Silicon carbide (SiC) has been discussed as a promising material for high ... transition metals generate defect levels deep in the band gap of SiC, which trap ...
Link¨ oping studies in science and technology, Dissertation No. 1388
D EEP LEVELS IN S I C
F RANZISKA C. B EYER
Semiconductor Materials Division Department of Physics, Chemistry and Biology Link¨ oping University, SE - 581 83 Link¨ oping, Sweden Link¨ oping, 2011
c Franziska C. Beyer 2011 Copyright unless otherwise stated
ISBN: 978-91-7393-100-7 ISSN: 0345-7524
Printed by LiU-Tryck, Sweden 2011
F¨ ur meine Familie
”...one should not give up too soon and one should not always look for gold at the end of new rainbows...”1 - William Shockley
1 Introduction
to the first Silicon Carbide Conference held in Boston, April 2-3, 1958
Abstract Silicon carbide (SiC) has been discussed as a promising material for high power bipolar devices for almost twenty years. Advances in SiC crystal growth especially the development of chemical vapor deposition (CVD) have enabled the fabrication of high quality material. Much progress has further been achieved in identifying minority charge carrier lifetime limiting defects, which may be attributed to structural defects, surface recombination or point defects located in the band gap of SiC. Deep levels can act as recombination centers by interacting with both the valence and conduction band. As such, the defect levels reduce the minority charge carrier lifetime, which is of great importance in bipolar devices. Impurities in semiconductors play an important role to adjust their semiconducting properties. Intentional doping can introduce shallow defect levels to increase the conductivity or deep levels for achieving semi-insulating (SI) SiC. Impurities, especially transition metals generate defect levels deep in the band gap of SiC, which trap charge carriers and thus reduce the charge carrier lifetime. Transition metals, such as vanadium, are used in SiC to compensate the residual nitrogen doping. It has previously been reported that valence band edges of the different SiC polytypes are pinned to the same level and that deep levels related to transition metals can serve as a common reference level; this is known as the L ANGER -H EINRICH (LH) rule. Electron irradiation introduces or enhances the concentration of existing point defects, such as the carbon vacancy (VC ) and the carbon interstitial (Ci ). Limiting the irradiation energy, Eirr , below the displacement energy of silicon in the SiC lattice (Eirr < 220 keV), the generated defects can be attributed to carbon related defects, which are already created at lower Eirr . Ci are mobile at low temperatures and using low temperature heat treatments, the annealing behavior of the introduced Ci and their complexes can be studied. Deep levels, which appear and disappear depending on the electrical, thermal and optical conditions prior to the measurements are associated with metastable defects. These defects can exist in more than one configuration, which itself can have different charge states. Capacitance transient investigations, where the defect’s occupation is studied by varying the depletion region in a diode, can be used to observe such occupational changes. Such unstable behavior may influence device performance, since defects may be electrically active in one configuration and inactive after transformation to another configuration. This thesis is focused on electrical characterization of deep levels in SiC using deep level transient spectroscopy (DLTS). The first part, papers 1-4, is dedicated to defect studies of both impurities and intrinsic defects in as-grown material. The second part, consisting of papers 5-7, is dealing with the defect content after electron irradiation and the annealing behavior of the introduced deep levels. In the first part, transition metal incorporation of iron (Fe) and tungsten (W) is discussed in papers 1 and 2, respectively. Fe and W are possible candidates to compensate the residual nitrogen doping in SiC. The doping with Fe resulted in one level in n-type material and two levels in p-type 4H-SiC. The capture process is strongly coupled to the i
lattice. Secondary ion mass spectrometry measurements detected the presence of B and Fe. The defects are suggested to be related to Fe and/or Fe-B-pairs. Previous reports on tungsten doping showed that W gives rise to two levels (one shallow and one deep) in 4H- and only one deep level in 6H-SiC. In 3C-SiC, we detected two levels, one likely related to W and one intrinsic defect, labeled E1. The W related energy level aligns well with the deeper levels observed in 4H- and 6H-SiC in agreement with the LH rule. The LH rule is observed from experiments to be also valid for intrinsic levels. The level related to the DLTS peak EH6/7 in 4H-SiC aligns with the level related to E7 in 6HSiC as well as with the level related to E1 in 3C-SiC. The alignment suggests that these levels may originate from the same defect, probably the VC , which has been proposed previously for 4H- and 6H-SiC. In paper 3, electrical characterization of 3C-layers grown heteroepitaxially on different SiC substrates is discussed. The material was of high quality with a low background doping concentration and S CHOTTKY diodes were fabricated. It was observed that nickel as rectifying contact material exhibits a similar barrier height as the previously suggested gold. A leakage current in the low nA range at a reverse bias of -2 V was achieved, which allowed capacitance transient measurements. One defect related to DLTS peak E1, previously presented in paper 2, was detected and suggested to be related to an intrinsic defect. Paper 4 gives the evidence that chloride-based CVD grown material yields the same kind of defects as reported for standard CVD growth processes. However, for very high growth rates, exceeding 100 µm/h, an additional defect is observed as well as an increase of the Ti-concentration. Based on the knowledge from paper 2, the origin of the additional peak and the assumed increase of Ti-concentration can instead both be attributed to the deeper and the shallower level of tungsten in 4H-SiC, respectively. In the second part of the thesis, studies of low-energy (200 keV) electron irradiated as-grown 4H-SiC were performed. In paper 5, bistable defects, labeled EB-centers, evolved in the DLTS spectrum after the annihilation of the irradiation induced defect levels related to DLTS peaks EH1, EH3 and the bistable M-center. In a detailed annealing study presented in paper 6, the partial transformation of M-centers into the EBcenters is discussed. The transition between the two defects (M-centers → EB-centers) takes place at rather low temperatures (T ≈ 400 ◦ C), which suggests a mobile defect as origin. The M-center and the EB-centers are suggested to be related to Ci and/or Ci complex defects. The EB-centers anneal out at about 700 ◦ C. In paper 7, the DLTS peak EH5, which is observed after low- and high-energy electron irradiation is presented. The peak is associated with a bistable defect, labeled F-center. Configuration A exists unoccupied and occupied by an electron, whereas configuration B is only stable when filled by an electron. Reconfiguration temperatures for both configurations were determined and the reconfiguration energies were calculated from the transition kinetics. The reconfiguration B → A can also be achieved by minority charge carrier injection. The F-center is likely a carbon related defect, since it is already present after low-energy irradiation.
ii
arvetenskaplig sammanfattning Popul¨ ¨r b¨ aller elektriska studier av kiselkarbid. Dess egenskaper a attre Avhandlingen inneh˚ anpassade f¨ or h¨ ogtemperatur, h¨ ogeffekt och h¨ ogfrekvens (h¨ oghastighets) till¨ ampninger ¨n t.ex. kisel. Idag finns det elektriska komponenter framst¨ ¨verallt a allda av kisel n¨ astan o ¨r att runt omkring oss i k¨ oket, i bilen, i telefon och dessutom i datorer... Dagens filosofi a ¨ka deras prestanda. Detta inneb¨ minska alla komponentdelar och samtidigt o ar m˚ anga nya utmaningar... Ju mer man krymper elektroniken, desto b¨ attre m˚ aste materialet leda bort v¨ armet, ¨n kiselkarsom utvecklas i drift. Eftersom kisel har en l¨ agre termisk ledningsf¨ orm˚ agan a ¨r kylningen ett stort problem. Kisel a ¨r dock ganska enkelt att framst¨ bid, a alla utg˚ aende ¨r en blandning av kisel och kol, finns fr˚ an SiO2 , som finns i naturen. Kiselkarbid, som a inte naturligt, s˚ a det m˚ aste man tillverka. Eftersom bindningen mellan de tv˚ a elementen ¨r stark, a ¨r det sv˚ a art att tillverka materialet. Kiselkarbid klarar h¨ oga temperaturer, men det beh¨ ovs ocks˚ a v¨ aldigt h¨ oga temperaturer f¨ or att skapa materialet. Kiselkarbid best˚ ar av flera lager kisel och kol. Efter det f¨ orsta dubbellagret (Si-C), har de f¨ oljande lagren olika m¨ ojligheter att s¨ atta sig p˚ a. Det bildas s˚ a kallade polytyper, som alla best˚ ar av samma Si-C skikt men som har olika ordningsf¨ oljd, vilket medf¨ or olika egenskaper. I den h¨ ar avhandlingen beskrivs mest elektriska egenskaper av 4H-, 6H- och 3C¨r ett halvledande material, vilket betyder att det har ett bandgap. SiC. Kiselkarbid a Det betyder att elektrisk ledning sker n¨ ar elektroner f˚ ar s˚ a pass mycket energi att de ¨ver gapet till ledningsbandet, d¨ exciteras o ar de kan r¨ ora sig n¨ astan som fria elektroner. I valensbandet efterl¨ amnas ett h˚ al, som beter sig som en elektron fast med motsatt ¨r mycket stort, f¨ laddning. Eftersom bandgapet a or 4H-SiC n¨ astan tre g˚ anger st¨ orre ¨n i kisel, finns det inte tillr¨ a acklig med termisk energi och d¨ arf¨ or f¨ orekommer ingen elektrisk ledning vid rumstemperatur. Genom att dopa kristallen med st¨ oratomer som antingen kan ge elektroner (donator) eller binda elektroner (acceptor) kan man styra ledningsf¨ orm˚ agan. Under tillv¨ axtprocessen skapas ocks˚ a defekter djupare i bandgapet, som oftast p˚ averkar materialets beteende. Laddningsb¨ arare (elektron eller h˚ al) kan f˚ angas in i dessa niv˚ aer och d¨ arf¨ or bidrar de inte l¨ angre till ledningen. S˚ adana djupa niv˚ aer kan vara relaterade till Si-C gittret sj¨ alv (n˚ agon atom sitter p˚ a fel plats) eller ett helt annat element (f¨ ororening) som sitter p˚ a en Si eller C- gitterplats. Man kan m¨ ata koncentrationen av de djupa niv˚ aer och deras position i bandg˚ apet med hj¨ alp av deep level transient spectroscopy (DLTS). I f¨ orsta delen av avhandlingen diskuteras defekter som skapas under tillv¨ axten, b˚ ade intrinsiska och f¨ ororeningsrelaterade defekter (extrinsiska defekter). I de f¨ orsta tv˚ a artiklarna beskrivs inkorporering av j¨ arn och wolfram i SiC. I DLTS spektrat visade sig en ¨ ¨vre delen av bandgapet. Aven j¨ arnrelaterad topp med aktiveringsenergi i o tv˚ a niv˚ aer i nedre delen av bandgapet f¨ oresl˚ as vara orsakad av j¨ arninkorporering. Man vet att wolfram skapar tv˚ a defektniv˚ aer i 4H- och en i 6H-SiC. Valensbanden f¨ or de olika SiC poly¨r olika stora - st¨ typerna ligger ungef¨ ar p˚ a samma niv˚ a. Eftersom bandgapen a orst f¨ or ¨r konstant, hamnar den 4H-SiC och avst˚ andet mellan defektniv˚ aer och valensbandet a grundare W niv˚ an, som detekterades i 4H-SiC, i ledningsbandet f¨ or 6H- eller 3C-SiC. Tidigare ber¨ akningar visar ocks˚ a en niv˚ a i 3C-SiC, som nu kunde m¨ atas experimentellt. iii
¨verg˚ Det visade sig att niv˚ aer relaterat till o angsmetaller som W ligger p˚ a samma niv˚ a ¨ven f¨ i alla SiC-polytyper, den s˚ a kallade L ANGER -H EINRICH regel. Regeln st¨ ammer a or intrinska defekter relaterat till samma defekt. Tredje artikeln behandlar intrinsiska defekter i 3C-SiC. Det visade sig en topp i DLTS spektrat, som f¨ oresl˚ as vara relaterat till en intrinsisk defekt. I Link¨ oping har det utvecklats en klorbaserat tillv¨ axtmetod. Elektrisk karakterisering av material odlat med den tekniken visar att den nya metoden inte skapar nya defekter relaterade till klor (papper 4). Den h¨ oga tillv¨ axthastigheten kan dock medf¨ ora att f¨ ororeningar av wolfram detekteras med DLTS. I andra delen av avhandlingen studerades defekter, som skapats efter elektronbestr˚ alning. Elektronbestr˚ alning anv¨ ands f¨ or att unders¨ oka hur komponenter p˚ averkas i processningen och f¨ or grundl¨ aggande defektstudier. SiC material blir bombarderat av elektroner med h¨ og hastighet. Kollisioner med gitteratomer skapar punktdefekter genom att atomen sparkas ut fr˚ an deras position i atomgittret. V¨ armebehandlingen till˚ ater defektatomer att flytta sig tillbaka till deras urprungliga l¨ age. Defektanalyser i avhandlingen fokuserar p˚ a l˚ agenergi elektronbestr˚ alat material. Vid denna energi p˚ a¨r verkar mest kolatomer i SiC kristallen, eftersom tr¨ oskelr¨ orelseenergin f¨ or Si-atomer a h¨ ogre. D¨ arf¨ or kan de observerade defektniv˚ aerna relateras med kol. M-centret i 4H-SiC, ¨r observerat efter h¨ som tidigare a ogenergibestr˚ alning, blev detekterat redan i l˚ agenergi elektronbestr˚ alat material. Efter v¨ armebehandling f¨ orsvinner M-centret och flera nya ¨r relaterade till nya defekter, de s˚ DLTS toppar uppst˚ ar, som a a kallade EB-centerna. En koppling mellan defekterna kunde bekr¨ aftas. B˚ ada M och EB-centerna visar sig instabilt mot p˚ alagd sp¨ anning och temperatur. De uppvisar ett metastabilt beteende, eftersom de har flera konfigurationer och kan byta mellan de olika konfigurationerna, om tillr¨ acklig ¨r n¨ ¨verkomma aktiveringsbarri¨ ¨r energi a arvarande f¨ or att o arerna. Metastabilt beteende a viktigt att studera, eftersom komponenter inte b¨ or visa instabiliteter i drift. Efter ytterligare v¨ armebehandling vid h¨ ogre temperaturer, f¨ orsvinner EB-centerna. I sista pappret studeras ytterligare en metastabil defekt, den s˚ a kallade F-center och dess rekonfigu¨r n¨ rationsbeteende. Det betyder vilken temperatur och d¨ armed energi a odv¨ andigt f¨ or att byta konfiguration och hur snabbt konfigurations¨ overg˚ angen sker. Alla defekter: ¨r relaterade till kol som befinner sig i en mellangitM-center, EB-centerna och F-center a ¨r n¨ terposition eller komplex av kol-interstitialer, eftersom de a arvarande efter l˚ agenergi elektronbestr˚ alning och de f¨ orsvinner vid v¨ armebehandling vid relativt l˚ aga temperaturer.
iv
Preface This thesis is presenting the experimental work which was conducted during the years 2006 to 2011 in the Semiconductor Materials division at the Department of Physics, Chemistry and Biology at Link¨ oping University, Sweden. The thesis is divided into three parts, representing a short introduction to the scientific field of the characterized material and the applied characterization techniques, a summary of the presented work and finally the collection of the included seven papers. The papers, which are included in this thesis are listed on the next page and my contribution to each paper is subsequently specified. In the following pages, my journal articles as well as my conference contributions, which are related to the work but not included in this thesis, will be presented.
v
Included papers: allstr¨ om, A. Henry, and [1] F. C. Beyer, C. G. Hemmingsson, S. Leone, Y.-C. Lin, A. G¨ E. Janz´en. Deep levels in iron doped n- and p-type 4H-SiC. Submitted to J. Appl. Phys., 2011. [2] F. C. Beyer, C. G. Hemmingsson, A. G¨ allstr¨ om, S. Leone, H. Pedersen, A. Henry, and E. Janz´en. Deep levels in tungsten doped n-type 3C-SiC. Appl. Phys. Lett., 98:152104, 2011. [3] F. C. Beyer, S. Leone, C. Hemmingsson, A. Henry, and E. Janz´en. Deep levels in hetero-epitaxial as-grown 3C-SiC. AIP Conf. Proc., 1292:63–66, 2010. [4] F. C. Beyer, H. Pedersen, A. Henry, and E. Janz´en. Defects in 4H-SiC layers grown by chloride-based epitaxy. Mater. Sci. Forum, 615-617:373–376, 2009. [5] F. C. Beyer, C. G. Hemmingsson, H. Pedersen, A. Henry, J. Isoya, N. Morishita, T. Ohshima, and E. Janz´en. Bistable defects in low-energy electron irradiated ntype 4H-SiC. Phys. stat. sol. - RRL, 4:227–229, 2010. [6] F. C. Beyer, C. Hemmingsson, H. Pedersen, A. Henry, E. Janz´en, J. Isoya, N. Morishita, and T. Ohshima. Annealing behavior of the EB-centers and M-center in low-energy electron irradiated n-type 4H-SiC. J. Appl. Phys., 109:103703, 2011. [7] F. C. Beyer, C. G. Hemmingsson, H. Pedersen, A. Henry, E. Janz´en, J. Isoya, N. Morishita, and T. Ohshima. Capacitance transient investigations of the bistable F-center in electron irradiated n-type 4H-SiC. Manuscript.
My contribution to the papers: paper 1 - 6 Planned and conducted all the electrical characterization, analyzed the data, wrote the manuscript and finalized the paper. paper 7 Planned and conducted the electrical characterization for the low-energy irradiated samples, analyzed the data for the low-energy irradiated samples, wrote the manuscript and finalized the paper.
vii
Not included journal articles: [1] S. Leone, F. C. Beyer, H. Pedersen, O. Kordina, A. Henry, and E. Janz´en. Growth of stepbunch free 4H-SiC epilayers on 4 ◦ off-axis substrates using chloride-based CVD at very high growth rates. Mater. Res. Bulletin, 46:1272–1275, 2011. [2] S. Leone, F. C. Beyer, A. Henry, C. Hemmingsson, O. Kordina, and E. Janz´en. Chloride-based SiC epitaxial growth toward low temperature bulk growth. Cryst. Growth Des., 10:3743– 3751, 2010. [3] S. Leone, F. C. Beyer, A. Henry, O. Kordina, and E. Janz´en. Chloride-based CVD of 3C-SiC epitaxial layers on 6H (0001) SiC. Phys. stat. sol. - RRL, 4:305–307, 2010. [4] S. Leone, F. C. Beyer, H. Pedersen, S. Andersson, O. Kordina, A. Henry, and E. Janz´en. Chlorinated precursors study in low-temperature CVD of 4H-SiC. Thin Solid Films, 519:3074– 3080, 2011. [5] S. Leone, F. C. Beyer, H. Pedersen, O. Kordina, A. Henry, and E. Janz´en. High growth rate of 4H-SiC epilayers grown on on-axis substrates with different chlorinated precursors. Cryst. Growth Des., 10:5334–5340, 2010. [6] P. Carlsson, N. T. Son, F. C. Beyer, H. Pedersen, J. Isoya, N. Morishita, T. Ohshima, and E. Janz´en. Deep levels in low-energy electron-irradiated 4H-SiC. Phys. stat. sol. - RRL, 4:121–123, 2009. [7] H. Pedersen, F. C. Beyer, A. Henry, and E. Janz´en. Acceptor incorporation in SiC epilayers grown at high growth rate with chloride-based CVD. J. Cryst. Growth, 311:3364–3370, 2009. [8] H. Pedersen, F. C. Beyer, J. Hassan, A. Henry, and E. Janz´en. Donor incorporation in SiC epilayers grown at high growth rate with chloride-based CVD. J. Cryst. Growth, 311:1321– 1327, 2009. [9] H. Pedersen, S. Leone, A. Henry, F. C. Beyer, V. Darakchieva, and E. Janz´en. Very high growth rate of 4H-SiC epilayers using chlorinated precursor methyltrichlorsilane (MTS). J. Cryst. Growth, 307:334–340, 2007.
viii
Not included conference articles: allstr¨ om, B. Magnusson, F. C. Beyer, A. Gali, N. T. Son, S. Leone, I. G. Ivanov, A. Henry, [1] A. G¨ C. G. Hemmingsson, and E. Janz´en. The Electronic Structure of Tungsten in 4H-, 6H- and 15R-SiC. Submitted for publication in Mater. Sci. Forum, 2011. - invited oral presentation by A. G¨ allstr¨ om [2] S. Kotamraju, B. Krishnan, F. C. Beyer, A. Henry, O. Kordina, E. Janz´en, and Y. Koshka. Electrical and Optical Properties of High-Purity Epilayers Grown by the Low-Temperature Chloro-Carbon Growth Method. Submitted for publication in Mater. Sci. Forum, 2011. - oral presentation by Y. Koshka [3] S. Leone, H. Pedersen, F. C. Beyer, S. Andersson, O. Kordina, A. Henry, A. Canino, F. La Via, and E. Janz´en. Chloride-Based CVD of 4H-SiC at High Growth Rates on Substrates with Different Off-Angles. Submitted for publication in Mater. Sci. Forum, 2011. - oral presentation by S. Leone [4] A. Henry, S. Leone, F. C. Beyer, H. Pedersen, O. Kordina, S. Andersson, and E. Janz´en. SiC epitaxy growth using chloride-based CVD. Accepted for publication in Physica B: Cond. Matter, 2011 - oral presentation by A. Henry [5] A. G¨ allstr¨ om, B. Magnusson, F. C. Beyer, A. Gali, N. T. Son, S. Leone, I. G. Ivanov, A. Henry, C. G. Hemmingsson, and E. Janz´en. Optical Identification and Electronic Configuration of Tungsten in 4H- and 6H-SiC. Accepted for publication in Physica B: Cond. Matter, 2011. - oral presentation by E. Janz´en [6] F. C. Beyer, C. Hemmingsson, H. Pedersen, A. Henry, J. Isoya, N. Morishita, T. Ohshima, and E. Janz´en. Observation of bistable defects in electron irradiated n-type 4H-SiC. Mater. Sci. Forum, 679 - 680:249–252, 2011. - poster presentation by F. C. Beyer [7] A. Henry, S. Leone, F. C. Beyer, S. Andersson, O. Kordina, and E. Janz´en. Chloride-based CVD of 3C-SiC on (0001) α SiC. Mater. Sci. Forum, 679 - 680:75–78, 2011. - oral presentation by A. Henry [8] S. Leone, Y. C. Lin, F. C. Beyer, S. Andersson, H. Pedersen, O. Kordina, A. Henry, and E. Janz´en. Chloride-based CVD at high rates of 4H-SiC on-axis epitaxial growth. Mater. Sci. Forum, 679 - 680:59–62, 2011. - oral presentation by S. Leone [9] S. Leone, F. C. Beyer, A. Henry, O. Kordina, and E. Janz´en. Chloride-based CVD of 3C-SiC Epitaxial layers on On-axis 6H (0001) SiC Substrates. AIP Conf. Proc., 1292:7–10, 2010. - oral presentation by S. Leone [10] F. C. Beyer, C. Hemmingsson, H. Pedersen, A. Henry, J. Isoya, N. Morishita, T. Ohshima, and E. Janz´en. Defects in low-energy electron irradiated n-type 4H-SiC. Physica Scripta, T141:014006, 2010. - oral presentation by F. C. Beyer ix
NOT INCLUDED CONFERENCE ARTICLES:
[11] F. C. Beyer, C. Hemmingsson, H. Pedersen, A. Henry, J. Isoya, N. Morishita, T. Ohshima, and E. Janz´en. Metastable defects in low-energy electron irradiated n-type 4H-SiC. Mater. Sci. Forum, 645-648:435–438, 2010. - oral presentation by F. C. Beyer [12] H. Pedersen, S. Leone, A. Henry, F. C. Beyer, A. Lundskog, and E. Janz´en. Chloride-based SiC epitaxial growth. Mater. Sci. Forum, 615-617:89–92, 2009. - oral presentation by H. Pedersen [13] H. Pedersen, S. Leone, A. Henry, F. C. Beyer, V. Darakchieva, and E. Janz´en. Very high epitaxial growth rate of 4H-SiC using MTS as chloride-based precursor. Mater. Sci. Forum, 600-603:115–118, 2009. - oral presentation by H. Pedersen [14] S. Hahn, F. C. Beyer, A. G¨ allstr¨ om, P. Carlsson, A. Henry, B. Magnusson, J. R. Niklas, and E. Janz´en. Contactless electrical defect characterization of semi-insulating 6H-SiC bulk material. Mater. Sci. Forum, 600-603:405–408, 2009. - poster presentation by S. Hahn [15] A. G¨ allstr¨ om, B. Magnusson, P. Carlsson, N. T. Son, A. Henry, F. C. Beyer, M. Syv¨ aj¨ arvi, R. Yakimova, and E. Janz´en. Influence of cooling rate after high temperature annealing on deep levels in high-purity semi-insulating 4H-SiC. Mater. Sci. Forum, 556-557:371–374, 2007. - oral presentation by A. G¨ allstr¨ om [16] A. Henry, J. ul Hassan, H. Pedersen, F. C. Beyer, J. P. Bergman, S. Andersson, E. Janz´en and P. Godignon. Thick Epilayers for Power Devices. Mater. Sci. Forum, 556-557:47–52, 2007. - invited oral presentation by A. Henry -
x
Acknowledgements I would like to express my gratitude to all the people having made this dissertation possible: • Prof. Erik Janz´ en - my supervisor. I am truly thankful for the opportunity to be part of your research group. Thanks for the unquestionable confidence you had in me. • Dr. Carl Hemmingsson - my second supervisor, who helped me out of all confusing problems. Thanks for your support whenever I needed it. • Prof. Karin Sundblad-Tonderski - my mentor. Thank you for listening to me during our shared lunch and coffee breaks. It was a pleasure to talk with you about work and life. • I am obliged to all my colleagues in the Semiconductor Materials group, who supported me throughout the years, especially Dr. Anne Henry for her continuous encouragement. • Arne Eklund, Eva Wibom and Louise Gustafsson Rydstr¨ om - thanks for all technical and administrative help and for the nice working environment. ˚ forsk, LM Ericssons stiftelse f¨ • A or fr¨ amjande av elektroteknisk forskning and SiCED Student Support thanks for financial support. • Sven Andersson - thanks for your support during the tender procedure and the installation of our Venus. It was a great time with you. • Andreas G¨ allstr¨ om - thanks for your invaluable feedback. I like your view of life! • Dr. Patrick Carlsson, Dr. Stefano Leone, Dr. Daniel Dagnelund, Dr. Henrik Pedersen and Dr. Jawad ul Hassan - thanks for discussions on work and life in general. Thanks to all the other PhDs or PhD-students, for the nice time we spent together, especially during conferences. • My friends here in Sweden but also in Germany, who made life so joyful. • My parents, parents-in-law and my french family, who always believed in me and my work and who pushed me up when I needed encouragements. Thanks for all the love and for the time you spent with me and my family, especially with my kids. Thanks to the rest of my family; even though there were large distances between us, we were always feeling close. • Last but not least... I want to thank my love Jan and my dear daughters, Samira and Camilla, who made and still make my life so wonderful and meaningful. Without you, I could not imagine to exist anymore... Franziska xi
Contents 1 Motivation
1
2 Silicon Carbide 2.1 SiC: crystal structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Electrical properties of SiC . . . . . . . . . . . . . . . . . . . . . . . . . .
5 5 7
3 Point defects in the bandgap 3.1 Classification . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1 Energetic properties . . . . . . . . . . . . . . . . 3.1.2 Interaction with the bands . . . . . . . . . . . . . 3.2 Doping . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Electrically active intrinsic defects in SiC . . . . . . . . . 3.3.1 Irradiation induced defects investigated by DLTS 3.4 Transition metal related levels in SiC . . . . . . . . . . . 3.5 Defect annealing . . . . . . . . . . . . . . . . . . . . . . 3.6 Metastability . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
9 10 10 11 12 13 13 15 15 16
4 Electrical measurements 4.1 Depletion region . . . . . . . . . . . . . . 4.2 Current-voltage characteristic . . . . . . . 4.3 Capacitance-voltage measurements . . . . 4.4 Transient spectroscopy . . . . . . . . . . . 4.4.1 Rate equations . . . . . . . . . . . 4.4.2 Capture processes . . . . . . . . . 4.4.3 Emission process . . . . . . . . . . 4.4.4 Time dependent trap concentration 4.4.5 DLTS . . . . . . . . . . . . . . . . . 4.4.6 MCTS . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
19 20 21 23 25 25 26 28 29 32 34
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
5 Summary of papers
37
Bibliography
41
1 Motivation At the first International Conference on Silicon Carbide in 1958, William Shockley announced that silicon carbide (SiC) will be the prominent semiconductor to follow silicon (Si)[1] long before diamond-based electronics will become commercially available. Almost 200 years have passed since the first synthesis of SiC was published by J. J. Berzelius in 1824[2]. In 1891, E. G. Acheson discovered a process to manufacture SiC, which he patented in 1892[3]. He labeled the dark SiC crystals carborundum. The abrasive properties of this material were extraordinary and its hardness (M OHS hardness of 9.5) were inferior only to diamond (M OHS hardness of 10). Shortly after the commercial manufacture of SiC, B. W. Frazier discovered that the SiC crystals had different crystal symmetries[4], the concept polytypism was born. A consistent notation of the different polytypes classified by the stacking and orientation of the double layers in the unit cell was later introduced by L. S. Ramsdell[5] in 1947. In 1893, F. H. Moissan found (1893) and analyzed (until 1904) the only natural SiC crystal, labeled after his discoverer moissanite, in meteorite impact material in the Arizona desert [6]. In 1906, H. H. C. Dunwoody patented the first solid-state detector, a point contact diode made of SiC to receive radio-waves[7]. The first current-voltage (IV) characteristics showing a typical diode behavior, labeled as ”unilateral conductive”, were presented in the following year by G. W. Pierce[8]. Measuring IV led to the next step shortly afterwards, the observation of colored luminescence, known as electroluminescence, by applying a voltage to the diode by H. J. Round[9]. The invention of the light emitting diode (LED) started using the same material and in the same year by a young Russian scientist, O. V. Losev[10, 11]. Besides IV measurements, the first temperature dependent resistivity measurements on SiC crystals were performed by K. Lehovec in 1953[12]. The improvement of the Acheson process by J. A. Lely in 1955[13] led to a more controlled process. However, the grown crystals were small and non-uniform in shape. For a commercial breakthrough, large amounts and high quality SiC crystals were necessary. As William Shockley stated at the conference in 1958: ”The SiC situation suffers from the very same thing that makes it good. The bond is very strong and so all processes go on at a very high temperature.” Thus suitable processes had to been discovered to surmount this inconvenience. During the coming years, important milestones pushed the SiC growth. In 1973, Yu. M. Tairov and V. F. Tsvetkov used a small SiC crys1
tal grown with the L ELY-method as seed for their sublimation growth process[14–16], which led to bulk crystals. This physical vapor transport (PVT) growth technique was further improved by PhD students at the North Carolina State University, who founded Cree Research Inc.[17] in 1987. The first single crystal SiC wafers were commercially available in the beginnings of 1990s. Along with the sublimation growth, Japanese researchers introduced SiC vapor phase epitaxial (VPE) growth processes[18–20] on Si and discussed the growth mechanisms[21]. Further developments of epitaxial chemical vapor deposition (CVD) using a hot-wall setup[22] and now-a-days chloride based epitaxy[23, 24] resulted in high quality layers. The advantages of high purity gas precursors were adapted to the known sublimation reactors and a combined system, labeled high temperature CVD, was developed by O. Kordina[25]. All these prerequisites made it possible to increase the substrate diameter and the growth rate, while decreasing the costs. The quality of the epitaxial layers as well as device structures was enhanced and hence in 2001, the first S CHOTTKY diode was commercially available by Infineon and shortly after also by Cree. Along with the progress in bulk crystal growth and the availability of low-doped epitaxial layers, electrical characterization, such as deep level transient spectroscopy (DLTS), became an important tool to study intrinsic and extrinsic defects in SiC. One of the first DLTS studies was performed on 3C-SiC grown on Si[26]. Later-on, 3C investigations stagnate until promising high quality material was heteroepitaxially deposited on other SiC-polytypes and thus the high background-doping was lowered[27]. Important fields of electrical investigation were and are the identifications of shallow dopants, deep levels related to transition metal impurities, irradiation/implantation induced defects and their annealing behavior as well as intrinsic defects, which are limiting the minority charge carrier lifetime. The shallow donor levels, such as nitrogen[28] and phosphorous[29], and acceptor levels, e.g. aluminum[30] and boron[31], were studied in the middle of the 90s. Deep levels attributed to metal incorporation, especially transition metal incorporation[32, 33] (Cr, Ti, V, W, and Ta) were investigated using radio-tracer DLTS[34, 35], which observes the radioactive decay from implanted isotopes to daughter-isotopes. Among the transition metals, vanadium was studied most due to its ability to compensate the residual nitrogen doping[36–38]. Particle (e.g. ions, electrons and protons) irradiation is used to investigate defects related to the impinging particle or intrinsic defects. By such kind of artificial irra2
diation, device processing steps can be simulated. High-energy irradiation will create highly defective areas in both Si and C-sublattice. Important DLTS studies were done by Hemmingsson et al.[39, 40], Doyle et al.[41], Dalibor et al.[42] and others[43–46]. Restricting the irradiation energy, Eirr , below the energy needed for displacing the Si atom in the SiC crystal (Eirr < 220 keV), will give the possibility to relate irradiation induced defects to the carbon-atom and/or C-related defects. Danno et al.[47] and Storasta et al.[48, 49] contributed with important studies in this field. The defective material after irradiation can in most instances be cured by annealing[44, 49–52]. Low temperature annealing will activate carbon interstitials to move around and to annihilate or to form complexes, whereas less mobile defects, e.g. carbon vacancies need more thermal activation energy and related deep levels can still be detected after annealing at T > 1500 ◦ C [51, 53, 54]. With the help of thermal oxidation[55, 56] or carbon implantation and annealing [57], the prominent Z1/2 level in 4H-SiC detected by DLTS, which is attributed to be the minority life-time killing defect[58], was significantly reduced and thus the assignment to the carbon vacancy was strengthened. In 2011, Sasaki et al.[54] presented a comparative DLTS study between intrinsic defects detected in 4H- and 6H-SiC and their behavior after thermal oxidation and annealing. The intrinsic defect levels align in the two different SiC-polytypes (Z1/2 and EH6/7 in 4H-SiC with E1/2 and E7 in 6H-SiC, respectively) in a similar way as earlier observed for the transition metal defect levels in semiconductor heterostructures[59], known as the L ANGER -H EINRICH rule. The Z1/2 in 4H-SiC and the E1/2 in 6H-SiC show a negative U behavior[60, 61], i.e. both defect levels capture two electrons and the binding energy of the second electron is larger than that for the first electron. Their correlation was firstly suggested by Hemmingsson et al.[61]. Manufactured devices need to be stable during operation. They should not only withstand high temperatures, voltages and frequencies, but also varying electric fields. There are defects, commonly known as metastable defects, which depend strongly on the electrical, thermal and optical environment and can exist in more than one configuration. During device operation, the conditions may change and the defects, which previously were in an electrically inactive state transform to another configuration, which now will trap charge carriers and thus change the device performance. Metastable defects in SiC have been observed for the first time in high-energy electron irradiated 6H-SiC[62]. In 4H-SiC, the M-center[63, 64] was detected in high-energy-proton im3
planted material. Defect characterization of manufactured devices combined with basic defect studies on simple S CHOTTKY diodes, as presented in this thesis, may finally help to find the defects responsible for bad device behavior. Recently, the electrical behavior of processed transistors[65] was studied after radiation damage and subsequent annealing. The authors suggested that other defects than the Z1/2 and EH6/7 in 4H-SiC are also crucial for the device performance, since complete recovery was achieved at such low temperatures (T < 800 ◦ C [66]), where Z1/2 and EH6/7 are usually unaffected. Coming back to William Shockley; he finished his introductory words at the conference: ”... the (growth) approach might very well be different. Perhaps something on a really large scale has to be done, so that large crystals will result.” He was right; on the last European conference held in Oslo 2010, Cree showed the first 6 inch SiC wafer... Now in 2011, Cree launched a 1700 V S CHOTTKY diode[67] for e.g. variable speed motor drives or power converters in wind energy systems. Replacing Si-components by SiC devices lead to more efficient energy conversion processes and reduced switching losses. The first SiC MOSFET is now commercially available[68] for 1200 V at 80 mΩ on-resistance. If we use the generated energy as efficient as possible, we can shut down more and more nuclear power plants without fearing an energy collapse... The aim of this thesis is to deepen and expand the knowledge of defects in SiC and to use the gained information to obtain a common view of extrinsic and intrinsic defects in different SiC polytypes using electrical characterization techniques. In the following sections, first structural and electrical properties of silicon carbide will be presented, how the crystal is build up and which properties come along with the crystal structure. Afterwards, defects and their origin will be discussed in detail. At last, an overview of possible deep level characterization techniques, which were applied in this thesis, will be given.
4
2 Silicon Carbide 2.1
SiC: crystal structure
Silicon carbide is build up equally of silicon and carbon atoms. One C atom is bound to four neighboring Si atoms, which are placed in the corners of a tetrahedron. Vice versa is the Si atom bound similarly to four C atoms. The Si-C bond is very strong ˚ ) and its sp3 hybridization; it has a nearly covalent due to its short bond length (1.89 A character. However, slightly different electronegativity values for C and Si (EN(C) = 2.55 and EN(Si) = 1.9 after PAULING) imply a small ionic contribution (≈ 10 %) to the SiC bond[69]. The stacking of the double layers, composed of one Si and one C layer, reaches a close packed structure, if the subsequent layer (B) is shifted with regards to the first layer (A), see figure 2.1. The following layer can take position A again or a new one, C. The combination of the three possible positions results in hundreds of different theoretically possible sequences. In this thesis, the presented work is restricted to the most common ones (ABC), (ABCB) and (ABCACB). From the different stacking sequences in one direction (c-axis), different crystal modifications, labeled polytypes for SiC[70], are possible. The phenomenon of a material to exist in more than one crystal structure is known as polymorphism; for SiC the polymorphism is restricted to one dimension[71]. A common notation of all the polytypes composed of a number and a letter, was introduced by L. S. Ramsdell[5] in 1947. The number stands for the amount of double layers in the unit cell and the letter represents structural information; C - cubic, H - hexagonal and R - rhombohedral. All lattice sites are equivalent in case of 3C-SiC, meaning that the local environment of each tetrahedron is purely cubic, labeled k after the Jagodzinski notation [72]. With its zincblende structure, the lattice ˚ [73]) is determined by the edge of the cube. 2Hparameter of 3C-SiC (a = 4.3596 A SiC, which is very difficult to synthesize as stand-alone crystals[74], has a wurtzite structure and all lattice sites have a pure hexagonal environment, labeled h. All the other polytypes have both lattice sites with a quasi-cubic environment and lattice sites with a local hexagonal environment. In case of a hexagonal structure (e.g. 4H- and 6HSiC), the lattice parameter corresponds to the edge of the basal plane. The hexagonal site implies a twist (180◦ ) between adjacent bi-layers, which can be seen as turning point in figure 2.1. Depending on the stacking sequence, the fraction of hexagonal lattice sites in the unit cell, changes from theoretically 100 % in 2H, to 50 % in 4H (one 5
2.1. SIC: CRYSTAL STRUCTURE
[0001]
h
Si C
k k
[1120]
h k
k2 k1
[1100]
A B C
C B A 3C−SiC (ABC)
4H−SiC (ABCB)
6H−SiC (ABCACB)
Figure 2.1: Stacking sequence of the most common SiC polytypes.
k and one h), 33 % in 6H (k1 , k2 and h) and 0 % in 3C-SiC. Along with hexagonality, physical properties of the polytypes change, such as the size of the band gap. Goldberg et al.[75] determined following band gaps at room temperature: Eg (4H-SiC) = 3.23 eV Eg (6H-SiC) = 3.08 eV Eg (3C-SiC) = 2.36 eV Choyke and co-authors[76] found an empirical relation between the size of the band gap, Eg , of different SiC polytypes and the fraction of hexagonality. Eg increases for polytypes with high amount of quasi hexagonal lattice sites. However, the measured band 6
2.2. ELECTRICAL PROPERTIES OF SIC
gap of 2H-SiC (Eg (2H-SiC) = 3.33 eV[77]) should be even larger if following this relation. Backes et al.[78] tried to theoretically explain this relation by a one-dimensional Kronig-Penney-like model. The inequivalent sites can be resolved by photoluminescence measurements due to different ionization energies related to emissions from donors or acceptors residing on different sites[79]. It is much more difficult to detect defects located at different lattice sites by electrical measurements, such as conventional DLTS, where the energy resolution is only about 10 meV.
2.2
Electrical properties of SiC
Silicon carbide is discussed as promising material for high temperature, voltage, power and frequency applications due to the large band gap compared to Ge, Si and GaAs. Table 2.1 summarizes some important electrical parameters of different semiconducting materials. Larger band gaps imply higher thermal energy needed for intrinsic conductivity and thus lower intrinsic charge carrier concentrations, ni , are predominant. ni itself is a crucial parameter for the reverse leakage current of a diode. The intrinsic charge carrier concentration at ambient conditions, ni , is much larger for Si than for the wider band gap materials such as SiC, GaN and diamond. High power applications demand a high electric field breakdown strength, Eb , which is given for large band gap and in case of SiC, Eb is 10 times larger than the one for Si. The thermal conductivity, κ, which is the ability to transport away heat, is at least twice better for SiC than for Si and GaAs and thus less cooling is needed for devices and consequently, the devices can withstand much larger power densities and harsher environments. The charge carrier mobilities, µe and µh , characterize the movement of charge carriers in an electric field. High mobilities and high electron saturation velocities νs (maximum electron velocity at high electric fields) determine the maximal speed/frequency information and responses can have and thus are important parameters for high frequency devices[80]. Comparing the wide band gap semiconductors, diamond has the best properties. However, the crystal growth of mono-crystalline diamond is still at a research level, whereas SiC started to step into commercial applications. GaN, which has a direct band gap, has advantages in light emitting applications and with the even higher νs also for high frequency applications.
7
2.2. ELECTRICAL PROPERTIES OF SIC
Table 2.1: Electrical properties of various semiconductors. A range of reported values from different sources[81–86] is presented.
semiconductor
8
Si
GaAs
4H-SiC
GaN
Diamond
Eg (eV)
1.12
1.42-1.43
3.2-3.3
3.4-3.5
5.5-5.6
ni (cm−3 )
≈ 1010
1.8 × 106
≈ 10−7
1.9 × 10−10
1.6 × 10−27
Eb (MV/cm)
0.3-0.6
0.35-0.60
2.0-3.0
2.0-3.3
5-10
µe
(cm2 /Vs)
1200-1500
6000-8600
700-1000
900-2000
1600-1900
µh
(cm2 /Vs)
420-480
250-320
115-200
νs (cm/s)
1×107
(1.2-2.0)×107
2.0×107
(1.5-2.7)×107
2.7×107
κ (W/cm·K)
1.3-1.5
0.45-0.80
3-5
1.3-2.0
10-30
3 Point defects in the bandgap Defects are crystal imperfections of different dimensions; point defects are 0-dimensional defects, line defects, such as dislocations, disturb one dimension and planar defects, which extend over two dimensions, are caused by stacking faults. During growth, subsequent cooling and device processing, defects will be created intentionally or unintentionally in the SiC lattice. Defects can be intentionally introduced by impurity doping to increase conductivity. The introduction of deep levels, which serve as trapping centers, decreases the minority charge carrier lifetime or increases the resistivity by compensation effects. Point defects are point-like defective volumes, limited roughly to the size of a unit cell of the crystal structure, such as a substitional impurity, vacancy, interstitial or antisite. An early review on point defects in SiC was given by Schneider et al.[87] in 1993. Substitutional impurity means that an atom is replaced by another atom not belonging to the crystal lattice. Vacancies occur if one atomic lattice position is left empty in the crystal. Atoms not placed in an ordinary lattice site but in between are called interstitials. If they are of the same kind as the crystal lattice, they are called self-interstitials otherwise impurity interstitials. An interstitial atom together with a vacant lattice site is labeled F RENKEL pair. If in a compound semiconductor, such as SiC, C is taking a Si-lattice site, or Si is sitting on a C-place, the defect is called antisite. Defects including two or slightly more atoms form complexes or complex structures. A vacancy paired with an impurity atom form a vacancy-impurity complex. A split interstitial is given when two atoms take the lattice place of only one atom[88] and thus disturbing the lattice locally. In case of SiC, if two carbon interstitials are close to a carbon atom and in addition shift the carbon atom from its original position, a dumbbell self interstitial complex is formed[89]. Increasing the numbers of joining atoms, the defect can be assigned to defect clusters, such as Ci -aggregates[88]. Finally, if whole atomic planes are shifted, the defect is no longer a point defect but belongs to the class of extended defects, which are not going to be discussed in detail in this thesis. A defect is of intrinsic origin, if the defective volume is composed of the same atoms as the undisturbed crystal lattice. If foreign atoms, such as doping impurities, take part in the defect, the defect is said to be of extrinsic character. 9
3.1. CLASSIFICATION
3.1
Classification
Defects in the band gap can be classified according to their energetic properties in the gap; whether they are shallow - hydrogenic impurities or deep. Deep levels can be further characterized by their interaction strength with the bands, as traps or recombination centers. A detailed description can be found in [90].
3.1.1
Energetic properties
Defects are often divided into two groups: shallow and deep levels. Depending on the size of the band gap, a level may be regarded by its energetic location as deep in Ge or Si but may be shallow in a wide-band gap semiconductor. Shallower levels have a large interaction with one band, which is due to the large extension of its electron wave-functions. The electrons are loosely bound to the impurity and thus move similar to free electrons but with a different mass. The thermal energy at room temperature is enough to ionize a large amount of the levels, meaning emission of charge carriers to the adjacent bands; thus they are used as donor or acceptor levels. The potential of shallow levels can be approximated by a hydrogen-like C OULOMB potential screened by dielectric permeability, ε, of the host lattice (figure 3.1) and with modified effective masses, m∗ . Effective mass theory calculates very well the energetic position of the excited states but deviates from the observed ground state energies. Electrons in the ground state feel more the core environment, which is different from that of a point charge. A central cell correction or chemical shift taking the chemical identity of the impurity into account has to be included. Shallow levels are used to control the F ERMI level in the semiconductor, thus they can be intentionally introduced into the semiconductor to define the donor- and acceptor concentrations, ND and NA , respectively, i.e. the available charge carriers needed for conductivity. The shallow-level density can be measured from H ALL- or capacitance-voltage investigations, whereas the energetic location is determined by temperature-dependent H ALL- or photoluminescence measurements. Deep levels have a short steep potential (figure 3.1). Their electron wave functions are largely localized at the defect site. The spatial localization allows a de-localization in the ~k-space. The potential can be approximately described using tight-binding theory. The linear combination of atomic orbitals (LCAO) results in the bonding and antibonding states of the host and the defect molecule. Additional lattice relaxation or distortion due to impurity incorporation can occur. Often higher charge states can be 10
3.1. CLASSIFICATION
approximated by the screened C OULOMB potential. Charge carriers will be trapped and only released if enough thermal energy is available. The trap concentration is often only a fraction of the net doping concentration, unless deep levels are used for compensation of the residual shallow level concentration to obtain semi-insulating (SI) material. However, in most cases the deep level density is too small to affect the electron density in the bands. U 0
Coulombic potential
Square−well potential x
Figure 3.1: One dimensional potential of shallow and deep levels (based on figure 3.3. in [91]).
3.1.2
Interaction with the bands
Deep levels can further be classified after their primary interaction with the bands. If the electron capture rate, cn , is much larger than the hole capture, c p (cn c p ), then the defect is called electron trap; vice versa if the hole capture is larger than the electron capture, then the defect is acting as a hole trap (c p cn ). If both capture rates are almost similar (cn ≈ c p ), then the defect interacts with the same strength with both bands and is regarded as a generation-recombination center (G-R center), see figure 3.2. EC Et
EV cn
cp
>>
cn >>c p
cn ~ ~ cp
Figure 3.2: Capture and emission characteristics of traps and recombination centers.
11
3.2. DOPING
3.2
Doping
Since SiC is a compound semiconductor composed of two group IV elements, p-type doping can be achieved by atoms from group III, such as Aluminum (Al) and Boron (B), whereas n-type doping is done by introducing group V elements (Nitrogen (N) or Phosphorus (P)) into the SiC crystal lattice. The close packed stacking and the short bond length of the Si-C layers hinders efficient diffusion of impurities at low temperatures[92], thus ion implantation is the only suitable post-growth process. However, post-implantation annealing (Tanneal ≈ 1700 ◦ C ) steps are needed to activate the dopants and to cure from introduced intrinsic point defects[93]. The damage during implantation depends on the size and the weight of the implanted atom. B, which is lighter than Al causes less damage and thus diffusion is facilitated. For SiC polytypes, the dopants prefer either the C, Si or both lattice sites[94]. This phenomenon can be used during epitaxial growth by choosing certain conditions to facilitate or hinder the impurity incorporation, as described in detail by Larkin et al.[95]. The valence band edges, EV , of the different SiC polytypes are pinned to the same absolute level[96], thus hole ionization energies related to substitutional impurity defect levels are almost independent of the polytype. In contrast, large differences are observed for donor levels referred to the conduction band edges, EC , in n-type SiC of different polytypes and for different lattice sites[94]. N, which replaces C[97], introduces two shallow levels in 4H-SiC and three levels in 6H-SiC. The acceptor levels of Al, residing on a Si-lattice site[98], are almost identical for the inequivalent sites, thus only one level is formed in 4H- and two in 6H-SiC. The exact ionization energies for the shallow levels related to N, P and Al in 4H-SiC were determined by the emission from donor-acceptor pairs (see [99, 100] and references therein). Additionally, Ikeda et al.[94] found that the ionization energy for an impurity sitting on a quasi-cubic site is larger than on a hexagonal site. However, this is not true in case of P on a Si lattice site[100]. The p-type doping is more difficult; since the acceptor levels of both Al (Ea = 229 meV[101]) and the shallow B (Ea = 300 meV) are relatively deep, complete ionization occurs at rather high temperatures. Boron is assumed to occupy both C and Si- sites[97]. The shallow level is associated with the Si-site[102]. Calculations and experimental results differ for the preferred lattice site and the identification of the deep B level (Ea = 580 meV), which is believed to be a complex with an intrinsic defect (BSiVC [103], BSi SiC [104], BCCSi [105] and for a review see [106, 107]). Other possible impurity levels used for doping, such as Ga and In are discussed in detail in [106]. 12
3.3. ELECTRICALLY ACTIVE INTRINSIC DEFECTS IN SIC
3.3
Electrically active intrinsic defects in SiC
This section will describe the most important levels prior to and after electron-irradiation in 4H-SiC investigated by DLTS. Levels observed in other polytypes are described elsewhere: impurity levels ([34, 42]), irradiation induced ([40, 108, 109]) and comparison of defect levels in 4H- and 6H-SiC([54]). In as-grown n-type 4H-SiC, two dominant levels are usually observed as peaks in DLTS measurements; the Z1/2 [42] (EC − 0.68 eV) and the EH6/7[39] (EC − 1.65 eV). Growth studies of Danno et al.[110] reported that the defect concentrations of these levels depend on the C/Si ratio and on the temperature, but not on the growth rate. Carbon rich conditions and lower growth temperatures hinder the defect formation, thus the carbon vacancy (VC ) is suggested as origin for these defects. Calculations[111] on the behavior of VC and its formation energy show similar results. Earlier reports[112], assigned the Z1/2 to a nitrogen-Ci complex. However, it was shown[48, 113] that N is not involved in the formation of the Z1/2 . Electron irradiation, see section 3.3.1 and annealing investigations have been performed in order to understand the origin of the underlying defect. There are only few DLTS investigations done on as-grown p-type 4H-SiC. Danno et al.[47, 114] detected several DLTS peaks related to defects with different annealing behavior, which were denoted HK2, HK3 and HK4 ranging from EV + (0.84 ··· 1.44) eV. The mid gap level HK4 is suggested to be a C-related complex defect[47]. Storasta et al.[115, 116] observed one level, HS1 (EV + 0.35 eV) using minority charge carrier injection, which correlates with the DI defect[117] in photoluminescence measurements.
3.3.1
Irradiation induced defects investigated by DLTS
As mentioned before, irradiation is used to intentionally introduce intrinsic and extrinsic defects and to study their recovery process by annealing. Additionally, the defect concentration of the intrinsic defects observed in as-grown material will be enhanced. Irradiation doses and energy will influence the damage degree of the lattice. Electron irradiation generates mainly point defects, since electrons have a very low mass, me , compared to the target atoms, MSi or MC , in the SiC crystal (me MC < MSi ). The simple bowl collision model as described by R UTHERFORD cannot directly be applied to obtain the energy transfer process, but has to be adjusted using relativistic terms. Thus, the maximal energy, T , that the electron transfers to a C-atom with mass, MC , in case of 13
3.3. ELECTRICALLY ACTIVE INTRINSIC DEFECTS IN SIC
straight incidence is written as[118]: me E T =2 MC
E +2 me c2
(3.1)
High energies, E, are needed to generate large crystal damages, which occur, if the transmitted energy, T , is larger than the displacement energy, Td , of both types of atoms. Lower energies cause the formation of F RENKEL pairs and damages can be limited to the C-sub-lattice, in case of SiC, if Td (C) < T < Td (Si). Pioneering DLTS characterization investigations after electron irradiation were done by Hemmingsson et al.[39, 40]. After high-energy electron irradiation of 4H-SiC, four additional DLTS peaks (EH1, EH3, EH4 and EH5) related to irradiation induced defects were detected in the DLTS spectrum besides the defects known from as-grown material: Z1/2 and the EH6/7 level. EH1 and EH3, labeled S-center elsewhere[119, 120], show a similar annealing behavior. Thus they were attributed to the same defect in different charge states. The annealing takes place at rather low temperature, which implies a mobile defect, such as Ci , as origin. In p-type irradiated 4H-SiC, one level HH1[39] or HS1[115, 116] is prominent besides impurity related defects. A negative U behavior occurs when a defect, which can capture two electrons, binds the second electron stronger to it than the first one. Thus more ionization energy is needed to emit the second electron. Such a behavior was detected for the Z1/2 level in 4H-SiC[40]. The repulsive C OULOMB force is overcome by a local distortion of the lattice. The same behavior was observed for the E1/2 level in 6H-SiC[61]. Both levels have most probably the same origin. High-energy electron irradiation will affect both Si and C atoms in the SiC lattice. Calculations[121] and experiments[48, 122] have determined a critical energy value for Si-displacements (Eirr > 220 keV). Low-energy irradiation will mainly generate Crelated defects. However, during recovery processes by annealing, the mobile Ci may form complexes with Si related defects. Most defects created by high-energy irradiation mentioned above, were observed also after lower-energy irradiation and thus are likely carbon related defects. The EH6/7 DLTS peak is associated with two levels, which react differently to irradiation. For high-energy electron irradiation, a broad peak is visible[39], whereas low-energy irradiation enhanced predominantly EH7 and EH6 smears out at the low temperature side of the peak[48]. EH6 is suggested to be built up of a complex defect structure[45] and EH7 may be one charge state of the VC . In papers 5-7, defects detected after low-energy irradiation and their annealing behavior will be discussed. Minority charge carrier injection give information about the levels 14
3.4. TRANSITION METAL RELATED LEVELS IN SIC
in the lower part of the band gap. Storasta et al.[48] detected an additional trap HS2, which they assigned to a F RENKEL pair and which shows a recombination enhanced annealing process[49] (detailed description in [123, 124]). A elaborate study on process and irradiation induced defects in p-type 4H-SiC is presented by Danno et al.[47].
3.4
Transition metal related levels in SiC
Transition metals (TM) introduce shallow and deep levels in SiC, which are crucial for the charge carrier concentration and the minority charge carrier life-time. The most studied TMs in SiC are Titanium (Ti), Vanadium (V), Chromium (Cr) and Tungsten (W). Ti, Cr and V are dominant background impurities due to their presence in parts of SiC growth (both sublimation and epitaxial CVD) reactors. V attracts attention for its ability to compensate the residual nitrogen doping to obtain semi-insulating SiC[36– 38] by a deep donor level in the middle of the gap (EC − 1.59 eV). An additional level is detected at EC − 0.97 eV [32]. V is discussed having an amphoteric character, i.e. both donor and acceptor levels in the SiC band gap. Ti introduces two shallow levels (EC − 0.13 eV and EC − 0.17 eV) in the band gap of 4H-SiC, which accounts for the two inequivalent lattice sites. For detailed description and information about Cr review articles are recommended [42, 106, 125]. Deep levels related to TMs form a common reference level in semiconductor heterostructures[59], i.e. the level is aligned in such materials, labeled as the L ANGER H EINRICH rule (LH). Dalibor et al.[42] and Achtziger et al.[125] concluded that deep levels related to the investigated TMs (Ti, Cr, V, Ta and W) in 4H-, 6H- and 3C-SiC follow LH as well. Transition metal incorporation will be further discussed in in paper 1 for iron (Fe) doping and in paper 2 for W in 3C-SiC. W incorporation was previously studied in 4H- and 6H-SiC using implanted radioactive W and radio tracer DLTS[33] or unintentionally by impurity contamination during growth[126]. Two levels are observed in 4H (EC − 0.17 eV and EC − 1.43 eV) and only one in 6H-SiC(EC − 1.16 eV), which is explained by the LH rule[33].
3.5
Defect annealing
The generated point defects may recover by different processes; (I) movement to defect sinks at surfaces, dislocations or grain boundaries, (II) direct recombination with 15
3.6. METASTABILITY
its counterpart and (III) cluster formation by reaction with another defect[127]. The clusters may dissociate at higher annealing temperatures. The annealing process, i.e. the time-dependent change of the defect concentration, Nt , occurs according to: dNt = − K Ntα dt
(3.2)
The pre-factor, K, is a reaction velocity constant and α represents the order of reaction; α = 1 single defect reaction (migration or dissociation of defects) and α = 2 two defects of same concentration recombine, for further details see [128]. The pre-factor K can be written as:
K = K0 exp
∆E kB T
(3.3)
with K0 the frequency factor related to the vibration frequency of the crystal and kB is the B OLTZMANN constant. Equation 3.3 shows the temperature dependence of K. The activation energy, ∆E, of the annealing process can be determined using isochronal and isothermal annealing studies. The Z1/2 and the EH6/7 level, are those who are most thermally stable and anneal out about T ≈ 2000 ◦ C [51, 53, 54], whereas other irradiation induced defects disappear at T < 1200 ◦ C . Annealing studies were performed in paper 6.
3.6
Metastability
Deep levels, as explained in section 3.1.1 often have strong interactions with the lattice and generate local distortions. The defect may gain energy by changing its position after a charge state change. The easiest way, to explain such processes is the usage of a configuration coordinate diagram (CC-diagram), which plots the total defect energy (elastic and electronic contributions) versus the generalized coordinate, taking care of all displacement changes (defect atom and lattice distortion) compared to a reference configuration. Lattice-defect interactions will cause small lattice vibrations, which are approximated by harmonic oscillations (parabolas). The electronic, optical and thermal history may allow a defect to occur in more than one configuration. The defect then has a multistable behavior[129] and can change reversibly between the configurations. In case of only two possible configurations, the defects are referred to as bistable defects. One configuration itself can have more than one available charge state (An and An−1 in figure 3.3). The defect shows bistability if after change of charge state, a different configuration is favored than before. In 16
3.6. METASTABILITY
figure 3.3, the defect is stable in configuration B, if occupied by an electron, but favors configuration A after electron emission. The two configuration states are separated by energy barriers (shown for the occupied defect: Ea (A → B) and Ea (B → A)). The electronic, optical or thermal conditions stabilize the defect in one configuration. If the conditions change, another configuration will become more stable and thermal energy
B n +e −
Ea(B
B)
A)
An +e −
Ea(A
Total energy (electronic + elastic)
may be available to surmount the barrier.
A n−1
B n−1 generalized coordinate
Figure 3.3: Coordinate configuration diagram of a bistable defect with configurations A and B, which both exist in two charge states. Configuration B is the stable configuration, when the defect is occupied (Bn−1 ) and configuration A after electron emission (An + e−1 ); n is the charge of the defect. Reconfiguration barriers for both transitions (A → B and B → A) are indicated.
Often temperature is used to provide the necessary energy for reconfiguration, but also optical excitation can be used. The charge states can be influenced by an electric field. Depletion regions, as in S CHOTTKY diodes or in a pn-junction, are ideal for probing metastable behavior. The applied bias will change the width of the depletion region and thus the fraction of occupied and empty defect levels, as explained in detail in section 4.4. Depending on the bias and the temperature prior to DLTS measurement, the defects are stabilized in one configuration, which gives rise to one (several) peak(s) in the DLTS spectrum. If the conditions change, the configuration may be transformed to another configuration and the previously observed peak(s) disappear(s) and another peak(s) evolve(s). The overall defect concentration stays constant. The config17
3.6. METASTABILITY
uration behavior can be studied by isochronal (time constant, temperature increasing) and isothermal (temperature constant, time increasing) annealing. From the reconfiguration rates, R, the reconfiguration energies, EA (activation barriers), can be determined:
R = R0 exp
−EA kB T
(3.4)
The pre-factor, R0 , can give information about the underlying reconfiguration process, such as an atomic jump or free charge carrier emission[130]. Reconfiguration studies were used in papers 5-6 to characterize the newly detected EB-centers and the transformation between the M-center[63, 64] and the EB-centers. In paper 7, the bistable defect, labeled F-center, is analyzed and minority charge carrier injection is used to force reconfiguration. Metastable defects are crucial for the device performance, since during operation, the conditions may favor a configuration with an electrically active state in the band gap, whereas the defect may be electrically inactive in the other configuration, when no bias is applied to the device or vice versa.
18
4 Electrical measurements One requirement for electrical measurements is the formation of suitable contacts to measure the desired properties, such as the capacitance of a depletion region. From capacitance changes, information about the density and the depth distribution of fixed shallow impurities (donors and acceptors) and deep defect centers can be obtained. Deep levels can be monitored as capacitance transient response to an electrical (deep level transient spectroscopy - DLTS) or optical (minority charge carrier transient spectroscopy - MCTS) pulse (see section 4.4). The concentration of ionized impurities in a depleted region can give information about the net doping concentration, Nd − Na or Na − Nd determined by capacitance-voltage (CV) investigations (see section 4.3). Capacitive studies are complementary to free charge carrier transport measurements (H ALL), which determine the free charge carrier density, n or p, and charge carrier mobilities, µe or µh , respectively. Figure 4.1 shows schematically for which part of the band gap the different methods can be applied. Hall CV
EC E F Ed − shallow donors
+ + + + + + + + +
DLTS
ET1 ET2 ET3
Ei
deep levels
MCTS HT1 Ea − shallow acceptors
Hall CV EV
Figure 4.1: Electrical measurement techniques and their applicable energetic range in the band gap with Ei the intrinsic F ERMI level, EF the F ERMI level in a doped semiconductor, EC and EV the edges of the conduction and valence band, respectively. The arrows indicate possible electron emission and capture processes. Figure based on [131].
The shallow donor-(acceptor) levels emit their electrons (holes) to the conduction band, EC , (valence band, EV ) where they travel freely and contribute to the conductivity. The free charge carriers may be trapped by deeper levels, where they are frozen out unless enough thermal energy is provided to excite them back to the bands again. 19
4.1. DEPLETION REGION
Deeper located levels can interact with both bands easily and the free charge carriers may use these defect states as recombination path. The donors electron can also interact directly with a hole, thus the charge carriers recombine without contributing to conduction, charge carrier compensation takes place.
4.1
Depletion region
Most parts of the thesis are dealing with metal-semiconductor devices, thus the following sections and discussions will be focused on the characteristics of S CHOTTKY diodes. S CHOTTKY diodes are formed after metal deposition onto the surface of a semiconducting material. The work functions of the metal, Φm , and the semiconductor, Φs , are defined as the energies needed to promote electrons at the F ERMI level (EFm and EF , respectively) to the vacuum level, see figure 4.2.
vacuum level
χs Φs
Φm +
ΦB E Fm
Vbi + + E + + + + + + + + + C Ed EF EV
− − − − −
depleted
neutral xd
Figure 4.2: Energy diagram of a S CHOTTKY barrier (based on figure 5.1 in [132]).
At the vacuum level, the free electron possesses no kinetic energy. The electron affinity, χ, represents the energy, which is needed to remove an electron from the conduction band edge, EC , to the vacuum level. The work functions of the metal and the semiconductor are usually different, which forces a band bending with a characteristic barrier height, ΦB , to maintain F ERMI level continuity after contact formation. The barrier height is defined as the difference between the metal work function, Φm , and the electron affinity of the semiconductor (ΦB = Φm − χ), neglecting the very small bending in 20
4.2. CURRENT-VOLTAGE CHARACTERISTIC
the metal side, which is caused by image forces close to the interface and influences due to surface states[133]. During contact formation, since Φm > Φs , the electrons diffuse out of the contact region to the metal, leaving fixed ionized donors behind, which build up a positive space charge, thus causing an electric field. In the metal, a corresponding negative space charge is formed. Thermal equilibrium will be recovered, if the diffusion current equals the drift current caused by the built-in field. A depletion region emptied of free carries is formed. The depletion width, xd , is defined as distance from the interface to normal bulk behavior without band bending and electric field[132]. The band bending introduces a voltage drop (qVbi ), frequently labeled built-in voltage, across the device, which is determined by the difference of the barrier height and (EC − EF ). Due to the existence of the barrier, the current flow shows a rectifying characteristics, i.e. blocked in the reverse bias direction and allowed in the forward bias case, as explained more in detail in the following section 4.2. If the semiconductor is heavily doped, the depletion region will be thin enough to allow free charge carrier tunneling, labeled field emission, and the contact behaves almost resistor-like; an ohmic contact is formed.
4.2
Current-voltage characteristic
Current-voltage (IV) measurements were used to judge the quality of the fabricated S CHOTTKY diodes. By forward biasing the diode, the internal electric field will decrease and majority charge carrier will diffuse into the depletion region. Reaching flat band conditions, i.e. compensation of the band bending, and even higher voltages, the current increases exponentially until the current starts to be limited by the series resistivity of the diode. In the reverse direction, the applied voltage will increase the depletion width, thus the current flow is blocked. The forward characteristics of the diodes may deviate from an ideal exponential behavior due to series resistance resulting from bad contacts (inter facial layer) or undepleted material[132]. The reverse breakdown voltage is the voltage, for which the built-up electric field will cause the generation of charge carriers in the semiconductor conduction band either by tunneling from the metal or by ionization of the host atoms in the depletion region. The electrons will be accelerated in the field and knock out more electrons (impact ionization). The device heats up and thus even more electrons are generated until the diode breaks catastrophically. The reverse leakage current is important for capacitance measurements and should as a rule of thumb not exceed Ileakage < 10 µA[132]. For moderately doped semiconductors, 21
4.2. CURRENT-VOLTAGE CHARACTERISTIC
thermionic emission, i.e. thermal activation of charge carriers, which causes a current flow surmounting the barrier, results in the following relation between current, I, voltage, V , and barrier height, ΦB ,: I = A A∗ T 2 exp
−q ΦB kB T
qV exp −1 n kB T
(4.1) (4.2)
where A is the area of the diode, n its ideality factor and A∗ the R ICHARDSON constant, which depends on the effective mass of the semiconductor. A definition of A∗ and
rs > 0
ln
I
1 − exp (− qV/k B T )
experimental values for some semiconductors can be found in [134].
slope ~
q n kB T
Is
Vf >> kBT/e
V
Figure 4.3: Schematic IV characteristic of a S CHOTTKY diode in forward bias direction.
ΦB and n of the diode can be determined from the semi-logarithmic IV -plot (figure 4.3). ΦB is proportional to IS (saturation current), which represents the intercept with the y-axis: ΦB =
∗ 2 AA T kB T ln q IS
(4.3)
n can be obtained from the slope of the linear part of the forward characteristics. For an ideal diode n = 1. Additional series resistances, which may be caused by the formation of an insulating oxide layer during or prior to metal deposition, increase n. For a homogenous material, ΦB calculated from IV and capacitance voltage (CV) measurements, as explained in the following section 4.3 should be similar. In paper 3, detailed IV analysis of diodes manufactured on n-type 3C-SiC is discussed. 22
4.3. CAPACITANCE-VOLTAGE MEASUREMENTS
4.3
Capacitance-voltage measurements
The depletion layer capacitance, C, i.e. the available charge Q per applied voltage, V , of a diode is measured by the variation of the depletion width with the applied bias. The charge density, ρ, in the depletion region can be determined solving the P OISSON equation for the electrostatic potential, Φ,: d2Φ 1 =− ρ(x) dx2 ε ε0
(4.4)
where x is the distance from the contact interface, ε the dielectric constant and ε0 the relative permittivity. Figure 4.4 illustrates schematically the charge density of a S CHOTTKY
diode. The absolute value of the space charge built up in the semiconductor QS
equals the absolute value of charge close to the contact interface in the metal, Qm . metal
semiconductor
ρ
abrupt
QS
real
xd
x
Qm
Φ
xd
x
Figure 4.4: Schematic of the charge density profile, ρ(x), and the electrostatic potential, Φ(x), across a S CHOTTKY diode.
Using the boundary conditions (ρ = 0 in a bulk semiconductor (x xd ) and in the metal (x < 0) as well as ρ 6= 0 close to the contact interface; 0 < x < xd ) and assuming a homogeneous n-type doped semiconductor (ρ = q Nd ), the total voltage over the diode, V , is: 1 q Nd 2 xd 2 εε s 0 2 ε ε0 = V q Nd
V =
(4.5)
xd
(4.6) 23
4.3. CAPACITANCE-VOLTAGE MEASUREMENTS
The total voltage, V , is the sum of Vbi and the applied bias, Va . The depletion layer capacitance can thus be calculated using equation 4.6 and dQ = q Nd A xd : r dQ dQ dxd ε ε0 A ε ε0 q Nd =A C= = = dV dxd dV xd 2V
(4.7)
The built-in potential, Vbi , and the doping concentration, Nd , can be obtained from a 1/C2 versus V plot as interception with the x-axis and from the slope, as indicated in
−2
−2
C (pF )
figure 4.5.
slope ~
−Vbi
0
2 A 2 ε ε0 q Nd Vr (V)
Figure 4.5: Schematic CV dependence of a S CHOTTKY diode.
The S CHOTTKY barrier height, ΦB , can be determined from Vbi : kB T NC kB T − ∆Φ ≈ Vbi + ln ΦB = Vbi + (EC − EF ) + q q ND where the potential due to image forces, ∆Φ, is neglected as well as
(4.8) kB T q .
NC is the
effective density of states in the conduction band. The dependencies are valid under certain conditions, which are summarized as the depletion approximation: • the depletion region is completely free of free charge carriers (Va kB T /q ≈ 0) (neglected D EBYE tail of electrons close to depletion edge) • abrupt transition between the depleted and the neutral semiconductor at xd • the concentration of deep levels, Nt Nd For a compensated material, Nd is replaced by the net doping concentration, Nd − Na . The depletion capacitance is measured by increasing a DC voltage superimposed by a small AC (10 kHz - 1 MHz) voltage signal (VAC ≈ 100 mV), which induces changes of the free charge carrier density by small fluctuations at the depletion edge, as schematically shown in figure 4.6. 24
metal
4.4. TRANSIENT SPECTROSCOPY
Nd
W Vr
dW dVr
Figure 4.6: Schematic measurement circuit for CV investigations of a S CHOTTKY diode.
4.4
Transient spectroscopy
The deep impurity level concentration is often much less than the shallow impurity density. Transient techniques are applied to study capture and emission processes of deep levels in a depletion region, where shallow impurities are ionized. The relaxation from / response to a voltage or optical pulse, which disturbs the thermal equilibrium is studied. In a first part, possible capture and emission processes are discussed using S HOCKLEY-R EAD -H ALL statistics[135, 136] and afterwards applied to deep levels in a depletion region for a change of capacitance, ∆C(t).
4.4.1
Rate equations
The concentrations of electrons, n, and holes, p, in the conduction and valence band, respectively, are determined by the shallow impurity doping. In figure 4.7, the possible emission and capture processes in the semiconductors band gap are shown: (1) electron capture to the empty trap (pT ) from the conduction band (EC ) with the electron capture rate cn (2) electron emission from the occupied trap level (nT ) back to EC with the electron emission rate en (3) hole capture to the occupied trap level from the valence band (EV ) with the hole capture rate c p (4) hole emission from the emptied level to EV ; corresponding to electron capture from EV with the hole emission rate e p The combined processes (1)+(3) and (2)+(4) represent interactions with both EC and EV and are known as recombination and generation of charge carriers, respectively. In 25
4.4. TRANSIENT SPECTROSCOPY
(1)
(3)
(2)
(4) n
cn
EC
en
+ pT
+
ET
nT ep
cp
+
p
EV
Figure 4.7: Possible electron and hole capture or emission processes for one trap level located at ET in the semiconductors band gap, where nT is the concentration of the occupied trap and pT the concentration of the unoccupied trap.
contrast, the combination of (1)+(2) and (3)+(4) are typical trapping processes, i.e. interaction either with EC or EV . Depending on the F ERMI-level, temperature, capture cross section and the energetic location of the trap in the band gap, the capture and emission rates will vary, as explained in section 3.1.2.
4.4.2
Capture processes
The charge carrier recombination can occur directly from the conduction band to the valence band by photon emission or indirectly by radiative or non-radiative paths via defects. SiC has an indirect band gap thus recombination via defects is much more probable than direct band-to-band recombination. In this section, non-radiative capture and recombination of charge carriers will be discussed. Possible non-radiative capture processes are cascade phonon capture, multi-phonon emission assisted capture (MPE) and capture due to AUGER recombination. The three processes show different temperature dependencies[137] and can thus be distinguished from each other.
Cascade capture The cascade phonon capture is the predominant capture process for shallow impurities with hydrogen-like potential or ionized deep levels with excited states close to the bands. The free charge carrier is attracted by the screened C OULOMB potential of the 26
4.4. TRANSIENT SPECTROSCOPY
impurity. The charge carrier relaxes through the excited states down to the ground state under emission of single phonons with phonon energies according to the energy separation between the states (e.g. h¯ Ω1 and h¯ Ω2 ), as shown in figure 4.8. Re-emission from excited states to the band depends strongly on the temperature and can occur due to more accessible thermal energy (kB T ) at higher temperatures. For deeper located energy levels the separation between the ground and the excited states will be increased and more than one phonon is needed to overcome it (e.g. h¯ Ω3 and h¯ Ω4 ). A multiparticle process, which is less probable, is needed. The cascade phonon capture shows a T −x (x = -3 ··· -1) dependence[138] controlled by the involved phonons (acoustic or
energy
optical).
hΩ 1
hΩ 2
hΩ 3 + h Ω 4
x Figure 4.8: Schematic drawing of a cascade phonon capture (based on figure 5.2. in [91]).
Multi-phonon emission assisted capture As described in section 3.1.1, deep levels are localized in the real lattice, but delocalized in ~k-space, thus interactions with phonons of different phonon energies are possible[139]. The free electron (hole) states in the conduction (valence) band are localized in ~k-space at Eex (Q) (E0 (Q)) due to their delocalization in the real space. For a free electron capture process, the lattice starts to vibrate about its equilibrium position, Eex (Q), (purple arrow in figure 4.9) at elevated temperatures and if enough thermal energy is available (activation barrier Ebn ), the delocalized conduction band state of the free electron (Eex (Q)) will cross with the localized defect state E• (Q) and thus the free electron will be trapped by the defect. The filled state, E• (Q), relaxes down to its 27
4.4. TRANSIENT SPECTROSCOPY
equilibrium position at Q− displaced from Q0 under emission of multiple phonons (red arrow). The recombination can be completed by a hole capture (electron release to valence band), if the valence band interacts in the same way with the defect. The charge carrier capture cross section increases exponentially with temperature for MPE: σ (T ) = σ∞ exp − kEBσT . The required activation energies to overcome the
barriers for electron (Ebn ) and hole capture (Ebp ) are indicated in the figure 4.9. At higher temperatures more phonons are available, which enhance the capture of electrons or
total energy (lattice + defect)
holes, respectively.
Eex(Q) E (Q) n Eb p
Eb
Eg E 0 (Q) Q0
Q − generalized coordinate
Figure 4.9: CC-diagram for MPE capture (based on figure 5 in [137]).
AUGER recombination For an AUGER recombination, the capture cross section decreases weakly with temperature, whereas a strong dependence on the charge carrier concentration is observed[137]. AUGER recombination is the least probable capture process in the characterized unintentionally or low-doped SiC material, since the interaction of three particles (electronhole recombination excites a third particle) requires large charge carrier densities[139].
4.4.3
Emission process
The thermal emission can be enhanced by lowering the barrier by δ E due to the applied electric field, see figure 4.10. Observable effects are detected for electric fields exceeding 105 V/cm[140]. The energy barrier lowering is referred as P OOL -F RENKEL emission 28
4.4. TRANSIENT SPECTROSCOPY
and explained in detail in [91, 141]. If additional, the electron is excited by phonons and tunnels through the barrier (phonon - assisted tunneling), than even less energy is needed for electron emission from the deep level. In n-type material, donor-like states will be neutral after electron capture and positively charged after electron emission. Acceptor-like states will in contrast be charged if occupied by an electron and be neutral after electron emission. Thus the enhancement of the charge carrier emission by the applied electric field will only affect donor-like states. Lef`evre et al.[142] developed a technique, labeled double correlated DLTS (DDLTS), to distinguish donor-like from acceptor-like behavior using different electric field strength.
(a)
δE
EC (b) hΩ
ET Figure 4.10: Electric field enhanced emission (a) and phonon assisted tunneling (b) from the deep level, ET , to the conduction band, EC (based on figure 5.6 in [134]).
4.4.4
Time dependent trap concentration
The change in trap occupancy due to emission and capture processes is time-dependent: dnT = (cn + e p )(NT − nT ) − (en + c p ) nT dt
(4.9)
whereas NT is the total defect concentration, NT − nT is the concentration of empty (pT ) and nT of filled traps. In thermal equilibrium, the principle of detailed balance, i.e. capture equals emission process related to electrons (en nT = cn (NT − nT )) and related to holes (c p nT = e p (NT − nT )) must be fulfilled at the same time, i.e. charge carrier generation by band-to-band interactions is forbidden. The trap occupancy in thermal equilibrium can thus be written 29
4.4. TRANSIENT SPECTROSCOPY
as:
ep cn nT = = NT cn + en e p + c p
The trap occupancy can also be derived from F ERMI -D IRAC statistics: −1 nT g0 ET − EF = 1 + exp NT g1 kB T
(4.10)
(4.11)
The ratio g0 /g1 represents the degeneracy between empty and filled trap. More detailed information can be found in [132] chapter 7. Combining equation 4.10 and 4.11, the emission rate, en , can be obtained: en = cn
g0 ET − EF exp g1 kB T
(4.12)
The electron capture rate, cn , can be expressed by: cn = σn hvn i n
(4.13)
where σn is the charge carrier capture cross section, hvn i the mean thermal velocity and n the density of electrons in the conduction band. n is determined by: EC − EF n = NC exp − kB T
(4.14)
with NC the effective density of states in the conduction band. Thus the temperature dependent emission rate, en (T ), can be rewritten using equation 4.13 and 4.14: g0 EC − ET (4.15) en (T ) = σn hvn i NC exp − g1 kB T hvn i and NC are temperature dependent: 1 3 kB T 2 m∗n 3 2 π m∗n kB T 2 = 2 MC h2
hvn i = NC
(4.16) (4.17)
where MC is the number of equivalent conduction band minima of the semiconductor, h the P LANCK constant and m∗n the effective electron mass. Putting all temperature √ 32 ∗ 2 mn kB MC , equation 4.15 can be rearindependent variables together in γ = 2 3 2hπ2 ranged: g0 EC − ET 2 en (T ) = σn γ T exp − g1 kB T 30
(4.18)
4.4. TRANSIENT SPECTROSCOPY
In a semi-logarithmic plot ln
en (T ) T2
versus
1 T,
the trap energy, ET , can be obtained
from the slope and the charge carrier capture cross section, σn , from the intercept with the y-axis. EC − ET and σn are affected if the capture process is strongly temperature dependent, as explained in section 4.4.2. Thermodynamic approaches (a detailed description can be found in [132, 134] and references therein) result in the dependence of the G IBBS free energy, ∆G(T ), on the thermal emission rate, en (T ),: ∆G(T ) en (T ) = σn (T ) hvn i NC exp − kB T
(4.19)
where ∆G(T ) is given by the enthalpy ∆H and the entropy ∆S: ∆G(T ) = ∆H − T ∆S and thus
∆H(T ) en (T ) = χn σn (T ) hvn i NC exp − (4.20) kB T The entropy factor, χn = exp ∆S kB , includes changes in entropy, ∆S, during charge carrier emission. If the capture cross section is temperature dependent, an additional term σn (T ) = σ∞ exp − kEBσT has to be taken into account: ∆H(T ) + Eσ en (T ) = χn σ∞ (T ) hvn i NC exp − kB T
(4.21)
The activation energy for charge carrier emission from the trap is a sum of the enthalpy change, ∆H = EC − ET , and the energy barrier for charge carrier capture, Eσ . Direct capture cross section measurements can be used to determine the temperature-dependent capture cross section, σn (T ), and the thermal barriers, Eσ , see investigations in paper 1. The temperature dependent concentration of the filled traps, nT (T ), can be determined solving equation 4.10 using the steady state occupancy, nT (∞),: cn + e p dnt NT for cn + e p + c p + en dt t=∞ t nT (t) = nT (∞) − (nT (∞) − nT (0)) exp − τ
nT (∞) =
(4.22) (4.23)
nT relaxes back to thermal equilibrium after a perturbation at t = 0 with a time con1 , which depends on all the four reaction rates. By using appropriate stant, τ = cn +e p +c p +en
measurement conditions in a space charge region (n-type and ND NT ) the calculations can be much simplified and the time constant depends only on en or cn , if the reverse bias is applied or zero, respectively. 31
4.4. TRANSIENT SPECTROSCOPY
4.4.5
DLTS
Deep level transient spectroscopy (DLTS) is used to gain information about the electrical signature of defects, However, defect identification needs additional characterization techniques, such as electron spin resonance. Conventional DLTS, which is applied mainly for the electrical characterization of SiC in this thesis, uses the described perturbation of the equilibrium electron concentration (section 4.4.1) to obtain information about deep levels in the band gap (ET , σn and the defect concentration, NT ). The relaxation to thermal equilibrium conditions is measured as capacitance transient response to a voltage pulse. The principle is shown in figure 4.11. If no bias (Vr = 0 V) is applied to the S CHOTTKY diode, the depletion region is small, as explained in section 4.1 and the deep levels located below the F ERMI level are filled with electrons. In this situation, the capture rate, cn , is larger than the emission rate, en . Increasing the reverse bias (Vr = −10 V), increases the band bending and results in a widened depletion region (x = 0 ··· xd ). The deep levels will now be located above EF and thus en cn and the electrons will be emitted to the conduction band and rapidly flow out from the depletion region due to the applied electric field. Close to the edge of the depletion region (x0 for Vr = 0 and xd for Vr = -10 V), there exist a free charge carrier tail, which extends a distance λ into the space charge region. In this transition region, the traps stay filled because cn > en . The distance λ is determined by the shallow doping concentration, Nd , and the energetic position of the trap, ET ; s λ=
2 ε ε0 (EF − ET ) q2 Nd
(4.24)
Thus the transition region becomes less important for high electric fields, where xd λ and for deep levels close to the band edges. The thermally dependent concentration of the filled traps in the depletion region for Vr = 0 is given by equation 4.23. The voltage pulse Vr = −10 V will cause the traps to emit their charge carriers and thus the charge density in the depletion region changes, which implies a change in the width of the depletion region. The process continues until all traps are emptied (nT (∞) = 0). The change in depletion width will further cause a change in capacitance, as discussed in section 4.3 under the condition NT Nd . In the depletion region, the time constant will be determined by the electron emission from electron traps: nT = NT exp(−en t), since nT (∞) = 0 in equation 4.23. Thus the trap 32
4.4. TRANSIENT SPECTROSCOPY
cn >> en
electron energy
EC
metal
ET
EV
λ
x2 electron energy
EF
x0
Vr = 0 V
x
en >> cn eVr
en
EC
metal
ET
EV
λ
x1
EF
xd
x
Vr = −10 V Figure 4.11: Band diagram of a
SCHOTTKY
contact on n-type material (based on figure 7.6 in
[132]). Occupancy of one deep level under different reverse bias conditions is indicated. The shallow donor levels are not shown.
concentration can be determined from the capacitance change[132]: ∆C(t) = C(t) −C(∞) = −∆C0 exp(−en t) ∆C0 1 x12 − x22 nT = C 2 Nd xd2
(4.25) (4.26)
where ∆C0 = C(∞) −C(0), see figure 4.12, x1 = xd − λ and x1 = x0 − λ , see figure 4.11. If the transition region is neglected (λ xd ), then equation 4.26 simplifies to: ∆C0 1 nT = C 2 Nd
(4.27)
Capacitance transients (C(t) = C(∞) + ∆C0 exp(−en t)) are measured at different temperatures, which change the emission rate. If the capacitance change is analyzed by 33
4.4. TRANSIENT SPECTROSCOPY
C t C(
)
∆C( t )
∆C0
C( 0 )
Figure 4.12: Diode capacitance transient ∆C(t) compared to steady state capacitance C(∞).
multiplication with a weighting function (
R en (T ) 0
w(t) dt = 0), the emission from a defect
level will be maximized for a certain rate window at a certain temperature and thus a peak will occur in the spectrum (DLTS signal versus temperature). Changing the rate window, τre f , the emission peak occurs at a different temperature. In case of a double boxcar correlation function (∆C = C(t2 ) −C(t1 ))[143], the DLTS signal, S, is: S ∝ [exp(−t1 /τ) − exp(−t2 /τ)]
(4.28)
and the emission takes place at: τ(T ) = τre f =
t2 − t1 ln tt21
(4.29)
Deep levels, in n-type materials located above the middle of the band gap in the case SiC, can be detected by a temperature scan from 85 to 700 K. In this thesis a standard double boxcar technique, a lock-in amplification simulator and a three-point-correlation window[144] are used. A detailed view on each weighting function is given in [91].
4.4.6
MCTS
To study the response of minority charge carriers, one can use the advantage of an asymmetrical doped p+ n- junction by injecting minority charge carriers by forward biasing the diode. However, much effort lies on the fabrication of the mesa-structures. Normal S CHOTTKY diodes are easily manufactured, but allow only majority charge carrier spectroscopy. If the front-side S CHOTTKY contact has the ability to transmit light, then photons can be used to excite electrons from the valence band to the conduction band even in the depletion region. Brunwin et al.[145] developed such a technique, known as minority charge carrier transient spectroscopy (MCTS). 34
4.4. TRANSIENT SPECTROSCOPY
Minority charge carriers are injected in the depletion region by illumination with light exceeding the band gap, h ν Eg , and thus generating free electrons and holes in the corresponding bands. The generation of electron-hole pairs, G, can be described by [132]: (4.30)
G = α T Φ0 exp(−αx)
where T accounts for the transparency of the contact, Φ0 for the photon flux and α for the absorption coefficient. The illumination should be efficient enough to reach a steady state between hole emission and capture. The minority charge carrier concentration is larger than the majority charge carrier concentrations in the depletion region, since the generated electrons will drift out and holes are pulled into the depleted region due to the applied reverse bias, see figure 4.13. A hole diffusion current towards the contact interface is built up and should exceed the electron drift current for efficient hole trapping. This requirement in fulfilled if the hole diffusion length, L p , and the absorption distance, α −1 , exceed the depletion width, xd , (L p > α −1 xd ). electron energy
Vr = −10 V
drift
h ν > Eg
metal
eVr
cn
EC EF
+
ET
cp+ drift
+
+
xd diffusion
EV
Figure 4.13: Band diagram of a semi-transparent S CHOTTKY contact on a n-type with a minority trap level, ET (based on figure 10.31 in [132]). Electron-hole pairs are generated by illumination energy exceeding the band gap. Diffusion and drift currents of holes as well as electron drift current due to the applied electric field are indicated.
The concentration of the trapped holes depends on both hole capture and hole emission rates, as derived in [132]. Thus care has to be taken in analyzing the trap concentration, since the capacitance change depends on both rates.
35
5 Summary of papers Paper I
Deep levels were detected in Fe-doped n- and p-type 4H-SiC using deep
level transient spectroscopy (DLTS). The Fe1 level (EC − 0.39 eV) in n-type material is very likely Fe related. DLTS spectra of p-type 4H-SiC show two dominant peaks (EV + 0.98 eV and EV + 1.46 eV). The temperature dependence of the capture cross sections indicates a multi-phonon capturing process. SIMS analysis proved the incorporation of Fe in both n- and p-type epilayers. Fe is also suggested as origin to the two levels observed in p-type 4H-SiC. Comparison to Fe-related defects detected in Si indicates that the observed DLTS peaks are probably related to Fe and Fe-B pairs. Paper II
Tungsten was incorporated in SiC and W related defects were investigated
using DLTS. In agreement with literature, two levels related to W were detected in 4HSiC (EC − 0.18 eV and EC − 1.40 eV), whereas only the deeper level, W 2, was observed in 6H-SiC(EC − 1.15 eV) and for the first time also in 3C-SiC (EC − 0.47 eV). The activation energy of W 2 in 3C-SiC agrees with the predicted ionization energy and aligns with the W 2 levels in 4H- and 6H-SiC in agreement with the Langer-Heinrich rule. Tungsten can thus be regarded as a common reference level in SiC. Additionally, the alignment of intrinsic defect levels was observed, which suggests that the E1 level (EC − 0.57 eV) in 3C-SiC is related to EH6/7 in 4H-SiC and E7 in 6H-SiC, respectively, and thus to the carbon vacancy. Paper III
3C-SiC grown hetero-epitaxially on 4H- or 6H-SiC using a standard or
a chloride-based CVD process was electrically characterized using IV, CV and DLTS. Ni and Au Schottky contacts were deposited directly after a HF dip or after an additional UV illumination step, respectively, which is known to saturate surface defects by local oxidation. Therefore, the reverse leakage currents of the Au-Schottky diodes were reduced to lower than 10−8 A at a reverse bias of -2 V . The Schottky barrier heights of the Ni and Au contacts were determined by IV measurement to be φB = 0.575 eV and φB = 0.593 eV, respectively. One dominant DLTS peak (EC − 0.60 eV) was observed in the 3C-epilayers independently of the substrate. This electron trap is supposed to be an intrinsic defect, like the Z1/2 or the EH6/7 in 4H-SiC layers. However, a direct correlation between the peak detected in 3C-SiC and Z1/2 or the EH6/7 in 4H-SiC cannot be concluded from our results. 37
Paper IV The intrinsic defects Z1/2 and EH6/7 dominate the DLTS spectrum of n-type 4H-SiC epilayers, grown by chloride-based CVD growth process. For growth rates exceeding 100 µm/h an additional peak occurred in the DLTS spectra. In MCTS measurements, the shallow and the deep boron complexes as well as the HS1 defect were observed. With increasing C/Si ratio the intrinsic defect concentrations decreased until C/Si = 1. The influence of the Cl/Si ratio on the trap concentrations was less pronounced. The PL spectra were completely dominated by the near band gap (NBG) emission, which confirmed high quality epilayers. Only the PL line related to the D1 center, which corresponds to the HS1 defect in the MCTS spectra, was weakly observed. The addition of chlorine in the growth process gives the advantage of high growth rates without the introduction of additional defects for moderate growth rates.
Paper V Epitaxial n-type 4H-SiC layers were irradiated at room temperature by low-energy electrons (200 keV). The irradiation induced defects EH1 and EH3 annihilate at about at 650 K. Simultaneously, three new bistable centers, labeled EB-centers, were detected in the DLTS spectrum. Controlling the bias and temperature conditions, the defects can be repeatedly stabilized in either configuration I or II. The reconfigurations of the EB-centers (I → II and II → I) take place at room temperature with a thermal reconfiguration energy of about 0.95 eV. The origin of the EB-centers is suggested to be related to carbon complex defects, since the defects were formed after low-energy electron irradiation below the energy threshold for displacing a Si atom.
Paper VI The bistable M-center, previously reported to be introduced after highenergy proton implantation, is detected in the DLTS spectrum of low-energy electron irradiated epitaxial n-type 4H-SiC. The annealing behavior of the M-center is confirmed and an enhanced recombination process is suggested. The annihilation process is coincidental with the evolution of the bistable EB-centers in the low temperature range of the DLTS spectrum. The annealing energy of the M-center is similar to the generation energy of the EB-centers. However, the defect concentrations are different, thus a partial transformation of the M-center to the EB-centers is suggested. The EB-centers themselves annealed out at about 700 ◦ C. The M-center, EH1 and EH3, as well as the EB-centers are present after low-energy electron irradiation and their annihilation temperatures are relatively low, therefore the attributed defects are related to the C atom in SiC, most probably to carbon interstitials, Ci , and their complexes. 38
Paper VII
A bistable defect, labeled F-center, with two configurations (A and B) is
observed after high- and low-energy electron irradiation of n-type 4H-SiC. In configuration A, the EH5 peak (Ea = EC − 1.07 eV) is detected in the DLTS spectrum. In the accessible temperature range of the DLTS setup, no peak is detected for configuration B. Isochronal investigations revealed the transition temperatures to be TA→B > 730 K and for the opposite process TB→A ≈ 710 K. The energy needed to conduct the transformations were determined to EA (A → B) = (2.1 ± 0.1) eV and EA (B → A) = (2.3 ± 0.1) eV, respectively. The pre-factor indicated that the transformation A → B is an atomic jump process and the opposite transition B → A is dominated by a charge carrier emission. Minority charge carrier injection enhanced the transformation B → A, since configuration B is unstable when occupied by a hole. Since the bistable F-center is present after low-energy electron irradiation, its origin is likely related to carbon.
39
Bibliography [1] W. Shockley, Introduction to the first Silicon Carbide Conference held in Boston, April 2-3, 1958. ¨ ber Flussspaths¨ [2] J. J. Berzelius. Untersuchung u aure und deren merkw¨ urdigen Verbindungen. Annalen der Physik und Chemie, 1:169–230, 1824. [3] E. G. Acheson, U.S. patent no. 17911, 1892. [4] B. W. Frazier. Report of an examination of crystals furnished by Mr. E. G. Acheson, president of the carborundum company. J. Franklin Inst., 1893. p. 287. [5] L. S. Ramsdell. Studies on silicon carbide. Am. Mineral., 32:64–82, 1947. [6] F. H. Moissan. Nouvelles recherches sur la m´et´eorite de Ca˜ non Diablo. Comptes rendus, 139:773– 786, 1904. [7] H. H. C. Dunwoody, U.S. patent no. 837,616, 1906. [8] G. W. Pierce. Crystal Rectifiers for Electric Currents and Electric Oscillations. Part I. Carborundum. Phys. Rev. (Series I), 25:31–60, 1907. [9] H. J. Round. A Note on Carborundum. Electric. World, 49:309, 1907. [10] O. V. Losev. Luminous carborundum detector and detection with crystals. Telegrafiya i Telefoniya bez Provodov, 44:485–494, 1927. [11] N. Zheludev. The life and times of the LED — a 100-year history. Nature Photon., 1:189–192, 2007. [12] K. Lehovec. Energy of Trapped Electrons in Ionic Solids. Phys. Rev., 92:253–258, 1953. [13] J. A. Lely. Darstellung von Einkristallen von Silicium Carbid und Beherrschung von Art und Menge der eingebauten Verunreinigungen. Berichte der Deutschen Keramischen Gesellschaft, 32:229–236, 1955. [14] Yu. M. Tairov, V. F. Tsvetkov, and I. I. Khlebnikov. Growth of silicon carbide crystals by vapourliquid-solid (VLS) mechanism in the sublimation method. J. Cryst. Growth, 20:155 – 157, 1973. [15] Yu. M. Tairov and V. F. Tsvetkov. Investigation of growth processes of ingots of silicon carbide single crystals. J. Cryst. Growth, 43:209 – 212, 1978. [16] Yu. M. Tairov and V.F. Tsvetkov. General principles of growing large-size single crystals of various silicon carbide polytypes. J. Cryst. Growth, 52:146 – 150, 1981. [17] http://www.fundinguniverse.com/company-histories/Cree-Inc-Company-History.html, 2011.
41
BIBLIOGRAPHY
[18] H. Matsunami, S. Nishino, and H. Ono. IVA-8 heteroepitaxial growth of cubic silicon carbide on foreign substrates. IEEE T. Electron Dev., 28:1235 – 1236, 1981. [19] S. Nishino, J. A. Powell, and H. A. Will. Production of large-area single-crystal wafers of cubic SiC for semiconductor devices. Appl. Phys. Lett., 42:460–462, 1983. [20] N. Kuroda, K. Shibahara W. Yoo, S. Nishino, and H. Matsunami. Step controlled VPE growth of SiC single crystals at low temperatures. In Conference on Solid State Devices and Materials, pages 227–230, 1987. [21] T. Kimoto and H. Matsunami. Surface kinetics of adatoms in vapor phase epitaxial growth of SiC on 6H-SiC{0001} vicinal surfaces. J. Appl. Phys., 75:850–859, 1994. [22] O. Kordina, A. Henry, J. P. Bergman, N. T. Son, W. M. Chen, C. Hallin, and E. Janz´en. High quality 4H-SiC epitaxial layers grown by chemical vapor deposition. Appl. Phys. Lett., 66:1373–1375, 1995. [23] H. Pedersen. Chloride-based Silicon Carbide CVD. Link¨ oping studies in science and technology, dissertation no. 1225, Link¨ oping University, 2008. [24] S. Leone. Advances in SiC growth using chloride-based CVD. Link¨ oping studies in science and technology, dissertation no. 1340, Link¨ oping University, 2010. [25] O. Kordina, C. Hallin, A. Ellison, A. S. Bakin, I. G. Ivanov, A. Henry, R. Yakimova, M. Touminen, A. Vehanen, and E. Janz´en. High temperature chemical vapor deposition of SiC. Appl. Phys. Lett., 69:1456–1458, 1996. [26] P. Zhou, M. G. Spencer, G. L. Harris, and K. Fekade. Observation of deep levels in cubic silicon carbide. Appl. Phys. Lett., 50:1384–1385, 1987. [27] S. Leone, F. C. Beyer, A. Henry, O. Kordina, and E. Janz´en. Chloride-based CVD of 3C-SiC epitaxial layers on 6H(0001) SiC. Phys. Status Solidi Rap. Res. Lett., 4:305–307, 2010. [28] T. Kimoto, A. Itoh, H. Matsunami, T. Nakata, and M. Watanabe. The effects of N+ dose in implantation into 6H-SiC epilayers. J. Electron. Mater., 24:235–240, 1995. [29] T. Troffer, C. Pepperm¨ uller, G. Pensl, K. Rottner, and A. Sch¨ oner. Phosphorus-related donors in 6H-SiC generated by ion implantation. J. Appl. Phys., 80:3739–3743, 1996. [30] T. Kimoto, A. Itoh, H. Matsunami, T. Nakata, and M. Watanabe. Aluminum and boron ion implantations into 6H-SiC epilayers. J. Electron. Mater., 25:879–884, 1996. [31] T. Troffer, M. Schadt, T. Frank, H. Itoh, G. Pensl, J. Heindl, H. P. Strunk, and M. Maier. Doping of SiC by Implantation of Boron and Aluminum. Phys. Stat. Solidi A, 162:277–298, 1997. [32] N. Achtziger and W. Witthuhn. Band gap states of Ti, V, and Cr in 4H–silicon carbide. Appl. Phys. Lett., 71:110–112, 1997.
42
BIBLIOGRAPHY
[33] N. Achtziger, G. Pasold, R. Sielemann, C. H¨ ulsen, J. Grillenberger, and W. Witthuhn. Tungsten in silicon carbide: Band-gap states and their polytype dependence. Phys. Rev. B, 62:12888–12895, 2000. [34] N. Achtziger, D. Forkel Wirth, J. Grillenberger, T. Licht, and W. Witthuhn. Identification of deep bandgap states in 4H- and 6H-SIC by radio-tracer DLTS and PAC-spectroscopy. Nucl. Instrum. Meth. Phys. Res. B, 136-138:756–762, 1998. [35] N. Achtziger, J. Grillenberger, and W. Witthuhn. Radiotracer spectroscopy of deep levels in the semiconductor band gap. Hyperfine Interact., 120-121:69–79, 1999. [36] H. McD. Hobgood, R. C. Glass, G. Augustine, R. H. Hopkins, J. Jenny, M. Skowronski, W. C. Mitchel, and M. Roth. Semi-insulating 6H-SiC grown by physical vapor transport. Appl. Phys. Lett., 66:1364– 1366, 1995. [37] J. R. Jenny, J. Skowronski, W. C. Mitchel, H. M. Hobgood, R. C. Glass, G. Augustine, and R. H. Hopkins. Deep level transient spectroscopic and Hall effect investigation of the position of the vanadium acceptor level in 4H and 6H SiC. Appl. Phys. Lett., 68:1963–1965, 1996. [38] T. Kimoto, T. Nakajima, H. Matsunami, T. Nakata, and M. Inoue. Formation of semi-insulating 6H-SiC layers by vanadium ion implantations. Appl. Phys. Lett., 69:1113–1115, 1996. [39] C. Hemmingsson, N. T. Son, O. Kordina, J. P. Bergman, E. Janz´en, J. L. Lindstr¨ om, S. Savage, and N. Nordell. Deep level defects in electron-irradiated 4H SiC epitaxial layers. J. Appl. Phys., 81:6155–6159, 1997. [40] C. Hemmingsson, N. T. Son, O. Kordina, E. Janz´en, and J. L. Lindstr¨ om. Capture cross sections of electron irradiation induced defects in 6H–SiC. J. Appl. Phys., 84:704–708, 1998. [41] J. P. Doyle, M. K. Linnarsson, P. Pellegrino, N. Keskitalo, B. G. Svensson, A. Schoner, N. Nordell, and J. L. Lindstrom. Electrically active point defects in n-type 4H–SiC. J. Appl. Phys., 84:1354–1357, 1998. [42] T. Dalibor, G. Pensl, H. Matsunami, T. Kimoto, W. J. Choyke, A. Sch¨ oner, and N. Nordell. Deep Defect Centers in Silicon Carbide Monitored with Deep Level Transient Spectroscopy. Phys. Stat. Solidi A, 162:199–225, 1997. [43] G. Alfieri, E. V. Monakhov, and B. G. Svensson. Evidence for a deep two charge state defect in high energy electron irradiated 4H-SiC. Mater. Sci. Forum, 457-460:481–484, 2004. [44] G. Alfieri, E. V. Monakhov, M. K. Linnarsson, and B. G. Svensson. Capacitance Spectroscopy Study of High Energy Electron Irradiated and Annealed 4H-SiC. Mater. Sci. Forum, 483-485:365–368, 2005. [45] G. Alfieri, E. V. Monakhov, B. G. Svensson, and A. Hall´en. Defect energy levels in hydrogenimplanted and electron-irradiated n-type 4H silicon carbide. J. Appl. Phys., 98:113524, 2005.
43
BIBLIOGRAPHY
[46] A. Castaldini, A. Cavallini, L. Rigutti, F. Nava, S. Ferrero, and F. Giorgis. Deep levels by proton and electron irradiation in 4H-SiC. J. Appl. Phys., 98:053706, 2005. [47] K. Danno and T. Kimoto. Deep level transient spectroscopy on as-grown and electron-irradiated p-type 4H-SiC epilayers. J. Appl. Phys., 101:103704, 2007. [48] L. Storasta, J. P. Bergman, E. Janz´en, A. Henry, and J. Lu. Deep levels created by low energy electron irradiation in 4H-SiC. J. Appl. Phys., 96:4909–4915, 2004. [49] L. Storasta, F. H. C. Carlsson, J. P. Bergman, and E. Janz´en. Observation of recombination enhanced defect annealing in 4H–SiC. Appl. Phys. Lett., 86:091903, 2005. [50] A. Castaldini, A. Cavallini, L. Rigutti, and F. Nava. Low temperature annealing of electron irradiation induced defects in 4H-SiC. Appl. Phys. Lett., 85:3780–3782, 2004. [51] G. Alfieri, E. V. Monakhov, B. G. Svensson, and M. K. Linnarsson. Annealing behavior between room temperature and 2000 ◦ C of deep level defects in electron-irradiated n-type 4H silicon carbide. J. Appl. Phys., 98:043518, 2005. [52] A. Castaldini, A. Cavallini, L. Rigutti, S. Pizzini, A. Le Donne, and S. Binetti. Diffusion length and junction spectroscopy analysis of low-temperature annealing of electron irradiation-induced deep levels in 4H-SiC. J. Appl. Phys., 99:033701, 2006. [53] Y. Negoro, T. Kimoto, and H. Matsunami. Stability of deep centers in 4H-SiC epitaxial layers during thermal annealing. Appl. Phys. Lett., 85:1716–1718, 2004. [54] S. Sasaki, K. Kawahara, G. Feng, G. Alfieri, and T. Kimoto. Major deep levels with the same microstructures observed in n-type 4H–SiC and 6H–SiC. J. Appl. Phys., 109:013705, 2011. [55] T. Hiyoshi and T. Kimoto. Reduction of Deep Levels and Improvement of Carrier Lifetime in n-Type 4H–SiC by Thermal Oxidation. APEX, 2:041101, 2009. [56] T. Hiyoshi and T. Kimoto. Elimination of the Major Deep Levels in n- and p-Type 4H–SiC by TwoStep Thermal Treatment. APEX, 2:091101, 2009. [57] L. Storasta, H. Tsuchida, T. Miyazawa, and T. Ohshima. Enhanced annealing of the Z1/2 defect in 4H–SiC epilayers. J. Appl. Phys., 103:013705, 2008. [58] P. B. Klein, B. V. Shanabrook, S. W. Huh, A. Y. Polyakov, M. Skowronski, J. J. Sumakeris, and M. J. O’Loughlin. Lifetime-limiting defects in n- 4H-SiC epilayers. Appl. Phys. Lett., 88:052110, 2006. [59] J. M. Langer and H. Heinrich. Deep-Level Impurities: A Possible Guide to Prediction of Band-Edge Discontinuities in Semiconductor Heterojunctions. Phys. Rev. Lett., 55:1414–1417, 1985. [60] C. G. Hemmingsson, N. T. Son, A. Ellison, J. Zhang, and E. Janz´en. Negative-U centers in 4H silicon carbide. Phys. Rev. B, 58:R10119–R10122, 1998.
44
BIBLIOGRAPHY
[61] C. G. Hemmingsson, N. T. Son, and E. Janz´en. Observation of negative-U centers in 6H silicon carbide. Appl. Phys. Lett., 74:839–841, 1999. [62] C. G. Hemmingsson, N. T. Son, O. Kordina, and E. Janz´en. Metastable defects in 6H–SiC: experiments and modeling. J. Appl. Phys., 91:1324–1330, 2002. [63] D. M. Martin, H. Kortegaard Nielsen, P. L´evˆeque, A. Hall´en, G. Alfieri, and B. G. Svensson. Bistable defect in mega-electron-volt proton implanted 4H silicon carbide. Appl. Phys. Lett., 84:1704–1706, 2004. [64] H. K. Nielsen, A. Hall´en, and B. G. Svensson. Capacitance transient study of the metastable M center in n -type 4H − SiC. Phys. Rev. B, 72:085208, 2005. ¨ [65] A. Hall´en, M. Nawaz, C. Zaring, M. Usman, M. Domeij, and M. Ostling. Low-Temperature Annealing of Radiation-Induced Degradation in 4H-SiC Bipolar Junction Transistors. IEEE Electron Dev. Lett., 31:707 –709, 2010. [66] M. Usman, M. Nour, A.Yu. Azarov, and A. Hall´en. Annealing of ion implanted 4H-SiC in the temperature range of 100-800 ◦ C analysed by ion beam techniques. Nucl. Instrum. Meth. B., 268:2083 – 2085, 2010. [67] http://www.cree.com/products/power 1700V.asp, 2011. [68] http://www.cree.com/products/power mosfet.asp, 2011. [69] W. F. Knippenberg. Growth phenomena in silicon carbide. Philips Res. Reports, 18:161–274, 1963. ¨ ber die Kristalle des Carborundums. Z. Kristallogr., 50:33–39, 1912. [70] H. Baumhauer. U [71] C. J. Schneer. Polymorphism in one dimension. Acta Crystallogr., 8:279–285, 1955. [72] H. Jagodzinski. Eindimensionale Fehlordnung in Kristallen und ihr Einfluss auf die R¨ ontgeninterferenzen. I. Berechnung des Fehlordnungsgrades aus den R¨ ontgenintensit¨ aten. Acta Crystallogr., 2:201–207, 1949. [73] http://www.ioffe.ru/SVA/NSM/Semicond/SiC/, 2011. [74] M. Imade, T. Ogura, M. Uemura, F. Kawamura, M. Yoshimura, Y. Kitaoka, Y. Mori, T. Sasaki, M. Yamazaki, and S. Suwabe. Growth of 2H-SiC Single Crystals in a C-Li-Si Ternary Melt System. Mater. Sci. Forum, 600-603:55–58, 2009. [75] Y. Goldberg, M.E. Levinshtein, and S.L. Rumyantsev. Properties of Advanced Semiconductor Materials: GaN, AlN, SiC, BN, SiC, SiGe. John Wiley & Sons, Inc., New York, 2001. [76] W. J. Choyke, D. R. Hamilton, and L. Patrick. Optical Properties of Cubic SiC: Luminescence of Nitrogen-Exciton Complexes, and Interband Absorption. Phys. Rev., 133:A1163–A1166, 1964.
45
BIBLIOGRAPHY
[77] L. Patrick, D. R. Hamilton, and W. J. Choyke. Growth, Luminescence, Selection Rules, and Lattice Sums of SiC with Wurtzite Structure. Phys. Rev., 143:526–536, 1966. [78] W. H. Backes, F. C. de Nooij, P. A. Bobbert, and W. van Haeringen. Kronig-Penney-like description for band gap variation in SiC polytypes. Phys. B Condens. Matter, 217:207 – 211, 1996. [79] L. Patrick. Inequivalent Sites and Multiple Donor and Acceptor Levels in SiC Polytypes. Phys. Rev., 127:1878–1880, 1962. [80] H. Matsunami and T. Kimoto. Step-controlled epitaxial growth of SiC: High quality homoepitaxy. Mater. Sci. Eng. R., 20:125 – 166, 1997. [81] P. G. Neudeck. Encyclopedia of Materials: Science and Technology, chapter Silicon carbide electronic devices, pages 8508–8519. Elsevier Ltd., 2001. [82] R.J. Trew. SiC and GaN transistors - is there one winner for microwave power applications? IEEE J. Proc., 90(6):1032 – 1047, jun 2002. [83] A. R. Powell and L. B. Rowland. SiC Materials—Progress, Status, and Potential Roadblocks. Proceedings of the IEEE, 90:942–955, 2002. [84] R.T. Kemerley, H.B. Wallace, and M.N. Yoder. Impact of wide bandgap microwave devices on DoD systems. IEEE J. Proc., 90:1059 – 1064, 2002. [85] T. Paul Chow. High-voltage SiC and GaN power devices. Microelectron. Eng., 83:112 – 122, 2006. [86] M. Willander, M. Friesel, Q. ul Wahab, and B. Straumal. Silicon carbide and diamond for high temperature device applications. J. Mater. Sci.: Mater. Electron., 17:1–25, 2006. [87] J. Schneider and K. Maier. Point defects in silicon carbide. Physica B, 185:199–206, 1993. [88] M. Bockstedte, A. Mattausch, and O. Pankratov. Ab initio study of the annealing of vacancies and interstitials in cubic SiC: Vacancy-interstitial recombination and aggregation of carbon interstitials. Phys. Rev. B, 69:235202, 2004. [89] F. Gao, M. Posselt, V. Belko, Y. Zhang, and W. J. Weber. Structures and energetics of defects: a comparative study of 3C- and 4H-SiC. Nucl. Instrum. Meth. B., 218:74 – 79, 2004. [90] P. Y. Yu and M. Cordona. Fundamentals of Semiconductors, Graduate Texts in Physics, chapter Electronic Properties of Defects, pages 160–202. Springer Berlin / Heidelberg, 4 edition, 2010. [91] C. Hemmingsson. Deep Levels in Electron-Irradiated and As-grown SiC Power Devices Material. Link¨ oping studies in science and technology, dissertation no. 546, Link¨ oping University, 1998. [92] B. G. Svensson, A. Hall´en, J. Wong Leung, M. S. Janson, M. K. Linnarsson, A. Yu. Kuznetsov, G. Alfieri, U. Grossner, E. V. Monakhov, H. K.-Nielsen, C. Jagadish, and J. Grillenberger. Ion implantation processing and related effects in SiC. Mater. Sci. Forum, 527-529:781–786, 2006.
46
BIBLIOGRAPHY
[93] P. De´ ak, B. Aradi, A. Gali, and U. Gerstmann. ”Some like it shallower” - p-type doping in SiC. Phys. Stat. Solidi B, 235:139–145, 2003. [94] M. Ikeda, H. Matsunami, and T. Tanaka. Site effect on the impurity levels in 4H, 6H, and 15R-SiC. Phys. Rev. B, 22:2842–2854, 1980. [95] D. J. Larkin. SiC Dopant Incorporation Control Using Site-Competition CVD. Phys. Stat. Solidi B, 202:305–320, 1997. [96] V. V. Afanas’ev, M. Bassler, G. Pensl, M. J. Schulz, and E. Stein von Kamienski. Band offsets and electronic structure of SiC/SiO2 interfaces. J. Appl. Phys., 79:3108–3114, 1996. [97] H. H. Woodbury and G. W. Ludwig. Electron Spin Resonance Studies in SiC. Phys. Rev., 124:1083– 1089, 1961. [98] V. Heera, D. Panknin, and W. Skorupa. p-Type doping of SiC by high dose Al implantation problems and progress. Appl. Surf. Sci., 184:307–316, 2001. [99] I. G. Ivanov, B. Magnusson, and E. Janz´en. Optical selection rules for shallow donors in 4H − SiC and ionization energy of the nitrogen donor at the hexagonal site. Phys. Rev. B, 67:165212, 2003. [100] I. G. Ivanov, A. Henry, and E. Janz´en. Ionization energies of phosphorus and nitrogen donors and aluminum acceptors in 4H silicon carbide from the donor-acceptor pair emission. Phys. Rev. B, 71:241201, 2005. [101] N. I. Kuznetsov and A. S. Zubrilov. Deep centers and electroluminescence in 4H—SiC diodes with a p-type base region. Mater. Sci. Eng. B, 29:181 – 184, 1995. [102] A. van Duijn Arnold, J. Mol, R. Verberk, J. Schmidt, E. N. Mokhov, and P. G. Baranov. Spatial distribution of the electronic wave function of the shallow boron acceptor in 4H- and 6H-SiC. Phys. Rev. B, 60:15829–15847, 1999. [103] A. v. Duijn Arnold, T. Ikoma, O. G. Poluektov, P. G. Baranov, E. N. Mokhov, and J. Schmidt. Electronic structure of the deep boron acceptor in boron-doped 6H-SiC. Phys. Rev. B, 57:1607– 1619, 1998. [104] B. Aradi, A. Gali, P. De´ ak, E. Rauls, T. Frauenheim, and N. T. Son. Boron Centers in 4H-SiC. Mater. Sci. Forum, 353-356:455–458, 2001. [105] M. Bockstedte, A. Mattausch, and O. Pankratov. Boron in SiC: Structure and Kinetics. Mater. Sci. Forum, 353-356:447–450, 2001. [106] A. Lebedev. Deep level centers in silicon carbide: A review. Semiconductors, 33:107–130, 1999. 10.1134/1.1187657. [107] K. Mochizuki. One-dimensional Models for Diffusion and Segregation of Boron and for Ion Implantation of Aluminum in 4H-Silicon Carbide, chapter 2, pages 29–54. InTech, 2011.
47
BIBLIOGRAPHY
[108] V. S. Ballandovich. Deep-level transient spectroscopy of radiation-induced levels in 6H-SiC. Semiconductors, 33:1188–1192, 1999. [109] X. D. Chen, C. L. Yang, M. Gong, W. K. Ge, S. Fung, C. D. Beling, J. N. Wang, M. K. Lui, and C. C. Ling. Low Energy Electron Irradiation Induced Deep Level Defects in 6H-SiC: The Implication for the Microstructure of the Deep Levels E1 =E2. Phys. Rev. Lett., 92:125504–1–125504–4, 2004. [110] K. Danno, T. Hori, and T. Kimoto. Impacts of growth parameters on deep levels in n-type 4H-SiC. J. Appl. Phys., 101:053709, 2007. [111] L. Torpo, M. Marlo, T. E. M. Staab, and R. M. Nieminen. Comprehensive ab initio study of properties of monovacancies and antisites in 4H-SiC. J. Phys.: Condens. Matter, 13:6203, 2001. [112] I. Pintilie, L. Pintilie, K. Irmscher, and B. Thomas. Formation of the Z1/2 deep-level defects in 4H-SiC epitaxial layers: Evidence for nitrogen participation. Appl. Phys. Lett., 81:4841–4843, 2002. [113] A. Castaldini, A. Cavallini, and L. Rigutti. Assessment of the intrinsic nature of deep level Z1/Z2 by compensation effects in proton-irradiated 4H-SiC. Semicond. Sci. Tech., 21:724–728, 2006. [114] K. Danno and T. Kimoto. High-Temperature Deep Level Transient Spectroscopy on As-Grown P-Type 4H–SiC Epilayers. Jpn. J. Appl. Phys., 45:L285–L287, 2006. [115] A. P. Zhang, L. B. Rowland, E. B. Kaminsky, V. Tilak, J. C. Grande, J. Teetsov adn A. Vertiatchikh, and L. F. Eastman. Correlation of device performance and defects in AlGaN/GaN high-electron mobility transistors. J. Electron. Mater., 32:388–394, 2003. [116] L. Storasta, F. H. C. Carlsson, S. G. Sridhara, J. P. Bergman, A. Henry, T. Egilsson, A. Hall´en, and E. Janz´en. Pseudodonor nature of the DI defect in 4H-SiC. Appl. Phys. Lett., 78:46–48, 2001. [117] T. Egilsson, J. P. Bergman, I. G. Ivanov, A. Henry, and E. Janz´en. Properties of the D1 bound exciton in 4H − SiC. Phys. Rev. B, 59:1956–1963, 1999. [118] F. Redmann. Untersuchung von Siliziumkarbidkristallen mit Hilfe der Positronen-AnnihilationsSpektroskopie. PhD thesis, Martin-Luther-Universit¨ at Halle-Wittenberg, 2003. ˚ berg, L. Storasta, A. Hall´en, and B.G. Svensson. Implantation Temperature Dependent Deep [119] D. A Level Defects in 4H-SiC. Mater. Sci. Forum, 353-356:443–446, 2001. [120] M. L. David, G. Alfieri, E. M. Monakhov, A. Hall´en, C. Blanchard, B. G. Svensson, and J. F. Barbot. Electrically active defects in irradiated 4H-SiC. J. Appl. Phys., 95:4728–4733, 2004. [121] H. J. von Bardeleben, J. L. Cantin, L. Henry, and M. F. Barthe. Vacancy defects in p-type 6H − SiC created by low-energy electron irradiation. Phys. Rev. B, 62:10841–10846, 2000. [122] J. W. Steeds, F. Carosella, A. G. Evans, M. M. Ismail, L. R. Danks, and W. Voegeli. Differentiation between C and Si Related Damage Centres in 4H- and 6H-SiC by the Use of 90-300 kV Electron Irradiation Followed by Low Temperature Photoluminescence Microscopy. Mater. Sci. Forum, 353356:381–384, 2000.
48
BIBLIOGRAPHY
[123] D. V. Lang and L. C. Kimerling. Observation of Recombination-Enhanced Defect Reactions in Semiconductors. Phys. Rev. Lett., 33:489–492, 1974. [124] D. Stievenard and J. C. Bourgoin. Defect-enhanced annealing by carrier recombination in GaAs. Phys. Rev. B, 33:8410–8415, 1986. [125] N. Achtziger and W. Witthuhn. Slicon Carbide: Recent Major Advances, chapter Radiotracer Deep Level Transient Spectroscopy, pages 537–561. Springer Verlag Berlin, Heidelberg, New York, 2004. [126] K. Irmscher, I. Pintilie, L. Pintilie, and D. Schulz. Deep level defects in sublimation-grown 6H silicon carbide investigated by DLTS and EPR. Physica B, 308-310:730–733, 2001. [127] S. Eichler. Untersuchungen zu leerstellenartigen Kristalldefekten nach Ionenimplantation in Halbleitern. PhD thesis, Martin-Luther-Universit¨ at Halle-Wittenberg, 1998. [128] M. Moll. Radiation damage in Silicon Particle Detectors. PhD thesis, Universit¨ at Hamburg, 1999. [129] A. Chantre. Configurationally bistable C center in quenched Si:B: Possibility of boron-vacancy pair. Phys. Rev. B, 32:3687–3694, 1985. [130] A. Chantre. Introduction to Defect Bistablity. Appl. Phys. A, 48:3–9, 1989. [131] G. Pensl. Investigastion of deep impurities. Presentation at the tutorial day of the ICSCRM2009, 2009. [132] P. Blood and J. W. Orton. The Electrical Characterization of Semiconductors: Majority Carriers and Electron States. Academic Press, 1992. [133] A. M. Cowley and S. M. Sze. Surface States and Barrier Height of Metal-Semiconductor Systems. J. Appl. Phys., 36(10):3212–3220, 1965. [134] D. K. Schroder. Semiconductor Material and Devices. Wiley-Interscience, IEEE, 2006. [135] W. Shockley and W. T. Read. Statistics of the Recombinations of Holes and Electrons. Phys. Rev., 87:835–842, 1952. [136] R. N. Hall. Electron-Hole Recombination in Germanium. Phys. Rev., 87:387, 1952. [137] C. H. Henry and D. V. Lang. Nonradiative capture and recombination by multiphonon emission in GaAs and GaP. Phys. Rev. B, 15:989–1016, 1977. [138] M. Lax. Cascade Capture of Electrons in Solids. Phys. Rev., 119:1502–1523, 1960. [139] H. Okushi, Y. Tokumaru, S. Yamasaki, H. Oheda, and K. Tanaka. Energy dependence of electroncapture cross section of gap states in n-type a-Si:H. Phys. Rev. B, 25:4313–4316, 1982. [140] A. F. Tasch and C. T. Sah. Recombination-Generation and Optical Properties of Gold Acceptor in Silicon. Phys. Rev. B, 1:800–809, 1970.
49
BIBLIOGRAPHY
[141] P. A. Martin, B. G. Streetman, and K. Hess. Electric field enhanced emission from non-Coulombic traps in semiconductors. J. Appl. Phys., 52:7409–7415, 1981. [142] H. Lef`evre and M. Schulz. Double correlation technique (DDLTS) for the analysis of deep level profiles in semiconductors. Appl. Phys. A - Mater., 12:45–53, 1977. [143] D. V. Lang. Deep-level transient spectroscopy: A new method to characterize traps in semiconductors. J. Appl. Phys., 45:3023–3032, 1974. [144] K. Dmowski. A new correlation method for improvement in selectivity of bulk trap measurements from capacitance and voltage transients. Rev. Sci. Instrum., 61:1319–1325, 1990. [145] R. Brunwin, B. Hamilton, P. Jordan, and A. R. Peaker. Detection of minority-carrier traps using transient spectroscopy. Electron. Lett., 15:349–350, 1979.
50
