UV Solid-State Light Emitters and Detectors

NATO Science Series A Series presenting the results of scientific meetings supported under the NATO Science Programme. The Series is published by IOS Press, Amsterdam, and Kluwer Academic Publishers in conjunction with the NATO Scientific Affairs Division Sub-Series I. II. III. IV. V.

Life and Behavioural Sciences Mathematics, Physics and Chemistry Computer and Systems Science Earth and Environmental Sciences Science and Technology Policy

IOS Press Kluwer Academic Publishers IOS Press Kluwer Academic Publishers IOS Press

The NATO Science Series continues the series of books published formerly as the NATO ASI Series. The NATO Science Programme offers support for collaboration in civil science between scientists of countries of the Euro-Atlantic Partnership Council. The types of scientific meeting generally supported are “Advanced Study Institutes” and “Advanced Research Workshops”, although other types of meeting are supported from time to time. The NATO Science Series collects together the results of these meetings. The meetings are co-organized bij scientists from NATO countries and scientists from NATO’s Partner countries – countries of the CIS and Central and Eastern Europe. Advanced Study Institutes are high-level tutorial courses offering in-depth study of latest advances in a field. Advanced Research Workshops are expert meetings aimed at critical assessment of a field, and identification of directions for future action. As a consequence of the restructuring of the NATO Science Programme in 1999, the NATO Science Series has been re-organised and there are currently Five Sub-series as noted above. Please consult the following web sites for information on previous volumes published in the Series, as well as details of earlier Sub-series. http://www.nato.int/science http://www.wkap.nl http://www.iospress.nl http://www.wtv-books.de/nato-pco.htm

Series II: Mathematics, Physics and Chemistry – Vol. 144

UV Solid-State Light Emitters and Detectors edited by

Michael S. Shur Rensselaer Polytechnic Institute, Troy, NY, U.S.A. and

Artu¯ras Z˘ukauskas Vilnius University, Vilnius, Lithuania

Kluwer Academic Publishers Dordrecht / Boston / London Published in cooperation with NATO Scientific Affairs Division

Proceedings of the NATO Advanced Research Workshop on UV Solid-State Light Emitters and Detectors Vilnius, Lithuania 17–21 June 2003 A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN 1-4020-2035-X (PB) ISBN 1-4020-2034-1 (HB) ISBN 1-4020-2103-8 (e-book)

Published by Kluwer Academic Publishers, P.O. Box 17, 3300 AA Dordrecht, The Netherlands. Sold and distributed in North, Central and South America by Kluwer Academic Publishers, 101 Philip Drive, Norwell, MA 02061, U.S.A. In all other countries, sold and distributed by Kluwer Academic Publishers, P.O. Box 322, 3300 AH Dordrecht, The Netherlands.

Printed on acid-free paper

All Rights Reserved © 2004 Kluwer Academic Publishers No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed in the Netherlands.

Contents

Contributing Authors Preface Basic Device Issues in UV Solid-State Emitters and Detectors M. S. SHUR AND A. ŽUKAUSKAS

ix xiii 1

HVPE-Grown AlN-GaN Based Structures for UV Spectral Region A. S. USIKOV, YU. MELNIK, A. I. PECHNIKOV, V. A. SOUKHOVEEV, O. V. KOVALENKOV, E. SHAPOVALOVA, S. YU. KARPOV, AND V. A. DMITRIEV

15

GaN-Based Laser Diodes S. EINFELDT, S. FIGGE, T. BÖTTCHER, AND D. HOMMEL

31

Quaternary AlInGaN Materials System for UV Optoelectronics E. KUOKSTIS, G. TAMULAITIS, AND M. ASIF KHAN

41

III-Nitride Based UV Light Emiting Diodes R. GASKA, M. ASIF KHAN, AND M. S. SHUR

59

UV Metal Semiconductor Metal Detectors 77 J- L. REVERCHON, M. MOSCA, N. GRANDJEAN, F. OMNES, F. SEMOND, J- Y. DUBOZ, AND L. HIRSCH

vi Characterization of Advanced Materials for Optoelectronics by Using UV Lasers and Four-Wave Mixing Techniques K. JARAŠINjNAS

93

Quantum Phospors A. P. VINK, E. VAN DER KOLK, P. DORENBOS, AND C.W.E. VAN EIJK

111

Optical Measurements Using Light-Emitting Diodes A. ŽUKAUSKAS, M. S. SHUR, AND R. GASKA

127

Novel AlGaN Heterostructures for UV Sensors and LEDs M. STUTZMANN

143

Nitride Photodetectors in UV Biological Effects Studies E. MUÑOZ, J. L. PAU, AND C. RIVERA

161

Promising Results of Plasma Assisted MBE for Optoelectronic Applications A. GEORGAKILAS, E. DIMAKIS, K. TSAGARAKI, AND M. ANDROULIDAKI Low Dislocations Density GaN/Sapphire for Optoelectronic Devices B. BEAUMONT, J.-P. FAURIE, E. FRAYSSINET, E. AUJOL, AND P. GIBART

179

189

Stimulated Emission and Gain in GaN Epilayers Grown on Si 199 A. L. GURSKII, E. V. LUTSENKO, V. Z. ZUBIALEVICH, V. N. PAVLOVSKII, G. P. YABLONSKII, K. KAZLAUSKAS, G. TAMULAITIS, S. JURSENAS, A. ZUKAUSKAS, Y. DIKME, H. KALISCH, A. SZYMAKOWSKI, R. H. JANSEN, B. SCHINELLER, AND M. HEUKEN Materials Characterization of Group-III Nitrides under High-Power Photoexcitation S. JURŠƠNAS, G. KURILýIK, S. MIASOJEDOVAS, AND A. ŽUKAUSKAS Small Internal Electric Fields in Quaternary InAlGaN Heterostructures S. ANCEAU, S. P. àEPKOWSKI, H. TEISSEYRE, T. SUSKI, P. PERLIN, P. LEFEBVRE, L. KOēCZEWICZ, H. HIRAYAMA, AND Y. AOYAGI

207

215

vii

MOCVD Growth of AlGaN Epilayers and AlGaN/GaN SLs in a Wide Composition Range W. V. LUNDIN, A. V. SAKHAROV, A. F. TSATSUL'NIKOV, E. E. ZAVARIN, A. I. BESULKIN, A. V. FOMIN, AND D. S. SIZOV

223

Gallium Nitride Schottky Barriers and MSM UV Detectors B. BORATYNSKI AND M. TLACZALA

233

III-Nitride Based Ultraviolet Surface Acoustic Wave Sensors D. ýIPLYS, A. SEREIKA, R. RIMEIKA, R. GASKA, M. SHUR, J. YANG, AND M. ASIF KHAN

239

Optically Pumped InGaN/GaN/AlGaN MQW Laser Structures V. YU. IVANOV, M. GODLEWSKI, H. TEISSEYRE, P. PERLIN, R. CZERNECKI, P. PRYSTAWKO, M. LESZCZYNSKI, I. GRZEGORY, T. SUSKI, AND S. POROWSKI

247

High Power LED and Thermal Management A. MAHLKOW

253

Detection of Blue Light by Self-Assembled Monolayer of Dipolar Molecules O. NEILANDS, N. KIRICHENKO, I. MUZIKANTE, E. FONAVS, L. GERCA, S. JURSENAS, R. VALIOKAS, R. KARPICZ, AND L. VALKUNAS

261

Atomic and Molecular Spectroscopy with UV and Visible Superbright LEDs 271 G. PICHLER, T. BAN, H. SKENDEROVIû, AND D. AUMILER Semi-Insulating GaN and its First Tests for Radiation Hardness as an Ionizing Radiation Detector J. V. VAITKUS, W. CUNNINGHAM, M. RAHMAN, K. M. SMITH, AND S. SAKAI Towards the Hybrid Biosensors Based on Biocompatible Conducting Polymers A. RAMANAVICIENE AND A. RAMANAVICIUS

279

287

viii Optically Pumped UV-Blue Lasers Based on InGaN/GaN/Al2O3 and InGaN/GaN/Si Heterostructures G. P. YABLONSKII, A. L. GURSKII, E. V. LUTSENKO, V. Z. ZUBIALEVICH, V. N. PAVLOVSKII, A. S. ANUFRYK, Y. DIKME, H. KALISCH, R. H. JANSEN, B. SCHINELLER, AND M. HEUKEN

297

Key Word Index

305

Author Index

307

Contributing Authors

Key Lectures

EINFELDT Sven

Institute of Solid State Physics, University of Bremen, P.O. Box 330440, 28334 Bremen, Germany [email protected]

GASKA Remis

Sensor Electronic Technology, Inc., 1195 Atlas Road, Columbia, SC 29209, USA [email protected]

JARAŠINjNAS KĊstutis

Institute of Materials Science and Applied Research, Vilnius University, Saulơtekio 9-III, 2040 Vilnius, Lithuania [email protected]

KUOKSTIS Edmundas

Department of Electrical Engineering, University of South Carolina, Columbia, SC 29208, USA [email protected]

ix

x

MUÑOZ Elias

Institute for Systems Optoelectronics and Microtechnology and DIE ETSI Telecomunicación, Universidad Politecnica de Madrid, 28040 Madrid, Spain [email protected]

REVERCHON Jean-Luc

Thales Research & Technology, 91404 Orsay Cedex, France [email protected]

SHUR Michael

Center for Broadband Data Transport, Rensselaer Polytechnic Institute, CII 9017, 110 8th street, Troy, NY 12180, USA [email protected]

STUTZMANN Martin

Walter Schottky Institut, Technische Universität München, 85748 Garching, Germany [email protected]

TAMULAITIS Gintautas

Institute of Materials Science and Applied Research, Vilnius University, Saulơtekio 9-III, 2040 Vilnius, Lithuania [email protected]

USIKOV Alexander

TDI, Inc., 12214 Plum Orchard Dr., Silver Spring, MD 20904, USA [email protected]

VINK Arjan

Interfaculty Reactor Institute, Delft University of Technology, Mekelweg 15, 2629 JB Delft, The Netherlands [email protected]

ŽUKAUSKAS Artnjras

Institute of Materials Science and Applied Research, Vilnius University, Saulơtekio 9-III, 2040 Vilnius, Lithuania [email protected]

xi

Poster Presentations

BORATYNSKI Boguslaw

Faculty of Microsystem Electronics and Photonics, Wroclaw University of Technology, Janiszewskiego 11/17, 50-372 Wroclaw, Poland [email protected]

ýIPLYS Daumantas

Department of Radiophysics, Vilnius University, Saulơtekio 9-III, 2040 Vilnius, Lithuania [email protected]

GEORGAKILAS Alexandros

Institute of Electronic Structure and Laser, Foundation for Research and TechnologyHellas, P.O. Box 1527, 71110 Heraklion, Crete, Greece [email protected]

GIBART Pierre

Lumilog, 2720, Chemin de Saint Bernard, Les Moulins I, 06220 Vallauris, France [email protected]

GURSKII Alexander

Stepanov Institute of Physics of NAS Belarus, F. Skaryna Ave. 68, 220072 Minsk, Belarus [email protected]

IVANOV Vitalii

Institute of Physics, Polish Academy of Sciences, Al. Lotników 32/46, 02-668 Warsaw, Poland [email protected]

JURŠƠNAS Saulius

Institute of Materials Science and Applied Research, Vilnius University, Saulơtekio 9-III, 2040 Vilnius, Lithuania [email protected]

xii

àEPKOWSKI Sáawomir

UNIPRESS, Polish Academy of Sciences, Sokoáowska 29/37, Warszawa, Poland [email protected]

LUNDIN Wsevolod

A.F. Ioffe Physico-Technical Institute of the Russian Academy of Science, 194021 St.Petersburg, Russia [email protected]

MAHLKOW Adrian

Optotransmitter Umweltschutz Technologie e.V., Berlin, Germany [email protected]

MUZIKANTE Inta

Institute of Physical Energetics, Aizkraukles Str. 21, LV-1006 Riga, Latvia [email protected]

PICHLER Goran

Institute of Physics, Bijeniþka cesta 46, P. O. Box 304, HR-10001 Zagreb, Croatia [email protected]

RAMANAVIýIUS Arnjnas

Laboratory of Bioanalysis, Institute of Biochemistry, Mokslininkǐ 12, 2600 Vilnius, Lithuania [email protected]

VAITKUS Juozas

Institute of Materials Science and Applied Research, Vilnius University, Saulơtekio 9-III, 2040 Vilnius, Lithuania [email protected]

YABLONSKII Genadii

Stepanov Institute of Physics of NAS Belarus, F. Skaryna Ave. 68, 220072 Minsk, Belarus [email protected]

Preface

Infrared and visible light LEDs and photodetectors have found numerous applications and have become a truly enabling technology. The promise of solid state lighting has invigorated interest in white light LEDs. Ultraviolet LEDs and solar blind photodetectors represent the next frontier in solid state emitters and hold promise for many important applications in biology, medicine, dentistry, solid state lighting, displays, dense data storage, and semiconductor manufacturing. One of the most important applications is in systems for the identification of hazardous biological agents. Compared to UV lamps, UV LEDs have lower power consumption, a longer life, compactness, and sharper spectral lines. UV LEDs can provide a variety of UV spectra and have shape and form factor flexibility and ruggedness. Using conventional phosphors, UV LEDs can generate white light with high CRI and high efficiency. If quantum cutter phosphors are developed, white light generation by UV LEDs might become even more efficient. Advances in semiconductor materials and in improved light extraction techniques led to the development of a new generation of efficient and powerful visible high-brightness LEDs and we expect that similar improvements will be achieved in solid-state UV technology. NATO Advanced Research Workshop UV Solid-State Light Emitters and Detectors took place on June 17–21, 2003 in Vilnius, Lithuania (see http://www.natoarw-uv.ff.vu.lt). It brought together leading researchers in semiconductor UV technology and systems applications. The topics covered at the workshop ranged from basic device issues to substrates, epitaxial growth, materials characterization, nitride quaternary alloys, doping, strain energy band engineering, quantum phosphors, ohmic contacts and Schottky xiii

xiv barriers, UV LED and solar blind photodetector device design and performance, thermal management, and applications for biological hazardous agent sensing, solid state lighting, environmental control, and optical measurements. All these issues are presented in these Proceedings, and we hope that this book will be useful for students, engineers, scientists, and researchers interested in solid state light emitters and detectors and in wide band gap semiconductor technology. We gratefully acknowledge the support of the workshop by NATO, the US Defense Advanced Research Projects Agency, Ministry of National Defense of the Republic of Lithuania, Ministry of Education and Science of the Republic of Lithuania, Lithuanian State Foundation of Science and Studies, Vilnius University, Center of Broadband Data Transport Science and Technology at Rensselaer Polytechnic Institute, Sensor Electronic Technology, Inc., and EKSPLA Ltd.

Michael S. Shur and Artnjras Žukauskas Troy, NY, USA – Vilnius, Lithuania

BASIC DEVICE ISSUES IN UV SOLID-STATE EMITTERS AND DETECTORS M. S. SHUR 1 and A. ŽUKAUSKAS 2 1

Center for Broadband Data Transport, Rensselaer Polytechnic Institute, CII 9017, 110 8th street, Troy, New York 12180, USA 2 Institute of Materials Science and Applied Research, Vilnius University, Saulơtekio 9-III, LT-2040 Vilnius, Lithuania

Abstract:

UV light emitting diodes (LEDs) and lasers are expected to find numerous applications in biotechnology, medicine, dentistry, home security, food and air safety technology, short-range covert communications, industry, and solidstate lighting. 340–400-nm LEDs are already available commercially and milliwatt power 285-nm LEDs have been demonstrated in a laboratory. In parallel, AlGaN alloys with large molar fractions of Al for UV solar blind Schottky barrier, p–n junction and MSM detectors have been demonstrated. Recent work on surface-acoustic-wave (SAW) UV detectors revealed their potential for remote solar blind detection applications. However, with decreasing wavelengths, UV LEDs power is dropping and challenges in growing high quality nitride heterostructures with a high aluminum molar fraction are becoming more formidable. The solutions to device problems lie in using better substrates (with bulk AlN substrates in non-polar orientations being especially promising), using better epitaxial growth techniques, improving device design and using better contact technology and design.

Key words:

ultraviolet LEDs, solar blind photodetectors, aluminum gallium nitride

1.

INTRODUCTION

When Monsanto introduced the first commercial visible LEDs in 1968, they produced only red light with the intensity of approximately 10–3 lumen, barely visible under ambient light. Much brighter LEDs with colors ranging from red to yellow and green have been developed between 70’s and mid 90’s. As a result, infrared and visible LEDs as well as their counterparts, 1 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 1–13. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

M. S. Shur and A. Zukauskas

2

photodetectors, have found numerous applications and have become a truly enabling technology. Pioneering work of Pankove, Akasaki, and Nakamura has led to the development of bright green and blue LEDs based on nitride semiconductors in the recent decade. Further development of the AlInGaN materials system resulted in an appearance of ultraviolet (UV) LEDs and solar blind photodetectors, which represent the next frontier in solid-state optoelectronics with a huge potential in biological, medical and environmental instrumentation, dense data storage, disinfection, deodorization, communications, and solid-state lighting. UV spectral region spans from 100 to 400 nm and is usually divided into three subregions based on the absorption in the atmosphere and biological action of radiation. UV radiation with wavelengths from approximately 315 to 400 nm is referred to as UVA; UV radiation with wavelengths from 280 to 315 nm is referred to as UVB; and UV radiation with wavelengths from 100 to 280 nm is referred to as UVC (see Fig. 1). UVA penetrates the atmosphere without substantial absorption and causes minor biological action, mainly premature aging of the skin. UVB is partially absorbed in the atmosphere; it results in sunburn and may cause skin cancer. UVC doesn’t reach the earth’s surface due to absorption in ozone contained in the upper atmosphere and is highly dangerous for live organisms because of strong absorption in proteins.

UVC 100-280 nm

UVB

UVA 280- 315 315-4 00 nm nm

UV radia tion Visible radiation (light ) Infr ared radiation (IR)

100 - 40 0 n m 400 - 760 nm > 760 nm

Figure 1. Classification of spectral ranges.

Figure 2 (from [1]) shows the energy gaps of semiconductor materials and the corresponding wavelengths. As seen, the group-III nitride materials family spanning the direct energy gaps from 0.8 eV to 6.2 eV is ideal for applications in UV emitter and detector technology for UVA, UVB, and, partially, UVC regions. In particular, the band gap of InGaN alloy covers a part of the UVA region and AlGaN alloy can be tailored to wavelengths ranging form 360 to 200 nm depending on the Al molar fraction.

Basic Device Issues in UV Solid-State Emitters and Detectors

3

W avelength (nm ) 2000 AlN ZnS

800 600 500 400

300

200

C

GaN ZnO SiC( 4H)) SiC(6H) ) ZnSe CdS AlP CdO SiC(6H)) GaP Zn Te AlAs InN CdSe Al Sb CdTe GaAs InP Si GaSb Ge InAs InSb 0

UV

IR

1

2

3

4

5

6

7

B and Gap Energy (eV )

Figure 2. Semiconductor bandgaps and corresponding radiation wavelengths. Human eye sensitivity curve (arbitrary logarithmic scale) is also shown [1].

2.

UV LIGHT-EMITTING DIODES

Compared to UV lamps, UV LEDs have lower power consumption, a longer lifetime, compactness, and sharper spectral lines. UV LEDs can provide a variety of UV spectra and have shape and form factor flexibility and ruggedness. In particular, UV LEDs are expected to be used for the disruptive technology of solid-state lighting [2]. Using conventional phosphors, UV LEDs can generate white light with high color rendering properties and high efficiency. If quantum cutter phosphors are developed, white light generation by UV LEDs might become even more efficient. Another important application of UV LEDs is fluorescence excitation [2,3]. Based on this technique, novel and cost-efficient instruments for detection and characterization of biochemical compounds and biological agents, including hazardous agents, can be developed. Soon after the invention of the p–n junction GaN UV LED by Akasaki et al. in 1992 [4], a tremendous progress in solid-state sources of UV light was achieved. Present UV LEDs are based on heterostructures developed using nitride materials systems GaN/AlGaN [5], InGaN/AlGaN [6], AlGaN/AlGaN [7], and quaternary AlInGaN (for a review, see Ref. 8). For mature UV LEDs, the main device issues to be addressed are almost the

4

M. S. Shur and A. Zukauskas

same as those for advanced visible LEDs [2]: the chips must feature electronic structure that facilitates high efficiency of carrier injection into the active layer, the internal quantum efficiency should be maximized by enhancing radiative recombination and suppressing the nonradiative recombination, and light generated within the chip must be efficiently extracted. Reduction of the dislocation density and preventing cracking of epitaxial layers mismatched to the substrate is one of the most important issues. Most of fabrication approaches employ growth on sapphire substrate, which has a 16% lattice mismatch with GaN. This drawback is being bypassed through dislocation filtering by epitaxial lateral overgrowth and by using superlattices, strain-compensating layers and quaternary AlInGaN alloys (strain engineering approach). To substitute sapphire, novel substrates are being searched for UV LEDs. An example of such a substrate is bulk AlN, which offers identical crystal structure, close lattice and thermal expansion match to high Al-content nitride alloys, and refraction index favorable for UV light extraction. A UV LED grown over bulk AlN was recently reported by Xerox PARC and Crystal IS (see Fig. 3).

Figure 3. UV LED on bulk AlN substrate (courtesy Crystal IS, Inc.).

To increase the internal quantum efficiency, optimization of quantumwell structures is required through selecting composition and doping profiles of the well and barrier layers, shaping of the interfaces, and engineering of the built-in electric field to avoid the quantum-confined Stark effect. In addition, basic research for unveiling the routes of nonradiative recombination in AlGaN alloys with high molar fraction of aluminum is needed. Wide band gap of semiconductors used in UV LEDs have high ionization energies for shallow impurities, especially for acceptors. This results in difficulties with p-doping and increased resistivity of the layers and contacts.

Basic Device Issues in UV Solid-State Emitters and Detectors

5

Output power, mW

To overcome these difficulties, novel doping approaches including piezoelectric and superlattice doping, as well as co-doping are being developed. Finally, specific issues related to UV light extraction must be addressed. Conventional plastics used in visible-LED domes absorb UV radiation and should be substituted by new materials. New plastic materials, optical couplers to silica windows, as well as novel transparent contacts are required for further promotion of UV LED technology. First commercial 375-nm LEDs were introduced by Nichia [6]. Typically, these devices feature 1.5–2 mW optical power and are available with the outcoupling optics for narrow-angle (20º) and wide-angle (110º) radiative pattern. Cree introduced the first near-UV LED for use in the illumination market in 2001 (12 mW, 405-nm and 395-nm UV InGaN on SiC substrate devices). These LEDs have a geometrically enhanced vertical chip structure to maximize light extraction efficiency and require only a single wire bond connection. Recently, considerable progress in penetration into the UVC spectral region was achieved. SET/USC/RPI team has already reported on UV LEDs with the wavelength as short as 265 nm [9,10]. Deep UV LEDs with the peak optical powers of 3 mW (1 A) at 280 nm and 10 mW (1 A) at 325 nm were fabricated and characterized (see Fig. 4). Sandia National Laboratories have demonstrated UV LEDs with 290 nm wavelength with 1.3 mW of output power and with 275 nm wavelength with 0.4 mW of output power [11].

10

1

pulse 1A

dc 100 mA

0.1 270 280

290 300

310 320

330 340

350

Wavelength, nm Figure 4. Output power of UV LEDs versus wavelength [9].

A high-power 365-nm UV LED is being developed by Nichia and is expected to enter the market in 2004. The chip containing an active InGaN/AlGaN multiple quantum well structure is separated from its sapphire substrate and mounted on a CuW heat sink. The device has an output power

M. S. Shur and A. Zukauskas

6

of 210 mW when driven at dc 500 mA and 4.3 V and has an external quantum efficiency of 12.4% [12].

3.

SOLAR BLIND PHOTODETECTORS

Equally impressive progress has been achieved in solar blind nitride based UV detectors. Figure 5 shows different types of nitride based UV photodetectors. ohmic t ransparent Schot t ky Schot t ky

phot oconduct ive GaN

phot oconduct ive AlGaN

Met al-A lGaN- Met al

n- GaN

i - Ga N

n-AlGaN p- GaN

t ransparent Schottky contact

Met al-GaN-Met al

opt oelect ronic HFET

p-i-n

junct ion

phot oconduct ive AlGaN

Figure 5. Schematic representation of photodetectors realized using the GaN materials system [13].

Direct band-gap AlxGa1–xN layers with x t 0.4 exhibit a sharp transmission cut-off at Ȝ < 280 nm. Devices using such layers found applications in jet engines, furnaces, environmental monitors, and missile detection systems. Both AlGaN based photoconductive and photovoltaic detectors have been explored [14–16]. Photoconductive devices have large gains but slow response times. Also, photoconductive devices cannot operate at zero bias and, therefore, have an extra noise coming from the dark current. AlGaN p–i–n photodiodes have several drawbacks [17–20] related to the difficulties of ptype doping of AlGaN layers with a high Al content [18] and to a high resistance of ohmic contacts to p-type AlGaN layers. This resistance can be decreased by using p-GaN as the contact layer [17,19,20] with i-AlxGa1–xN (x > 0.4) as the active layer. However, the contact GaN layer absorbs a significant fraction of the optical beam reducing the device responsivity and

Basic Device Issues in UV Solid-State Emitters and Detectors

7

deteriorating UV/visible selectivity. Also, to avoid cracking, i-AlGaN active layer thickness has to be well below 2000 Å. A metal–semiconductor–metal (MSM) design [21] does not require ohmic contacts. However, the MSM devices cannot operate at zero bias, which increases the noise. Also, at moderate bias values, the photo response of MSM diodes has a significant slow photoconductive component, since the space charge width in the AlGaN layer is smaller than the electrode spacing. Lateral geometry transparent Schottky barrier photodetectors avoid most of the above problems [22]. However, this design requires an n-doped Al0.4Ga0.6N layer, and such doping attempts using Si had resulted in insulating material. Carrano et al. [23] studied the current transport mechanisms in GaNbased MSM photodetectors. They concluded that thermionic and thermionicfield emissions were the dominant transport mechanisms. The traps affecting the current transport seemed to be surface defects (such as threading dislocations) and deep defect states, which are within the tunneling distance from the heterointerface. Adivarahan et al. [24] reported on a new In–Si co-doping approach to obtain n-Al0.4Ga 0.6N active layers with resistivity as low as 0.16 ohm·cm. In addition to a significantly increased doping efficiency, the introduction of a small concentration of In also allows for the direct deposition of a crack-free 0.5 Pm thick Si-doped Al0.4Ga0.6N layer over a 200 ǖ thick AlN buffer layer on basal plane sapphire substrates. They also demonstrated the potential of using these In–Si co-doped layers for a lateral geometry, true solar-blind Schottky barrier detector (Ȝ cut-off at 278 nm). The increased n-type doping due to the addition of In can result from the introduction of a shallow impurity level. Indium incorporation might also reduce the defect formation as indicated by the improved structural quality and morphology of the grown films [25]. Indium might counteract the incorporation of defects responsible for the self-compensation of high Al mole fraction AlGaN layers, such as DX centers and cation vacancies [26]. Figure 6 (from [24]) shows optical transmission and photoluminescence spectra of In co-doped AlGaN layer at room temperature. Figure 7(a) shows the current-voltage characteristics measured between two 50 Pm × 150 Pm transmission line model (TLM) pads separated by a 2 Pm gap [24]. The characteristics are linear due to the low sheet resistivity of the AlGaN layer. The TLM measurements yielded the specific contact resistivity to be 2.5u10–3 ohm·cm. This was the first ever-reported data on ohmic contact resistivity to thick AlGaN epilayers with Al fraction of about 40%. Figure 7(b) shows the dark I–V characteristics for the transparent Schottky barriers fabricated by Adivarahan et al. [24]. As seen, the turn-on voltage and the forward differential resistance were approximately 1.2 V and 500

M. S. Shur and A. Zukauskas

8

ohm, respectively, and the reverse leakage current at a bias of –3 V was as low as 6 nA. The effective Schottky barrier height extracted from temperature measurements was 0.64 V.

60

40

20

260

270

280

290

300

Transmitivity (%)

PL Intensity (arb. units)

80

0 310

Wavelength (nm)

Figure 6. Spectra of optical transmission and photoluminescence of the In co-doped AlGaN epilayer with approximately 40% of Al [24].

5

(a)

3.0 In co-doped

2.5 2.0

zero In flux 0

-5

-2.5

0 -2.5

2.5

5

Current, mA

Current, mA

2.5

(b)

1.5 1.0 6.3 nA

0.5 0.0

-5

Voltage,V

-3

-2

-1 0 1 Voltage, V

2

3

Figure 7. Current-voltage (I–V) characteristics of the ohmic contacts to In co-doped AlGaN (a) and of the AlGaN Schottky photodiode (b). Dashed line in the left figure shows the I–V curve for the AlGaN layer with no In co-doping [24].

Figure 8 shows the measured responsivity of this In co-doped Schottky photodiode.

Basic Device Issues in UV Solid-State Emitters and Detectors

9

Responsivity (A/W)

0.1

0.01

1E-3

1E-4

1E-5

260

280

300

320

340

360

380

400

Wavelength (nm) Figure 8. Photoresponsivity spectrum of the AlGaN photodiode [24].

Rumyantsev et al. [27] studied low-frequency noise in Schottky barrier Al0.4Ga0.6N diodes. At forward bias, the low-frequency noise is a superposition of the 1/f and generation-recombination noise. The spectral noise density, SI, of current fluctuations increases as SI ~ I1.5 at low currents I and as SI ~ I2.5 at high currents (see Figure 9). The measured dependencies of noise on forward current show that the noise is a superposition of the noise from Schottky barrier and from the series resistance of the contacts and/or the base. At high current densities, when the noise from the base or contacts is dominant, the upper bound of the Hooge parameter in AlGaN was estimated as D < 10. However, this high value of D does not present an obstacle for practical applications of these photodetectors, since their detectivity is primarily limited by the thermal noise of the load resistance. Osinsky et al. [28] demonstrated visible-blind GaN Schottky barrier detectors grown on Si (111). The spectral response of the lateral Schottky barrier detectors had a cutoff at 365 nm with peak responsivities of 0.05 A/W at zero gate bias and 0.1 A/W at –4 V bias. This work demonstrated a possibility of integrating GaN-based photodetectors with Si electronics. Van Hove et al. [29] fabricated GaN and AlGaN 1u10 photodetector arrays for high temperature sensing applications. The device epitaxial layers were grown on sapphire substrates by RF atomic nitrogen plasma molecular beam epitaxy (MBE). At room temperature, GaN p–i–n photodetectors had a peak sensitivity of 0.198 A/W at 360 nm (which corresponded to the internal quantum efficiency of 85%). The devices operated up to 400 oC.

M. S. Shur and A. Zukauskas

10

2

1

10

2

Current density (A/cm )

Current density (A/cm ) -5 -3 -1 10 10 10 -10

SI (A /Hz)

10

-14

2

10

-18

10

1

10 10

-3

10

-7

2

I forward bias

-11

10 0

4

8

12

Voltage (V)

16

reverse bias

1.5

I

-22

10

SIT=4kT/Req

SIc=2qI

-26

10

10

-10

-8

10

-6

-4

10 10 Current (A)

-2

10

Figure 9. Dependence of spectral noise density of current fluctuations SI on current. T = 300 K, f = 1 Hz. Open and closed symbols show experimental data for forward and reverse biases, respectively. Dashed line shows the level of thermal noise SIT = 4kT/Req (Req = 50 :). Dotdashed line shows the level of the shot noise SIc = 2qI (q is the electron charge). Dotted line shows the slope of current dependence of noise for the reverse bias. Inset shows current– voltage characteristic of Schottky barrier Al0.4Ga0.6N diode (device area 9u10–4 cm2) [27].

Khan et al. [30,31] reported on a photodetector based on a 0.2-Pm gate AlGaN/GaN heterostructure field effect transistor (HFET). The epilayer structure and processing details for the gated photodetectors are similar to those for short gate AlGaN/GaN HFETs. Ciplys et al. [32–34] reported on a SAW-based UV GaN sensor by placing a SAW element into an oscillator feedback loop. The output of such sensor was a radio signal with UV radiation-dependent frequency, which made this sensor attractive for remote sensing applications. Several review papers and book chapters discuss nitride based UV photodetectors, see, for example, Refs. 13,35–38.

4.

CONCLUSIONS

Further progress in nitride based UV emitters and photodetectors will depend on several materials and device issues ranging from using large area substrates, improved homo- and heteroepitaxy of nitrides, improved p-type and n-type doping of high Al molar fraction epilayers, better light extraction designs for UV LEDs, thermal management, and UV resistant packaging.

Basic Device Issues in UV Solid-State Emitters and Detectors

11

Quantum phosphors (quantum cutters) might be essential for applications of UV LEDs in solid-state lighting. And, of course, reaching higher yields and reducing costs will be important. The development of UV semiconductor lasers and moving toward shorter wavelengths, especially into the UVC region, which is important for protein excitation, will stimulate biophotonics applications such as clinical screening, point-of-care medical instrumentation, environmental control, and biological hazardous agent detection.

REFERENCES 1. 2. 3. 4. 5.

6.

7. 8. 9.

10. 11. 12.

13. 14.

15.

M. S. Shur, Introduction to Electronic Devices (Wiley, New York, 1996). A. Žukauskas, M. S. Shur, and R. Gaska, Introduction to Solid State Lighting (Wiley, New York, 2002). A. Žukauskas, M. S. Shur, and R. Gaska, “Optical measurements using light-emitting diodes,” this volume. I. Akasaki, H. Amano, K. Itoh, N. Koide, and K. Manabe, “GaN-based UV/blue light emitting devices,” Inst. Phys. Conf. Ser. 129, pp. 851–856 (1992). J. Han, M. H. Crawford, R. J. Shul, J. J. Figiel, M. Banas, L. Zhang, Y. K. Song, H. Zhou, and A. V. Nurmikko, “AlGaN/GaN quantum well ultraviolet light emitting diodes,” Appl. Phys. Lett. 73, pp. 1688–1690 (1998). T. Mukai, M. Yamada, and S. Nakamura, “Current and temperature dependences of electroluminescence of InGaN-based UV/blue/green light-emitting diodes,” Jpn. J. Appl. Phys. 37, pp. L1358–L1361 (1998). T. Nishida, and N. Kobayashi, “346 nm emission from AlGaN multi-quantum-well light emitting diode,” Phys. Stat. Sol. A 176, pp. 45–48 (1999). E. Kuokstis, G. Tamulaitis, and M. Asif Khan, “Quaternary AlInGaN materials system for UV optoelectronics,” this volume. A. Chitnis, V. Adivarahan, J. P. Zhang, M. Shatalov, S. Wu, J. Yang, G. Simin, M. Asif Khan, X. Hu, Q. Fareed, R. Gaska, and M. S. Shur, “Milliwatt power AlGaN quantum well deep ultraviolet light emitting diodes,” Phys. Stat. Sol. A 200, pp. 99–101 (2003) R. Gaska, A. Khan, and M. S. Shur, “III-nitride based UV light emitting diodes,” this volume. M. Hatcher, “Sandia UV LEDs emit record power,” Compound Semiconductor 20, November 2003. D. Morita, M. Sano, M. Yamamoto, M. Nonaka, K. Yasutomo, K. Akaishi, S. Nagahama, and T. Mukai, “Over 200 mW on 365 nm ultraviolet light emitting diode of GaN-free structure,” Phys. Stats. Sol. A 200, pp. 114–117 (2003). M. S. Shur and M. A. Khan, “GaN/AlGaN heterostructure devices: Photodetectors and field effect transistors,” MRS Bull. 22, pp. 44–50 (1997). M. A. Khan, J. Kuznia, D. T. Olson, M. Blasingame, and A. R. Bhattarai, “Schottky barrier photodetector based on Mg-doped p-type GaN films,” Appl. Phys. Lett. 63, pp. 2455–2456 (1993). M. A. Khan, Q. Chen, C. J. Sun, M. S. Shur, M. F. Macmillan, R. P. Devaty, and J. Choyke, “Optoelectronic devices based on GaN, AlGaN, InGaN homoheterojunctions and superlattices,” Proc. SPIE 2397, pp. 283–293 (1995).

12

M. S. Shur and A. Zukauskas 16. A. Osinsky, S. Gangopadhyay, B. W. Lim, M. Z. Anwar, M. A. Khan, D. V. Kuksenkov, and H. Temkin, “Schottky barrier photodetectors based on AlGaN,” Appl. Phys. Lett. 72, pp. 742–744 (1998). 17. D. Walker, V. Kumar, K. Mi, P. Sandvik, P. Kung, X. H. Zhang, and M. Razeghi, “Solar-blind AlGaN photodiodes with very low cutoff wavelength,” Appl. Phys. Lett. 76, pp. 403–405 (2000). 18. S. C. Jain, M. Willander, J. Narayan, and R. Van Overstraeten, “III–nitrides: Growth, characterization, and properties,” J. Appl. Phys. 87, pp. 965–1006 (2000). 19. G. B. Parish, S. Keller, P. Kozodoy, J. P. Ibbetson, H. Marchand, P. T. Fini, S. B. Fleischer, S. P. DenBaars, U. K. Mishra, and E. J. Tarsa, “High-performance (Al,Ga)N-based solar-blind ultraviolet p–i–n detectors on laterally epitaxially overgrown GaN,” Appl. Phys. Lett. 75, pp. 247–249 (1999). 20. E. J. Tarsa, P. Kozodoy, J. Ibbetson, B. P. Keller, G. Parish, and U. Mishra, “Solarblind AlGaN-based inverted heterostructure photodiodes,” Appl. Phys. Lett. 77, pp. 316–318 (2000). 21. T. Li, D. J. H. Lambert, A. L. Beck, C. J. Collins, B. Yang, M. H. Wong, U. Chowdhury, R. D. Dupuis, and J. C. Campbell, “Solar-blind AlxGa1–xN-based metal–semiconductor–metal ultraviolet photodetectors,” Electron. Lett. 36, pp. 1581–1583 (2000). 22. V. Adivarahan, G. Simin, J. W. Yang, A. Lunev, M. Asif Khan, N. Pala, M. Shur, and R. Gaska, “SiO2-passivated lateral-geometry GaN transparent Schottky-barrier detectors,” Appl. Phys. Lett. 77, pp. 863–865 (2000). 23. J. C. Carrano, T. Li, P. A. Grudowski, C. J. Eiting, R. D. Dupuis, and J. C. Campbell, “Current transport mechanisms in GaN-based metal– semiconductor–metal photodetectors,” Appl. Phys. Lett. 72, pp. 542–544 (1998). 24. V. Adivarahan, G. Simin, G. Tamulaitis, R. Srinivasan, J. Yang, M. Asif Khan, M. S. Shur, R. Gaska, S. L. Rumyantsev, and N. Pala, “Indium–silicon co-doping of high aluminum content AlGaN for solar blind photodetectors,” Appl. Phys. Lett. 79, pp. 1903–1905 (2001). 25. G. Tamulaitis, K. Kazlauskas, S. Juršơnas, A. Žukauskas, M. A. Khan, J. W. Yang, J. Zhang, G. Simin, R. Gaska, and M. S. Shur, “Optical bandgap formation in AlInGaN alloys,” Appl. Phys. Lett. 77, pp. 2136–2138 (2000). 26. Stampfl and C. G. Van de Walle, “Doping of AlxGa1–xN,” Appl. Phys. Lett. 72, pp. 459–461 (1998). 27. S. L. Rumyantsev, N. Pala, M. S. Shur, R. Gaska, M. E. Levinshtein, M. Asif Khan, G. Simin, X. Hu, and J. Yang, “Low frequency noise in Al0.4Ga0.60N based Schottky barrier photodetectors,” Appl. Phys. Lett. 79, pp. 866–868 (2001). 28. A. Osinsky, S. Gangopadhyay J. W. Yang, R. Gaska, D. Kuksenkov, H. Temkin, I. K. Shmagin, Y. C. Chang, J. F. Muth, and R. M. Kolbas, “Visible-blind GaN Schottky barrier detectors grown on Si(111),” Appl. Phys. Lett. 72, pp. 551–553 (1998). 29. J. M. Van Hove, P. P. Chow, R. Hickman, II, J. J. Klaassen, A. M. Wowchak, and C. J. Polley, “GaN and AlGaN photodetectors for high temperature sensing applications,” in Abstracts of Materials Research Society Conference (December 1997, Boston, MA), D19.5. 30. M. A. Khan, M. Shur, and Q. Chen, “High transconductance AlGaN/GaN optoelectronic heterostructure field effect transistor,” Electron. Lett. 31, pp. 2130–2131 (1995).

Basic Device Issues in UV Solid-State Emitters and Detectors

13

31. M. A. Khan, M. S. Shur, Q. Chen, J. N. Kuznia, and C. J. Sun, “Gated photodetector based on GaN/AlGaN heterostructure field effect transistor,” Electron. Lett. 31, pp. 398–400 (1995). 32. D. Ciplys, R. Rimeika, M. S. Shur, R. Gaska, A. Sereika, J. Yang, and M. Asif Khan, “Radio frequency response of GaN-based SAW oscillator to UV illumination by the Sun and man-made source,” Electron. Lett. 38, pp. 134–135 (2002). 33. D. Ciplys, R. Rimeika, M. S. Shur, S. Rumyantsev, R. Gaska, A. Sereika, J. Yang, and M. Asif Khan, “Visible-blind photoresponse of GaN-based surface acoustic wave oscillator,” Appl. Phys. Lett. 80, pp. 2020–2022 (2002). 34. D. Ciplys, A. Sereika, R. Rimeika, R. Gaska, M. S. Shur, J. Yang, M. A. Khan, “IIInitride based ultraviolet surface acoustic wave sensors,” this volume. 35. M. Razeghi and A. Rogalski, “Semiconductor ultraviolet detectors,” J. Appl. Phys. 79, pp. 7433–7473 (1996). 36. M. S. Shur and M. A. Khan, “GaN and AlGaN devices: Field effect transistors and photodetectors,” in GaN and Related materials II, ed. by S. J. Pearton, Optoelectronic Properties of Semiconductors and Superlattices, Vol. 7 (Gordon and Breach Science Publishers, Amsterdam, 2000), pp. 47–92. 37. H. Morkoç, A. Di Carlo, and R. Cingolani, “GaN-based modulation doped FETs and UV detectors,” in Condensed Matter News, ed. by Patrick Bernier, Vol. 8, issue 2, pp. 4–46 (2001). 38. H. Morkoç, “Wurtzite GaN based modulation doped FETs and UV detectors,” in Handbook of Thin Film Devices: Hetero-Structures for High Performance Devices, ed. by M. H. Francombe, Chapter 5 (Academic Press, San Diego, 2000), pp. 193– 216.

HVPE-GROWN AlN-GaN BASED STRUCTURES FOR UV SPECTRAL REGION A. S. USIKOV 1, Yu. MELNIK 1, A. I. PECHNIKOV 1, V. A. SOUKHOVEEV 1, O. V. KOVALENKOV 1, E. SHAPOVALOVA 1, S.Yu. KARPOV 2, and V. A. DMITRIEV 1 1

TDI, Inc., 12214 Plum Orchard Dr., Silver Spring, MD 20904 Soft-Impact, Ltd., P.O. Box 83, 27 Engels av., St. Petersburg, 194156 Russia E-mail: [email protected], Phone: +1 (301) 572 7834, Fax: +1 (301) 572 6438 2

Abstract:

In this paper we describe ultraviolet light emitting diodes (LEDs) emitting in the spectral range from 305 to 340 nm based on AlGaN/AlGaN multi-layer submicron heterostructures grown by hydride vapor phase epitaxy (HVPE). The developed HVPE process possesses unique features such as ability (i) to combine deposition of thick low-defect layers and thin device multi-layer structures in the same growth run and (ii) to easily grow high-quality AlGaN layers in the whole composition range. HVPE is carbon-free growth technique producing GaN materials with very low background impurity concentrations. For a packaged LED with the peak wavelength of 340 nm, an optical output power of 2 mW was achieved at pulsed injection currents of 110 mA. The obtained results prove the developed HVPE technique to have a significant potential for production of device epitaxial wafers, particularly for fabrication AlGaN-based light emitters.

Key words:

HVPE, AlGaN/GaN heterostructures, UV LED

1.

INTRODUCTION

Group-III nitride semiconductors (AlGaN) with high aluminum content could be developed as light emitters operating in the ultra-violet (UV) spectral range (350–220 nm). High performance AlGaN-based UV light emitters can find a lot of applications in UV optoelectronics for military, industrial and medicine needs. These devices would allow the implementation of miniaturized and inexpensive system for biological agent detection. Being 15 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 15–29. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

16

A. Usikov et al

realized through fluorescence excitation of 1–5 Pm diameter sampled particles by UV light (emission wavelength should be shorter than 340 nm), biodetection of vanishing concentration of dangerous biological agents will be effective and reliable with high optical power of the UV light emitters. In addition, these devices can operate in the solar blind region of the spectrum (240–280 nm) where the earth’s atmosphere is opaque. A particular interest is in high efficiency light emitters for 280–360 nm wavelength range. These devices would enable satellite communication secure from the ground and non-line-of-sight covert communication. Underwater submarine communication would also be possible. Civilian applications include medicine and solid-state lighting. UV emission could be converted into white light utilizing luminescent polymers and the resulting white light sources can be much more efficient, compact, and rugged than conventional fluorescent lamps. To date, the main technological method to fabricate UV light emitters based on AlGaN materials is metal organic chemical vapor deposition (MOCVD). This method is proven to be a reliable fabrication tool for blue/green GaInN/GaN-based light-emitting diodes (LEDs) [1,2] and violet laser diodes (LDs) [3,4]. The LDs with GaN active region emitting at 366 nm under pulsed current injection were also fabricated by MOCVD [3]. To shift to the shorter wavelength range (

1E+20

1E+07

Ga->

1E+06

1E+19

1E+05

1E+18

Eg). In-plane resolution is limited by the number of grating lines in the excited spot and vary from 50 Pm to 1 mm. Development of the FWM technique requires specific knowledge of dynamic holography, nonlinear optics, and semiconductor physics. Table I lists the quantitative relationships, which bridge the light-induced modulation of optical properties and diffraction efficiency with optical and electrical parameters of a semiconductor. From the point of view of materials research, the most important part in this loop is the set of equations, which join quantitatively the generation, recombination, and transport processes with the measured light diffraction characteristics, i.e. grating kinetics at various peTable I. Optical nonlinearity: a bridge between the optical and electrical parameters. Field What it provides Dynamic x A configuration to create and monitor spatial complex refractive holography index modulation n*=n+ik: n(x,t) = n0+'n(t)cos(Kx) k(x,t) = k0+'k(t)cos(Kx). x A type of diffraction grating: phase, amplitude, or mixed one. x Relationships between the diffraction efficiency K = I1/IT and modulation amplitude of 'n or 'k: K = (S'nd/O)2 + (S'kd/O)2 Nonlinear Optics x Modulation mechanisms and their coefficients 'nEO = neo ESC. 'nFC = neh 'Ne,h Semiconductor x Nonequilibrium carrier density 'Ne,h or SC field ESC and their Physics spatio-temporal evolution, governed by carrier generationtransport-recombination: i.e. by optical (D, E) and electric parameters (D, WR, S). Quantitative relationships follow from the solution of a set of continuity and current density equations.

K. Jarasiunas

96

riods and excitation intensities, dependence of diffraction efficiency on excitation intensity. It goes without saying that the a priori knowledge of the dominant refractive index modulation mechanism in the conditions of experiment is of the utmost importance. As a rule in the spectral region below the band gap, the free-carrier nonlinearity is dominant with the refractive index modulation 'nfc prevailing over light-induced absorption index change 'kfc = 'DfcO/4S < 'nfc. The refractive index modulation by free carriers 'nfc = neh 'Ne,h is given by the well-known Drude- Lorentz model with the refractive index modulation coefficient by one electron-hole pair, neh , [3,4,6] 'nfc(x,t) = (–e2 /2nH0Z2)[Zg2/(Zg2 – Z2)] u u ('Ne(x,t)/me + 'Nh(x,t)/mh).

(1)

It is seen that the electrons dominate in refractive index modulation, even at bipolar carrier generation, because me < mh. The value of refractive index modulation neh varies in the range of (4–7)u10–21 cm–3 (for CdTe, GaAs, GaN) at the wavelength of 1.06 Pm of a probe beam. This coefficient together with the thickness of excited region d determines the sensitivity of the technique, i.e. the minimum value of the product ('nd), which can be detected in diffraction. Usually, the signal/noise ratio of | 1 is reached at the diffraction efficiency of K # 10–5, what corresponds to ('nd) t 0.001 cm, and this limit corresponds to the lowest carrier density of 1015 cm–3 (in a few mm thick bulk crystals) or 1018 cm–3 (in a 1-Pm thick epitaxial layer) which can be assessed by the FWM technique.

3.

CONFIGURATIONS AND INSTRUMENTATION OF TIME-RESOLVED FWM

Two following figures (Figs. 2 and 3) present configurations of optical schemes and instrumentation, used for free carrier grating recording and probing in bulk crystals and in layered structures, correspondingly. These schemes present two common configurations of FWM interaction, involving two combinations of four waves. The case of degenerate FWM (Fig. 2) involves four waves whose wavelengths are equal: the forward s-polarized waves record a grating in the bulk and the backward p-polarized delayed probe wave monitors the grating decay. The diffracted wave counterpropagates to one of the recording beams at the Bragg-matched angle and is extracted by a Glan prism. The scheme allows extraction of all the diffracted beam intensity, which may reach 100% of the probe beam for light diffrac-

Characterization of Advanced Materials for Optoelectronics

97

Figure 2. Scheme of a four-wave mixing setup. The elements shown are: 1. Sample; 2. Translation stages; 3. Glan prism; 4. Delay time of the probe beam; 5. Shutter; 6. Attenuator; 7. Detectors; 8. Beam splitters; 9. Dielectric mirrors.

tion on a thick grating. The DFWM is used to study cases of weak interaction in bulk crystals, as deep impurity-related carrier generation in semiinsulating materials or two-photon interband transitions. If a sample is thin and the grating period is large enough to satisfy condition that the factor Q = 2SOd/(n/2) d 1, then the regime of light self-diffraction takes place, and the diffracted pump beams can be observed in the far field of diffraction (the latter case was used to monitor homogeneity of defect distribution in GaAs wafers [7]). The second arrangement is used for investigation of highly absorbing thin samples (epitaxial layers, heterostructures, multiple quantum well structures). Here the beams at wavelengths above the Eg are used to record the free-carrier grating, while the probe beam wavelength is in the transparency region of a sample. The carrier grating is created at the very surface with the excitation depth deff being close to the inverse absorption coefficient. The further bipolar diffusion of carriers to the sample depth significantly expands

Figure 3. Configuration for a surface grating recording.

K. Jarasiunas

98

the excited area in bulk crystals, while a potential barrier in heterostructures may localize carriers in the front layer. A delayed probe beam monitors the electron–hole density as a function of time via the time-varying diffraction efficiency of the grating. This configuration has been applied initially to determine surface recombination velocities in GaAs and InP crystals [8]. The intensities of recording (I0), probing (Ip), transmitted probe (IpT) and diffracted (I1) beams are controlled by Si photodiodes. The data acquisition system measures the instantaneous diffraction efficiency of the grating K(t) = (I1 – Is)/IpT (the scattered background signal Is is extracted from the diffracted beam) and the transmission of the sample T(t) = IpT/Ip. The system measures all the signals, and the LabView system presents the kinetics in a required intensity window. The FWM technique requires coherence of the laser beams that record the dynamic grating. In our experiments, we used a picosecond mode-locked Nd:YAG laser model PL-2143 (Ekspla Co.), which also emits the second (532 nm) and third (355 nm) harmonics, used for recording “surface” gratings in III-V compounds and nitrides, respectively. A parametric generator PG-401 (Expla Co.) emitting in the spectral range from 420 to 2000 nm was also used to record gratings in II-VI heterostructures and the grating decay was monitored by the delayed beam of PL-2143 at 1064 nm. The grating decay kinetics is measured at various grating periods to plot an “angular” dependence of the grating decay time WG versus grating period. The decay time of the grating efficiency K(t) ~ exp(–2t/WG) corresponds to the time interval t = WG,in which the K value decreases by e2 times, while the carrier modulation decreases by e times in this time interval. The plot in a form 1/WG vs. (2S//)2 allows separation of diffusion processes from carrier recombination, as the inverse of WG(t) is the sum of two decay process of spatial carrier modulation: 1/WG = 1/WR + 1/WD

(2)

where WR is the recombination time and WD = ȁ2/(4ʌ2D) is the diffusion time of the grating erasure. A slope of the plot provides D value, while the intersection with the ordinate axis yields the 1/WR value. An “exposure” characteristic of diffraction is the dependence of diffraction efficiency on excitation beam intensity, K(I0), which is also called a “lux-diffraction” (LDCh) characteristic (like “lux-ampere” characteristic in measurements of photoconductivity). The slope of a LDCh in log–log plot, J, reveals carrier generation rate vs. excitation: for linear carrier generation J = 2, while at two-photon absorption J = 4. At deep-trap assisted carrier generation, the value of J may vary from J = 2 to J < 1, indicating an ex-

Characterization of Advanced Materials for Optoelectronics

99

hausting of the traps with excitation, or increase up to J = 4 at high excitations, when two-step transitions via impurity states dominate [6]. The one more characteristic is a mapping of a wafer, which provides planar distribution of defects responsible for carrier generation or recombination. For mapping, the sample is positioned on a translation stage and inplane variation of the diffraction efficiency K(x,y) is measured while other parameters (the excitation energy, grating period, and delay time of probe beam) are fixed. The mapping of commercial 3-inch GaAs wafers in regime of self-diffraction revealed W-shape distribution of deep donor EL2, which correlated well with the dislocation density [7].

4.

SAMPLES

We investigated various heterostructures by FWM technique with an attempt to reveal peculiarities of electrical properties in the epitaxial layers and in the vicinity of interfaces, which are very sensitive to lattice mismatch and related changes in structural/optical properties. Here we present some results of recent studies carried out in MOCVD-grown GaN layers on sapphire, SiC, or Si substrates, InGaN/GaN, as well MOVPE and hydrogen-transport-VPE grown II-VI heterostructures CdTe/GaAs, and in heavily C-doped GaAs layers embedded in a double-heterostructure (DHS). In bulk crystals, the application of FWM technique for investigation of electrical activity of deep defects and modification of their charge state by co-doping as well as related carrier transport features were the objects of recent studies in semi-insulating II-VI compounds [6,7,9–11]. Defect related properties in semi-insulating bulk crystals of GaAs:EL2, CdTe and CdZnTe, doped by deep vanadium impurity and co-doped by shallow donors or acceptors (Cl,As) as well as in proton-irradiated GaAs:Cr crystals allowed studies of impurity-related carrier generation, determination of carrier lifetimes, type of dominant carriers, photoexcited from deep traps, density of residual defects, and control uniformity of commercial GaAs and CdTe wafers. The studies of bulk SiC and GaN wafers grown for the electronics and optoelectronics industry are in progress.

5.

EXPERIMENTAL RESULTS AND DISCUSSION

5.1

GaN-Based Heterostructures

The experiments were performed on a 2.6 Pm undoped GaN sample grown at 1075 oC by MOCVD on 25-nm thick GaN buffer layer, deposited at low

K. Jarasiunas

100

S (cm/s): 1x104 5x104 1x106

0.1

10.5 0.01

b) I = 0.4 mJ cm -2 / = 10.5 P m

0 ps 1019

Carrier density, cm -3

Diffraction efficiency, a.u.

a) 1

100 ps 300 ps

1018

1 ns 1017

5.9 3.1 0

500

3.6 1000

0.0

0.5

1.0

1.5

z, P m

Delay time, ps

Figure 4. Grating decay kinetics in 2.6 Pm-thick GaN epilayer on sapphire at various grating periods, fitted by numerical calculations using Eq. 3 (a) and carrier in-depth profiles for various probing times (b).

temperature (450 oC) on sapphire (0001) substrate [12]. The thickness of the buffer layer was optimized to reduce the residual strain in the GaN epilayer. The InxGa1–xN samples used in this study were grown on a basal-plane (0001) sapphire substrate by MOCVD [13]. Prior to the InxGa1–xN alloy layer, a 1000-nm thick GaN epilayer was deposited at 980 oC under a pressure of 76 Torr and served as a buffer layer. All InGaN samples were nominally undoped. The thickness of InxGa1–xN epitaxial layers was 50 nm. The averaged In fraction x was estimated by X-ray diffraction measurements. In this study, we show the results of three representative samples with In content varying from 8 to 15%. The transient free carrier gratings were recorded by exciting the front layer of the structure by two coherent 25-ps duration laser pulses at 355 nm of the Nd:YAG laser directed to the sample at a certain angle, which could be varied from 2 to 7 degrees. This allowed us to change the dynamic grating period from 3.1 to 10.5 Pm. We probed this free carrier grating by the delayed probe laser pulse at 1064 nm through measuring the intensity of its first diffraction order. For comparison, we also measured carrier dynamics the same way in the vicinity of internal buffer-epilayer interface excited through the sapphire. Figure 4a shows the grating decay kinetics measured in the 2.6-Pm thick GaN epilayer at various grating periods. A typical behavior of a surface grating in bulk crystals [8,14], a very fast initial decay that at later times evolves into a slower exponential process, was observed. The decay indicated contri-

Characterization of Advanced Materials for Optoelectronics

101

bution of the surface and nonlinear recombination in the initial stage followed by a linear recombination and diffusion at later times ('t t 500 ps). Using the WG values extracted from the tail parts of kinetics, we plotted a dependence of 1/WG vs. (2S//)2 and determined the bipolar diffusion coefficient Da = 1.7 cm2/s and effective carrier lifetime Weff = 670 ps (the effective time accounts for all channels of recombination; at given conditions we assume that both linear and surface recombination contribute to its value). The Da value allows estimation of hole diffusion coefficient Dh | Da/2, hole mobility Ph = 32 cm2/V·s, and hole diffusion length Lh = 0.28 Pm. The surface recombination velocity was evaluated by solving twodimensional continuity equation

w N ( x, z, t ) G( x, z, t )  ’[ D( N )’N ( x, z, t )] wt 

N ( x, z, t )

WR

(3)

 BN 2 ( x, z, t )

with the boundary conditions for a semi-infinite media, as the carrier diffusion length is much less than the epilayer thickness:

wN ( x, z, t ) wz z

0

S N ( x,0, t ), N ( x, f, t ) D( N )

0,

(4)

where G(x,z,t) is the carrier generation rate by light interference pattern, B is the bimolecular carrier recombination rate, and S is the surface recombination velocity. Using the determined value of Da = 1.7 cm2/s and the coefficient B = 4.7u10–11 cm3/s at 300 K, we fitted the numerical calculations to the experimentally measured set of kinetics (Fig. 4a), and determined the values of S = 5u104 cm/s and linear recombination time of WR = 950 ps [15]. The numerical simulation of carrier distribution within the epilayer is also useful in determining the effective thickness of the photoexcited region and evaluation of in-depth resolution of the technique. In Fig. 4b, we present the calculated in-depth profiles of carrier density for various probing times, using the values of D, S, and WR determined above. It is seen that diffusion into the bulk and surface recombination rapidly dilute the carrier plasma generated initially in the depth deff close to the inverse absorption coefficient 1/D = 130 nm. The deff value increases with time up to 1 Pm, and the carrier density simultaneously decreases to a value of (0.5–1)u1018 cm–3, thus allowing results being independent on nonlinear recombination. Comparison of the determined D and WR values of the given GaN epilayer with published ones have shown the high quality of the used layer. The stud-

K. Jarasiunas

102

Diffra ction efficiency (a.u.)

a) Excitation: InGaN (layer) GaN (substrate )

b)

470 ps ~ 50 p s

0.0

0.5 1.0 Probe delay time (ns )

Figure 5. Carrier dynamics in InGaN/GaN structure and in the buffer layer of GaN (a) and the angular dependence of grating decay times (b).

ies of 7-Pm thick GaN epilayer at its homogeneous in-depth excitation by 532 nm a provided trap-assisted carrier recombination lifetime of 100 ps and 1.1 ns in a bulk [16]. The determined hole diffusion length of 0.28 Pm at a 1Pm distance from the interface is comparable with that of Lh = 0.25 Pm at a distance of ~8 Pm from the GaN/sapphire interface for thick HVPE-grown quasi-bulk n-GaN samples [17]. The carrier dynamics in MOCVD grown ~1-Pm thick GaN/sapphire layers exhibited a relaxation time of ~50 ps and bipolar diffusion coefficient of Da d 0.16 cm2/s [18], thus confirming essential contribution of dislocations as centers of nonradiative recombination in GaN [19]. The 50-nm thick InxGa1–xN epitaxial layers present a case of very thin single layers, in which in-plane diffusion and recombination of carriers have been studied by FWM technique at 300 K. At excitation by UV beams at 355 nm, the photoexited carriers are confined in the front layer, since the InGaN/GaN interface presents potential barriers both for electrons and holes. Nevertheless, the incident light reaches the 1-Pm thick GaN buffer and creates free carriers, which also contribute to light diffraction. More detailed analysis of the carrier spatial distribution and its evolution, using Eq. (3), allowed evaluation of the contribution of InGaN and GaN layers to the diffraction signal [20]. Their initial contribution to diffraction signal, which is proportional to the squared carrier density, equals to the ratio of about 2:1. With time, the contribution of GaN to diffraction vanished rapidly due to essentially faster carrier decay in the buffer than in the InGaN layer (Fig. 5a). This simplified the analysis of diffraction kinetics, as at delay time 't > 100–150 ps the diffraction is solely determined by the carrier modulation dynamics merely in the InGaN layer. The grating decay kinetics, in the range of / from 3 to 10.5 Pm allowed determination of the bipolar diffusion

Characterization of Advanced Materials for Optoelectronics

Diffraction efficiency (a.u.)

10 0

I II

1.6

10 -1

III

103

Content of In: I8% II10% III15% IVGaN/sapphire interface

2.2 1.7

IV 1 Excitation (m J/cm 2 )

Figure 6. Exposure characteristics in InGaN layers with different In content and in the GaN/sapphire interface.

coefficient Da = 2.1r0.2 cm2/s and the effective recombination time WR eff = 480 ps for highly excited InGaN layer (x = 8%), see Fig.5b. Simulation of carrier dynamics in the 50-nm thick InGaN layer [20] has shown that the nonexponential part of decay that is present on experimental curves of InGaN/GaN (see, e.g. Fig. 5a) describes nonlinear processes in the front layer and can be attributed to radiative bimolecular recombination of carriers in InGaN. This conclusion was strongly supported by the experimental data that the non-exponential part in grating decay was becoming faster with excitation (in the range from 0.4 to 1 mJ/cm2) as well as with increasing In content due to carrier localization effects, which stimulate processes of radiative recombination [21,22]. The dependence of diffraction efficiency vs. excitation intensity sensitively revealed the carrier density dependent changes of the recombination rate in InGaN and GaN layers. Nearly linear carrier recombination was found at the excitation below 1 mJ/cm2, while at higher excitations the characteristic exhibited a tendency for strong saturation of diffraction efficiency (Fig. 6). We attribute the latter behavior to a decrease of carrier lifetime to values shorter than the laser pulse duration due to reaching the threshold of radiative recombination, which prevails over other recombination channels at high carrier densities. Figure 6 shows that this threshold is reached most easily in the layer with 10% of In content, thus indicating the optimal In content to localize carriers. The excitation of internal GaN/sapphire interface shows that carrier capture at dislocations dominate with short enough carrier lifetime, thus no signature of stimulated emission is present (see curve IV in Fig. 6). The correlation between the grating decay time and the saturation efficiency was also observed in GaN/Si layers: the longer was the carrier lifetime, the more pronounced was saturation of the diffraction efficiency.

K. Jarasiunas

104

Deeper analysis of carrier recombination mechanisms vs. excitation requires studies of temporally and spectrally resolved photoluminescence.

5.2

Heavily Doped Double GaAs:C Heterostructures

Investigation of carrier dynamics at extremely high doping concentrations (>1019 cm–3) by time-resolved FWM is demonstrated below. Carrier transport and recombination have been investigated in heavily carbon-doped GaAs layer (p0 = (1–2)u1019 cm–3), embedded in a double-heterostructure. The carriers were injected into the 1-Pm thick p-GaAs layer sandwiched between 50-nm thick AlGaAs:C (p0 = 1018 cm–3) or InGaP:Si layers, using light interference pattern of two picosecond laser pulses at 532 nm. Due to the density gradient along the z-axis, carriers diffuse away from the surface and become confined in the GaAs:C layer. The measurements were carried out at 300 K at the excitation intensity which corresponds to nonequilibrium carrier density of about 1018 cm–3 in the GaAs layer. The carrier dynamics in heavily-doped DHS (Fig. 7) was found very different from that in the bulk GaAs crystals. The fast component nearly followed the laser pulse temporal shape in both samples but was more pronounced in the DHS with the AlGaAs:C barriers. After this fast decay, a slower exponential relaxation with a time constant of ~1 ns was dominant in both samples. Assuming that the latter decay component resembles the diffusion and recombination of carries confined in the quantum wells [11], we measured its decay time constant WG at various values of / and determined the in-plane diffusion coefficient D as well as the carrier lifetime WR in the pdoped GaAs:C layers (Fig. 7b): D = 35 cm2/s and WR = 1.5 ns for the layer with p0 = 2u1019 cm–3, and D = 27 cm2/s, WR = 2 ns for another one with p0 = 1u1019 cm–3. The values of D correspond to minority electron diffusion, as the measb)

Diffraction efficiency (a.u.)

10 1 Grating period / : 7.1 P m 9.5 P m 19.1 P m

10 0

1.1 ns 10 -1

3 Wg-1 (ns -1 )

a)

2

p = 1x10 19 cm -3 p = 2x10 19 cm -3

D = 35 cm 2 /s WR = 1.5 ns D = 27 cm 2 /s WR = 2.0 ns

400 ps 1

300 ps

10 -2 0

250

500

0.2

0.4 0.6 4 S 2/ / ( P m -2)

0.8

Probe delay (ps)

Figure 7. Grating decay kinetics in GaAs:C heterostructure with AlGaAs barriers (a) and the determined carrier plasma parameters in the heavily doped p-GaAs layers.

Characterization of Advanced Materials for Optoelectronics

105

urement regime ensured the condition that the doping density was essentially higher than that of photoexcited electrons: p0 >> 'Ne. The corresponding electron mobilities in heavily doped GaAs layers, estimated from the measured D values, are of about 1100 cm2/V·s and 1450 cm2/V·s. The increase of mobility in the heavier doped layer (with p0 = 2u1019 cm–3) is in quantitative agreement with carrier scattering features in heavily-doped GaAs: at the hole density of 2u1019 cm–3, the Fermi energy is 60 meV in the valance band and many states for holes are occupied. This effect of degeneracy effectively reduces the electron–hole scattering according to Pauli exclusion principle, and Monte Carlo calculations predict an increase of minority electron mobility in highly doped p-GaAs at p > 1019 cm–3 [23,24]. The performed FWM studies complementary with time-resolved PL decay time measurements in heavily doped layers [25] will provide access to minority carrier lifetimes vs. doping. More detailed evaluation of carrier capture in the front barrier and a role of interface is in progress [26].

5.3

CdTe and ZnTe based heterostructures

Similarly to given above studies of GaAs heterostructures, we investigated II-VI hereostructures grown by different techniques. In Fig. 8, carrier dynamics in MOVPE-grown ~1-Pm thick CdTe on ZnTe/GaAs buffer and in 36-Pm thick HVPE-grown CdTe epilayer on GaAs are compared. FC grating kinetics in the thin epilayer revealed 870-ps decay times [14], while the 36Pm thick H2T-grown layers exhibited very fast decay followed by a weak relaxation tail of 300 ps time [27]. The latter decay was found very similar to

I 870 ps

10 -1 II I 1 (a.u.)

Diffraction signal (a.u.)

10 0

270 ps

10 -2

slope J = 1.6; N ~ I 0.5

III

10 -3

Excitation energy (a.u.) 0

300 Delay time (ps)

600

Figure 8. Free carrier grating kinetics in as-grown (I) and aged (III) CdTe/ZnTe/GaAs heterostructure and in HVPE-grown thick CdTe layer (II). The inset shows an exposure characteristic of diffraction in the thick layer.

106

K. Jarasiunas

dynamics in the aged 1.6-Pm thick CdTe/ZnTe/GaAs layer, where migration of defects from the strained interface towards the epilayer surface caused the epilayer degradation: the carrier lifetime in the aged layer decreased to 60 ps. High number of nonradiative recombination centers in as-grown HVPE layers leads to very fast carrier recombination at dislocations, which attract both carriers and also cause band-gap modulation by dislocation stress and strain fields [28]. A slope J = 1.6 of exposure characteristic corresponds to sublinear carrier density dependence on excitation (N ~ I1/2, see inset of Fig. 8) and confirms the bimolecular carrier recombination at dislocations in the HVPEgrown ZnTe epilayer. Recent studies of homoepitaxial ZnTe layers, grown on ZnTe:P substrates, have been carried out using FWM with wavelength-tunable picosecond pulses and revealed different recombination features [29]: the decrease of carrier lifetime with excitation in the epilayers, while the increase of WR at the surface of the substrate was caused by saturation of residual traps. This tendency, together with ns-duration lifetimes showed by the epilayers, can be clearly ascribed to a dominant bimolecular recombination rate and a low density of trapping defects in the epilayers. The high value of D = 11 cm2/s in ZnTe epilayer was very close to that in the substrate, pointing out towards an occurrence of similar density (~1018 cm–3) of ionized impurities in both epitaxial and bulk crystals.

6.

HOLO-DEVICES FOR NONDESTRUCTIVE CONTROL OF SEMICONDUCTORS

There is a constant need of novel techniques able to control the material electrical parameters, in spite that commercially available devices (e.g. TEM, AFM, XRD, SIMS, EPR, FTIR) allow control of a semiconductor material morphology, defect density, strain, density of impurities, their charge state, etc. Studies of photoelectrical properties under laser excitation in Vilnius University resulted in development of light-induced transient grating technique and its applications for nondestructive studies of technologically important materials, as Si, GaAs, CdTe, etc. We foresee useful applications of the time-resolved FWM technique for control of wide bang gap materials, both bulk wafers and heterostructures. Recently, we implemented the picosecond FWM technique into devices and assembled (jointly with Ekspla Co, www.ekspla.com) HOLO modules, which can be attached to a pulsed laser source. The novel devices (Fig. 9) are able to control bulk materials or heterostructures, using one or two fixed wavelengths and advanced data acquisition system.

Characterization of Advanced Materials for Optoelectronics

107

Figure 9. A pilot device HOLO-1 (Ekspla Co.) for nondestructive control of bulk crystals, based on DFWM at 1.06 Pm.

In conclusion, the feasibility of picosecond FWM technique on freecarrier gratings for studies of carrier dynamics in heterostructures has been demonstrated. Carrier diffusion and recombination in differently grown GaN, InGaN, heavily doped GaAs:C and CdTe-based heterostructures or ZnTe homoepitaxial structures have been investigated and a number of carrier parameters were determined. The studies allowed implementation of the FWM technique into novel HOLO devices (Ekspla Co.) for nondestructive control and metrology of nonequilibrium carrier parameters in bulk crystals and layered structures.

ACKNOWLEDGEMENTS The research was sponsored by NATO’s Scientific Affairs Division in the framework of the Science for Peace Programme (Project SfP-974476), European Commission (Contract No. G5MA-CT-2002-04047), and Lithuanian State Science and Studies Foundation. The author would like to thank co-workers Dr. M. Sudzius, Dr. V. Gudelis, R. Aleksiejunas, T. Malinauskas for substantial contribution, and acknowledges the cooperation of colleagues from Rensselaer Polytechnic Institute, Troy, USA (Prof. M. S. Shur), Sensor Electronic Technology, Inc. (Dr. R. Gaska), Lecce University, Italy (Prof. N. Lovergine), University of Tokushima, Japan (Prof. S. Sakai), FerdinandBraun-Insitut für Höchstfrequentztechnik, Germany (Dr. M. Weyers), and Institute of Physics, Belorus (Prof. G. Yablonskii).

K. Jarasiunas

108

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

21. 22. 23. 24. 25.

J. Shah, Ultrafast Spectroscopy of Semiconductors and Semiconductor Nanostructures (Springer, Berlin, 1999). Nonlinear Optics in Semiconductors, I & II, ed. by E. Garmire and A. Kost, Semiconductors and Semimetals Vol. 59 (Acad. Press, New York, 1999). H. J. Eichler, P. Gunter, and D. Pohl, Light-Induced Dynamic Gratings, Springer Series in Optical Sciences Vol. 50 (Springer, Berlin, 1986). R. K. Jain and M. B. Klein, in Optical Phase Conjugation, ed. by R. A. Fisher (Acad. Press, New York, 1983), chap. 10. Special Issue on Dynamic Gratings and Four-Wave Mixing, IEEE J. Quant. Electr. QE-22, No. 8 (1986). M. Snjdžius, R. Aleksiejnjnas, K. Jarašinjnas, D. Verstraeten, and J. Launay. Semicond. Sci. Technol. 18, 367 (2003). J. Vaitkus, K. Jarašinjnas, E. Gaubas, and M. Petrauskas, Semicond. Sci. Technol. 7, A131 (1992). C. A. Hoffman, K. Jarašinjnas, H. J. Gerritsen, and A. Nurmikko, Appl. Phys. Lett. 33, 536 (1978). K. Jarašinjnas, L. Bastienơ, J. C. Launay, P. Delaye, G. Roosen, Semicond. Sci. Technol. 14, 48 (1999). K. Jarašinjnas and N. Lovergine, Mater. Sciece & Eng. B91-92, 100 (2002). K. Jarašinjnas, V. Mizeikis, S. Iwamoto, M. Nishioka, T. Someya, K. Fukutani, Y. Arakawa, T. Shimura, and K. Kuroda, Jpn. J. Appl. Phys. 39, 5781 (2000). T. Wang, T. Shirahama, H. B. Sun. H.X. Wang, J. Bai, S. Sakai, and H. Misawa, Appl. Phys. Lett. 76, 2220 (2000). E. Kuokstis, J. W. Yang, G. Simin, M. Asif Khan, R. Gaska, and M. S. Shur, Appl. Phys. Lett. 80, 977 (2002). V. Mizeikis, K. Jarašinjnas, N. Lovergine, and P. Prete, J. Cryst. Growth 214/215, 234 (2000). R. Aleksiejnjnas, M. Snjdžius, T. Malinauskas, J. Vaitkus, K. Jarašinjnas, and S. Sakai, Appl. Phys. Lett. 83, 1557 (2003). B. Taheri, J. Hays, J. J. Song, and B. Goldenberg, Appl. Phys. Lett. 68, 587 (1996). L. Chernyak, A. Osinsky, A. Schulte, Solid State Electron. 45, 1687 (2001). H. Haag, B. Honerlage, O. Briot, and R. L. Aulombard, Phys. Rev. B 60, 11624 (1999). T. Sugahara, H. Sato, M. Hao, Y. Naoi, S. Kurai, S. Tottori, K. Yamashita, K. Nishino, L.T. Romano, and S. Sakai, Jpn. J. Appl. Phys. 37, L398 (1998), Part 2. R. Aleksiejunas, V. Gudelis, M. Sudzius, K. Jarasiunas, Q. Fareed, R. Gaska, M. S. Shur, J. J. Zhang, J. Yang, E. Kuokstis, and M. A. Khan, Phys. Stat. Sol. (c) 0 (2003) (in press). E. Kuokstis, J. W. Yang, G. Simin, M. Asif Khan, R. Gaska, and M. S. Shur, Appl. Phys. Lett. 80, 977 (2002). L. H. Robins, A. J. Paul, C. A. Parker, J. C. Roberts, S. M. Bedair, E. L. Piner, N.A. El-Masry, MRS Internet J. Nitride Semicond. Res. 4S1, G3.22 (1999). T. Furuta and M. Tomizava, Appl. Phys. Lett. 56, 824 (1990). J.R. Lowney and H.S. Bennett, J. Appl. Phys. 69, 7102 (1991). A. Maasdorf, S. Gramlich, E. Richter, F. Brunner, M. Weyers, G. Trankle, J.W. Tomm, Y. I. Mazur, D. Nickel, V. Malyarchuk, T. Gunther, Ch. Lienau, A. Barwolff, and T. Elsaesser. J. Appl. Phys. 91, 5072 (2002).

Characterization of Advanced Materials for Optoelectronics

109

26. K. Jarasiunas, R. Aleksiejunas, V. Gudelis, M. Sudzius, A. Maaȕdorf, F. Brunner, and M. Weyers, Abstracts of 10th Int. Conf. on Defects DRIP-X (September 2003, Batz-sur Mer, France). 27. K. Jarašinjnas, E. Gaubas, R. Aleksiejnjnas, M. Snjdžius, V. Gudelis, T. Malinauskas, P. Prete, A.M. Mancini, and N. Lovergine, Phys. Stat. Sol. (a) 195, 238 (2003). 28. V. Kažukauskas, J. Storasta, and J. Vaitkus, J. Appl. Phys. 80, 2269 (1996). 29. R. Aleksiejunas, T. Malinauskas, M. Sudzius, K. Jarasiunas, N. Lovergine, M. Traversa, P. Prete, A. M. Mancini, and T. Asahi, Proc. 10th European Workshop on MOVPE (June 2003, Lecce, Italy), paper PS.II.06.

QUANTUM PHOSPORS Observation of the photon cascade emission process for Pr3+doped phosphors under vacuum ultraviolet (VUV) and X-ray excitation A.P. VINK 1,2, E. VAN DER KOLK 1, P. DORENBOS 1, and C.W.E. VAN EIJK 1 1

Radiation Technology Group, Interfaculty Reactor Institute, Delft University of Technology, Mekelweg 15, 2629 JB Delft, The Netherlands 2 Chemical Sciences, Netherlands Organisation for Scientific Research, P.O. Box 93470, 2509 AL The Hague, The Netherlands

Abstract:

In luminescent-tube lighting (TL), mercury is used to excite (Ȝmax = 254 nm) three phosphors, resulting in white light. The use of mercury however gives environmental problems and causes an undesired delay in lamp startup. If mercury is replaced by xenon, which is already gaseous at room temperature and harmless to the environment, both problems are solved. Xenon however emits at higher energy (Ȝmax = 172 nm) and the phosphors used in mercurybased tube lighting show a less efficient absorption to this vacuum ultra violet (VUV) radiation. Therefore, much effort is put in developing new phosphors. The Pr3+ (4f2) ion shows strong absorption in the VUV range, which can be assigned to the 4f2ĺ4f15d1 transition. Another interesting effect is that in some hosts the Pr3+ ion can, after excitation into the 4f15d1 bands, show a two-step relaxation to the ground state. This process is called photon cascade emission (PCE) or quantum cutting and could result in a quantum efficiency larger than 100%. The emitted photons are typically in the violet and green spectral region. This contribution describes the principles of Pr3+ quantum cutting and presents methods to select host materials in which Pr3+ shows quantum cutting. Furthermore, several Pr3+-doped hosts are presented, which show the PCE process both under VUV and host excitation.

Key words:

TL lighting, quantum cutting, photon cascade emission process, Pr3+, host materials 111

˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 111–126. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

A. P. Vink et al

112

1.

NEW GENERATION LIGHTING

For ages, mankind has sought for ways to make light when there is only darkness. These efforts resulted in the discovery of fire many thousands of years ago. More recently (in 1879) Thomas Alva Edison (1847–1931) invented the light bulb, which makes use of the blackbody radiation of hot carbon. Although the light bulb still is commonly used for lighting applications, the disadvantage is that it produces mainly heat instead of light. Just before the second world war, a new kind of lighting was invented with the main purpose to get more light than heat. In this type of lighting, emission originating from a mercury discharge is used to excite phosphor material, which is located on the inside (C in Fig. 1) of the glass tube (A). The emission from the mercury discharge is mainly at 254 nm. This type of lighting is called TL after the French name “tube luminescent”.

Figure 1. Schematic representation of TL lighting.

Phosphors of the early period (1938–1948) were based on mixtures of two phosphors (MgWO4 and (Zn,Be)2SiO4:Mn2+) to make white light. MgWO4 shows broad-band emission over the whole spectral region, whereas the Mn2+-doped silicate covers the green to red spectral region [1]. As the silicate material is not very stable under discharge conditions and the used beryllium is highly toxic, alternatives for these phosphors were investigated. The mix of two phosphors was replaced by a single phosphor, of which emission covers almost the whole visible spectral region. This phosphor was based on the halophosphate Ca(PO4)3F/Cl and was doubly doped with Sb3+ and Mn2+. The trivalent antimony emits around 500 nm. Part of the energy is however transferred to the Mn2+ ion, which emits around 600 nm. The manganese ion itself does not absorb the 254-nm light from the mercury

Quantum Phosphors

113

discharge. These halophosphate phosphors show high brightness and high color rendering (CRI of about 60) [1]. At the start of the seventies, a set of three phosphors was developed which generally not show broad-band emission, but narrow line emission. The commonly used phosphors are BaMgAl10O17:Eu2+ (blue emitting), (Ce,Gd)MgB5O10:Tb3+ (green emitting) and Y2O3:Eu3+ (red emitting). A mix of these three phosphors, emitting around 450, 550 and 610 nm respectively, results in white light with an even higher color rendering (CRI of 80–85) than the halophosphate phosphors [1]. All three commonly used phosphors are based on the lanthanides. The lanthanide group is located at high-atomic number end of Mendeleev’s periodic table and is characterized by a partially filled 4f-shell. The 4f-shell is an inner shell, which is shielded from its surrounding by the already-filled 5s2 and 5p6 shell. Therefore, the position of the energy levels of the lanthanides is mainly determined by the electrostatic interaction between the electrons and not by the static and dynamic crystal field. More recently, effort came to replace the commonly used mercury by a noble gas. The advantages for getting rid of mercury are based on environmental considerations and the need to extend the application for TL lighting. The applications for mercury-based TL lighting are limited as a certain startup time was found. The startup time is present because the mercury first has to evaporate before it can emit UV light. The delay in startup is unwanted for possible applications of TL lighting in photocopiers and brake lights. The three phosphors present in the commonly used TL lighting

1,0

8.3eV 0,9

7.2eV

0,8

Intensity (a.u.)

0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 120

130

140

150

160

170

180

190

200

Wavelength (nm)

Figure 2. Emission spectrum of a Xe discharge. The emission is situated in the vacuum ultraviolet (VUV) spectral region.

A. P. Vink et al

114

are optimized for the ultraviolet discharge, but not for the xenon discharge, which shows only emission below 200 nm (see Fig. 2).

2.

QUANTUM CUTTING

The major disadvantage of using xenon in the new-generation TL lighting is the low energy efficiency. At 254 nm, about 50% of the excitation energy is lost, whereas at 172 nm this loss is more than 70%. A solution to this problem could be the introduction of phosphors, which can emit two photons per absorbed photon. The dopant (e.g. a lanthanide) however needs specific energy levels to make a two-step emission to the ground state possible. The two-step emission process is called Photon Cascade Emission (PCE) or quantum cutting. In the past, several different quantum cutting systems where found, which are described below. In 1974 Piper [2] and Sommerdijk [3] independently found that if Pr3+doped YF3 is irradiated with a zinc or mercury lamp, emission from the highenergy 1S0 level occurs. This results in the emission of two photons at 404 nm and 480 nm. More recently, research led to a quantum cutting system based on energy transfer. It was found that excitation of the LiGdF4:Eu3+ phosphor led to the emission of two red photons. At around 50,000 cm–1 (200 nm), the 6GJ levels of Gd3+ can be populated. After a Gd-Eu cross-relaxation step, the 5D0 level of Eu3+ can be populated and emission of the first red photon occurs. The remaining energy (Gd3+ in the 6PJ level) can be transferred to the Eu3+, where the 5D0 level can be populated. From this level, emission of a second photon can occur. A schematic representation of the Gd-Eu quantum cutting system is presented in Fig. 3 [4,5]. Although the theoretical quantum efficiency can be as high as 200%, the actual quantum efficiency was found to be much lower. The low value of 32% (including quantum cutting) for the measured quantum efficiency can be ascribed to weak absorption in the Gd3+ ions. The 8S7/2ĺ6PJ transition is parity forbidden and therefore competes with the host absorption in the same region. The addition of co-dopants, with a strong absorption in the 50,000 cm–1 region, could increase the quantum efficiency [5]. Another concept of quantum cutting without activators is exciting of a semiconductor (or insulator) at twice the band gap energy. Via Augerinteraction, an electron high up in the conduction band and a hole at the bottom of the valence band generate two electron–hole pairs located at the top of the valence band (holes) and the bottom of the conduction band (electrons) [5].

Quantum Phosphors

115

Although this process sounds very promising, major limitations are set to the use of these materials. It was discovered that the Auger effect is efficient at energies of about 2.5 times the band gap. As the xenon discharge emission is located around 7.2 eV (see Fig. 2), materials with a band gap smaller than 2.9 eV are needed. Such materials however are not transparent to visible light. Using such higher excitation energy would again imply a low energy efficiency.

Figure 3. Quantum cutting using the Gd-Eu system.

In this contribution, the focus will be on the Pr3+-doped materials. Although the Pr3+ quantum cutting system show major disadvantages, like emitted photons with different energies and photons in the violet spectral region, it offers the advantages that it shows a strong absorption of the Xe discharge emission region and quantum cutting without the use of codopants.

3.

PHOTON CASCADE EMISSION WITH PR3+

3.1

Introduction

In the PCE process, a high-energy photon is absorbed, what results in twostep emission to the ground state. This process was already predicted in 1957

A. P. Vink et al

116

60

3

-1

Energy (10 cm )

by Dexter. He called this process “photon splitting” [6]. The first actual example for this process was discovered in 1974 independently by two different research groups [2,3]. The group of Sommerdijk from Philips Research Laboratories found the PCE process in YF3:Pr3+ and Į-NaYF4:Pr3+ under excitation with a zinc lamp [3], whereas Piper and co-workers from General Electric found the PCE process in a large number of Pr3+-doped hosts, also including YF3:Pr3+ and ĮNaYF4:Pr3+, by excitation with a mercury lamp [2]. Piper already stated that, to obtain the PCE process, the 4f2 1S0 level must be located below the 4f15d1 bands. Piper also measured Pr3+-doped materials, like LiYF4 where only 4f15d1ĺ4f2 emission was found [2]. In Fig. 4, the energy level scheme of the Pr3+ ion is shown. The energy level scheme consists of 13 Pr3+ (4f2) levels, which are mainly situated in the energy region up to about 22,000 cm–1. Around 47,500 cm–1, the isolated 1S0 level can be found. PCE can occur when this level is populated. This level can be populated by excitation with ultraviolet (UV) light of about 215 nm, but the 3H4ĺ1S0 transition is inefficient as it is parity forbidden. A more efficient population of the 1S0 level can be achieved by pumping into the 4f15d1 bands. If the bands are situated above the 1S0 level, efficient nonradiative relaxation leads to population of the 1S0 level. From this level, two-step emission to the ground state can occur.

1

1

4f 5d 50 1

S0

40

30 3

P2 3

3 1

1

10

1

P1, I6 P0

20 D2

G4 3

3

F3, F4 3 H6, F2

3 3

0

H5 H4

3

Figure 4. Energy level scheme of the Pr3+ ion showing the PCE process.

Quantum Phosphors

117

The most efficient two-step emission route is via 1S0ĺ1I6 and 3P0ĺ3H4 transitions. This results in photons with a wavelength around 404 nm (1S0ĺ1I6 transition) and 480 nm (3P0ĺ3H4 transition). Loss processes are other 1S0ĺ2S+1LJ transitions, which are visible as emissions in the UV region. Nonradiative relaxation from the 3P0 to the 1D2 leads to red emission. The process of nonradiative relaxation is more efficient for hosts with higher phonon energies. In principle, fluoride quantum cutters show strong green emission, whereas the Pr3+ emission for oxide-based hosts is situated in the red spectral region. This red emission is however weak as the emission from the 1D2 level is quenched by cross-relaxation at low Pr3+ concentrations. In Fig. 5, an excitation- and emission spectrum of SrAlF5:Pr3+, a typical quantum cutter, is shown. The different emissions are assigned [7].

1,0

3

1

1

1

H4 -> 4f 5d

Oem=404 nm Oexc=189 nm

1

S0 -> I6

Intensity (a.u.)

0,8

3

0,6

3

P0 -> H4

0,4

host

0,2

1

1

S0 -> G4 1 1 S0 -> D2

1

3

D2 -> H4 3

3

P0 -> H6

0,0 50

100 150 200 250 300 350 400 450 500 550 600 650 700

Wavelength (nm)

Figure 5. Excitation- (Ȝem = 404 nm) and emission (Ȝexc = 189 nm) spectrum of SrAlF5:Pr3+ measured at T = 10 K.

Many Pr3+-doped hosts show quantum cutting. It was found for Pr3+doped SrAl12O19 [8], LaB3O6 [9], LaMgB5O10 [10], KMgF3 [11], NaMgF3 [12], LiCaAlF6, LiSrAlF6 [13], LuF3, BaMgF4 [14], LaZrF7, Į-LaZr3F15 [15], BaSiF6 [16], Sr0.7La0.3Al11.7Mg0.3O19, SrB4O7 [17], SrSO4, and BaSO4 [18]. Not all Pr3+-doped hosts show PCE. The occurrence of the PCE process is determined by the position of the high-energy levels. As the 4f orbitals are shielded form their surrounding by the filled 5s2 and 5p6 orbitals, it can be expected that the location of the 4f2 energy levels, like the 1S0 level, is almost independent of the host in which the lanthanide is doped. The interaction between the host and 5d1 orbitals is much stronger, however. Therefore it is

A. P. Vink et al

118

expected that the position of the 4f15d1 energy levels show a strong variation with respect to the host in which the Pr3+ ion is doped. This is clearly illustrated in Figs. 6 and 7 where the excitation- and emission spectra of BaSO4:Pr3+ and CaSO4:Pr3+ are shown [19]. 1,0

1

Excitation (Oem=403 nm) Emission (Oexc=187 nm)

1

S0-> I6

Intensity (a.u.)

0,8 0,6 0,4

1

1

0,2

1

3

S0-> F4

1

S0-> D2

1

3

3

D2-> H4 + 3 3 P0-> F2

1

S0-> G4

3

P0-> H4

0,0 100

150

200

250

300

350

400

450

500

550

600

650

Wavelength (nm) Figure 6. Excitation- (Ȝem = 403 nm) and emission (Ȝexc = 187 nm) spectrum of BaSO4:Pr3+ measured at T = 10 K.

1,0

Intensity (a.u.)

0,8

1

1

3

a: 4f 5d -> H4 1 1 3 b: 4f 5d -> H5

1

a b

1

3

Excitation (Oem=230 nm) Emission (Oexc=190 nm)

4f 5d -> F2 1 1 3 4f 5d -> F3

0,6 0,4 1 1

1

1

4f 5d -> D2 1 1 3 1 4f 5d -> G4 4f 5d -> PJ, I6 (J:0,1,2)

0,2

1

1

1

3

D2-> H4 + 3 3 P0-> F2

0,0 100

150

200

250

300

350

400

450

500

550

600

650

Wavelength (nm) Figure 7. Excitation- (Ȝem = 230 nm) and emission (Ȝexc = 190 nm) spectrum of CaSO4:Pr3+ measured at T = 10 K.

The emission spectrum of BaSO4:Pr3+ is that typical of an oxide-based quantum cutter. It shows a strong 1S0ĺ1I6 emission but almost no green and red emission. The absence of these emissions can be ascribed to the high

Quantum Phosphors

119

phonon energy of the sulphate host, favoring nonradiative relaxation from the 3P0 to the 1D2 level and quenching of the 1D2 level by cross-relaxation [20]. The emission spectrum of CaSO4:Pr3+, however, shows no 4f2 line emission but the broad-band emission, which is mainly located far in the UV spectral region. For CaSO4:Pr3+ the lowest-energy 4f15d1 band is located below the 1S0 level. Therefore, excitation into the 4f15d1 bands yields a parityallowed 4f15d1ĺ4f2 emission and no population of the 1S0 level.

3.2

Selecting Hosts

To predict which Pr3+-doped hosts show the PCE process, data from the optical properties of Ce3+ can be used. The Ce3+ ion only has one electron in the 4f shell. This gives rise to only two 4f1 levels: 7F5/2 ground state and the 7F7/2 excited state. The 4f05d1 bands are located at higher energy. Like for Pr3+, the energy of these bands is strongly dependent on the host in which the lanthanide is doped. The Ce3+ 4f05d1 bands are however located at much lower energy that the Pr3+ bands. In Fig. 8A and 8B, the excitation spectra of CaSO4:Ce3+,Na+ and CaSO4:Pr3+ are shown. B

Intenstity

Intenstity

A

60,0 57,5 55,0 52,5 50,0 47,5 45,0 42,5 40,0 37,5 35,0 32,5 30,0 72,5 70,0 67,5 65,0 62,5 60,0 57,5 55,0 52,5 50,0 47,5 45,0 42,5 3 -1 3 -1 Energy (10 cm ) Energy (10 cm )

Figure 8. Excitation spectra of CaSO4:Ce3+,Na+ (A, Ȝem = 326.5 nm ) and CaSO4:Pr3+ (B, Ȝem = 230 nm). Note the difference in the energy scale.

From Fig. 8, it can be clearly observed that the structure of the 5d bands is roughly the same. An analysis of the position of the 4fn–15d1 bands in many different lanthanide host shows that the energy difference between the lowest-energy 4fn–15d1 band of Ce3+ and Pr3+ is around 12,240 cm–1 [21,22]. This fixed energy difference can be used to predict Pr3+-based quantum cutters from Ce3+ data. The position of the Ce3+ 4f05d1 bands is known in many hosts as these bands are generally located in the UV spectral region and are therefore relatively easy to measure.

A. P. Vink et al

120

The position of the 4fn–15d1 bands in general is roughly determined by two independent factors: (i) by the centroid energy EC, which is mainly determined by type of ligands in the host and (ii) by the crystal field splitting İcfs which is determined by the symmetry and coordination number (CN) of the dopant in the host [21,22]. To get quantum cutting in a Pr3+-doped host, a host material is needed in which the lanthanide has a high centroid energy and a small crystal field splitting. Most of the Pr3+-doped materials are based on fluorides, which have a large centroid energy. The BaSO4:Pr3+ material, on the other hand, shows the PCE process (see Fig. 6), because İcfs is rather small.

3.3

Two Types of Emissions

Some Pr3+-doped materials show both broad-band 4f15d1 emission and 4f2 [1S0] emission (resulting in quantum cutting) under excitation into the 4f15d1 bands. An explanation for this behavior can be the presence of two different cation sites, like in CaF2:Pr3+ [23]. This behavior can however not explain the emission behavior of the BaSO4:Pr3+ material, where only one Pr-site site is expected. 1.0

1

T= 10K T= 292K

1

S0 -> I6

Intensity (a.u.)

0.8

1

1

S0 -> D2

0.6 1

1

3

1

4f 5d -> H4 - G4 0.4 1

1

S0 -> G4

0.2

1

1

1

3

4f 5d -> D2

3

P0 -> H4

0.0 200

250

300

350

400

450

500

Wavelength (nm)

Figure 9. Emission spectrum (Ȝexc = 188 nm) of BaSO4:Pr3+ at T = 10 K (dashed line) and 292 K (solid line). The 4f2 and 4f15d1 emissions are assigned.

The intensity ratio between all 4f15d1 and 4f2 emissions was found to be strongly temperature dependent as is shown in Fig. 9. Also, a decrease of more than three times was found for the decay time of the 1S0 emissions (from 190 to 56 ns) [24].

Quantum Phosphors

121

This behavior can be explained by assuming thermal population of the lowest energy 4f15d1 band from the 1S0 level. This process is temperature dependent as at higher temperatures more electrons can cross the energy barrier between the 1S0 and the 4f15d1 band. By fitting both the intensity ratio and the decay time for the different temperatures, a value for the energy barrier ǻE of about 0.04±0.006 eV (325±50 cm–1) was found [25]. The process of thermal population is present for all Pr3+-doped materials, which show the PCE process, but thermal population can only be observed at room temperature for materials where the 4f15d1 bands are relatively close to the 1S0 level. This thermal population process is a serious concern for the application of oxide-based Pr3+-doped quantum cutters.

3.4

Quantum Cutting with X Rays

The observation of the PCE process is not only limited for excitation into the 4f15d1 bands using high-energy vacuum ultraviolet (VUV) light. Recent literature also shows quantum cutting for Pr3+-doped hosts under X ray excitation [25,26]. In Fig. 10 the emission spectrum of SrAlF5:Pr3+ at T = 100 K and T = 350 K under X-ray excitation is shown. From this figure, it can be observed that emission from 1S0 level is absent at 100 K, whereas it is present at 350 K. 0.5

T=100K T=350K

Intensity (a.u.)

0.4

0.3

0.2

0.1

0.0 200

300

400

500

600

700

800

Wavelength (nm)

Figure 10. X-ray excited emission spectra of SrAlF5 :Pr3+ measured at T = 100 K (dotted line) and T = 350 K (solid line). The emission spectra are corrected for the response of the measuring system and photo multiplier.

122

A. P. Vink et al

Other emission originating from the lower-lying Pr3+ levels, like 3P0 and 1D2, is however clearly visible at both temperatures. This must mean that a route other than the 1S0ĺ1I6 emission populates the 3P0 level. Figure 10 also shows that at 100 K, a broadband emission around 450 nm is visible. This emission can be assigned to emission from a localized electron–hole pair (SelfTrapped Exciton, STE). At higher temperature, energy transfer from the STE to the Pr3+ ion can occur. Energy transfer from the STE to the Pr3+ can explain the presence of Pr3+ emission from the lower-lying 3P0 and 1D2 levels, which is present at both temperatures in the region from 480 to 720 nm (see Fig. 10). It is however crucial for resonant energy-transfer to the 1S0 level that the STE band extends up to the energy corresponding to the 3H4ĺ1S0 transition at a 215 nm. As can be observed from Fig. 10 the STE starts to emit from 240 nm. The process, which is responsible for the typical quantum cutting behavior under host excitation at different temperatures, is the direct recombination (without the formation of a STE) of the electron and the hole with the Pr3+ ion. It is however expected that the process of direct recombination is temperature independent. Two possible explanations for the temperature dependence of the recombination process can be proposed: (1) The electron can recombine with Pr3+ forming Pr2+ and the hole can be trapped in a VK center and (2) the hole recombines with Pr3+ to form Pr4+ and the electron is trapped for example at an anion vacancy, forming an F center. The hole (for process 1) or the electron (for process 2) can be de-trapped at elevated temperature and recombine at Pr4+ and Pr2+ resulting in population of the 4f15d1 bands. Using a temperature study of the 1S0 emission intensity, assuming Arrhenius behavior for the intensity, the activation energy ǻE could be determined. The value of ǻE (450 cm–1 or 0.06 eV) is so small that it can be interpreted as an electron trap. The activation energy for migration of holes is much higher, typically in the order of tenths of electronvolts [27]. Therefore it seems that the observation of the PCE process originates from process (2) [7]. The different excitation, energy transfer and emission processes described above are visualized in Fig. 11. The band gap of SrAlF5 is about 90,000 cm–1, corresponding to 11 eV. The different processes resulting in Pr3+ emission are visualized in Fig. 11. Excitation of electrons into the conduction band results in holes in the valence band (1). The first process is the formation of an STE, which is shown as (2a) and (2b). The STE can either emit radiatively (3a) or can transfer its energy to Pr3+ (3b). This energy transfer is more efficient at higher temperatures as the STE becomes mobile. Mi-

Quantum Phosphors

123

CB

90 5b

3

-1

Energy (10 cm )

gration to Pr3+ is followed by energy transfer, populating the lower-lying (3PJ, 1I6 and 1D2) Pr3+ levels. Emission from the 3P0 is shown as (4) [8]. The other process, which leads to Pr3+ emission, is population of the lanthanide without an intermediate exciton state. Here, the hole is trapped on Pr3+ (5a) and the electron in an electron trap (5b). This situation does not result in any emission from praseodymium at temperatures lower than 150 K. Above this temperature, the electrons in the shallow traps are released and populate the higher 4f15d1 bands of Pr3+ (6). This population of the 4f15d1 bands results in two-photon emission, which is shown as 1S0ĺ1I6 (7a) and 3P0ĺ3H4 emission (7b) [8].

6

electron trap

75 1

2a

1

4f 5d 1

60 1

S0

7a

45

3a 1

3 1

30

4 7b

3b 3

15

G4

H5

3

STE

1

PJ (J:0,1,2), I6 D2

5a

Pr

3

F3, F4

3

H4

2b

3

3

H6, F2

3+

0 VB Figure 11. Schematic description of the different excitation, emission, and energy transfer processes. The processes, described in the text, are numbered accordingly.

3.5

Energy Transfer of the 1S0 Emission

For lighting applications, the emission of the first PCE step (1S0ĺ1I6) is too much on the short-wavelength side of the visible spectral region. For application of quantum cutting phosphors in lighting, it is highly preferable to convert the violet photon to a visible photon. A possibility for energy con-

A. P. Vink et al

124

60

3

-1

Energy (10 cm )

version is to add a co-dopant, which can convert the energy from the 1S0ĺ1I6 emission to the visible spectral region. A possible candidate could be the Mn2+ ion as the 6A1 ĺ4A1,2E absorption bands show strong overlap with the 1 S0ĺ1I6 emission (see Fig. 12).

1

1

4f 5d

50 1

S0

4

T2 T 4 1 A2

4

40

4

T1 E 4 T2 4 2 A1, E 4 T2 4 T1 4

30 3

P2

3

1

P1, I6 P 1 0 D2

20

3

1

10

G4

3

3

F3, F4 3 H , F2 3 6 H5 3 H4

3

0

Pr

3+

6

Mn

2+

A1

Figure 12. Energy level schemes of both Pr3+ and Mn2+ showing the possibility of Pr-Mn energy transfer.

Measurements on SrAlF5:Pr3+,Mn2+ however showed no typical green Mn2+ emission under excitation into the Pr3+ 4f15d1 bands. X-ray excited emission measurements however show the presence of built-in Mn2+ ions. Up till now it is not understood why the energy transfer does not occur. It was suggested that selection rules for exchange interaction apply, making the energy transfer from the Pr3+ 1S0 state to Mn2+ a forbidden transition [5]. Up till now no theoretical background was found for this claim.

4.

CONCLUSIONS

It was shown that quantum cutting is possible for Pr3+ in many, mainly fluoride-based, hosts. Furthermore, a possibility to predict whether Pr3+, in a cer-

Quantum Phosphors

125

tain host, shows PCE was explained. This method makes use of the large amount of data, which is available on the optical properties of the Ce3+ ion. Measurements on an oxide-based quantum cutter (BaSO4:Pr3+) showed that the quantum efficiency in the visible spectral region is severely lowered at higher temperatures by an increase of UV emission. This makes the application of oxide-based quantum cutting phosphors more difficult. Using X-ray excitation, quantum cutters were studied under band gap excitation, creating electrons and holes. It was shown that at low temperatures, broadband STE emission is visible. Only at higher temperatures, PCE becomes visible. It is expected that for fluoride-based quantum cutters, no band-to-band transitions occur at typical xenon discharge excitation energies. For oxide-based quantum cutters, host lattice excitation is efficient and gives rise to efficiency losses. The fluoride-based quantum cutters cannot be applied in new-generation TL lighting, as the energy of one of the emitted photons is too far into the UV side of the visible spectral region. Efforts to convert this photon more to the visible region were not successful up till now. It is highly unlikely that the lanthanide-based quantum cutting phosphor will be applied for (TL) lighting in the near future. Both systems (Gd–Eu and Pr) still have major difficulties, which have to be overcome first. It must also be noted that the application of these phosphors in Light Emitting Diodes (LEDs) is not possible as the emission of LEDs is too much on long wavelength side to excite the high-energy levels of the quantum cutting phosphors.

ACKNOWLEDGEMENTS The authors thank Dr. M. Kirm (HASYLAB, DESY Hamburg) for his assistance during the experiments performed on the SUPERLUMI set-up and Dr. M. Weil (Institute for Chemical Technology and Analytics, Vienna University of Technology) for sample preparation. The investigations were supported by the Dutch Technology Foundation (STW) and by the IHP-Contract HPRI-CT-1999-00040 of the European Commission.

REFERENCES 1. 2. 3.

G. Blasse and B. C. Grabmaier, Luminescent materials (Springer-Verlag, Berlin, Heidelberg, New York, 1994). W.W. Piper, J. A. deLuca, and F. S. Ham, J. Lumin. 8, 344–348 (1974). J. L. Sommerdijk, A. Bril, and A. W. de Jager, J. Lumin. 8, 341–343 (1974).

A. P. Vink et al

126 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

R. T. Wegh, H. Donker, K. D. Oskam, and A. Meijerink, Science 283, 663–666 (1999). C. Ronda, J. Lumin. 100, 301–305 (2002). D. L. Dexter, Phys. Rev. 108, 630–633 (1957). A. P. Vink, P. Dorenbos, J. T. M. de Haas, H. Donker, P. A. Rodnyi, A. G. Avanesov, and C. W. E. van Eijk, J. Phys.: Condens. Matter 14, 8889–8899 (2002). A. M. Srivastava and W. W. Beers, J. Lumin. 71, 285–290 (1997). A. M. Srivastava and D. A. Doughty, J. Electrochem. Soc. 144, L190–192 (1997). A. M. Srivastava and D. A. Doughty, J. Electrochem. Soc. 143, 4113–4115 (1996). I. Sokólska and S. Kück, Chem. Phys. 270, 355–362 (2001). N. J. M. Le Masson, A. P. Vink, P. Dorenbos, A. J. J. Bos, J. P. Chaminade, and C. W. E. van Eijk, J. Lumin. 101, 175–183 (2003). S. Kück and I. Sokólska, J. Electrochem. Soc. 149, J27–J30 (2002). S. Kück and I. Sokólska, Chem. Phys. Lett. 364, 273–278 (2002). E. van der Kolk, P. Dorenbos, and C. W. E. van Eijk, Opt. Commun. 197, 317–326 (2001). E. van der Kolk, P. Dorenbos, C. W. E. van Eijk, A. P. Vink, C. Fouassier, and F. Guillen, J. Lumin. 97, 212–223 (2002). E. van der Kolk, P. Dorenbos, and C. W. E. van Eijk, J. Phys.: Condens. Matter 13, 5471–5486 (2001). E. van der Kolk, P. Dorenbos, A. P. Vink, R. C. Perego, C. W. E. van Eijk, and A. R. Lakshmanan, Phys. Rev. B 64, 195129 (2001). A. P. Vink, E. van der Kolk, P. Dorenbos, and C. W. E. van Eijk, J. Alloys Comps. 341, 338–341 (2002). A. P. Vink, P. Dorenbos, and C. W. E. van Eijk, J. Solid State Chem. 171, 308–312 (2002). P. Dorenbos, J. Lumin. 91, 91–106 (2000). P. Dorenbos, J. Lumin. 91, 155–176 (2000). K. D. Oskam, A. J. Houtepen, and A. Meijerink, J. Lumin. 97, 107–114 (2002). A. P. Vink, P. Dorenbos, and C. W. E. van Eijk, Phys. Rev. B 66, 075118 (2002). A. M. Srivastava and S. J. Duclos, Chem. Phys. Lett. 275, 453–456 (1997). P. A. Rodnyi, S. B. Mikhrin, P. Dorenbos, E. van der Kolk, C. W. E. van Eijk, A. P. Vink, and A. G. Avanesov, Opt. Commun. 204, 237–245 (2002). K. S. Song, Solid State Commun. 9, 1263–1266 (1971).

OPTICAL MEASUREMENTS USING LIGHT-EMITTING DIODES A. ŽUKAUSKAS 1, M. S. SHUR 2, and R. GASKA 3 1

Institute of Materials Science and Applied Research, Vilnius University, Saulơtekio 9-III, LT-2040 Vilnius, Lithuania 2 Center for Broadband Data Transport, Rensselaer Polytechnic Institute, CII 9017, 110 8th street, Troy, New York 12180, USA 3 Sensor Electronic Technology, Inc., 1195 Atlas Road, Columbia, South Carolina 29209, USA

Abstract:

Recent advances in optical measurements using light-emitting diodes (LEDs) are reviewed. The review covers applications of LEDs as stable and compact sources of light, fluorometry including fluorescence lifetime measurements, and spectroscopic applications (photoluminescence line shape, absorption and absorption correlation, surface-plasmon resonance, photoreflection, and Raman measurements).

Key words:

light-emitting diodes, optical measurements, fluorescence sensing, spectroscopy

1.

INTRODUCTION

The advances in semiconductor materials and in improved light extraction techniques led to the development of a new generation of efficient and powerful high-brightness LEDs [1]. Red AlGaInP LEDs and violet InGaN LEDs demonstrated efficiencies approaching 60% [2] and in excess of 40% [3], respectively. Visible colored and white phosphor-conversion LEDs are already available in the electrical power range of 1 to 5 W with the optical power output of hundreds of milliwatts [2,4,5]. Further progress in the development of AlInGaN materials system has resulted in an appearance of ultraviolet (UV) LEDs with the wavelengths as short as 265 nm [6]. Highpower near-UV LEDs with the output power of 200 mW have been reported recently [7]. 127 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 127–142. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

128

A. Žukauskas, M. S. Shur and R. Gaska

In addition to wide spread use of LEDs in signals, full-color video displays and lighting, advanced LEDs find many new applications in optical measurement technology, where they substitute for conventionally used sources of light, such as incandescent or discharge lamps and even lasers. Optical measurement technology benefits from advantages of LED technology, such as broad range of available emission wavelengths, high efficiency and power, stable output, long lifetimes (100,000 h and more), durability, low driving voltage, small dimensions, low self heating, reliability, and low cost. In combination with extremely low noise of the radiant flux [8] and with possibility of high-frequency modulation [9] and subnanosecond pulse generation [10], these advantages resulted in the novel instrumentation based on a variety of optical techniques ranging from simple transmission measurements to a more sophisticated spectroscopy. This paper reviews instrumental applications of LEDs. Section 2 describes applications requiring compact, reliable, and stable sources of light with a narrow-continuum spectrum. Section 3 deals with LED-based fluorometry, including fluorescence lifetime and fluorescence anisotropy measurements. Section 4 briefly describes spectroscopic applications.

2.

LEDS AS COMPACT AND STABLE SOURCES OF LIGHT

LEDs are widely used as compact and stable narrow-continuum light sources for optical transmittance and reflectivity measurements, photodetector calibration, and as well as for generation of light pulses with stable parameters. One of the most famous transmittance applications is in pulse oxymetry, a noninvasive method for monitoring arterial oxygen saturation. The method is based on different absorption spectra of hemoglobin saturated with oxygen and desaturated hemoglobin. Typically, the sensor probe consisting of two LEDs (660 nm and 910 nm, respectively) and a photodetector is mounted on a finger of the patent. An ac component of the photodetector signal, which is due to blood pulsation, is used to estimate the amount of oxygen saturation. Another example of simple application of LEDs is optical thermometer (Fig. 1). Using a glass fiber, light emitted by the LED is guided to a conventional low-pass color filter that is placed in the environment under testing. Another fiber is used to bring the light passed through the filter to a photodetector. The wavelength of the LED is matched with the absorption edge of the filter, which contains semiconductor particles dispersed in a glass matrix. A temperature change shifts the absorption edge of the filter due to semiconductor band gap variation and, as a result, the amount of the transmitted light is altered. The all-glass thermometer probe can be used in

Optical Measurements Using Light-Emitting Diodes

129

strong electric and magnetic fields, under nuclear radiation, and in corrosive chemical environments. LED Fiber

Photodetector Color glass filter

Figure 1. Schematic of the LED-based optical thermometer.

Direct switching of the LED driving current can yield light flashes of precise duration and intensity. Such flashes are widely used in machine vision to obtain images of moving objects. A remarkable application of LEDgenerated stable light pulses is in optical dating of sediments [11]. The method relies on freeing the electrons produced by radiation of natural radioisotopes and trapped at defect sites of the sediment. Some of these electrons do not relax on a geological time scale unless excited by light. The number of trapped electrons depends on the radiation dose accumulated since the last exposure to light. Under illumination with a precise portion of light, the sediment might exhibit characteristic anti-Stokes luminescence with the amount of photons emitted being proportional to the time since the sediment deposition. Dating of quartz and feldspars was demonstrated using highbrightness 525-nm LEDs. An array of LEDs was shown to successfully substitute for much more expensive argon-ion- or organic-dye laser systems. LEDs are narrow-continuum sources with the linewidth of the emission spectrum in the range of 10 nm. The relevant coherent length is of the order of 10 Pm. This implies that LEDs can be used in “white-light” interferometers for measurement of small absolute displacements. In comparison with interferometers based on monochromatic or long-coherence-length sources, LED-based interferometers offer simple central-fringe identification. A LED-based fiber-optic interferometer was described in Ref. 12. The interferometer featured a two-arm Michelson design and a three-LED broadspectrum light source for improved central-fringe identification. The device was successfully tested as a strain sensing system.

A. Žukauskas, M. S. Shur and R. Gaska

130

3.

FLUOROMETRY

Using advanced LEDs, detectable levels of fluorescence can be excited in a variety of objects. The development of blue and UV chips resulted in a dramatic increase of LED-based fluorometry applications, especially in the fields of biochemistry, life sciences, and environmental control. This section reviews LED-based fluorescence sensing as well as measurements in timeand frequency domain.

3.1

Fluorescence Sensing

Simple high-precision fluorometers consist of an LED operating in a continuous regime, an optical filter (filtering the fluorescence), and a photodetector [13]. An additional optical filter can be used to spectrally isolate the detector from the excitation emission. Optical fibers that deliver both the excitation light and the fluorescence signal are used in remote configurations. To separate the fluorescence signal from the ambient light, the LED is pulsed and amplitude-modulation [14] or phase-sensitive [15] detection technique is used. Owing to low-noise output of LEDs, such inexpensive fluorometers can exhibit high sensitivity and precision, comparable with that of state-of-the-art systems that employ bulky xenon lamps. Compact and simple LED-based fluorometers can be useful for detection of various organic and inorganic compounds in biotechnology, chromatography, water purity control, and hazardous biological agent detection. LED-based chemical sensors with fluorophores that are either quenched or activated by the substance detect gases, proteins, nucleic acids, etc. [1]. Advanced fluorometric systems employ arrays of different LEDs for multi wavelength excitation that results in partially selective fluorescence measurement [16]. Under appropriate signal processing, such systems were demonstrated to resolve the individual contributions in multicomponent mixtures. By using an array of 370 to 640 nm LEDs, correct identification and quantification of six fluorescent dyes in two to six component mixtures has been achieved [16].

3.2

Time-Domain Fluorescence Measurements

LEDs operating in a short-pulse generation regime offer an attractive alternative to costly and bulky lasers or pulsed arc lamps. In particular, timeresolved fluorescence measurements can be implemented using repetitive short pulses of LED light. The time-domain measurements trace the decay of fluorescence intensity and polarization after excitation with a short pulse, with the obtained fluorescence transient being used to extract the decay time.

Optical Measurements Using Light-Emitting Diodes

131

A simple LED-based system for fluorescence lifetime measurements on the nanosecond time scale [10] contained an avalanche-transistor-based current driver that pumped a high-brightness blue AlInGaN LED by pulses with a peak current of 2 A at a repetition rate of 10 kHz. In this regime, the LED generated 4-ns UV light pulses with the output of 40 mW. Fluorescence decay in quinine sulfate solution was recorded using a conventional timecorrelated single-photon counting system and the fluorescence lifetime of 19.5 ns was extracted from the decay kinetics. Another example of LEDs replacing pulsed lasers and arc flash lamps is time-resolved polarization anisotropy measurements [17]. The fluorophore molecules are excited with their optical transition vectors being parallel to the polarization plane of the exciting light. The initial biased population of molecules becomes increasingly randomized with time due to Brownian rotational diffusion. Since the polarization plane of a fluorescence photon is determined by the molecule orientation, the fluorescence polarization measurement provides information on molecular interaction. An LED-based technique was applied for time-resolved fluorescence anisotropy measurements in dilute solution of Coumarin 6 in ethylene glycol. The arrangement consisted of a blue LED producing highly reproducible 680-ps pulses at a repetition rate of 10 MHz, a low-pas optical filter, sheet polarizers, and a timecorrelated single-photon counting system. By measuring the fluorescence intensity decay in two polarizations with high statistical precision, the transient of the anisotropy was extracted and the anisotropy decay time of 2.1 ns was determined. The simplicity of driving LEDs in a pulsed regime and availability of LEDs over a broad range of wavelengths allow one to compose experimental protocols for even more sophisticated investigation of fluorescence transients using the series of excitation and sampling pulses of different wavelengths and polarizations with an independent control of intensity and of temporal profile [18].

3.3

Frequency-Domain Fluorescence Lifetime Measurements

Fluorescence lifetime measurements provide useful information on structure, environment, and transient evolution of molecular compounds. Frequencydomain measurements allow one to measure the lifetime without a detailed analysis of the fluorescence decay kinetics, in contrast to time-domain measurements, which require numerous data points to be processed for extraction of the lifetime values. Figure 2 illustrates the principle of the frequencydomain measurement [19]. The excitation source is modulated by a sinusoidal waveform at an angular frequency Z. This results in the modulation of

A. Žukauskas, M. S. Shur and R. Gaska

132

the fluorescence signal with the same frequency. However because of a finite fluorescence lifetime W , the fluorescence signal has a phase shift I and the modulation depth is decreased by a factor m. For a single-exponential decay, the phase shift and the relative modulation depth are given by

tan I m

ZW ,

1  Z W 2 2

1 2

I

.

Excitation Fluorescence b

Intensity

B

m=

a

BA ba

A

Time Figure 2. Time variation of the excitation and fluorescence intensity in frequency-domain lifetime measurements. (After Ref. 19.)

These relations imply that by measuring the phase shift and modulation depth, the lifetime can be extracted in two independent ways. Typically, a dependence of the phase shift and/or modulation depth is measured as a function of frequency and the lifetime is extracted by fitting the data points to a simulated dependence. The frequency-domain method uses narrowfrequency-band electronics (typically, standard lock-in amplifiers) and, therefore, has a higher precision than the time-domain measurements relying on wide-band electronics. LEDs ideally fit the requirements for light sources in frequency-domain fluorescence lifetime measurements on the nanosecond time scale, since they can be directly modulated up to frequencies of hundreds of megahertz. The use of LEDs in frequency-domain measurement regime resulted in a substantial decrease in complexity and in the cost reduction of the instrumentation. An example of fluorescence lifetime measurement in standard fluorophore, fuorescein disodium salt, was described in Ref. 9. Blue and green

Optical Measurements Using Light-Emitting Diodes

133

LEDs were biased at 5 mA current, which was modulated with a radiofrequency power of 4 mW. The measured fluorescence decay time of 3.51 ns indicated that the inexpensive LED-based instrumentation is capable to substitute for much more complex phase-modulation technique with argon-ion laser modulated with a Pockels cell. Introduction of cost-efficient frequency-domain measurement technology led to the development of numerous sensors using chemically quenched fluorophores, such as ruthenium and platinum ligand complexes. In particular, LED-based fluorescent pH indicators and gas sensors have been developed [1]. Use of UV LEDs offers even more possibilities, since most organic compounds and biological agents exhibit excitation spectra in the UV range. Compact LED-based sensors can be assembled into arrays to monitor several chemical or biological species simultaneously. An example is a multichannel detection system that is able to monitor fluorescence lifetimes of many samples in real time [20]. The system detects LED-excited fluorescence at different wavelengths by means of a multianode photomultiplier and resolves lifetime changes of different fluorophores using phase meter software.

4.

LED BASED SPECTROSCOPY

Benefits offered by advanced LEDs, such as direct modulation, stability, small dimensions, low cost, and unique spectral properties were successfully utilized in a variety of spectroscopic applications. In this section, we briefly review recent applications of LEDs in photoluminescence spectroscopy, absorption and absorption correlation spectroscopy, surface-plasmon resonance sensing, photoreflection, and Raman spectroscopy.

4.1

Photoluminescence spectroscopy

Owing to low noise and high stability, LEDs can be used in photoluminescence measurements for precise characterization of the spectral features. An example of application of LEDs in luminescence spectroscopy is the investigation of photoluminescence in an InGaN alloy, the key material for fabrication of efficient green to near-UV LEDs. Figure 3 shows temperature dependences of the luminescence band linewidth and peak position in an InGaN epitaxial layer. The InGaN layer was photoexcited using a 375-nm LED with the active layer based on the InGaN alloy with a smaller indium molar fraction. The luminescence spectra were recorded using a double monochromator and a photon-counting system.

A. Žukauskas, M. S. Shur and R. Gaska

134

3.08 0.110

InGaN

0.108

Peak position Linewidth

3.07

3.06

0.104 0.102

3.05

0.100 3.04

0.098 0.096

Peak position (eV)

Linewidth (eV)

0.106

3.03 0.094 0.092

3.02

0.090 0

50

100

150

200

250

300

Temperature (K) Figure 3. Temperature dependence of the luminescence band linewidth and peak position in an InGaN epilayer excited by an InGaN LED.

Application of a low-noise light source for photoexcitation resulted in a highly precise determination of linewidth and peak position of the luminescence band. (Note that the experimental points for the linewidth are scattered within | 1 meV whereas the linewidth value is of about 100 meV.) This precision resulted in revealing of tiny nonmonotonous behavior of the measured temperature dependences. This nonmonotonous behavior is a signature of an intricate character of the exciton motion over the band potential fluctuations in InGaN alloys [21]. It is worth noting that the price of the LED light source is smaller by a factor of 103 than the price of a He-Cd laser commonly used in such experiments. A deeper penetration of LEDs into the UV region [6] may result in partial substitution of excimer lasers and with a commensurate price reduction by a factor of 104.

4.2 Absorption and Absorption Correlation Spectroscopy Absorption spectroscopy provides with a simple and powerful insight into many physical phenomena. Absorption measurements yield the energies of electronic transitions with high precision and quantitatively characterize the transition probabilities based on the absorption coefficients. Recently, shortwavelength LEDs received attention as light sources for absorption measurements in the near UV region. The interest in use of LEDs for near-UV absorption spectroscopy is due to several reasons. First, conventional incandescent lamps generate a relatively weak flux in this region what diminishes

Optical Measurements Using Light-Emitting Diodes

135

the signal-to-noise ratio, and discharge lamps suffer from inherent instabilities. Second, a narrow-band continuum is highly desirable in spectrally resolved absorption measurements in order to avoid stray light in other spectral regions. One of the first demonstrations of LEDs for absorption spectroscopy was performed in dense cesium vapor [22]. A high-brightness blue LED with the driving current increased over its standard value was used as a source of narrow-band continuum in the spectral range around 390 nm. Owing to a lownoise output of the LED, a few new spectral features were distinguished in the absorption spectrum. LED1

LED2

Reference cell

Measurement cell

Measurement detector

Reference detector Figure 4. System schematic for detection of gases by correlation spectroscopy (After Ref. 23).

Based on absorption measurements, an elegant method for target-specific detection of gases, which takes advantage of the full multi-line structure of the absorption spectrum, can be implemented using LEDs [23]. The proposed correlation spectroscopy method is based on two complementary sources, an LED with narrow-continuum emission covering the target absorption spectrum and a similar LED with the emission partially absorbed by target gas contained in a reference cell (see Fig. 4). The LEDs are modulated in anti-phase and the driving currents are balanced to produce no net intensity modulation at the reference photodetector, which measures the sum intensity of these two sources. The two beams are passed through the measurement cell and detected by the second photodetector. Once the target gas is present in the measurement cell, the intensity balance between the two beams is violated, since the original beam is absorbed to a larger extent than the partially-absorbed one. The resulting modulation is detected by the second photodetector. A non-specific absorption with non-matching spectral lines attenuates both beams in the same proportion and no modulation occurs. The method was shown to be capable of detecting various concentrations of O2, CO, and CO2 gases.

A. Žukauskas, M. S. Shur and R. Gaska

136

4.3

Surface-Plasmon Resonance

Surface-plasmon waves are guided optical modes, which can be excited in the interface between a conductor and a dielectric. The frequency of the waves is very sensitive to the refractive index of the dielectric. When the wave vector of the incident light matches with that of the surface-plasmon mode, the reflected light is strongly attenuated and a dip in the reflectivity spectrum occurs indicating the presence of a resonance. As a result, small changes in the refraction index can be monitored with the accuracy similar to that of the resonance-frequency measurement. Owing to high sensitivity, surface-plasmon resonance (SPR) is widely used in a variety of gas-, liquid-, and bio-sensors. Typically, halogen incandescent lamps are used in SPR sensors. The use of LEDs reduces the size and cost of SPR sensors and makes them portable, much more reliable, and suitable for long-lasting real-time measurements. Both narrow-continuum color LEDs [24] and wide-spectrum white LEDs [25] have been already demonstrated as light sources in SPR measurements. An example of white-LED based SPR sensor is schematically shown in Fig. 5.

Monochromator

Lock-in amplifier

White LED

Metal film

Collimator Polarizer

Figure 5. Schematic setup for the LED-based surface-plasmon resonance measurements. (After Ref. 25.)

The sensor contains a right-angle prism with the sensing surface coated with a 50-nm gold film. The sample fluid is run over the sensing surface

Optical Measurements Using Light-Emitting Diodes

137

within a chamber attached to the prism. A parallel beam of light emitted by the LED is formed by a two-lens collimator and polarized. The reflected light is collected and focused to the entrance slit of the monochromator that disperses the spectrum. An essential part of the setup is a lock-in amplifier that both modulates the driving current of the LED and implements phasesensitive detection. Like in other LED-based optical measurements, direct modulation of the optical source improves the measurement precision and results in a substantial advantage over halogen-lamp–based systems. The described white-LED based SPR sensor was shown to be able to reliably extract values of the refractive index of glycerin-water solutions with four-digit accuracy.

4.4

Photoreflectance

Photoreflectance is a kind of modulation spectroscopy that produces sharp derivative-like features in the reflectance spectra. Free carriers, which are photoexcited by a modulated source of light, modulate the reflectance by altering the refractive index. Photoreflectance is widely used for sensitive characterization of band structure, surface properties, and built-in electric fields in semiconductors. Typically, the modulated photoexcitation is produced by mechanical chopping of emission from lasers or arc lamps, and the reflectance signal is processed by a lock-in amplifier synchronized by an additional photodetector. Bright and easy-to-modulate LEDs provide a cost-efficient alternative for modulated sources of light in photoreflectance experiments. LEDs offer solutions with a variety of photoexcitation wavelengths, high stability, broad frequency range, and absence of mechanical parts. No additional photodetector is required for synchronization, since the LED can be driven by an internal oscillator contained in most lock-in amplifiers. Advantages of the LEDbased photoreflectance technique were demonstrated in Ref. 26. Blue and green AlInGaN LEDs (with no dc bias) were driven at 5 kHz frequency and 100% modulation depth. The LEDs shielded by colored glass filters were positioned next to the samples without focusing optics. Reliable photoreflectance spectra of GaAs structures were measured.

4.5 Raman Measurements Raman spectroscopy is widely used for characterization of vibrational spectra in molecules and crystals. Typically, laser radiation is used to produce narrow spectral lines caused via inelastic scattering of light by vibrations. Although the linewidth of LED emission is too large for resolving narrow Raman lines, the low-noise Raman signal can be used as an intrinsic inten-

A. Žukauskas, M. S. Shur and R. Gaska

138

sity standard for scaling fluorescence signals. An example of such an application is measurement of Raman spectra in water for the normalization of fluorescence [27]. Water has a Raman feature with a huge shift of about 3300 cm–1 and a large intrinsic width (| 400 cm–1) that is comparable with the linewidth of emission from LEDs (typically 380 cm–1 for AlGaInP red and amber LEDs and around 1000 cm–1 for green, blue, and near-UV AlInGaN LEDs, respectively). The feature is due to five overlapped O–H-intramolecular stretching modes of tetrahedral hydrogen bonded structure of liquid water. A Raman spectrum was recorded using a blue AlInGaN LED with the emission line peaked at 465 nm and with the 5 W electrical and 0.5 W optical powers. The LED driven by a constant current was mounted directly on the steel optical table that also served as a heat sink. To remove the longwavelength wing of the LED emission due to localization of carriers at bandtail energy states of the semiconductor alloy and to narrow the excitation spectra in the short-wavelength region, the emission was passed through band-pass and long-pass color-glass filters. The resulting spectrum of LED emission is shown in Fig. 6 by dashed line indicating a peak at 21600 cm–1 and a full width at half magnitude of ~800 cm–1. The emission was focused on water contained in a fused-silica cell by an optical grade acrylic collimator and a lens. The scattered light was collected by a condenser, polarized perpendicular to the scattering plane by a sheet polarizer, and projected on the entrance slit of a double monochromator equipped with a photomultiplierWavenumber / cm 16500

18000

19500

-1

21000

22500 LED

Counts / s

-1

6000

Plumbing water 4000 Fluorescence

2000 Distilled water 0 -6000

-4500

-3000

-1500

Raman shift / cm

0

1500

-1

Figure 6. Raman spectra of distilled and plumbing water recorded under high-power light emitting diode excitation (points) [27]. Solid line, fluorescence background; dashed line, the Rayleigh spectrum.

Optical Measurements Using Light-Emitting Diodes

139

based photon counting system. Pure distilled water as well as potable water from Vilnius city water supply was investigated at room temperature. Lower points in Fig. 6 depict a spectrum recorded for the distilled water sample. The minimal value of the measured Raman shift (2000 cm–1) is limited by the transmission threshold of the band-pass filter used. The feature at 3300 cm–1 is seen to be clearly resolved with the linewidth of about 900 cm–1 due to the broad line of the excitation source. The standard deviation of the Raman signal exactly equaled N s , where N s is the number of photon counts. This deviation was entirely due to the randomness of the spontaneous photon emission within Poisson distribution, and the limiting value of the signal to noise ratio of N s was achieved. This suggests that LEDs can serve as the lowest-noise sources for Raman spectroscopy, similarly to their applications in other optical measurements. Regular city water exhibits a spectrum with a pronounced fluorescence background (upper points in Fig. 2; the extracted background is shown by solid line). Since the Raman feature is much narrower than the fluorescence band and can still be clearly resolved, it can be used to scale the fluorescence signal. This technique was introduced in airborne laser fluorometry of water to correct the fluorescence signal for water transmittance [28] and is still widely used for laser-based fluorescence analysis of wastewater. The Raman feature due to O–H stretching modes makes the normalization convenient and reliable because of the high cross-section of the scattering, large shift, and thermal stability of the central frequency. LEDs make this kind of measurements much more cost efficient, since the price of the light source drops by 3 to 4 orders of magnitude as compared to lasers. In addition, LEDs offer higher stability, lower noise and a longer lifetime. The drawbacks that prevent the using of LEDs in conventional Raman spectroscopy (a high radiation divergence and a large bandwidth) are not essential in the particular case. The sensitivity of Raman-normalized fluorescence measurements can be substantially improved by introduction of LEDs emitting UV light, which is more favorable for excitation of fluorescence in many organic compounds and biological agents. In addition, Raman-normalization technique might be useful in LED-based fluorometers with multi wavelength excitation where fluorescence signals are excited by different-wavelength LEDs with different output power. In water purification systems based on UV LEDs, Raman and fluorescence signals excited by emission of the LEDs can be used for the control of the purification process.

A. Žukauskas, M. S. Shur and R. Gaska

140

5.

SUMMARY

Numerous advantages of LEDs over conventional sources of light and their unique properties have already been exploited in a variety of optical measurement techniques. Cumulative progress in LED-based optical measurements is expected with deeper penetration of solid-state technology into the UV region and utilization of multiple-wavelength LED arrays. In particular, novel cost-efficient bio-optical applications might be anticipated with development of LEDs emitting below 280 nm.

ACKNOWLEDGEMENT The work at Vilnius University was supported by the Lithuanian State Foundation of Science and Studies and European Commission supported SELITEC Center (contract No.G5MA-CT-2002-04047). A. Žukauskas acknowledges the Lithuanian Ministry of Education and Science for his Fellowship.

REFERENCES 1. 2.

3.

4.

5.

6.

A. Žukauskas, M. S. Shur, and R. Gaska, Introduction to Solid-State Lighting (Wiley, New York, 2002). M. R. Krames, M. Ochiai-Holcomb, G. E. Höfler, C. Carter-Coman, E. I. Chen, I.-H. Tan, P. Grillot, N. F. Gardner, H. C. Chui, J.-W. Huang, S. A. Stockman, F. A. Kish, M. G. Craford, T. S. Tan, C. P. Kocot, M. Hueschen, J. Posselt, B. Loh, G. Sasser, and D. Collins, “High-power truncated-inverted-pyramid (AlxGa1–x)0.5 In0.5P/GaP light-emitting diodes exhibiting >50% external quantum efficiency,” Appl. Phys. Lett. 75, pp. 2365–2367 (1999). H. Kudo, Y. Ohuchi, T. Jyouichi, T. Tsunekawa, H. Okagawa, K. Tadatomo, Y. Sudo, M. Kato, and T. Taguchi, “Demonstration of high-efficient InGaN-based violet light-emitting diodes with an external-quantum efficiency of more than 40%,” Phys. Stat. Sol. A 200, pp. 95–98 (2003). J. J. Wierer, D. A. Steigerwald, M. R. Krames, J. J. O’Shea, M. J. Ludowise, G. Christenson, Y.-C. Shen, C. Lowery, P. S. Martin, S. Subramanya, W. Götz, N. F. Gardner, R. S. Kern, and S. A. Stockman, “High-power AlGaInN flip-chip light-emitting diodes,” Appl. Phys. Lett. 78, pp. 3379–3381 (2001). D. A. Steigerwald, J. C. Bhat, D. Collins, R. M. Fletcher, M. Ochiai Holcomb, M. J. Ludowise, P. S. Martin, and S. L. Rudaz, “Illumination with solid state lighting technology,” IEEE J. Select. Topics Quantum Electron. 8, pp. 310–320 (2002). A. Chitnis, V. Adivarahan, J. P. Zhang, M. Shatalov, S. Wu, J. Yang, G. Simin, M. Asif Khan, X. Hu, Q. Fareed, R. Gaska, M. S. Shur, “Milliwatt power AlGaN quantum well deep ultraviolet light emitting diodes,” Phys. Stat. Sol. A 200, pp. 99–101 (2003); R. Gaska, A. Khan, M. S. Shur, this volume.

Optical Measurements Using Light-Emitting Diodes 7.

8.

9.

10.

11.

12. 13. 14.

15.

16. 17. 18. 19. 20.

21.

22.

23. 24.

141

D. Morita, M. Sano, M. Yamamoto, M. Nonaka, K. Yasutomo, K. Akaishi, S. Nagahama, and T. Mukai, “Over 200 mW on 365 nm ultraviolet light emitting diode of GaN-free structure,” Phys. Stat. Sol. A 200, pp. 114–117 (2003). L. Callegaro and E. Puppin, “Lasers and light-emitting diodes as sources for fixedwavelength magneto-optical phase modulated ellipsometry,” Rev. Sci. Instrum. 66, 5375–5376 (1995). J. Sipior, G. M. Carter, J. R. Lakowicz, and G. Rao, “Single quantum well light emitting diodes demonstrated as excitation sources for nanosecond phasemodulation fluorescence lifetime measurements,” Rev. Sci. Instrum. 67, pp. 3795– 3798 (1996). T. Araki and H. Misawa, “Light emitting diode-based nanosecond ultraviolet light source for fluorescence lifetime measurements,” Rev. Sci. Instrum. 66, pp. 5469– 5472 (1995). R. B. Galloway, D. G. Hong, and H. J. Napier, “A substantially improved greenlight-emitting diode system for luminescence stimulation,” Meas. Sci. Technol. 8, 267–271, (1997). L. Yuan, “White-light interferometric fiber-optic strain sensor from three-peakwavelength broadband LED source,” Appl. Optics 36, pp. 6246–6250. B. W. Smith, B. T. Jones, and J. D. Winefordner, “High-precision fluorimetry with a light-emitting diode source,” Appl. Spectrosc. 42, pp. 1469–1472 (1988). Karsten U, Klimant I, and G. Holst, “A new in vivo fluorimetric technique to measure growth of adhering phototrophic microorganisms,” Appl. Environ. Microbiol. 62, pp. 237–243 (1996). B. D. Mac Craith, G. O’Keeffe, C. McDonagh, and A. K. McEnvoy, “LED-based fibre optic oxygen sensor using sol-gel coating,” Electron. Lett. 30, pp. 888–889 (1994). S. J. Hart and R. D. JiJi, “Light emitting diode excitation emission matrix fluorescence spectroscopy,” Analyst 127, pp. 1693–1699 (2002). P. Kapusta, R. Erdmann, U. Ortmann, and M. Wahl, “Time-resolved fluorescence anisotropy measurements made simple,” J. Fluorescence 13, pp. 179–183 (2003). M. Trtílek, D. M. Kramer, M. Koblízek, and L. Nedbal, “Dual-modulation LED kinetic fluorometer,” J. Lumin. 72–74, pp. 597–599 (1997). J. R. Lakowicz, Principles of Fluorescence Spectroscopy (Kluwer Academic/Plenum, New York, 1999). E. Rabinovich, M. J. O’Brien, S. R. J. Brueck, and G. P. Lopez, “Phasesensitive multichannel detection system for chemical and biosensor arrays and fluorescence lifetime-based imaging,” Rev. Sci. Instrum. 71, pp. 522–529 (2000). K. Kazlauskas, G. Tamulaitis, A. Žukauskas, M. A. Khan, J. W. Yang, J. Zhang, G. Simin, M. S. Shur, and R. Gaska, “Double-scaled potential profile in a group-III nitride alloy revealed by Monte Carlo simulation of exciton hopping,” Appl. Phys. Lett. 83, pp. 3722–3724 (2003). T. Ban, H. Skenderoviü, S. Ter-Avetisyan, and G. Pichler, “Absorption measurements in dense cesium vapor using UV-blue light emitting diode,” Appl. Phys. B 72, pp. 337–341 (2000); G. Pichler, T. Ban, H. Skenderoviü, and D. Aumiler, this volume. J. P. Dakin, M. J. Gunning, P. Chambers, and Z. J. Xin, “Detection of gases by correlatioin spectroscopy,” Sens. Actuators B 90, pp. 124–131 (2003). J. Melendez, R. Carr, D. U. Bartholomew, K. Kukanskis, J. Elkind, S. Yee, C. Furlong, and R. Woodbury, “A commercial solution for surface plasmon sensing,” Sens. Actuators B 35/36, pp. 212–216 (1996).

142

A. Žukauskas, M. S. Shur and R. Gaska 25. H. P. Ho, S. Y. Wu, M. Yang, and A. C. Cheung, “Application of white lightemitting diode to surface plasmon resonance sensors,” Sens. Actuators B 80, pp. 89–94 (2001). 26. K. Gaskill, R. T. Holm, and O. J. Glembocki, “GaN light emitting diode as a photoreflectance pump source,” Rev. Sci. Instrum. 71, pp. 4341–4343 (2000). 27. A. Žukauskas, A. Noviþkovas, P. Vitta, M. S. Shur, and R. Gaska, “Raman measurements in water using a high-power light-emitting diode,” J. Raman Spectrosc. 34, pp. 471–473 (2003). 28. M. P. Bristow, D. Nielsen, D. Bundy, and R. Furtek “Use of water Raman emission to correct airborne laser fluorosensor data for effects of water optical attenuation,” Appl Optics 20, pp. 2889–2906 (1981).

NOVEL AlGaN HETEROSTRUCTURES FOR UV SENSORS AND LEDS M. STUTZMANN Walter Schottky Institut, Technische Universität München, 85748 Garching, Germany

Abstract:

The use of novel AlGaN/GaN heterostructures for UV applications is reviewed. Multiple AlGaN layers can be employed to realize spectrally selective narrow-band UV sensors. Epitaxial heterostructures of n-type AlN on p-type diamond were grown by MBE and exhibit surprisingly good electronic properties, suggesting a possible application for future UV light-emitting diodes. Finally, the use of AlGaN/GaN heterostructures for biosensors is briefly discussed.

Key words:

narrow-band UV detectors, UV light-emitting diodes, AlGaN/GaN biosensors

1.

INTRODUCTION

The ternary AlGaN-alloy system is particularly suited for optoelectronic devices in the ultraviolet spectral region. As indicated by Fig. 1, the band gap of AlGaN layers spans the entire spectral region between 3.4 (350 nm) and 6.2 eV (200 nm), thus encompassing the historically defined UVA (380–315 nm), UVB (315–280 nm), and UVC (280–200 nm) spectral ranges of ultraviolet radiation. Because of the onset of ozone formation by UV dissociation of oxygen for wavelengths smaller than 200 nm, this so-called vacuum ultraviolet (VUV) region is of less importance for the applications to be discussed in the following. Because of their tunable band gap and their direct band structure, AlGaN alloys are very favourable materials for UV sensors [1]. Here we concentrate on particular sensor structures which allow a narrow-band detection of specific UV spectral lines, e.g. the 320 nm OH-emission line for combustion control purposes, or the 250–270 nm spectral range for ozone detection and the monitoring of UV radiation which gives rise to maximum DNA damage. 143 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 143–159. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

M. Stutzmann

144

As can be deduced from Fig. 1, this corresponds to Al contents in AlGaN of 30 at.% or more. Unfortunately, for such high Al concentrations, substitutional doping of AlGaN becomes increasingly difficult and, in the case of ptype doping, actually has not been realized so far. Therefore, UV sensors in the spectral range of interest here are commonly based on photoconductor or Schottky diode structures [2]. 6.5

PIMBE AlxGa1-xN

6.0

T = 300 K

Bandgap E g [eV]

5.5

5.0

UVC 4.5

UVB 4.0

3.5

UVA

3.0 0.0

0.2

0.4

0.6

0.8

1.0

Al Content

Figure 1. Experimental values for the room temperature optical band gap of AlGaN with different Al contents. Differently shaded regions indicate the spectral ranges of UVA, UVB, and UVC radiation. PIMBE refers to the growth method, namely plasma-induced molecular beam epitaxy.

Because of the p-type doping problem, also bipolar optoelectronic devices based on AlGaN at present are limited to the range of Al concentration below about 15 at.%. This is a particular handicap for UV light emitting diodes in the UVB and UVC region. As a potential solution of this problem, we will discuss the properties of heterostructures consisting of n-type AlGaN-layers epitaxially grown on p-type diamond substrates. This combination allows to overcome the fundamental doping problems in these two wide gap semiconductor systems and, in addition, provides an attractive synergy between the possibility of UV band gap engineering in the AlGaN alloy system on one hand and the excellent thermal, chemical and mechanical properties of diamond on the other hand.

Novel AlGaN Heterostructures

145

Finally, we will have a brief look at the use of AlGaN/GaN heterostructures for future biosensor applications. A very attractive feature in this context is the optical transparency of AlGaN in the spectral range between 360 nm and 800 nm, which is commonly employed in fluorescence investigations and microscopy of biological systems. Making use of the spontaneously formed two-dimensional electron gas at AlGaN/GaN heterointerfaces, this can be combined to realize a new generation of optoelectronic sensors for the investigation of electronic and ionic processes in biology and medicine.

2.

AlGaN ULTRAVIOLET SENSORS

UV sensors based on simple AlGaN layers have already been studied and optimized extensively. Depending on the Al content of the sensitive layer, various sensor designs have been used: photoconductors, Schottky diodes, MSM detectors, p–n diodes, avalanche detectors, and phototransistors [3–5]. The spectral sensitivity of such detectors is generally determined by the optical absorption coefficient of the respective AlGaN layer and the surface recombination of photoexcited carriers in the case of strongly absorbed light. A typical example of the spectral sensitivity of AlGaN Schottky diodes with a Pt-Schottky contact is shown in Fig. 2. At the direct band gap of the respective alloy film, the sensitivity drops by several orders of magnitude and is determined by defect related absorption in the subgap region. Above the band edge, the sensitivity remains constant within an order of magnitude, depending on the particular metal used for the Schottky contact, details of the surface preparation, etc. In order to achieve a narrower sensitivity characteristics, a more sophisticated heterostructure has to be used. As shown in Fig. 3, the most simple narrow-band detector structure consists of a three-layer sequence with different Al contents. The first layer on the transparent sapphire substrate acts as an optical filter, which only transmits light with a photon energy below the respective band gap. Thus, this layer determines the high-energy cutoff of the sensitivity curve. The filter layer is followed by an electrical isolation layer with a much higher Al content. This isolation layer prevents the spillover of photoexcited carriers created in the filter layer into the uppermost photoconductor layer. The band gap of the photoconductor layer finally determines the low energy cutoff of the spectral sensitivity, similar to the curves in Fig. 2 [6].

146

M. Stutzmann

Figure 2. Spectral sensitivity of different Pt/AlGaN Schottky diodes at 300 K.

For the specific structure shown in Figure 3, the filter layer with an Al content of 40 at.% is expected to absorb all UV radiation with a photon energy above 4.15 eV. The isolation layer with an Al content of 60 at.% has a band gap of 5.1 eV and, therefore, serves as an energy barrier of many times kT at room temperature, both for the photoexcited electrons and holes. The photoconductive layer with an Al content of 33 at.% absorbs efficiently for photon energies above 3.95 eV, so that the entire three-layer-structure is expected to exhibit a significant response only in the wavelength region between 314 nm and 295 nm. Within experimental accuracy, this is confirmed by the linear spectral response curve in Fig. 4.

Novel AlGaN Heterostructures

147

Figure 3. Layer structure of a narrow-band AlGaN UV sensor. The function of the three different layers as optical filter, electrical isolation, and photoconductor is described in the text in more detail.

Figure 4. Linear spectral response curve of the AlGaN-heterostructure shown in Figure 3.

The experimental sensitivity curve shown in Fig. 4 agrees very well with the expected spectral sensitivity discussed above. A significant response is only observed for wavelengths between 290 and 310 nm. On the low-energy

148

M. Stutzmann

side, the UV sensor already exhibits a noticeable response at 340 nm due to structural imperfections in the photoconductive layer. On the high energy side, a sensitivity cutoff at about 290 nm is observed which, however, depends strongly on the thickness of the filter layer. This is shown in more detail in Fig. 5, where the spectral response of several UV sensors with different thicknesses of the filter layer is plotted on a semilogarithmic scale.

Figure 5. Normalized photocurrent (on a logarithmic scale) of UV sensor structures as the one shown in Fig. 3, with different thicknesses of the AlGaN filter layer.

As can be seen from Fig. 5, the sensor structure without a filter layer exhibits a spectral sensitivity comparable to that of the simple AlGaN Schottky diodes in Fig. 2. With increasing filter layer thickness, the narrow-band characteristics of the sensor become more and more pronounced. For the highest filter layer thickness of 1400 nm investigated here, the suppression

Novel AlGaN Heterostructures

149

of radiation detection outside of the desired wavelength region exceeds three orders of magnitude. In addition, the thick filter layer acts as a buffer layer for the final photosensitive AlGaN layer, leading to a considerable reduction of defect related absorption in the lower photon energy range [7]. The results presented above have been obtained without a systematic optimization of the structural quality and the design of the AlGaN heterostructures involved and, thus, should only be viewed as a proof of concept. Obviously, real sensor devices with much better spectral selectivity and sensitivity will be possible, if such an optimization is performed for a specific application.

3.

AlGaN/DIAMOND HETEROSTRUCTURES

Although in the last decade a lot of progress has been made in the preparation of high quality wide band gap semiconductors such as GaN, ZnO, or CVD diamond, both as homoepitaxial as well as heteroepitaxial layers, the fabrication of high efficiency light-emitting diodes (LEDs) or laser diodes (LDs) in the UVB and UVC spectral range (cf. Fig. 1) still remains an elusive dream. The reason for this setback is the general problem of the bipolar dopability of wide band gap semiconductors, which is a conditio sine qua non to transform a wide gap material of sufficient structural quality into an exciting new material for semiconductor applications. Usually, all wide band gap semiconductors can be doped quite efficiently by at least one shallow dopant. Thus, the n-type doping of GaN or ZnO is fairly straightforward, whereas the p-type doping of these materials still remains problematic or at least a nuissance for the realization of efficient bipolar devices such as LEDs. In the case of AlGaN, both n- and p-type doping become increasingly difficult with increasing Al-content. Similarly, p-type doping of diamond by boron has been known and employed for many years [8], whereas n-type doping by, e.g. phosphorous or sulphur has been very problematic and difficult to realize [9]. In particular, the application of wide band gap semiconductors for UVLEDs or LDs requires a complete understanding and control of both, n- and p-type doping, since otherwise the achievable efficiency of UV light emission will be limited by the capability to inject electrons or holes into the optically active region of such devices. The problems encountered so far with both, the n- and p-type doping of AlGaN with high Al contents thus explain why despite of the favourable electronic and optical properties of the AlGaN materials system (in particular the possibility to produce quantum wells, waveguides, Bragg mirrors, etc.) no deep UV light emitters have been realized so far.

150

M. Stutzmann

According to the present state of knowledge, in particular the p-type doping of AlGaN rapidly becomes inefficient with increasing Al concentration. Whereas GaN can be doped p-type by Mg acceptors up to free hole concentrations at room temperature of the order of 1018 cm–3, all acceptor atoms tested so far turn into deep defects already at Al contents of 20–30%. Fortunately, the situation is considerably better for n-type doping of AlGaN with Si, at least in the case of MBE grown material. As shown in Fig. 6, Si forms a shallow donor with an ionization energy increasing linearly from about 20 meV in GaN to 320 meV in AlN [10].

Figure 6. Conductivity activation energy of different MBE-grown AlGaN films doped with silicon. The dashed line shows the corresponding results for nominally undoped films containing a residual concentration of oxygen contamination.

Since the solubility of substitutional silicon in AlGaN is very high due to the similar atomic sizes and bond energies of Al and Si with nitrogen, good n-type contacts required for UVC LEDs can in principle be realized with Sidoped AlN layers. The DX-behaviour observed for Si-doped AlGaN with Al contents above 50% [10] only constitutes a problem at low temperatures (760–3000)

Irradiance (W/m2) 6.4 21.1 85.7 532 722

0.5 1.5 6.3 38.9 52.8

0.5 6.3 38.9 54.3

the UVB band (increasing) and the ozone concentration (decreasing), with the shortest wavelength significantly enhanced (amplification factor) if the ozone layer thins down. As an example, Fig. 2 shows that relationship, as determined by ground measurements in Thessaloniki [3].

Figure 2. Variation of total ozone (Dobson units) solar UV irradiance reaching the earth´s surface. They have been measured at Thessaloniki [3].

Very small changes in the ozone concentration cause significant increases in the UVB radiation reaching us [4]. This key role of ozone has been a matter of serious concern as the ozone column has been decreasing in the recent decades, and dramatically proven on the Antarctic continent. Society has been alerted about ozone depleting substances, mainly halogens.

Nitride Photodetectors in UV Biological Effects Studies

165

This stratospheric ozone layer, if being under normal conditions of temperature and pressure, would be a layer just about a 3-mm thick (300 Dobson units). Table II summarizes the well-known risks and benefits in humans of the solar UV radiation. As in all biological processes, damage and healing mechanisms go together, in a very complicated scheme. Somewhere, there is a balance between too much sun and melanoma risk, or too little sun and autoimmune disease for instance. Table II. Risks and benefits in humans of solar UV radiation. Risks Sunburn (red skin) Suntanning (immediate and delayed pigmentation) Photoaging Solar kerastoses Skin cancer (basal and squarnous cell carcinoma) Malignant melanoma Local and systemic inmune suppression Photosensitivity diseases Drug related phototoxic and photoallergic reactions Cataracts Benefits Vitamin D synthesis (bones) Phototherapy and photochemiotherapy Psychological comforts Lower risk of certain autoimmune system diseases

There is presently a significant concern about the solar UV radiation, derived from the important effects of such wavelengths on the biological ecosystem. All the biological consequences of the solar UV radiation are wavelength dependent and are characterised by unique action spectra (describing the relative effectiveness of each wavelength). Thus action spectra depend on the specific biological process being studied (skin-sunburn or erythema, DNA damage, chloroplast activity, plant damage, germicidal effects, bacteria killing, skin cancer, etc.). The UV bioaction is the weighted product of the UV irradiance by the action spectrum of the specific species or biological process being illuminated, as illustrated in Fig. 3.

E. Munoz et al

166

Figure 3. The biological weighted effect of the UV radiation depends on the specific action spectrum and on the impinging solar irradiance [3].

Some representative biological actions of the solar UV are illustrated in Fig. 4. There are also well known cases where the UV exposure produces very beneficial effects, and again these positive UV effects are wavelength

Figure 4. Some illustrative biological action spectra [3].

Nitride Photodetectors in UV Biological Effects Studies

167

dependent. Because of its practical importance for human beings, the erythema action (skin sunburning) has been extensively studied, and agreed internationally (International Commission on Illumination, CIE, Commission Internationale de l’Eclairage). Such response is the most widely used UV biological action response. Commercial instruments were developed, and this availability made that other biological actions were also monitored by using such erythema weighted detectors. Current solutions and AlGaNbased devices are described in the last part of this chapter. The key biological action is DNA damage. The various nucleotides have peak absorptions in the 240–270 nm range, resulting in DNA absorption maximum at about 250 nm. On the other hand, chloroplast activity determines plant photosynthesis. Photosynthesis is due to the visible spectrum, being stronger for blue and green photons (photoactinic region, PAR). These two basic action spectra (benchmark spectra) are shown in Fig. 5 [5]. By similar considerations, the way UV radiation interferes plant photosynthesis does not follow a single pattern. The effects of the increasing UVB radiation on agriculture have received attention worldwide [6]. In many cases, an enhanced UVB radiation causes photosynthesis to be inhibited.

Figure 5. Biological action spectra for DNA dimers and chloroplasts [5].

From Figs. 4 and 5 it is clear that although all the effects of the UV radiation are stronger at the shortest wavelengths, and these effects decrease as photon energy becomes smaller, the details depend markedly on the specific process, the specific species being studied and on their previous UV history.

E. Munoz et al

168

Living species have developed their own UV protection schemes (furs, pigments, flavonoids in plants, mycosporinelike aminoacids in algae, etc.). At the medical level, the main concerns about UV radiation refer to its effects on the skin and the eyes. The importance of erythema action lies in the fact that it may be the first step towards skin cancer (Fig. 6) [7]. The action spectrum of human skin cancer has revealed to have a rather non-monotonic structure (Fig. 7), probably due to the complex interactions of the repairing mechanisms. Information about the daily erythema dose (for how long a sunbath can be taken), is being provided to the public, in several countries, through the UV index [8,9]. To give some numbers, but for each individual there are various modifying factors (skin type, frequency of UV dose, etc.), the minimum erythema dose for red skin is 210 J/m2 . Ranking from 1 to 16, this UV index is obtained from the weighted erythematic radiation, in steps of 25 mW/m2 , just multiplying by 40.

Figure 6. Skin cancer cascade [7].

The absorption of UV radiation in the various regions of the eye is displayed in Fig. 8 [10]. The lens absorbs strongly the UVA and remaining UVB regions, leading to cataracts. In the search for new, efficient lighting sources, it is important to minimize any hazard coming from the residual UV emitted from the lamp. In this context, it is important to notice that semiconductor light sources do not follow any black body radiation scheme, and even white LEDs have a very much smaller UV hazard as compared to most traditional white light sources [11].

Nitride Photodetectors in UV Biological Effects Studies

169

Figure 7. Important UV action spectra for human beings.

Figure 8. UV absorption of the cornea, aqueous humour and lens [10].

At this point, one may conclude that the UV radiation is a key one for life, its damage and healing. Since the big bang, in the long process for life generation, organisms have been adapting to the solar radiation, in fact, to Earth’s atmosphere. By natural selection, genetic variants, repairing enzymes, etc., life today has adapted to the radiation received from the Sun. There is a perfect coupling between ozone absorption and the wavelengths

E. Munoz et al

170

that severely damage DNA (250 nm). Living species have developed their own protection and healing–repairing mechanisms; life has adapted [12]. One may also recall that clean water absorption has the lowest absorption region in the 580–380 nm band, approximately [13]. This is important for marine life, amphibians, and also for detection of contaminants.

3.

UV RADIATION IN BIOPHOTONICS TECHNIQUES

The role of optical techniques in biology and medical research has increased spectacularly in recent years. Being non-invasive in nature, and allowing all wonders of digital image processing, biophotonics is a leading field in life sciences. Fluorescence is a key technique today. Well known marking dyes have been joined by fluorescent proteins (initially, the green fluorescent protein was found in aequora victoria fish), fluorescent microspheres (filled with dyes), and semiconductor quantum dots to allow imaging of nucleic acids, protein tagging, cell marking, immunodiagnostic assays, DNA damage measurements, cancer cell imaging, etc. (Fig. 9) [14]. Concerning UV, one is limited by its potential damage indicated in the previous section. Intrinsic UV fluorescence from DNA seems to be rather weak, but it has been shown that the presence of metallic particles enhances DNA UV emission, probably due to charge transfer mechanisms [15]. In tagging applications,

Figure 9. The excitation and emission spectra of enhanced blue, cyan, green, yellow and DsRed fluorescent proteins [14].

Near-UV photons, typically above 350 nm, are used as excitation (absorption) for blue-green markers. Its helps to significantly broaden the range of fluorescence colours, allowing multi-tagging applications and softening

Nitride Photodetectors in UV Biological Effects Studies

171

bandpass filter requirements. Hence, nitride UV and blue sources open new possibilities in biophotonics. The use of semiconductor quantum dots (QD) as fluorophores has created a lot of expectations as reviewed in [16], and have started to be offered commercially. Quantum dots offer high photoluminescence response, long life, easy to excite, stability, and the possibility of tuning emission colours as needed (by size). Difficulties are seen as individual quantum dots have to be produced, later to be covered with a hydrophilic coating and carboxylic acid groups for bioconjugation, so they can be incorporated into the cells or used in DNA sequencing (Fig. 10). Being initially prepared from II-VI semiconductors (CdSe, 3–6 nm for instance), other schemes like AlGaN/GaN nanocolumns grown by MBE may open new possibilities for semiconductor nanotechnology in biophotonics [17] by extending the emission range into the UV.

Where both nitride UV emitters and detectors may have a strong role is

Figure 10. Unkown DNA containing fluorescent dyes is analyzed by being mixed with microbeads containing a color code of quantum dots. DNA that matches the sequence on the outside of the microbead sticks to it [16].

Where both nitride UV emitters and detectors may have a strong role is in lab-on-a-chip, DNA/protein microarrays. The micro optics integration (arrays) of LEDs, LDs and photodetectors, from the near UV to the green region, all based on nitrides, will have benefits in multifunctional microarrays, softening also optical filter requirements. Some views are given in [18]. Figure 11 shows AlGaN/GaN nanocolumns, with heights in the 200-nm range and 10–20 nm in diameter, grown by MBE under nitrogen rich conditions [17]. Smaller sizes and proper coating of such individual nanocolumns seems to be plausible.

172

E. Munoz et al

Figure 11. AlGaN/GaN/AlGaN nanocolumns grown by MBE [17].

4.

AlInGaN-BASED PHOTODETECTORS IN BIOPHOTONICS APPLICATIONS

In the two previous sections, a range of biophotonic applications for nitride UV and VIS detectors has been delineated: UV biological action studies, fluorescence from dyes and proteins as markers, photoluminescence from quantum dots, detection by fluorescence of hazardous species and contaminants, etc. All they impose different requirements on the photodetectors to be used. From the wavelength viewpoint, detection of chemicals and biospecies by fluorescence goes down to 250 nm or even lower, while in tagging techniques, damage issues advice not to go lower than 320 nm. Solar UV biological action studies should cover the whole UVA, UVB bands (280 nm). In the UVB range, AlGaN detectors have been reported with very good responsivity and noise characteristics. Let us also remember that fluorescence emission is usually made in the visible for marking applications. Other differences come from the detector frequency or time domain operation. Solar biological effects studies are made under quasi-dc conditions, being stability and reproducibility for long periods of time the main requirements. Besides, sensitivity is usually not a problem, the solar signal is not difficult to detect. On the other hand, fluorescence imaging and fluorescence detection of species require photodetectors working under ac conditions and a high sensitivity is needed. Time-domain fluorescence and fluorescence resonance energy transfer (FRET) have also revealed as very interesting techniques in biophotonics. To consider detector structures for the above applications, let us summarize some issues about nitride UV photodetectors that would help to bind

Nitride Photodetectors in UV Biological Effects Studies

173

their applications in biophotonics. Various detailed reviews on the developments of GaN UV photodetectors has been published recently [19]. Photoconductive, p–n junction, p–i–n devices, Schottky barrier (SB), metal– semiconductor–metal (MSM), metal–insulator–semiconductor structures (MIS), phototransistors, detector arrays, avalanche detectors, etc., have already been reported. UV imaging using AlGaN arrays has also been clearly demonstrated. All these developments have been tightly linked with progresses in the quality of the AlGaN materials, and with the development of reproducible processing technology. The lack of appropriate substrates is a pending problem in the nitrides device effort. From the physics viewpoint, photodetectors work under very low currents and very low photon fluxes, usually, and they reflect even more pronouncedly some of the current problems in AlGaN layer quality, mainly the presence of dislocations. In AlGaN layers slow trapping processes are usually present at very low currents, giving rise to persistent photoconductivity effects (PPC). This PPC is not yet fully understood, and it may show up in AlGaN photodetector devices, depending on their contact structure, materials quality and device operation. In fact, while junction photodiodes are fast with a good UV/VIS contrast and a linear operation, AlGaN photoconductors are slow, have internal large gain, non-linear operation and a poorer UV/VIS contrast. A hybrid behaviour is found in many cases in photoconductive and MSM structures. Details of the metal contacts, free carrier concentration and dislocation density cause to have photodetectors where some significant photoconductive gain coexist with a reasonable speed and a good UV/VIS contrast. On the other hand, Schottky, MSM and p–n junction detectors may show photoconductive gain and deviations from linearity under reverse bias conditions, reflecting also processing and layer quality conditions [19]. The (InGaNAl)N detectors shown below are adequate to properly cover the UVA and the UVB region, to match any weighted action spectrum. Fluorescence ac techniques require high sensitivity detectors. It is interesting to consider gain mechanisms in AlGaN detectors. Two mechanisms have been studied, photoconductive gain and avalanche multiplication. Photoconductive gain is easy to achieve in AlGaN photodetectors, although its control and reproducibility is poor. Photoconductors and MSM hybrid devices are thus adequate for fluorescence applications, although some external optical filter may be required to increase UV/VIS contrast. Avalanche multiplication requires further developments, and only small area devices have been reported. In biological effects studies, erythema weighted detectors are widely used for solar UV monitoring. One of the reasons has been the availability of commercial solutions presently achieved by using low gap photodiodes (Si, GaAs, GaAsP), and a series of filters and phosphor coatings to be inserted in

174

E. Munoz et al

the optical path (Fig. 12) trying to match the erythema weighting function (Fig. 13) [20]. All these elements make the sensor system bulky, less reliable, needing a temperature-controlled chamber, more expensive (although affordable) and prone to degradation. This technical solution for sun-burning UV detectors was already suggested in 1976. Today, as shown in Fig. 13, by proper layer composition and shallow centers control, the erythema weight action is approximated by a single AlGaN photodiode, with about 30% of Al composition, and no filters [20]. Note that a key region is the device response below the band edge. In this application, not the best AlGaN epilayer has to be used. This is the most important point to insure spectral reproducibility in the erythema detector fabrication. A key nitride advantage is its radiation hardness. In solar monitoring, UV detectors receive a significant UVA and UVB dose that tend to degrade optical filters (solarization), as required in erythema radiometers shown in Fig. 12. The stability and reproducibility requirements of erythema detectors are very stringent in the UVA region. Besides, there is a need for nitride-based UVA and VIS detectors. (In,Ga,Al)N quantum well-based (QW) photodetectors are being developed at ISOM in order to benefit from the larger absorption of QWs and from the compatibility with QW emitters. Figure 14 (b) shows the spectral responses of MOVPE QW photodetectors with detection edges in the blue and the UVA band, respectively. The layer structure shown in Fig. 14 (a) allows the emitter–detector integration in microarrays. The use of QW (In,Ga,Al)N photodetectors allows developing multidetector (multiband) sensors, able to match any spectral weighting function (action spectrum) needed for quantifying effects of UV radiation. In nitride-based high-sensitivity detectors for fluorescence applications, significant advances in materials quality and device processing have been produced. As reviewed in [19], small area MSM and p–i–n detectors with time responses in the 300 ps range have been reported, internal quantum efficiencies of 86% have been achieved, very low noise structures with detecti-

Nitride Photodetectors in UV Biological Effects Studies

175

Figure 12. Optical filters and phosphorous layer required to achieve the erythema weighting action spectrum, due to the use of low bandgap semiconductor photodiodes [20].

UV-B

UV-A

Relative Response

100

CIE Erythema Action Commercial Detector AlGaN Detector

10-1

10-2

10-3

10-4 260

280

300

320

340

360

380

Wavelength (nm) Figure 13. Erythema weighting action spectrum achieved with AlGaN photodetectors (Au semitransparent Schottky diodes), and no other optical component [21].

E. Munoz et al

176

p - PAD pPAD

1

MQW n-GaN Substrate( ) Sapphire

0,01

Responsivity (a.u.)

GaN AlxGa1-xN

n- PAD nPAD

A4Y 1-7 FU145B 2-5

0,1

1E-3 1E-4 1E-5 1E-6 1E-7 1E-8 210

240

270

300

330

360

390

420

450

480

Wavelength (nm) a)

b)

Figure 14. a) The generic structure of InGaN/AlGaN MQW p–i–n diodes is shown. Three QWs have been mainly used. b) The spectral responsivity of devices, intended to have their detection edge in the blue region and at 400 nm, respectively.

vities in the 1013 cm·W–1 Hz1/2 range have been fabricated, and UV/visible contrasts between three and five orders of magnitude are today obtained in various laboratories. These devices are quite appropriate for biophotonics applications.

5.

CONCLUSIONS

There is a perfect coupling between ozone absorption and the wavelengths that severely damage DNA (250 nm). All the biological consequences of the solar UV radiation are wavelength dependent and are characterised by unique action spectra (describing the relative effectiveness of each wavelength). Representative action spectra have been presented. Erythema action (red skin) detectors are reference detectors in a variety of medical and biological effects studies. Their nitride implementation (AlGaN) is commercially interesting. For other action spectra, to fit any spectral weighting function, multiband (In,Ga,Al)N detectors, properly weighted channels, are needed. Radiation hardness offered by nitrides is an important point. A range of biophotonic applications for nitride UV and VIS detectors has been delineated. Fluorescence is presently a key tool in medical and biological research, and III-nitrides offer the unique possibility to integrate a range of emitters and detectors from the green to the UV, softening also filter requirements. Lab-on-a-chip, DNA and protein microarrays will benefit from such integration based on nitrides. Significant advances in materials quality

Nitride Photodetectors in UV Biological Effects Studies

177

and device processing have been produced, but extra efforts are needed to achieve gain for high-sensitivity fluorescence detection. Nitrides nanotechnology may also open new possibilities in fluorescent tagging applications.

REFERENCES 1. 2. 3.

4. 5.

6. 7. 8. 9. 10. 12. 13. 14. 15. 16. 17. 18. 18. 19. 20. 21.

M. Razeghi and A. Rogalsky, J. Appl. Phys. 79, 7433 (1996). University of Thessaloniki, LAP; Greece. P. C. Simon and C. S. Zerefos, “UV radiation at the Earth´s surface”, in The contribution of EASOE and SESAME to our current understandingt of the ozone layer, European Commission, DG XII, EUR 16986 (1997). G. Seckmeyer, Instruments to measure solar ultraviolet radiation. Part I, WMO TD No. 1066 (2000). P. J. Neale, “Spectral weighting functions for quantifying effects of ultraviolet radiation in marine ecosystems” in S.J. de Mora et al, Effects of UV Radiation on Marine Ecosystems (Cambridge Univ. Press, 2000). D. S. Bigelow, J. R. Slusser, A. F. Beaubien, and J. H. Gobson, Bulletin of the American Meteorological Soc., 79, 601–615 (1998). T. B. Fitzpatrick, J. Dermatol, 23, 616–20 (1996). Report on the WMO-WHO Meeting of Experts on Standardization of UV Indices and their Dissemination to the Public, WMO/TD-No. 921, WMO, Geneva, 1998. International Commission on Non-Ionizing radiation Protection, Global Solar Index, ICNIRP-1/95 (1995). World Health Organization, The effects of solar UV radiation on the eye (WHO, Geneva, 1994). A. Zukauskas, M. Shur, and R. Gaska, Introduction to Solid State Lighting (Wiley, New York, 2002). J. Withgott, Natural History 38 (2001). R. A. J. Litjens, T.I. Quikenden, and C.G. Freeman, Appl. Optics 38 1216–1223 (1999). Clontech Laboratories Inc., Palo Alto, California, USA. J.R. Lakowicz, I. Gryczynski, Y. Shen, J. Malicka, and Z. Gryczynski, Photonics Spectra 96–104 (2001). B. D. Butkus, editor, Biophotonics Intl., p. 60 (Dec. 2001) and p. 68 (Dec. 2002). J. Ristic, M.A. Sánchez-García, J.M. Ulloa, E. Calleja, J. Sanchez-Paramo, J.M. Calleja, U. Jahn, A. Trampert, K.H. Ploog, Phys. Stat. Solidi (a) 234, 717 (2002). N. D. Lamontagne, editor, Biophotonics Intl., 42–46 (Jan-Feb 2003). E. Muñoz, E. Monroy, J.L. Pau, F. Calle, F. Omnès, and P. Gibart, J. Phys.: Condens. Matter 13, 7115–7137 (2001). Yankee Environmental Systems, Inc., Turner Falls, MA 01376 USA. E. Muñoz, E. Monroy, F. Calle, F. Omnès, and P. Gibart, J. Geophys. Res. 105, 4865 (2000).

PROMISING RESULTS OF PLASMA ASSISTED MBE FOR OPTOELECTRONIC APPLICATIONS A. GEORGAKILAS, E. DIMAKIS, K. TSAGARAKI, and M. ANDROULIDAKI Microelectronics Research Group (MRG), Institute of Electronic Structure and Laser (IESL), Foundation for Research and Technology-Hellas (FORTH), P.O. Box 1527, 71110 Heraklion-Crete, Greece and University of Crete, Physics Department, Heraklion-Crete, Greece

Abstract:

Plasma-assisted MBE (PAMBE) is not yet considered as an epitaxial growth method that could produce device quality material for optoelectronic applications. However, several results suggest that PAMBE can grow high quality heterostructures on GaN templates and even directly on sapphire substrates. High quality N-face material can be probably grown on sapphire only by PAMBE. N-face GaN is grown after sapphire nitridation at low temperature and a layer-by-layer growth seems to occur immediately after the initiation of the GaN growth on the sapphire surface. Recently, high quality In-containing alloys, such as quaternary InAlGaN alloy thin films and quantum well heterostructures, have also been grown by PAMBE with good control of the In incorporation. The layers exhibited strong PL up to room temperature and lasing under optical pumping.

Key words:

GaN, III-nitrides, III-nitride semiconductors, MBE, plasma-assisted MBE, InGaN, InAlGaN, quaternary alloys, N-face GaN, sapphire nitridation, GaN polarity

1.

INTRODUCTION

III-nitride devices for light emission and detection have been commercialized using heterostructure materials grown by metalorganic vapor phase epitaxy (MOVPE). Molecular Beam Epitaxy (MBE) of III-nitrides has been investigated from the early stages of III-nitrides research [1–4] but the method has not yet been developed to the level required for device applica179 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 179–188. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

180

A. Georgakilas et al

tions. Two different techniques of III-nitride MBE growth can be distinguished, depending on the type of the source used to supply active nitrogen species to the substrate; the plasma assisted MBE (PAMBE) and the NH3 gas source MBE, also called as reactive MBE (RMBE). PAMBE uses a compact plasma source to activate N2 gas, such as the electron cyclotron resonance (ECR) microwave source or the radio frequency (RF) source. ECR sources [1,2] were used frequently in the first MBE experiments but the inductively coupled RF plasma sources [3,4] soon dominated due to the reduced ion damage in the grown layers. An extensive review of the PAMBE growth of III-nitride materials has been given elsewhere [5]. This paper intends to present some important aspects of PAMBE and recent results suggesting that PAMBE with nitrogen RF plasma source (RFMBE) could be used efficiently for the growth of state-of-the-art IIInitride heterostructure materials [5–14]. Three different growth tasks will be considered: (a) homoepitaxial GaN growth [5,6], (b) heteroepitaxial GaN growth on Al2O3 (0001) [11–17], and (c) growth of InxAlyGa1–x–yN alloys and quantum well heterostructures [12–14].

2.

GALLIUM NITRIDE HOMOEPITAXY

GaN substrates have yet not been commercially available like in the conventional III-V semiconductors. However, significant experience for the GaN PAMBE growth has been gained by using templates consisting of GaN thin films grown on Al2O3 (0001) or SiC (0001) substrates by MOVPE or hydride vapor phase epitaxy (HVPE) [5]. Such experiments have helped to understand the fundamentals of GaN PAMBE growth. It has been well documented both theoretically [15] and experimentally [16] that the GaN growth mode depends on the concentration of surface adatoms. A surface Ga bilayer [15,16] can be formed using a Ga/N flux ratio above unity and favors the layer-by-layer growth of GaN by the step-flow growth mechanism. Figure 1 shows an AFM micrograph for a 0.5-µm GaN layer grown by RFMBE on a ~2-µm GaN/Al2O3 (0001) MOVPE template. The Ga/N flux ratio was ~1.3 and the substrate temperature was 700 oC. Under these conditions, step-flow growth occurs with no accumulation of any Ga droplets on the surface. The surface rms roughness was 0.26 nm, quite comparable to that of the initial MOVPE material. It has been found that the crystalline defects of a RFMBE GaN film are continuation of defects propagating from the GaN template [5]. No additional dislocations are formed at the interface, although several reports suggest that residual impurities may remain on the GaN surface after the usual substrate cleaning treatments [17].

Promising Results of Plasma Assisted MBE

181

Figure 1. AFM micrograph of 0.5 µm GaN grown by PAMBE under step-flow growth mode on a ~2.0 µm GaN/Al2O3 (0001) MOVPE template. The scan size is 2 × 2 µm and the z-axis full scale is 5 nm.

MOVPE template

(a) 3

2 Oexc=325nm

T=17K

1 3.35

3.40

3.45 3.50 Energy(eV)

3.55

3.60

Photoluminescence (a.u.)

Photoluminescence (a.u.)

4 3

MBE sample

(b)

2 Oexc=325nm

T=17K

3.35

3.40

3.45 3.50 Energy(eV)

3.55

3.60

Figure 2. 17-Ȁ PL spectra for (a) ~2-µm GaN/Al2O3 (0001) MOVPE template and (b) 0.5-µm GaN grown by RFMBE on the MOVPE template.

Finally, the low temperature photoluminescence (PL) spectra of the RFMBE layers exhibit excitonic peaks with full width at half maximum (FWHM) values similar to those of the MOVPE templates, as shown in Fig. 2. This indicates that the RFMBE material quality is not limited by ioninduced damage.

182

3.

A. Georgakilas et al

HETEROEPITAXY ON (0001) SAPPHIRE

PAMBE is generally inferior of MOVPE in the field of direct GaN heteroepitaxial growth on a different substrate, such as the commonly used Al2O3 (0001). The PAMBE weakness, according to the authors’ opinion, is linked to two important stages of GaN/Al2O3 heteroepitaxy: (1) the GaN nucleation and (2) the buffer layer growth. GaN films with the (0001) orientation or Ga-polarity are generally preferable than films with the (000-1) orientation or N-polarity, because they can be grown easily with smooth surface morphology (Fig. 1) and the direction of polarization fields is convenient for the fabrication of AlGaN/GaN high electron mobility transistors (HEMTs) [6]. MOVPE seems able to nucleate easily high purity Ga-face material, possibly as a result of significant Al2O3 nitridation under exposure in NH3 flux at high substrate temperature. In addition, large three-dimensional (3D) GaN islands may be initially nucleated on the Al2O3 (0001) surface [18] but eventually a flat GaN/Al2O3 buffer layer can be grown, apparently due to the high lateral growth speed and adatom surface mobility which is possible in MOVPE. This allows to increase the size of the initial crystalline grains (3D islands) and so to increase the percentage of the good crystallinity material that occurs within the grains. In PAMBE, we have found [5,10,11] a significant difference in the amount of sapphire nitridation at the high nitridation temperature (HNT) of ~750 qC and the low nitridation temperature (LNT) of ~200 qC and that this controls the polarity of overgrown GaN layers, when a GaN buffer layer is used. A substantial surface nitride layer, with average thickness of 1.5 nm, was formed only after 100 min nitridation at HNT, while the nitridation should be limited to a surface atomic plane at LNT. The HNT results to Gaface material, while the LNT results to N-face GaN. The Ga-face polarity is attributed to the formation of an AlN/sapphire interface by nitridation at HNT (although a GaN buffer layer is then used), since Ga-face material is also grown when an AlN buffer layer is used on a substrate nitridated at LNT. We have grown, by RFMBE, Ga-face GaN/Al2O3 films with surface smoothness [5] and PL comparable to MOVPE. Figure 3 shows the 17-K PL spectrum of a 1.8-µm GaN/Al2O3 sample. The FWHM of the excitonic peak is 7 meV and there is no deep levels’ PL below 3.2 eV. However, it has been found rather difficult [5,7–9] to achieve reproducibly high purity, low defect density and smooth Ga-polar films, similar to MOVPE. The higher density of threading dislocations may be the result of the misfit strain relaxation during a nucleation-growth process that deviates from the ideal layer-by-layer MBE growth.

Photoluminescence (a.u.)

Promising Results of Plasma Assisted MBE

183

6

4 7 meV

2 Oexc=325nm

T=17K

0 3.0

3.1

3.2 3.3 3.4 Energy(eV)

3.5

3.6

Figure 3. 17-K PL spectrum of a 1.8-µm Ga-face GaN layer grown by PAMBE on Al2O3 (0001) using a 50-nm AlN buffer layer.

N-face GaN is grown on sapphire nitridated at LNT, using a GaN buffer layer [10,11]. A streaky RHEED pattern is observed continuously during GaN nucleation, indicating a layer-by-layer growth mode. Transmission electron microscopy (TEM) studies revealed [5,11] that the N-face layers have a low defect density compared to the usual structure of Ga-face GaN [8,9]. “Cubic pockets” were observed in the initial GaN buffer layer, while no such cubic phase material was monitored in all the Ga-face samples. However, a significantly lower density of threading dislocations and inversion domain boundaries (IDBs) appeared in the N-face samples, indicating that cubic GaN embedded in the hexagonal matrix is related to the reduction of the defect content of the overgrown layers. The TEM results were in agreement with high-resolution X-ray diffraction (HRXRD) measurements (Fig. 4). HRXRD rocking curves for the symmetric (00.2) and asymmetric (11.4) reflections of 1.3-µm N-face GaN/Al2O3 (0001) exhibited record low values of FWHM, equal to 54 and 135 arcsec, respectively. These results suggest that PAMBE is particularly suitable for the growth of N-face GaN/Al2O3 and this has not yet been exploited.

A. Georgakilas et al

184

2.5

(a)

2.5

Intensity (x 10 cps)

(00.2) FWHM = 54 arcsec

2.0 1.5 1.0 0.5

2.0

(11.4) FWHM = 135 arcsec

1.5 1.0 0.5 0.0

0.0 -1000

(b)

4

5

Intensity (x 10 cps)

3.0

-500

0 500 ǻȦ (arcsec)

1000

-1000

-500

0 500 ǻȦ (arcsec)

1000

Figure 4. X-ray diffraction Ȧ-scans for the reflections: (a) symmetric (00.2) and (b) asymmetric (11.4) in grazing incidence, for a 1.3-µm N-face GaN/Al2O3 (0001) layer

4.

GROWTH OF INDIUM-CONTAINING ALLOYS AND QUANTUM WELLS

The growth of In-containing alloys, InxAlyGa1–x–yN and quantum well heterostructures may be the strongest point of PAMBE. The energetic nitrogen species that are supplied by the plasma source make possible to grow IIInitrides at very low substrate temperatures, where the thermal decomposition of NH3 gas would be impossible. This facilitates the incorporation of In, which is limited from the low thermal stability of InN. We have investigated the temperature dependent incorporation of In atoms in quaternary InxAlyGa1–x–yN alloys [12]. The fluxes of Al (FAl) and Ga (FGa) atoms were kept constant and corresponded to an AlGaN composition (without incorporation of In) of approximately Al0.40Ga0.60N. The flux of active nitrogen species (FN) was also kept constant and it was higher than the growth rate of Al0.40Ga0.60N. The incorporation of In atoms in the InxAlyGa1–x–yN alloys was a sensitive function of the growth temperature, as shown in Fig.5, which gives a plot of the In mole fraction (x) as determined by RBS versus the growth temperature in the range of 510–610 ºC. The results show that the amount of incorporated In flux matches the excess nitrogen flux {FN – (FAl + FGa)} at a substrate temperature of 510–530 ºC. The preferential incorporation of Al and Ga atoms compared to In atoms is expected considering the relative strength of their bonds with N [19]. It should be noticed, however, that a recent careful analysis suggests that additional In atoms may be incorporated instead of Al and Ga atoms at temperatures lower than ~530 ºC [20].

Promising Results of Plasma Assisted MBE

185

We have studied also the control of the InxAlyGa1–x–yN composition by the flux of In atoms, which is controlled by the temperature of the In cell. As shown in Fig. 6, for a constant substrate temperature of 540 ºC and an N flux of 2.6u1013 cm–2s–1 the percentage of the incoming In flux that is incorporated is constant and equal to ~17%.

In mole fraction

0.10 0.08 0.06 0.04 0.02 0.00 500

520 540 560 580 600 620 o Growth Temperature ( C)

640

Figure 5. Dependence of the In mole fraction in quaternary layers of InxAl0.40(1–x)Ga0.60(1–x)N on the growth temperature. The alloy composition was determined by RBS. (Reprinted from Journal of Crystal Growth, Vol. 251, E. Dimakis, A. Georgakilas, M. Androulidaki, K. Tsagaraki, G. Kittler, D. Cengher, E. Bellet-Amalric, D. Jalabert, N.T. Pelekanos, “Plasma-Assisted MBE Growth of Quaternary InAlGaN Quantum Well Heterostructures with Room Temperature Luminescence”, pages 476 – 480, Copyright 2003, with permission from Elsevier.)

Further work has leaded to the optimization of the growth of InxAlyGa1–x–yN alloy thin films and quantum well heterostructures [12–14]. Ternary and quaternary alloy layers with smooth surfaces and excellent optoelectronic properties were grown at 530–540 oC. Indicative 17-K PL results for different InxAlyGa1–x–yN alloy compositions are given in Fig. 7. Both ternary and quaternary alloys exhibited very strong band edge luminescence with the best FWHM values of 60–65 meV for emission at 3.0–3.15 eV, which are record low values for PAMBE grown alloys. As shown in Fig. 7, we have grown ternary InxGa1–xN alloys with x up to 0.26 without any evidence of phase separation, according to HRXRD measurements. The In0.26Ga0.74N layer was characterized by PL emission at 2.46 eV and FWHM of 127 meV. For quaternary InxAlyGa1–x–yN alloys, we have reached reproducibly In composition up to x = 0.135 for y = 0.26–0.28 and these layers exhibited 17-K PL emission at ~3.1 eV with FWHM of 64–68 meV. The alloys also maintained strong PL emission at room temperature what indicates relatively weak nonradiative processes in the material.

A. Georgakilas et al

186

incorporated In flux non-incorporated In flux

-2 -1

5

13

In flux (x10 cm s )

6

4 3

13

-2 -1

N-flux: 2.6x10 cm s

2 1 0

2

3 4 5 13 6-2 -1 Incident In flux (x 10 cm s )

7

Photoluminescence(a.u)

Figure 6. The incorporated In flux (squares) and the non-incorporated In flux (circles) have been plotted versus the incident In flux on the substrate surface, for constant N flux of 2.6u1013 cm–2s–1.

x=0.001, y=0.37

x=0.07, y=0.33

x=0.13, y=0.28 x=0.26, y=0

Oexc=325nm T=17K

2.0 2.4 2.8 3.2 3.6 4.0 4.4 4.8 Energy(eV)

Figure 7. 17-K PL spectra of different InxAlyGa1–x–yN samples grown by RFMBE.

The quaternary InxAlyGa1–x–yN alloys exhibited a particularly large value for the In bowing coefficient [13]. This means that the bandgap of the InAlGaN alloys could be larger or smaller than that of GaN, depending on the In mole fraction of InAlGaN. Multiple Quantum Well (MQW) GaN/InAlGaN structures were also grown, where the well material was either GaN or quaternary InAlGaN. The growth was optimized to ensure the optimum surface smoothness of all the layers and the abruptness of the interfaces [12]. HRXRD [12] and highresolution TEM [21] measurements confirmed the good quality of the MQW structures. The polarization-induced electric fields within the GaN/InAlGaN QWs were also investigated and the potential for zero field QWs was demon-

Promising Results of Plasma Assisted MBE

187

strated, as a consequence of polarization matching between the InAlGaN and GaN layers [13,14]. Finally, laser emission under optical pumping at room temperature has been obtained on structures containing several InAlGaN active QWs [12,14]. The threshold power density for the InAlGaN/GaN QWs was lower than that of comparable GaN/AlGaN QWs, indicating the positive effect of reduced internal electric field [14].

5.

CONCLUSIONS

We have discussed several results suggesting that PAMBE is a technique with great potential for the development of state-of-the-art III-nitride heterostructure-nanostructure materials, similarly to what the MBE method used to be in the development of other III-V semiconductors. However, full exploitation of its capabilities requires the availability of GaN substrates, which are under development with a variety of technological approaches. A significant advantage of PAMBE is anticipated in the area of In-containing alloys and heterostructures. In addition, heteroepitaxial N-face GaN/Al2O3 (0001) of excellent structural quality can be probably grown only by PAMBE.

ACKNOWLEDGEMENTS Our work has been funded from several programs of the General Secretariat for Research and Technology (GSRT) of the Greek Ministry of Development and from the European Commission projects IST-FET-26464 and 38982. Support from University of Crete is also acknowledged. We are also grateful to N. Pelekanos, G. Konstantinidis, Ph. Komninou, Th. Karakostas, M. Calamiotou, E. Bellet-Amalric and D. Jalabert for their valuable collaboration.

REFERENCES 1. 2. 3. 4.

T. Lei, M. Fanciulli, R. J. Molnar, T. D. Moustakas, R. J. Graham, and J. Scanlon, Appl. Phys. Lett. 59, 944 (1991). R. J. Molnar and T. D. Moustakas, J. Appl. Phys. 76, 4587 (1994). W. E. Hoke, P. J. Lemonias and D. G. Weir, J. Cryst. Growth 111, 1024 (1991). J. M. Van Hove, G.J. Cosmini, E. Nelson, A. M. Wowchak, P. P. Chow, J. Cryst. Growth 150, 908 (1995).

A. Georgakilas et al

188 5.

6. 7.

8.

9.

10. 11.

12.

13.

14.

15. 16. 17. 18. 19. 20. 21.

A.Georgakilas, H. M. Ng and Ph. Komninou, in Nitride Semiconductors, Handbook on Materials and Devices, ed. by P. Ruterana, M. Albrecht, and J. Neugebauer, Chapter 3 (2003, Wiley-VCH, Berlin, 2003), pp. 107–191. M. Zervos, A. Kostopoulos, G. Constantinidis, M. Kayambaki, and A. Georgakilas, J. Appl. Phys. 91, pp. 4387–4393 (2002). K. Amimer, A. Georgakilas, M. Androulidaki, K. Tsagaraki, M. Pavelescu, S. Mikroulis, G. Constantinidis, J. Arbiol, F. Peiro, A. Cornet, M. Calamiotou, J. Kuzmik, and V. Yu. Davydov, Mater. Sci. Eng. B 80, pp. 304–308 (2001). G. P. Dimitrakopoulos, Ph. Komninou, J. Kioseoglou, Th. Kehagias, E. Sarigiannidou, A. Georgakilas, G. Nouet and Th. Karakostas, Phys. Rev. B 64, 245325 (2001). Ph. Komninou, Th. Kehagias, Th. Karakostas, G. Nouet, P. Ruterana, K. Amimer, S. Mikroulis and A. Georgakilas, in Proceedings of the 2000 MRS Fall Meeting, Symposium G: GaN and Related Alloys (November 27 – December 1, 2000, Boston, USA) Mat. Res. Soc. Symp. Proc. 639, G3.47 (2001). S. Mikroulis, A. Georgakilas, A. Kostopoulos, V. Cimalla, E. Dimakis, and Ph. Komninou, Appl. Phys. Lett. 80, pp. 2886–2888 (2002). A. Georgakilas, S. Mikroulis, V. Cimalla, M. Zervos, A. Kostopoulos, M. Androulidaki, Ph. Komninou, Th. Kehagias and Th. Karakostas, Phys. Stat. Sol. (a) 188, pp. 567–570 (2001). E. Dimakis, A. Georgakilas, M. Androulidaki, K. Tsagaraki, G. Kittler, D. Cengher, E. Bellet-Amalric, D. Jalabert, N.T. Pelekanos, J. Cryst. Growth 251, pp. 476–480 (2003). M. Androulidaki, N. T. Pelekanos, E. Dimakis, F. Kalaitzakis, E. Aperathitis, F. Bellet-Amalric, D. Jalabert, K. Tsagaraki, and A. Georgakilas, Phys. Stat. Solidi (c) 0, pp. 504–507 (2002). F. Kalaïtzakis, M. Androulidaki, N. T. Pelekanos, E. Dimakis, E. Bellet-Amalric, D. Jalabert, D. Cengher, K. Tsagaraki, E. Aperathitis, G. Konstantinidis, and A. Georgakilas, Phys. Stat. Sol. (a) 195, 2003 (in print). J. E. Northrup, J. Neugebauer, R. M. Feenstra, and A. R. Smith, Phys. Rev. B 61, 9932 (2000). B. Heying, R. Averbeck, L. F. Chen, E. Haus, H. Riechert, and J. S. Speck, J. Appl. Phys. 88, 1855 (2000). S. W. King, J. R. Barnak, M. D. Bremser, K. M. Tracy, C. Ronning, R. F. Davis, and R. J. Nemanich, J. Appl. Phys. 84, 5248 (1998). R. Chierchia, T. Bottcher, H. Heinke, S. Einfeldt, S. Figg, and D. Hommel, J. Appl. Phys. 93, 8918 (2003). R. Averbeck and H. Riechert, Physica Status Solidi (a) 176 301 (1999). E. Dimakis, Master Thesis, Physics Department, University of Crete, July 2003. Ph. Komninou, unpublished results.

LOW DISLOCATIONS DENSITY GaN/SAPPHIRE FOR OPTOELECTRONIC DEVICES B. BEAUMONT, J.-P. FAURIE, E. FRAYSSINET, E. AUJOL, and P. GIBART Lumilog, 2720, Chemin de Saint Bernard, Les Moulins I, 06220 Vallauris, FRANCE

Abstract:

It is nowadays well established that threading dislocations (TDs) are degrading the performances and the operating lifetime of optoelectronic GaN-based devices (LDs and UV-LEDs). GaN/sapphire layers have been grown by Metal Organic Vapor Phase Epitaxy (MOVPE). An amorphous silicon nitride layer is deposited using a SiH4/NH3 mixture prior to the growth of the low temperature GaN buffer layer. Such a process induces a 3D nucleation at the early beginning of the growth, resulting in a kind of ELO process with intrinsic random opening sizes. This produces a significant decrease of the TDs density compared to the best GaN/sapphire templates. GaN layers with TD density as low as 7u107 cm–2 were obtained, as measured by atomic force microscopy (AFM), cathodoluminescence (CL) and transmission electron microscopy (TEM). The two-step epitaxial lateral overgrowth technology (2S-ELO) allows decreasing the TDs around 107 cm–2. These templates are suitable for fabricating LDs. Regrowth by HVPE on this ELO GaN/sapphire further decreases the TDs density below 106 cm–2.

Key words:

threading dislocations, epitaxial lateral overgrowth, MOVPE, HVPE, UV detectors

1.

INTRODUCTION

Bulk GaN is intrinsically very difficult to grow because of the high vapour pressure of nitrogen at the melting point of GaN. Growth in molten metals (Ga, Na) is currently under development and has so far only produced small high quality crystals. Therefore, all the development of nitride based devices has been made on foreign substrates. GaN is grown in the form of epitaxial layers on either sapphire or 6H-SiC. The lattice parameters and the thermal 189 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 189–197. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

B. Beaumont et al

190

expansion coefficients of sapphire and SiC are not well matched to GaN. The epitaxial growth therefore generates huge densities of dislocations (109 to 1011 cm–2). These dislocations propagate up to the surface, deteriorating the performance of optical and electronic devices. Heteroepitaxy on sapphire requires several steps including the nitridation of the sapphire substrate, the deposition of a low-temperature buffer layer and the heat treatment of this nucleation layer. The main goal of heteroepitaxy is to limit the generation of threading dislocations. But how far should we reduce the TDs in GaN? GaN epilayers with TDs densities in the mid 108 range can be produced easily by MOVPE. This quality has proven to be good enough for fabricating standard LEDs. However, new generation of LEDs or UV LEDs need dislocation densities much lower than 108 cm–2. Furthermore, laser diodes (LDs) for Blue-Ray DVD require free standing GaN with the TDs density in the low 106 cm–2 range.

2.

GaN/SAPPHIRE WITH TD DENSITY Ex, Ex is the exciton binding energy) and intense photoexcitation (estimated carrier density of the order of 1019 cm–3, thus essentially higher than the Mott density [6,10]), predetermine formation of free electron–hole pair system, where excitonic states are thermally ionized and/or screened by carriers. Radiation emitted in the backward direction is mainly due to spontaneous emission of plasma, since the thickness of the excited region in GaN is very small (dg | 0.1 Pm) and stimulated emission in the backward direction is inefficient [6]. However at high excitation, carrier system becomes degenerated and laterally stimulated emission usually is strong [8]. Some traces of scattered stimulated emission at 3.3 eV can be resolved at an early delay time [Fig. 1(b)]. At high optical excitation, the system of nonequilibrium carriers can be brought out of thermal equilibrium with the lattice due to an excess energy supplied and/or due to many-body recombination processes. Room temperature experiments with resonant and off-resonant excitation of GaN have demonstrated that broadening of the high-energy tail of the EHP luminescence band is due to the excess photon energy supplied to nonequilibriumcarrier and longitudinal optical phonon system [7]. Analysis of the highenergy wing of the recorded EHP band within a simple one-particle ap-

Materials Characterization of Group III-Nitrides

209

proach [13] (simulated spectra are shown by lines in Fig. 1) indicates that the carrier temperature reaches a value of 750 K at zero delay and relaxes exactly to the equilibrium value within the first 100 ps. Carrier heating up to 1200 K has been observed in GaN epilayers [7,8].

4

GaN/GaN

(a)

4

10

3

10

2

10

750 K

1

435 K

10

3.2

26.7 ps 66.7 ps 173 ps

3.3

3.4

298 K

3.5

3.6

Photon energy (eV)

3.7

Luminescence intensity (arb. units)

Luminescence intensity (arb. units)

10

GaN/Al2O3

(b)

3

10

2

10

700 K 1

10

375 K

26.7 ps 86.7 ps 206.7 ps

280 K

0

10

3.8

3.2

3.4

3.6

3.8

Photon energy (eV)

Figure 1. Luminescence spectra in homo- (a) and heteroepitaxial (b) GaN layers for the excitation density of Ig = 1.1 mJ/cm2 recorded for backward geometry at different delay time. The spectra are arbitrary shifted along the vertical axis. Points, experiment; solid lines, calculation. The deduced carrier temperature is indicated at each spectrum.

Carrier temperature is one of the crucial parameters of a dense nonequilibrium quasiparticle system. It determines the degree of degeneration of the carrier system and controls the rates of stimulated and spontaneous emission as well as the radiative and nonradiative recombination rates [8,14]. Variation in plasma temperature can also result in changes in the peak position of the emission band due to temperature-dependent band-gap renormalization [7] and band-filling effect. Cooling of the initially hot EHP results in a nonlinear carrier recombination dynamics [8]. In particular, the rate of lateral stimulated emission is highly sensitive to the carrier temperature. Usually at room temperature, a rapid carrier thermalization process takes place on the initial stage of relaxation. Owing to a decrease in plasma temperature, the carrier system becomes degenerated, thus stimulated emission rises up and results in a rapid decrease of the plasma density. After the stimulated emission is exhausted, spontaneous radiation is observed [8]. To characterize a material by time-resolved luminescence at high excitation conditions, carrier heating and laterally stimulated emission effects have to be considered [8].

210

3.

S. Jursenas et al

MATERIALS CHARACTERIZATION BY ELECTRON-HOLE PLASMA LUMINESCENCE DECAY

Transients of EHP luminescence can provide important information on materials quality. In the later stage of the relaxation (>100 ps), when carrier and nonequilibrium phonon thermalization is completed and the inverse population is exhausted, one usually observes an exponential decay, which is related to capture of carriers by deep centers of nonradiative recombination [8,9]. The number of nonradiative traps is one of the crucial parameters that controls the efficiency of semiconductor light emitters. Under intense photoexcitation when the nonequilibrium electron and hole concentrations ('n and 'p, respectively) are large compared to the density of deep traps, Nt, and the equilibrium carrier density, n0, ('n, 'p >> Nt, n0), for typical asymmetry of the electron and hole capture cross sections (say, bh > be), the traps are saturated by holes. Thus a deep-trap saturation regime is established [9]. In this regime, carrier recombination is controlled by the electron capture time We = (beNt)–1 provided that the bimolecular recombination rate, br, is negligible (brn > beNt), the luminescence decay becomes nonexponential and plasma-density dependent [8]. Figure 2 shows luminescence transients obtained in various GaN epilayers at the peak position of the emission band (~3.4 eV). It is evident that luminescence decays almost exponentially and the characteristic decay time varies in samples prepared by different growth procedures. The largest luminescence decay time WLU = 445 ps is characteristic of a homoepitaxial GaN epilayer grown by metalorganic chemical vapor deposition (MOCVD, (solid points). A GaN epilayer grown under similar conditions over a sapphire substrate shows a significantly lower value of WLU = 195 ps (open points). Lowquality GaN epilayers can have a luminescence decay time below 10 ps (squares). Our study shows that thick GaN epilayers (~1 Pm) grown by hydride vapor-phase epitaxy (HVPE) can have rather large luminescence decay time of WLU = 205 ps. Lines in Fig. 2 show results of calculation of the EHP spontaneous luminescence decay transients obtained by solving a system of three coupled rate equations for nonequilibrium electrons, holes, and recombination centers, respectively, under standard assumptions [9]. To fit the experimental points, different electron capture times were used: We = 195 ps (dotted line), 440 ps

Materials Characterization of Group III-Nitrides

211

Luminescence int. (arb. units)

(dashed line), and 970 ps (dot-dashed line). (Note that the values of We are slightly larger than 2WLU, because of bimolecular recombination taken into account.) For a reasonable value of the electron recombination coefficient be = 1u10–8 cm3/s [9], the density of deep traps can be estimated: Nt = 5.1u1017 cm–3 for a GaN epilayer grown on sapphire, Nt = 2.1u1017 cm–3 for HVPE-grown GaN and Nt = 1.0u1017 cm–3 for homoepitaxial GaN. In GaN grown on sapphire, these deep traps are most likely related to a larger number of threading dislocations occurring at the epilayer–substrate interface.

10

no defects GaN homoepitaxial GaN HVPE GaN on sapphire low-lifetime GaN on sapphire

3

bimolecular rec.

10

WLU = 445 ps

2

WLU = 205 ps

10

1

WLU < 10 ps

0

WLU = 95 ps

200 400 600 800 1000 Delay time (ps)

Figure 2. Transient behavior of the normalized luminescence intensity at the peak position of emission band (~3.4 eV) in various GaN crystals. Points, experiment; lines, calculation for various electron capture times: dotted line, We = 195 ps; dashed line, We = 440 ps; dot-dashed line, We = 970 ps; solid line, intrinsic band-to-band recombination (no nonradiative capture).

Solid line in Fig. 2 shows a luminescence decay transient for intrinsic band-to-band recombination with the bimolecular coefficient of br = 3u10–11 cm3/s (an average of the data reviewed in Ref. 15). It is evident that although homoepitaxy significantly reduces the number of threading dislocations, the quality of crystals can still be improved to reach a luminescence decay time of about 1.6 ns expected for purely bimolecular recombination at a carrier density of 1019 cm–3.

S. Jursenas et al

212

4.

CHARACTERIZATION OF InGaN/GaN MULTIPLE QUANTUM WELLS UNDER SCREENED BUIT-IN ELECTRIC FIELD

Ternary InGaN-based multiple quantum wells (MQWs) are the key structures for near-UV to green light emitting diodes and laser diodes [1,2]. Luminescence spectroscopy is one of the most widely used methods for characterization of the MQWs. However, the characterization is usually ambiguous because of a difficulty in distinguishing between blue shifts due to the quantum-confined Stark effect caused by built-in electric field and filling of localized-states in a partially disordered InGaN alloy [16,17]. Here, we applied a high-excitation regime to screen the built-in electric field by free carriers and to reveal the impact of In-segregation related disorder. InGaN/GaN MQW structures of various well thicknesses have been characterized by backward and lateral luminescence at low and high excitation density [18,19]. The undoped MQWs consisted of five 10–nm thick GaN barriers and In0.15Ga0.85N wells with the thickness of 2 to 4 nm in different samples. Figure 3 shows the well-width variation of the peak position of luminescence obtained at various excitation conditions. stimulated lateral (high exc.) spontaneous backward (high exc.) spontaneous backward (low exc.)

Peak position (eV)

3.1

3.0

2.9

2.8

2.7

InGaN/GaN MQWs 2.6 2

3

4

Well layer thickness (nm) Figure 3. Well-width dependence of the emission peak position at different excitation conditions. Solid points, backward spontaneous emission at low intensity cw excitation; open squares, backward spontaneous emission at high excitation obtained after 100 ps time delay introduced for exhaustion of stimulated emission; solid squares, lateral stimulated emission at the threshold of stimulation. Lines are guides for eye.

Solid points in Fig. 3 show the luminescence peak position obtained at room temperature for low-intensity cw excitation. With increasing well width from 2 nm to 4 nm, a red shift of 310 meV is observed. The shift re-

Materials Characterization of Group III-Nitrides

213

sults not only from the quantum size effect but also from the quantumconfined Stark effect and from intricate thermodynamics of In segregation that is due to lattice mismatch between InN and GaN [1,2,16]. To eliminate the built-in electric field, high-excitation conditions were applied (Ig = 1 mJ/cm–2). To avoid distortion of the spectra by stimulated emission that occurs at high excitation, backward spontaneous luminescence was measured at a 100-ps time delay, i.e. when stimulated emission is exhausted [19]. Open squares show the obtained variation of the peak position with well width. The red shift is seen to be reduced to 210 meV but it is still larger than 125 meV, the value expected to be due to finite-barrier quantum confinement in our MQW structures. We attribute the excess red shift of 85 meV to a difference in magnitude of the potential fluctuations that are due to formation of spatially separated In-rich and In-poor regions in the quantum well layers. Such fluctuations might be larger in thicker layers, since a longer growth time facilitates In segregation in strained InGaN system [20]. This assumption is confirmed by measurements of the peak position of the stimulated-emission band (solid squares in Fig. 3). With increasing well width, the energy separation between the stimulated and spontaneous emission bands is seen to increase from 25 meV to 220 meV. Since stimulated emission originates from carriers in the vicinity of the mobility edge [3,4] and spontaneous emission usually appears from the lowest occupied localized states, the energy separation between the peak energies of stimulated and spontaneous emission bands is close to the double magnitude of the potential fluctuation. The difference in potential fluctuation estimated from the well-width dependence of the separation (~100 meV) is in reasonable agreement with that deduced earlier from the excess red shift of the spontaneous band (85 meV).

5.

CONCLUSIONS

Group-III nitride materials properties have been studied by time-resolved luminescence spectroscopy under high-power photoexcitation conditions that are close to semiconductor-laser operation regime. An approach for estimation of GaN crystal quality by carrier lifetime under a deep-trap saturation regime has been demonstrated. The method was applied for GaN epilayers grown on sapphire, HVPE grown GaN, and highquality homoepitaxial GaN layers. Advantages of a high-excitation regime that enabled us to characterize In-segregation related disorder in InGaN/GaN MQWs under conditions of built-in electric field screening by a high-density carrier system were demonstrated.

S. Jursenas et al

214

ACKNOWLWDGEMENTS The authors would like to thank to S. Porowskii, P. R. Hageman, and C. C. Yang for high quality GaN epilayers and InGaN/GaN MQWs. The research was partially supported by the Lithuanian State Science and Education Foundation, by the joint Lithuanian-Latvian-Taiwan grant, by and European Commission supported SELITEC center Contract No.G5MA-CT-200204047. A. Ž. acknowledges the Lithuanian Ministry of Education and Science for his Fellowship.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

S. Nakamura, G. Fasol, The Blue Laser Diode: GaN Based Light Emitters and Lasers (Springer, Berlin, 1997). Žukauskas, M. S. Shur, and R. Gaska, Introduction to Solid-State Lighting (Wiley, New York, 2002). S. Chichibu, T. Sota, K. Wada, and S. Nakamura, J. Vac. Sci. Technol. B 16, 2204 (1998). Y.-H. Cho, T. J. Schmidt, S. Bidnyk, G. H. Gainer, J. J. Song, S. Keller, U. K. Mishra, and S. P. DenBaars, Phys. Rev. B 61, 7571 (2000). S. Juršơnas, G. Kurilþik, and A. Žukauskas, Phys. Rev. B 58, 12937 (1998). F. Binet, J. Y. Duboz, J. Off, and F. Scholz, Phys. Rev. B 60, 4715 (1999). S. Juršơnas, G. Kurilþik, G. Tamulaitis, A. Žukauskas, R. Gaska, M. S. Shur, M. A. Khan, and J. W. Yang, Appl. Phys. Lett. 76, 2388 (2000). S. Juršơnas, G. Kurilþik, N. Kurilþik, A. Žukauskas, P. Prystawko, M. Leszczynski, T. Suski, P. Perlin, I. Grzegory, and S. Porowski, Appl. Phys. Lett. 78, 3776 (2001). S. Juršơnas, G. Kurilþik, N. Kurilþik, A. Žukauskas, and P. R. Hageman, Appl. Phys. Lett. 83, 66 (2003). S. Hess, R. A. Taylor, J. F. Ryan, B. Beaumont, and P. Gibart, Appl. Phys. Lett. 73, 199 (1998). J. F. Müller and H. Haug, J. Luminescence 37, 97 (1987). E. A. Meneses, N. Jannuzzi, J. G. P. Ramos, R. Luzzi, and R. C. C. Leite, Phys. Rev. B 11, 2213, (1975). G. Lasher and F. Stern, Phys. Rev. 133, 553 (1964). S. Juršơnas, G. Kurilþik, and A. Žukauskas, Phys. Rev. B 54, 16706 (1996). A. Dmitriev and A. Oruzheinikov, J. Appl. Phys. 86, 3241 (1999). N. A. Shapiro, P. Perlin, C. Kisielowski, L. S. Mattos, J. W. Yang, and E. R. Weber, MRS Internet J. Nitride Semicond. Res. 5, 1 (2000). E. Kuokstis, J. W. Yang, G. Simin, M. A. Khan, R. Gaska, and M. S. Shur, Appl. Phys. Lett. 80, 977 (2002). S. Miasojedovas, S. Juršơnas, G. Kurilþik, A. Žukauskas, S.-W. Feng, C. C. Yang, H.-W. Chuang, C.-T. Kuo, and J.-S. Tsang, Phys. Status Solidi C 0, 483 (2002). S. Juršơnas, S. Miasojedovas, G. Kurilþik, A. Žukauskas, S.-W. Feng, Y.-C. Cheng, C. C. Yang, C.-T. Kuo, and J.-S. Tsang, Phys. Satus Solidi C 0 (2003), in press. D. Doppalapudi, S. N. Basu, K. F. Ludvig, Jr., and T. D. Moustakas, J. Appl. Phys. 84, 1389 (1998).

SMALL INTERNAL ELECTRIC FIELDS IN QUATERNARY InAlGaN HETEROSTRUCTURES S. ANCEAU 1,2, S. P. àEPKOWSKI 1,*, H. TEISSEYRE 1, T. SUSKI 1, P. PERLIN 1 P. LEFEBVRE 2, L. KOēCZEWICZ 2, H. HIRAYAMA 3, and Y. AOYAGI 3 1

UNIPRESS, Polish Academy of Sciences, Sokoáowska 29/37, Warszawa, Poland GES, Universite Montpellier II, F-34095 Montpellier Cedex 5. France 3 RIKEN (Institute of Physical and Chemical Research) 2-1Hirosawa, Wako-shi, Saitama 351-0198, Japan * Corresponding author; e-mail: [email protected] 2

Abstract:

A study of the internal electric field contribution to the light emission mechanism in InAlGaN-based multiple quantum wells was performed. We used two sets of structures with different quantum well width and different Al content in the barriers (series A with 30% of Al in the barriers and series B with 60% of Al in the barriers). To determine the magnitude of the built-in electric field we employed several methods: i) theoretical estimation of piezoelectric and spontaneous polarizations, ii) analysis of the emission energy as a function of quantum well width, iii) hydrostatic pressure experiments, and finally iv) measurements of photoluminescence decay. The performed calculations gave high magnitudes of the built-in electric field for both series of samples. On the contrary, from the experimental results, we concluded that the built-in electric field is negligible in the samples of series A and is rather weak in those of series B. Possible reasons for the controversies between theory and experiment are suggested.

Key words:

quaternary InAlGaN compounds, quantum wells, piezoelectric polarization, built-in electric fields, time-resolved spectroscopy, high pressure experiments

1.

INTRODUCTION

Recently, there is an increasing demand for light-emitting diodes and semiconductor lasers operating at around 350 nm. Such emitters have wide applications in many fields, in particular, environment control and medical in215 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 215–222. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

S. Anceau, S. P. Lepkowski et al

216

struments as well as illumination. Highly efficient emitters in the UV range were recently developed using quaternary InAlGaN alloys or GaN/AlGaN quantum wells (QWs) [1–3]. In the case of InGaN/GaN or GaN/AlGaN QWs [grown along the (0001) direction], we always deal with internal strain in the system, which results in a piezoelectric polarization and thus in a built-in electric field in the QW layers. Moreover, spontaneous polarization, which exists in hexagonal nitrides, has different values for AlN, InN and GaN [4]. This leads to an additional increase in the built-in electric field in both InGaN/GaN and GaN/AlGaN QWs. The resulting large internal electric field invokes the Quantum Confined Stark Effect (QCSE), which results in the separation of the electron and hole wavefunctions, and reduces the optical transition matrix element. On the contrary, in InAlGaN-based QWs compressive or tensile strain can be engineered. Particularly, one may expect that, for properly chosen compositions of barriers and QWs, the internal electric field is negligible since a term coming from spontaneous polarization cancels the term originating from the piezoelectric effect. In such a case no reduction of the optical transition probability due to QCSE occurs. In practice, it is difficult to realize any arbitrary composition of the quaternary alloys using epitaxial growth. There are some interrelations between In- and Al-incorporation in these compounds (see, e.g. Ref. 3). Thus, the condition for the cancellation of built-in electric field (which theoretically is close to lattice matching between barriers and wells) is not easy to achieve. The purpose of this work is to determine the magnitude of built-in electric fields in quaternary InAlGaN QWs showing intense light emission. We realize this task by performing time-resolved photoluminescence experiments and high-pressure studies of the light emission on two series of samples containing InAlGaN QWs of different width. The structures in the first set of samples (series A) have lower Al content in the barriers than the samples belonging to the second series (B). With increasing Al content in the barriers, we increase both the lattice mismatch between the barriers and the wells (i.e. piezoelectric effect) and the difference in spontaneous polarizations between the barriers and the wells. Both effects should lead to a higher magnitude of the built-in electric field in the samples from series B compared to those from series A.

2.

SAMPLES

The studied InAlGaN structures were grown by metalorganic vapor phase epitaxy (MOVPE) on the Si-face of an on-axis 6H-SiC(0001) substrate. Each sample consists of 3 quantum wells separated by barriers of 5-nm width. In

Small Internal Electric Fields in Quaternary InAlGaN

217

series A, the composition of QWs and barriers is In0.05Al0.20Ga0.75N and In0.02Al0.3Ga0.68N, respectively. The QW width varies from 1.3 nm up to 4.0 nm in different samples. In series B, the composition of QWs is the same as in series A while the composition of barriers is different, i.e. In0.02Al0.6Ga0.38N. The QW width varies from 0.9 nm up to 4.7 nm in different samples of series B. In each of the samples, the MQWs were grown on a structure consisting of: i) a 400-nm thick Al0.2Ga0.8N buffer layer, ii) a 30-nm thick In0.02Al0.3Ga0.68N layer (barrier composition), iii) a strain reduction layer, and iv) a 20-nm thick In0.02Al0.3Ga0.68N layer. Additionally, a reference sample containing a thick epitaxial layer of In0.05Al0.20Ga0.75N (120 nm) was also grown on a 400-nm thick Al0.2Ga0.8N buffer layer deposited on on-axis 6HSiC(0001). The alloys composition in the samples of series A was measured by Rutherford backscattering spectrometry, while for series B, Al and In contents were estimated from the technological procedure. More details of the growth procedure are given elsewhere [3].

3.

DETERMINATION OF BUILT-IN ELECTRIC FIELD

3.1

Emission Energy versus QW Width

Figure 1 shows the dependence of the PL peak position on QW width L for series A (squares) and B (circles). The emission energy for the thick reference layer corresponding to the QW material is also shown (triangle). In hexagonal ternary nitride QWs, the dependence of the emission energy, EE vs. L has been often used to determine the magnitude of the internal electric field. However, there exist three factors, which can contribute to this dependence. Namely, i) quantum confinement, which in narrow QWs causes a blue shift of EE in respect of the band gap, EG, of the well material, ii) the built-in electric field, which leads to a red shift of EE in respect of EG with the effect more pronounced in wider wells, and iii) in InGaN based structures, In fluctuations, which cause a decrease of a “local” value of EG and EE in respect to the average band gap of the QW material. In the case of quaternary nitride structures, the above-described analysis is even more complicated. In contrary to ternary QWs, determination of the quantum confinement term is presently impossible. This is due to the lack of knowledge of the band offsets and the effective masses for quaternary welland barrier materials. Moreover, though there are strong suggestions in favor of the importance of In-fluctuations in the material studied, their structural details require further investigations.

S. Anceau, S. P. Lepkowski et al

218

4.0

T=80K

series A series B thick layer

Peak position [eV]

3.9 3.8 3.7 3.6 3.5 3.4 3.3 1

2

3

4

120

Quantum well width [nm]

Figure 1. Photoluminescence peak positions taken at 80 K, as a function of quantum well width for samples of series A (squares) and B (circles) and for 120-nm thick layer of In0.05Al0.20Ga0.75N (triangle).

Taking these arguments into account for the case of quaternary QWs, a reliable determination of the magnitude of the built-in electric field from the dependence of EE vs. L should be treated with caution. However, one may suggest that from the fact that EE is weakly dependent on L, what is observed particularly for the series A, the magnitude of the built-in electric field is small. Additionally for all samples from series A, EE is higher than the emission energy for the thick layer (see Fig. 1), thus this suggests that the decrease of EE with increasing L is mainly due to quantum confinement. For the samples from series B we observe a stronger decrease of EE with increasing L, and for the widest QW (4.7 nm) EE is smaller than that obtained for the thick reference sample. This may indicate a presence of the built-in electric field in the latter case. However, in both cases the decrease of EE with increasing L is rather small comparing with the values obtained for ternary GaN/AlGaN and InGaN/GaN QWs [5,6]. It may suggest that in the case of the studied quaternary based QWs (series A and B) we deal with the presence of rather small internal electric fields. A contrary observation comes from theoretical estimation of the magnitude of internal electric field in our samples. Surprisingly, application of the recently published results of ab initio calculations for spontaneous and piezoelectric polarizations in ternary nitride compounds [4] to our quaternary QWs (assuming Vegard-like behavior, i.e. no additional bowing for quaternary alloys) brings the magnitude of the internal electric field between 1.3 and 1.7 MV/cm for series A and between 3.2 and 3.7 MV/cm for series B, respectively.

Small Internal Electric Fields in Quaternary InAlGaN

219

In order to clarify the situation, we analyze in considered structures: i) the pressure behavior of the PL peak position and ii) the PL decay times. Both effects are very sensitive to the presence of the internal electric field in the QW systems. The PL decay time, W and the pressure coefficient of the peak position, dEE/dP are known to show a drastic variation (an increase of W and a decrease of dEE/dP) with QW width in a QW system with strong builtin electric field. This refers to wurtzite InGaN/GaN and GaN/AlGaN QWs [5,6]. On the contrary,ҏ W and dEE/dP remain almost independent of QW width when the internal electric field in QWs is negligible [7].

3.2

Time-Resolved Photoluminescence Measurements

PL Decay Time [ns]

Time-resolved photoluminescence was measured at T = 8 K by using frequency tripled laser pulses from a Ti-sapphire cavity (hQ = 4.77 eV) with a typical pulse width of 2 ps and a repetition rate of 82 MHz. The typical power density for the luminescence excitation was of the order of 100 W/cm2, subject to some changes for different samples.

Series A Series B

10

T = 8K

1 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Quantum well width [nm] Figure 2. PL decay time as a function of QW width for samples of series A and B.

For the studied samples, the PL decays show nonexponential character. For this reason, the PL decay time, W is determined as a time after which the maximum of PL intensity drops by factor of 10. Figure 2 depicts the measured values of W for the samples from series A and B. For series A, the values of Ware seen to locate between 1 and 2 ns, and show no sensitivity to QW width. This observation supports the finding of weak internal fields in samples of series A. The estimated magnitude of the built-in electric field is below 0.1 MV/cm [7]. For the samples of series B, W is seen to increase by

S. Anceau, S. P. Lepkowski et al

220

one order of magnitude with increasing L from 0.9 nm to 4.7 nm. In this series, the estimated magnitude of the internal electric field is of about 0.6 MV/cm [7].

3.3

High Pressure Experiments

High pressure experiments were performed for the samples from series A. The measurements were carried out at T = 80 K in a low-temperature diamond anvil cell filled with liquid argon, which served as a pressure transmitting medium. PL was excited by a He–Cd laser (hQ = 3.8 eV) with the power of 2 mW. The emission from the sample was collected in a backscattering geometry, dispersed by a SPEX500M spectrometer and detected by a GaAs photomultiplier. In Fig. 3, we compare the measured dEE/dP for the samples of series A (solid circles) with the pressure coefficients of wurtzite (solid triangles) and cubic (solid squares) InGaN/GaN QWs taken from Refs. 5 and 8, respectively. For InAlGaN-based QWs, dEE/dP varies from 34 meV/GPa to 36 meV/GPa. Taking into account that the experimental error of dEE/dP determination is r2 meV/GPa, the obtained values of dEE/dP can be treated as almost independent of quantum well width. A weak dependence of dEE/dP on QW width (or a small nonlinear dependence associated with the quantum confinement effect) was observed in the case of cubic InGaN/GaN QWs where internal electric field is absent, for symmetry reasons. A very small variation of dEE/dP in quaternary InAlGaN-based QWs (wurtzite structure) is in contrast to dramatic changes of dEE/dP in the case of hexagonal InGaN/GaN QWs. In the latter case, the internal electric field can reach 2.4 MV/cm (at ambient pressure) for the In content of 20% [5]. The observed strong, linear decrease of dEE/dP with QW width in these structures was explained by the pressure-induced increase of the magnitude of piezoelectric field [9]. The described results suggest that in the studied quaternary QWs of series A, the built-in electric field is very small.

Pressure coefficient [meV/GPa]

Small Internal Electric Fields in Quaternary InAlGaN

221

35 30 25 20 15 10 cubic InGaN/GaN QWs wurtzite InGaN/GaN QWs wurtzite InAlGaN (series A)

5 0 -5 0

1

2

3

4

5

Quantum well width [nm] Figure 3. Pressure coefficients of the peak position as a function of QW width for the samples of series A (circles) compared to results obtained for hexagonal (triangles) and cubic (squares) InGaN/GaN QWs (Refs. 5 and 8, respectively).

Our preliminary studies of dEE/dP in the series-B samples show that there is some tendency to a reduction of the pressure coefficient with quantum well width. The sample with L = 3.3 nm shows dEE/dP=32±2 meV/GPa whereas the sample with L = 4.7 nm is characterized by dEE/dP = 25±2 meV/GPa. This result supports the conclusion drown from the time resolved measurements that in the samples of series B, a small built-in electric field is present.

4.

CONCLUSIONS

Measurements of the PL decay time and evolution of PL with pressure have been used to determine the magnitude of built-in electric field in two sets of InAlGaN-based QWs showing intense light emission. For the first set of samples having lower Al content in the barriers (series A), both the pressure coefficient of the PL peak energies, dEE/dP, and the PL decay time, W do not change with the thickness of the QW. This observation can be explained by assuming a very weak (100 keV neutrons was investigated. No significant degradation in CCE was seen in the X-ray irradiated device, whilst a reduction in CCE to 78% was observed in the neutron-irradiated sample.

Key words:

semi-insulating GaN, ionizing radiation detectors, charge collection efficiency, radiation hardness

1.

INTRODUCTION

Rapid progress of epitaxial growth techniques for GaN makes this material attractive for applications in high-temperature/high-power electronic devices operating at high frequencies as well as in blue-UV optoelectronic devices [1]. Semi-insulating GaN is a promising material for application in the detection of UV radiation but also, due to its high density, stability and high 279 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 279–286. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

J. V. Vaitkus et al

280

threshold voltage, seems to be promising for ionising radiation detectors (comparison of different materials for fast ionising radiation detectors is presented in Table I). The proposed upgrade of the CERN Large Hadron Collider to ten times brighter luminosity (1035 cm–2/s) poses severe challenges to semiconductor detectors within the CERN experiments. Table I. Comparison of materials for fast radiation hard detectors at 300 K.

Z Density, g/cm3 Bandgap '(, eV

Electron mobility, cm2/V·s Hole mobility, cm2/V·s Breakdown field, MV/cm Displacement energy, eV e–h pair energy, eV ** predicted values

Si 14 2.33 [2] 1.1 [2]

1500 [2] 600 [2] 0.3 [2]

13–20 [6] 3.6 [6]

SiC 14/6 3.2 [2] 3.26 (4H) 3.0 (6H) [2] 1000 (4H) 370 (6H) [2] 50 [2] 2.4 (6H) 2.0 (4H) [2] 22 [2] 7.8 (4H) [2]

C 6 3.5 [2] 5.45 [2]

GaAs 31/33 5.32 [2]

GaN 31/7 6.15 [4]

AlN 13/7 3.23 [4] 6.2 [4]

1.43 [2]

3.39 [3]

2200 [2] 1600 [2] 10 [2]

4000 [3] 400 [2]

1000 [3] 30 [3]

135 [4] 14 [4]

0.6 [2]

17.3** [5]

25.3 ** [5]

43 [6] 13 [2]

9.8 [4] 4.21 [2]

>15 [7] 19r2 [8]

Recently the epitaxial layers of semi-insulating GaN (SI-GaN) were grown [9] and some peculiarities of differently compensated samples properties, dark current temporal dependence and photoconductivity transient behaviour, were investigated in our recent paper [10]. In the present work, particular attention is paid to a review of first measurements of the charge collection efficiency (CCE) in as-grown material and in samples irradiated by different types of ionising radiation samples.

2.

EXPERIMENTAL

The epitaxial GaN layers used in this study were grown by MOCVD on Al2O3 (0001) substrates. The properties of the layers were changed by variation of the substrate temperature and the trimethylgallium (TMGa) flow rate during growth [9]. The buffer n*-GaN layer was thin, low temperature grown GaN (carrier concentration 3.3u1016 cm–3, electron mobility 610 cm2/V·s). An epitaxial 2–2.5-Pm thick capping layer was grown at a TMGa

Semi-Insulating GaN as an Ionizing Radiation Detector

281

-6

dE/dx (MeV/10 m)

flow rate of 88 µmol/min, at 925 ºC. Pad detector test structures were fabricated using 1.5-mm diameter evaporated gold Schottky contacts. The I–V characteristics were measured by contacting one Schottky contact, deposited on top of the sample and highly doped GaN by conductive glue (Fig.1). The devices demonstrated good rectifying behaviour of the Schottky contact, with a room temperature reverse leakage current density of 1u10–8 A·cm–2 at V = 10 V in the as-grown samples. The CCE measurements were performed by applying bias to the two Au contact areas, so that the real sample structure was “Schottky contact–(SIGaN)–(highly doped GaN)–(SI-GaN)– Schottky contact”. A 241Am alpha particle source was used for non-equilibrium carrier excitation. The 241Am emits 5.48 MeV alpha particles. The detector and source were housed in a vacuum chamber in which the pressure was less 20 mbar, to ensure negligible particle energy loss. The measurement setup consisted of a charge-sensitive preamplifier Figure 1. A schematic crossand a shaper amplifier with a shaping section of the sample. time of 1 Psconnected to a pulse-height analyser. Energy calibration of the detection system was carried out using a Si surface barrier diode assumed to have 100% CCE. By using the same gain and shaping time, it is possible to convert from channel number to energy for the Si diode. Further, by correcting for the difference in the mean energy to create an electron–hole pair between Si and GaN, it is possible Energy deposited by ionisation 0.30 to assign energies to the peaks of the observed GaN spectra. In order to calculate the energy deposited within the ~2 microme0.25 ters thick SI layer, the SRIM program [11] was used to calculate the energy deposition versus 0.20 depth dependence. This integra0.0 0.5 1.0 1.5 2.0 2.5 Depth (Pm) tion gave a total energy deposition of 553 keV within the 2 Pm Figure 2. Energy deposition vs. penetration detection layer (Fig.2). From depth of 5.48 MeV alpha particles in GaN, this, it was possible to calculate calculated using the SRIM code. the CCE of the detector.

J. V. Vaitkus et al

282

3.

RESULTS AND DISCUSSION a 6000

I, pA

4000 45 min 2h 4h 6h

2000

0 -2 -15

-10

-5

0

5

U, V

b

0.9

I, nA

0.6 0 min 30 min 60 min

0.3 0.0 -0.3 -15

-10

-5

0

5

10

U, V

30 20 0 min 30 min 60 min

10

I, pA

0 -10

c

-20 -30 -40 -15

-10

-5

0 U, V

5

10

15

Figure 3. Current -voltage I–V characteristics in asgrown (a), irradiated by X-rays (b) and by neutrons (c). Numbers in the inserts are the time duration after bias of sample.

The I–V dependencies presented in Fig. 3 for the Au–SI-GaN–n*-GaN samples irradiated by different ionising radiation way show a few peculiarities of the changes related to the irradiation. One is the very noisy reverse current and the initial part of the forward current in as-grown samples. This noise current can be related to tunneling through the complicated layer structure that creates a percolation in the SI-GaN conductivity. The properties of these samples are analyzed in [12]. The irradiation by X-rays created a load resistance that shifted the region of the carriers injection to higher forward bias and created defects. Polarization effects are seen in Fig. 3 and cause the I–V characteristics to be time dependent. The irradiation by neutrons transforms the Schottky barrier type of sample into a photoresistor but the introduced defects maintain the high resistivity state of the sample. The analysis of radiationinduced defects and radiation damage properties

Semi-Insulating GaN as an Ionizing Radiation Detector

283

will be published elsewhere. All samples were of high enough resistivity to permit testing of the alpha particle induced signals. Exposure of the GaN samples to 241Am alpha particles resulted in spectra given in Fig. 4–6 for non-irradiated, X-rays- and neutron-irradiated samples, 0V 1V 5V 8V 12 V 20 V 24 V 28 V Gauss fit C.C.E. 92.7% (peak at 0.514 MeV)

400

counts, a.u.

300

200

100

0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

E, MeV

Figure 4. Alpha particle pulse height spectra from a non-irradiated SI-GaN sample. The spectra were recorded at various bias voltages, which are indicated in the inset.

respectively. The spectra were recorded at various bias voltages as indicated in the insets. Fluence 600 Mrad of 10 keV X-rays

Counts, a.u.

600

0V 1V 8V 10 V 14 V 18 V 24 V 30 V Gauss fit C.C.E. 93 % (Peak at 0.514 MeV)

400

200

0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

E, MeV Figure 5. Alpha particle pulse height spectra from a SI-GaN sample irradiated by 600 MRad of 10 keV X-rays. The bias voltages are indicated in the inset.

Well-defined signals were observed in all unbiased samples, showing that the samples are polarized and an internal field exists. The half-width and

J. V. Vaitkus et al

284

2000

14

0V 1V 10 V 5V 24 V 16 V 30 V 26 V Gauss fit C.C.E. 78 % (Peak at 0.434 eV)

1500 Counts, a.u.

-2

Fluence 5 x 10 cm of fast neutrons

1000

500

0 0.0

0.1

0.2

0.3 0.4 0.5 0.6 E, MeV Figure 6. Alpha particle pulse height spectra from the SI-GaN sample irradiated by fast neutrons (E > 100 keV), fluence 5×1014 cm–2. The bias voltages are indicated in the inset.

peak position values of the pulse height spectra increase with bias. An increase of the bias voltage also leads to growth of the CCE, which in nonirradiated and X-rays irradiated samples reached a value of approximately 93%. In neutron-irradiated sample the CCE was smaller, at around 78%, what is still much better than in other promising materials at the same level of irradiation. A comparison of our results and those presented in [13] is shown in Fig.7. 100

C.C.E., %

80 60

GaAs GaN

40 20 0 0

1

2

3 14

4

5

6

2

F, 10 /cm

Figure 7. Schematic comparison of CCE dependence on neutron fluence in GaAs according [13] and GaN. The lines are only a guide to the eye.

Semi-Insulating GaN as an Ionizing Radiation Detector

285

So far, the GaN detectors that we have tested have shown very good CCE (~93%) with a small reduction only after very high fluences of neutrons. The observed change of ~15% is to be compared to the ~60% reduction observed for GaAs at a similar neutron fluence [13]. These first tests of charge collection in irradiated GaN open a desire to grow thicker (0.05 – 0.2 mm thick) SI-GaN epitaxial layers or wafers for testing material corresponding to the requirements of the big experiments related to the Large Hadron Collider (LHC), ATLAS, CMS, LHCb and ALICE [14]. The GaN parameters (as well as AlN) listed in Table I suggest the possibility of developing a detector system with the working parameters required for an upgrade of the LHC [6,14]. The pixel readout time should be of the order of 12 ns. The signal will not be worse than in silicon detectors due to the higher breakdown field and saturation velocity [2]. Also, the radiation hardness of material has to be tested up to 1 MeV neutron equivalent 1u1016 cm–2 fluence. The pixel sizes, e.g. for ATLAS detector, have to be 0.04 u 0.04 mm2 and a standard microstrip sensor chip is around 6 u 5 cm2. The total area of ionising radiation detectors for the LHC detectors combined has to be more than hundreds square meters. Therefore the design of large-area GaN wafer growth technology and reduction of the wafer cost become important.

4.

CONCLUSION

Semi-insulating GaN has to be shown to be promising as a radiation-hard detector of ionizing radiation.

AKNOWLEDGEMENTS This work has been partly performed in the framework of the CERN RD50 collaboration. The authors express their gratitude to the Royal Society and to PPARC for a long-term grant that supported this work. We thank M.Mikuz and E.Noah for their assistance with the sample irradiations.

REFERENCES 1. 2. 3.

Nakamura S. and Chichibu S.F. (Ed.). Introduction to nitride semiconductor blue lasers and light emitting diodes. London: Taylor and Francis, 2000. Bertuccio G., Casiraghi R., Study of SiC for X-ray detection and spectroscopy. IEEE Transactions on Nuclear Science 2003; 50:175–185. Shur M. S. GaN based transistors for high field power application. Solid State Electronics 1998; 42:2131–8.

J. V. Vaitkus et al

286 4.

5. 6.

7. 8.

9.

10.

11. 12.

13.

14.

Levinshtein M. E., Rumyantsev S. L. and Shur M.S. (Ed.). Properties of Advanced semiconductor materials: GaN, AlN, InN, BN, SiC, SiGe. New York: John Wiley and Sons, 2001. Allam J. “Universal” dependence of avalanche breakdown on bandstructure: choosing materials for high-power devices. Jpn. J. Appl. Phys. 1997; 36:1529–42. Moll M., on behalf of the RD50 collaboration. Development of radiation hard sensors for very high luminosity colliders – CERN – RD50 project. Computer physics communications, 2003 (to be published). Look D. S., Reynolds D. S., Hemsky J. W., Sizelove J. R., Jones R. L., and Molnar R. J. Defect donor and acceptor in GaN. Phys. Rev. Lett. 1997; 79:2273–6. Ionascut-Nedlcescu A., Carlone C., Haudayer A., von Bardeleben H.J., Cantin J.-L., Raymond S. Radiation Hardness of Gallium Nitride. IEEE Transactions of Nuclear Science 2002; 49:2733–8. Wang T., Shirahama T., Sun H. B., Wang H. X., Bai J., and Sakai S. Influence of buffer layer and growth temperature on the properties of an undoped GaN layer grown on sapphire substrate by metalorganic vapor deposition. Appl. Phys. Lett. 2000; 76:2220–2. Vaitkus J. V., Gaubas E., Sakai S., Lacroix Y., Wang T., Smith K. M., Rahman M., and Cunningham W. Role of potential barriers in epitaxial layers of semi-insulating GaN layers. Solid State Phenomena 2003; 93–93:301–6. Zeigler J. F., Biersack J. P., IBM-Research, Yorktown Heights, NY, USA, 1996. Vaitkus J., Cunningham W., Gaubas E., Rahman M., Sakai S., Smith K. M., and Wang T. Semi-insulating GaN and its evaluation for alpha particle detection. Nuclear Instruments and Methods in Physics Research A 2003; 509:60–4. Bates R. L., Da Via C., D'Auria S., O'Shea V., Raine C., Smith K. M., and RD8 collaboration. The effects of radiation on gallium arsenide radiation detectors. Nuclear Instruments and Methods in Physics Research A, 1997; 395:54–59. Price M. J. The LHC project. NIMA, 2002; 478: 46–61.

TOWARDS THE HYBRID BIOSENSORS BASED ON BIOCOMPATIBLE CONDUCTING POLYMERS

A. RAMANAVICIENE 1 and A. RAMANAVICIUS * 2,3,4 1

Laboratory of Ecological Immunology, Institute of Immunology of Vilnius University, Moletǐ pl. 29, Vilnius, Lithuania 2 Department of Analytical and Environmental Chemistry, Vilnius University, Naugarduko 24, Vilnius, Lithuania 3 Sector of Immunoanalysis and Informatics, Institute of Immunology of Vilnius University, Moletǐ pl. 29, Vilnius, Lithuania 4 Laboratory of Bioanalysis, Institute of Biochemistry, Mokslininkǐ 12, Vilnius Lithuania. *Corresponding author e-mail: [email protected] and/or [email protected]

Abstract:

The effective combination of biological and physical methods in analytical device could provide the basis for direct detection of wide range of analytes with great sensitivity and specificity. The main aim of this study is to briefly present the current state in construction of biosensors based on conducting polymers. Current state of research on S–S conjugated polymer polypyrrole based biosensors is presented. The biological recognition parts of reviewed biosensors were based on polypyrrole doped or covalently modified by enzymes and other proteins. Future perspectives of polypyrrole application in hybrid biosensors based on dual (optical and electrochemical) detection system are discussed. Investigations of polypyrrole show that florescence of this S–S conjugated polymer is undetectable. It can be successfully exploited as the immobilization matrix in fluorescence based biosensors, however.

Key words:

biosensors, conducting polymers, polypyrrole

287 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 287–296. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

A. Ramanaviciene and A. Ramanavicius

288

1.

INTRODUCTION

The most powerful alternative to conventional analytical techniques, harnessing the specificity and selectivity of biological systems in small, lowcost devices is biosensor technology. Biosensor is described as a compact analytical device, incorporating a biological or biomimetic sensing element to, or integrated within, a transducer system [1]. The detection is based on specific complementary binding or catalytic conversion of analyte of interest by bio-recognition element immobilized on the suitable signal transducer. The specific interaction of analyte with bio-recognition element results in a change of one ore more physicochemical properties (electron transfer, capacity, optical properties, etc.); those are detected and can by measured by the signal transducer. Biosensors and affinity-sensors are usually defined as sensing devices consisting of a biological recognition element in intimate contact with a suitable transducer, which is able to convert biological recognition reaction or, eventually, the biocatalytic process into a measurable electronic signal. Major indispensable condition during the creation of affinity-sensors is that one of the able-to-bind reagents is immobilized and at least one must be found in the sample [2]. It means that immobilisation of biologically sensitive compounds is one of the main questions during the creation of affinity- and biosensors [3]. Here, conducting polymers can be considered as effective immobilization materials [4]. Polyaniline and polythiophene are often used as electrocatalysts and immobilizers for biomolecules [5]. However, the necessity to detect bio-analytes at neutral pH range leads to electro-inactivity of the deposited films, discouraging the use of polyaniline and polythiophene as biosensing materials. Polyacetylene can be synthesized from gas phase only in the presence of the catalysts. Highly sophisticated procedure and instability of polyacetylene in humid environment makes this polymer less attractive for application in sensor design. Among the other conducting polymers, polypyrrole (Ppy) is one of the most extensively studied materials. Ppy can be easily synthesized by chemical and electrochemical polymerization approaches. This polymer has attractive features, such as excellent conductivity and stability on various substrate materials, even in a neutral pH region. The electrochemical properties of Ppy strongly depend on their redox states, and overoxidation of Ppy occurring at positive potentials leads to lowering of its conductivity as well as to dedoping of anionic molecules. Overoxidized Ppy has been used in some electroanalytical applications that utilize its permselectivity. Polypyrrole-based selective and stimulus-responsive biopolymers prove to be promising as new materials in affinity-sensors especially for biomedical application where direct detection of analyte is desirable [6]. Polypyrrole is often used in biosensors and affinity-sensors because of the

Hybrid Biosensors Based on Biocompatible Conducting Polymers

289

best biocompatibility and the easy ways for immobilization of various biologically active compounds [7]. From the analytical point of view, Ppy has some very attractive characteristics: (i) is biocompatible and, hence, causes minimal and reversible disturbance to the working environment; (ii) is capable of transducing the energy arising from interaction between immune reagents into electrical signals that are easily monitored; (iii) protects electrodes from fouling and interfering materials such as electroactive anions; (iv) can be modified in situ in a controlled fashion. Characteristics mentioned above possess great application possibilities of Ppy in affinity-sensors devoted to direct detection of analyte (4). Depending on the method of signal transduction, biosensors are divided into different groups: electrochemical, optical, thermometric, piezoelectric or magnetic. Electrochemical biosensors are the most commonly reported class of biosensors [8]. The main advantages of electrochemical transduction systems are low cost, simple operation and the use of disposable electrodes. Limitations of electrochemical transducers include interference from electroactive compounds and, as usually, low sensitivity. The problems mentioned here can be solved if alternative detection method is applied additionally. The most powerful alternative is optical signal transduction. The optical transducers may be involved to detect the analytical signal simultaneously with the electrical transducer to extend quality/quantity off data received. The main aim of the work presented here is to show: (i) current state in application of conducting (S–S conjugated) polymers and especially polypyrrole in biosensor design, (ii) perspectives of polypyrrole application in biosensors based on dual detection (optical and electrochemical) system.

2.

DISCUSSION

2.1

Discovery and Structure of Conducting Polymers

Conducting polymer polyacetylene was discovered by MacDiarmid, Shirakawa, and Heeger. They brought the unique properties of conjugated polymers to the fore in 1977 when they discovered that chemical doping of these materials resulted in increases in electronic conductivity over several orders of magnitude [9]. Since then, electronically conducting materials based on conjugated (conducting) polymers have been applied in diverse items such as sensors, biomaterials, light-emitting diodes, polymeric actuators, and corrosion protection agents. Some conducting polymers like polyaniline, polytiophene or polypyrrole are biocompatible and cause minimal and reversible disturbance to the working environment and protect electrodes from

290

A. Ramanaviciene and A. Ramanavicius

fouling and/or interfering with electrochemically active materials [10]. Currently the potential of these new materials for a wide range of nano-science and nano-technological applications is being demonstrated, as well [11]. General structure of conducting polymers is based on a framework of alternating single and double carbon–carbon (sometimes carbon–nitrogen) bonds. Single bonds are referred to as V-bonds, and double bonds contain a V-bond and a S-bond. All conjugated polymers have a V-bond backbone of overlapping sp2 hybrid orbitals. The remaining out-of-plane pz orbitals on the carbon (or nitrogen) atoms overlapped with neighboring pz orbitals to give Sbonds. The chemical structures of these materials usually are represented as sequences of consecutively alternating single and double bonds. In reality, the electrons that constitute the S-bonds are delocalized over the entire molecule [12].

2.2

Methods Used for Fabrication of Polypyrrole Films

Polypyrrole is usually synthesized by electrochemical and chemical oxidative polymerization techniques. Chemical polymerization occurs after oxidation of pyrrole monomers and oligomers by oxidators to the cation-radicals those are recombining and forming polymeric structure of polypyrrole. Chemical methods are difficult to use for miniaturization, construction of sensor arrays, or optimization of surface microenvironments. Electropolymerization is a more modern and elegant method of polymeric film deposition. It is achieved if an initial electron transfer takes place that permits coupling reactions to occur leading to additional chain growth as additional electrons are transferred. Polypyrrole films are generally formed by electrochemical anodic oxidation of a monomer, and are insoluble, conducting or, in some cases, insulating polymer films that coat the electrode surface [13]. This is especially attractive, since the oxidation of monomer solution under the appropriate conditions results in a film deposited on the surface of the electrode, and enables control of growth rate and film thickness. The electrochemical formation of conducting-polymer films has found increasing interest in the development of bio- and immuno-sensors since they allow non-manual reproducible formation of modified electrode surfaces with integrated biological recognition elements [14]. The use of conducting polymers for immobilizing a biological compound in sensor applications has the advantage, compared to conventional immobilization procedures, because the amount of deposited material can be readily controlled and the immobilizing matrix can conduct electricity allowing switching between the conducting and isolating states. Moreover, useful copolymeric structures can be developed if differently modified monomers are

Hybrid Biosensors Based on Biocompatible Conducting Polymers

291

copolymerized [15,16]. Electrochemical Ppy film formation can be performed using potential cycling methods, fixed potential techniques, pulsed potential approaches, and galvanostatic techniques.

2.3

Application of Conducting Polymers for Immobilization of Biologically Active Compounds

A number of techniques to immobilize biologically active compounds (BCs) on the electrode surface are available including adsorption, covalent attachment, cross-linking and entrapment within polymeric chain. Conducting polymers such as polypyrrole, polythiophene and polyaniline are proved to be the most useful molecular structures for these applications. Adsorption is the simplest way to immobilize a BC on the surface of an electrode coated by polypyrrole. The electrode can be electrochemically coated by polypyrrole using one of the mostly used electropolymerization methods. By using this method, BC from single proteins (antigens or antibodies) uptol whole cells can be adsorbed on the CP surface. Covalent attachment is the next approach that has been used for immobilization of BC on the surface of conducting polymer. Covalent attachment of BC on the surface of CP results in BC activity higher than that obtained previously by adsorption and is responsible for enhanced stability of analytical system during continous measurements. The next very important advantage of this method is ordered BC orientation, which is especially important for efficiency of affinity sensors. In many cases, irreversible adsorption or covalent binding of an antigen or antibody does not lead to the effective biosensing system due to the instability of BC molecules, low BC loading, or the potential loss of activity during covalent immobilization. Entrapment of BC molecules within conducting polymer backbone is more favourable approach for sensor design. Polypyrrole is suitable for this purpose because it can be easily prepared on miniaturized components; besides, it has a high conductivity and is relatively stable. Usually entrapment of biologically active compounds BC within the polymeric chains of conducting polymer is carried out during electrochemical polymerization of monomer in the presence of conducting polymer [17]. This method is very attractive because the formation of biologically active layer can be performed during “one-step” procedure and is promising for the formation of multi-array biosensors and affinity sensors. On the other hand, polypyrrole is biocompatible and, hence, causes minimal and reversible disturbance to the working environment. Ppy is also capable to transfer energy as electrochemical transducer. Stabilization of the biological response currently is the major problem, with almost every reported sensor exhibiting a gradual deg-

292

A. Ramanaviciene and A. Ramanavicius

radation in the electrical signal during continuous measurements. This is due to instability of biomolecules used in design of biosensors. Resolution of this problem and the production of robust designs, vital for medical and environmental monitoring applications, can be based on creation of synthetic molecular recognition systems. Artificial receptors have been gaining in importance as a possible alternative to immobilized biomolecules based systems. Molecular imprinting is increasingly becoming recognized as a versatile technique for the preparation of artificial receptors based on molecularly imprinted conducting polymers (MIPs) containing tailor-made recognition sites. MIP is another class of substances of great interest in the field of chemical sensor technology. These highly stable synthetic polymers possess molecular recognition properties due to cavities in the polymer matrix that are complementary to the analyte (ligand) both in shape and in positioning of functional groups [18]. It is the reason why development of synthetic recognition systems is of great interest to workers in the field of sensor technology. Moreover, some of these polymers have shown very high selectivity and affinity constants fully comparable to naturally occurring recognition systems, such as antibodies, what makes them especially suitable for use in artificial receptors. Overoxidized polypyrrole exhibits an improved selectivity, which is attributed to the removal of positive charges from Ppy films due to introduction of oxygen functionality, such as carbonyl groups. The nanopores and nano-cavities complementary to removed dopant can arise during dedoping process. Sensors based on mPpy for serotonin and 1naphthalensufonate were reported [19]. Molecular imprinting is a technology for the manufacture of synthetic polymers with predetermined molecular recognition properties. The preparation of molecularly imprinted polymers requires polymerization around print species using monomers those are selected for their capacity to form specific and definable interactions with the print species. BC molecules entrapped within Ppy can be removed by solvent extraction and the molecularly imprinted polymer is ready for use. Cavities are formed in the polymer matrix, which are images of the size and shape of the print molecules. Furthermore, chemical functionalities of the monomer residues become spatially positioned around the cavity in a pattern, which is complementary to the chemical structure of the print molecule [20]. These imprints constitute a permanent memory for the print species and enable the imprinted polymer to selectively rebind the print molecule from a mixture of closely related compounds. Finally, the print molecules are removed by solvent extraction and the molecularly imprinted polymer is ready to be used. In some instances very high selectivities and affinity constants have been reported, fully comparable to naturally occurring recognition system such as antibodies. Some of these synthetic polymers have been shown

Hybrid Biosensors Based on Biocompatible Conducting Polymers

293

to be useful in sensor applications, exhibiting tolerance towards acid, base, high temperature, and organic phases.

2.4

General Detection Methods Used in Biosensors

The conversion of the binding event into a measurable signal at low concentrations of analyte, the regenerability, and the reusability are, among other topics, major challenges in sensor development research. According to application of additional labels, two main detection types are applied in electrochemical sensor design: (i) indirect electrochemical detection – the binding reaction is visualized indirectly via an auxiliary reaction by a labeled compound; (ii) direct detection – no additional electrochemical labels are required. The effective combination of bio-chemistry and electrochemistry in an analytical device could provide the basis of direct electrical detection of wide range of analytes with great sensitivity and specificity [21]. Directdetection affinity sensors are the most attractive because they require no additional chemicals. Moreover, it allows real-time measurement without any additional hazardous reagents. For direct conversion of the binding event into measurable signal, one can use: (i) optical, for example, surface plasmon resonance; (ii) piezoelectric, for instance, quartz crystal microbalance; (iii) surface scanning, like atomic force microscopy (AFM), and (iv) electrochemical transducers. Because of possibilities to operate in non-transparent solutions, miniaturization, and simple signal transduction, the electrochemical affinity sensors are the most suitable for direct detection of the analyte. In direct affinity sensors, the electrochemical detection of the analyte binding to a biologically active layer can be performed with alternating current (ac), impedance or potential-step methods by measuring differences in capacitance and/or resistance of electrode. Pulsed amperometric detection (PAD) techniques are such techniques where sensor can be used for analyte detection in static or flow injection mode by applying pulsed potentials between the working electrode (sensor surface) and the reference electrode. In this method changes in charge densities or conductivities are used for transduction and no auxiliary reaction is needed. It seems that the mechanism of the analyte binding reaction at mPpy based electrodes involves the variation in the capacitive properties of the polymer. During interactions of analyte with biologically-active layer, differences in capacitance and/or resistance arising in electrochemical system can be converted into electrical signals that are easily monitored [22]. The current obtained can be directly related to the concentration of the analyte in solution [23]. Over the past years advances in the application of Ppy for construction of bio-sensing components and the fast response times offered by electrochemical techniques has created a demand for fast bio-sensing infor-

294

A. Ramanaviciene and A. Ramanavicius

mation-collecting systems that can be easily integrated into instruments, such as microprocessor based electronics. The PAD detection seems to be one of the most promising among the other electrochemical techniques applied in Ppy based affinity sensors.

2.5

Application Potential of LED Based Transducers in Design of Hybrid Biosensors

Multiple-transducer based biosensors offer great advantages over conventional mono-transducers based techniques. The multi-functionality of transducing part of biosensor offers the opportunity for development of highly specific devices for real-time analysis in complex mixtures, without the need for extensive sample pre-treatment or large sample volumes. Since those biosensors promise to be highly sensitive, rapid, reproducible, simple tooperate and multipurpose analytical tools, they will find great application not only in analytical fields but in scientific investigations as well. Such systems are especially required in the fields of proteomics and bioinformatics where a wide spectrum of information about interacting biological objects is desired to be collected in real-time. In recent years, a key stimulus for the development of optical biosensors has been the availability of high-quality and diverse LED’s, fibres and other optoelectronic components. The electrochemical/optical biosensor format may involve direct detection of analyte of interest or indirect detection through labeled probes. The optical transducers may detect changes in fluorescence, luminescence, absorbance, polarisation, refractive index etc. The advantages of optical transducers are their speed, the immunity of signal to electrical or magnetic interference and the potential for higher sensitivity and advanced information content. However, several years ago the main drawback was high cost of some optical instrumentation. Currently, the costs of optical components dramatically decreased, what offers a great opportunity for construction of biosensors suitable for mass production.

3.

CONCLUSIONS AND OUTLOOK

The background presented illustrates that polypyrrole is very attractive, versatile material suitable for preparation of various enzymatic-, immuno-, DNA-sensors. The presented overview of experimental results shows that electrochemical affinity sensor based on molecularly imprinted polypyrrole could have a great potential for direct electrochemical sensing. Since, according to our initial investigations, fluorescence of polypyrrole is undetectable, we believe that it can be successfully exploited as the immobilization

Hybrid Biosensors Based on Biocompatible Conducting Polymers

295

matrix in fluorescence based biosensors as well. Registration of binding/desorption of template molecule and quantification of analyte can be performed by very simple PAD method. Optical transducers based on fluorescence detection of protein-protein and DNA–DNA binding events are under development.

ACKNOWLEDGEMENT This work was financially supported by Lithuanian State Science and Studies Foundation project number C 03047, and COST program Project number 853.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

14. 15. 16. 17.

Warsinke A., Benkert A., Scheler F. W. Electrochemical immunoassays. Fresenius J. Ana.l Chem. 2000; 366: 622–634. Cook C. J. Nature Biotechnol. 1997; 15: 467–471. Sargent A., Sadik O. A. Electrochim. Acta 1999; 44: 4667–4675. Ramanaviciene A., Ramanavicius A. Crit. Rev. Anal. Chem. 2002; 32: 245–252. Geise R. J., Adams J. M., Barone N. J., Yacynych A. M. Biosens. Bioelectron. 1991; 6: 51–160. Wang J., Jiang M., Fortes A., Mukherjee B. Anal. Chim. Acta 1999; 402 (1–2): 7– 12. Kwon I. C., Bae Y. H., Kim S. W. Nature 1991; 28: 291–293. Knichel M., Heiduschka P., Beck W., Jung G., Gopel W. Sens. Actuators 1995; B28: 85–94. Chiang C. K., Fincher C. R., Park Y. W., Heeger A. J., Shirakawa H., Louis E. J., Gau S. C., MacDiarmid A. G. Phys. Rev. Lett. 1977; 39: 1098–1101. Ramanavicius A. Biologija 2000; 2: 64–66. Shiigi H., Okamura, K., Kijima, D., Deore, B., Sree, U., Nagaoka, T. J. Electrochem. Soc. 2003; 150: H119–H123. Wallace G. G., Dastoor P. C., Officer D. L., Too C. O. Chem. Innov. 2000; 30: 14– 22. Ramanavicius A., Habermüller K., Razumiene J., Meskys R., Marcinkeviciene L., Bachmatova I., Csöregi E., Laurinavicius V., Schuhmann W. Prog. Colloid Polym. Sci 2000; 116: 143–148. Ramanavicius A., Habermüller K., Csöregi E., Laurinavicius V., Schuhmann W. Anal. Chem. 1999; 71: 3581–3586. Habermuller K., Ramanavicius A., Laurinavicius V., Schuhmann W. Electroanal. 2000; 12: 1383–1389. Mieliauskiene R.; Kurtinaitiene B., Bachmatova I., Ramanavicius A. Biologija 2000; 2: 42–44. Ramanavicius A., Kurtinaitiene B., Razumiene J., Laurinavicius V., Meskys R., Rudomanskis R., Bachmatova I., Marcinkevicienơ L. Biologija, 1998; Suppl. 1: 15– 17.

296

A. Ramanaviciene and A. Ramanavicius 18. Kroger S., Turner A. P. F., Mosbach K., Haupt K. Anal. Chem. 1999: 71: 3698– 3702. 19. Shiigi H., Okamura K., Kijima D., Hironaka A., Deore B., Sree U., Nagaoka T. Electrochem. Solid St. 2003; 6: H1–H3. 20. Spurlock L. D., Jaramillo A., Praserthdam A., Lewis J., Braither-Toth A. Anal. Chim. Acta 1996: 336: 37–40. 21. Laurinavicius V., Razumiene J., Kurtinaitiene B., Lapenaite I., Bachmatova I., Marcinkeviciene L., Meskys R., Ramanavicius A. Bioelectrochemistry 2002; 55: 29–32. 22. Ramanaviciene A., Vilkanauskyte A., Acaite J., Ramanavicius, A. Biologija 2000; 2: 67–69. 23. Ramanaviciene A., Vilkanauskyte A., Acaite J., Ramanavicius A. Acta Medica Lituanica 2000; 5: 49–59.

OPTICALLY PUMPED UV-BLUE LASERS BASED ON InGaN/GaN/Al2O3 AND InGaN/GaN/Si HETEROSTRUCTURES

G. P. YABLONSKII 1, A. L. GURSKII 1, E. V. LUTSENKO 1, V. Z. ZUBIALEVICH 1, V. N. PAVLOVSKII 1, A. S. ANUFRYK 1, Y. DIKME 2, H. KALISCH 2, R. H. JANSEN 2, B. SCHINELLER 3, and M. HEUKEN 3 1

Stepanov Institute of Physics of NAS Belarus, F. Skaryna Ave. 68, 220072 Minsk, Belarus E-mail: [email protected] 2 Institut fur Theoretische Elektrotechnik, RWTH, Aachen, Germany 3 AIXTRON AG, Aachen, Germany

Abstract:

Optically pumped lasing in GaN epitaxial layers and InGaN single, multiple and electroluminescence-test quantum well heterostructures grown on sapphire and silicon substrates are investigated as functions of temperature (80–650 K), the excitation density of the nitrogen- and dye laser radiation (10–1100 kW/cm2), excitation and operation wavelengths, and MOVPE growth conditions. Laser action was achieved in all types of the heterostructures from the near ultraviolet (370 nm) up to the blue spectral region (470 nm). The lowest laser threshold at room temperature was 35 kW/cm2, the maximal laser power was 80 W, and the half width of the laser line was 0.04 nm. The maximal operating temperature of 630 K was for InGaN/GaN/Si lasers. On the base of investigation of the temperature dependence of the laser threshold and photoluminescence spectra, photoluminescence and laser excitation spectra, conclusions about the role of localized and delocalized states in the optical gain mechanisms were made.

Key words:

lasing, InGaN, heterostructures, optical pumping, threshold, gain, buffer layers, MOCVD, sapphire substrate, silicon substrate

297 ˘ M.S. Shur and A. Zukauskas (eds.), UV Solid-State Light Emitters and Detectors, 297–303. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

G. P. Yablonskii et al

298

1.

INTRODUCTION

The fabrication of GaN based lasers for green and UV spectral region still remains a problem, in contrast to the light emitting diodes, which already occupy a much more wide spectral region than lasers [1]. The main problem in fabrication of green lasers is the phase separation in InGaN alloy [2,3], while UV lasers suffer mostly from low AlGaN dopability and high activation energies of acceptors [4]. While the violet lasers have lifetime up to 15000 hours and output power of about 30 mW [5], true blue junction lasers have been created only recently and have lifetime only about 2000 hours [6]. Thus, another problem connected with the two above-mentioned ones is the short lifetime of devices. In addition, especially for automotive and other applications in high-volume low-cost market segments, the choice of substrate is still an open question. Sapphire and SiC are commonly used as substrates because of the lack of large-area GaN bulk crystals. Sapphire suffers from a low thermal conductivity worsening the high-power operation of lasers, light emitting diodes and transistor devices due to heat-removal problems. Silicon carbide has a lower lattice mismatch to GaN and a higher thermal conductivity than sapphire, but is hampered by a very high price and a limited availability for the electrically insulating crystal modification. Silicon is a promising alternative substrate for GaN growth because of its low cost, excellent quality, large-area availability and the possibility to integrate GaN-based light emitting devices and high-power electronics with Si-based photodetectors and logical circuits. However, the main challenge connected with the use of silicon is the high lattice mismatch. Whereas the growth of GaN on sapphire requires one nucleation layer to form the crossover from sapphire to GaN, the growth of high-quality crack-free GaN on silicon requires a sophisticated growth procedure employing a combination of highand low-temperature GaN and AlGaN layers as well as more than one nucleation/recrystallization step [7]. Tensile strain, especially during the cooldown phase after the epitaxial deposition, causes cracks in the nitride layer stack. As a result, the devices grown on Si show usually worse characteristics compared to those grown on sapphire substrate. Study of the spontaneous and stimulated radiation of the structures at optical excitation is an excellent tool to understand the weak points of material and heterostructures and to establish a feedback between the growth technology and laser properties. In this paper, we report on the creation of the optically pumped lasers based on InGaN/GaN MQW heterostructures grown both on sapphire and Si and covering a wide spectral range, and on the study of the properties of these lasers.

Optically Pumped UV-Blue Lasers

2.

299

EXPERIMENTAL

All structures were grown in AIXTRON MOVPE reactors on 2-inch (0001)oriented Al2O3 and on (111)-oriented n-type Si substrates at low pressures (200 mbar or 50 mbar). Trimethylgallium (TMGa), Trimethylaluminum (TMAl), Trimethylindium (TMIn), ammonia (NH3) and silane (SiH4) were used as precursors. The thickness of the InGaN active layer was 3 nm, the thickness of the barriers was 4 nm and the thickness of the upper cap layer was 10-50 nm for the MQWs grown on sapphire (MQW/Al2O3). The structures grown on Si were free of cracks. Photoluminescence (PL) and lasing were excited by the radiation of a N2 laser (hQ = 3.68 eV, Iexc = 102– 106 W/cm2, f = 1000 Hz, Wp = 8 ns), a He–Cd laser (hQ = 3.81 eV) and by radiation of a dye laser with tuning frequency for direct excitation of the quantum wells. The quantum energy of the dye laser was lower than the band gap value of GaN. Figures 1 shows the typical design of the InGaN MQW/Al2O3 and InGaN MQW/Si heterostructures under study, respectively.

a)

b)

Figure 1. Design of InGaN/GaN heterostructures grown on sapphire (a) and silicon (b) substrate

3.

RESULTS AND DISCUSSION

Figure 2 shows the laser spectra of all studied heterostructures (both grown on Si and sapphire) excited by N2 laser radiation at room temperature. The spectra are seen to span over the spectral interval from the near-ultraviolet up to the blue region. The lasers based on InGaN/GaN SQW and GaN/AlGaN single heterostructures grown on sapphire cover the spectral region from O = 370 nm up to O = 470 nm. The lasers based on the structures grown on Si cover the region from O = 383 nm up to O = 462 nm. The best laser parameters were reached for the InGaN/GaN MQW/Al2O3 “violet”

300

G. P. Yablonskii et al

(O = 400–450 nm) lasers. The laser action was achieved up to very high temperature Tmax = 580 K. The minimal laser threshold at room temperature was Ithr = 35 kW/cm2, the full width at half maximum (FWHM) of the laser line near the laser threshold was 0.04 nm, the pulse energy was E = 630 nJ, the pulsed power was P = 80 W and the characteristic temperature was T0 = 164 K at T = 200 – 500 K and T0 = 530 K at T = 80–200 K. The laser parameters of the “blue” lasers were the following: O = 450–470 nm, FWHM = 0.05 nm, Tmax = 460 K, Ithr = 50 kW/cm2, E = 300 nJ, P = 40 W. Lasers based on the MQWs grown on Si have operating temperatures from 300 K to 630 K, characteristic temperatures T0 = 200 K at T < 500 K and T0 = 55 K at T > 500 K, Ithr = 30–50 kW/cm2 and P = 30 W at 300 K. It should be noted that the obtained threshold values for the structures grown on Si are only by an order of magnitude higher than the lowest values obtained for optimized structures grown on bulk GaN substrates (2.4–5.8 kW/cm2 [8]). On the base of measurements of the lasing, stimulated emission and PL spectra as functions of temperature and excitation intensity, it was concluded that at high excitation intensity (Iexc > 1000 kW/cm2) a considerable thermal overheating 'T of about 100 K of the active region takes place. It was shown that it is exclusively due to inherent InGaN/GaN laser radiation which power density on the laser mirrors was evaluated to be I > 5 MW/cm2. The spectral-angular distributions of laser emission in the InGaN/GaN MQW on sapphire substrates at T = 300 K and at the excitation density Iexc | Ithr (the threshold value) and at Iexc = 3.2Ithr are shown in Fig. 3. In this case, the cavity length was about 230 nm. From these patterns one can conclude, based on the calculations of the optical confinement factor *, that lasing takes place in high-order transverse modes. Numerical modeling showed that for heterostructures containing 5 QWs, the optical confinement factor * has its maximal value (about 4%) for closed and leaked modes. With increasing the wavelength O, the value of * increases for the fixed mode order. For heterostructures with 10 QWs, the value of * is about 6% for modes of the 9–10th order at O = 390 nm. For longer wavelengths, the maximum corresponds to the lower-order modes (the 7th order at O = 470 nm). From the spectral-angular distribution of laser emission for InGaN/GaN MQW grown on Si (Fig. 4) one may conclude (and it is confirmed by the preliminary calculations) that the optimum conditions for lasing take place for low-order modes (1st–4th order). Figure 5 shows the power characteristics of the 449-nm InGaN/GaN/Si MQW laser at different excitation levels. The maximum pulse energy achieved was 240 nJ, and the external laser quantum efficiency was 5%. This value is even higher than the maximal values we observed for InGaN/GaN MQWs grown on sapphire (about 3–4% for lasing wavelengths around 430 nm, for longer wavelengths the values were

Optically Pumped UV-Blue Lasers

301

smaller). This is evidence that the use of the set of several strain-reducing layer stacks between the Si substrate and GaN heterostructure permits achieving of the results comparable or even better than those for laser heterostructures grown on sapphire.

Wavelength [nm] 440

Blue

420

400

380

Violet

360

UV on silicon

Laser emission intensity [a.u.]

460

T=300 K

GaN on sapphire

InGaN/GaN MQW

2.6

2.7

2.8

2.9

3.0

3.1

3.2

3.3

3.4

Energy [eV] Figure 2. Laser spectra of the studied MQW heterostructures grown on Si (top) and on sapphire (bottom).

a)

b)

Figure 3. 3D-plots of the spectral-angular distribution of the intensity of laser radiation from InGaN/GaN MQW structure grown on sapphire at the threshold value of the excitation density (a) and above the threshold (b).

G. P. Yablonskii et al

302

InGaN/GaN/Si MQW

2.83

438 Lcav = 600 Pm

T = 290 K

437

N2-laser excitation

Energy [eV]

Wavelength [nm]

439

2.84

2

Iexc = 300 kW/cm 436

-40

-20

0

20

40

Angle[degree]

5

Laser pulse energy [nJ]

140 120

4

100 3

80 60

2

40 1 20 0

0

2000

4000

6000

Excitation pulse energy [nJ]

0 8000

External laser quntum efficiency [%]

Figure 4. Spectral-angular distribution of output radiation of the InGaN/GaN MQW laser grown on Si. 0 degrees corresponds to radiation in the plane of MQW.

Figure 5. Laser pulse energy (triangles) and external quantum efficiency (circles) of the InGaN/GaN/Si MQW laser at different excitation levels.

4.

CONCLUSIONS

Optically pumped lasing was achieved and investigated in a large number of InGaN/GaN MQW heterostructures grown both on sapphire and silicon substrates. The spectral range of lasing was 370–470 nm for heterostructures grown on sapphire and 385–460 nm for structures grown on Si. Lasing was

Optically Pumped UV-Blue Lasers

303

observed at T = 80–550 K for sapphire-grown structures, and at T = 300–630 K for structures grown on Si. The maximum observed output power at room temperature was 40 W and 30 W, respectively. While in MQW heterostructures grown on sapphire lasing takes place on the high order transverse and leaky modes, in heterostructures grown on Si the main operating modes were first–third order transverse modes. The external quantum efficiency of MQW InGaN/GaN heterostructure grown on Si was about 5% which is higher than that measured for MQW heterostructures grown on sapphire. Thus, the use of the set of strainreducing layer stacks containing low temperature AlN layers and graded composition layers for growth of MQW heterostructures on Si allows to achieve laser parameters comparable or better than those for analogous structures grown on sapphire. The achieved values of thresholds were 30–50 kW/cm2, the estimated characteristic temperatures were 55 K above 500 K and about 200 K at lower temperatures.

ACKNOWLEDGEMENTS This work was partially supported by the ISTC grant B-176.

REFERENCES 1. 2. 3.

4.

5. 6. 7. 8.

9.

Morkoc H., Nitride Semiconductors And Devices. Berlin/Heidelberg/New York: Springer, 1999, Chapter 11. Singh R., Dappalapudi D., Maussakas T. D., Romano L. T. Phase separation in InGaN thick films and formation of InGaN/GaN double heterostructures in the entire alloy composition. Appl. Phys. Lett. 1997; 70: 1089–1091. Romano L.T., McCluskey M. D., Van de Walle C. G., Northrup J. E., Bour D. P., Kneissl M., Suski T., Jun J. Phase separation in InGaN multiple quantum wells annealed at high nitrogen pressures. Appl. Phys. Lett. 1999; 75: 3950–3952. Kozodoy P., Hansen M., DenBaars S. P., Mishra U. K. Enhanced Mg doping efficiency in Al0,2Ga0,8N/GaN superlattices. Appl. Phys. Lett. 1999; 74: 3681–3683. Nagahama S., Yanamoto T., Sano M., Mukai T. Characteristics of Laser Diodes Composed of GaN-Based Semiconductor. Phys. Stat. Sol. (a) 2002 190: 235–246. Nagahama S., Yanamoto T., Sano M., Mukai T. Blue-violet nitride lasers. Phys. Stat. Sol. (a) 2002; 194: 423–427. Dadgar, A., Bläsing, J., Diez, A., Alam, A., Heuken, M., Krost, A. Metalorganic chemical vapour phase epitaxy of crack-free GaN on Si (111) exceeding 1 Pm in thickness. Jpn. J. Appl. Phys. 2000; 39: L1183–L1185. Ivanov V.Yu., Godlevski M., Teisseyre H., Perlin P., Czernecki R., Prystavko P., Leszczynski M., Grzegory I., Suski T., Porovski S. Ultralow threshold powers for optical pumping of homoepitaxial InGaN/GaN/AlGaN lasers. Appl. Phys. Lett. 2002; 81: 3735–3737.

Key Word Index

action spectrum, 161 AlGaInN/nitride-alloys, 161 AlGaN, 223, 233 AlGaN/GaN biosensors, 143 AlGaN/GaN heterostructures, 15 AlInGaN, 41 aluminum gallium nitride, 1, 59, 239 ambient lighting, 253 atomic spectroscopy, 271 biophotonics, 161 biosensors, 287 buffer layers, 199, 297 built-in electric field, 41 built-in electric fields, 215 cathodoluminescence, 247 CdTe, 93 charge collection efficiency, 279 chip-size, 253 conducting polymers, 287 diffusion, 93 donor-acceptor molecule, 261 electron capture time, 207 electron-hole plasma, 207 epitaxial growth, 59 epitaxial lateral overgrowth, 189 erythema, 161 exciton localization, 41 fluorescence, 261 fluorescence sensing, spectroscopy, 127 four-wave mixing, 93

GaAs, 93 GaN, 93, 179, 199, 207, 233, 247 GaN polarity, 179 GaN/nitrides, 161 general lighting, 253 group-III nitrides, 31 heat dissipation, 31 heterostructures, 41, 93, 297 HFET, 233 high pressure experiments, 215 high-power-SMD-LED, 253 host materials, 111 HVPE, 15, 189 IBICC, 77 III-nitride semiconductors, 179 III-nitrides, 41, 179 illumination, 253 InAlGaN, 179 InGaN, 93, 179, 297 InGaN/GaN MQWs, 207 integration density, 253 ionizing radiation detectors, 279 laser diode, 247 laser-diodes, 31 lasing, 199, 297 LED-module, 253 light source, 253 light-emitting diodes, 127 luminescence decay, 207 MBE, 179 305

306 medical application, 253 Metal-Semiconductor-Metal detectors, 77 MOCVD, 199, 223, 297 molecular spectroscopy, 271 MOVPE, 189 MSM, 233 narrow band UV detectors, 143 N-face GaN, 179 nonequilibrium carriers, 93 optical measurements, 127 optical nonlinearities, 93 optical output, 253 optical pumping, 199, 297 optical waveguiding, 31 PALE, 59 photoluminescence, 207 photon cascade emission process, 111 piezoelectric polarization, 215 plasma-assisted MBE, 179 polypyrrole, 287 power dissipation, 253 Pr3, 111 quantum cutting, 111 quantum wells, 161, 215 quaternary alloys, 179 quaternary compounds, 41 quaternary InAlGaN compounds, 215 radiation hardness, 279 recombination, 93 red skin, 161 sapphire nitridation, 179 sapphire substrate, 297 Schottky barriers, 233 screening of built-in field, 207

self-assembled monolayers, 261 semi-insulating GaN, 279 silicon substrate, 297 SiO, 199 solar blind photodetectors, 1 solar radiation, 161 solid-state lighting, 253 strain relief, 59 stress management in (Al,Ga)N, 77 superlattice, 223 surface acoustic wave, 239 surface potential, 261 surface recombination velocity, 93 switching effect, 261 thermal management, 253 thermal resistance, 253 thin isolating layer, 253 threading dislocations, 189 threshold powers, 247 threshold, gain, 199, 297 time-resolved spectroscopy, 215 TL lighting, 111 ultraviolet LEDs, 1 ultraviolet sensor, 239 UV damage, 161 UV detectors, 189, 233 UV emitter, 59 UV LED, 15, 271 UV light emitting diodes, 143 UV photodetectors, 161 UV solar blind detectors, 77 wide-band-gap semiconductors, 41 ZnTe, 93

Author Index

Anceau, S., 215 Androulidaki, M., 179 Anufryk, A. S., 297 Aoyagi, Y., 215 Asif Khan, M., 41, 59, 239 Aujol, E., 189 Aumiler, D., 271 Ban, T., 271 Beaumont, B., 189 Besulkin, A. I., 223 Boratynski, B., 233 Böttcher, T., 31 ýiplys, D., 239 Cunningham, W., 279 Czernecki, R., 247 Dikme, Y., 199, 297 Dimakis, E., 179 Dmitriev, A., 15 Dorenbos, P., 111 Duboz, J- Y., 77 Einfeldt, S., 31 Faurie, J-P., 189 Figge, S., 31 Fomin, A. V., 223 Fonavs, E., 261 Frayssinet, E., 189

Gaska, R., 59, 127, 239 Georgakilas, A., 179 Gerca, L., 261 Gibart, P., 189 Godlewski, M., 247 Grandjean, N., 77 Grzegory, I., 247 Gurskii, A. L., 199, 297 Heuken, M., 199, 297 Hirayama, H., 215 Hirsch, L., 77 Hommel, D., 31 Ivanov, V. Yu., 247 Jansen, R. H., 199, 297 Jarašinjnas, K., 93 Juršơnas, S., 199, 207, 261 Kalisch, H., 199, 297 Karpicz, R., 261 Karpov, S. Yu., 15 Kazlauskas, K., 199 Kirichenko, N., 261 KoĔczewicz, L., 215 Kovalenkov, O. V., 15 Kuokstis, E., 41 Kurilþik, G., 207 Lefebvre, P., 215 307

308

àepkowski, S. P., 215 Leszczynski, M., 247 Lundin, W. V., 223 Lutsenko, E. V., 199, 297 Mahlkow, A., 253 Melnik, Yu., 15 Miasojedovas, S., 207 Mosca, M., 77 Muñoz, E., 161 Muzikante, I., 261 Neilands, O., 261 Omnes, F., 77 Pau, J. L., 161 Pavlovskii, V. N., 199, 297 Pechnikov, A. I., 15 Perlin, P., 215, 247 Pichler, G., 271 Porowski, S., 247 Prystawko, P., 247 Rahman, M., 279 Ramanaviciene, A., 287 Ramanavicius, A., 287 Reverchon, J- L., 77 Rimeika, R., 239 Rivera, C., 161 Sakai, S., 279 Sakharov, A. V., 223 Schineller, B., 199, 297

Semond, F., 77 Sereika, A., 239 Shapovalova, E., 15 Shur, M. S., 1, 59, 127, 239 Sizov, D. S., 223 Skenderoviü, H., 271 Smith, K. M., 279 Soukhoveev, V. A., 15 Stutzmann, M., 143 Suski, T., 215, 247 Szymakowski, A., 199 Tamulaitis, G., 41, 199 Teisseyre, H., 215, 247 Tlaczala, M., 233 Tsagaraki, K., 179 Tsatsul'nikov, A. F., 223 Usikov, A. S., 15 Vaitkus, J. V., 279 Valiokas, R., 261 Valkunas, L., 261 Van der Kolk, E., 111 Van Eijk, C. W. E., 111 Vink, A. P., 111 Yablonskii, G. P., 199, 297 Yang, J., 239 Zavarin, E. E., 223 Zubialevich, V. Z., 199, 297 Žukauskas, A., 1, 127, 199, 207