| 1. |
|
|
| 2. |
- Badel, Xavier
(författare)
-
Development of macropore arrays in silicon and related technologies for X-ray imaging applications
- 2003
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Digital devices have started to replace photographic film inX-ray imaging applications. As compared to photographic films,these devices are more convenient to obtain images and tohandle, treat and store these images. The goal of the presentstudy is to develop macropore arrays and related silicontechnologies in order to fabricate X-ray imaging detectors formedical applications, and in particular for dentistry. Althougha few detectors are already available on the market, theirperformances, such as the X-ray sensitivity, can still beimproved. In addition, the image quality, defined by thespatial resolution and the signalto- noise ratio (SNR), shouldbe sufficiently high to enable diagnosis and, as regard to thepatient health, the X-ray dose should be reduced to aminimum.Three detector concepts were investigated. All of themrequire the formation of a macropore array in silicon as afirst step in the detector fabrication. Even though deepreactive ion etching was used to form these macropore arrays,silicon electrochemical etching in aqueous hydrofluoric acid(HF) solution has been more intensively studied. The porespacing was fixed to about 50 µm in order to achieve aspatial resolution of 10 lp/mm, as required in dentalapplication. Pore depths up to 420 µm with diameterranging from 10 to 40 µm, depending on the detectorconcept, have been achieved. Electrochemical etching of siliconis, indeed, a very promising technique to fabricate high aspectratio structures and damage-free macropore arrays. Thistechnique is based on a silicon dissolution reaction involvingthe species of the HF solution, silicon atoms and holes, thepositive charge carriers. As holes are the minority carriers inn-type silicon, they are usually photogenerated. However, wealso developed an alternative technique based on hole injectionfrom a forwardbiased p-n junction, and the possibility to formmacropore arrays and diverse threedimensional structures wasdemonstrated.The first detector concept investigated consists of asilicon charge-coupled device (CCD) in proximity with ascintillating guide screen. This screen is made of a siliconmacropore array filled with CsI(Tl), emitting photons at awavelength of 550 nm (green light) under X-ray exposure. Thevisible light is then reflected on the walls of the pores inorder to be detected by the CCD pixels. Both oxide and metalcan be used as a reflective layer. Such detectors were fullyfabricated and characterized, showing good spatial resolutionand comparable results with currently available detectorsconcerning the SNR and the X-ray dose. The second detectorstudied in this thesis uses photodiodes, instead of a CCD, inorder to detect the photons emitted from the scintillator. Thisconcept would lead to high charge collection efficiency sincethe diodes are formed in the silicon pore walls, making thedistance between the generation and detection points of thevisible photons short. However, this design implies two majordifficulties in the detector fabrication: formation of p-njunction in the pore walls and formation of contacts to thediodes. Thus, boron diffusion from a solid source andlow-pressure chemical vapor deposition of boron-dopedpoly-silicon were experimented. Both techniques were shown tobe successful. The last detector concept is based on thegeneration of electron/hole pairs in the semiconductor bulkunder X-ray exposure. The generated charges would then becollected by electrodes going through the bulk, requiringformation of deep and narrow pores. Siliconphoto-electrochemical etching was used and 425-µm deeppores with a diameter of 14 µm were formed, resulting inan aspect ratio of ~ 30 and an active area of 90 %.
|
|
| 3. |
- Badel, Xavier, 1977-
(författare)
-
Electrochemically etched pore arrays in silicon for X-ray imaging detectors
- 2005
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Digital devices have now been introduced in many X-ray imaging applications, replacing slowly traditional photographic films. These devices are preferred as they offer real time imaging, easy handling and fast treatment of the images. However, the performance of the detectors still have to be improved in order to increase the image quality, and possibly reduce the X-ray dose, a vital parameter for medical use. In this thesis, three different new detector concepts have been investigated. All designs use pore arrays, which are ideal starting structures to form pixellated detectors. Electrochemical etching of n-type silicon in aqueous hydrofluoric acid solution (HF) has been studied to form these pore arrays. A broad range of pores have been fabricated with diameters varying from 200 nm to 40 µm and with depths reaching almost the wafer thickness, thus leading to very high aspect ratios. The technique was also found to be suitable for the formation of other types of structures such as pillars and tubes on the sub micrometer scale. The etching is based on the dissolution of silicon in HF under anodic bias and a supply of positive electrical carriers (holes). As holes are the minority carriers in n-type silicon, they are usually photo-generated. In this work an alternative technique, based on hole injection from a forward-biased pn junction, has been successfully pioneered. The first X-ray imaging detector concept presented in the thesis consists of a silicon charge coupled device (CCD) in proximity with a scintillating screen. The screen is made from a pore array having reflective pore walls and filled with CsI(Tl), emitting photons at a wavelength of 550 nm under X-ray exposure. The secondary emitted photons are light-guided by the pore walls and then detected by the CCD pixels. Detectors were fully fabricated and characterized. This concept provides good spatial resolution with negligible cross talk between adjacent pixels. The dependences of the detector efficiency on pore depth and on the coating of the pore walls are presented. Although most of the produced detectors had a detective quantum efficiency of about 25%, some detectors indicate that efficient scintillating screens can be achieved approaching the theoretical limit as set by poissonian statistics of the X-ray photons. The two other detector designs require the formation of vertical pn junctions, i.e. in the pore walls. In one concept the secondary emitted photons are detected by photodiodes located in the pore walls. This would lead to high charge collection efficiency as the photons do not have to be guided to one end of the pore. However, high noise due to the direct detection of X-rays in the diodes is expected. The other concept is based on generation of electron-hole pairs in a semiconductor and the ‘3D’ detector, where an array of vertical electrodes is used to separate the charges via an electric field. To uniformly dope the inside of deep pores, both boron diffusion and low-pressure chemical vapor diffusion of boron-doped poly-silicon were shown to be successful techniques. This was confirmed by SIMS profiles taken through the pore wall thickness. Finally, the possibility to form individual junction in each pore was shown. The diodes were electrically characterized, demonstrating good rectifying behavior and sensitivity to light.
|
|
| 4. |
|
|
| 5. |
- Christensen, Jens S.
(författare)
-
Dopant diffusion in Si and SiGe
- 2004
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Dopant diffusion in semiconductors is an interestingphenomenon from both technological and scientific points ofview. Firstly, dopant diffusion is taking place during most ofthe steps in electronic device fabrication and, secondly,diffusion is related to fundamental properties of thesemiconductor, often controlled by intrinsic point defects:self-interstitials and vacancies. This thesis investigates thediffusion of P, B and Sb in Si as well as in strained andrelaxed SiGe. Most of the measurements have been performedusing secondary ion mass spectrometry on high purityepitaxially grown samples, having in-situ incorporated dopantprofiles, fabricated by reduced pressure chemical vapordeposition or molecular beam epitaxy. The samples have beenheat treated both under close-to-equilibrium conditions (i. e.,long time annealings in an inert ambient) and conditions whichresulted in non-equilibrium diffusion (i. e., vacuum annealing,oxidation, short annealing duration, and protonirradiation).Equilibrium P and B diffusion coefficients in Si asdetermined in this thesis differ from a substantial part ofpreviously reported values. This deviation may be attributed toslow transients before equilibrium concentrations of pointdefects are established, which have normally not been takeninto account previously. Also an influence of extrinsic dopingconditions may account for the scattering of the diffusivityvalues reported in literature. B and Sb diffusion in Si underproton irradiation at elevated temperatures was found to obeythe so-called intermittent model. Parameters describing themicroscopic diffusion process were derived in terms of theintermittent diffusion mechanism, and it was found also thatthe presence of Sb strongly affected the B diffusion and viceversa.In relaxed Si1-xGex-alloys, which has the same lattice structure as Sibut a larger lattice constant, P diffusion is found to increasewith increasing Ge content (x≤ 0.2). In Si/SiGe/Si heterostructures, wherethe SiGe layer is biaxially strained in order to comply withthe smaller lattice parameter of Si, P diffusion in thestrained layer is retarded as compared with relaxed materialhaving the same Ge content. In addition, P is found tosegregate into the Si layer via the Si/SiGe interface and thesegregation coefficient increases with increasing Ge content inthe SiGe layer.
|
|
| 6. |
- Doyle, James P.
(författare)
-
Copper germanide schottky contacts to silicon and electrically active defects in n-type 6H-SiC and 4H-SiC epitaxial layers
- 1997
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Metallization for contacts to silicon devices presents amajor challenge as the linewidths are further reduced into thesub-micron regime. Copper germanide due to its relatively lowroom temperature resistivity ( ~ 10 µΩ - cm) has beenexamined as a potential contact metallixation. In addition, dueto the growing interest in silicon carbide as a potentialwide-bandgap semiconductor candidate for high voltage, power,and frequency devices, a characterization of electricallyactive deep levels was conducted on as-grown and electron andion irradiated layers as such defects can influence theelectrical behaviour of active devices. N-type 6H- and 4H-SiChave been characterized by deep level transient spectroscopy(DLTS) in conjunction with MeV electron beam irradiation andkeV hydrogen and deuterium implantation. Intrinsic levels havebeen identified in both polytypes through an examination ofas-grown and irradiated samples.It is demonstrated that silicon outdiffusion into the coppergermanide contact causes an increase in the electricalresistivity. Additionally, deviation from ideal behaviour isobserved in current-voltage (IV) measurements an indicationthat this outdiffusion of silicon produces recombinationcenters primarily in the upper part of the bandgap near to thecontact-semiconductor interface. Additionally, copper germanideis found to be morphologically unstable during theamorphous-to-crystalline transformation. The high diffusivityof copper in amorphous germanium at temperatures as low as 200°C coupled with the independence of the barrier height onthe germanium fraction used during contact formation indicatesthat the interface between copper germanide and silicon issimilar to that of copper silicide and silicon.A number of electrically active deep levels have beencharacterized in both the 4H- and 6H-SiC polytypes. In 6H-SIC,many of the defects are found to be intrinsic as they areobserved to grow as a function of increasing irradiation dose.In contrast, as-grown 4H-SiC epitaxial layers, exhibit only asingle acceptor-like level at 0.70 eV below the conduction bandedge (Ec). After irradiation the level is not observed toincrease in concentration, but two new levels are found. Theselevels are found to be unstable at room temperature in contrastto the 6H-SiC defects. In 6H-SiC, the defect concentration isfound to limit the average carrier lifetime. Through the use ofthermal annealing, SIMS, and electron irradiation, a model oftheir Origin is proposed.Keywords:Deep level transient spectroscopy (DLTS),secondary ion mass spectrometry (SIMS), amorphous germanium,electron irradiation, ion irradiation, wide-bandgap
|
|
| 7. |
- Fröjdh, Christer
(författare)
-
Schottky barriers and Schottky barrier based devices on Si and SiC
- 1998
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This work is devoted to a study of the formation andcharacterisation of Schottky Barriers on differentsemiconductors with an extension to the development of devicesbased on Schottky barriers. The major part of the work has beendone on 6H-SiC, but Si and SiGe alloys have also been used.Different metals have been deposited on p-type and n-type6H-SiC and the properties of the created barriers have beeninvestigated by CV, IV and photoelectric techniques. Data fromother research groups have been collected in order toinvestigate the correlation between the barrier height and themetal work function. The conclusion is that for a number ofmetals there is a strong correlation between the barrier heightand the metal work function, while other metals showsignificant deviation from the Schottky-Mott theory. Largescatter in the data exists between different investigationsindicating that the method of sample preparation is veryimportant for the results. This is not surprising since theSchottky Barrier is mainly a surface device, which makes itvery sensitive to variations in the surface conditions.Extensive work has been done in order to explain thebehaviour of Schottky diodes fabricated on SiC and to qualifythe measurement techniques in presence of different anomaliesin the devices. A highly resistive interfacial layer presentbetween the bulk wafer and the epitaxial top layer on certainp-type 6H-SiC wafers has been found.A process for fabrication of buried Schottky and ohmiccontacts in Silicon by wafer bonding has been developed usingCo as a buried metal layer. During the heat treatment the metalreacts with the Silicon and forms CoSi2. The processed contacts exhibit the sameproperties as similar contacts produced by other techniques.Schottky contacts formed by deposition of W and Ti on differentSiGe alloys have been characterised. The pinning of the Fermilevel in these alloys is stronger than in pure Si.The device work included fabrication of Schottky diodes foruse as photodetectors and Permeable Base Transistors (PBT).Grid shaped Ti based Schottky diodes were fabricated in 6H-SiC.Optical characterisation shows that the diodes are sensitive inthe UV-range with a peak sensitivity around 300 nm. The diodesare insensitive to visible light. Diodes on p-type materialshow higher sensitivity than diodes on n-type material. This isexpected due to the higher barriers on p-type and the highermobility of electrons, increasing the contribution fromdiffusion.A process for the fabrication of PBT:s on Silicon using selfaligned CoSi2contacts was developed. Devices were fabricatedand DC-characterised. The first PBT on 6H-SiC was fabricatedusing a process based on etching of epitaxial layers. Ni wasused as etch mask and ohmic contact. Ti was used as gatemetal.
|
|
| 8. |
- Grivickas, Paulius, 1974-
(författare)
-
Optical studies of carrier transport and fundamental absorption in 4H-SiC and Si
- 2004
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The Fourier transient grating (FTG) technique and a novelspectroscopic technique, both based on free carrier absorption(FCA) probing, have been applied to study the carrierdiffusivity in 4H-SiC and the fundamental absorption edge in4H-SiC and Si, respectively. FTG is a unique technique capable of detecting diffusioncoefficient dependence over a broad injection interval rangingfrom minority carrier diffusion to the ambipolar case. In thiswork the technique is used for thin epitaxial 4H-SiC layers,increasing the time- and spatial-resolution of the experimentalsetup by factors of ~100 and ~10, respectively, in comparisonto the established Si measurements. It is found that thediffusion coefficient within the detected excitation range inn-type 4H-SiC appears to be lower than the analyticalprediction from Hall-mobility data. To explain this, it issuggested that the minority hole mobility is reduced withrespect to that of the majority one or that the hole mobilityvalue is in general lower than previously reported. Observeddifferences between the temperature dependency of the ambipolardiffusion and the Hall-prediction, on the other hand, areattributed to the unknown Hall factor for holes and theadditional carrier-carrier scattering mechanism in Hallmeasurements. Furthermore, at high excitations a substantialdecrease in the ambipolar diffusion is observed andadditionally confirmed by the holographic transient gratingtechnique. It is shown that at least half of the decrease canbe explained by incorporating into the theoretical fittingprocedure the calculated band-gap narrowing effect, taken fromthe literature. Finally, it is demonstrated that numerical datasimulation can remove miscalculations in the analytical Fourierdata analysis in the presence of Auger recombination. Measurements with variable excitation wavelength pump-probeare established in this work as a novel spectroscopic techniquefor detecting the fundamental band edge absorption in indirectband-gap semiconductors. It is shown that the techniqueprovides unique results at high carrier densities in doped orhighly excited material. In intrinsic epilayers of 4H-SiC,absorption data are obtained over a wide absorption range, atdifferent temperatures and at various polarizations withrespect to the c-axis. Experimental spectra are modeled usingthe indirect transition theory, subsequently extracting thedominat phonon energies, the approximate excitonic bindingenergy and the temperature induced band-gap narrowing (BGN)effect in the material. Measurements in highly dopedsubstrates, on the other hand, provide the first experimentalindication of the values of doping induced BGN in 4HSiC. Thefundamental absorption edge is also detected in highly dopedand excited Si at carrier concentrations exceeding theexcitonic Mott transition by several orders of magnitude. Incomparison to theoretical predictions representing the currentunderstanding of absorption behavior in dense carrier plasmas,a density dependent excess absorption is revealed at 75 K.Summarizing the mainfeatures of the subtracted absorption, itis concluded that an excitonic enhancement effect is present inSi.
|
|
| 9. |
|
|
| 10. |
- Janson, Martin, 1968-
(författare)
-
Hydrogen diffusion and ion implantation in silicon carbide
- 2003
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Secondary ion mass spectrometry (SIMS) has been employed tostudy the spatial distributions resulting from mass transportby diffusion and ion implantation in single crystal siliconcarbide (SiC). By a systematic analysis of this data,fundamental processes that govern these phenomena have beenderived.The acceptor atoms Al and B are known to be electricallypassivated by H in SiC. By studying the thermally stimulatedredistribution of implanted deuterium (2H) in various acceptordoped structures, it is found that hydrogen forms complexeswith the doping atoms, and also interacts strongly withimplantation induced defects. A comprehensive understanding ofthe formation and dissociation kinetics of these complexes hasbeen obtained. The extracted effective capture radius for theformation of 2H-B complexes is in good agreement with thatexpected for a coulomb force assisted trapping mechanism. Thelarge difference of 0.9 eV in the extracted dissociationenergies for the 2H-Al and 2H-B complexes suggests that theatomic configurations of the two complexes are significantlydifferent. Furthermore, by studying the migration behavior of Hin the presence of built-in electric fields, it is concludedthat all of the mobile H is in the positive charge state inp-type SiC.A large number of implantations have been performed withrespect to ion mass, energy, fluence, and crystal orientation.The electronic stopping cross sections in the low velocityregime for ions with atomic numbers 1 ≤ Z1 ≤ 15have been extracted from the ion range distributions. Theydisplay both Z1-oscillations and a smaller than velocityproportional stopping for ions with Z1 ≤ 8, in agreementwith previous reports for other materials. Furthermore, thedegree of ion channeling in various major axial and planarchannels of the 6H and 4H-SiC crystal has been explored. Twotypes of ion implantation simulators have been developed. Onebased on a statistical, data-base approach, and one atomisticsimulator, based on the binary collision approximation (BCA).By fitting BCA simulated profiles to the experimental profiles,detailed information about the electronic stopping andimplantation induced damage is extracted. In addition, thevacancy-related damage caused by the implantations has beeninvestigated by positron annihilation spectroscopy (PAS). Twotypes of implantation induced positron traps have been isolatedand are tentatively identified as a Si vacancy (VSi) and a Si-Cdivacancy (VSiVC). The extension of detected VSi is in goodagreement with that predicted by BCA simulations, and forimplantations with heavier ions VSi are revealed at far greaterdepths than the mean projected ion range due to deeplypenetrating channeled ions.
|
|
| 11. |
- Josyula, Lalita
(författare)
-
Stability of point defects in silicon induced by high energy low dose ion implantation
- 1997
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Ion implantation is a key' process for the introduction ofdopants in semiconductor technology. It involves bombarding thesubstrate material with energetic ions. One of the principleside effects of particle irradiation into crsystallinesemiconductor materials, like Si, is the disruption of theoriginal lattice caused by collisions with incident ionscreating atomic displacements. Point defects mediate dopantdiffusion in semiconductors, at temperatures where ordinarythermal diffusion is negligible. They also introduce energylevels into the forbidden gap of semiconductors.These energy levels can serve as recombination centres forelectrons and holes. While recombination centres are unwantedin, for example silicon photo detectors, where carrier lossshould be kept at a minimum, they are deliberately introducedin silicon power devices to control their switching propertiesas they are effective 'lifetime killers'. For this defectengineering in silicon, proton irradiation is widely usedbecause of the fact that localised damage regions are createdand the damage profile can be precisely controlled by justadjusting the incident ion energy.Implantation induced damage is a function of variousparameters like the energy of the incident ion, mass, dose,dose rate, substrate temperature among others. In this work anattempt has been made to study the effects of variation of theabove mentioned factors on the resultant damage.Very low dosesof ions have been employed and hence, defects formed constitutea very dilute regime of concentration. Deep level transientspectroscopy (DLTS) has been used for sample characterisationafter implantation. The most common defects observed in allroom temperature implanted silicon samples, irrespective ofmass, are the divacancy (V2) and the vacancy-oxygen(VO) centres in n-type,the divacancy (V2) and the carbon-oxygen (CO centres in p-type. Theconcentration of defects increases linearly with increasingdose, provided doses are not very high where these defectsinteract with each other forming higher order complexes. Doserate studies have revealed that the concentration of pointdefects decreases with increasing dose rate and this isatrributed to the fact that the rapidly diffusing siliconself-interstitials (I's) annihilate vacancies created inadjacent ion tracks, due to overlap of cascades, therebyreducing vacancy concentration and thus preventing theformation of vacancy type defects. The dose rate effect is evenmore pronounced for heavy' ion implantation and occurs at lowerdose rates owing to the larger size of an individual collisioncascade and a high density of I's. Implantation at elevatedsubstrate temperatures have indicated a relaxation of thelattice strain created due to ion bombardment and favoured theproduction of unperturbed V2centres as well as VO centres.Implantation induced defects like V2and VO centres which involve broken or danglingbonds are easily passivated at room temperature by hydrogen andcopper. If introduced in a controlled fashion, copper andhydrogen can form. electrically active complexes withdefects.Annealing studies show that V2and VO centres are less stable in Czochralskigrown (CZ)ion implanted silicon compared to electron irradiatedfloat zone silicon (FZ). This is because of a largeconcentration of impurities in CZ silicon like oxygen andcarbon and highly disordered regions in ion implanted sampleswhich act as effective traps for migrating point defects.
|
|
| 12. |
|
|
| 13. |
- Juhasz, Robert, 1975-
(författare)
-
Silicon nanowires, nanopillars and quantum dots : Fabrication and characterization
- 2005
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Semiconductor nanotechnology is today a very well studied subject, and demonstrations of possible applications and concepts are abundant. However, well-controlled mass-fabrication on the nanoscale is still a great challenge, and the lack of nanofabrication methods that provide the combination of required fabrication precision and high throughput, limits the large-scale use of nanodevices. This work aims in resolving some of the issues related to nanostructure fabrication, and deals with development of nanofabrication processes, the use of size-reduction for reaching true nanoscale dimensions (20 nm or below), and finally the optical and electrical characterization to understand the physics of the more successful structures and devices in this work. Due to its widespread use in microelectronics, silicon was the material of choice throughout this work. Initially, a fabrication process based on electron beam lithography (EBL) was designed, allowing controlled fabrication of devices of dimensions down to 30 nm, although, generally, initial device dimensions were above 70 nm, allowing the flexible but low-throughput EBL, to be replaced by state-of-the-art optical lithography in the case of industrialization of the process. A few main processes were developed throughout the course of this work, which were capable of defining silicon nanopillar and nano-wall arrays from bulk silicon, and silicon nanowire devices from silicon-on-insulator (SOI) material. Secondly, size-reduction, as a means of providing access to few-nanometer dimensions not available by current lithography techniques was investigated. An additional goal of the size-reduction studies was to find self-limiting mechanisms in the process, that would limit the impact of variations in the size and other imperfections of the initial structures. Thermal oxidation was investigated mainly for self-limited size-reduction of silicon nanopillars, resulting in well-defined quantum dot arrays of few-nm dimensions. Electrochemical etching was employed to size-reduce both silicon nanopillars and silicon nanowires down into the 10-nm regime. This being a novel application, a more thorough study of electrochemical etching of low-dimensional and thin-layer structures was performed as well as development of a micro-electrochemical cell, enabling electrochemical etching of fabricated nanowire devices with improved control. Finally, the combination of nanofabrication and size-reduction resulted in two successful device structures: Sparse and spatially well-controlled single silicon quantum dot arrays, and electrically connected size-reduced silicon nanowires. The quantum dot arrays were investigated through photoluminescence spectroscopy demonstrating for the first time atomic-like photoemission from single silicon quantum dots. The silicon nanowire devices were electrically characterized. The current transport through the device was determined to be through inversion layer electrons with surface states of the nanowire surfaces greatly affecting the conductance of the nanowire. A model was also proposed, capable of relating physical and electrical properties of the nanowires, as well as demonstrating the considerable influence of charged surface states on the nanowire conductance.
|
|
| 14. |
- Kaplan, Wlodek
(författare)
-
The use of self-aligned Ti silicide in integrated Si technology
- 1998
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The performance and cost efficiency of integrated circuits(IC) are constantly improved by a miniaturization of theindividual device dimensions. As a consequence, the materialand electrical properties of conductors and contacts becomecritical, and fabrication technology development meets newchallenges from the continuous reduction of devicedimensions.Thin film refractory metal silicides have been widely usedbecause of their great importance for industrial applicationsin very large scale integration (VLSI) complementarymetal-oxide semiconductor (CMOS) circuits. The use of silicidesallows the formation of low resistance source, gate and draincontacts which can significantly reduce the resistance of aCMOS gate and the source/drain series resistance compared tonon-silicided structures, and hence improves speedperformance.Most of the silicide applications are realized using aSelf-ALIgned siliCIDE (SALICIDE) process based on titanium(Ti). Titanium silicide (TiSi2) is one of the most attractive materials among therefractory metal silicides due to its relatively hightemperature stability and low reported resistivity. Thesuccessful application of the Ti SALICIDE process has beenreported for most of the IC fabrication technologies. Atsub-micron dimensions the process window for the formation ofTiSi2is very small and rapid thermal processing (RTP)has been developed and successfully used in many applications.For deep sub-micron linewidths a new technique forenhancingTiSi2formation has to be used because the processwindow for ordinary RTP formation of silicide becomes toonarrow.The implementation of the Ti SALICIDE technology in astandard Si technology was investigated. The study focused onsilicide formation in different ambient, etch selectivity,bridging and the Ti - SiO2interaction. Moreover, sheet resistance andcontact resistance measurements were made and yield statisticson fabricated devices were studied in order to fullycharacterize the Ti SALICIDE process. A new etch procedure forself-aligned Ti silicide was proposed. Developed technology wassuccessfully applied in the laboratory scale device fabricationprocesses.The realization of sub-micron VLSI circuits operating atlower voltages suffers from the high resistance of dopedpolysilicon. Hence the Ti SALICIDE process is still one of themost important technologies to enhance circuit performance.Observed difficulties with the formation of the low resistiveC54 titanium silicide phase on sub-micron polysilicon linesoutlined and inspired further studies. A novel method toenhance TiSi2formation from Mo/Ti bilayer was investigated. Aone step Ti SALICIDE process to form TiSi2from Mo/Ti/TiN was proposed. Based on the recentliterature, novel silicide concepts to maintain fabrication ofshallow junctions under the silicide; e.g. elevatedsource/drain approach or selective deposition of TiSi2are described. Progress in these fields indicatesthat TiSi2is still one of the strongest candidates forapplications in a future deep sub-micron CMOS technology.Key words:very large scale integration, VLSI,self-aligned silicide, SALICIDE, titanium silicide, TiSi2, silicidation, bridging, etch selectivity,sub-micron, Mo/Ti bilayer, TiSi2applications.
|
|
| 15. |
- Karlin, Tord E.
(författare)
-
Process integration issues for high-performance bipolar technology
- 1997
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The work in this thesis has been focused on processintegration issues for high-performance bipolar technologyincluding experimental work on self-aligned silicides,ion-implanted andin situdoped polysilicon emitters, strained silicongermanium for heterojunction bipolar transistors and physicalprocess and device simulation.Key issues for the self-aligned silicidation of small devicefeatures such as the influence of dopants, silicon morphologyand line width on titanium disilicide formation, phasetransformation and temperature stability, have been addressed.Pre-amorphization and ion beam mixing by arsenic implantationprior to the suicide formation was shown to extend the use oftitanium silicide into the sub-micron line width range, wheretransformation to the low-resistive phase is otherwise impeded.The temperature stability of cobalt disilicide has also beenstudied.An epitaxial silicon germanium base was integrated into adeep trench isolated double polysilicon high-frequency bipolarprocess. The integrity of the boron and germanium profiles inthe silicon germanium base was investigated for processrelevant furnace and rapid thermal annealings. It was shownthat the outdiffusion of boron and germanium can besignificantly lowered by the use ofin situdoping of the polysilicon emitter compared tothe conventional ion implantation.Physical process and device simulation was utilized as apowerful tool for device design and the development of a shortturn-around time high-frequency bipolar transistor researchprocess. Simulations were used both for prediction andanalysis. Analysis by simulation revealed that the non-linearforward common emitter current gain, observed for the firsttransistors fabricated in the research process, originated froma high interface charge carrier recombination velocity belowthe emitter oxide spacers. Measurements on fabricated devices,with different ratios between the emitter interface area andoxide spacer area, verified the simulation results. It wasshown that passivation with hydrogen could improve thelinearity and peak value of the forward common emitter currentgain as well as the maximum transition frequency.The workhas been carried out within the high-frequencybipolar project at the Department of Electronics at KTH.Keywords:process integration, bipolar technology,silicon, silicon germanium, titanium silicide, phasetransformation, pre-amorphization,in situdoping, physical process simulation, physicaldevice simulation.
|
|
| 16. |
- Kortegaard Nielsen, Hanne
(författare)
-
Capacitance transient measurements on point defects in silicon and silicol carbide
- 2005
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Electrically active point defects in semiconductor materials are important because they strongly affect material properties like effective doping concentration and charge carrier lifetimes. This thesis presents results on point defects introduced by ion implantation in silicon and silicon carbide. The defects have mainly been studied by deep level transient spectroscopy (DLTS) which is a quantitative, electrical characterization method highly suitable for point defect studies. The method is based on measurements of capacitance transients and both standard DLTS and new applications of the technique have been used. In silicon, a fundamental understanding of diffusion phenomena, like room-temperature migration of point defects and transient enhanced diffusion (TED), is still incomplete. This thesis presents new results which brings this understanding a step closer. In the implantation-based experimental method used to measure point defect migration at room temperature, it has been difficult to separate the effects of defect migration and ion channeling. For various reasons, the effect of channeling has so far been disregarded in this type of experiments. Here, a very simple method to assess the amount of channeling is presented, and it is shown that channeling dominates in our experiments. It is therefore recommended that this simple test for channeling is included in all such experiments. This thesis also contains a detailed experimental study on the defect distributions of vacancy and interstitial related damage in ion implanted silicon. Experiments show that interstitial related damage is positioned deeper (0.4 um or more) than vacancy related damage. A physical model to explain this is presented. This study is important to the future modeling of transient enhanced diffusion. Furthermore, the point defect evolution in low-fluence implanted 4H-SiC is investigated, and a large number of new defect levels has been observed. Many of these levels change or anneal out at temperatures below 300 C, which is not in accordance with the general belief that point defect diffusion in SiC requires high temperatures. This thesis also includes an extensive study on a metastable defect which we have observed for the first time and labeled the M-center. The defect is characterized with respect to DLTS signatures, reconfiguration barriers, kinetics and temperature interval for annealing, carrier capture cross sections, and charge state identification. A detailed configuration diagram for the M-center is presented.
|
|
| 17. |
- Lalic, Nenad
(författare)
-
Light emitting devices based on silicon nanostructures
- 2000
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Although silicon is the dominant semiconductor today, lightemitting devices are currently based on compound semiconductorsdue to their direct band-gap, which promotes fast radiativerecombination. However, in nanometer-size silicon structures,carrier confinement enhances the radiative recombination,while, at the same time, suppresses diffusion to non-radiativerecombination centra, resulting in a significant increase inlight emission efficiency. Moreover, the band-gap is wideningas the crystal size is reduced (quantum confinement), enablinglight emission in the visible range. In this work, twodifferent approaches to manufacture a light emitting diode(LED) in silicon have been investigated. The first type ofsilicon LED's is based on porous silicon (PSi) and manufacturedby electrochemical etching of a previously formed pn diodestructure. After optimizing the etching process, PSi LED's wereproduced with an external quantum efficiency of ~0.2% underpulsed excitation, more than an order of magnitude higher thanpreviously reported. Tunability of the emission wavelength inthe range 1.6-2eV was demonstrated by varying the etchingparameters. The EL wavelength is determined by the band-gap ofthe nanocrystals, i. e. their size, as evidenced by a lowerthreshold for longer EL wavelengths, due to lower barriers forinjection into larger crystallites. The EL decay after the biaspulse follows a stretched exponential shape, in agreement witha model involving exciton migration in partially interconnectednanocrystals. Under constant bias, the EL and forward currentare decreasing, due to charging, caused by carrier trapping inthe porous network. After the etching the hydrogen passivatedporous silicon surface is being gradually oxidized, resultingin increased barriers, permanent conductivity reduction and ELdegradation. To improve stability, the second LED approach,based on Si nanocrystals embedded in SiO2, was studied. Nanocrystals were formed by theimplantation of Si into thermally grown SiO2and by subsequent annealing at high temperatures(mostly 1100°C). Photoluminescence investigation showedthat luminescence properties are dependent on nanocrystal sizeand similar to those of PSi. However, decay shapes and timeconstants revealed a stronger isolation of the nanocrystalsthan in PSi. For the EL, good current transport properties werenecessary. That required a thin SiO2layer and efficient injection, realized using anin-situ doped poly-Si cap layer. The Si nanocrystal LED's werestable, although the total light intensity was lower than inPSi, as a consequence of a thin active layer.Key words: Electroluminescence, photoluminescence, lightemitting diode, porous materials, nanostructured materials,silicon, etching, anodized layers, ion implantation.
|
|
| 18. |
|
|
| 19. |
- Lundberg, Nils
(författare)
-
Thermally stable electrical contacts to 6H silicon carbide
- 1996
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Silicon Carbide (SiC) are at present being developed for use in high temperature (≥ 500◦C) and high-power environments under which conventional semiconductors can hardly perform. Thermally stable electrical contacts are essential to the ability of SiC to function under such extreme conditions. In this thesis work the effect of high temperature heat treatments of electrical contacts to 6H-SiC have been investigated. Thermally induced reactions between thin films of Co, Co and Si, W and WC, and silicon face single-crystal 6H-SiC in vacuum at temperatures ranging from 300 to 1100◦C for 1 to 5 hours are investigated by Rutherford backscattering spectrometry (RBS), X-ray diffraction (XRD), X-ray photoelectron spectrometry (XPS), auger electron spectrometry (AES), scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The electrical characteristics of the contacts are characterized using current-voltage (I-V) and capacitance voltage (C-V) measurements.RBS showed that Co reacts with 6H-SiC after 1 hour at 600◦C forming a new phase identified by XRD as Co2Si. CoSi formation started during anneal at 900◦C and that reaction was completed after 1 hour at 1000◦C. CoSi2 was not observed to form even at 1350◦C. Co does not form carbides; thus, the carbon must precipitate somewhere. A depth profile of carbon revealed a homogenous carbon distribution through the film with a distinct accumulation at the surface. TEM investigation of the CoSi film revealed a rough morphology and a very jagged interface. I-V measurements carried out on n-type Schottky contacts after annealing showed an increase in the ideality factor and the leakage current upon reaction. The estimated barrier heights for the contacts were 1.10 eV (n-type) and 1.70 eV (p-type) before reaction and roughly 1.3 – 1.5 ev for both types after Co2Si formation. The specific contact resistivity (rC) of n-epilayers with carrier concentration of 7x1018cm-3 was ≤ 2.5x10-3 Wcm2 after CoSi formation. Corresponding rC values for p-type were 2x1019 cm-3 and 1.05x10-3 Wcm2.To fabricate CoSi2 /SiC contacts, stochiometric deposited films of Co and Si were annealed at 500, 700 and 900◦C, for 5, 2 and 2 hours, respectively. During the 500◦C annealing CoSi2 was formed via a diffusion-controlled reaction yielding a mirror-like surface. The annealings at 700◦C and 900◦C had 2 purposes: a) enabling the silicon deficient CoSi2 film to react with the SiC substrate leading to improved adhesion and b) to test the temperature stability of the contacts. Moreover, TEM investigation of the CoSi2 film revealed a smooth morphology and islands of an amorphous interface layer. I-V measurements after 900◦C anneal revealed an increase in the ideality factor and the leakage current density. On low doped epilayers of SiC the obtained barrier height and ideality factors were 1.05 eV, 1.15 (n-type) and 1.90 eV, 1.38 (p-type) and on highly doped epilayers 1.4x1019 cm-3 and 2x1019 cm-3, rC was 3.0±0.4x10-5 Wcm2 and 4.0±0.7x10-6 Wcm2 for n-type and p-type, respectively.Schottky contacts to low-doped 6H-SiC were fabricated via chemical vapor deposition (CVD) at 400◦C followed by reactive ion etch to pattern the contacts. RBS and TEM verified a distinct W/SiW interface after a two-hour vacuum annealing at 800◦C. I-V and C-V measurements revealed a low n-type barrier height = 0.79 eV, promising for low ohmic applications on high-doped n-type epi-layers. On p-type a high FBp (1.89 eV at 500◦C) was observed, suitable for rectifying applications since it displayed a current rectification ratio at 200◦C of 107 at ±10 V.WC films of 200nm thickness were deposited by CVD from a WF6/C3H8/H2 mixture at 900◦C onto low-doped epi-layers of 6H-SiC. TEM cross sections revealed the film to be poly crystalline with small grains yielding a high electrical resistivity. The interface was clear and unreacted. The barrier height and ideality factor were 0.79 eV, 1.05 (n-type) and 1.81 eV, 1.13 (p-type) at room temperature. For the p-type contact a low reverse current density was found after six hours at 500◦C in room atmosphere, and no signs of contact delamination or contact oxidation were observed.The results of this thesis suggest that, by stoichiometric deposition of refractory silicides and carbides, smooth metallic films to SiC are obtained. The films are thermally stable up to at least 500◦C and reveal useful electrical characteristics. Hence, they are suitable as contacting layers in device mentalizations to SiC intended for high temperature and high-power operation.I-V (current -voltage), C-V (capacitance-voltage), rectifying, ohmic, interface, silicide, carbide, epilayer, Schottky barrier height (FB), current rectification ratio (CRR), specific contact resistivity, transmission line model (TLM), contact resistance RC), transfer length (LT).
|
|
| 20. |
|
|
| 21. |
|
|
| 22. |
|
|
| 23. |
- Mouroux, Aliette
(författare)
-
The reactive formation of TiSi2in the presence of refractory metals
- 1999
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Titanium disilicide (TiSi2) has been the favoured material for contactmetallisation in recent Si devices. The formation of TiSi2usually begins with the high resistivity C49 phaseas a result of the Ti-Si interaction at about 300-550 °Cand finishes with the low resistivity C54 phase through theC49-C54 phase transformation at about 700 °C. The C49-C54transformation becomes increasingly difficult as the devicedimensions are scaled down and remains a technologicalchallenge for dimensions below 0.5 µm. In this thesis, athin layer of Mo deposited between Ti film and Si substrate isused to promote the formation of the C54 phase at temperature100 °C lower than for the C49-C54 transformation. Onnarrow Si lines down to 0.25 µm width, the influence of Moon the formation of TiSi2is even more pronounced than on blanketsubstrates; lower sheet resistance with smaller scatter isobtained in the presence of Mo than without. The enhancement inthe formation of the C54 phase is interpreted as a consequenceof epitaxial effects where the formation of C40 (Mo,Ti)Si2plays a key role. The validity of the template mechanism isverified by replacing Mo with Ta and Nb. The idea of using Taand Nb comes from the fact that TaSi2and NbSi2have the same crystallographic structure andcomparable lattice parameters as (Mo,Ti)Si2. The epitaxial mechanism is confirmed by latticeimaging with a high-resolution microscope. In order to gain aninsight into the thermodynamics of the ternary systems, thepseudo binary phase diagram of TiSi2-NbSi2is studied. Three phase domains are identified,i.e. 1) C54 (Ti,Nb)Si2with Nb varying from 0 to 10 % at the metal sites,2) a mixture of C54 and C40 (Ti,Nb)Si2with Nb being 10 to 25 % at the metal sites, and3) C40 (Ti,Nb)Si2with Nb varying from 25 to 100 % at the metalsites. The resistivity of (Ti,Nb)Si2C54 increases by 1.2 µΩ cm per at. % Nbwhen the Nb concentration varies from 0 to 10 % at the metalsites. The presence of the refractory metals (Ta, Nb or Mo) atthe Si/Ti interface modifies the energetic factors for theformation of C54 TiSi2. The formation of C49 TiSi2is hindered and that of C54 is enhanced. With a Moor Nb interposed layer, the phase of C54 TiSi2can be obtained at temperatures as low as 450°C. Moreover, if a continuous silicide layer in the C40structure is formed at the Si/TiSi2interface, it is the Si diffusion through thisinterfaced layer that is the controlling factor for the C54TiSi2growth. Furthermore, the use of an interposedlayer of Mo, Ta or Nb generally improves the surface morphologyand morphological stability: the TiSi2 formed has a smoothersurface and interface and is more resistant toagglomeration.Key words: Titanium disilicide TiSi2, contact metallisation, phase formation,refractory metals, sub-micron technology, template growth, verylarge scale integration, VLSI, interconnection.
|
|
| 24. |
|
|
| 25. |
- Nordell, Nils
(författare)
-
Metall organic vapour phase epitaxy for advanced III-V devices
- 1993
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Metalorganic vapour phase epitaxy (MOVPE) has proven to be a successful method for growth of structures for advanced optoelectronic semiconductor devices in III-V compounds. This thesis deals with technological and process related aspects of MOVPE from an experimental perspective. Special attention is given to three main questions: uniformity over large areas, p-type dopant diffusion and redistribution in structures for heterostructure bipolar transistors, and planar embedding of high mesas for buried heterostructure lasers.A uniformity within a few per cents for thickness, alloy composition, and dopant incorporation over large areas is fundamental for growth of advanced device structures. We present a new horizontal reactor with a large width-to-height aspect ratio operating at reduced pressure, and with a rotating susceptor for single wafers with a diameter of up to 75 mm. The obtained uniformity over 40 mm diameter is within 1 % for thickness, composition and doping, and within 1 nm for the wavelength of quaternary InGaAsP at = 1560 nm. The doping distribution has been used to estimate the temperature gradient over the wafer to a few tenths of a degree, and the uniformity is most probably limited by gas phase diffusion and depletion of the reactants.Abrupt doping profiles are important for some devices. An example is the high frequency performance of the n-p-n heterostructure bipolar transistor (HBT), which is improved by a very narrow and highly doped p-type base. We have compared Zn and Mg for this application. The memory effect of Mg in the reactor gives doping tails towards the surface, but with Zn we obtained an abruptness of three decades over 60 nm at a maximum doping level exceeding 21019 cm-3. However, a highly n-doped layer of AlGaAs adjacent to the Zn-doped region gives rise to a significant re-distribution of Zn into the AlGaAs layer. At a reduced growth temperature this effect is diminished and the maximum doping level is simultaneously increased.Growth on partly masked and non-planar substrates has become an important field for realisation of devices designed in three dimensions. We have focused on planar regrowth of semi-insulating InP around high mesas of buried heterostructure lasers. Unwanted growth over the masked mesa top usually occurs. These depositions are reduced at high growth temperatures and for growth around low mesas (< 2 m). Addition of CCl4 in the MOVPE process considerably improves the morphology, especially for growth around high mesas. When chlorine is provided to the growing surface nucleation is prevented on the phosphorus-faced {111}B lattice planes and the mask, and this effect permits reproducible planar regrowth not limited by the mesa height.
|
|
| 26. |
- Palmgren, Pål
(författare)
-
Initial stages of metal- and organic-semiconductor interface formation
- 2006
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This licentiate thesis deals with the electronic and geometrical properties of metal-semiconductor and organic-semiconductor interfaces investigated by photoelectron spectroscopy and scanning tunneling microscopy. First in line is the Co-InAs interface (metal-semiconductor) where it is found that Co is reactive and upon adsorption and thermal treatment it alloys with the indium of the substrate to form metallic islands, about 20 nm in diameter. The resulting broken bonds causes As entities to form which are loosely bond to the surface and evaporate upon thermal treatment. Thus, the adsorption of Co results in a rough interface. Secondly the metal-free phthalocyanine (H2PC) - titanium dioxide interface (organic-semiconductor) is investigated. Here it is found that the organic molecules arrange themselves along the substrate rows upon thermal treatment. The interaction with the TiO2 is mainly with the valence Π-electrons in the molecule causing a relatively strong bond, but this interaction is short range as the second layer of molecules retains their molecular character. This results in an ordered adsorption but limited mobility of the molecules on the surface prohibiting well ordered close packed layers. Furthermore, the hydrogen atoms inside the cyclic molecule leave the central void upon thermal treatment. The third case is the H2PC-InAs/InSb interface (organic-semiconductor). Here ordered overlayer growth is found on both substrates where the molecules are preferentially adsorbed on the In rows in the [110] direction forming one-dimensional chains. The InSb-H2PC interface is found to be weakly interacting and the bulk-like molecular character is retained upon both adsorption and thermal treatment. On the InAs-H2PC interface, however, the interaction is stronger. The molecules are more affected by the surface bond and this effect stretches up a few monolayers in the film after annealing.
|
|
| 27. |
|
|
| 28. |
|
|
| 29. |
|
|
| 30. |
|
|
| 31. |
- Tornblad, Olof
(författare)
-
Physical modeling of on-state losses in bipolar Si and SiC power devices
- 1998
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Power losses affect both the installation- and long-term cost of power electronic systems. The installation cost is related to the fatt that power losses in silicon power devices generate heat and make installation of heat sinks and water cooling necessary. If self-heating effects are strong, lotal overheating can eventually lead to device failure. To improve the design of power device systems, more accurate calculations of the Safe Operating Area (SOA) of power devices are desirable. Power semiconductor devices with lower losses are also needed. With the emerg ing SiC technology, much higher drift temperatures can be tolerated and much lower power losses can be achieved for the very high voltage range. In this thesis, on-state losses in bipolar Si and SiC power devices have been studied by comparing measurements to numerical simulations. Carrier distributions under high-leve1 injection were mesured utilizing the technique of Free Carrier Absorption (FCA). Measurements were performed for elevated temperatures under static equilibtium for Si power diodes and Insulated Gate Bipolar Transistors (IGBTs). Potential distributions in power diode structures were meas ured by contacting the samples with a tungsten probe tip. A set of physical models for accurate simulation of bipolar Si power devices is proposed; special attention was drawn to the modeling of minority carrier transport in emitters. Measurements of carrier distributions were canied out also for 4H-SiC power diode structures and the results were compared with simulation. Physical models for simulation of 4H- and 6H- SiC bipolar power devices are suggested. It was found that anisotropic material properties are important for the operation of bipolar 6H-SiC devices.Finally, various contributions to the heat generation term of a recently improved theory were evaluated under extreme, but realistic conditions. It was concluded that heat generation in bipo lar Si power devices, both stationary and transient, can be modeled accurately by only taking the Joule heat and the recombination heat terms into account.
|
|
| 32. |
- Wolborski, Maciej, 1978-
(författare)
-
Termination and passivation of Silicon Carbide Devices.
- 2005
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Silicon carbide rectifiers are commercially available since 2001, and MESFET switches are expected to enter the market within a year. Moreover, three inch SiC wafers can be purchased nowadays without critical defects for the device performance and four inch substrate wafers are announced for the year 2005. Despite this tremendous development in SiC technology, the reliability issues like device degradation or high channel mobility still remain to be solved. This thesis focuses on SiC surface passivation and termination, a topic which is very important for the utilisation of the full potential of this semiconductor. Three dielectrics with high dielectric constants, Al2O3, AlN and TiO2, were deposited on SiC with different techniques. The structural and electrical properties of the dielectrics were measured and the best insulating layers were then deposited on fully processed and well characterised 1.2 kV 4H SiC PiN diodes. For the best Al2O3 layers, the leakage current was reduced to half its value and the breakdown voltage was extended by 0.5 kV, reaching 1.6 kV, compared to non passivated devices. As important as the proper choice of dielectric material is a proper surface preparation prior to deposition of the insulator. In the thesis two surface treatments were tested, a standard HF termination used in silicon technology and an exposure to UV light from a mercury lamp. The second technique is highly interesting since a substantial improvement was observed when UV light was used prior to the dielectric deposition. Moreover, UV light stabilized the surface and reduced the leakage current by a factor of 100 for SiC devices after 10 Mrad γ ray exposition. The experiments indicate also that the measured leakage currents of the order of pA are dominated by surface leakage.
|
|
| 33. |
- Zetterling, Carl-Mikael
(författare)
-
Silicon dioxide and aluminium nitride as gate dielectric for high temperature and high power silicon carbide MOSFETs
- 1997
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Silicon carbide (SIC) is a wide bandgap semiconductor thathas been suggested as a replacement for silicon in applicationsusing high voltages, high frequencies, high temperatures orcombinations thereof. Several basic process steps need to bedeveloped for reliable manufacturing of long-term stableelectronic devices. One important process step is the formationof an insulator on the silicon carbide surface that may be usedas a) a gate dielectric, b) for device isolation or c) forpassivation of the surface. Silicon dioxide and aluminumnitride have been suggested for these purposes. This thesiscovers the investigation of some formation methods for boththese materials on 4H and 6H silicon carbide, and theelectrical characterisation of the resulting films.Commercially available n-type and p-type 4H and 6H SICwafers have been used, and both the silicon face and the carbonface have been investigated. Silicon dioxide has been formed byseveral methods: a) dry thermal oxidation with or without theaddition of TCA (trichloroethane), b) wet oxidation inpyrogenic steam or with awater bubbler, c) oxide deposition byPECVD (plasma enhanced chemical vapor deposition) or LPCVD (lowpressure chemical vapor deposition) and d) oxidation of aevaporated or LPCVD deposited sacrificial layer of silicon. Theinfluence of various cleaning methods prior to oxidation hasbeen studied, as well as post-oxidation and post-metallisationannealing. The aluminum nitride films were grown by MOCVD(metal organic chemical vapor deposition) under various processconditions.Oxidation kinetics have been studied for dry thermaloxidation at 1200 0C. The redistribution of aluminum (p-typedopant in SiC) during dry thermal oxidation has beeninvestigated using SIMS (secondary ion mass spectrometry). Themorphology of the aluminum nitride was determined using x-raydiffraction rocking curves, RHEED (reflection high energyelectron diffraction) and AFM (atomic force microscopy). Thequality of the silicon dioxide used as gate dielectric has beendetermined using breakdown field measurements. High frequencycapacitance-voltage measurements have been used on bothinsulators to a) verify thickness measurements made with othermethods, b) to determine fixed oxide charges by measuring theflatband voltage shifts and c) to quantitatively compare theamount of interface states.For electrical characterisation either aluminum, titanium ordoped polysilicon circular gate contacts of various sizes wereformed on the insulator surface. Flat MOS capacitors weremainly used for the electrical characterisation. U-grooved MOScapacitors, manufactured by RIE (reactive ion etching), wereused to test the quality of oxides grown on vertical surfaces.Two types of MOSFETs (metal oxide semiconductor field effecttransistors) have been fabricated: vertical U-grooved andlateral devices.Keywords:silicon carbide, thermal oxidation, silicondioxide, metal organic chemical vapor deposition (MOCVD),aluminum nitride, capacitance-voltage measurements, MOSFET.
|
|
| 34. |
- Zimmermann, Uwe, 1970-
(författare)
-
Design, Processing and Characterization of Silicon Carbide Diodes
- 2003
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Electronic power devices made of silicon carbide promisesuperior performance over today's silicon devices due toinherent material properties. As a result of the material'swide band gap of 3.2eV, high thermal conductivity, itsmechanical and chemical stability and a high critical electricfield, 4H-silicon carbide devices have the potential to be usedat elevated temperatures and in harsh environments. Shortercarrier lifetimes and a reduction in the necessary width of thelow-doped drift zone in silicon carbide devices compared totheir silicon counterparts result in faster switching speedsand lower switching losses and thus in much more efficientpower devices.High-voltage 4H-silicon carbide diodes have been fabricatedin a newly developed processing sequence, using standardsilicon process equipment. Epitaxial layers grown by chemicalvapor deposition (CVD) on commercial 4H-silicon carbidesubstrates were used as starting material for both mesa-etchedepitaxial and implanted p+n-n+ planar diodes, Schottky diodesand merged pn-Schottky (MPS) diodes, together with additionaltest structures. The device metallization was optimized to givea low contact resistivity on implanted and epitaxial layers anda sufficiently high Schottky barrier with a singlemetallization scheme. Different high-field termination designshave been tested and breakdown voltages of up to 4 kV onimplanted, field-ring terminated diodes were achieved,corresponding to 80% of the critical electric field. A 5kVepitaxial diode design with a forward voltage drop of 3.5V at acurrent density of 100Acm-2 equipped with an implanted junctiontermination extension (JTE) was also fabricated.A new measurement setup was designed and built with thecapability of measuring current-voltage and capacitance-voltagecharacteristics of semiconductor devices at reverse biases upto 10kV. Together with these electrical measurements, theresults of other characterization techniques were used toidentify performance limiting defects in the fabricated siliconcarbide diodes. Increased forward voltage drop of bipolardevices during on-state operation was studied and it was shownthat the stacking faults causing forward degradation arevisible in scanning electron microscopy. With the help ofsynchrotron white-beam X-ray diffraction topographs (SWBXT),electron beam induced current (EBIC) and electroluminescencemeasurements of silicon carbide diodes, the role of screwdislocations as a dominant source of device failure in the formof localized microplasma breakdown was identified. Screwdislocations with and without open core have been found tocause a 20-80% reduction in the critical electric field of4H-silicon carbide diodes, both for low-voltage (150V) andhigh-voltage (~5kV) designs. While micropipes have almost beeneliminated from commercial silicon carbide material,closed-core screw dislocations are still abundant withdensities in the order of 10000cm-2 in state-of-the-art siliconcarbide epitaxial layers.
|
|
| 35. |
|
|
| 36. |
|
|
| 37. |
- Österman, John, 1971-
(författare)
-
Characterization of electrical properties in 4H-SiC by imaging techniques
- 2004
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- 4H-SiC has physical properties supremely suited for a variety of high power, high frequency and high temperature electronic device applications. To fully take advantage of the material's potential, several problems remain to be solved. Two of the most important are (1) the characterization and understanding of crystallographic defects and their electrical impact on device performance, and (2) the introduction of acceptor dopants, their activation and control of the final distribution of charge carriers. Two main experimental methods have been employed in this thesis to analyze 4H-SiC material with respect to the issues (1) and (2): electron beam induced current (EBIC) and scanning spreading resistance microscopy (SSRM), respectively. EBIC yields a map of electron-hole-pairs generated by the electron beam of a scanning electron microscope and collected in the depleted region around a junction. EBIC is conducted in two modes. In the first mode the EBIC contrast constitutes a map of minority carrier diffusion lengths. Results from these measurements are compared to white beam syncrotron x-ray topography and reveal a one-to-one correlation between lattice distortions and the electron diffusion length in n+p 4H-SiC diodes. In the second EBIC mode, the junction is highly reverse biased and local avalanche processes can be studied. By correlating these EBIC results with other techniques it is possible to separate defects detrimental to device performance from others more benign. SSRM is a scanning probe microscopy technique that monitors carrier distributions in semiconductors. The method is for the first time successfully applied to 4H-SiC and compared to alternative carrier profiling techniques; spreading resistance profiling (SRP), scanning electron microscopy (SEM) and scanning capacitance microscopy (SCM). SCM successfully monitors the doping levels and junctions, but none of these techniques fulfill the requirements of detection resolution, dynamic range and reproducibility. The SSRM current shows on the other hand a nearly ideal behavior as a function of aluminum doping in epitaxially grown samples. However, the I-V dependence is highly non-linear and the extremely high currents measured indicate a broadening of the contact area and possibly an increased ionization due to sample heating. Finite element calculations are performed to further elucidate these effects. SSRM is also applied to characterize Al implantations as a function of anneal time and temperature. The Al doping profiles are imaged on cleaved cross-sections and the measured SSRM current is integrated with respect to depth to obtain a value of the total activation. The evaluation of the annealing series shows a continuous increase of the activation even up to 1950 °C. Other demonstrated SSRM applications include local characterization of electrical field strength in passivating layers of Al2O3, and lateral diffusion and doping properties of implanted boron.
|
|
