor [96] . T he first MCPR reactor design is designated as Reactor A and it has been used for polycrystalline deposition over the past two decades . [96] The major focus was to develop the expe rimental condition s to deposit polycrystallin e diamond uniformly on Si substrate at pressures 80 - 140 Torr [38]. Later the operation was successfully scaled up to a deposition at 180 Torr over an area of 75 mm diameter [96] . However, to increase the growth rate and to move towards single crystal diamond deposition , reactor A (which was a generic design) was modified to operate at 180 - 260 Torr. The major changes happened to the diam eter of the cooling stage (R3) and the diame ter of the substrate holder (R4) with minor adjustments at the substrate height (i.e. adjusting L1 and L2). This new design allowed more concentrated plasma discharge (i.e. higher energy density) with higher grow th rate but with a cost of sh rinking coverag e area. This new reactor design is named as Reactor B. This hybrid reactor`s (cylindrical for z>0 and coaxial waveguide for z<0 ) design allowed excit ation of the TM 013 and TEM 001 modes simultaneously in the cylin drical and coaxial waveguide sections, as sh own in the figure 5. 7 . Experimental work on PCD deposition and SCD deposition on this reactor B can be found in the thesis work of K adek H e m a wan, Jing Lu and Shreya Nad and associated publications [97 103] . 94 Figure 5. 7 . (a) Hybrid microwave cavity applicator (cross sectional view) with cylindrical and coaxial intersecting at z=0 plane (b) cross sectional view of standing wave pattern in Reactor B (smaller substrate holder) [104] (a) ( b ) 95 To visualize the field distribution inside the cavity it is helpful to have idea about the guided wavelength g , which is the wavelength of the propagating wave inside the boundary of the cylindrical waveguide. Here g is the could be defined as w here is the corresponding cutoff frequency of the TM (transverse magnetic) or the TE (transverse electric) modes in a cylindrical resonator free space and f is operating microwave frequency, i.e. f = 2.45 GHz here . In th e cylindrical resonator a stand ing wave will be produced when the two ends are enclosed by conducting plates g /2. To further explore diamond growth in a higher - pressure regime, i.e. 220 - 300 Torr, a new design was est ablished. The design evolved b ased on knowledge and operational experience of earlier reactors and it has been designated as Reactor C [105] . A detailed description of the background and developmental stages of this reactor can be found in the thesis work of Yajun Gu [106] . The generalized goal of this specific design was to develop an applicator which is a combination of different cylindrical coaxial and cylindrical waveguide sections. In additio n, the adjustable probe and short plate capability were kept intact to have precise load matching control in order to minimize microwave power loss. The critical idea of the new cavity design is to properly guide the microwave power (i.e. to spatially focu s, defocus and finally refocus) towards the substrate holder ( z =0 plane) to get a m ore intense plasma. The radius of the cooling stage and the substrate holder remained unchanged in this reactor as compared to Reactor B. However, to compensate the power lo ss from the intense plasma being close to the quartz bell jar, the bell jar`s 96 diame ter and height was increased from the earlier designs. A final prototype of the reactor design is shown in figure 5. 8. Figure 5. 8 . Cross section of generalized design of Reactor C [106] A d etailed description of the sections 1 , 2, 3 and 4 shown on the right side of figure 5. 8 can be found in reference [106] . In summary, this reactor excites differen t electromagnetic modes ( as compared to reactor s A and B) in different sections based on the dimensions selec ted for the design. Hence a new MCPR reactor became available from MSU that was capable of synthesizing single crystal diamond at a higher - pressure range. A summary of the different MCPR diamond deposition reactors is given in table 5. 1. 97 Table 5. 1. Differ ences in important design parame ters between Reactor A, B and C (Adapted from reference [97] ). Applicator d imensions (cm) Cooling stage dimensions (cm) Substrate position (cm) Quartz bell jar dimensions (cm) Excita tions modes R1 R2 R3 R4 Zs Radius Height Reactor A 8.9 7.0 4.1 5.1 Fixed 13.0 9.5 TM 013 Reactor B 8.9 7.0 1.9 3.2 Variable 13.0 9.5 TM 013 + TEM 001 Reactor C 15.2 10.2 1.9 3.2 Variable 21.6 10.9 TM 01n + TEM 001 Operating P rocedures of MPA CV D reactor (Reactor B): This section will cover the detailed operation of the specific CVD reactor used to conduct the work in this thesis. This specific reactor is known as Diamond S ystem 3 or DS - 3 in the Fraunhofer - MSU center, which wa s established based o n a MCPR reactor - B conf iguration and operated at a high pressure (240 Torr) regime. A brief coverage of the microwave cavity structure was already covered in the previous section. The other important necessary attachment s to the cavity are; (a) microwave p ower supply and the pow er transmission waveguide , (b) the vacuum pump system and the pressure control system, (c) the gas flow mechanism, (d) the substrate holders and water flow system for necessary cooling and (e) a computer to run the automation program (LabVIEW) to control t he entire deposition process. The following schematic in figure 5. 9 shows a schematic of the entire CVD system for diamond growth, followed by a table (5.2.) showing all important attachments for the operation of the reactor. 98 Figu re 5. 9 . A schematic of the entire diamond deposition sys tem based upon reactor B design 99 Table 5. 2. Description of the different sections/attachments of reactor - B 1. Gas cylinders (Ar, H 2 , CO 2 and CH 4 ). 7. Short plate. 13. 2.45 GHz, 6 kW Cob er power supply. 19. Four channel Mass flow controller (MKS 247C) 2. Mass flow controllers for each gas cylinders. 8. Probe assembly. 14. One color pyrometer . 20. Dotted line represents all the communication cables from any instrument to computer. 3 . Main gas needle valve to flow gas in the chamber. 9. Rectangular coaxial waveguide connector. 15. Sensor for dome temperature reading. 21. Vacuum chamber door. 4. Stainless steel baseplate. 10. Flexible waveguide. 16. Computer for running LabView. 5. Quartz dome. 11. 90° bend wave guide 17. Baffle valve. 6. Cylindrical microwave cavity. 12. 60 dB power meter. 18. Mechanical pump. (a) Microwave power supply & power transmission line subsystem : This section covers the functionalities of th e microwave power supply and the process of microwave power transmission to ignite and maintain the discharge. A Cober power source (model S6F) is the 2.45 GHz microwave energy source. It has maximum power rating up to 6 kW. A circulator, a magnetron, a du mmy load are the major parts of the microwave power supply which are enclosed inside the power supply cabinet. The circulator and the magnetron are con tinuously water cooled during its operation. The generated power is fed to the cavity through a series of wave guides, i.e. a S - band rigid rectangular waveguide, a bi - directional coupler (with 60 dB attenuation) which are also connected to two HP 8482H pow er sensors connected to HP 435 B power meters, a 90° bend waveguide, a flexible waveguide and finally an adapter to connect a coaxial waveguide. The two HP 435B power meters are used to measure the parameters; incident power (P inc ) and reflected 100 power (P re fl ) which are used to determine the absorbed power in the reactor, as well as the power efficiency of the system and the power loss. Also, the reflected power is used to tune the reactor to operate it at maximized efficiency, which will be discussed in a l ater section. (b) Vacuum and pressure control subsystem : The pressure inside the chamber is maintained w ith the combination of a vacuum pump and a throttle valve. The Alcatel 2010 SD pump is directly connected via a throttle valve and an oil trap. The cha mber is normally evacuated for 8~12 hrs to reach a low pressure (3~5 mTorr). The pressure in the vacuum c hamber is measured with two Baratron capacitance manometers (one measures 0.001 Torr to 10 Torr and another one from 10 Torr to 100 Torr) and they are read by an MKS - PDR - 2C module. The operating pressure is maintained by the MKS - 651 pressure controller, whi ch remotely adjusts the throttle valve`s position to hold the set pressure (~ 240 Torr for most of the processes here) within ± 0.5 Torr of the set val ue. The pressure is created by allowing the feed gases H 2 , CH 4 , and O 2 to flow in the chamber at manually controlled flow rate s and eventually computer - controlled rates , once the pressure reaches its set point. Generally, a leak check is conducted by closing the throttle valve and measuring the increase of pressure over a set amount of time. This is done to e nsure no, or only very minimal external leak s exist. The DS - 3 system is a tight system which has a leak rate of 3 0 ~ 40 mTorr/week. Another source of nitrogen getting into reactor when the chamber is center to atmosphere. Nitrogen gas is also used to dilut e the hydrogen in the exhaust line from the pump to the external vent outside the building. This is done to minimize any chance of explosion. (c) Gas flow subsystem : The gas flow in the system is controlled initially during a run by manual controlled flow and then finally computer - controlled flow by the MKS 627C (4 channel readout). The base pressure is monitored by the MKS 627B pressure controller which also adjusts the throttle valve position. During this study DS - 3 ha d 4 gas cylinders attached to it, i. e. CH 4 (99.999% pure), 101 H 2 (research grade, 99.9995% pure), CO 2 and Ar gas. All four gas sources are connected through flexible metallic tubes to i ndividual mass flow controllers (MFC) of different max imum flow ranges. Each MFCs are controlled by the MKS 62 7C which comes with 4 channels for read out. The CO 2 and Ar gases are generally used only for the quartz dome cleaning purpose or in some specific condition where the growth takes place with a little CO 2 added in the feed gas. H 2 gas is used both for the substrate etching (@ 180 Torr) or the growth process with 5% CH 4 added (@ 240 Torr). The output of the MFC`s meet at a common point which has a pi n valve attached. The gases flow to the growth chamber through another flexible metallic tube only when the pi n valve is opened. From the final gas tube, the feed gases enter the quartz dome through a single hole in the bottom steel plate and then flows th rough equally spaced small holes in to the quartz bell jar through another supporting plate. The gas circulates inside the dome and flows out of the quartz dome through a base substrate holder with sixteen holes drilled close to its perimeter and next throu gh a quartz tube of diameter ~ 60 mm, just seating on the cooling stage. The gas flow reaches the throttle val ve and gets pumped out based upon the throttle valve`s position. As earlier mentioned, the gas flow rate is controlled manually from the front p anel of the MKS 247C to reach the target pressure through different stages. The pressure is put as a set value in the LabView program and the computer monitors the pressure. Once the pressure is reached, as read by the front panel of the MKS 627B as well a s the MKS 627C are all turned to remote control mode so that the LabView can take full control of the operatio n for the entire growth time. Also, the program can shut down (turns off the microwave power supply) the process if there is a pressure fluctuation by more than 1 Torr. This is a safety procedure programmed for the operation along with some other safety pr ocedures. 102 ( d ) The substrate holder and cooling sub system: As the deposition process in the reactor takes place at high pres sure and high - power density, the substrate holder and the diamond s ubstrate itself gets very hot unless it is efficiently cooled. T he reactor is connected with a Neslab CFT - 300 water chiller. The circulated water is maintained at 18°C. From the main water flow line in the reactor, the line splits into two parts. One part goes to the base plate of the cavity. The other part goes to the sliding short plate through a flow meter (to ensure a standard flow rate for every operation). The water flow rate normally varies between 0.45 gal/min to 0.50 gal/min. As a matter of maintenance, if the water flow rate drops below 0.25 gal/min, then the whole water chiller and the flow lines are cleaned, and fresh distilled water is put into the tank. From the sliding short pl ate, the water flows into the cooled substrate holder stage (located inside the vacuum chamber) and then it returns to the water ch iller. Hence the flow rate is also important to adjust the cooling stage temperature, which further controls the actual subst rate temperature. Apart from the substrate holder being water cooled, the inside of the microwave cavity is also cooled with an a ir blower . It blows in slightly above the quartz dome, where air flows through an inlet in the side of the microwave cavity s o that the air swirls around the outside of the quartz dome and then flows out through the cavity windows with small drilled holes. There are two additional fans directed towards the outside walls of the cavity, one cools around the antenna probe and anoth er flows air externally around the cavity. Next, focusing on the cooling stage, figure s 5.10 and 5.11 show the design of the cool ing stage and the substrate holder configuration. The substrate holder seats on top of the cooling stage shown in figure 5. 10 . L1 and L2 decide the distance of the top of the substrate holder from Z=0 plane. The thickness of a shim that goes between the co oling stage base and the cavity base plate is used to adjust the substrate position, denoted by Zs (= L1 - L2). The experiments for this work 103 were done using a shim with thickness 0.254 inches. The actual substrate holder that sits on the cooling stage is a combination of three molybdenum parts, a holder with 2.55 inches diameter , a supporting disc and the SCD recessed pocket holder (Fig. 5. 11) . The entire cooling stage shown in figure 5. 10 , that the substrate holder sits on , is made of stainless steel . The inlet and the outlet line of water flow in the cooling stage have diameter s of 0.375 inches. The conducting plate is used to act as a conduction path of the microwave field surface currents. It also defines the effective bottom cavity length to sustain the desired microwave mode. At the top of the cooling stage column there is a small indentation. This is to properly align the substrate holder on top of the cooling stage along the line of the probe. There are 8 holes drilled inside the circular conducting plate. The quartz tube that seats on this conducting plate hold s the substrate holder and guides the gases to flow through the tube and finally to a lower chamber connected to the vacuum pump. 104 Figure 5. 1 0. Design of the cooling stage for reactor B [107,108] 105 Figure 5. 11. Dimensions of a 1.8 mm deep pocket holder [108] (e) Computer control subsystem : LabView is installe d on a computer to automate the whole deposition process. The program has access to the MKS 651 pressure control unit to monitor the pressure, the 4 - channel read out to control the MFCs, a control to the microwave power supply to turn it on or off remotely , an d access to the power meters to read (and store in a log file) the incident and the reflected powers. The reflected power is an important parameter, as its value is 106 the deciding factor for the tuning of the reactor. At the growth pressure, with the des ired gas flow , microwave power, and standard probe length (Lp ~ 3.6 cm) set, t he short plate position, i.e Ls , is adjusted to find the minimum reflected power position. This is also guided by the selection of L1 and L2. However, L1 and L2 cannot be continu ousl y adjusted during the process. So, a selection of L1 and L2 needs to be done before the process initiates. With a fixed set of L1 and L2 selected, Lp and Ls are varied continuously to search for a point where the reflected power shows minimum value. Fo r th is work it is found that Ls ~ 21.5 cm allows a minimum reflected power point. More details about tuning process of the reactor could be found in reference [103] . Unless the process is not run with minimum reflected power it will add to more power l oss. Also, if the reflected power is more than 25% of the incident power the computer control program is designed to shut off the process. There are many recipes developed through labView by former researchers in this group. The most commonly used process e s are; (a) a 180 Torr hydrogen plasma etch process (b) a 240 Torr H 2 /CH 4 gas discharge process for single crystal diamond deposition and (c) a 60 Torr CO 2 /Ar gas discharge process to clean the quartz dome. As a standard sequence, processes (a) and (b) are always r un in a sequential combination for a growth process. In process (a) the top surface of the diamond seed is etched to remove any polishing damage and any impurities on the surface. The process is r un for 1.5 hr close to a temperature 1020°C ~ 10 30°C with a pure hydrogen discharge. After completion of this process, the chamber is pumped down for at least an hour to get rid of all the residu al gas in the chamber . The next process at 240 Torr is the one used for scd deposition. Any of these proc esse s can be run for different time s and at different pressures (these are the input parameters of the LabView program) unless other physical parameters of the reactors act as limitations. In process (b), multiple 107 s teps are added in the L abView program and the user can add or remove st eps with ease. In general, H 2 gas is flow ed in to the chamber to reach the growth pressure, and the power is also adjusted in steps to reach a temperature around 1050°C ~ 1060°C. Next the process continues at the stage for anot her 10 minutes for the temperature to settle down. Following this the methane is added in three stages. In stage - 1, 1% methane is flow ed for 3 min, then 2% methane for 3 min in stage - 2, next, 3.5% methane for 3 min. in stage - 3 and finally 5% methane for th e re st of the process (either 2.5, 12 or 24 hrs.). All these parameters can be controlled from the LabView program. However, the power is gradually adjusted so that there are no abrupt power changes. This sequence for the addition of methane helps to avoid any sudden surge of methane and therefore any abrupt change in temperature. The growth process is conducted at 1080°C ~1100°C. Overall , this sequence allows the temperature to rise very gradually and with minimal adjustments in the source power control kn ob. Substrate temperature measurement is done with an Arduino uno microcontroller programmed to record the temperature from the Ircon Pyrometer. T he data is then fed to a raspberry - pi computer which continuously displays the substrate temperature and the dome temperature (with room temperature also as a reference). The raspberry - pi also collects picture s of the top surface of the diamond substrate from a camera and updates it every 5 minutes in the display screen. There is an additional camera that records liv e picture of the plasma through one of the viewing ports. Overall it generates a log file with all the variable parameters being updat ed every 5 minutes. The camera looking directly at the top surface of the s ubstrate is attached to the view port of the in frared pyrometer. The operation of the pyrometer is fully controlled by the Arduino uno. There is another Arduino uno connected to the microwave power supply. This one is used to automatically ignite the plasma at 5 Torr pressures w ith 1 kW power. The uno adjusts 108 the position of a stepper motor which can move the power adjustment dial. For any further power adjustment, a rotary encoder is used to either jog or drop the power by moving the power adjustment dial through sending a signa l from the uno. 109 6.1. Overview of scd deposition on misoriented substrates A major focus in the earlier section of this dissertation was on diamond detector`s performance . One aspe ct of this was addressing the performance of MSU lab grown diamond in comparison with the commercial electronic grade diamond. One of the major drawback s of these MSU lab grown diamonds is their low charge collection distance (ccd) and therefore low charge collection efficiency (cce). As observed in section 3.2 , in comparison to the commercial diamond plates (even after irradiation with heavy ion beams) the MSU lab grown diamond plates show ed a narrow er transient current response time at high electric field s . The lab grown diamond plates also had high birefringence i.e. higher strain in the samples . These problems mainly arise because the MSU diamond plates have higher nitrogen impurities (1 00 ppb~ few ppm), as nitrogen vacancies act as charge trapping cente r s . Also, the strain generated from the defects (threading dislocations) also severely affect the electronic properties (low breakdown voltages, high leakage current) [109] . Hence key factors towards growing plates for electronic /detector applications are (a) deposition of the SCD at very low nitrogen or controlled nitrogen environment and (b) control led growt h conditions to minimize defects on the grown layer. While nitroge n is a detriment to the quality of diamond for electronic and detector applications, it has a critical influence on the growth of diamond. A s mall amount of nitrogen added in the gas phase helps to achieve grow th with low er strain during deposition and it can enhance the gro wth rate [110] , [111] . Wa t a nabe et a l. show ed that an atomically flat epilayer of diamond with both low defects and impurities could be grown with a very low concentration of 110 methane (0.025%) in the feedgas to the plasma [112] , [113] . However, to increase the growth rate from 50 nm/hr they used 0.15% to 0.5% methane which resulted in formation of uniepitaxial crystallites and pyramidal hillocks, as shown in figure 6.1 . The defects formed dur ing the ear ly stages of diamond growth and for low methane l evel s 0.15%) the surface is smoother [114] . Figure 6.1. Appearance of defects on scd surfaces with very low percentage of CH 4 addition [114] The authors claim that this ultra - low methane mix actually strikes a nice balance between etching of the surface defects as well as d eposition of a high - quality layer. Further investigation in the de fects shown that the defects were more localized at the unepitaxial defect regions . Similar non - epitaxial features and pyramidal hillocks were observed by Tsuno et a l. [115] while growing a . However, their growth condition had oxygen added in the feed gas. These non - epitaxial particles mostly formed due to small {111} facets forming on the { 001} substrates du ring growth. 111 Bauer et a l. [116] reported on grow ing high quality scd without any deliberate addition of nitrogen in the feed gas. I n the process they added almost 10% of methane and raised the s ubstrate temperature to 1200 °C to reach The result was some of their initial growth on (001) oriented surface s had non - epitaxial defects on the surface. Additionally , the edges of the grown samples had high PCD deposition, which blocked the lateral growth of the scd . The non - epitaxial particles laterally increased in size with longer duration of growth. Also, these structures gr e w at a higher rate compared to the epit axial growth and hence they often extend ed further above the flat portion of the substrate into the dep osition vapor allow ing faster nucleation in the ir inclined surfaces compared to the nucleation on the non - tilted (flat) surface. T heir conclusion was th ese particles may have been generated from impurities in the grown layer, rather than structural defect s penetrating from the interface. Alternatively , the elevated growth temperature could create a growth condition which was non - favorable for the (001) growth. Gauk rog er et a l. [117] studied X - ray t o pographs of CVD diamond grown on HPHT seeds to have a better explanation of clusters of dislocations that reach the surface and emanat e from the seed - grown layer interface. They found that the majority of the dislocations have a direction near [100] line directions. Generally, such dislocations propagate along the growth direction as the process minimizes total energy. Some d efects can originate from polishing damage . These were observed as 45° mixed dislocations wit h a Burger`s vector parallel to polishing line s . Overall X - ray topography is a powerful technique to understand the propagation of defects. It is also an effecti ve method to understand growth of a sample at different stages during the growth process , how ever, to p ography requires a relatively thick layer for analysis. Microstructures of thre a ding dislocations bundles were investigated using a combination of the etc h pit method and cross sectional TEM by Tsubouchi et a l . [118] . Cross sectional TEM allows better spatial 112 resolution as compared to X - ray topography. In the study by Tusbouchi t he etch pits were guiding point s for the locations where samples were carefully prepared for TEM analysis. The TEM micrographs showed the dislocation bundles were nearly parallel to film growth direction [001] and that they gradually diverge d . I n a different study from the same gr oup [119] , the authors focused on the evolution of threading dislocations during growth by performing an etch pit density analysis. High density dislocation bundles in the substrate generally act as t he source of dislocation defects that propagat e to the grow th s urface. These dislocation defect also show u p during birefringence pattern measurement as a butterfly pattern. During growth , many of these dislocations propagate in the growth direction, howev er some of these show more latera l extensions. The importance of H 2 plasma etching and a detailed description of etch pit analysis could be f ound in the report by Ivanov et al. [120] . This technique is relatively less complex in comparison to X - ray topography or TEM analysis. Temperature and misorientation angle can strongly influence etching rate and etch - pit shap es. Figure 6.2 shows an AFM image of a typical etch pit shape created on a (100) plane, presented by Ivanov. et al . [120] . 113 Figure 6.2. Appearance of an etch - pit , as viewed in AFM [120] Tallaire et al. further explored the commonly appearing defects in CVD grown diamond layer s by etching using a H 2 /O 2 plasma ( which can etch more effectively the defects ) and then performing regrowth [121] . The thre e major types of defects that appear ed during the process were unepitaxial crystals (UC), hillocks with flat top (FH) and the pyramidal hillocks (PH). A cross sectional TEM study by the authors of the grown layer revealed that UCs don`t have any direct roo t to the interface or the substrate, rather these defects nucleate on any contaminants (particles) that get into the layer during the growth process. The other two types, FHs and PHs are generate from extended crystallographic defects. The FHs and PHs con tain (111) microfacets. If twinning occurs during the growth, then a twinn ed crystal will be formed on top of the hillock and steps will get generated from the twin and flow along the side facets. The authors also found that the short 114 H 2 /O 2 etching method is a reliable and relatively easy technique to reveal dislocation defects and that it correlates with birefringence and cathode luminescence techniques . Further the etching technique also reveals shallow (both small and large) and deep (large) etch - pits. H owever, one has to estimate whether the etch pits or defects are fairly la rge as these shapes only relate to contaminants not the threading dislocations. The small square shaped etch pits ( deep and shallow) could be correlated to mixed and edge dislocatio ns propagating along <100> growth direction [122] . So, it is evident from the discussion above that t he growth of electronic grade diamond with both low nitrogen content and low dislocation defect density poses a hurdle . The current state of art shows that most of the CVD grown diamond still comes with dislocation densities ~ 10 4 /cm 2 as well as nitrogen i ncorporation levels above a few ppb . Lowering th ese number s r equires a systematic approach starting with using higher purity feed gases, a tight control on the growth parameters, a proper maintenance of the reactor chamber, and proper quality control towar ds seed selection. However, none of these processes allows on e to have a direct control over the defects especially during the initial nucleation stages. As mentioned earlier that an addition of nitrogen allows faster growth of diamond layers with fewer pr oblems originating from the dislocation defects, but this is at the cost of poorer electronic properties due to the incorporated nitrogen . Hence a way is needed to serve both purposes, i.e. outgrowing CVD layer s with low nitrogen without the formation of h illocks from the propagation of dislocations. Takeuchi et a l . proposed a roadmap explaining how misorientation angle and CH 4 /H 2 ratio variation during PACVD could produce different textured surface s over a small range of methane percentage (up to 2%) [123] . In their study the authors s how ed how to successfully grow atomically flat surface with a very minimal methane flow (~0. 02 5%). However, to achieve a higher 115 g ro wth rate they increase d the methane (up to 0.5%) while keeping the other growth condition the same , but this ended up having more defects (UC, FHs and PHs) on the surface. The study results reve a led that for low percentage methane flow (0.15 ~0.5%), UCs a nd PHs defects appear for low misorientation angles (<1.5 ° ) of the top growth surface relative to the (001) crystallographic plane . They also found that w ith an increase in the misorientation angle , i.e. 2~3° , and by adding slightly more methane ( up to 1%) , the growth occurred with more macro bunching steps but the surface was free of hillock features. Tsuno et a l. studied homoepitaxial film growth on misoriented substrate s , especially the dependence of the surface texture and growth rate on off - cut angle . The influence of off - angle on growth rate was more evident at low methane concentration (~1%) tha n compared to higher methane concentration (~6%). Also, the growth rate variation was very minimal at low temperature ( 8 00°C) but showed an increase with low to high off - angle for deposition at higher temperature (1000°C) with 1% methane . On the other hand, a growth at 6 % methane at lower temperature didn`t show much difference. The authors also achieved a smoother surface (hillock free) growth for a low offcu t angle with growth conditions of a low met h ane concentration and higher deposition temperature. 116 Figure 6.3. Growth rate vs misorientation angle toward [110] at (a) 6% CH 4 and 800°C, (b) 1% CH 4 and 800°C, and (c) 1% CH 4 and 1000°C as reported in referenc e [124] As a key observation reported by the au thors, at the low methane percent c ondition shown in figure 6.3 (c), a strong influence of the misorientation has been found. However, the influence is also dependent on temperature as shown in figure 6.3 (b) and (c) . A lower temperature and the same metha ne percentage as in 6.3 (c) showed that the effect of misorientation is nominal. Finally, in figure 6.3 (a) the effect of misorientation angle remains very minimal even with higher methane flow (6%). A possible explanation by the authors for the observation in figure 6.3 (c) is that there m ight be two - dimensional nucleation rate that is smaller compared to dimer formation at low methane percenta ge, which changes with the misorientation. The surface migration eventually saturates at higher angle s at this growth condition. Lee and Badzian [125] looked into surface morphologies by varying the me thane concentration from 1%, 2% and 6%. In deposition on misoriented samples (~ 15 °) the authors found that the 2% film produced lowest surface rough ness but had some pyramidal as well as flat head defects as seen in figure 6.4 . 117 Figure 6.4. DICM images of the surface morphologies of ( 001 ) homoepitaxial diamond films grown for 5h r at 875°C with CH 4 concentrations of ( a ) 1%, ,( b ) 2%, and ( c ) 6% in H 2 . [125] An increase in methane to 6% overgr e w these defects but ended up in more random features with a general flow along the misorientation direction. The growth with 1% methane show ed step flow growt h as adsorbates mainly diffuse into steps since the diffusion length of adso rbates are longer than terrace width. A 2% methane adsorbates start to form stable clusters. With further increase in methane to 6%, the adsorbates ha ve a much lower diffusion leng th and therefore random growth morphologies cover the entire surface. Van En ckevort et a l. performed a detailed study of diamond films deposited at misorientation angles relative to the [001] plane for diamond deposited by hot filament assisted CVD [126] . The authors concluded that homoepitaxial growth on [001] face is mostly guided by nucleation and propagation of <110> steps. However, in the process of (1) the nucleation happeni ng on dislocation defect sites , or (2) non - epitaxial particles being on the surface, the deposition cannot give rise to large hillocks. 118 Theije et al . [127] studied the eff ect of nitrogen on step bunching for misoriented substrates ( 1.5° and 7°) directed towards [110]. The authors successfully verified their experimental observation by a simulation to reach a good agreement between experiment and s imulation. According to the ir report, nitrogen`s presence causes a growth rate increase as well as step bunching. The step propagation on the surface become s slower due to the nitrogen present on surface, however the sub - surface nitrogen enhances the growt h by activating surface rec onstruction bonds with N - donor electrons. I t was already discussed in reference [51] earlier about the hillock structures observed by Bauer et a l. in their growth process on nominally [001] surface s . In the same article, the auth ors further discussed on ho w growth on misoriented substrates (up to 8°) worked as a better solution in achieving a defect free surface. A surface with a certain inclination angle from the crystalline normal plane can act as a source of new crystal lattice planes. In presence of vic inal planes, the lateral spread of the planes can help towards overgrowth on top the defects and provide a higher quality layer. The quality of the layers grown on misoriented substrates were further confirmed by narrow full widt h half max value derived fr om rocking curve measurement. I n more recent work, Tallaire et a l. extended their study on etching of misoriented substrate by using a H 2 /O 2 plasma to understand new surface features. They create d seeds with high offcut angles from the (100) plane ( up to 20° offcut) . The etching of diamond with a H 2 /O 2 plasma forms clear etch pits (generally square in shape) when the surface being etched is close to the (100) crystal plane, however the etch pit shape changes and r ecognizable etc h pit s becomes challenging for the misoriented surfaces . As the etch rate on the misoriented surface s are strongly influenced b y the angle of miscut, the steady condition to create a regular etch pit shape cannot be attained. Instead, the et ch process beco mes highly directional and etches more along the [110] direction 119 s . But, on an interesting note the authors found that for samples with high misorientation along the <110> di rection the etc h process produces surfaces with low surface roughness. Based upon their observation of the etching of misoriented substrates, Tallaire et a l. further continued to grow thick SCD layers on different misoriented samples, where the angle of misorie ntation va ried from 3.5° to 20° from both [ 100 ] . a nd [ 110 ] and a thick As a n interesting outcome, none of the final grown layer s had any impression carried from the misoriented substrate, which had different surface features due to etching. All the grown layers appeared smooth and defect free under the microscope , however the minimum thickness of growth to achieve a nearly flat surface varied largely with the misorientation angle (larger the misorientation the longer it is req uired to grow a flat surface). The thickness of the flat layer which directly correlates with the misorientation angle was further studied through a 3D modelling to idealize the different misorientation along [ 110 ] and [ 100 ] at different stages of growth. A detailed geometric modeling on the different growth shapes of CVD grown diamond in terms of growth rate of the main planes, [100], [110], [111] and [113] can be found in t he report published by Silva et a l . [128] . The model can be applied to growth of any crystal orientation to maximize deposition area. However, the model doesn`t address any defect growth that can deviate the process. The etch pit density of the CVD layers grown on the misoriented substrates by Ta llaire et a l. was quite high (10 6 ~ 10 7 /cm 2 ) as compared to generally reported value for CVD layers (~10 4 /cm 2 ). Also, the comparison between etch pits on a long [ 110 ] and [ 100 ] misoriented study was not fully conclusive. Therefore, even though etch pits can reveal a defect rich zone of a CVD dia mond layer and it can be correlate d with higher birefringence layers grown on misoriented substrates. 120 6 . 2 . Aim of current study Deposi tion of single crystal diamond on the (100) surface is a complex process at low nitrogen environment as in most cases t he surface gets covered with defects arising from dislocations and twinning present in the substrate. All these defects partially or full y slow down any lateral step flow growth on the surface. Hence the primary solution to tackle this problem is to grow s ingle crystal diamond on misoriented substrates. A misoriented substrate with offcut direction along the <001> or <110> directions offers adsorbates steps to be absorbed at so that a lateral flow growth occurs. Hence the higher lateral growth rate ensures that the lateral step flow grows over the defects (hillocks, flat tops) and eventually ends up as a defect free surface. Although these surfaces will have different surface structures (step - terrace distribution , step bunch ing) unless grown quite thick ( 50 It is also found in literature that on a surface with steps and terraces that impurities generally incorporate more into steps than the terraces. This may affect the electronic propert ies of the grown layer. So, the main focus of this sect ion of study is to understand the distribution (the mean values of step heights and terrace width) of step heights and terrace width with respect to different misorientation angles ( 2.5°, 5° and 10°) along <001 > and <110> directions . According to the lite rature, this is expected to generate a wide range of step and terrace distribution s with respect to the angle while keeping the other growth parameters fixed. These distributions or rather surfaces are very impo rtant to decide what kind of growth condition and misorientation angle one needs to pick up to achieve a dopant distribution for any specific device fabrication. This study only covers the effect of misorientation angle on step height and terrace width dis tribution s , which otherwise could also get mo dified with methane percentage and temperature. 121 6 . 3 . Experimental p rocedure 6 . 3 .1. SCD deposition process The single crystal diamond depositions performed in this study were conducted in a 2.45 GHz microwave cavity plasma reactor (MCPR). HPHT 1b type s ingle crystal diamond substrates from Sumitomo Carbides were used for all the depositions. Initially all the s ubstrates had dimension of 3.5 3.5 1.4 mm, with both sides polished. The substrates were checked with X - ray diffraction to ensure that their initi al miscut angle from the (100) crystallographic place was below 1°. The samples were next cut and polished to the desired offcut shape as shown in figure 6.5 . For bevel shape only, the top surface was cut (either along the [100] or the [110] direction) and polished to the desired angle. For the parallel substrate structure two sides were cut and polished. The fina l orientation of the crystallographic planes was verified again with a XRD measurement to ensure the desire offcut stays within ± 1°. The XRD meas urements were carried out on a Bruker - AXS D8 instrument. The schematics of the different offcut substrates are shown in figure 6.5 with a desired orientation of the offcut angle plane. For offcut directions along [110] only a half of the surface was tilted as otherwise it is difficult to cut and polish the entire top surface along the diagonal. After the proper la ser cut of the surface, the surface wa s polished to get an average surface roughness of 3 nm or less. 122 Figure 6.5 . Orientations of offcut substr ates along different directions from the (100) plane The samples were next cleaned in boiling sulfuric and nit ric acid mixture (1:1) for 20 minutes, hydrochloric acid for 20 minutes, followed by ultrasonication in acetone and methanol for 20 minutes each t o remove any impurities from the surface. Next, the prepared samples were loaded in a pocket recess of a molyb denum substrate holder as shown in figure 6.6 and placed inside the MCPR. The holder is built to minimize polycrystalline growth along the rim of the substrate. The two different parameters, i.e. depth d of the substrate pocket recess below the top surface of the substrate holder and the gap w from t he edge of the substrate to the edge of the pocket recess holding the substrate, are important to con trol the growth rate of diamond and minimize the growth of unwanted polycrystalline diamond. In these work the d values used were, 1.8 mm (for 24 hours lo ng growth on bevels), 1.6 mm (for 2.5 hours short growth) and 0.9 mm (for parallel offcut substrate s), whereas for all growths w was kept fixed at 1.5 mm. These holders are made from molybdenum and polished on both sides. The deposition reactor was pum ped down to reach a low vacuum and the leak rate was checked for each run to minimize any external le ak from the atmosphere to ensure that the growth takes place in a low nitrogen impurity environment. As a pre - growth treatment, the substrate was plasma etched for 1 - 1.5 hour at 180 Torr by a hydrogen discharge to get rid of (a) any damage caused by mechanical polishing, and (b) any 123 impurities that could still be present on the surface. Some of the established etching methods to remove defects gene rated from polishing damage include reactive ion etching and hydrogen plasma etching. For the etching step, care needs to be taken to keep the surface relatively smooth. H 2 /O 2 plasma has also been used as a pre - treatment method before growth [129] , but occasionally this method leaves deep etch - pits formed on the surface. Hence in all our growth process we settled for a 90 minutes H 2 plasma etching process (at T ~ 1030º C) as a pretreatment. This resulted in the removal of close t The growth chamber is then pumped down for at least an additional hour b efore moving to the actual growth process. All depositions of single crystal diamonds for this work were conducted at 240 Torr. The microwave power is in itiated to the reactor at ~5 Torr and the pressure/power is increased in suitable steps to reach a su bstrate temperature (1080°C~1100°C) range and maintained further to have a fixed temperature during the entire process. A short 10 min hydrogen plasma etch process at 240 Torr preceded the addition of methane that was then flowed into the deposition reacto r. The methane was slowly added (1% for 2 min, 2% for 2 min, 3.5% for 3 min and 5% for the rest of the run) to avoid any sudden spike of methane flow at th e initial stage of the growth. The deposition was then performed at a constant temperature, fixed tot al gas flow rate (5% CH 4 added to H 2 gas with the total flow rate of 420 sccm) and fixed pressure condition. The input power was maintained around 2 kW wi th minor adjustments to achieve the desired substrate temperature. The temperature was read by a one - color pyrometer (Ircon Ultimax, emissivity at 0.1). The introduction of the methane often gives rise to temperature increases and therefore the power cont rol dial need ed to be frequently adjust ed to get the right temperature during startup . This process s uffered from having lack of good control over temperature . There were two approaches taken to fix this issue; (i) an Arduino uno was paired with a stepper motor and a rotary 124 encoder to precisely rotate the power dial and (ii) methane was fed into the chamb er in a controlled amount as 1% for 2 minutes, 2% for 3 minutes, 3.5% for 3 minutes and then 5% for rest of the growth time. This recipe with a precise con trol in the dial has been quite successful in avoiding any sharp rise in temperature and allowing the temperature to slowly rose to its target value. After completion of the deposition process, the final substrates were cleaned in boiling nitric and sulfuric acid mixture (1:1) for an hour followed by ultrasonication in acetone and methanol. Before and a fter each deposition the thickness of the substrate s was measured by linear encoder (Solartron DR600) and were weighed using a micro balance (Mettler Toledo XS105DU, with accuracy of 0.0001 g). The thickness was measured by linear encoder near the edge of the substrate and at the center of the substrate be fore and after the growth process to estimate the distribution of the growth rate, especially for the bevel cut substrate offcut along the (100) direction. For (110) offcut direction substrates all the thi ckness was measured along the diagonal of the subst rate. The same procedure was followed for thickness measurements in parallel offcut substrates. The deposition duration times used were 2.5 hrs, 12 hrs and 24 hrs. The growth rates at different points of the substrates were calculated by the average grown layer thickness divided by the growth duration. A Nikon ME600 optical microscope and an SZM stereo microscope are used to look for the surface texture and features. The surface textures were further quant ified using a Dektak surface profilometer and Hitac hi 5100N atomic force microscopy. 125 Figure 6.6. Schematic of a diamond deposition process showing the pocket recess t hat holds the diamond substrate 6 . 4 . Optical c haracterization of offcut grown samples 6.4.1. Operation of optical microscope An optical microscope is an important tool to analyze features that is present on the surface of the substrate or grown layer (when a snapshot of the top surface is collected from the light reflecting from the surfac e) or even the strain present inside a f ree - standing layer ( through the light transmitting through the sample and its cross polarization). The first method is applied either for dark field imaging or bright field imaging or occasionally differential interf erence contrast imaging (DICM) where in all these cases, lig ht is illuminated from the top of the sample through a selective lens and the reflected light is focused through the objective lens. For DICM a prism is brought in the pathway of the incident ligh t which based on its position can split the beam into to par tial beam separated by wavelength. 126 In the light transmission method, i.e. mostly for birefringence analysis, light from the bottom of the sample travels through it and two polarizers are placed w ith their polarization axis at 90 ° to each other on either s ide of the sample . Any non - refracted light will be blocked by the polarizers and will appear dark in the eyepiece or any captured image. However, any strain in the substrate/sample will result in additional refraction and therefore the refracted light will propagate through the second polarizer. These strained region s will appear very bright in the captured image. Quantitative b irefringence analysis has not been conducted in this work , however qual itative images are used in some selected cases. For the opt ical characterization in t o this work, a Nikkon eclipse ME600 optical microscope was used. Th is particular instrument is configured with 2.5X, 10X, 50X and 100X objective len s es . The following pair of schematics in figure 6.7 (a & b) show the two basic oper ational mode of the optical microscope. The first schematic shows how a birefringent prism (4) is adjusted to split the incident beam into two beams differed by wavelength. When the two beams stri ke a rough surface, due to their lateral displacement reflects back with a path difference. Once the beams are converged back through the prism and 127 F igure 6.7. (a) S chematic of differential interference contrast imaging (DICM) m ode (b) Dark field im aging mode [130] the analyzer lens (7) the path difference gets co nverted into grey values and appears as distinct feature in the eyepiece or collected image. The auxiliary plate (7a) can be adjusted to change the (b) (a) 128 grey values into different colors. In the next schematic, i.e. Dark F ield (DF) imaging, a condenser aperture is inserted to block the di rect beam. A 45 ° mirror is used to focus the secondary beam on the object and then finally the reflected light ent ers the objective lens and is directed to the ey e piece or camera to capture the image. 6. 4 .2. Observation s and discussions A standard 3.5 3.5 mm 2 commercial HPHT substrate (from Sumitomo Carbide) , and a commercially available CVD substrate are shown in figure 6.8 after the acid cleaning process. Generally, these substrates come with a thickness around 1.4 mm , however based upon the offcut preparation approach the thickness varied. Th ese images show that the substrates are of good quality i.e. without any visible defects and properly polished along < 100> or <110> directions . This image is captured with a 25X magnification setting (the The corresponding pair picture shows the birefringence image captured using the cross polarizer. 129 Figure 6.8. An image (25x) of a polished HPHT substrate and corresponding birefringence (top row) and image (25x) of commercial CVD pl ate and correspon ding birefringence (bottom row) It is evident from the bottom pair of pictures that surface damage occurred during polishing on the CVD plate (numerous scratches are found along the d iagonal and this has also resulted into strain present closed to the surface (red circles). There are some threaded dislocations also seen on the surface (which appears almost like four lobes 90 ° apart, as seen in red circle d locations ). Altogether these p air of diagrams help one to decide about the feasibility of a substrate selection, and an expectation of a good quality of deposition. However, there are other parameters that also play crucial role s during growth of a defe ct free thick diamond layer nam ely; substrate temperature, energy density of plasma, purity of feed gases as well as amount of nitrogen present in the growth environment. There ha ve been multiple reports [111,13 1] that show how additional nitrogen in the growth process aids towards higher 1 130 quality of surface finishing as well as higher growth rates. H owever, any deliberate addition of nitrogen comes with the cost of a substantial de terioration of electronic prope rties in the crystal. Therefore, in a low nitrogen environment. But growing thin epi - layer or thick layer of diamond in low nitrogen environment is challenging as the surface in most occasions gets filled up with defects, as seen in figure 6.9 . Figure 6.9 (a) shows a 25X magnification and (b) shows 100X Figure 6.9. An optical micrograph ( 25 X ) of an epilayer deposi ted on a commercial CVD plate ( left) and the same surface looked at hig her magnif ication (100X) magnification - layer deposited on a commercial CVD substrate. As evident from the higher magnified image the surface has p yramidal and flat top structures. The flat top structures or the pyramidal structures mainly appear from extended defects, i.e. edge di slocations [122] . In general, these defects propagate all the way to the grown layer especially when gr own in a very low nitrogen impurity environment and they end up as a pyramidal or flat top type structures on the deposited layer. A remedy to this problem is to deposit diamond on a misoriented substrate. A growth mechanism on a misoriented surface versus an almost flat surface had been discussed by Lee at. al [132] . 131 To avoid the pyramidal and flat top structures at low nitrogen growth conditions this study used offcut angles of 2.5, 5 and 10 degrees. The optical microgr aphs in figure 6.10 shows the surface profiles of substrates grown at different offcut angles for different durations. 2.5° 5° 10° 2.5 hr (a) (b) (c) 12 hr (d) (e) (f) 24 hr (g) (h) (i) Figure 6.10 . Top surface of SCD de posited layers for different offcut angles ([100] direction bevel) & different growth durations. The thinner edge of each beveled sample i s facing the bottom of each micrograph All the substrates in figure . 6.10 had very minimal polycr ystalline diamond gro wn on the top surface or at the edges of the top surface. However different surface features evolved with longer duration of growth and higher offcut angle. Also, it is important to note that corners of substrates (a, b and c) were pur posefully laser cut/m arked and therefore the grown top surface`s appearance and the black corner are not related. However, there are two distinct features observed in the figures 6.10 (f) and 6.10 (h) where a dark structure has encroached partially in the grown sample. By 132 care fully adjusting the focusing on the microscope it was found that the dark patch is present close to the interface where the growth began, and not on the top growth surface. A possible explanation of this dark patch is that a layer of c racking starts at the edge of the sample and propagates inward. These darker patches do not appear to interfere with the diamond growth, so the cracking probably occurs at the end of the run when the substrate cools after the deposition. This phenomenon is not observed on the other substrates. Further depositions were carried out on two different types of offcut substrates as shown in figure 6.11 , including (1) on substrates with a bevel structure offcut along the [110] direction (Figs. 6.11 (a) and (c)) an d (2) on parallel cut substrates with the offcut along the [100] direction (Fig. 6.11 (b) and (d)). The thickness of these grown layers on the Off - cut type /Off - cut angle (110) bevel cut (100) parallel plate 5.0 ° (a) (b) 10.0 ° (c) (d) Figure 6.11 . Top surface of deposited layer on [110] beve l and [100] parallel substrates A feature is found along the thinner edge of 10° offcut bevel [110] direction and [100] direction substrates ( figures . 6.11 (c) and 6 .10 (i)), i.e. the thinner edge appears rougher than the 133 rest of the samples. A possible cause could be that the plasma doesn`t cover the thinner edge uniformly during deposition. This effect is very minimal on the parallel substrate growth ( figures . 6.11 (b) and 6.11 (d)). An interesting feature that is observed on the top surface of all the bevel cut [100] samples i s a patch of area that grows out flatter (closer to the (100) crystallographic plane) on the section that has the highest crystallographic plan es. This indicates the addition of new crystal planes occurs slower than the lateral growth along the step - terrac e region of the sloped offcut surface. Fig. 6.12 (a) & (b) shows the top and side views of 24 hours growth runs with the offcut of 2.5° at the top right corner to 10° at the bottom left. The lateral growth becomes more prominent on the side view of the sur faces. Figure 6.12. The top surface and side view (a & b) of three different offcut angles (2.5°, 5° and 10° along [100)] bevel) showing t he lateral growth of the deposited layers. All pi ctures are for 24 hours growth (a) 134 Figure 6.12 ( cont`d ) The lateral area also partially extends out from the substrate to the left in the figure 6.12 (b) pictures, which indicates an area gain from the orig inal substrate top surface to the growth surface. The notion of lateral area growth/increase becomes clear from the magnified side view of all the grown surfaces as shown in figure . 6.12 (b). Referring to the top grown surface, the 2.5° shows a low lateral growth (half compared to the other offcut angles), which is possibly due to the substrate being 150 inside the pocket recess of the holder as shown in figure 6.6. The other two offcut growth experiments in figure . 6.12 show nearly the same length of lateral growth (however diff erent surface texture). We also looked into the vertical growth rate (V P ) and the lateral growth rate (V L ) of these three substrates. ( b ) 135 Table 6.1 . Lateral and vertical growths on di fferent offcut angle substrates Offcut angle Lateral growth rate (V L ) in Vertical growth rate (V p V L /V P 2.5° 16. 3 8.0 2.03 5.0° 33.7 14.9 2.25 10.0° 29.4 14.6 2.02 To measure the lateral growth rate, the corresponding length of lateral expansion (from figure 6.12 (b)) is indicated on the top of the side view of the sample. A line was drawn from the tip of the length to the edge of the substrate to find the segment th at had grown laterally. This length is used to estimate the lateral growth length and then the lateral growth rate. The ratio of lateral gro wth rate and the vertical growth rate shows a nearly constant value for this particular pocket depth as seen in tabl e 6. 1. Hence it can be stated that the ratio stays almost independent of the offcut angle for this set of growth conditions. The overall v ertical growth rates of the substrates with different offcut angles are summarized in figure 6.13 . As mentioned earl ier the thickness were measured by a Solartron DR600 linear encoder. For each substrate, data points were collected at multiple positions (a ll edges and the center for beveled [100] offcut direction substrates and along the diagonal for [110] offcut direct ion substrates) on the surface and averaged over three times at each location . This was done both before deposition and after deposition to get the growth rate at each point. 136 Figure 6.13 . Vertical growth rate of deposited layer on different offcut samples measured at different locatio ns for 24 hours deposition Figure 6.13 shows there is minimal variation in the growth rates at different lo cations on the substrate. The slight variation in the growth rates are partially due to surface morpholog y arising from the angled substrate and the overall roughness . However, there is no large variation found in the growth rates within each individual su bstrate. One difference is the marginally higher growth rate is observed in the [100] bevel direction gr owths (top right plot in figure 6.13 ) ). This possibly occurred because of a slightly higher N 2 presence in the plasma during deposition. The reactor, e specially the gas flow system had gone through some minor upgrades after the [100] bevel direction runs a nd before the other runs. Typically, a slightly higher nitrogen impurity level in the deposition produces a higher growth rate if all other parameters are held unchanged. 137 6 . 5 . Study of d istribution of step and terrace growth The growth of homoepitaxial ( 001) diamond is guided mainly by two processes, formation of new islands (planes or terraces) and propagation of steps at the end of crystal planes (te rraces) [126] . It is noteworthy that the formation of new islands (planes) of diamond also is enhanced due to presence of defects. This island formation is at a faster rate in de fect regions a s compared to low defect concentration region s [133] . The defect region can be from either threading dislocations or from foreign particles present on the substrate surface. A surface with a low offcut angle (< 0.5°) is alwa ys prone to en d up having pyramidal or hillock type defects since in such cases there are few steps, and the island growth from defects takes a dominant role. In general, the suppression of these hillock structures happens for offcut angles above 2°. In th is section we look into how steps and terraces arise out of different offcut angles and the effect of the offcut angle on lateral growth (100) in comparison to perpendicular growth (001). Step and terrace structure appearance is a common phenomenon for cry stalline sampl es. In general, any crystalline surface is comprise d of both micro and macro steps . M acro steps mainly grow due to step bunching of micro steps. Also, the offcut from the (100) plane initiates the micro steps and based upon the angle of incli nation and the exposure of the top surface to the plasma, the growth proceeds by further propagation of the steps in the [100] and [110] directions. 138 Figure 6.14. High magnification (500x) pictures of step - terrace formation (a) 5° bevel (c) 10° with the bevel along the [100] direction and the corresponding [110] direction bevel at (b) 5° and ( d) 10° The behavior of the step and terrace size distribution as well as the bunching is determined by the local growth conditions of the deposition plasma. This growth is affected by the shape of the top surface inside the pocket recess with respect to the depos ition plasma, i.e. the bevel cut shown in figure 6.14 can experience different plasma conditions across its surface as compared to a parallel cut substrat e due to non - uniform plasma interaction due to the slope. As evident from figure 6.14 an offcut angle close to 10° ends up to a very rough surface irrespective of its direction of inclination. 139 Figure 6.15 . Step and terrace propagation along (a) 5° and (b) 10° offcut parallel substrate growth However, looking into the surfaces of parallel substrate s with offcut angles, it is evident that a uniform plasma - substrate interface creates a more uniform flow of steps and terraces. Although there is a significa nt difference among the two surfaces shown in figure 6.15 , both show a steady terrace and step propagation along [100] direction across the entire surface. However, the separation of the terraces and their bunching looks much different in the 5° compared t o 10°. The 10° offcut parallel substrate shows some micro - steps on top of macro structures. Also, the micro - steps are more densely situated in the 10° offcut sample. This makes the 10° offcut substrate having higher surface roughness compared to the 5° off cut substrate growth. Therefore, a collective observation of all these different kinds of offcut angle growth s generates a keen interest to 140 understand more about the distribution of the steps and terraces with respect to the type of inclination of the surf aces. A quick observation of all the surfaces suggests th at the terrace s and steps grow very differently for bevel structures of different angles and therefore the local impurity/doping concentration variations in the grown layer may be dissimilar on all t hese different surfaces. Hence this section will focus on surface topography characterization to quantify the overall distribution of step heights and terrace widths. Figure 6.16 shows a first attempt to quantify the steps and terraces, an AFM scan of the thicker flat region of the sample in Fig. 6.10 (i). A Hit achi 5100N atomic force microscopy with 20 nm tip is used in DFM mode for this scan. A 20 2 area is scanned on the flat part of the sample. The average surface roughness for this area is ~ 40 nm. In the same set, figure 6.16 shows the step and terra ce distribution along selected line scans shown as red and blue lines . The step height s vary between 90 nm to 150 nm and the only, which allows selected pre cise detection of steps and terraces but not a statistically significant full distribut ion. It was difficult to measure areas outside this flat region as the surface is too rough and sloped for the AFM instrument. 141 F igure 6.16 . Surface roughness profile of a selected area (f lat region) of 10° bevel growth A Dektak profilometer was then use d to measure the rougher surface s . A Veeco Dektak 6M Before pro ceeding to the next part, it is important to have a brief discussion about how a stylus based DEKTAK profilometer works. 6 . 5 .1. Basic operation of DEKTAK profilometer The DEKTAK surface profilomet er is a common instrument to analyze surface characteristics where a stylus with a small diameter, made mostly from diamond or other material moves over a selected scan length (in contac t mode). The displacement of the spring loaded stylus 142 is translated to a Z variation (height measurement) of the features under study. An el e ctro - mechanical transducer is attached to the stylus which converts any variation of mechanical force to electrical signal, better known as linear va riable differential transformer [134] . Figure 6.17 shows a schematic o f the basic Figure 6.17 . A schematic of the operation of st ylus - based surface profilometer [134] operation of the tran sducer. The sensitivity of the transducer is approximately 1 Å, where the stylus i s attached to a ferrite core which ultimately forms a n AC bridge. Any movement in the stylus causes change in the balance of the bridge. Any demodulation or amplification of the signal is converted to a voltage which is finally converted by analog to digital converte r to read th e value that is then stored by software. The sensitivity of the measurement largely depends on the stylus tip and the stylus force. Typical tip diameters are 2 ~ 10 . Any broader tip otherwise will miss finer sur face feature s and higher force could be detrimental both 143 for the film under measurement as well as lifespan of the tip. Figure 6.18 shows the adjustable stage and other controls of a Veeco 6M profilometer [135] , which is used for this study. Figure 6.18 . Image of the measurement stage of Veeco 6M profilometer [135] A software allows the stylus to be l ower ed down to or raised up from a sample. T he x/y - direction knobs help to center the sample (especially for small size samples). The scan length, scan time, and resolution c an be adjusted by the software program . For the measurement s of th is project a th the time of scan set to 140 sec to get the maximum 144 6 . 5 .2. Observations and discussions diameter tip and 5 mg force was configured to the inclination of the surface. The steps and terraces are scanned with the probe tip and then the step height and terrace width are calculated as shown in figure 6.19 . This figure shows the scanning over a rougher area. Often the traced curved was leveled to identify and quantify the terraces and steps as shown in figure 6.19 (b). Measuring micro - steps was challenging for this tip, and therefore software plots the raw data and the step and terrace sizes were determined from the plots manually. Figure 6.20 shows the distributi on of terrace and steps over a small scanned range, from the sample with 10° offcut (same sample as shown in figure 6.10 (i)). This figure basically serves to illustrate the process followed for step/terrace tracing and their height and width calculations. Five scans at different spot s of the sample (mostly at/near the center, unless specified) were performed and a distribution frequency of the step height and terrace width was constructed as shown in figure 6.20 . Each bin in the x axis of figure 6.20 repre sents either a range of terra ce width s or step height s . The y - axis shows the distribution frequency of the corresponding bin. 145 Figure 6.19 . Schematic showing leveling (from (a) to (b)) and measurement of step height and terrace wid th on a 10° offcut bevel sample Figure 6.20 . S tep - terrace distributions (bar diagrams) at different sections of the 10° offcut [1 00] bevel direction grown layer ( a) ( b ) 146 Figure 6.20 shows an example of the step height and terrace width distribution analyzed and represe nted in a diagram from different regions ( as pointed by the arrows) of the substrate grown at a 10° offcut and beveled along the [100] direction. The microscope picture itself shows differences among the selected regions of the sample. As seen in the thick er and flatter portion of the substrate, t he step heights are maximum around the 50~100 nm bin , and terrace widths have a distribution that is highest in the 0 - bin (inset picture). The other two distribution plots required that the step height distrib 3 nm) scale to accommodate all the distribution, which cle arly indicates these regions are rougher. A notable feature is that comparing the mid region to the thin (right) region, one finds the bar diagram has wider distributed tail for the thin region (in other words, it is slightly rougher from the mid region). Also, the terrace e overall di stribution helps to quantify the observation from optical micrographs that a 10° offcut sample becomes rougher at the longer run times. 147 F igure 6.21 . Step - terrace distributions (bar diagrams) at different sections of a for a 5° offcut [100] bevel grown layer The distribution is different and more uniform when we investigate the 5° bevel offcut from the [100] direction substrate growth, as sh own in figure 6.21 . The same approach (i.e. looking into three regions of the sample) sho ws that the step height distribution marginally changes from the thicker edge, which otherwise makes the overall sample much smoother compared to the 10° offcut substr ate. Also, this results in the terrace width distribution being more uniform. (except at the extreme corner on the thinner end). This general uniformity is likely due to the fact that in the pocket, for a 5° offcut bevel, the plasma can cover more uniforml y the full sample and hence a better plasma - substrate interaction for diamond growth. 148 In figures 6.22 and 6.23 a comparison is made of the diamond growth for the [100] bevel direction, [110] bevel direction and [100] parallel substrates for 5° and 10° off cut angles. In the first set of optical micrographs in figure 6.22 the results for 5° offcu t experiments are presented. The two substrates in Fig ure 6.22 (a) and (c) are both for an offcut direction of [100] with either a bevel or a parallel cut. They show b asically the same step flow growth surface morphology, whereas figure 6.22 (b) shows a [110] oriented offcut growth with a slightly different morpholog y. Comparing the optical micrographs figures . 6.22 (a) and 6.22 (c), as well as their corresponding step - terrace distribution, it can be seen that the mean terrace height for both these surfaces are around 200 nm, whereas for the figure 6.22 (b) case it is close 300 nm, which makes the [110] offcut direction beveled surface slightly rougher. For the [100] and [110] oriented bevel structures ( figure . 6.22 (a) & (b) ) for the [100] para llel substrate growth, terrace width distributions are slightly extended with a 149 Figure 6.22 . Step - terrace structures and the ir distributions (bar diagrams) for 5° offcut (a) (100) bevel (b) (110) bevel and (c) (100) pa rallel substrate thick growth On the other hand, if we look at the 10° offcut samples in figure 6.23 arranged in the same sequence as the 5° offcut substrates in figure 6.22 , a major difference is that the bevel offcut substrates in figure 6.23 (a) and (b) are very rough. Both of these grown layers have mean step the surf aces very rough. But the [100] parallel cut sample in figure 6.23 (c) is smoother. The mean st ep height drops back to 150~200nm, even though it has some long terraces but overall it is comparable with the 5° substrates. Further comparing, the [100] 5° parallel substrate (Fig. 6.22 (c)) and [100] 10° parallel substrate (Fig. 6.23 (c)), both have near ly the same range of step height and 150 terrace width. In the case of 10° beveled substrates in figure 6.23 (a) and (b) the dominant effect is believed to be the non - uniform plasma - substrate interaction because of the sloped top surface of the diamond substra te. Again, it shows that up to 5° for these 3.5x3.5 mm substrates the effect of such non - uniformity is not that severe. One can also notice that although figures 6.22 (b) and 6.23 (b) represents growth on same kind of substrates (5° and 10° [110] bevel di rection cut) however the result is very different. The 10° offcut growth (Fig. 6.23 (b)) doesn`t clearly follow any strong direction of lateral growth, whereas the 5° offcut (Fig. 6.22 (b)) shows a clear directional orientation of the step flow. The non - un iform plasma coverage might have affected the 10° [110] direction bevel as it could have done for the [1 00] direction 10° bevel growth (Fig. 6.23 (a)), which makes the current growth condition non - favorable for 10° or higher bevel cut in any direction. 151 Figure 6.23 . Step - terrace structures and their distributions (bar diagrams) for 10° offcut (a) (100) be vel (b) (110) bevel and (c) (100) p arallel substrate thick growth Finally, all the average step height and terrace width values for different deposition s are listed in table 6. 2. For the 2.5 ° offcut case, the [110] bevel and [100] parallel substrate exper iments were not conducted since the 5° experiments for the different cases showed little difference. Both [100] and [110] bevel directions for the 10° o ffcut substrates ended up having very rough surfaces urfaces had long terraces. The terrace width also depends on the duration of growth, as it is seen that all of the 24 hr grown substrates have smoother surface s (as step heights are generally less than 400 nm). The longer terrace width is an indication of lateral growth; however, accumulation of more micro steps decides whether it will 152 be rougher or smoother, which happened for the 10° bevel growths. Also, the n umbers suggest that [110] bevel direction growth comes with slightly more surface roughness as compared to [100] bevel direction growth. The 5° and 10° parallel substrates grown samples (Figs. 6.22 (c) and 6.23 (c)) do not show any major differences in ter ms of the numbers and are relatively smooth. Table 6.2 . Average step height and terrace width distribut ion for different offcut growth Offcut type & angle Growth durations 2.5° 5° 10° [100] bevel /2.5 hr Step height 24 nm 8 nm 261 nm Terrace width [100] bevel /12 hr Step height 53 nm 26 nm 234 nm Terrace width [100] bevel /24 hr Step height 326 nm 176 nm Terrace width [110] bevel /24 hr Step height 317.1 nm Terrace width [100] parallel /24 hr Step height 181.4 nm 202.6 nm Terrace width 153 In first part of this work, the performance of detector s made of single crystal diamonds were tested by irradiating them with swift heavy ion (SHI) beams. These beams consist ed of energetic ions from heavy nuclei with energy in the range of 100 - 150 MeV/nucleon. In such a harsh environment, established Si based detector technology will not survive for long when exposed to high fluence , hence diamond is a better choice as a detector material i n such cases due to its outstanding material capability. 7.1. Conclusions In this work we have looked into how gradual degradati on of single crystal diamond happens while interacting with SHI beams. The generated response in the material to high energy particles is converted into a readable voltage through an external electrical circuit. A signal strength (amp litude) drop was obser ved versus particle fluence for detectors made from both commercial and MSU lab grown diamond. The major conclusions made from the degradation measurements are; The percentage of amplitude drop as a function of particle fluence follo wed a standard paramet ric equation (hyperbolic) of the same nature that followed the distribution of CCD vs particle fluence, proposed by the RD42 group [3] . After irradiating with particle fluence up to 3 10 13 / cm 2 the detectors made of MSU lab grown di amond still retained 5 5% of the initial signal (amplitude) strength. Post irradiat ion the diamond - based detectors were analyzed by their transient current response, where the charge transport of the lightly and heavily irradiated segments were compared. A s a 154 standard reference the response of a commercially available non - irradiated electronic grade sample was used. The key observations made from this section are, The lightly irradiated and heavily irradiated segment s of the commercial diamonds showed near ly the same charge dri ft time, however the amplitude of the transient current signals were slightly lower in the heavily irradiated segment. The carrier lifetime dropped to almost 40 - 60% in the heavily irradiated segment as compared to the lightly irradia ted segment. This coul d be due to vacancies or traps created by the irradiation in th at region of the diamond. No significant s tructural damage was measured. Hence it is concluded that structural damage happens at even hig her fluence levels. I t is also concluded that the perfo rmance degradation is mainly due to the creation of vacancies /traps . A measurement of dark current or leakage current show ed that the leakage current is o n the order of 10 - 12 A within the range of applied voltage ± 400 V. However, a non - linear rise in the current takes place above +400V. This behavior appears to be similar to leakage current observed for polycrystalline based pixel detectors. The non - linear rise could happen due some dislocation defects in the diamond. It was also observed that the leakage current varied in the presence of light from a red LED placed on top of the detector. In all the tests, a transient response occurred when the red LED was turned on i.e. there was a quick rise o f the current which slowly decay ed back to a normal value. Th e transient response decayed faster in the heavily irradiated segment than compared to the lightly irradiated segment. This could be possibl y due to the presence of more charge traps/ vacancies present in the heavily irrad iated segment. 155 To have any idea o f degradation in terms of optical properties or structural in the irradiated samples, their optical response (% Transmission or absorbance) were measured by UV - Vis and FTIR spectroscopy. Generally, a type IIA single crysta l diamond show s nearly 70% transmi ssion in the visible and partly into the UV range other than a strong transition at the bandgap energy nea r 22 5 nm. As found from the % Transmission curve of both heavily irradiated segment s , a strong absorption is found around 350 nm, which is due to for mation of color centers in the diamond. None of these irradiated diamonds showed any additional absorption peaks in infrared region, other than the characteristic sp 3 peaks. Further the samples were characterized in high resolution XRD (HRXRD) and Raman s pectroscopy to understand if there had been any indication of structural damage in the samples. For HRXRD, any change in the ( 004 ) peak position could be related to as the onset of structural swelling or damage. As per our observations made in this section , There had been no appreciable change in the ( 004 ) peak position while measured across different spots in the lightly and heavily irradiated segments of the beam irradiated diamond plates. However, within the resolution of the beam spot dimension, a mar ginal increase in the full width half maximum (FWHM) of the ( 004 ) ] peak was noticed in the heavily irradiated segment . However, this requires further investigation by adjusting to a better collimation in the X - ray source side to measure a smaller beam spot . For Raman spectroscopy, the initial state of a lab grown diamond was a strong 1332 cm - 1 peak and some additional p eaks around 1430 cm - 1 and close to 1750 cm - 1 . The peak around 1400 cm - 1 could have appeared from some defective phases formation at the ear ly stage of 156 CVD growth (also observed at polycrystalline CVD). There is another around either 1730 cm - 1 or 1755 cm - 1 . Any peak present in the first position is due to C=O pea k which might have become stretched [ 136] . Any peak in the second position normally happens due to any defect during ion implantation process, however se cond case has almost very remote chance. An overall observation from this work is that the MSU lab grown diamonds ha d higher nitrogen impur ities than suitable for th is detector application. Therefore, it was difficult to study the transient current response i n these samples. Most of the generated charges got trapped resulting in a narrow transient current signal width. Therefore, for lab grow n diamonds at MSU to be applicable for detector operation, the growth needs to be done in a low er nitrogen impurity environment. Hence for further growth of substrates for detectors , a reacto r that has a very low external leak rate ( 30 ~ 40 mTorr/week) was se lected and diamond was deposited. Some o f the major outcomes from this section of the work are; It is extremely difficult to grow a scd layer on a substrate which has its top growth surface orientation very close to [001] crystallographic plane . Defects f rom dislocations and twinnings propagate perpendicularly and end up showing on the final la yer as pyramidal structures and/or defects . Growth on misoriented substrates showed step flow growth and it allowed growth defect free thick layer s of scd. The fin al surface of misoriented growth appears to have steps and terraces. Based upon the misorientation angle, methane percent, and growth temperature the step heights and terrace width will vary. In this work we studied the distribution in step height a nd terr ace 157 width with varying misorientation angle s from the (001) plane along the [ 100 ] and [ 110 ] directions . Also, it was found in literature that dopants like boron and phosphorus tend to settle and migrate to steps more than the terrace region. [137] , [138] . This can lead to nonuniform doping and impurity distributions in the grown diamond. So, it may happen occasionally that, in presence of an exte rnal el ectric field there could be non - uniform charge transport across the sample. Especially if the terraces regions are far apart , then any device made from such layers could suffer from reduced/non - uniform breakdown voltage. 7.2. Future i nvestigations E lect ronic grade diamonds offer high charge collection efficiency and superior charge collection distance in high energy particle detectors. In order to grow the higher quality diamond the process involves optimization of different physical parameters , e .g. in the misoriented substrate process one needs to ensure that there will be reasonable vertical and lateral growth with steering of defects at the same time. Therefore, it is important to establish a systematic investigation of defect propagation with change in misoriented angle substrates. There are some limited studies already i n the literature about how X - ray topography or TEM could guide such a study. However, both processes are quite complex. Hence etch - pit analysis comes as a n easier solution. There are only a very few literature reports that ha ve tried to analyze etch pit di stribution with varying misorientation angle. This needs to be explored to understand how the etch pit densities change under different misorientation angle. In terms of growing elec tronic grade diamond, it will be also an important study to check if the charge transport properties have any direct influence under misorientation angle . Hence one 158 immediate extension of this work will be to try to grow diamond for longer time to reach a point where the surface starts to appear flat, which depends upon the mis orientation angle. Finally, for a superior quality of diamond, a substrate with four fold bevel cut (otherwise defined as pyramidal - shape substrate , [139] which is expected to steer the dislocations towards the edges, needs to be explored. 159 BIBLIOGRAPHY 160 BIBLIO G RAPHY [1] https://www.puregrowndiamonds.com/blog/the - birth - of - a - diam ond . [2] R.S. Balmer, J.R. Brandon, S.L. Clewes, H.K. Dhillon, J.M. Dodson, I. Friel, et al., Chemical vapour deposition synthetic diamond: materials, technology and applications., J Phys Condens Matter. 21 (2009) 364221. doi:10.1088/0953 - 8984/21/36/364221 . [3] W. Trischuk, & RD42, Recent Advances in Diamond Detectors, (20 08). [4] [5] A. Oh, Diamond particle detectors systems in high energy physics, J. Inst. 10 (2015) C04038 C04038. doi:10.1088/1748 - 0221/10/ 04/C04038. [6] D.K. Avasthi, Modification and Characterisation of Ma terials by Swift Heavy Ions, Defence Science Journal. 59 (2009) 401 412. [7] D iamond Structures, Http://www.hardasrocks.info/diamond - structures.htm. [8] J.J. Gracio, Q.H. Fan, J.C. Madale no, Diamond growth by chemical vapour deposition, J. Phys. D, Appl. Phys. 43 (2010) 374017. doi:10.1088/0022 - 3727/43/37/374017. [9] NRC: Radiation Basics, Radiation Basics, United States Nuclear Regulatory Commission. (2017). https://www.nrc.gov/about - nrc/ radiation/health - effects/radiation - basics.html (accessed Novem ber 30, 2018). [10] G.F. Knoll, Chapter 4. General properties of radiation detectors, in: Radiation Detection and Measurement , 3rd ed., 2000: pp. 103 119. [11] R adiation D etectors and D osimetry , http://www.phy sics.usyd.edu.au/super/ life_sciences/ AN/AN5.pdf . [12] O regonstate .edu/instruct/ch374/ch 418518/Chapter 18 Radiation Detectors.pdf . [13] M. Marcel, Detectors of Radiation , oldweb.reak.bme.hu/Wigner_Course/2004/ WignerManuals/Bratislava/detectors.pdf. [14] I nteraction of radiation with matter,www.physics.edu.au/super/life - sciences/A N/AN4.pdf [15] J.F. Ziegler, Stopping of energetic light ions in elemental matter, J. Appl. Phys. 85 (1999) 1249 1272. doi:10.1063/1.369844. [16] T. Stefan, Interactions of Particles in Matter, S. Tavernier, ExperimentalTechniques in Nuclear and Particle P hysics, 2010. 161 [17] The stopping of charged particles in matter, gapsing2.uniroma1.it/aschiavi/Pubblicazioni/ /PhD/Chapter_4.pdf. [18] E. Garutti, The Physics of Particle Detectors, Le cture Notes, http://www.desy.de/~garutti/ LECTURES/ParticleDetectorSS12/L1_Introduction_HEPdetectors.pdf [19] G.F. Knoll, Chapter 2. Radi ation Interactions , in: Radiation Detection and Measurement, 3rd ed., 2000: pp. 29 57. [20] K.A. Olive, Passage of particles through matter 1, (2014) , http://pdg.lbl.gov/ 2014/ reviews/ rpp2014 - rev - passage - particles - matter.pdf . [21] S. N. Ahmed, Section 2.4 ., Interaction of Heavy Charged Particles with Matter, in: Physics and Engineering of Radiation Detectors, 1st ed., 2007: pp. 105 121. [22] S. Kumar , Proceedings of the DAE - BRNS Symp. on Nucl. Phys. 61 (2016 ) [23] T.A. Grotjohn, J. Asmussen, Chapter 7. Microwave Plasma Assisted D iamond Film Deposition, in: J. Asmussen, D.K. Reinhard (Eds.), Diamond Films Handbook, MarcellDekker, Inc, United States of America, 2002: p. 211. [24] C. J. H. Wort, R. S. Balmer, Diamond as an electronic material, Materials Today. 11 (2008) 22 28. doi:10.10 16/S1369 - 7021(07)70349 - 8. [25] K. E. Holbert, Radiation effects and damage, http:// h olbert.faculty.asu.edu/eee560/ RadiationEfffectsDamage.pdf. [26] D.M. Trucchi, P. Allegrini, P. Calvani, A. Galbiati, K. Oliver, G. Conte, Very Fast and Primingless Single - Crystal - Di amond X - Ray Dosimeters, IEEE Electron Device Lett. 33 (2012) 6 15 617. doi:10.1109/LED.2012.2185476. [27] I. Ciancaglioni, R. Consorti, M.C. Rossi, G. Conte, Polycrystalline diamond detectors compared with silicon X - ray dosimeters for clinical use, in: IEEE Symposium Conference Record Nuclear Science 2004., IEEE, 2004 : pp. 4445 4447. doi:10.1109/NSSMIC.2004.1466871. [28] N. Tartoni, M. Angelone, M. Pillon, S. Almaviva, M. Marinelli, E. Milani, et al., X - Ray Detection by Using CVD Single Crystal Diamond Detec tor, IEEE Trans Nucl Sci. 56 (2009) 849 852. doi:10.1109/TNS.2 008.2010705. [29] J.W. Keister, J. Smedley, Single crystal diamond photodiode for soft X - ray radiometry, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometer s, Detectors and Associated Equipment. 606 (2009) 774 779. doi:10.1016/j.nima.2009.04.044. 162 [30] A. Balducci, M. Bruzzi, A. De Sio, M.G. Donato, G. Faggio, M. Marinelli, et al., Diamond - based photoconductors for deep UV detection, Nuclear Instruments and Me thods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment. 567 (2006) 188 191. doi:10.1016/j.nima.2006.05.159. [31] W. de Boer, J. Bol, A. Furgeri, S. Müller, C. Sander, E. Berdermann, et al., Radiation hardness o f diamond and silicon sensors compared, Phys. Stat . Sol. (a). 204 (2007) 3004 3010. doi:10.1002/pssa.200776327. [32] J.W. Tsung, M. Havranek, F. Hügging, H. Kagan, H. Krüger, N. Wermes, Signal and noise of diamond pixel detectors at high radiation fluences , J. Inst. 7 (2012) P09009 P09009. doi:10.1088/174 8 - 0221/7/09/P09009. [33] H. Kagan, Recent advances in diamond detector development, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equip ment. 541 (2005) 221 227. doi:10.1016/j.nima.2005. 01.060. [34] H. Kagan, Diamond radiation detectors may be forever!, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment. 546 (2005) 222 227. doi:10.1016/j.nima.2005.03.128. [35] W. hard diamond sensors for future tracking applications, Nuclear Instruments and Methods in Physics Research Section A: Acc elerators, Spectrometers, Detectors and Associated Equipment. 565 (2006) 278 283. doi:10.1016/j.nima.2006.05.127. [ 36 ] M. Pomorski, E. Berdermann, A. Caragheorgheopol, M. Ciobanu, M. Ki , A. Martemiyanov, et al., Development of single - crystal CVD - diamond detectors for spectroscopy and timing, Phys. Stat. Sol. (a). 203 (2006) 3152 3160. doi:10.1002/pssa.200671127. [37] D. Asner, M. Barbero, V. Bellini, V. Belyaev, J.M. Brom, M. Bruzzi, et al., Diamond pixel modules, Nuclear Instruments and Methods in Physic s Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment. 636 (2011) S125 S129. doi:10.1016/j.nima.2010.04.096. [38] A. Stolz, M. Behravan, M. Regmi, B. Golding, Heteroepitaxial diamond detectors for heavy ion beam tracking, Di amond and Related Materials. 15 (2006) 807 810. do i:10.1016/j.diamond.2005.12.019. [39] J. Pietraszko, T. Galatyuk, V. Grilj, W. Koenig, S. Spataro, M. Träger, Radiation damage in single crystal CVD diamond material investigated with a high current relativ istic 197Au beam, Nuclear Instruments and Methods in Physics Research Section A: 163 Accelerators, Spectrometers, Detectors and Associated Equipment. 763 (2014) 1 5. doi:10.1016/j.nima.2014.06.006. [40] G. García, M. Díaz - Híjar, V. Tormo - Márquez, I. Preda, O. Peña - Rodríguez, J. Olivares, Structural damage on single - crystal diamond by swift heavy ion irradiation, Diamond and Related Materials. 58 (2015) 226 229. doi:10.1016/j.diamond.2015.08.014. [41] A. Lohstroh, P.J. Sellin, S.G. Wang, A.W. Davies, J.M. Parkin , Mapping of polarization and detrapping effects i n synthetic single crystal chemical vapor deposited diamond by ion beam induced charge imaging, J. Appl. Phys. 101 (2007) 063711. doi:10.1063/1.2653669. [42] H. Pernegger, High mobility diamonds and particl e detectors, Phys. Stat. Sol. (a). 203 (2006) 3299 3314. doi:10.1002/pssa.200671404. [ 43 ] H. Pernegger, S. Roe, P. Weilhammer, V. Eremin, H. Frais - K lbl, E. Griesmayer, et al., Charge - carrier properties in synthetic single - crystal diamond measured with th e transient - current technique, J. Appl. Phys. 97 ( 2005) 073704. doi:10.1063/1.1863417. [44] J. Isberg, M. Gabrysch, A. Tajani, D.J. Twitchen, Transient current electric field profiling of single crystal CVD diamond, Semicond. Sci. Technol. 21 (2006) 1193 1 195. doi:10.1088/0268 - 1242/21/8/035. [45] Y. Gu, J . Lu, T. Grotjohn, T. Schuelke, J. Asmussen, Microwave plasma reactor design for high pressure and high power density diamond synthesis, Diamond and Related Materials. 24 (2012) 210 214. doi:10.1016/j.diamo nd.2012.01.026. [46] P. Moritz, E. Berdermann, K. Blasche, H. Stelzer, B. Voss, Broadband electronics for CVD - diamond detectors, Diamond and Related Materials. 10 (2001) 1765 1769. doi:10.1016/S0925 - 9635(01)00434 - 4. [47] M. Guthoff, K. Afanaciev, A. Dabrow ski, W. de Boer, W. Lange, W. Lohmann, et al., Rad iation damage in the diamond based beam condition monitors of the CMS experiment at the Large Hadron Collider (LHC) at CERN, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spec trometers, Detectors and Associated Equipment. 730 (2013) 168 173. doi:10.1016/j.nima.2013.05.041. [48] Leakage current measurements of a pixelated polycrystalline CVD dia mond detector, J. Inst. 9 (2014) E03001 E03001. do i:10.1088/1748 - 0221/9/03/E03001. [49] S. Müller, W. deBoer, M. Schneider, A. Sabellek, M. Schmanau, C. Rühle, et al., Study of leakage currents in PC CVD diamonds as function of the magnetic field, Phys. St at. Sol. (a). 206 (2009) 2091 2097. doi:10.1002/ps sa.200982221. 164 [50] www.wardsci.com/www.wardsci.com/images/BeerLamb_Colorimeter.pdf, [51] M.E. Thomas, W.J. Tropf, O ptical P roperties of diamond , Johns Hopkins APL Tech Dig. (1993). [52] Introduction to Fourier Transform Infrared Spectrometry, https://www.labx.com/ product/nicolet - ftir [53] R. Linares, P. Doering, Properties of large single crystal diamond, Diamond and Related Materials. 8 (1999) 909 915. doi:10.1016/S0925 - 9635(98)00382 - 3. [54] Bruker - AXS High Resolution X - ray Diffracto meter, W. M. Keck Microfabrication Facility, (n.d. ). https://kmf.pa.msu.edu/KMF/Equipment/equip13.asp (accessed September 27, 2018). [55] Evaluation of the crystallinity (tilt and twist distributions) of a GaN thin film by the X - ray rocking curve method | Rigaku, (n.d.). http://www.rigaku.com/en/products/ xrd/ smartlab/app025 (accessed October 3, 2018). [56] Raman Spectroscopy | Science Facts Raman Spectroscopy Science Facts, (2017). http://www.sciencefacts.net/raman - spectroscopy.html) (accessed October 3, 2018). [57] M.G. Donato, G. Faggio, M. Marinelli, G. Messina, E. Milani, A. Paoletti, et al., High quality CVD diamond: a Raman scattering and photoluminescence study, Eur. Phys. J. B. 20 (2001) 133 139. doi:10.1007/s100510170291. [58] S . Prawer, R.J. Ne manich, Raman spectroscopy of diamond and doped diamond., Philos. Trans. A, Math. Phys. Eng. Sci. 362 (2004) 2537 2565. doi:10.1098/rsta.2004.1451. [59] Diamond and Th eir Annealing : Experiment and Simulation, Physica Status Solidi (a). (1999). [60] G. Yang, K. Livingston, R. Jones, F. Klein, High resolution X - ray diffraction study of single crystal diamond radiators, Phys. Stat. Sol. (a). 209 (2012) 1786 1791. doi:10.10 02/pssa.20120 0017. [61] J.C. Angus, Chapter 2. A short History of Diamond Synthesis, in: J. Asmussen, D.K. Reinhard (Eds.), Diamond Films Handbook, Marcel Dekker, Inc., United States of America, 2002: p. 17. [62] P. W. Bridgman, Synthetic Diamonds, Scientific American 193 ( 1955) 42 - 47. [63] M. Simpson, M aterials with Diamond Touch, New Sci. (1988). 165 [64] H.P. Bovenkerk, F.P. Bundy, H.T. Hall, H.M. Strong, R.H. Wentorf, Preparation of Diamond, Nature. 184 (1959) 1094 1098. doi:10.1038/1841094a0. [65] W.G. Eversole, Synthesis of Diamond, 3030187 , 1962. [66] B.V. Spitsyn, Origin, State of the Art and Some Prosp ects of the Diamond CVD , Brazilian Journal of Physics. (2000). http://dx.doi .org/10.1590/ S0103 - 97332000000300002 (accessed July 2, 2018). [67] B.V. Deryagin, B.V. Spitsyn, A Technique of Diamond Growth on a Diamond Face, 339134, 1980. [68] J.C. Angus, H.A. Will, W.S. Stanko, Growth of diamond seed crystals by vapor deposition, J. Appl. Phys. 3 9 (1968) 2915 2922. doi:10.1063/1.1656693. [69] S. Matsumoto, Y. Sato, M. Kamo, N. Setaka, Vapor Deposition of Diamond Particles from Methane, Jpn. J. Appl. Phys. 21 (1982) L183 L185. doi:10.1143/JJAP.21.L183. [70] G. - R. Lai, E.N. Farabaugh, A . Feldman, L. H. Robins, Deposition of diamond films in a closed hot filament CVD system., J Res Natl Inst Stand Technol. 100 (1995) 43 49. doi:10.6028/jres.100.004. [71] H. Liu, D.S. Dandy, Chapter 3. Diamond CVD Techniques, in: Diamond Chemical Vapor Depo sition , 1995 : p. 14. [72] http://www.chm.bris.ac.uk/pt/diamond/jamespthesis/chapter1.htm . [73] M. Konuma, Film Deposition By Plasma Techniques, Springer - Verlag, New York, 1992. [74] Y.F. Li, J.J. Su, Y.Q. Liu, M.H. Ding, X.L. Li, G. Wang, et al., Design of a new TM02 1 mode cavity type MPCVD reactor for diamond film deposition, Diamond and Related Materials. 44 (2014) 88 94. doi:10.1016/j.diamond.2014.02.010. [75] CSIR Central Glass & Ceramics Research Institute, 196 Raja S C Mullick Road, Kolkata, 70003 2, West Benga l, India, A. Mallik, Microwave plasma CVD grown single crystal diamonds A review, J. Coat. Sci. Technol. 3 (2016) 75 99. doi:10.6000/2369 - 3355.2016.03.02.4. [76] T.A. Grotjohn, J. Asmussen, Microwave plasma - assisted diamond film deposition, Diamond Films Handbook. (2002). [77] F. Silva, K. Hassouni, X. Bonnin, A. Gicquel, Microwave engineering of plasma - assisted CVD reactors for diamond deposition., J Phys Condens Matter. 21 (2009) 364202. doi:10.1088/0953 - 8984/21/36/364202. 166 [78] M. Füner, C. Wild, P. Koi dl, Novel microwave plasma reactor for diamond synthesis, Appl. Phys. Lett. 72 (1998) 1149 1151. doi:10.1063/1.120997. [79] J.E. Butler, Y.A. Mankelevich, A. Cheesman, J. Ma, M.N.R. Ashfold, Understanding the chemical vapor deposition of diamo nd: recent pr ogress., J Phys Condens Matter. 21 (2009) 364201. doi:10.1088/0953 - 8984/21/36/364201. [80] M. Schwander, K. Partes, A review of diamond synthesis by CVD processes, Diamond and Related Materials. 20 (2011) 1287 1301. doi:10.1016/j.diamond.2011. 08.005. [81] A. Gicquel, K. Hassouni, F. Silva, J. Achard, CVD diamond films: from growth to applications, Current Applied Physics. 1 (2001) 479 496. doi:10.1016/S1567 - 1739(01)00061 - X. [82] R.A. Khmelnitskiy, Prospects for the synthesis of large single - cry stal diamonds , Phys. - Usp. 58 (2015) 134 149. doi:10.3367/UFNe.0185.201502b.0143. [83] T. Teraji, Chemical vapor deposition of homoepitaxial diamond films, Phys. Stat. Sol. (a). 203 (2006) 3324 3357. doi:10.1002/pssa.200671408. [84] J.E. Butler, R.L. Woodin , Thin film d iamond growth mechanisms, Philos. Trans. R. Soc. Lond. A. (1993). [85] P.K. Bachmann, H. - J. Hagemann, H. Lade, D. Leers, F. Picht, D.U. Wiechert, et al., Diamond Chemical Vapor Deposition: Gas Compositions and film Properties, MRS Proc. 339 (1 994) 267. doi:10.1557/PROC - 339 - 267. [86] K.E. Spear, M. Frenklach, High - temperature chemistry of CVD (chemical vapor deposition) diamond growth, Pure Appl. Chem. 66 (1994) 1773 1782. doi:10.1351/pac199466091773. [87] M. Frenklach, H. Wang, Detailed surface and gas - p hase chemical kinetics of diamond deposition, Phys. Rev. B. 43 (1991) 1520 1545. doi:10.1103/PhysRevB.43.1520. [88] D.T. Stibbe, J. Tennyson, Near - threshold electron impact dissociation of H 2 within the adiabatic nuclei approximation, New J Phys. 1 (1999) 2 2. doi:10.1088/1367 - 2630/1/1/002. [89] D.T. Stibbe, J. Tennyson, Rates for the electron impact dissociation of molecular hydrogen, Astrophys. J. 513 (1999) L147 L150. doi:10.1086/311918. [90] r depositi on. I. Diamond surface chemistry, J. Appl. Phys. 74 (1993) 6888 6894. doi:10.1063/1.355063. 167 [91] D.G. Goodwin, Theory of Diamond Chemical Vapor Deposition, in: M.A. Prelas, G. Popovici, L.K. Bigelow (Eds.), Handbook of Industrial Diamonds and Dia mond Films , Marcel Deker, Inc., 1998: p. 527. [92] J.C. Angus, Diamond synthesis by chemical vapor deposition: The early years, Diamond and Related Materials. 49 (2014) 77 86. doi:10.1016/j.diamond.2014.08.004. [93] R.M. Hazen, The Diamond Makers , Cambrid ge Univers ity Press, 2003. [94] M. Kamo, Y. Sato, S. Matsumoto, N. Setaka, Diamond synthesis from gas phase in microwave plasma, J. Cryst. Growth. 62 (1983) 642 644. doi:10.1016/0022 - 0248(83)90411 - 6. [95] Seocal | Seki - ASTeX: Diamond CVD, Carbon Nanotube, Laser Spec trometer, (n.d.). http://www.seocal.com/seki_products_detail.html (accessed July 4, 2018). [96] S.S. Zuo, MICROWAVE PLASMA - ASSISTED CVD POLYCRYSTALLINE DIAMONDFILMS DEPOSITION AT HIGHER PRESSURE CONDITIONS, Michigan State University, 2009. [97] J . Lu, Y. G u, T.A. Grotjohn, T. Schuelke, J. Asmussen, Experimentally defining the safe and efficient, high pressure microwave plasma assisted CVD operating regime for single crystal diamond synthesis, Diamond and Related Materials. 37 (2013) 17 28. doi:10. 1016/j.dia mond.2013.04.007. [98] K. - P. Kuo, J. Asmussen, A n experimental stu dy of high p ressure s ynth esis of diamond films using microwave cavity p lasma reactor, Diamond and Related Materials 6 (1997) 1097 - 11 05. [99] S.S. Zuo, M.K. Yaran, T.A. Grotjohn, D.K. Reinhard, J. Asmussen, Investigation of diamond deposition uniformity a nd quality for freestanding film and substrate applications, Diamond and Related Materials. 17 (2008) 300 305. doi:10.1016/j.diamond.2007.12.069. [100] K.W. Hemawan, T.A. Grotjohn, D.K. Reinhard, J. Asmussen, Improved microwave plasma cavity reactor for di amond synt hesis at high - pressure and high power density, Diamond and Related Materials. 19 (2010) 1446 1452. doi:10.1016/j.diamond.2010.07.005. [101] S. Nad, J. Asmussen, Analyses of single crystal diamond substrates grown in a pocket substrate holder via MPACVD, Di amond and Related Materials. 66 (2016) 36 46. doi:10.1016/j.diamond.2016.03.007. 168 [102] S. Nad, Y. Gu, J. Asmussen, Growth strategies for large and high quality single crystal diamond substrates, Diamond and Related Materials. 60 (2015) 26 34. doi :10.1016/j .diamond.2015.09.018. [103] S. Nad, Y. Gu, J. Asmussen, Determining the microwave coupling and operational efficiencies of a microwave plasma assisted chemical vapor deposition reactor under high pressure diamond synthesis operating conditions., Rev. Sci. Instrum. 86 (2015) 074701. doi:10.1063/1.4923092. [104] K.W. Hemawan, INVESTIGATION OF MICROWAVE CAVITY APPLICATORS FOR PLASMAASSISTED CVD DIAMOND SYNTHESIS AND PLASMA ASSISTEDCOMBUSTION, Doctoral dissertation, Michigan State University, 2010. [1 05] J. Asm ussen, Y. Gu, T.A. Grotjohn, MICROWAVE PLASMA REACTORS, PCT/US12/37555, 2013. [106] Y. Gu, THE NEW GENERATION MICROWAVE PLASMA ASSISTED CVD REACTOR FOR DIAMOND SYNTHESIS , Doctoral dissertation, Michigan State University, 2011. [107] J. Lu, SINGL E CRYSTAL MICROWAVE PLASMA ASSISTED CHEMICAL VAPORDIAMOND SYNTHESIS AT HIGH PRESSURES AND HIGH POWERDENSITIES, Doctoral dissertation, Michigan State Universty, 2013. [108] S. Nad, GROWTH AND CHARACTERIZATION OF LARGE, HIGH QUALITY SINGLE CRYSTAL DIAMOND SU BSTRATES V IA MICROWAVE PLASMA ASSISTED CHEMICAL VAPOR DEPOSITION, Doctoral dissertation, Michigan State University, 2016. [109] N. T atsumi , K. I keda , H. U mez awa , S. S hikata , Development of Diamond Schottky Barrier Diode, SEI Technical Review 68 (2009) 5 4 - 61. [110] Wörner, F. Fuchs, C. Wild, P. Koidl, Nitrogen induced increase of growth rate in chemical vapor deposition of diamond, Appl. Phys. Lett. 68 (1996) 759 760. doi:10.1063/1.116733. [111] C. - S. Yan, Y.K. Vohra, H. - K. Mao, R.J. Hemley, Very high growth rate che mical vapo r deposition of single - crystal diamond., Proc. Natl. Acad. Sci. USA. 99 (2002) 12523 12525. [112] H. Watanabe, S. - G. Ri, S. Yamanaka, D. Takeuchi, H. Okushi, High - Quality Homoepitaxial Diamond Film Growth, (n.d.). 169 [113] H. Watanabe, D. Takeuchi, S. Yamanaka, H. Okushi, K. Kajimura, T. Sekiguchi, Homoepitaxial diamond film with an atomically flat surface over a large area, Diamond and Related Materials. 8 (1999) 1272 1276. doi:10.1016/S0925 - 9635(99)00126 - 0. [114] Takeuchi, Watanabe, Yamanaka, Okush i, Kajimura, Defects in Device Grade Homoepitaxial Diamond Thin Films Grown with Ultra - Low CH4/H2 Conditions by Microwave - Plasma Chemical Vapor Deposition, (n.d.). [115] T. Tsuno, T. Imai, N. Fujimori, Twinning structure and growth hillock on diamond (001) ep itaxial film, Jpn. J. Appl. Phys. 33 (1994) 4039 4043. doi:10.1143/JJAP.33.4039. [116] T. Bauer, M. Schreck, H. Sternschulte, B. Stritzker, High growth rate homoepitaxial diamond deposition on off - axis substrates, Diamond and Related Materials. 14 (2005 ) 2 66 271. doi:10.1016/j.diamond.2004.10.043. [117] M.P. Gaukroger, P.M. Martineau, M.J. Crowder, I. Friel, S.D. Williams, D.J. Twitchen, X - ray topography studies of dislocations in single crystal CVD diamond, Diamond and Related Materials. 17 (2008) 262 2 69. doi:10.1016/j.diamond.2007.12.036. [118] N. Tsubouchi, Y. Mokuno, Microstructures of threading dislocation bundles included in CVD homoepitaxial diamond plates, Diamond and Related Materials. 78 (2017) 44 48. doi:10.1016/j.diamond.2017.07.007. [119] N. Ts ubouchi, S. Shikata, Evaluation method for grown - in dislocations in CVD single crystal diamond using plasma surface treatment, Jpn. J. Appl. Phys. 53 (2014) 068010. doi:10.7567/JJAP.53.068010. [120] O.A. Ivanov, A.B. Muchnikov, V.V. Chernov, S.A. Bogdan ov, A.L. Vikharev, J.E. Butler, Experimental study of hydrogen plasma etching of (100) single crystal diamond in a MPACVD reactor, Mater Lett. 151 (2015) 115 118. doi:10.1016/j.matlet.2015.03.073. [121] A. Tallaire, M. Kasu, K. Ueda, T. Makimoto, Origin of gr owth defects in CVD diamond epitaxial films, Diamond and Related Materials. 17 (2008) 60 65. doi:10.1016/j.diamond.2007.10.003. [122] A. Tallaire, T. Ouisse, A. Lantreibecq, R. Cours, M. Legros, H. Bensalah, et al., Identification of dislocations in syn the tic chemically vapor deposited diamond single crystals, Cryst. Growth Des. 16 (2016) 2741 2746. doi:10.1021/acs.cgd.6b00053. [123] Takeuchi, Watanabe, Yamanaka, Okushi, Kajimura, Homoepitaxial diamond films grown by step - flow mode in various misorientat ion angles of diamond substrates, (n.d.). [124] T. Tsuno, H. Shiomi, Y. Kumazawa, S. Shikata, S. Akai, Growth rate and surface morphology of diamond homoepitaxial films on misoriented (001) substrates, Jpn. J. Appl. Phys. 35 (1996) 4724 4727. doi:10.1143/J JAP .35.4724. 170 [125] N. Lee, A. Badzian, Effect of methane concentrations on surface morphologies and surface structures of (001) homoepitaxial diamond thin films, Appl. Phys. Lett. 67 (1995) 2011 2013. doi:10.1063/1.114770. [126] W. J. P. van Enckevort, Janssen, Sch erm er, Giling, Step - related growth phenomena on exact and misoriented {001} surfaces of CVD - grown single - crystal diamonds, Diamond and Related Mater ials 4 (1995) 250 - 255. [127] F.K. de Theije, J.J. Schermer, W.J.P. van Enckevort, Effects of nitrogen impurities on the CVD growth of diamond: step bun ching in theory and experiment, Diamond and Related Materials. 9 (2000) 1439 1449. doi:10.1016/S0925 - 9635(00)00261 - 2. [128] F. Silva, X. Bonnin, J. Achard, O. Brinza, A. Michau, A. Gicquel, Geometric modeling of homoepitaxial CVD diamond growth: I. The {10 0}{111}{110}{113} system, J. Cryst. Growth. 310 (2008) 187 203. doi:10.1016/j.jcrysgro.2007.09.044. [129] A. Tallaire, J. Barjon, O. Brinza, J. Achard, F. Silva, V. Mille, et al., Dislocations and impurities introduced from etch - pits at the epitaxial gr owt h resumption of diamond, Diamond and Related Materials. 20 (2011) 875 881. doi:10.1016/j.diamond.2011.04.015. [130] Optical Microscope Nikon Eclipse ME600 | PoliFAB, (n.d.). http://www.polifab.polimi.it/equipments/nikon - eclipse - me600 - optical - microscop e/ (accessed October 3, 2018). [131] A. Chayahara, Y. Mokuno, Y. Horino, Y. Takasu, H. Kato, H. Yoshikawa, et al., The effect of nitrogen addition during high - rate homoepitaxial growth of diamond by microwave plasma CVD, Diamond and Related Materials. 13 ( 200 4) 1954 1958. doi:10.1016/j.diamond.2004.07.007. [132] N. Lee, A. Badzian, A study on surface morphologies of (001) homoepitaxial diamond films, Diamond and Related Materials. 6 (1997) 130 145. doi:10.1016/S0925 - 9635(96)00698 - X. [133] M.P. Everson, M.A. Ta mor, Investigation of growth rates and morphology for diamond growth by chemical vapor deposition, J Mater Res. 7 (1992) 1438 1444. doi:10.1557/JMR.1992.1438. [134] R.M. Silver, J.A. Soons, J.F. Song, A.X. Zheng, T. Brian Renegar, T.V. Vorburger, NIST S URF ACE ROUGHNESS AND STEP HEIGHT CALIBRATIONS, Measurement Conditions and Sources of Uncertainty , (n.d.). [135] Veeco Dektak 6M Profilometer, https://www.cnfusers.cornell.edu/sites/default/files/ Equipment - Resources/Veeco%20Dektak%206M%20protoc ol% 20102011.pdf (accessed December 12, 2018). 171 [136] A.M. Zaitsev, 4.2. Raman Scattering , in: Optical Properties of Diamond, A Data Handbook, 2001: pp. 69 131. [137] M. Ogura, H. Kato, T. Makino, H. Okushi, S. Yamasaki, Misorientation - angle dependence of boron incorporation into (001) - oriented chemical - vapor - deposited (CVD) diamond, J. Cryst. Growth. 317 (2011) 60 63. doi:10.1016/j.jcrysgro.2011.01.010. [138] M.A. Pinault - Thaury, T. Tillocher, D. Kobor, N. Habka, F. Jomar d, J. Chevallier, et al., Phosphoru s donor incorporation in (100) homoepitaxial diamond: Role of the lateral growth, J. Cryst. Growth. 335 (2011) 31 36. doi:10.1016/j.jcrysgro.2011.06.007. [139] A. Tallaire, J. Achard, O. Brinza, V. Mille, M. Naamoun, F. S ilva, et al., Growth strategy for c ontrolling dislocation densities and crystal morphologies of single crystal diamond by using pyramidal - shape substrates, Diamond and Related Materials. 33 (2013) 71 77. doi: 10.1016/j.diamond .2013.01.006.