Konakova Book-microondas Cuanticas

8/10/2019 Konakova Book-microondas Cuanticas

1/192

1

NATIONAL ACADEMY OF SCIENCES OF UKRAINEINSTITUTE OF SEMICONDUCTOR PHYSICS

A.E. Belyaev, E.F. Venger, I.B. Ermolovich,R.V. Konakova, P.M. Lytvyn, V.V. Milenin,I.V. Prokopenko, G.S. Svechnikov,

E.A. Soloviev, L.L. Fedorenko

EFFECT OF MICROWAVE AND LASER RADIATIONS

ON THE PARAMETERS

OF SEMICONDUCTOR STRUCTURES

!"#$ %&'() 2002


2/192

2


3/192

3

UDC 621.382.2

!"#$! %!& ' #()*+%,-, !$"+,.'%/!(%0 %( "(+(.*1+$ !"#$! %!& ' #()*+%,-, !$"+,.'%/!(%0 %( "(+(.*1+$ !"#$! %!& ' #()*+%,-, !$"+,.'%/!(%0 %( "(+(.*1+$ !"#$! %!& ' #()*+%,-, !$"+,.'%/!(%0 %( "(+(.*1+$ %("'!"+,!'2%$3,!$4 51+%("'!"+,!'2%$3,!$4 51+%("'!"+,!'2%$3,!$4 51+%("'!"+,!'2%$3,!$4 51+6 66 6316+ .316+ .316+ .316+ .

787878789:8; , .< .,9:8; , .< .,9:8; , .< .,9:8; , .< ., =>?@. ABC.-DE@. FEG?, HI>A .;!J!J!J!JFFFFKJI < .L .,KJI < .L .,KJI < .L .,KJI < .L ., M9JF-?>I . %(% 6?IENFO ,=>?@. ABC.-DE@. FEG?, HI>A .;9>;OM ' .7 .,9>;OM ' .7 .,9>;OM ' .7 .,9>;OM ' .7 ., =>?@. ABC.-DE@. FEG?;3>FE?>;E + .! .,3>FE?>;E + .! .,3>FE?>;E + .! .,3>FE?>;E + .! ., =>?@. @JPF. FEG?, HI>A .;#O@;OF " .. .,#O@;OF " .. .,#O@;OF " .. .,#O@;OF " .. ., ?EF= . ABC.-DE@. FEG?;.B98FBF ! .! .,.B98FBF ! .! .,.B98FBF ! .! .,.B98FBF ! .! ., ?EF= . ABC.-DE@. FEG?;"I>?>HJF?> ' .! .,"I>?>HJF?> ' .! .,"I>?>HJF?> ' .! .,"I>?>HJF?> ' .! ., =>?@. ABC.-DE@. FEG?;5;8MFB?>; - .5 .,5;8MFB?>; - .5 .,5;8MFB?>; - .5 .,5;8MFB?>; - .5 ., ?EF= . ABC.-DE@. FEG?;5555>>>>9>;Q>; < ., .,9>;Q>; < ., .,9>;Q>; < ., .,9>;Q>; < ., ., ?EF= . @JPF. FEG?;

LJ=>IJF?> # .# .,LJ=>IJF?> # .# .,LJ=>IJF?> # .# .,LJ=>IJF?> # .# ., ?EF= . ABC.-DE@. FEG?

6 D>F>KIEABN ROR@JDE@OC>;EFB @E GCEKE9SFJFB IJCG9S@E@O =>R9B=TJFF: D>=OAB?EUBN ;9ER@O;>R@JQ VOI>?>K> ?9ERG @JPFBMF> ;ET9O;OP FEHB;HI>-;B=FO?B; (GaAs, GaP, InP, InSb, Cd x Hg 1-x Te, Si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

.>F>KIEAB: HIOCFEMJFE =9: FEG?>;UB; B AEPB;UB; G KE9GCB FEHB;HI>;B-=FO?>;>N J9J?@I>FB?O . !>FE D>TJ R@E@O ; HIOK>=B @E?>T ;O?9E=EMED,ERHBIEF@ED B R@G=JF@ED R@EIVOP ?GIRB; , :?OD =>;>=O@SR: DE@O RHIE;G C ;OYJ;?ECEFODO HI>X9JDEDO .

For a wide range of semiconductors (GaAs, GaP, InP, InSb, Cd x Hg 1-x Te,SiC), as well as contact structures based on them, that are of importance inengineering, the monograph gives, in a systematized and generalized form,the results of investigation of their properties modification under action ofhigh-power microwave radiations and nanosecond laser pulses. The physicalmodels for production, annihilation and annealing of point defects and impu-rity-defect complexes in irradiated materials are considered. An analysis is


4/192

4

made of the features of radiation-stimulated mass transport, chemical reac-tions, formation of new phases in various contact structures, and they arerelated to the parameters of surface-barrier structures. Much space is givento the elastic stress relaxation due to microwave radiation, depending on theinitial impurity-defect condition of materials, their surface morphology andirradiation mode. The authors substantially lean on their own experience inthis area. They pay special attention to the use of microwave and laser proc-essing of semiconductor materials for improvement of their structural perfec-tion and development, on their basis, of ohmic and barrier contacts havingimproved parameters.

The monograph is intended for researchers and specialists engaged insemiconductor electronics. It may be of use also to lecturers, post-graduatesand undergraduates dealing with the above problems.

! D>F>KIEAOO ROR@JDE@OCOI>;EFZ O >X>XYJFZ IJCG9S@E@Z ORR9J=>-;EFO: D>=OAO?EUOO R;>QR@; VOI>?>K> ?9ERRE @JPFOMJR?O ;ETFZP H>9G -HI>;>=FO?>; (GaAs, GaP, InP, InSb, Cd x Hg 1-x Te, Si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

.>F>KIEAO: HIJ=FECFEMJFE =9: FEGMFZP IEX>@FO?>; O RHJUOE9OR@>; ; >X9ER@O H>9GHI>;>=FO?>;>Q [9J?@I>FO?O . ,FE D>TJ@ HIOK>=O@SR: @E?TJ HIJH>=E;E@J9:D , ERHOIEF@ED O R@G=JF@ED R@EIVOP ?GIR>; , ?>@>-IZD HIOP>=O@R: ODJ@S =J9> R G?ECEFFZDO ;>HI>REDO .

ISBN 966-02-2541-5 ! @J?R@@J?R@@J?R@@J?R@ 789:8; , .< ., !JFKJI < .L .,9>;OM ' .7 ., 3>FE?>;E +.! .,#O@;OF " .. ., .B98FBF ! .! .,"I>?>HJF?> ' .! ., 5;8MFB?>; - .5 .,5>9>;Q>; < ., ., LJ=>IJF?> # .# .2002


5/192

5

CONTENTS

PRINCIPAL ACRONYMS AND SYMBOLS ....................................... 8 INTRODUCTION................................................................................ 12 Part 1. INVESTIGATION OF BOUNDARIES

BETWEEN PHASES IN HETEROGENEOUSSTRUCTURES AND THEIR MODIFICATIONUNDER MICROWAVE RADIATION..................................... 16

INTRODUCTION ..................................................................................... 161.1. EXPERIMENTAL TECHNIQUES.................................................... 181.2. EFFECT OF HIGH-POWER ELECTROMAGNETIC

RADIATION ON THE DEFECT STRUCTUREOF SEMICONDUCTOR MATERIALS............................................ 20 1.2.1. III V semiconductor compounds................................... 20

1.2.2. Cd x Hg 1-x Te semiconductor solid solutions.................. 31 1.3. INTERACTIONS BETWEEN PHASES

IN Me GaAs (InP) CONTACTS STIMULATEDBY MICROWAVE IRRADIATION................................................... 361.3.1. Effect of magnetron irradiation on the physico-

chemical processes in Me GaAs (InP) contacts .......... 36 1.3.2. Structural-chemical processes in Me GaAs

contacts stimulated by gyrotron irradiation ................ 44 1.3.3. Effect of microwave treatment

on the electrophysical parametersof surface-barrier structures ............................................ 50


6/192

6

1.4. EFFECT OF MICROWAVE RADIATION ON THEPARAMETERS OF RESONANT TUNNELING DIODES ............. 61

1.4.1. GaAs tunneling diode with two -layers in SCR ......... 61

1.4.2. Two-barrier resonant tunneling diode........................... 66 1.5. THE FEATURES OF STRUCTURAL CHANGES

IN GaAs SINGLE CRYSTALSUNDER MICROWAVE IRRADIATION.......................................... 67

1.6. EFFECT OF MICROWAVE IRRADIATION ON THEPROCESSES OF INTRINSIC STRESS RELAXATIONIN GaAs EPITAXIAL STRUCTURES .............................................. 82

1.7. EFFECT OF MICROWAVE RADIATIONON THE GaAs MESFET CHARACTERISTICS............................. 89

1.8. EFFECT OF MICROWAVE RADIATIONON THE PARAMETERS OF DISCRETE DEVICES

AND INTEGRATED CIRCUITS....................................................... 94

Part 2. LASER METHODS FOR DIAGNOSTICS ANDMODIFICATION OF SEMICONDUCTOR COMPOUNDS(GaAs, InSb, SiC), AS WELL AS FORMATION AND

DIAGNOSTICS OF STABLE OHMIC CONTACTS............. 101 INTRODUCTION ................................................................................... 1012.1. MODEL FOR DEFECT PRODUCTION

IN III V SEMICONDUCTOR COMPOUNDSBY NANOSECOND LASER PULSES............................................ 104

2.2. TECHNIQUES AND EQUIPMENT .............................................. 108

2.2.1. A plant for laser pulse sputtering ................................ 108 2.2.2. A plant for laser modification of semiconductors ..... 109

2.3. LASER MODIFICATION OF GaAs AND InSb ........................... 1112.4. LASER DIAGNOSTICS OF NONEQUILIBRIUM

CHARGE CARRIERS IN SEMICONDUCTORS.......................... 121

2.4.1. A contactless technique for determinationof bipolar diffusion coefficientfor nonequilibrium charge carriers............................... 121


7/192

7

2.4.2. A complex technique for diagnosticsof electron-hole plasma .................................................. 123

2.4.3. Photoconductivity of InSb

under picosecond excitation ......................................... 130 2.5. DEVELOPMENT OF STABLE OHMIC CONTACTS

TO n - -SiC AND p -GaAs USING LASER TECHNOLOGY....... 1382.5.1. Contacts to n - -SiC......................................................... 138 2.5.2. Contacts to p -GaAs ......................................................... 144

2.6. LASER-THERMAL DIAGNOSTICS OF METAL SEMICONDUCTOR CONTACT STRUCTURES ........................ 149

CONCLUSION................................................................................... 162 REFERENCES .................................................................................... 163


8/192

8

PRINCIPAL ACRONYMS AND SYMBOLS

A - acceptor

AES - Auger electron spectroscopyD - donorG-(ir)radiation - gyrotron (ir)radiationIC - integrated circuitLI - laser irradiationLM - laser modificationMBE - molecular-beam epitaxyMCE - magnetoconcentration effectMESFET - metallized semiconductor field-effect transistorM-(ir)radiation - magnetron (ir)radiationNCC - non-equilibrium charge carriersPC - photoconductivityPL - photoluminescencePR - plasma resonanceQFR - quasi-forbidden reflection

RTD - resonant tunneling diodeSB - Schottky barrierSCR - space-charge regionTBRTD - two-barrier resonant tunneling diodeTIQFR - total intensity of quasi-forbidden reflectionsUSP - ultrashort pulsesXPS - x-ray photoelectron spectroscopyXRD - x-ray diffraction

( * - Richardson constantB - magnetic inductionC - capacitancec - concentration in atomic %D - diffusion coefficient


9/192

9

d - thicknessd 0 - crystal lattice constantd f - film thickness

d spacing between two -layersd - depth measured from the sample surfaceE - energy; electric fieldE a - activation energyE b - electron binding energyE F - Fermi energyE g - semiconductor energy gapE j - electric field in the p -n junctionE p - pulse energy* t - energy position of recombination center;

impurity energy levelE th - electric degradation thresholdF - forcef - magnetron (gyrotron) frequencyH - magnetic field

H 1/2 - PL band half-widthh (= 2 ! ) - Plancks constantI - radiation intensity; currentI c - cutoff currentI ch - channel currentI d - drain currentI ex - excess currentI i - flow of diffusing ions of the i -th typeI PL - intensity of photoluminescenceI R - reverse currentI th - threshold value of radiation intensityK - light absorption coefficientk - Boltzmann constantL - NCC bipolar diffusion length


10/192

10

L D - Debye shielding lengthL p - hole diffusion lengthl - mean free path

l j - width of p -n junctionN A - acceptor concentrationN D - donor concentrationN d - dislocation densityN i - concentration of ions of i -th typeN p - number of laser pulsesN t - concentration of recombination centers;

impurity concentrationN v - vacancy concentrationN - impurity concentration in the -layersn - electron concentration; ideality factorn i - intrinsic charge carrier concentrationn s - surface electron concentrationP - power

- hole concentration

Q* - heat of vacancy migrationq 0 - elementary chargeR c - contact resistanceR H - Hall constantr - bulk recombination rater, , z - cylindrical coordinatesr AD - distance between D and AS - transconductanceS b - laser beam cross sectionS c - contact areaS r - surface recombination velocityS s - sample surface areaT - temperaturet - time


11/192

11

t G(M) - duration of gyrotron (magnetron) irradiationt p - pulse durationV - voltage

V c - cutoff voltageV R - reverse voltageW - irradianceW B - potential barrier height W th - thermal degradation thresholdw - width of space-charge region - depth of electromagnetic field penetration - electron energy 0 - semiconductor permittivity - quantum efficiency of photoionization -wavelength -charge carrier mobility 0 -semiconductor permeability = E/ " - frequency - resistivity

- conductance hkl - specific surface energy I - hole lifetime - potential ! - Schottky barrier height - angular frequency


12/192

12

INTRODUCTION

The scientific and technological progress in solid-state elec-tronics and related areas derives, to a great extent, from use

of high technologies. Among them are the so-called beamtechnologies, namely, laser and microwave ones [1-8]. Theirfeature is a wide range of applications involving both militaryand non-military ones. To illustrate, the above technologieshave been used when developing the means for functionaldamage (the Star Wars and Strategic Defense Initiative pro-grams) [9-11]. As to their non-military applications, they in-

volve, among others, controlled synthesis of new materials

[12-16], novel gettering techniques [17,18] and formation ofohmic and barrier contacts [19-23]. Besides, the above tech-nologies are used in fundamental investigations of interactionbetween focused beams and nonuniform media [24,25] (in par-ticular, studies of diffusion and mass transport [26,27]). Simu-lation of laser and microwave actions on discrete semiconduc-tor devices and integrated circuits [28,29] is intensely used to

predict reliability of the developed element base.Microwave and laser treatments are used for rapid an-nealing of semiconductor layers after ion implantation [30-33].Many authors have studied various effects of such treatmentson the properties of GaAlAs/GaAs/GaAlAs/GaAl thin-filmmultilayer structures [34], electrical parameters of threadlikeSi crystals, Ge Si solid solutions, GaAs and GaAsP, strength of


13/192

13

GaAs [35] and Si [36] crystals, photoelectric properties ofnonuniform CdS single crystals and nonequilibrium effects inCdS photodetectors [37,38], possibilities for control over the

defect structure of Si and Ge single crystals [39].Numerous practical aspects of using microwave and la-

ser radiations in technology of semiconductor materials anddevices are protected by many patents issued in differentcountries. To illustrate, let us note only those dealing withprocessing of semiconductor materials and devices [40-44],silicon ingot slicing [46], doping of semiconductor materialsand formation of low-resistance contacts [19]. One can seethat they cover practically all stages of manufacturing ofsemiconductor materials, structures and devices.

Application of microwave and laser processing becomesparticularly efficient when producing devices of submicronsizes. These techniques make it possible to get rid of a num-ber of technological flaws inherent in traditional thermalprocessing. In this case one can exert control over the micro-

device characteristics and correct them using electromagneticradiation. In recent years such experiments have been per-formed for tunnel diodes [47,48] and planar-epitaxial multiply-ing diodes 2A604 [49-52]. These experiments demonstratedthat negative differential resistance might appear in the abovedevices. This was concluded from the form of the forwardbranch of I V curves of tunnel diodes (at voltages below thepeak values) and multiplying diodes. The above effect is ofimportance for understanding of interaction between micro-

wave radiation and such complex semiconductor objects.The physical effects occurring in thin semiconductor

films and at semiconductor surfaces exposed to laser radia-tion have been discussed as early as 30 35 years ago by

Vavilov and Galkin [53,54] and Fairfield, Schwuttke, Harper


14/192

14

and Cohen [55,56]. In that time many research centers allover the world have enjoyed the laser boom. As a result,the efforts of numerous investigators have been focused on

laser physics and its applications in solid state physics. Thepioneer works on laser pulse annealing of ion-implantedsemiconductor layers have been made by Khaibullin (Phy-sico-Technical Institute of the Kazan Division of the Acad-emy of Sciences of the USSR) and Smirnov et al. (Institute ofSemiconductor Physics of the Siberian Division of the Acad-emy of Sciences of the USSR) [57-61]. Various practical as-pects have been considered by Hess, Olson, Gibbons andother researchers who were at the outset of application oflaser techniques in electronic industry [62-64]. Similar inves-tigations have been also made at the Institute of Semicon-ductors of the Academy of Sciences of the Ukrainian SSR(now Institute of Semiconductor Physics of the National

Academy of Sciences of Ukraine) by Litovchenko, Glinchuk,Lysenko, Malyutenko et al. [65-67], Institute of Physics of

the Academy of Sciences of the Ukrainian SSR by Brodin etal. [68,69] and Institute of Applied Problems of Mechanicsand Mathematics by Tovstyuk, Kiyak et al. [70-72]. It is prac-tically impossible to name all the research teams that havecontributed to understanding of physical foundations of lasertechnologies and their application for manufacturing ofsemiconductor materials and devices.

However, despite the fact that the technologies usingelectromagnetic radiation seem rather simple and our knowl-edge of the corresponding processes is extensive, a number ofproblems, both fundamental and technological, still remainunsolved. They involve, in particular:! micrometallurgical processes at semiconductor surfaces and

interfaces enhanced by electromagnetic radiation;


15/192


16/192

16

Part 1. INVESTIGATION OF BOUNDARIES BETWEEN PHASESIN HETEROGENEOUS STRUCTURES AND THEIRMODIFICATION UNDER MICROWAVE RADIATION

INTRODUCTIONRequirements for the semiconductor element base involve itshigh tolerance to various beam actions, in particular, to high-power microwave irradiation. So far most of theoretical andexperimental studies have been aimed at elucidation of therole of static (or rather low-frequency) actions in degradationprocesses that occur in semiconductor materials and devicesduring their operation. The problem of interaction betweenhigh-power microwave electromagnetic fields and semicon-ductor materials and devices that would combine approachesof both physics and materials science has not been studied upto the present. There are practically no data relating electro-physical properties of the above objects to the features oftheir microstructure and chemical composition.

At the same time it is known that exposure of devicestructures and finished products (both discrete and integrated)to microwave irradiation results, in a number of cases, to theirdegradation and even failure [1 4]. Of all the components ofany circuit, semiconductor elements are most sensitive tohigh-power microwave radiation [2]. Depending on the ab-sorbed power, frequency, temporal parameters of action andobject tolerance, one can observe different stages of devicedegradation, namely, (i) short-duration failures - at low levels


17/192

17

of action; (ii) irreversible changes of parameters and character-istics with relative retention of efficiency at moderate levelsof action; (iii) complete failures (accompanied by thermal ef-

fects and electric breakdown) leading to loss of efficiency at high levels of action. The power (frequency) of radiationthat leads to device degradation ranges between several Wand several hundreds W (50 MHz and 10 GHz).

On the other hand, some authors have indicated at apossibility for microwave treatment application to modify de-fect structure in semiconductors [5 8]. To illustrate, it wasshown in [5] that microwave radiation of frequency f =

2.5 GHz could be used for post-implantation annealing of sili-con, and such treatment proved to be more efficient thanthermal annealing. The authors of [6] stressed that high-powermicrowave radiation is promising for rapid contactless anneal-ing of GaAs. They reasoned that no abrupt temperature gradi-ents and diffusion stresses appeared in this case.

A considerable modification of dislocation structure dueto microwave annealing ( f = 37 GHz, power P = 0.1 kW) hasbeen found in [7] for silicon wafers. The kinetics of residualstress relaxation was found to substantially differ from that inthe case of traditional thermal annealing in the air or vacuum.It was characterized by complicated oscillations with pro-nounced non-thermal sections. In [8] microwave relaxometryhas been applied to demonstrate that microwave irradiation ofn -Si results in changing parameters of impurity aggregations

and growth of charge carrier lifetime.The above results served as a reason for performing

some experimental studies whose objective was to determineeffect of microwave radiation beams on structural-chemicaland electrophysical characteristics of both semiconductor ma-terials and device structures. The results obtained are given in

what follows.


18/192

18

1.1. EXPERIMENTAL TECHNIQUES

The investigations were concerned with different subjects. Firstof all, we dealt with single-crystalline wafers of gallium ar-senide, indium phosphide and gallium phosphide, with freeelectron concentration n from about 10 16 up to 2 1017 cm -3.They were prepared for production of device structures usingstandard technological procedures. Besides, we studied n -Cd x Hg 1-x Te ( x = 0.21 0.24) wafers that have been cut out ofingots and put through mechanical treatment and etching.

The second class of subjects used by us involved

metal semiconductor structures of III V type. They were pre-pared in a vacuum (pressure of 10 -4 Pa) using thermal or elec-tron-beam evaporation of metals onto chemically cleansed sur-faces of GaAs, GaP and InP. The metals used for contact for-mation enabled us to obtain either chemically inert or chemi-cally active metal semiconductor interfaces [9]. In the lattercase it was possible, by proper choice of heteropairs, providedomination of either diffusion processes at interfaces orchemical reactions between the contact pair components.

Contact structures with antidiffusion barriers based onmetal nitrides and borides belonged to the third class of sub-

jects that we investigated. They were fabricated using magne-tron sputtering [10]. And still another group of subjects wasmade by Al x Ga 1-x As GaAs resonant tunneling homo- and het-erostructures that were formed using molecular-beam epitaxy

(MBE) [11].The samples (wafers) studied were exposed to directional

irradiation from cm and mm wavelength ranges under free-spaceconditions. The output signal power P for magnetron irradiation(M-irradiation) was 5 kW at a frequency f = 2.45 GHz. That forgyrotron irradiation (G-irradiation) was 50 kW at f = 84 GHz. To


19/192

19

perform comprehensive studies of the features of interaction be-tween microwave radiation and subjects of investigation, we ap-plied a complex of techniques that enabled us to obtain informa-

tion on electrophysical parameters of samples, their defect struc-ture and chemical composition.

The spectra of local states due to structural defects werefound from analysis of photoluminescence (PL) curves takenat a temperature T = 77 K in the 0.6 2.0 eV spectral range.The samples were illuminated with light ( h > 2 eV) from ahigh-power incandescent lamp !" -100.

Measurements of Hall effect and dark conductivity gaveus data on electron concentration n and mobility in thesamples studied. The lifetime p of minority charge carriers(holes), as well as concentration N t and energy position E t ofrecombination centers, were determined from temperature de-pendence of photoconductivity (PC) [12]. An analysis ofsteady-state I V curves [13] enabled us to determine the fol-lowing characteristics: from forward branches the contact

barrier heightc

c B I

S T A q kT

2

0ln

= (here I c is the cutoff cur-

rent, A * is Richardson constant, k is Boltzmann constant, q 0 isthe elementary charge, S c is the contact area) and ideality fac-

tor( )I

V kT q

n ln

0

= ; from backward branches effective life-

time of minority charge carriersR

c i p I

wS n q 2

0= (here I R is the

reverse current, w is the space-charge region (SCR) width, n i is the intrinsic charge carrier concentration).

The concentration depth profiles of the heterostructurecomponents were determined using Auger electron spectros-copy (AES) [14] combined with layer-by-layer etching by Ar + ions (energy of 1 keV). The Auger spectra were taken in the


20/192

20

differential mode. The primary electron energy was 3 keV.The component concentrations in the transition layer werecalculated using the elemental sensitivity coefficients [14].

The information concerning phase composition of thesubjects studied was obtained using x-ray photoelectron spec-troscopy (XPS) [14]. Mg K -emission ( h = 1253.6 eV) servedas source of excitation.

1.2. EFFECT OF HIGH-POWER ELECTROMAGNETICRADIATION ON THE DEFECT STRUCTURE

OF SEMICONDUCTOR MATERIALS

1.2.1. III V semiconductor compounds Shown in Figs.1.1 1.5 are the PL spectra of some III V semi-conductor compounds, as well as band intensity I PL, half-widthH 1/2 and peak position h m as function of time t of exposureto magnetron and gyrotron treatments. The treatment modesare given in figure captions [15].

One can see that there are two overlapping bands in thePL spectra of initial GaAs crystals. Their peak positions 1m h and 2m h , as well as ratio between intensities, depend on thedopant type and sample surface orientation.

GaAs:Sn (111). In these crystals the spread of 2,1m h values forthe observed PL bands is bigger than those in other samples.It was possible to separate the samples studied into threegroups, namely, those with 1m h = 1.150, 1.200 and 1.186 eV.The 2m h values for the second band demonstrated smallerspread; they were 2m h = 0.993 1.010 eV. The intensity ofband 1 was by a factor of 2 5 bigger than that of band 2. Thehalf-width of band 2 was, as a rule, less than that of band 1(100 150 and 200 230 meV, respectively).


21/192

21

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.5

1.5

1.5

1.5

1.5

1.5

0.5

0.5

I , arb.u.1.20eV

1.002eV

1.01eV

1.01eV

1.22eV

1.28eV

1.28eV

1.185eV

1.185eV

1.1

1.4()

(d)(e)

(f)

(b)

(c)

1

12

2

PL

I , arb.u.PL

Fig.1.1. PL spectra taken at T = 77 K for GaAs:Sn (111) (a c) and GaAs:Sn(100) (d f) samples before (a, d) and after magnetron (exposure timet M = 10 s (b), 60 s (c)) and gyrotron (exposure time t G = 1 s (e1), 10 s(e2), 4 s (f1), 40 s (f2)) irradiation [15].


22/192

22

The band 1 parameters are more affected by microwaveirradiation. In the case of M-irradiation even at exposure timet M < 2 s the 2,1m h values become the same for the samples

from all groups: 1m h = 1.185 eV, 2m h = 1.010 eV. These are just the values that are observed for the initial samples fromthe third group. The 2,1m h values change unevenly; for thesamples belonging to the first group both band intensity andhalf-width change just in this way.

For the crystals from the first and second groups the in-tensity of band 1 grows slightly with t M , and this band be-

comes narrower, while the parameters of band 2 remain prac-tically the same. For the crystals belonging to the secondgroup at t M = 10 and 20 s 2m h = 1.280 eV, and the intensityof band 2 becomes higher than that of band 1. M-irradiationdoes not affect the parameters ob both bands in crystals fromthe third group.

G-irradiation of the crystals belonging to the firstgroup strongly decreases the intensity of band 1, whileslightly affecting that of band 2. The case of the mode withtime of exposure t G = 40 s is an exception: the intensity ofband 1 drops even more abruptly, but that of band 2 in-creases and becomes higher than the intensity of band 1. Incrystals from the third group the PL peak positions do notshift (just as in the case of M-irradiation), but intensities ofboth bands grow (that of band 2 grows somewhat stronger).

As a rule, M- and G-irradiations of the crystals from allgroups lead to a decrease of the half-widths of both bands.

When the h m values change abruptly, then the half-widthsof both PL bands also change abruptly at the same valuesof t M,G .


23/192

23

H , meV

1

2

0

0

010

10

10

10

10

10

30

30

30

30

30

30

50

50

50

40

4080

40

50

50

80

0

120

120

120

100

100

200

150

150

100

200

200

200

200

200

160

220

220

220

180

180

180

180

180

140

180

4

4

4

4

4

4

5

3

3

3

3

3

3

2

2

2

2

2

2

1

1

1

1

1

1

t, s

t, s

t, s

t, s

t, s

t, s

+

+

+

+

+

+ + + + +

+

+

+

+

+

+

+

+

+

+

+++

+

+

+

+

+

+

+

+

+

+

+

+

()

(e)

(c)

(b)

(f)

(d)

h , eV m

h , eV m

h , eV m

I , arb.u.PL

I , arb.u.PL

I , arb.u.PL

I , arb.u.PL

I , arb.u.PL

I , arb.u.PL

H , meV h , eV m

H , meV h , eV m

H , meV

H , meV

H , meV

Fig.1.2. Band intensity I PL (curves 1, 2), half-width H 1/2 (curve 3) and peak position h m (curves 4, 5) as function of exposure time t for bands 1(curves 1, 3, 4) and 2 (curves 2, 5) taken at T = 77 K for GaAs:Sn(111) samples from groups 1 (a, b), 2 (c, d) and 3 (e, f). a, c, e magnetron irradiation, b, d, f gyrotron irradiation [15].


24/192

24

GaAs:Sn (100). At this orientation of the GaAs:Sn crystal sur-face two overlapping PL bands are also observed. In this case,

however, the band 2 with 2m h = 0.996 1.002 eV is more in-

tense. For the band 1 1m h = 1.220 1.240 eV for differentsamples in the initial state (see Fig.1.3). At t M = 2 s the in-tensity of band 2 drops abruptly, and the peak positions be-come 1m h = 1.240 eV and

2m h = 1.010 eV. For times of ex-

posure t M lying in the 2 30 s range 2,1m h retain their values;the intensity of band 1 decreases, while that of band 2 grows.

At t M

= 60 s the1m h value becomes equal to 1.280 eV, as

that for the GaAs:Sn (111) samples.

2

1

0 10 1030 3050

40

80

0

120

200

160

200160

4 5

3 3

2

21

1

t, s t, s

++ + ++

++

+

+

+++

+

++

(a) (b)

+ ++

+

+

+++

h , eV m

h , eV m

H , meV h , eV mI , arb.u.PLI , arb.u.PL

Fig.1.3. Band intensity I PL (curves 1, 2), half-width H 1/2 (curve 3) and

peak position h m (curves 4, 5) as function of exposure time t for bands 1 (curves 1, 3, 4) and 2 (curves 2, 5) taken at T = 77 K forGaAs:Sn (100) samples. a magnetron irradiation, b gyrotron ir-radiation [15].


25/192

25

At t G = 1 s the intensity of band 1 (2) grows (de-

creases); the 1m h value does not change and is 1.240 eV,

while 2m h value becomes 1.010 eV. At t G = 10 s the 1m h

value becomes 1.280 eV and remains the same for t G up to40 s. In this case the intensity of band 1 is higher than that ofband 2, and such relationship between them retains up to themaximal value (40 s) of the time of exposure t G used Fig.1.3b.

GaAs:Te (111). The M-radiation practically does not affectboth the half-width and m h value for the PL band 1.200 eV,

while its intensity grows with time of exposure t M (Fig.1.4b).G-radiation does not influence the band form too, while itsintensity changes non-monotonically: it grows at t G up to 10sand then decreases down to the initial value (at t G = 40 s) Fig.1.4c.

InP (100). Three PL bands are observed in the PL spectrum of

the initial crystals: that with 1m h = 1.410 eV, a band at1.150 eV with structural features and a band at 0.820 eV. Themost intense is the band at 1.410 eV - see Fig.1.5. Both M-and G-radiations do not change the peak positions for the ob-served bands, but that at 1.150 eV becomes devoid of itsstructure. The intensity of band 1.410 eV slightly drops atsmall times of exposure t M and then becomes to grow. Theintensities of the rest of bands grow too but much moreslightly than that of the band 1.410 eV see Fig.1.5b, c.


26/192

26

1.0

1.0 1.5

0.5

0

()

(b) ()

0 010 1030 3050

80

160

240

t, s

I , arb.u.PL

I , arb.u.PL

Fig.1.4. a - PL spectra taken at T = 77 K for GaAs:Te (111) samples before(full curve) and after (dots) gyrotron irradiation (exposure time t G =40 s); b (c) band intensity I PL as function of exposure time for mag-netron (gyrotron) irradiation [15].

The above results evidence that the initial impurity-defect state of the GaAs near-surface layers is determined bythe type of dopant and crystallographic orientation of the wa-fer face, as well as by unintentional impurities that can con-taminate the crystal surfaces during their treatment (for in-stance, from the etchant). In addition, one should take into

account that intrinsic stresses (that appear in the GaAs wafersduring single crystal slicing and further chemical-mechanicaltreatment) can also affect the defect composition of the near-surface layers through their enrichment or depletion with va-cations.


27/192


28/192

28

It is known from literature that the centers responsiblefor the band 1 in GaAs:Sn with 1m h = 1.150 1.220 eV are thedonor-acceptor (D A) complexes ( V Ga + Sn Ga ), while thoseresponsible for the band 2 are the individual acceptors Cu Ga (see, e.g., [16]). Our experimental results, in particular the factthat the band 2 is rather narrow as compared to the band 1,are not in contradiction with the above suggestions. A spread

of the 2,1m h values that has been observed by many authors,as well as by us for different initial samples, could result fromboth different distances r AD between donor (D) and acceptor

(A) in the above complexes and a possibility of nonequivalentD and/or A positions in the crystal lattice. The latter factormay stem from the presence of other defects, intrinsicstresses, dislocations, etc. in the nearest neighborhood of Dand A. All this determines the so-called nonuniform broaden-ing of the PL bands.

The ratio between the intensities of bands 1 and 2 in theGaAs:Sn wafers that have been cut of the same single-crystalline ingot and exposed to the same chemical-mecha-nical treatment but have different orientation of crystallo-graphic planes may vary from sample to sample. This may berelated to high V Ga concentration in the near-surface layers ofthe GaAs (111) face as compared to the case of the (100) face.It is known that the energy properties of III V crystals areanisotropic. This means that the surface energy values for

principal crystallographic planes are different [17]: 100 > 110 > 111 , where hkl is the specific surface energy. So one couldassume that concentration of intrinsic defects (vacancies) ishigher at the crystal face with smaller surface energy.

Both abrupt change of the peak position for band 1 inGaAs:Sn (111) and change of the band half-width occur at


29/192

29

very short times of exposure t M . This fact indicates that evensuch short-term microwave actions lead to lattice ordering andmake the r AD values in complexes equal for different samples.

The centers responsible for 1m h = 1.185 eV were found tobe the most stable D A complexes. It will be recalled that justthe same value of 1m h was observed in the PL spectra for thesamples belonging to the third group. Their PL properties didnot change during M-irradiation up to the highest exposures(t M = 60 s). It seems likely that the GaAs:Sn (111) crystalsfrom the third group were the most perfect (uniform) among

all the crystals we studied.The effect of M-irradiation on the band 1 in GaAs:Sn

(111) crystals is characterized by some selectivity. This factindicates that when lattice becomes more ordered during mi-crowave irradiation in the modes used, then the channel ofradiationless recombination is not affected. Otherwise the in-tensities of both bands would change in similar ways( 1=

!i i R under the same conditions of PL excitation before

and after microwave action; here R i is the fraction of the re-combinating charge carriers flow through the i -th channel ofradiation(less) recombination [18]). Just such similar behavior ofboth bands has been observed for the GaAs:Sn (111) crystalsfrom the third group under G-irradiation. This fact indicatesthat G-irradiation reduces the contribution from the radia-

tionless recombination channel to the total recombination flow.Some small growth of the band 1 intensity with time ofexposure t M for the GaAs:Sn (111) crystals from the first andsecond groups seems to result from the fact that M-radiationeither enhances V Ga concentration growth in the near-surfacelayers of the above crystals or favors more intense formationof complexes by V Ga and Sn Ga that are present there. (The lat-


30/192

30

ter process is realized through lattice ordering by removingother defects that could prevent the complex formation.) Thefirst of the above ways should be favored because, if the sec-

ond way would realize, then one could observe transformationof a PL band related to V GA into the 1.185 eV band, and thishas not been experimentally found. The experiments also evi-dence that the band 2 is related to the unintentional individ-ual Cu Ga impurities. In the case of GaAs:Sn (111) M-radiationpractically does not affect this band.

In the GaAs:Sn (100) crystals the PL band with 1m h =1.240 eV was observed. The centers responsible for this bandare acceptors Sn As [16]. At t

M = 60 s and t G 10 s the V Ga concentration increases, and the (Cu Ga + V Ga ) complexes re-sponsible for the 1.280 eV band are formed [16]. That complexformation occurs at the above irradiation modes is confirmedby the fact that the intensity changes for both bands are pro-portional (Fig.1.3).

In the GaAs:Te (111) crystals the centers ( V Ga + Te Ga )

responsible for the only band with m h = 1.200 eV [18] arestable. Neither M- nor G-radiation changes the form of thisband. And the band intensity change with t M,G seems to resultfrom the effect of irradiation on the radiationless recombina-tion centers.

No unambiguous interpretation of the band 1.410 eV inInP exists in literature. Some authors relate it to the Si unin-tentional impurity that is encountered most often, while theothers relate it to the V P (vacancies of phosphorus) or band-to-band recombination of the nonequilibrium charge carriers(NCC). The asymmetric band in the 1.060 1.150 eV range thatoften has a phonon structure is related to the (Fe In + V P)complexes [16]. The fact that the 1.410 eV band is relativelynarrow (as compared to the 1.150 eV band) enables one to


31/192

31

prefer the model of individual acceptors ( V P) for centers re-sponsible for that band. An increase in the intensities of bothbands after M- and G-irradiations indicates at a possible

growth of the concentration of phosphorus vacancies V P (re-sponsible for the 1.410 eV band) that are among the centersresponsible for the second band. The individual acceptors V P have larger capture cross-section for nonequilibrium holes ascompared to the case of D A complexes [18]. Thereforegrowth of the V P concentration has to make a strong effect onthe intensity of the 1.410 eV band. This is just what has beenobserved experimentally.

Thus we have shown that microwave irradiation of GaAsand InP crystals affects their defect structure and modificationof centers responsible for radiative recombination. The fea-tures of this influence depend on the dopant type and waferface orientation. We have demonstrated an important role of

vacancies in transformation of local centers in the near-surfaceregions of differently oriented GaAs wafers. The origin of cen-

ters responsible for PL in GaAs has been made clear. The factsthat PL bands become narrower and PL peak positions are thesame under M- and G-irradiations seem to be due to localheating of the nonuniformity regions.

The structure of GaP local centers practically did notchange under microwave irradiations of cm and mm wave-length ranges that have been used in the cases of GaAs andInP [19].

1.2.2. Cd x xx x Hg 1111----x xx x Te semiconductor solid solutions Generally both the initial impurity-defect composition andmorphology of semiconductor material influence the characterof changes in its electrophysical parameters due to microwaveirradiation. Indeed such influence has been confirmed by the


32/192

32

results of our studies of the narrow-gap semiconductors basedon Cd x Hg 1-x Te solid solutions [19].

To illustrate, presence in the initial sample (#1 - see

Table 1) of nonuniformities capable of leading to anomalieson the Hall constant vs temperature curve R H (T ) (see Fig.1.6)manifests itself after microwave treatment through conductiontype conversion and abrupt change in the minority chargecarrier lifetime p (Fig.1.7). For the rather uniform samples(#2, 3), however, the character of changes is qualitatively dif-ferent. Microwave treatment of such samples leads to a sub-stantial rise of p . In this case the Hall constant R H growsslightly in the impurity conduction region, and mobility somewhat decreases. The results obtained from the tempera-ture dependence of p evidence that the concentration of theactive recombination centers whose energies are in the upperhalf of the gap decreases.

Table 1.1. Effect of microwave irradiation ( f = 2.45 GHz, W = 5 kW/cm 2,t M = 5 s) on the electrophysical parameters (free electron concen-tration n and mobility , impurity ionization energy E t andconcentration N t ) of Cd xHg 1- xTe at T = 77 K [19].

Sample#

Composi-tion ( x )

Type oftreat-ment

n , cm -3 ,cm 2/V s

p ,s

E t ,eV

N t , cm-3

1 0.24 I*MI

6.9 10 15

2.54 10 16 210 4

610 3 0.61

0.2462 0.21 I

MI5.12 10 14

4.9 10 14 210 5

1.5 10 51.42.4

0.10.1

1.1 10 14

3.6 10 13 3 0.22 I

MI4.82 10 14

4.6 10 14 1.4 10 5

9.6 10 4 2.13.0

0.70.7

6.25 10 13

2.5 10 13 * I initial sample; MI that after microwave irradiation.


33/192

33

10

10

10

2

3

3H

4

2

2

4 6 8 10 12

1

141000/T, 1/K

R , cm /C

Fig.1.6. Hall coefficient R H vs inverse temperature curves for sample 1(Cd 0.24Hg 0.76Te) before (1) and after microwave irradiation for 5 s (2)[19].

10-5

10-6

10-7

2 4 6 8 10 12

1000/ T, 1/K 14

, sp

12

Fig.1.7. Minority charge carrier lifetime p vs inverse temperature curvesfor sample 2 (Cd 0.21Hg 0.79Te) before (1) and after microwave irradia-tion for 5 s (2) [19].


34/192

34

Investigations of the effect of plastic deformation on theproperties of samples from this group were performed bothbefore and after microwave treatment. They have revealed an

effect of improved tolerance of the electrophysical parametersof the irradiated samples for plastic deformation. We realizedplastic deformation of the samples by indentation with a

Wickers prism. The indenter load was 10 g; the indentationdensity was N ind = 10

4 cm -2. One can conclude from Fig.1.8that after microwave treatment the temperature dependenciesof p taken for the samples exposed to indentation were thesame as those before microwave treatment [20].

10

10

2

-6

2

3

4 6 8 10 12

1

14

1000/T, 1/K

, sp

Fig.1.8. Minority charge carrier lifetime p vs inverse temperature curvesfor Cd 0.2Hg 0.8Te samples before (1) and after microwave treatment(2); 3 18 h after microwave treatment and indentation. Dotted linecorresponds to the case of indentation load of 10 g and indentationdensity of 10 4 cm -2 [20].


35/192

35

No R H changes in the impurity conduction region havebeen observed, and the ionization energy of the active recom-bination centers remained the same, contrary to the case of

the samples that were not exposed to microwave treatment.If one assumes that the main energy dissipation

mechanism in the irradiated samples is heating, then the ob-served features of the changes in material parameters couldresult from the enhanced migration of the active recombina-tion point defects and impurities to the energy-stable drains,such as low-angle boundaries, dislocations, etc. One shouldnot also exclude decay of the metastable defects that arefrozen at room temperature and diffusion of their compo-nents to the thermally stable extended defects. Such en-hanced gettering manifests itself through some p growth.The dislocations that are introduced under indentation pene-trate through the whole sample thickness rather quickly (the p values measured for both indented and non-indented sam-ple surfaces are the same). However, formation of their at-

mospheres occurs under conditions when the number of de-fects that are forming the impurity dislocation atmospheres issubstantially below that in the samples that were not ex-posed to microwave treatment. As a result, the dislocationsthat are introduced under indentation cannot form the impu-rity atmospheres that could qualitatively modify the recom-bination characteristics of a sample.

Thus microwave treatment of semiconductor materialsessentially influences the structure-sensitive characteristics fornot only wide-gap semiconductors but narrow-gap ones as

well. A degree of this influence depends on the nature andpast history of semiconductor material, as well as irradiationmodes. Microwave treatment can result in improvement of thematerial parameters (homogenization, defect gettering) and


36/192

36

their degradation as well. To give the ultimate answer con-cerning mechanisms for structural relaxation in semiconduc-tors exposed to microwave treatments, further investigations

are needed.

1.3. INTERACTIONS BETWEEN PHASESIN Me GaAs (InP) CONTACTS STIMULATEDBY MICROWAVE IRRADIATION

The structural-chemical processes that occur at boundariesbetween phases in contacts results in changes of the electro-

physical characteristics of contacts and affect their operatingstability at abrupt changes of the operating modes [21,22].Therefore the required structure of boundaries betweenphases, as well as possibility for its realization over large ar-eas, should be provided during device manufacturing.

In what follows the principal physico-chemical processesare considered that affect the characteristics of the transitionlayers in contact exposed to microwave treatment, as well asdependence of these processes on the nature of deposited ma-terial and deposition modes.

1.3.1. Effect of magnetron irradiationon the physico-chemical processesin Me GaAs (InP) contacts

Shown in Figs.1.9, 1.10 are the Auger concentration depth

profiles for the components of Al n -n + -GaAs and Al n -n + -InPcontacts taken both before and after microwave treatments.These profiles represent the character of the structural-chemical transformations that occur at boundaries betweenphases in the structures studied.


37/192

37

50

50

100

c , at. %

As

As

Ga

Ga

O

O O

Al

Al

(a)

(b)

0 1000 5000 10000 t, s

Fig.1.9. Auger concentration depth profiles for the components of Al n-n+-GaAs heteropair before (a) and after (b) magnetron irradiation (ir-radiance of 100 W/cm 2, exposure time t M = 1 min); t is duration ofion etching [20].


38/192

38

The profiles taken for the Al n -n + -GaAs contacts aremonotone, contrary to those measured for the Al n -n + -InPcontacts. Taking into account the character of non-monotony

for the concentration depth profiles of Al, Ga and As in the Al n -n + -GaAs contacts, one can assume that at the Al GaAsinterface exposed to microwave treatment the following sub-stitution reaction seems most likely to occur:

y Al + GaAs ( y x )Al + Al x Ga 1-x As + x Ga (1.1)

even though rather high ( > 725 K) temperature is needed forits intense proceeding. Besides, microwave treatment results information of an oxide layer on the Al surface (see Fig.1.9)[20].

Contrary to the above, one can conclude from Fig.1.10that in the case of the Al n -n + -InP contact no chemical reac-tions stimulated by microwave irradiation occur, and the tran-sition layer width somewhat decreases. Thus the effect of mi-crowave irradiation on the studied heteropairs Al n -n + -GaAs

and Al n -n +

-InP is realized through different mechanisms forinteractions between phases. In the Al n -n + -GaAs contact ex-change reactions seem to be predominant, while in the Al n -n + -InP heterostructure interdiffusion of the contact pair com-ponents occurs. Such character of processes and correspond-ing formation of the end products should lead to distinctionsin the effect of microwave irradiation on the parameters of

Al n -n + -GaAs and Al n -n + -InP Schottky barriers (SBs).Now let us consider what mechanism for interaction be-

tween phases is realized in the Au Cr GaAs contacts withchemically active boundaries between phases. In [23] thermo-chemical calculations have been made for the Cr n -n + -GaAssystem, and it was shown that Cr As compounds of differentstoichiometric ratios could be formed.


39/192

39

20

20

40

40

0

0i

0.02

0.02

0.04

0.04

0.06

0.06

0.4

0.4

0.8

0.8

1.2

1.2

1.6

1.6

60

60

80

80

100

100

P

P

In

In

Al(o)

Al(o)

O

O

C

C

Al (m)

Al (m)

(a)

(b)

d/d

c , at. %

Fig.1.10. Auger concentration depth profiles for the components of Al n-n+-InP contact before (a) and after (b) microwave irradiation ( d i is theinitial metal layer thickness).


40/192

40

A possibility for formation of compounds by Cr and Gahas not been considered. The results of study of the Cr, Ga and

As concentration depth profiles in the Au Cr GaAs contacts

exposed to microwave treatment (see Fig.1.11) indicate at highprobability of formation of compounds involving chromium, as

well as release of metallic gallium, even though such exchangeinteraction is not very intense [24]. One should also note thatconcentration of chromium atoms in the semiconductor near-surface layers is increased as a result of microwave irradiation.

As

As

Ga

Ga

O (x10)

Au

Au

(a)

( )b

t, s

Cr

Cr

20

20

0 500 1000 1500 2000 2500 3000 3500

60

60

100

100

c , at. %

O (x10)

Fig.1.11. Auger concentration depth profiles for the components ofAuCr n-n+-GaAs contact and unintentional O and C impurities be-fore (a) and after (b) microwave treatment [24].


41/192

41

Titanium is close to chromium in its chemical activity.The pronounced chemical reactions between Ti and GaAs areobserved at thermal annealing at a temperature of about

500 C. Specificity of these interactions under microwavetreatment is demonstrated by Fig.1.12. Along the concentra-tion depth profiles of the Au Ti n -n + -GaAs contact compo-nents, the changes in binding energies E b of core electrons(Ga 3 d , As 3 d , O 2 s and Ti 2 p ) are also presented there. Thesedata enable one to judge how the phase composition changesacross the transition region width [25].

From the above data it follows that chemical composi-tions of the transition layers after microwave irradiation re-main the same as those for the initial structures. The mostpronounced features of the transition layer induced by micro-

wave irradiations are (i) Ga 3 d -electron binding energy shifttowards higher values (as compared with its value for GaAs)and (ii) more clear structure of the Ti concentration depthprofile. In addition, the main distinction in the component dis-

tributions for irradiated and non-irradiated structures is pene-tration of gold atoms to the Ti n -n + -GaAs interface. Theseresults enable us to assume that change of the chemical com-position in the transition region occurs mainly through inten-sification of the oxidation processes (as in [26]). Judging fromthe character of binding energy change for Ga 3 d -electrons,their catalyst is AuGa alloy that takes part in oxygen atomsactivation.

Thus the chemical processes occurring between barrier-forming metal and semiconductor can be substantiallychanged when atoms of another type are present in the metal-lization composition. The role of microwave treatment in thiscase lies mainly in the enhancement of diffusion penetrationof the above atoms to the boundaries between phases.


42/192

42

E , eV E , eV() (b)b b

19.2 19.2

24.2 24.2

458 458

455 455454 454

23.8 23.8

19.4 19.4

19.0 19.0

41.3 41.3

41.2 41.2

41.1 41.1

Ga 3d Ga 3d

As 3d As 3d

O 2s O 2s

Ti 2 p Ti 2 p

Ti

23.4 23.4

+ + +++ ++ ++ ++ ++ ++ ++ ++ ++ ++ +

20 20

0 0 0 0 40 40

40 40

60 60

80 80

80 80 120 120160 160t, min t, min

+ +

TiO2

GaAs

GaAs

Ga O(20.3 eV)

2 3

Ga(18.7 eV)

c , at. % c , at. %

+++

Fig.1.12. Auger concentration depth profiles (bottom) and electron bindingenergies E b for the components of Au Ti n-n+-GaAs contact and un-intentional O impurity before (a) and after (b) microwave treatment; t

is duration of ion etching [25].

In conclusion let us dwell on some features of the proc-esses enhanced by microwave radiation that occur in contactsformed by amorphous alloys and GaAs. Tungsten silicide canserve as example of such alloy. The metallization using W x Si1-x


43/192

43

attracts attention because it provides contact tolerance to hightemperatures.

Shown in Fig.1.13 are the results of layer-by-layer Auger

analysis of the W x Si1-x n -n +

-GaAs contact made before andafter microwave treatment.

80

80

70

70

60

60

50

50

40

40

30

30

20

20

10

10

200

0

00 400 600 800 1000 1200

W

W

O

O

Si

Si

C

C

Ga

Ga

As

As

()

(b)

t, s

c , at. %

Fig.1.13. Auger concentration depth profiles for the components of W xSi1- x n-n+-GaAs contact and unintentional O impurity before (a) and after (b)microwave treatment.


44/192

44

The results obtained enable one to conclude that mi-crowave irradiation practically does not change distribution ofthe above contact pair components. It seems that no reactions

occur between W x Si1-x and GaAs. In other words, the consid-ered structure is stable in the metallurgical sense. Some smallredistribution of the components due to microwave treatmentshould be related to structural changes in the metallizationlayer, as well as relaxation of intrinsic stresses (they can reachsuch values in a film that are enough for its cracking).

1.3.2. Structural-chemical processes in Me GaAs contactsstimulated by gyrotron irradiation

Let us discuss the structural-chemical modification of theMe GaAs interfaces (Me stands for metal) at increased G-irradiation power. Shown in Figs.1.14, 1.15 are the results ofour investigations of the effect of microwave irradiation on thecontacts of two types (one- and two-layer metallization) butinvolving the same barrier-forming metal [20,27]. One can see

that microwave action leads to changes in chemical composi-tions of the contact transition layers. These changes are dif-ferent for contacts of different types.

For contacts with one-component metallization microwave treatment enhances diffusion intermixing of phases. As aresult, the gallium content in metal layer grows abruptly (upto 20%) [27]. In the multilayer Au Pt Cr Pt n -n + -GaAs struc-ture the metallurgical processes resulting from microwave ir-radiation are more pronounced. They may be represented bythe following equation:

Pt + GaAs PtAs 2 + (PtGaAs), (1.2)

where (PtGaAs) is an alloy.


45/192

45

t, s

50

0

As

Ga

Pt

()100

50

600

(b)

100

70 80 90

As

Ga

Pt

c , at. %

Fig.1.14. Auger concentration depth profiles for the components of Pt n-n+-GaAs contact before (a) and after (b) gyrotron treatment [27].

Many authors have comprehensively studied chemicalprocesses and kinetics of reactions at the Pt n -n + -GaAs inter-face. (The appropriate works are reviewed in [28].) Differentmodes of thermal annealing have been used as active actions


46/192

46

on the Pt n -n + -GaAs contact. It seems to be of interest tocompare how two different external actions (thermal annealingand microwave irradiation) affect the contact characteristics.

50

0

()100

30

Au PtCr

Pt GaAs

50

0

(b)100

30

Au Pt Cr Pt GaAs

t, s

c , at. %

Fig.1.15. Auger concentration depth profiles for the components ofAuPtCr Ptn-n+-GaAs contact before (a) and after (b) gyrotrontreatment [20].


47/192

47

1. It follows from the studies of temperature depend-ence of the reaction kinetics in the Pt n -n + -GaAs structurethat the activation energy of the reaction between platinum

and gallium arsenide is rather high ( 2.3 eV) [23]. This isabout 1.5 2 times bigger than the values that are usuallycharacteristic of atom diffusion along the grain boundaries.In other words, for chemical interaction between Pt andGaAs to occur, rather high temperatures are needed that ex-ceed, by a factor of 1.5 2, those realized in the above ex-periments (150 200 C).

2. As in the case of thermal annealing at rather moder-ate (about 400 C) temperatures and thick ( 200 nm) Pt films,the reaction does not reach its final stage.

3. Contrary to the case of thermal annealing, no predo-minant Ga migration from GaAs and its uniform distributionover the Pt film thickness occur.

4. Microwave treatment does not lead to formation of alayered contact structure with clearly separated regions that

involve products of Pt reaction with Ga and As (as it takesplace at thermal annealing).

All the above facts evidence that microwave anneal-ing is not adequate to thermal one. This conclusion hasbeen also confirmed by the results of studies performed forthe W n -n + -GaAs structure (see Fig.1.16). One should notethat distribution of its components is qualitatively similar tothat in the case of the Pt n -n + -GaAs structure discussedabove, although the thermal characteristics that describechemical activity toward GaAs are essentially different for Wand Pt. Nevertheless, non-monotonic behavior of the contactpair components distribution indicates that intense chemicalprocesses occur at boundaries between phases in the irradi-ated W n -n + -GaAs contacts. It was shown in [29,30] that in-


48/192

48

terface in this contact is tolerant to high-temperature actions.Chemical reactions between W and GaAs (leading to produc-tion of the W 2 As 3 phase) are observed only at annealing

temperatures over 700 C.

80

100

60

40

20

0

W

OC

Ga As

()

t, s

80

100

60

40

20

0 0 5 10

W

OC

Ga

As (b)

15 20 25 30

c , at. %

Fig.1.16. Auger concentration depth profiles for the components of W n-n+-GaAs contact and unintentional O and C impurities before (a) andafter (b) microwave treatment.


49/192

49

Degradation processes in W n -n + -GaAs contacts at lowertemperatures were related to intensification of mass transportbetween phases. But the concentration depth profiles obtained

by us for the W n -n +

-GaAs contacts after microwave irradia-tion cannot be explained by the diffusion mechanism only, al-though the temperatures realized during our experiments havebeen much below those required for chemical reactions to oc-cur. So the results obtained for the W n -n + -GaAs structureevidence that, when analyzing the behavior of contacts exposedto high-power G-irradiation, one has also to take into accountsome non-thermal factors that affect the structural-chemicalmodification of the boundaries between phases.

Thus, depending on the irradiation mode, either kineticmechanism for metal semiconductor interface formation orchemical diffusion may occur. A discrete and localized (intime and space) character of heat release during irradiationdetermines the characteristics of atomic intermixing betweenphases. This process is determined by the balance between

the thermal excitation energy for the local regions of the sub-strate, rate of interdiffusion at the interface and rate of heatremoval to the substrate.

A radical difference between microwave annealing andthermal one is related to the peculiarities of temperature dis-tribution in the heterocontacts. It seems that for microwaveannealing an internal local overheating occurs at smalldeviations of the average temperature in the heterocontactfrom room one. Such overheating of local regions serves tointensify chemical reactions. Another factor that affects bothstructure and phase composition of the boundaries betweenphases is related to the action of the electric field of electro-magnetic wave. This electric field may reach values over thebreakdown one.


50/192

50

1.3.3. Effect of microwave treatment on the electrophysicalparameters of surface-barrier structures

Microwave irradiation leads to structural-chemical modification

of the boundaries between phases in metal semiconductor sys-tems, as well as changes in the characteristics of local centersin such systems. These processes are to be accompanied bychanges in the electrophysical parameters of SB.

In [22] it was shown that ion diffusion in a crystal latticecould be enhanced by the crystal electron subsystem. Notonly phonons take part in ion hopping from one position toanother but electrons as well. A deficiency of energy requiredfor a diffusing particle to execute hopping is supplied fromthe electron subsystem. In this case the diffusing particlechanges its charge state. Thus, on the one hand, the heightW B of the potential barrier that the diffusing particle has toovercome is changed; on the other hand, the charged particlegets an additional portion of energy from the electron sub-system. This portion is equal to the energy of the local level

resulting from the particle transfer from a lattice site to an-other state of equilibrium.

According to conservation of energy, it is electrons ofhigh ( > kT ), and not medium, energies that actively influ-ence hopping [31]. Therefore one can substantially intensifyion hopping (in other words, the diffusion coefficient) by in-creasing the number of high-energy electrons through excita-tion of the electron subsystem (in our case by using micro-

wave irradiation). Action of microwave radiation on a solid can lead to

crystal lattice heating with Joule heat. This disturbs the ori-ented ion motion. The characteristic time of lattice heating ismuch greater than that of electron gas heating. One can prac-tically exclude the crystal lattice heating by applying micro-


51/192

51

wave power in the pulse, and not continuous, mode. In thiscase the electron gas manages to heat up, and the crystal lat-tice does not. The characteristic time of the electron gas heat-

ing in GaAs is about 10-9

s, while it takes several minutes foratom transfer through the near-contact layers due to Jouleheating.

From [31,32] it follows that the ion diffusion coefficientD is given by the following expression:

( ) ( )[ ].2//1/exp 000 kT T T h kT W D D B += a (1.3)

Here D 0 = d 02 ( is the atomic vibration frequency and d 0 is

the crystal lattice constant); T is the temperature of electrongas (it differs from that of crystal lattice, T 0); a is a parameterthat characterizes interaction between the diffusing atom (ion)and crystal lattice. The typical values to be used in furthercalculations could be chosen as follows: D 0 10

-20 cm 2/s; (10 1310 15) s -1; W B 2 eV; 0.5 eV; ah 5 kT 0; T 0 300 K.

Strictly speaking, the concept of electron temperature

can be introduced only for the optical phonons at low tem-peratures or for the electron-electron interaction (the latter isof no significance at energy values that are of interest forus). To obtain D in other cases, one should average the tran-sition probability not over the Maxwell distribution with atemperature T T 0 but over the distribution obtained fromthe exact solution of the kinetic equation. However, takinginto account that our further considerations are approximate,

we shall use the effective electron temperature (as it hasbeen done in [33]).

If the electric vector E of the microwave field is normalto the static built-in electric field E 1 of the bulk charge, thenone can neglect heating action of that static field and writedown the effective electron temperature as


52/192

52

( ) ( )20200 3/1 kT El q T T += . (1.4)

Here l is the electron mean free path; is the fraction of en-

ergy loss at an inelastic collision. At l = 10-6

cm, 1/25 1/30and E = 10 4 V/cm the additional term ( ) ( )2020 3/ kT El q isequal to unity. This means that the effective temperature in-creases two-fold and, as a result, the diffusion coefficient in-creases by a factor of 10 10. (Of course, this value is overesti-mated because the energy that electron has got from the elec-tric field is spent not only for particle hopping.)

In [34] the change in the SCR width w under microwave irradiation has been calculated. The calculation wasbased on solving a set of equations that involved equations forflows I i of diffusing ions (whose type was labeled by i ), conti-nuity of these flows and the Poissons equation:

x N

EN I i i i

+= , (1.5)

0=

+

x

I

t

N i i , (1.6)

D N x

=

12

2

. (1.7)

This set of equations is written down for dimensionless quantities.The units of measurement of the flows I i are D i i L N D / where N i is the concentration of ions of i -th type (measured in units N D ,

i.e., the donor concentration).( )

2/12

0004/

D D N q kT L = is the

Debye shielding length (it also serves as unit of measurementof dimensionless distance x ) and 0 is the semiconductor per-mittivity. The potential is measured in kT / q , while the unitfor measurement of the electric field E is ( ) 0// =x D dx d L kT .

The unit for time measurement is 2/ D L D (it was noted above


53/192

53

that the diffusion coefficient D depends on the microwaveelectric field).

Strict approach to solving the above set of equations is

very complicated. So let us use a simplifying assumption thatthe diffusion front practically is not blurred. This enables us todetermine a change L 1,2 = L 1 + L 2 of the depletion layer

width (here L 1 ( L 2) is the change due to acceptor (donor) iondiffusion):

( )[ ] ( ) 2/112/112/112/112,1 erf /erf 22 t E t E E E L += , (1.8)

where erf ( x ) is the error function. (If the depletion layer width decreases due to ion diffusion, then one should take theright-hand side of expression (1.8) with the minus sign.) Itshould be stressed that the solution (1.8) is approximate. It

was obtained using the assumption that change in concentra-tion due to diffusion and corresponding penetration of atomsinto the depletion layer does not exceed the impurity concen-tration in that layer.

The experimental results presented in [34] correspondedto the case of microwave irradiation with angular frequency = 1 GHz and irradiance of 100 W/cm 2. The metallizationlayer thickness was less than skin depth, so one could assumethat almost all the above power was absorbed in the contactstructure and semiconductor. At the above irradiance valuethe microwave electric field is 10 4 V/cm. In this case the L 1,2

value is of the same order as the Debye shielding length L D

(or even may exceed it) if all microwave power is spent onchanging L 1,2 . At the above experimental conditions the transi-tion region width changed substantially.

At microwave irradiation of higher (about 10 2 GHz) fre-quency the time needed to heat up the electron gas becomescomparable to the period of the microwave electromagnetic


54/192

54

field. In addition, the skin depth goes down with electromag-netic field frequency. This results in decreasing the irradiationpower received by the contact.

After the microwave field is switched off, the diffusioncoefficient abruptly drops. Therefore the spatial distribution ofdiffusing particles that has resulted from the action of bothmicrowave irradiation and applied static electric field remainsfrozen for a long period.

Thus the microwave irradiation enhances diffusion ofpoint defects and impurities in the near-contact layers throughexcitation of the electron subsystem in a semiconductor. Thiseffect is to be more pronounced at application of pulsed mi-crowave irradiation, since in this case some side effects (say,

Joule heating of the crystal lattice) are excluded. The fre-quency of microwave radiation has to be of the order of in-

verse time of the electron gas heating, i.e., about 10 9 s -1. Soone should expect some distinctions between changes in theelectrophysical parameters for contacts exposed to M- and G-

irradiations.Shown in Fig.1.17a are typical I V curves of a surface-

barrier diode [19]. One can see that after microwave irradia-tion for 2 s the range of exponential portion of the forwardbranch of I V curve increased by an order of magnitude, thebarrier height B remained the same and the ideality factor n dropped. As to the backward branch of I V curve, the reversecurrent I R significantly decreased. This seems to result fromthe structural defect annihilation that increases the chargecarrier effective lifetime. This conclusion is supported by thefact that the dopant concentration did not change under mi-crowave irradiation this follows from the reverse capaci-tance squared vs voltage curves (1/ C 2 (V )) presented inFig.1.17b.


55/192

55

10-12

10-10

10

10

10

10

10

-8

-6

-4

-2

0

1

2

I, A

0 0.1 0.2 0.3 0.4 0.5 V , V

()

11019

2 -2

1

2

1/C , F

-1.0 0 0.5 1.51.0 2.0 V , V

(b)

-0.5 2.5 3.00

21019

31019

41019

51019

R

Fig.1.17. I V curves (a) and function 1/ C 2 = f (V R ) (b) for Schottky TiN n-n+-GaAs contact before (1) and after (2) microwave treatment [19].


56/192

56

The results of our studies of the effect of microwave ir-radiation on SBs formed by different metals with some III Vcompounds are summarized in Tables 1.2 and 1.3. An analy-

sis of these results enables one to draw the following conclu-sions:

a) Whatever the metallization type (the only exception beingPt n -n + -GaAs structures), short-term M-irradiations im-prove the electrophysical parameters of the contacts involv-ing GaAs;

b) Microwave irradiation affects the barrier height B slightlyand the ideality factor n somewhat stronger; the correspond-ing changes are bigger for those structures where the initialn value is higher. The reverse current I R changes most con-siderably under microwave irradiation. The semiconductorsused by us had low donor concentrations that did notchange under irradiation. So one can assume that in this

case the reverse current decrease is related to weakeningof the electron-hole pair generation in the depletion region.This means that microwave irradiation enhances partial an-nihilation of the initial structural defects in the near-contact layers and thus increases the charge carrier effec-tive lifetime;

c) Changes of the SB parameters under microwave irradiationare similar for structures based on GaP and GaAs. But forstructures based on InP both parameters (the barrier height B and ideality factor n ) decreased after irradiation.


57/192

57

Table 1.2. Changes under magnetron irradiation (frequency f = 2.45 GHz, power P = 5 kW, duration t M = 1s) in the electrophysical parame-ters of Schottky diodes (barrier height B, ideality factor n and

hole diffusion length L p) [20].

Contacts Type oftreatment

B, eV n L p , m

Al GaAs IMI

0.55 0.580.57 0.58

1.68 2.201.30 1.40

1.6 1.92.0

Pt GaAs IMI

0.88 0.950.88 0.89

1.12 1.371.18 1.24

2.1 2.22.1 2.2

Au Pt GaAsI

MI0.88 0.920.90 0.92

1.17 1.301.12 1.15

2.0 2.22.3 2.5

Cr GaAs IMI

0.73 0.750.76 0.77

1.17 1.241.08 1.09

6.5 7.07.2 7.4

Au Cr GaAs

IMI

0.70 0.760.75

1.12 1.171.04 1.08

0.9 1.11.0 1.3

Mo GaAs IMI

0.68 0.690.68 0.69

1.16 1.231.09 1.14

2.3 2.82.5 2.7

W GaAs IMI

0.65 0.660.69 0.70

1.20 1.401.09 1.12

1.7 2.02.1 2.2

Mo x Si1-x GaAs

IMI

0.70 0.710.72

1.15 1.171.09 1.10

2.4 2.52.6 2.7

W x Si1-x GaAs

IMI

0.72 0.730.74

1.15 1.171.07 1.09

2.2 2.42.5

TiN GaAs IMI

0.750.76

1.241.02

1.6 1.751.8 1.82

Cr InP IMI

0.67 0.690.63 0.65

1.5 1.81.2 1.4

1.52 1.61.7 1.75

Al InP IMI

0.77 0.800.75

1.73 2.731.14 1.73

1.55 1.71.73 1.8

Au GaP IMI

1.521.70

1.751.10

0.650.78

Cr GaP IMI

1.631.85

1.721.30

0.530.87


58/192

58

Table 1.3. Effect of gyrotron irradiation on the Schottky barrier parameters (barrier heightand hole lifetime ! ) [20].

Parameters and their spreadsBefore irradiation After irradia

ContactsB ,

eV ,eV

n n , p ,

10 -10 sB ,

eV ,eV

n

Pt GaAs 0.91 0.06 1.2 0.2 25.2 0.88 0.02 1.21 0

W GaAs 0.65 0.01 1.3 0.2 28.9 40 0.69 0.01 1.26

Al GaAs 0.56 0.03 1.9 0.4 19.4 0.59 - 2.5

Au Pt GaAs 0.92 0.07 1.2 0.13 23.4 0.86 0.08 1.5

Cr Pt GaAs 0.91 0.06 1.2 0.11 22.7 0.91 0.06 1.2 0

Au Ti GaAs 0.7 0.1 1.3 0.2 19.8 0.7 0.1 1.3

TiN Ti GaAs 0.72 0.09 1.3 0.14 19.8 0.73 0.08 1.3 0

5 8


59/192

59

Taking into account the results of investigations pre-sented in sections 1.2 and 1.3, one can assume that the prin-cipal factor responsible for the observed changes of SB pa-

rameters is structural ordering in the heterosystems exposedto microwave treatments. This ordering improves semicon-ductor material and makes the disordered heterogeneouscontact boundaries more uniform. The above conclusion isalso supported by the results on the effect of high-power G-irradiation on the Au Pt Cr Pt n -n + -GaAs multilayer con-tact (Fig.1.18).

16

14

12

10

8

6

4

2

0

16

14

12

10

8

6

4

2

0

() (b)

! n

N u m b e r o f d i o d e s i n a r r a y

Fig.1.18. Barrier height B (a) and ideality factor n (b) histograms forAuPtCr Ptn-n+-GaAs contact before and after (hatched area) mi-crowave annealing [35].


60/192

60

One can see that a clear tendency exists toward de-crease of parameter spread after microwave treatment for SBmade on the same wafer [35]. In Table 3 the results for other

contact pairs are given. From them one can conclude that G-irradiation either does not change the contact parameters oreven slightly impairs them, with some decrease in spread oftheir values.

Below we summarize the most probable mechanisms thatare responsible for the observed changes under microwaveirradiation in properties of both semiconductor materials andcontact structures based on them [15,19,20,24,25,27,34-45]:

Thermal mechanism that is related to heating of boun-daries between phases by absorbed microwave energy. Ouranalysis of the Auger concentration depth profiles for the con-tact structure components taken before and after microwaveirradiation, as well as their comparison with the results of thelayer-by-layer analysis of contacts after thermal annealing,evidence that this factor is insignificant.

Electrostatic mechanism that is related to an actualdrop in barrier voltage. It was shown in [2] that this mecha-nism can substantially affect spatial redistribution of the con-tact components through diffusion, even if there are no criti-cal electric fields that determine mechanisms for avalancheand tunnel breakdowns. An intense interdiffusion between ametal and gallium arsenide begins when the absorbed energylevel is as high as about 2/3 the critical value needed foravalanche breakdown to occur. However, according to esti-mations made in [2], this factor counts little in the used irra-diation mode.

Electrodynamic mechanism that is related to a departureof the electron subsystem of a semiconductor from the state ofthermodynamic equilibrium. This departure occurs due to ap-


61/192

61

pearance of hot charge carriers in the semiconductor near-surface layer that changes the impurity-defect composition ofthis layer [2]. The electrophysical parameters of the surface-

barrier structures exposed to microwave treatment are close tothose after 60Co -irradiation [46] (and it is known that in thelatter case the above mechanism is the governing factor). Thisfact indicates that it is the electrodynamic mechanism thatplays a decisive role in the structural-impurity transformationunder microwave actions.

Occurrence of non-steady-state stress gradients due to apractically instantaneous heating of disordered regions thatappear in a semiconductor during the contact structure forma-tion. In this case the concentration depth profiles for the con-tact pair components can be practically the same before andafter microwave irradiation. However, taking into account thepossibility for collective interactions in the stress fields, onecan substantially decrease the barrier for defect annihilationor defect complex transformation [47].

1.4. EFFECT OF MICROWAVE RADIATIONON THE PARAMETERS OF RESONANT TUNNELING DIODES

1.4.1. GaAs tunneling diode with two -layers in SCR In recent years many authors have been studying the tunneldiodes exposed to microwave actions. The emphasis wasmainly on a search for ways to exert control over their I V

characteristics. In this connection the works [48-59] should berecalled, to mention just a few. Along with technological prob-lems, their authors considered some physical aspects concern-ing interaction between microwave radiation and degeneratesemiconductors, as well as degradation mechanisms in the ir-radiated tunnel diodes. However, similar effects in more com-


62/192

62

plicated resonant-tunneling structures exposed to microwaveirradiation practically have not been studied yet.

An attempt to fill in the gap in this area has been made

in [60-63]. In the GaAs tunnel diodes studied the p -n junction width was l j = 10 nm, spacing between two -layers d 5 nm,impurity concentration in the -layers was N 10 13 cm -2, thosein the p + - and n + -regions were 10 20 and 10 19 cm -3, respectively.The Fermi energy in n + -GaAs was E F 0.023 eV and the elec-tric field in the p -n junction was E j = 2 10

6 V/cm. We investigated I V curves for the GaAs resonant

tunneling diodes (RTD) with two

-layers in SCR, both beforeand after microwave irradiation in the magnetron waveguide.The frequency of M-radiation was f = 2.5 GHz and the irradi-ance was W = 3 30 W/cm 2. A strong anisotropy of the mi-crowave field action has been observed. When the angle ofincidence of the electromagnetic wave was 90 (glazing inci-dence), then I V curves remained practically the same at timeof exposure t M 300 s, while at normal incidence the changes

occurred after t M = 10 s.The starting doses of microwave irradiation ( t M = 1020s,

W = 3 W/cm 2) have led to a shift (toward higher voltages) ofthe I V curve peak related to the resonant-tunneling currentflow via a level in the potential well formed by the -layers; be-sides, the excess current I ex has somewhat dropped (Fig.1.19).The qualitative explanation fo

Konakova Book-microondas Cuanticas

Documents

Transcript of Konakova Book-microondas Cuanticas