您好,欢迎来到爱go旅游网。
搜索
您的当前位置:首页半导体材料-欧姆接触

半导体材料-欧姆接触

来源:爱go旅游网


ISSN 1063-7826, Semiconductors, 2007, Vol. 41, No. 11, pp. 1263–1292. © Pleiades Publishing, Ltd., 2007.

Original Russian Text © T.V. Blank, Yu.A. Gol’dberg, 2007, published in Fizika i Tekhnika Poluprovodnikov, 2007, Vol. 41, No. 11, pp. 1281–1308.

REVIEWMechanisms of Current Flow

in Metal–Semiconductor Ohmic Contacts

T. V. Blank^ and Yu. A. Gol’dberg

Ioffe Physicotechnical Institute, Russian Academy of Sciences, Politekhnicheskaya ul. 26, St. Petersburg, 194021 Russia

^e-mail: tblank@mail.ioffe.ru

Submitted January 29, 2007; accepted for publication April 2, 2007

Abstract—Published data on the properties of metal–semiconductor ohmic contacts and mechanisms of cur-rent flow in these contacts (thermionic emission, field emission, thermal–field emission, and also current flowthrough metal shunts) are reviewed. Theoretical dependences of the resistance of an ohmic contact on temper-ature and the charge-carrier concentration in a semiconductor were compared with experimental data on ohmiccontacts to II–VI semiconductors (ZnSe, ZnO), III–V semiconductors (GaN, AlN, InN, GaAs, GaP, InP), Group IVsemiconductors (SiC, diamond), and alloys of these semiconductors. In ohmic contacts based on lightly dopedsemiconductors, the main mechanism of current flow is thermionic emission with the metal–semiconductorpotential barrier height equal to 0.1–0.2 eV. In ohmic contacts based on heavily doped semiconductors, the cur-rent flow is effected owing to the field emission, while the metal–semiconductor potential barrier height is equalto 0.3–0.5 eV. In alloyed In contacts to GaP and GaN, a mechanism of current flow that is not characteristic ofSchottky diodes (current flow through metal shunts formed by deposition of metal atoms onto dislocations orother imperfections in semiconductors) is observed.PACS numbers: 73.30.+y, 73.40.Cg, 81.40.EfDOI: 10.1134/S1063782607110012

1. A BRIEF HISTORY OF THE SUBJECTAn ohmic contact is a metal–semiconductor contactin which a potential barrier at the interface does notmanifest itself; this contact is an integral part of anysemiconductor device. Studies of ohmic contacts werestarted ~60 years ago when it was noticed that a poten-tial barrier exists at the Ni–CdS interface while there isno such barrier at the Al–CdS interface; Schottky [1]assumed that the barrier is not formed if the work func-tion Φm for electrons leaving the metal is smaller thanthe electron-affinity energy for the semiconductor Xs. In1940s–1950s, when the III–V semiconductors weredeveloped, Bardeen [2] noted that the presence of a bar-rier is often caused by the density and energy distribu-tion of surface states in a semiconductor rather than bythe work function of electrons leaving the metal; in theopinion of Spicer [3], the aforementioned states areformed due to the presence of extraneous atoms (forexample, oxygen atoms) at the semiconductor surface.The early stage of studies of ohmic contacts was con-sidered in review [4].

Subsequent studies were carried out in three direc-tions.

First, technological studies were conducted with theaim of decreasing the resistance of ohmic contacts to anextent that they do not manifest themselves in the char-acteristics of semiconductor devices. The ohmic con-tacts’ resistance per unit area was as high as Rc = 10–6–10–8 Ω cm2. This was accomplished either by variationin the chemical composition of the near-contact semi-conductor’s region or by additional doping.

Second, the composition of the phases formed at themetal–semiconductor interface was studied by theX-ray spectral analysis, chemical analysis, Auger spec-troscopy, and tunneling microscopy. On the basis ofthese studies, the optimal chemical composition of thesemiconductor’s near-contact region was defined; thiscomposition was supposed to ensure the lowest possi-ble resistance of the contact.

Third, the methods for determining the resistance ofan ohmic contact were developed; this resistance oftenamounts to a small fraction of the total resistance of astructure. The resistance of ohmic contacts was deter-mined from: (i) the dependence of potential differencebetween several contacts on the intercontact distance [5];(ii) the dependence of the resistance of the metal–semi-conductor-metal structure with two ohmic contacts onthe structure’s thickness [6]; (iii) an analysis of charac-teristics of contacts with various diameters [7]; and(iv) the transmission-line method [8].

This stage of the development of ohmic contactswas represented in general reviews [6, 9–13] and also inreviews concerned with specific semiconductors: ZnO[14]; ZnS [15]; GaAs, GaN, and ZnSe [16]; SiC [17];and C (diamond) [18].

Studies of the dependences of the ohmic contacts’resistance on temperature, the charge-carrier concen-tration, the band gap of semiconductor, and other fac-

1263

12

BLANK, GOL’DBERG

(a)

n-GaN

MgEvac

EvacΦm = 3.61 eVVXs = 4.1 eVe 1.0χEχcEg = 3.23 eVEv

EAu(b)vacn-GaPEvac

Φm = 5.15 eVXS = 3.8 eVqϕbEcχχEg = 2.26 eVEv

Fig. 1.semiconductors: (a) the semiconductor does not have sur- Energy diagrams of ohmic contacts to the n-typeface states and Φm < Xs (by the example of Mg–n-GaN);(b) the semiconductor contains high-density surface statesin the band gap (by the example of Au–the near-surface region is heavily doped. En-GaP); however,level and χ is the Fermi level.

vac is the vacuumtors were started in mid-1990s with the aim of estab-lishing the mechanism of current flow through anohmic contact. Both the current flow mechanisms char-acteristic of the Schottky barriers (thermionic emission,thermal field emission, and field emission) and othermechanisms (recombination, metal shunts) were con-sidered theoretically. The current flow mechanism wasdetermined by comparison of experimental data withtheoretical results. These studies were carried out forII–VI, III–Vbased on these semiconductors. This review is dedi-, and IV–IV semiconductors and for alloyscated to a systematization of the results of these studies.2. FORMATION OF AN OHMIC CONTACTcan be either rectyifying (barrier) if the potential barrierAs is well known, a metal–semiconductor contactbetween a metal and a semiconductor is tunneling-non-transparent or ohmic if the barrier is absent or is tunnel-ing-transparent for electrons.

(a)Vq(b)Vq(c)VqFig. 2.ductor ohmic contact: (a) thermionic emission of electrons Mechanisms of the current flow in a metal–semicon-above the barrier; (b) thermal field emission of electronsthrough the barrier top; and (c) tunneling (field emission) ofelectrons through the barrier.

An ohmic contact is typically formed in caseswhere:

(I) there is no potential barrier between a metal anda semiconductor, for example, if a metal with the workfunction for electrons smaller than the electron affinityin the semiconductor is chosen for an n-type semicon-ductor with a low density of surface states in the bandgap (Fig. 1a);

(II) there is a potential barrier but this barrier is nar-row (tunneling-transparent) which is attained by heavydoping of the near-contact region of the semiconductor(Figs. 1b, 2c); in this case, electrons traverse the inter-face through the barrier over its entire height (the fieldemission); and

(III) there is a potential barrier but it is low; as aresult, this barrier is easily surpassed by charge carriers.This situation is typically attained by varying the chem-ical composition of semiconductor near the contact, forexample, by formation of a narrow-gap near-contactlayer; in this case, electrons traverse the interface abovethe barrier (thermionic emission) (Fig. 2a).

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS1265

case electrons traverse the top of the barrier (Fig. 2b), isCombination of mechanisms II and III, in whichalso possible (this is the thermal field emission).

that the metal–semiconductor potential barrier isRecently, mechanisms have been seen which implypresent but:

shunts formed, for example, by deposition of metal(i) the space-charge region is shunted by metalatoms onto dislocations and other imperfections in thesemiconductor (this mechanism is characteristic ofalloyed ohmic contacts) and

near the contact, the lifetime of charge carriers is(ii) due to a large number of crystal-lattice defectsextremely short in this region, so that the ohmic contactis formed due to recombination of charge carriers (thismechanism rarely manifests itself).

3. RESISTANCE OF AN OHMIC CONTACTresistance per unit area. This resistance consists of:The main characteristic of an ohmic contact is itssemiconductor and

(I) the resistance of the near-contact region of thethrough the potential barrier.

(II) resistance related to the passage of electrons3.1. Resistance of the Near-Contact Region

tance of the heavily doped region combined with theThe resistance of the near-contact region is the resis-resistance of the tance of the heavily doped near-contact region is typi-n–n+ and p–p+ junctions. The resis-cally low in semiconductors with a high mobility ofcharge carriers. For example, the resistance of the~1 tion µm-thick n+n+-GaAs layer at an electron concentra-about 6 ≈ 1019 cm–3 and a mobility of ~103 cm2/(V s) isSiC and II–VI semiconductors, it is often the case that× 10–8 Ω cm2. At the same time, in the case ofheavy doping of the near-contact region cannot beattained and, in addition, the charge-carrier mobility isnot high (10–100 cm2tance of the near-contact region can be in the range of/(V s)). In this case, the resis-~(10–4–10–5) Ω cm2.

tions by looking at the example of the We consider the resistance of the n–n+ and p–p+ junc-This resistance is inversely proportional to the electronn+–n junction.concentration [19, 20] as

RLDn–n+≈-q---µ---------N-----c-N

----+-,(1)

nKNwhere LD is the Debye length in the n-type region, Nc isthe density of states in the semiconductor’s conductionband, µn is the electron mobility in the n-type region,N and n+ are the electron concentrations in the n- andn+extent the electron concentration at the Fermi level in-type regions, K is the coefficient that shows to whatthe n+-type region exceeds N+, and q is the elementary

SEMICONDUCTORS Vol. 41 No. 11 2007

Table 1. Calculated reduced resistance of the n–n+according to the model suggested in [21] junction

RN, cm–3

n–n+, Ω cm2GaAs

InPGaPGaN101610–61.8 × 10–610– × 10–6101710–71.8 × 10–710– × 10–71018

10–8

1.8 × 10–8

10–7

4 × 10–8

charge. In the case of an ohmic contact to GaAs, Rn–n+(1) makes the main contribution to the contact resis-tance at N < 5 × 1017 cm–3.

considered as a Schottky diode without a potential bar-It was shown in [21] that the n–n+ junction can berier and with a thermionic mechanism of current flow;the resistance of this junction is given by the formula

R⎛⎝-qA-----k--*----T-⎞⎠-ln-----[---1----+------exp---1

n–n+=------(--χ----/--kT-------)--]

,

(2)

where χ is the Fermi level energy, A* = 4πqm*k2120mប–3 =r A/(cm2 K2) is the effective Richardson constant,mcarriers r is the relative effective mass of the majority chargemr = m*/m0 (m0 is the mass of a free electron),and semiconductor we have k is the Boltzmann constant. If in a lightly dopedχ ӷ kT/q, then

RkNcn–n+=⎛⎝-qA-------*----TN-----⎞⎠

.(3)

the most widely used III–V semiconductors are listed inEstimates of the resistance of an n–n+ junction for

Table 1; the calculations were carried out on theassumption that the difference between the electronconcentrations in the at least two orders of magnitude.

n- and n+-type regions amounts to3.2. Resistance Related to the Transition Through

the Metal–Semiconductor Boundaryductor interface can occur:

Transition of electrons through the metal–semicon-(i) above the barrier (thermionic emission, Fig. 2a);Fig 2b); and

(ii) through the barrier top (thermal field emission,energy (tunneling, field emission, Fig. 2c).

(iii) through the barrier at the level of the Fermicific mechanism of the current flow manifests itself,In order to determine the conditions in which a spe-Padovani and Stratton [22] introduced the parameter that depends on the semiconductor type and the degree

E001266BLANK, GOL’DBERG

of its doping. In the case of an we have

n-type semiconductor,E2--ε-----N00=ប

-----d-m----*

-,

(4)

sε0where εs is the relative permittivity of the semiconduc-tor, εcentration of ionized donors in the semiconductor.0 is the free-space permittivity, and Nd is the con-by In the case of a semiconductor.

Np-type semiconductor, Nd is replaceda, the concentration of ionized acceptors in thecurrent flow are thermionic emission at high tempera-Calculations [10] show that the main mechanisms oftures (peratures (kT ӷ E00low temperatures (kT ≈), thermal field emission at medium tem- E00kT), and field (tunneling) emission at Ӷ E00).

theory of thermionic emission in the example of an 3.2.1. Thermionic emission. We now consider thetype semiconductor; all conclusions will also be validn-for ized donors is replaced with the concentration of ion-p-type semiconductors if the concentration of ion-ized acceptors and the density of states in the conduc-tion band is replaced by the density of states in thevalence band.

sity Dependence of the thermionic emission current den-as [10]

J on the voltage V and temperature T can be writtenJ=JqV

sexp⎛⎝-nkT--------–1⎞⎠

,

(5)

where

I2–q(ϕ–∆ϕs=JsS=A*ST--------------bkT

---------------b--)

-.(6)

Here, Is is the saturation current, n is the nonidealitycoefficient in the current-voltage (ϕI–V) characteristic,potential barrier height due to the mirror-image forcesb is the potential barrier height, ∆ϕb is a decrease in theand other factors, and resistance per unit area is in general represented as

S is the contact area. The contactRdV

dVc=-dJ------=-----(7)

s

dI-S

and, if V 0, is equal toRk

–q(ϕc=⎛⎝-qA-------*----T-⎞⎠exp--------------b

kT

--–-----∆ϕ--------b--)-;(8)

in the case of lowering of the barrier only by the mirror-image forces, the quantity ∆ϕb is given by

2πq2

N1/4

∆ϕb=--(---ε---------)

---d3-⎛V–kTd–V-----⎞

,(9)

sε0⎝q-⎠where Vference and d is the diffusion-related contact potential dif-V is the applied voltage.

governed by the thermionic emission, we observe theThus, if the current flow through an ohmic contact isfollowing features:

the potential barrier height (I) the contact resistance increases exponentially asϕb is increased;

is increased; the dependence (II) the contact resistance decreases as temperatureear on a semilogarithmic scale with the slope of thisRcT = f(1/T) should be lin-dependence proportional to the barrier height the cutoff proportional to A* at 1/T 0; andϕb andsemiconductor and decreases only slightly as the dop-(III) the contact resistance depends on the type ofing level is increased (∆ϕb ∝ ).

N1/4dcontacts under consideration, one reduces the height ofIn order to lower the ohmic-contact resistance in the

the metal–semiconductor potential barrier by varyingthe chemical composition of the semiconductor in itsnear-surface region.

ered by forming a near-surface narrow-gap layer, sinceFor example, the potential barrier height can be low-the potential barrier height typically decreases as theband gap of the semiconductor is decreased.

(W/Ni + In/Ni)–GaAs contacts at In particular, in the course of heat treatment of theinteracts with GaAs and forms NiT ≈ 300°C, nickelatoms mix with W and Ni2GaAs while Ni2GaAs [23]. The In0.6Ga0.4Asphase and NiAs islands appear at the Niinterface at 700°C, while an In2GaAs–GaAsformed on almost 90% of the GaAs surface area atxGa1 – xAs layer is900the potential barrier at an interface with a metal than°C. This material features a much smaller height ofdoes GaAs since the Fermi level in the conduction band rather than in the band gap.

n-InAs is pinned incontacts heated to 600An InxGa1 – xP alloy is formed in the Pd/In–bly the potential barrier with Pd [24] in comparison°C; this alloy reduces apprecia-n-GaPwith the height of the Pd–n-GaP barrier.

tance on the height of the metal–semiconductor poten-Theoretical dependences of the ohmic-contact resis-tial barrier in the case of thermionic mechanism of thecurrent flow are shown in Fig. 3.

ohmic contact is given by [25]

The smallest possible value of resistance of thisRmin

=-qA-----k--*----T--ln-----[---1----+------exp----1

c

-----(--–---χ----/--kT-------)---]

.(10)

Here, the Fermi level energy conduction-band bottom and, for a nondegenerateχ is measured from the

n-equal totype semiconductor in the case where χ ӷ kT/q, isRmin

=-kχqA-------*----T-exp⎛⎝-kT-----⎞

kNcc

⎠ = -qA-------*----T--N----- = ----2----π---m-----*---kT------.dqNd(11)

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

1267

Rc, Ω cm210010–1qϕqϕqqqbb = =0. ϕ10–2q10.1b 5=ϕ ϕ20b ϕ5 eV.=2 0.2b e e5=b = 0VV e V0.3 eV10–3.1 qeϕVb = 10–40.07q5ϕ eVb =10–5 0.05 eV10–6qϕb = 0.025 eV10–710–8

100

150

200

250300400

T, K

Fig. 3.tance R Calculated temperature dependences of the resis-c of the ohmic contact to GaAs at various values ofthe height ϕin the case of the current flow according to the theory ofb of the metal–semiconductor potential barrierthermionic emission.In the case of GaAs [26], the lowest possible value ofthe contact resistance is given by

15

Rminc[Ω c

m 2

] = 1.55×10

–5

-300--T-----10

---N------.(12)

d

At Nd ≈ 1015–1017 cm–3 and T = 300 K, in which casetunneling is unimportanrt, one can form ohmic con-tacts to GaAs with the resistance Rmin

c = 2 × 10–5–2 ×10–7 Ω cm2.

sion theory) [10] implies that the resistance of the3.2.2. Field emission. Tunneling theory (field emis-metal–semiconductor ohmic contact per unit area isgiven by

1

m*q2∞

-R----=------------∫

-------------------T----(--E----)-----------]----–----1

-dE,(13)

c2πប3

exp[(E–χ)/kT0where rier with an energy T(E) is the probability of passage of a charge car-lower than qϕE through the barrier with the heightb by the value of ∆E:

–2(∆E)

3/2

T(E)∝exp---3---E--------V----1/2

------,(14)

00d

where E00 is the Padovani–Stratton [22] parameter andVence. According to [27–29], we have

d is the diffusion-related (contact) potential differ-RkTc≈-qA-------*----T---2--E----00

-qV-------exp⎛qV--------d⎞d⎝E00⎠⎛(15)

×exp-χ

2εskT-----∝exp⎜⎝

------------εប---0--m-----*---ϕN----1/2-b---⎞⎟⎠.SEMICONDUCTORS Vol. 41 No. 11 2007

Rc, Ω cm2106104qϕb = 0.8 eVqϕb = 1.0 eV102qϕb = 0.6 eV10010–2qϕb = 0.4 eV10–4qϕb = 0.2 eV10–610–810–1010

171018101910201021Nd, cm–3

Fig. 4.tance R Calculated dependence of the ohmic-contact resis-c on the concentration of uncompensated donors Nand the height ϕdrier in the case of the current flow according to the theory ofb of the metal–semiconductor potential bar-field emission [29].

An analysis of this formula shows that, if the currentflow through the ohmic contact is governed by the fieldemission, we obtain the following results:

the square root of the charge-carrier concentration is(i) the contact resistance exponentially increases as decreased (i.e., as the potential barrier width isincreased);

the potential barrier height is increased; and

(ii) the contact resistance increases exponentially asof temperature.

(iii) the contact resistance is virtually independenttance on the charge-carrier concentration and theCalculated dependence of the ohmic-contact resis-metal–semiconductor barrier height is shown in Fig. 4in the case where field emission is the main mechanismof current flow.

theory of thermal field emission [30], the dependence3.2.3. Thermal field emission. According to theof the forward current density exponential

J on the voltage V isJ=JqV

sexp⎛⎝--E----⎞,

(16)

0⎠

where

EE0=E00coth⎛-----00⎝kT--⎞⎠

;(17)in this case, the saturation current density temperature as

Js depends on

JATπE00s=-------------k----coth-------(---q-(--ϕE----b---–-----qV---------+----χ----)

-00/kT)

(18)

×exp⎛⎝-kT-χ----–ϕ----b--+χT--------⎞

.0⎠

1268

BLANK, GOL’DBERG

R10c, Ω cm20Nd = 5 × 1018 cm–3(a)10–110–2Nd = 1019 cm–310–3Nd = 2 × 1019 cm–310–4Nd = 3 × 1019 cm–310–5–310–6Nd = 1020 cm–3Nd = 5 × 1019 cm50100150200250300100T, K10–1(b)10–210–310–4qϕb = 0.7 eV10–510–6qϕb = 0.5 eVqϕ10–7b = 0.3 eV10191020Nd, cm–3

Fig. 5. Resistance R(a) temperature sated donors NT for various concentrations of uncompen-c of the ohmic contact as a function ofd in GaAs and (b) the concentration Nd at var-ious heights of the metal–semiconductor potential barrier ϕby the example of nonalloyed Au/Ti–tures. The lines represent the calculated dependences basedn-GaAs ohmic struc-bon theory of thermal field and field emission, while thepoints represent the experimental data [21].

shows the following data:

Analysis of the formulas for thermal field emissionis exponential;

(I) the dependence of the forward current on voltagedence on the semilogarithmic scale is equal to 1/(II) at each temperature, the slope of this depen-quantity depends on the intrinsic parameters of theE0; thissemiconductor at a given temperature rather than on thebarrier’s properties; and

of current on voltage on the semilogarithmic scale(III) the cutoff on the vertical axis in the dependenceyields the value of saturation current, while the depen-dence of

J---s---coth-----------(-E00T

---------/--kT-------)-on

-E-1---0

is linear on the semilogarithmic scale; the slope of this

dependence corresponds to the height of the metal–semiconductor potential barrier.

V The contact resistance 0 is equal toRc = dV/dJs per unit area atRϕb

c∝exp⎛⎝-E---------------------------------------⎞.00coth(E00/kT)⎠

(19)

controlled by thermal field emission, then

Thus, if the current flow through an ohmic contact isthe potential barrier height (i) the contact resistance increases exponentially asϕb is increased and

is increased but to a much lesser extent than in the case(ii) the contact resistance decreases as temperatureof thermionic emission.

The contact resistance for nonalloyedAu/Ti-GaAs(2–10 nm)–n+-GaAs(n+ = 1020 cm–3)–n-tion to temperature and electron concentration in GaAsGaAs(n = 1018 cm–3)–GaAs ohmic structures in rela-was calculated by Nien-Po Chen et al. [31] taking intoaccount both the thermal field emission and field emis-sion (Fig. 5).

heterostrucrure with two-dimensional (2D) electronWe note a particular case of the ohmic contact to agas [32], for example, to GaN/Althe current flowing from the metal to the semiconductorxGa1 – xN. In this case,consists of the thermal emission current, longitudinaltunneling current, and the tunneling current caused byquantum wells with the latter current being predomi-nant. The probability of electron tunneling in theith subband with the energy Ei is given by

T(Eqϕ----b----–----qV------E---–----(---E----i--–-----χ---)

i)=exp–----,

(20)

00

while the contact resistance is equal to

R2qm–1

c=------π---2---*ប

---kT

3-----(Σ1+Σ2),

whereΣ1=---i-i--+-----1-00--⎛⎝1+-E2---kT-00----⎞qϕ⎠

exp–-------b----–---E-(---E----i--–-----χ---)-,E∑

E/E00i<χ

(21)

ΣE----i--/--E----002=

--ln⎛E----i---+----χ---⎞exp–q---ϕ----b----–----(---E----i--–-χ)

E∑

i+1⎝kT⎠E--------.

00

i>χ

In particular, the contact resistance of the Al/Ti/Ta–GaN/AlGan-decreases with temperature according to a law close tox1 – xN structures with a 2D electron gasexponential, from ~10–4at 300 K.

Ω cm2 at 77 K to ~10–6 Ω cm2metals used in calculations of the contact resistances.Table 2 lists the parameters of semiconductors andstudies [33, 34] that another mechanism of current flow3.2.4. Metal shunts. It was shown in our recent(current via metal shunts) can manifest itself in alloyedmetal–semiconductor contacts in the case where disso-lution of semiconductor in metal and recrystallization

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS1269

Table 2. Work functions Φm for electrons leaving the metal and the electron-affinity energies Xs in the semiconductor withenergy gap Eg and density of surface states Ds for various semiconductors [10, 12, 44–49]

MetalAgAlAuCoCrCuFeInMgMoNiPbPdPtSbSnTaTi

Φm, eV4.424.185.1–5.24.974.4–4..594.463.973.614.215.15–5.24.045.175.43–5.6.5.434.23.83–4.33

SemiconductorZnOZnSZnSeCdSAlN

GaN (wurtzite)InNGaAsInAsGaPInPGaSbInSbCSiGe4H-SiC6H-SiC

Eg, eV3.463.62.72.436.23.390.71.4250.32.261.3440.7260.175.46–5.61.120.6613.233.0

Xs, eV

Ds, eV–1 cm–2

–––1.6 × 1013

(1–2) × 1011 (for SiO2–GaN)

12.5 × 1013

–2.7 × 1013

––––2.7 × 1013

–~1013~1013

4.094.770..1–4.074.93.84.384.0.59–4.0.04.0.07

occur in the course of heat treatment. These shunts arerepresented by metal atoms deposited onto the imper-fection lines (for examples, dislocations) and shunt thespace-charge layer. In this case, the electric field is con-centrated at the edges of these “needles” and the currentflow is realized owing to field emission.

The presence of metal shunts in semiconductordevices was also assumed previously, in studies of theresistance of epitaxial films based on TiN [35] and ingaining insight into the mechanism of the reverse-cur-rent flow in the Ni-GaN Schottky diodes [36, 37]. It wasnoted recently [38] that indium diffuses over disloca-tions during thermal annealing of GaN light-emittingdiodes (LEDs) with contacts formed of an alloy of theindium and tin oxides (ITO). These shunts of indiumatoms were observed directly. Wang et al. [39] used atransmission electron microscope to study the interfacereactions in the Ti/Al/Mo/Au ohmic contacts to theAl/GaN heterostructures and showed that the reactionproduct is TiN. In this case, a correlation was observedbetween the appearance of TiN islands and the densityof dislocations in the semiconductor; these dislocationsacted as short circuit diffusion channels.

The current–voltage (I–V) and capacitance–voltage(C–V) characteristics of Schottky diodes based onGaAs and GaP under conditions of continuous heatingwere studied in [40–42]. It was established that the In-GaAs and In-GaP barrier contacts are transformed intoohmic contacts at temperatures that are much lowerthan the melting points for the metal or metal–semicon-SEMICONDUCTORS Vol. 41 No. 11 2007

ductor eutectic and prior to formation of any recrystal-lized layers (heavily doped or with graded-gap struc-ture); in this case, the ohmic contact is retained evenafter cooling of the structures.

One may assume that, in the course of heating,metal (for example, indium) atoms diffuse over disloca-tions or other imperfections; as a result, metal shuntsare formed. It is these shunts that give rise to an ohmiccontact.

In the case under consideration, if the atomic radiusof metal is smaller than the semiconductor’s lattice con-stant, the ohmic-contact resistance is given by

(ρ+αT)W-----------------------.Rc=------0

2πrP

(22)

0; α isHere, ρ0 is the resistivity of the metal at T the temperature coefficient of resistivity; W is the thick-ness of the space-charge layer; r is the atomic radius ofthe metal; and P is the density of dislocations or otherimperfections onto which metal atoms can be deposited.In Fig. 6, we show the calculated dependences of theohmic-contact resistance on the dislocation density inthe semiconductor in the case where shunts are respon-sible for formation of an ohmic contact. It can be seenthat, at a low dislocation density (104–106 cm–2) charac-teristic, for example, for GaAs crystals, the mechanismof current flow over metal shunts is unimportant. At thesame time, at a high dislocation density (108–1010 cm–2)characteristic, for example, for GaN [43], the resistance

1270

BLANK, GOL’DBERG

Rc, Ω × cm2100P = 106 cm–2

10–1(a)

107 cm–210–2108 cm–210–3109 cm–210–41010 cm–210–51011 cm–210–6

1012 cm–2

10–7

10020030040010010–1P = 105 cm–2106(b)

cm–2

10–2107 cm–2

10–3108 cm–2

10–4109 cm–2

10–51010 cm–2

1011 cm–210–61012 cm–2

10–7

100

200300

400T, K

Fig. 6. Calculated dependences of the resistance Rmetal–semiconductor ohmic contact on temperature c of thevarious densities of defects (dislocations) T atductor in the case of the current flowing through the metalP in the semicon-shunts in (a) GaAs and (b) GaN.

related to the metal shunts can become predominant inthe total contact resistance.

with a fairly low dislocation density, the mechanism ofIn addition, in alloyed contacts to semiconductorsshunting can also become important since the densityof imperfections (in particular, dislocations) increasesdrastically in the case of alloying metal to a semicon-ductor due to the difference between lattice constants ofthe materials in contact.

tor contact flows over metal shunts, the contact resis-Thus, if the current in an ohmic metal–semiconduc-tance increases as temperature is increased, which ischaracteristic of the metallic type of conductivity.

4. EXPERIMENTAL RESULTS4.1. Ohmic Contacts to II–VI SemiconductorsAlN nitrides), the surface-state density is low and anIn II–VI semiconductors (and also in the GaN andohmic contact is formed with the metals for whicheither the work function of electrons the electron affinity conductor or the work function for electrons from metal

XΦm is smaller thans for an electron in an n-type semi-Φsemiconductor and the band gap m is larger than the sum of electron affinities ductor); thus, we have

Ep-type semicon-Xs for the

g (a Φm < Xs in the case of an n-type semiconductor,(23)Φm > Eg + Xs in the case of a p-type

semiconductor.(24)

ing from metals The values of the work function for electrons escap-conductor Φm, the electron affinity of the semi-and the density of surface states Xs, the band gap of the semiconductor ductors are listed in Table 2 [10, 12, 44–49].

D for various semicon-Eg,spound began to be used in the fabrication of photode-In recent years, ZnSe and alloys based on this com-tectors for blue and ultraviolet optical radiation, mainlyfor detection of laser radiation; the above materials canbe easily grown on GaAs substrates.

example, with the use of In or a Ti-Pt-Au alloy (Ohmic contacts to n-ZnSe can be produced, for4.3 eV); in order to form the contact, it is only neces-Φm =sary to remove possible intermediate layers, which isaccomplished by heat treatment at T > 200°C [50].metals with a work function At the same time, in the case of Φp-ZnSe, there are nom that exceeds the sum ofthe electron affinity Xthe potential barrier always exists. Therefore, in orders and the band gap Eg; as a result,to form an ohmic contact, one typically changes thecomposition of the near-surface region, for example, bygrowing a graded-gap p-ZnSexTe1 – x layer on the sur-face or using HgSe (Xresult, the barrier height is reduced to 0.4 eV [51].s = 6.1 eV) as the metal film; as aExamples of ohmic contacts to ZnSe and ZnO are givenin Table 3 [52–65].

ZnSe and ZnO has not been studied in detail; however,The mechanism of current flow in ohmic contacts tothe tunneling mechanism of current flow is assumedsince the resistance of the contact to heavily doped ZnSe [52] and independent. Yang and Schetzina [66] assumed thatn-ZnO [] is typically temperature-n-tunneling combined with thermionic emission takesplace in p-ZnSe (p = 1017–1019 cm–3height surmounted by electrons estimated at ) with the barrier0.3 eV (for structures with p = 1017 cm–3).

ϕb =4.2. Ohmic Contacts to Semiconductor Nitridesohmic contact to A similar situation is also observed in the case of anrial for production of LEDs and photodetectors thatn-GaN; the latter is a promising mate-operate in the short-wavelength visible and ultravioletspectral regions. The Fermi level is almost not pinnedat the surface [67]; however, conventional chemicallystable metals have a work function larger than the elec-tron affinity for GaN (Table 4); the resistance of ohmiccontact to GaN increases drastically with an increasingwork function for electrons leaving the contact metal(Fig. 7) [68]. Therefore, in order to form an ohmic con-tact to the n-type semiconductors, multicomponent

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

Table 3. Ohmic contacts to II–VI semiconductors

MetalTi/Pt/AuInIn/Au

Si-As/BeTeBeTeBeTeCu/AuPd

p+-ZnTeTi/AuCuxO

CarrierAnnealing

Semicon-concentration, temperature,

ductor

cm–3°CZnSe

ZnSeZnSeZnSeZnSeZnSeZnSeZnSeZnSeZnOCdTe

n = 2 × 1019n = 2 × 1018n = 1017p

p = 1018

p = 2 × 1017p = 4.5 × 1018p

p = 7 × 1016n = 1019

250250–300

Contactresistance,Ω cm21.1 × 10–4

Compounds are not formed, In diffusion

10–21046 × 10–20.167

250

5 × 10–2

300

6 × 10–82.2 × 10–2

Formation of the Pd-Zn-Se phaseResonance tunneling of holesRoughening of the surface

The metal/Si-As/BeTe-ZnSe structureFormation of BexZn1 – xTeySe1 – y

Notes

1271

Refe-rences[52][][55][56][57][58][55][59, 62][53][][65]

Table 4. Ohmic contacts to n-GaN

MetalTi/Al

Carrier concent-ration, cm–3n = 1017n = 1017

n = 3.67 × 1018n = 7 × 1017n = 2 × 1017n = 4 × 1017Nanotubes

Annealing tem-perature, °C

Contact resis-tance, Ω cm28 × 10–68 × 10–68.63 × 10– × 10–61.19 × 10–78.9 × 10–81.8 × 10–27 × 10–63.86 × 10–610–55 × 10–4

Note

References[77][71][78][79][80][81][82][83][84][85][86]

Ti, TiN, Ti/TiNTi/Al/Ni/AuTi/Al/Pt/Au

Ti/Al/W2B/Ti/AuSi/TiGe/Cu/GeIn-Sn oxide

400–900Reaction of Ti and GaN

700–800800900

Lowering of the SiTi barrierand doping with donorsDonor-like VN vacancies

n = 1019

contacts are based on Ti and are formed in the course ofheat treatment of a compound with a low work-functionenergy. For example, Kim and Baik [69] formed ohmiccontacts to n-GaN (n = 15 × 1018 cm–3) by the alloyingof Si/Ti that produces titanium silicide as a result ofheating; the work function for electrons in this silicideis smaller than electron affinity for GaN (Φm ≈ 3.7 eVfor Ti5Si3).

Low resistance of the metal-GaN ohmic contact(as low as 10–6–10–7 Ω cm2 at high concentrations ofcharge carriers in the semiconductor) [70–72] is typi-cally related to the formation of nitrogen vacancies dueto interaction of GaN with the contact material, forexample, Ti. Such nitrogen vacancies form a damagedlayer under the contact; this layer acts as a heavilydoped layer.

SEMICONDUCTORS Vol. 41 No. 11 2007

In the case of p-GaN, there are also no metals withΦm > Xs + Eg = 7.5 eV; in this situation, either com-pounds with a high work function are used or a nar-row-gap near-surface layer is formed in the semicon-ductor [73].

For example, in the case of alloying Au/Ni to GaNat 600°C, several binary intermetallic phases areformed; these phases lower the potential barrier [74].Alloying of Au/C/Ni makes it possible to additionallydope the near-surface layer [75] since C atoms act asacceptors. The Ru/Ni contact with annealing in the O2atmosphere can also be used for p-GaN. The formedRuO2 compound reduces the effective height of thepotential barrier, while NiO acts as a barrier to diffusionof released Ga and N atoms. This contact features ahigh transmission of light (84.6%) and low resistance(4.5 × 10–5 Ω cm2) [76].

1272

Rc, Ω cm2101100Pd10–110–210–310–44.0AlTiSn4.55.05.5CuBLANK, GOL’DBERG

6.0Φm, eV

Fig. 7. Experimental dependence of the resistance Rc of themetal–n-GaN ohmic contact on the work function Φm forelectrons leaving the contact metal.

The most widely used contacts to n-GaN are givenin Table 4 [71, 77–86], and contacts to p-GaN are listedin Table 5 [79, 87–105].

In Table 6, data on the current flow mechanisms inohmic contacts to n- and p-GaN are given [34, 83, 84,

Table 5. Ohmic contacts to p-GaN

MetalPdPd/RuPd/AuPdAu

Pd/Au/InGaNNi

Ni/Au/InGaNNi/InNi/Pt/AuNi/AuPt/NiNi/AuNi/Au

In–Sn oxidePt/RuTa/TiNi/Cr/AuNi/Pd/AuIn0.19Ga0.18NAg/Pd

88, 91, 106–112]. In the case of heavily doped GaN, theresistance of an ohmic contact is nearly independent oftemperature (Fig. 8b, curve 1) and decreases as the con-centration of the majority charge carriers is increased(Fig. 8a), which substantiates the assumption that themain mechanism of the current flow is field emission(tunneling) [80, 106–108]. In the case of medium-doped GaN, the contact resistance decreases as temper-ature is increased, which makes it possible to assumethat the main mechanism of the current flow is the ther-mionic emission 69, 109, 110, 112]. It was shown byKwak et al. [112] that the ohmic-contact resistancedecreases as temperature is increased; however, thedependence of Rc on T was slight, and it was assumedthat the main mechanism of current flow is hoppingconductivity with involvement of deep-level centers.Thus, the resistance of an ohmic contact to GaN eitheris temperature-independent or decreases as temperatureis increased. Only Lu et al. [107] observed an increasein the contact resistance as temperature was increasedfrom 50 to 300°C. This behavior was attributed to orig-ination of imperfections in the GaN layer as tempera-ture was increased; these imperfections presumablybrought about an increase in the height of the metal–semiconductor potential barrier.

Carrier concentration, Annealing tem-Contact resis-cm–3perature, °Ctance Rc, Ω cm2p = (0.28–2.5) × 1017

p = 3 × 1017p = 3 × 1017p = 1018p

p > 1.7 × 1019p

p = 2 × 1017p = 9.4 × 1016p = 2 × 1017p = 1.7 × 1017ppp

p = (2–3) × 1017p = 7 × 1017p = 5 × 1017

500

2.4 × 10–5 10–41.5 × 10–61.1 × 10–6(12–6) × 10–3(8–9) × 10–32.1 × 10–24 × 10–68 × 10–3

Notes

Formation of acceptor-like states

References[87][88][][][90][91][92][93][94][95][96][97][98][79][99][100][101][102][103][104][105]

750, 950No annealing

Formation of p-In0.19Ga0.18NHeavy dopingHeavy doping

500600500

6 × 10–44.5 × 10–22.2 × 10–63 × 10–51.6 × 10–24.6 × 10–61.1 × 10–6(4–6) × 10–5

Formation of NiO or a-Ni-Ga-OFormation of Ga vacancies

Formation of p-In0.1Ga0.9 and Ga vacancies

p = 5 × 1017

330–530A decrease in the barrier height

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

Table 6. The current flow mechanisms in ohmic contacts to n- and p-GaN

Contact

n-GaN (theory)

Carrier concen-Temperature Concentration Barrier tration, cm–3dependencedependenceheight

Decreases

superlinearlyDecreasessublinearlyDecreases∝ 1/n

Current flowmechanism

1273

Refe-rences[108]

TunnelingTunnelingTunnelingTunnelingTunneling

Tunneling

Thermionic emissionat T ≥ 200 K

Tunneling at T ≤ 200 KThermionic emissionThermionic emissionThermionic emission

[106][77][107][82][91][109]

Ti/Ag–n-GaN

Ti/Ag/Ni/Au–n-GaNTi/Ag/Pd/Au–n-GaNTi/Al/W2B/Ti/Au–n-GaNNi–p-GaNPt–n-GaN

(1.5–1.7) × 1018

(4–30) × 1017The same6 × (1017–1020)IncreasesDecreases

Does not depend19>1.7 × 10The same

2 × 1017Decreases

∝ exp(T–1/4)(1.8–10) × 1017Decreases

∝ exp(1/T)Decreases ∝ exp(1/T)Decreases ∝ exp(1/T)10173 × 1017(2–22) × 1017

Increases

0.13

Pt–n-GaNSi/Ti–n-GaNAu/Ti/Si/Ti–n-GaN

[110][111][83][84][34][88]

In–n-GaNPd/Ru–p-GaNPd/Pt/Au–p-GaN

Decreases ∝ exp(T–1/4)

Metal shunts

Thermionic emission before annealing, tunneling after annealing at 500°CHopping conductivity[112]

It is noteworthy that all studies were carried out

using contacts with thin-film metal in which GaN ispractically not dissolved even at high annealing tem-peratures.

We previously observed [34] an increase in theresistance of the In–n-GaN ohmic contact with temper-ature, which is characteristic of metallic conductivity(Fig. 9); it was assumed that the ohmic contact ofalloyed In to GaN, during the formation of which dis-solution of semiconductor in metal occurs, can beformed due to the appearance of metallic shunts thatthread through the space-charge layer as a result of dep-osition of In atoms onto the near-surface dislocationsand other imperfections of the semiconductor. The den-sity of dislocations in conventional GaN crystals is 107–108 cm–2. The dislocation density determined from thetemperature dependence of the ohmic-contact resis-tance was estimated at ~108 cm–2; i.e., it was close tothe dislocation density in ordinary GaN crystals.

Alloys in the GaN-AlN system are very promisingfor ultraviolet photoelectronics; however, fabrication ofohmic contacts to AlxGa1 – xN alloys and, in particular,

SEMICONDUCTORS Vol. 41 No. 11 2007

to n-GaN, encounters serious difficulties since the elec-tron affinity in n-AlN is very small (0.6 eV), while theFermi level is virtually not pinned at the surface. Ohmiccontacts to GaN–AlN alloys are typically formed usingthe same scheme as in the case of GaN.

In Table 7, we list examples of ohmic contacts toGaN–AlN alloys [113–119].

Blank et al. [120] studied the mechanism of currentflow in the Pd–p-Al1 – xGaxN ohmic contact at composi-tions close to GaN (x = 0.94). It was found that, at aconcentration of uncompensated acceptors of 3 ×1018 cm–3, the ohmic-contact resistance decreasesexponentially as temperature is increased, which isaccounted for by the thermionic mechanism of currentflow. The potential barrier height determined from thedependence of Rc on T was found to be equal to 0.05 eV.However, at Na – Nd = 3 × 1018 cm–3, the value of Rc isalready independent of temperature, which makes itpossible to assume that the main mechanism of the cur-rent flow at the concentrations under consideration istunneling.

1274BLANK, GOL’DBERG

Rc, Ω cm210–1(a)10–210–310–410–500.51.01.52.010–1N–1/22.5,10–93.0cm3/2d10–2(b)210–310–4110–510–62.02.22.42.62.83.03.23.4103/T, K–1

Fig. 8.ohmic contact on the concentration of uncompensated

(a) Dependence of the resistance of the Ti/Ag–n-GaNdonors Ning (qϕd at 300 K. The mechanism of current flow: tunnel-Rb = 0.067 eV) [106]. (b) Dependence of the resis-tance c of the Pt–p-GaN ohmic contact on temperature:

(1) Concentration of uncompensated acceptors Na =

1018 cm–3 and the current flow mechanism is tunneling

(qϕb = 0.42 eV); (2) Na = 1.8 × 1018 cm–3 and the current flow

mechanism is thermionic emission (qϕb = 0.53 eV) [110].

tacts to (AlN)Guseœnov [115] studied the resistance of ohmic con-that, as the heterostructure’s band gap x(SiC)1 – x heterostructures; it was shown

contact resistance increases as

Eg increases, the

R2–9c[Ω c m ] = 2.7 × 10 exp ⎛E g[eV]⎝ - -- - - 0.28- - -- - - -- - - - - ⎞ ⎠

. (25) This dependence functionally corresponds to thedependence of Rc on Eg for GaAs1 – xPx alloys; thisdependence was reported previously [121].

ers are often used as substrates for the formation ofIn semiconductor devices, high-resistivity AlN lay-active layers in devices. Formation of thin-film ohmiccontacts to this semiconductor encounters serious diffi-culties; in this case, one can use the formation of

alloyed ohmic contacts with the mechanism of currentflow through metal shunts [34].

Rc, 10–3 Ω cm2201816141210820160180200220240260280300320340T, K

Fig. 9. The resistance Rcontact on temperature. The current predominantly flowsc of the alloyed In–n-GaN ohmicthrough metal shunts [34].

electronics. Since the band gap of InN is about 0.7 eVInN is also promising for medium-wavelength opto-the InN–GaN–AlN alloys can be used in solar cells and,photodetectors in the infrared, visible, and ultravioletspectral regions. Examples of ohmic contacts to InNand InGaN alloys are given in Table 8 [49, 122–127].

4.3. Ohmic Contacts to Semiconductor Arsenides,

Phosphides, and Antimonides

ductors (aside from nitrides and Semiconductors of Group IV and III–V semicon-concentration of surface states at the free surface; thesen-InAs) have a highstates are located deep in the band gap, which bringsabout a rigid pinning of the Fermi level at the surface.Therefore, in order to form an ohmic contact to thesemiconductors under consideration, one has to:(i) reduce the density of surface states,chemical composition of the near-contact region, or

(ii) reduce the potential barrier height by varying thenear-contact region.

(iii) increase the charge-carrier concentration in the

conductor electronics and optoelectronics as a material

At present, gallium arsenide is widely used in semi- for high-speed integrated circuits, microwave devices,detectors of nuclear particles, LEDs, and lasers. At the same time, gallium arsenide is a direct-gap model semi-conductor suitable for studies of electrical and opticalphenomena. The density of surface states in GaAs is

very high (higher than 1014of the metal–GaAs potential barrier (the Schottky bar- cm–2 eV–1), and the heightrier) is nearly independent of the metal’s nature, due topinning of the Fermi level at the surface, and equalsapproximately 0.9 eV.

be reduced by passivating the surface in solutions that

The density of surface states in a semiconductor cancontain sulfur ions. In the course of treatment of GaAsSEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

Table 7. Ohmic contacts to alloys in an AlN–GaN system

Metal

Ti/Al/Ti/AuAl/In

Al, Al/Cr, Al/N, Al/PtTi/Al/Mo/Au/SiTi/Al/Mo/AuNi/AuNiAl/Ti

Semiconductorn-AlGaN/AlNn-AlGaN/AlN(AlN)x(SiC)1 – xAlGaN/AlNAlGaN/GaNSuperlattice

p-Al0.15Ga0.85N/GaN(p = 1018 cm–3)Superlatticep-AlGaN/GaNn-AlGaN/GaN

Annealing tem-perature, °C

7001200800

Contact resis-tance, Ω cm25.6 × 10–510–4–10–2

10–6

Note

1275

Refe-rences[113][114][115][116][117][118][119][114]

Formation of TiSi and a dec-rease in the barrier heightFormation of TiN

650400–500900–950

4 × 10–69.3 × 10–45.6 × 10–6

Table 8. Ohmic contacts to InN and to InN-based alloys

MetalTi, Al, NTi/Pt/AuTi/Pt/AuW

Si0.44, Ti/Al

Semiconductor

Carrier concen-tration, cm–3

Annealing tempe-rature, °C

Contact resis-tance, Ω cm2

Note

Refe-rences

n-InN(1.5–2.3) × 1018n-InN5 × 1020InN

n-In0.17Ga0.83N1.63 × 1019n+-In0.65Ga0.35Nn-InN

n+-In0.75Ga0.25N

W, WSi0.44n+-In0.65Ga0.35N

n-InN

W, WSi0.44, Ti/Aln+-In0.65Ga0.35N

n-InN

n+-In0.75Ga0.25N

Without annealing1.4 × 10–7–10–6

300–4201.8 × 10–7380(5–6) × 10–79502.7 × 10–8

Without annealing(1–4) × 10–7Without annealing(1–10) × 10–7Without annealing10–4Without annealing(1–4) × 10–7Without annealing(1–10) × 10–7

900(1–4) × 10–7600(1–10) × 10–7

10–4

[122][123][123]

Formation of β-W2N[124]

[125]

Field emission[125]

[125]

Field emission[126]

[126]

Field emission[127]

[127][127]

in the Na2S or (NH4)2S solutions, the near-surfaceimpurity O atoms (responsible for the surface states)are replaced with S atoms that are kept at the surfacebecause of chemical adsorption. The energy released asa result (lower than 40 kJ/mol) is lower than the energyreleased during chemical adsorption (higher than100 kJ/mol) and is not high enough for formation ofintrinsic defects that give rise to pinning of the Fermilevel [128–131]. The density of surface states Ds inGaAs decreases to 2 × 1013 cm–2 eV–1 as a result oftreatment in (NH4)2S and to 1013 cm–2 eV–1 as a resultof treatment in (NH4)2Sx. A still lower density of sur-face states Ds in GaAs (7.6 × 1012 cm–2 eV–1) wasobtained in the case where a thin (3.5 nm) n-GaAs:Belayer was grown on the surface (n = 5 × 1016 cm–3)[132], as a result of which the Fermi level nearly wasnot pinned at the surface of the GaAs semiconductor.

SEMICONDUCTORS Vol. 41 No. 11 2007

It is noteworthy that passivation of semiconductorsin solutions containing ions of sulfur or other elementskept at the surface due to chemical adsorption reducesthe density of surface states for other materials as well.For example, treatment of the GaN surface in (NH4)2Sxbrought about a decrease in the density Ds from 1012 to

8.3 × 1010 cm–2 eV–1 [133]; in this case, the height of theNi/Cu–n-GaN Schottky barrier was 1.099 eV, which isclose to the theoretical limit (1.10 eV). Passivation ofthe Ti/Al–n-GaN:Si ohmic structures (n = 3 × 1018 cm–3)reduced the contact resistance approximately by twoorders of magnitude owing to removal of oxide fromthe GaN surface and to a shift of the Fermi leveltowards the conduction-band bottom [134]. Passivationof 4H-SiC in solutions containing the NO and NH3 ionsbrought about a decrease in the density Ds from ~1013to ~ 2 × 1012 cm–2 eV–1 [135].

1276BLANK, GOL’DBERG

Rc, Ω cm210–4qϕb = 0.6 eV10–5qϕb = 0.5 eV10–6qϕb = 0.4 eV10–710–800.51.01.52.02.53.03.5N–1/2

10–10cm

3/2d,Fig. 10. Dependences of the resistance Rc of the ohmic con-tact to acceptors p-GaAs on the concentration of uncompensatedNa in the initial material at 300 K. The currentflow mechanism is tunneling (qϕlines represent calculated dependences, while triangles,b = (0.4–0.6) eV). Thesquares, and circles represent the experimental data [209].

(1–3) The Al–in H× 10–4n- ΩGaAs ohmic contact with the resistance of cm2 was prepared by treating the surfaceface states in 2S solutions in order to reduce the density of sur-this method has not been widely recognized.

n-GaAs (n = 1018 cm–3) [136]; however,contacts to GaAs consists in formation of a heavilyThe most widely used method of forming the ohmicdoped (with the carrier concentration as high as 1020–1021metal–semiconductor potential barrier. Most often, this cm–3) surface layer, which narrows appreciably thelayer is formed owing to interfacial chemical reactionsthat bring about dissociation of GaAs and origination ofa new layer of heavily doped GaAs [74].

widely used Al/Ge/Ni contact, the latter at first reactsFor example, in the course of heat treatment of thewith GaAs at decomposes at T < 250°NiGaT > 250C and forms Ni°C with resulting formation ofxGaAs; this phasealso with Ge (giving rise to NiGe compound) and withy and NiAsz compounds. In this case, Ni reactsAl (giving rise to Nicooling, the concentration in this layer being as high as 10n+-GaAs:Ge layer grows with the electronxAly compound); in the course of201021 cm–3 [137].

–contacts, Ge first diffuses from the Au + Ge alloy to theIn the case of the Ni/[Au + Ge(27%)]/Ni/Au–n-GaAsupper layer, while Ni reacts with GaAS and formsNixGaAs. At temperatures of 375–400°C, Au reactswith Ga and forms the trates into Niβ-AuGa phase, while Ge pene-xGaAs and replaces Ga. The contactacquires the new heavily doped GaAs layer is formed as a result ofβ-AuGa/NiAs:Ge/GaAs structure, while acooling [138]. The resistance of these contacts ton-GaAs is as low as 3.6 × 10–6 Ω cm2 at n = 1016 cm–3

[139], 10–6 Ω cm2 at n = 2 × 1016 cm–3 [140], 5 ×10−7 Ω cm2 at n = 1.5 × 1017 cm–3 [141], and 4 ×10−7 Ω cm2 ar n = 2.2 × 1018 cm–3 [141].

the ohmic contact to GaAs was reduced by approxi-As the technology was improved, the resistance ofmately an order of magnitude in the decade [6]; thecontact resistance was inversely proportional to thecharge-carrier concentration [142].

GaAs consists in the formation of a narrow-gap (mostThe other method for attaining the ohmic contact tooften, Gaface Fermi level is pinned in the conduction band, which1 – xInxAs) near-surface layer. In InAs, the sur-appreciably reduces the height of the metal–semiconduc-tor potential barrier in the alloys of a GaAs–InAs systemcompared to the height of the metal–GaAs barrier. TheIn contact to of the above contact. This contact is formed as a resultp- and n-GaAs can serves as an exampleof heating to 300is observed after heating at 500°C; the lowest resistance of the contacttemperatures, evaporation of As atoms becomes notice-°C. At higher annealingable and increases the contact resistance.

described in Table 9 [143–172]; the data on ohmic con-Various types of ohmic contacts to n-GaAs aretacts to ohmic contacts to ternary alloys based on GaAs are rep-p-GaAs are given in Table 10 [173–182]; andresented in Table 11 [174, 181, 183–197]. Results of thestudies carried out in the last decade are given; earlierstudies were analyzed in reviews [6, 14].

of ohmic contacts to GaAs has attained very small val-As a result of technological efforts, the resistance Rcues (10–6–10–8 Ω cm2the mechanism of current flow in ohmic contacts to; see, for example, [21, 198–202]);GaAs was analyzed in [111, 203–211].

semiconductor in the metal does not occur were studiedThe contacts in which appreciable dissolution ofin [111, 148, 151, 1, 208–211] (Table 12): the con-tacts were either nonalloyed, or thin-film, or were sub-jected to a heat treatment at comparatively low temper-atures (~300tration of doping impurity (10°C). Contacts to GaAs with a high concen-18–1020 cm–3mainly studied in the aforementioned publications. The) werepotential barrier height in the above contacts was fairlylarge: 0.14–0.26 eV [205], 0.5 eV [209], and 0.3–0.5 eV[206]; the main current flow mechanism was tunneling,with the low contact resistance ensured by pronouncednarrowing of the potential barrier (Fig. 10).

near-contact region failed for a lightly doped GaAs (theAt the same time, attempts to form a heavily dopedelectron concentration lower than 1017the potential barrier height was reduced as a result of cm–3); therefore,interfacial reactions and was as low as 0.068 eV [111]and 0.09 eV [204]. This circumstance ensured a sub-stantial above-barrier current and thermionic mecha-nism of its flow. The thermionic mechanism of currentflow was also detected in the case of the Inalloy (Fig. 11); it is noteworthy that the current flow mech-0.53Ga0.47Asanism changed to that of thermal field emission [207].

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

Table 9. Examples of ohmic contacts to n-GaAs

Metal

Au/TiAu/TiAu/Al/TiAu/Pt/TiAu/Ti/W/TiAu/Ge/Ni/AuAu/Pt/Ti/GaSNi/In/GeNi/Ge/AuIn

n+-InxGa1 – xAsIn0.7Ga0.3As/Ni/W2Ni/WPd/InPd/GePd/Sn

Pd/Sn, Pd/Sn/Au

(1.6–1.8) × 1018

4 × 10172 × 1018Carrierconcentration,

cm–3

5 × 1018

Annealingtemperature,

°C–NonalloyedcontactWithoutannealing

–400

Contactresistance,Ω cm22 × 10–6

2 × 10–6 (40 K)3.7 × 10–33 × 10–45.5 × 10–65.6 × 10–.1 × 10–6

Intermediate GaAs eliminates pinning

Treatment of the surface in (NH4)2S

Formation of TiGaS

Formation of InxGa1 – xAsNi is used for wetting

Formation of an InxGa1 – xAs graded-gap layerTunneling

Note

1277

References[143][151][144][148][146][147][157][150][156][158][1][170][171][172][169][145][163][1][165][166][153][160][161][149][152][155][159][162][167][168]

2 × 1018

300650375Nonalloyedcontact

550

600360360–430

3 × 10–610–7

10–6Formation of InxGa1 – xAs4 × 10–73 × 10–5

8 × 10–6–3 × 10–5

Pd/Sn, Pd/Sn/AuPd/Ge/Ti/AuPd/Ge/Ti/PtPd/Ge/Ti/Pt

Pd/Ge/Au/Pd/AuPd/Ge/Au/Pd/AuPt/Ti/Ge/PdGe/CuWSi

5 × 10187 × 1017

Nonalloyedcontact

340400380–450400400450–600400400

2.8 × 10–62.4 × 10–6(2.4–5.3) × 10–6

2 × 10–62.1 × 10–6

Formation of the AuGa phaseFormation of Ga5Pd

7 × 10–76 × 10–6

Formation of AuGa and Pd5Ga2

Formation of Ge8Pd21

Formation of a heavily doped GaAs : Ge layer

Formation of β-AuGa and WSi2

Thermionic emission was also observed in thealloyed In–n-GaAs ohmic contact [203] (Fig. 12). Inthis case, the potential barrier height was very small(0.03 eV) due to formation of graded-gap InxGa1 – xAsalloys with a narrow-gap layer near the metal. The den-SEMICONDUCTORS Vol. 41 No. 11 2007

sity of dislocations and other imperfections is fairly low(~(102–104) cm–2) in contemporary GaAs crystals; as aresult, formation of metal shunts can be disregarded inthe case of origination of an ohmic contact. The resis-tance of ohmic contacts in alloyed structures decreased

1278

Table 10. Ohmic contacts to p-GaAs

MetalAu/Ti

Au/Ge, Au/TiAu/Ti/Pd

Au : Cu/Pt : Ir/TiNi/Pd/AuIn

InxGa1 – xAsPdIn

Pd/Ge/Ti/Pt, Ti/Pdε1-Cu3Ge

p+7 × 1018Carrier concen-tration, cm–31017

Semi-insulatingGaAs : Crρ = 106–107 Ω cm2

p+

BLANK, GOL’DBERG

Annealing tem-perature, °C300–320320–380400300

Contact resis-tance, Ω cm2

9 × 10–6

NoteReferences[173][181][180][174][175][182][178][176][179][177]

1.47 × 10–62 × 10–6

Pt–Ti barrier for the deep-level Cu impurity

Laser-assisted depositionCan be used for both n- andp-GaAs

Convenient contact for micro-wave devices

Semi-insulator

<400550400

10–7

5 × 10–6

as temperature was increased (the same as for nonal-loyed contacts), and the main mechanism of currentflow was the thermionic emission.

Gallium phosphide is another III–V semiconductorwith a high density of surface states. The Fermi level atthe GaP surface is pinned practically at the midgap.Gallium phosphide is widely used in LEDs for red, yel-low, and green lights and in photodetectors of UV radi-ation [212, 213]. Ohmic contacts to GaP were fabricatedin many studies (the examples are given in Table 13)

Rc, Ω cm210–4

350°C400°C[33, 214–221]; however, the current flow mechanismhas been studied in these contacts very little.

An anomalous temperature dependence of theohmic-contact resistance Rc for the In–n-GaP contactswas detected by Blank et al. [33]; this resistanceincreased as temperature was increased; it was con-cluded that the main mechanism of current flow in anohmic contact is the current flow through metallicshunts (Fig. 13). As a result, it becomes understandablewhy In forms an ohmic contact to GaP. In the case of Incontact to GaAs, a graded-gap InxGa1 – xAs phase canbe formed; it is noteworthy that it is not only a solidsolution but also n-InAs in which the Fermi level ispinned in the conduction band. At the same time, in the

Rc, Ω cm20.400.200.100.060.04As-deposited10–5

450°C10–6

300350400T, K0.020.012.53.03..0Fig. 11. Dependences of the resistance Rc of the Pt/Ti–p-In0.53Ga0.47As on temperature T at various conditions ofcontacts’ treatment. The mechanism of current flow for thecontacts not subjected to heat treatment is thermionic emis-sion (qϕb = 0.13 eV). As the annealing temperature isincreased, the current flow mechanism transforms into ther-mal field emission [207].

103/T, K–1

Fig. 12. The resistance Rc of the alloyed In–n-GaAs ohmiccontact on temperature T. The current flow mechanism isthermionic emission (qϕb = 0.03 eV) [203].

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

Table 11. Examples of ohmic contacts to InGaAs and AlGaAs

Metal

AuBe/Au,

Pd/AuBe/Pt/AuAu/Pt/Ti/WNxAu/Pt/Ti/WSiNAu/Ni/Au/Ge/PdAu/Ni/Au/Ge/PdPd/GePd/GePd/Si/Ti/PtPd/Ge/Ti/PtPd/Pt/Au/PdPd/Ga/Ti/AuPd/Si/Pd/Ti/AuPd/Ge(Si)/Pd/Ti/AuTi/Pt/AuBe–In/As/AuWNx/WN0.5x/W

SemiconductorAlGaAs/GaAsn-InGaAsn+-InGaAsn-InGaAsn-InGaAsAlGaAs/GaAsn-InGaAsn-InGaAsn-InGaAsp-InGaP

p = 3 × 1019 cm–3n-AlGaAs/InGaAsn-InGaAsn-InGaAsGaAs with vertical InGaAs layersAlGaAs–p-GaAsp = 3 × 1019 cm–3n-InGaAs

Annealing tem-perature, °C

400250–4504004004004200415–440

(2.3–95) × 10–6

300–400425

4.3 × 10–7(0.9–1.1) × 10–6(1–10) × 10–6

Formation of a superlattice, ϕb = 0.14–0.26 V

Contact resis-tance, Ω cm2

4 × 10–610–8–10–72 × 10–7

10–6

Note

1279

References[193][187][190][186][1][197][191][183][184][192][195][188][174][181][196][194]

(1–2) × 10–63.7 × 10–6

Formation of PdGe and dif-fusion of Ge

Formation of PdGa, PdAs2, and InxGa1 – xP

The contact is formed owing to Au2Al and β-AuGaFormation of the Pd2Si phase

400

2 × 10–7

[185]

Table 12. Mechanisms of current flow in ohmic contacts to GaAs

Contact

In–n-GaAs

Au/Pt/Ti–n-GaAsPt/Ti/p-GaAsInAs/n-GaAsAu/Ti–GaAs

n+-InxGa1 – xAs–n-GaAsTi/Pd, NiGeAu–GaAsTi/Pt/Ag–p-GaAs

Au/Ti–n-GaAs (theory)Ni/InxGa1 – xAs/n-GaAs

Charge-carrierconcentration, cm–3n = 4 × 1015n > 1018

p = (5–10) × 1018n = 1020>5 × 1018

Current flow mechanismsThermionic emissionThermionic emissionThermionic emissionTunnelingTunnelingTunnelingTunnelingTunneling

Tunneling between metallic state and the conduction band

Potential barrier height, eV0.030.090.0680.14–0.26<0.7

References[203][148][111][205][151][1][208][209][210][211]

p = 4 × 1020n = 1020n = 1018

0.50.3–0.50.13

case of GaP, formation of a graded-gap phase can onlybring about the appearance of InP on the surface. It isnoteworthy that the latter semiconductor has a fairlywide band gap with the Fermi level pinned deep in thisgap. We may assume that, in this case, In is depositedon dislocations and other imperfections and forms

SEMICONDUCTORS Vol. 41 No. 11 2007

metal shunts. The density of metal shunts in GaP crys-tals, determined from the temperature dependence ofthe contact resistance, was found to be nearly tempera-ture-independent and equaled 107 cm–2, which is closeto the dislocation density in GaP. In the case of GaAs,the dislocation density is usually lower by three–four

1280

Table 13. Ohmic contacts to GaP

MetalAi/Si/PdSi/PdIn

Pd/Zn/Pd

Semiconductorn-GaPn-GaPn-GaPp-GaP

5 × 10173 × 10172 × 1017

BLANK, GOL’DBERG

Carrier concen-Annealing tem-Contact resis-tration, cm–3perature, °Ctance, Ω cm2

550350–650

10–52 × 10–4

Note

Solid-phase epitaxy of

Si onto GaP

Solid-phase epitaxy of Si onto GaP

Metal shunts

Formation of the p-p+-ZnP2 and Zn3P2 phases

References[214][215][216][33][217][218][219][220][221]

550

6 × 10–5

Ni, Ti + Au

n-GaP : Bep-In0.49Ga0.51P

5 × 10192 × 1019

400

orders of magnitude, so that the shunts do not play animportant role in the current flow mechanism. As aresult, thermionic emission is the main mechanism ofcurrent flow in the In–GaAs contact, while metallicshunts are related to the main mechanism of currentflow in the In–GaP alloyed contact.

The InP compound and, especially, the InGaP alloysgrown on the InP substrates are used for the fabricationof field-effect transistors with a 2D electron gas and ahigh electron mobility. Examples of ohmic contacts ton-InP are given in Table 14 [222–236], while examplesof similar contacts to p-InP are given in Table 15[226, 237–248]. In Table 16 [249–257], we give theexamples of ohmic contacts to the InGaAs/InP andInAlAs/InP alloys.

The main mechanism of current flow in ohmic con-tacts to p-InP is the thermionic emission through thebarrier with qϕb < 0.2 eV [258], whereas tunneling rep-Rc, 10–4 Ω cm210820200250300350400T, K

resents the main mechanism of current flow through theAuZn(Ni) contact to heavily doped n-InP (n = 1.4 ×1020 cm–3) [259].

The GaSb compound and alloys on its basis are usedas materials for LEDs and photodetectors in the mid-IRspectral region [260]. The Fermi level at the GaSb sur-face is pinned near the valence-band edge [261]. Exam-ples of ohmic contacts to GaSb and alloys on its basisare given in Table 17 [261–272].

In Table 18 [273–278], we give examples of ohmiccontacts to narrow-gap InAs and HgCdTe semiconduc-tors. The current flow mechanism in these contacts wasstudied inadequately.

4.4. Ohmic Contacts to Semiconductors of Group IVOhmic contacts to Si and Ge have been already stud-ied for several decades and were considered in detail inthe monographs by Milnes and Foight [9] and by Rod-erick [10]; therefore, we will not consider those struc-tures and will restrict ourselves to dealing with ohmiccontacts to new semiconductors, diamond and siliconcarbide.

Diamond is the semiconductor with the widest bandgap among semiconductors used in optoelectronics(Eg = 5.5 eV, E0 ≈ 7.3 eV). Diamond can be used in pho-todetectors for the short-wavelength, especially, thevacuum (ultraviolet) spectral region. Ohmic contacts todiamond are typically formed on the basis of the Ti,Mo, or Ta metals that form carbides in the course ofannealing. Examples of ohmic contacts to diamond aregiven in Table 19 [279–298]. Detailed studies of thecurrent flow mechanisms in these contacts have notbeen carried out. In heavily doped diamond crystals ofp- and n-type, the main current flow mechanism isbelieved to be tunneling [280, 282, 291, 294, 298] onthe basis of the fact that the contact resistance is inde-pendent of temperature.

SEMICONDUCTORS Vol. 41 No. 11 2007

Fig. 13. Dependence of the resistance of the In–n-GaPohmic contact on temperature T. Current flows through themetal shunts [33].

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

Table 14. Ohmic contacts to n-InP

Metal

Au/Ge/NiAu/Ge/NiAu/Ge/PdAu/Pd/Te

Au/Si/Pt

Au/Ru/Au-Ge/NiAu/Pt/Au/GeNi/AuGe/Au

Ni/Au/Pt/Au/Pt/Ti/NiPd/Ge

Ge/Pt/Ge/PtAl/TiTi/Pt/AuW–Sb

W/Sb, W–In–SbWS0.79

Carrier concentra-tion, cm–3

8 × 101810174 × 101 × 1018

550400500300350500600Nonalloyedcontact400

400

1.4 × 1020 (InP : Te)

Annealing tem-perature, °C450250–400350

Contact resis-tance, Ω cm210–810–72.5 × 10–6

10–410–62.77 × 10–5

10–72.15 × 10–6

Note

1281

Refe-rences[226][230][233][231][225][236][232][224][234][233][235][228][222][223][227][229]

Formation of Ni2P + Al10In3Heavy doping with GeFormation of In2Te3

Formation of the Ni–P phase

4.2 × 10–67.71 × 10–73.4 × 10–6

10–610–610–6

Doping with Ge

Formation of Al/Ti/In/Ti–p-InPFormation of InN

P diffuses into W, and the In–Sb phase is formed

Table 15. Ohmic contacts to p-InP

MetalAu/Zn

Au-Be/Ru/AuAu/Zn/Au/Ti/AuNi, Pd/Zn/Ni, PdPd/Zn/Pd/AuPd/Zn/Pd(Pt)Pd/Sb/Zn/PdGe/Pd(Zn)Sb/Zn/PdSb/Zn/Au/NbZn3P2/InP

Carrier concen-Annealing tem-tration, cm–3perature, °C5 × 1018(1–4) × 1018

280

Contact resis-tance, Ω cm27 × 10–6(2–7) × 10–47 × 10–57 × 10–57 × 10–62 × 10–610–5–10–44 × 10–510–5

Formation of AuxIny and Au2P3Formation of Nd2.7InP and Pd2InPFormation of Pd2InP–PdP2Formation of the InSb phaseFormation of Pd2InP

Removal of oxide; diffusion of Au and Zn into InP

The Zn3P/InP heterojunction

Note

References[226][246][242][240][241][237][238][239][247][243][244][245][248]

400500400375325

Silicon carbide is a semiconductor that is activelyused in the fields of high-temperature, high-frequency,and high-power electronics. This semiconductor hasseveral polytypes; the most widely encountered ofthese are 4H-SiC (Eg = 3.23 eV, E0 ≈ 5 eV) and 6H-SiC(Eg = 3.0 eV, E0 ≈ 5.5 eV). Silicon carbide is a semicon-ductor with predominantly covalent bonding and thesurface states caused by the presence of carbon vacan-SEMICONDUCTORS Vol. 41 No. 11 2007

cies. The height of the potential barrier at the metal–SiCinterface depends heavily on the work function for elec-trons that leave the metal [299, 300].

Ohmic contacts to n-4H-SiC and n-6H-SiC are oftenfabricated on the basis of Ni (ΦNi = 5.15–5.2 eV). Thesecontacts can be formed both without annealing andusing a heat treatment, as a result of which oxides and

1282BLANK, GOL’DBERG

Table 16. Ohmic contacts to the InGaAs/InP and InAlAs/InP alloys

MetalTi/Pt/CuAuGe/Ni/AuTiNx

Pt/Zn/Pt/ZrB2/AuW(Zn)PdGe

Pd/Zn/Pd/AuPd/Zn/Pd/AuGe/Ag/Ni

Semiconductor

InGaAsInGaAsInGaAs/InPp-InGaAs/InP

p-InGaAs/InP (p = 1018 cm–3)n-In0.53Ga0.47As/InPp-In0.47Ga0.53As/InP

p-InGaAsP/InP (p = (1–1.5) × 1019 cm–3)n-InAlAs/InP

Annealing tem-perature, °C

350500

5 × 10–6

6.6 × 10–8–1.4 × 10–6

7.5 × 10–62 × 10–62.26 × 10–7Contact resistance,

Ω cm2

References[250][251][253][252][2][255][256][257][249]

425500400425

Table 17. Ohmic contacts to GaSb

MetalAu/Ge/PdAu/Ge/Ni,Ag/Au/Ge/NiPd(Ge, S)

Pd/Ge/Pd/In/PdPd/Ge/Pd

Pd/Ge/Au/Pt/AuPd, Pd/Sb,

Ge/PbGe/Pd/SbPd3In7/Pt(W, WSi2, WSiN)Te–AuTi/Pt/AuAu, AgZn–Au

Ti/Au, Pt/Au,Pd/Au, Ni/AuTi/Pt/Au

Semiconductorn-GaSbn-GaSbn-GaSbn-GaSb

n-GaSb

n-GaInAsSb/GaAsn-GaSbn-GaSbn-GaSb

n-GaAs1 – xSbxp-GaSbp-GaSbp-GaSbp-GaSb

6.6 × 10161.2 × 10185.6 × 10171.8 × 10181.3 × 10183 × 1018

400

Carrierconcentra-tion, cm–3

Annealing temperature,

°C

Contact resis-tance, Ω cm24.9 × 10–68 × 10–4–8 × 10–34 × 10–51.4 × 10–63.8 × 10–62 × 10–6–10–5

2 × 10–6(1–10) × 10–4

10–610–63 × 10–75 × 10–55 × 10–610–52.6 × 10–7

Note

Formation of poly-crystalline AuSb2

References[265][269][267][261][262][263][266][2][268][270][272][268][271][267]

350

300–325

2 × 1018

325–350Formation ofthe n-Ge/n-GaSbheterojunctionTunneling

10181019

250–500250–350

Nonalloyed contact

other layers are removed from the semiconductor sur-face. Examples of ohmic contacts to 4H-SiC are givenin Table 20 [301–328], while Table 21 lists examples ofohmic contacts to 6H-SiC [329–339].

Fabrication of contacts to p-SiC encounters seriousdifficulties due to the wide band gap of the semiconduc-tor. To this end, metals with a high work function(Ni, Pt, Pd, and Au) can be typically used; these metals,as a result of heating to 800–1000°C, form compounds

(of the Ti3SiC2 type) that reduce considerably the height ofthe metal–semiconductor barrier (see Table 20).The mechanism of current flow through ohmic con-tacts to silicon carbide was studied recently [314, 316,334, 340]. On the basis of temperature dependences ofthe contact resistance, it was assumed that the mainmechanism of current flow is the thermionic emission.The heights of the metal–SiC potential barrier as deter-SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

Table 18. Ohmic contacts to InAs and HgCdTe

Metal

Ti/Pt/Au

Au/Ge, Au/Te

graded-gap

Super-latticeconventionalTiPd/Pt/AuTi

Pd/Pt/Au

Semiconductorn-InAsn-InAs

n-InAs–n-GaAs

Carrier concen-Annealing tem-tration, cm–3perature, °C

350

325Nonalloyed contact

Contact resis-tance, Ω cm29.8 × 10–72.3 × 10–75 × 10–810–8–10–5

HgCdTep-InAsAlSb/InAs

175Nonalloyed contact

10–49.6 × 10–72.6 × 10–7

The barrier height 0.14–0.26 eVThermal field emission

Note

1283

Refe-rences[273][274][275]

1019

[276][277][278]

Table 19. Ohmic contacts to diamond

MetalAu, CuTi, AuAl2O3AuAu/TiAu/TiAu/TiAu/TiAu/Pt/TiAu/Pt/TiAu/TaAl/SiAl/Si

Al/Si, Ti/Au,TiWN/AuTi, Mo

Ti/AuTi/Au

TiC/Au, TaSi2/AuPt/Ti/Au

Semicon-ductorn-C : Bn+-C : BC nanotubesp-Cp-Cp-Cp-Cp-C : Bp-Cp-C : Bp-C : Bp-C : Bp-C : Bp-Cp-Cp-C : Bp-C : Bp-C : Bp-C : B

3 × 1018–3 × 1020

7 × 1020Carrier concen-tration, cm–3

5 × 1017

Tunneling

<10–31.2 × 10–61.6 × 10–62 × 10–410–41.3 × 10–5

Formation of TaC

450450

Tunneling

10–7

Tunneling, Rc decreases ~ 1/T

Formation of α-Mo2C, field and thermal field emissionFormation of TiC

Annealing tem-Contact resis-perature, °Ctance, Ω cm2

Note

Refe-rences[279]

[280][281][285][290][291][292][298][293][286][288][2][294][297][282][296][283][284][295][287]

1020

500

1020

500600

Formation of TiCTunneling

Formation of TiCTunneling

400–600850850850450

As low as 10–610–6–10–5

10–510–510–4

1018

mined from the above dependences differed by an orderof magnitude: from 0.097 eV for 4H-SiC [316] to0.97 eV for 6H-SiC [334]. In [314], temperature depen-dences of the contact resistance were not exactly con-sistent with the thermionic-emission theory, so it wasassumed [314] that, in these cases, the main currentflow mechanism is thermal field emission.

SEMICONDUCTORS Vol. 41 No. 11 2007

Crofton et al. [340] studied the dependence of thecontact resistance on the hole concentration in 4H-SiC.Experimental data are consistent with theoreticalresults if we assume that the tunneling effective massand the effective mass of the density of states are equalto the mass of a free electron; in this case, the height ofthe barrier overcome by electrons was found to be equal

1284

Table 20. Ohmic contacts to 4H-SiC

MetalNiNiNi

Ni

Ni/Ti/Au

Ni, Ni/W, Ni/Ti/WNi/Cr/W, Cr/MoWCr

Pt/Ti/WSi/NiCo/Si/CoAl/TiAl/TiAl/SiAl/SiAl/SiAl/Ti/Ni

Au/Al/Si, Au/Ti/AlAu/Ti/Au, Ti/AuAu/Ti/Al, Au/Pd/AlTi/AlTi/Al/GeNi/Ti/AuPdPdSi/PtGe/Ti/Au

BLANK, GOL’DBERG

Carrier concen-Annealing

tration (cm–3) or

temperature,

the conductivity

°C

typennnn+

n = 1019nnnnp

p+ (>1020)ppp

p = 1019pppp

p

p = 4.5 × 1018pppp

950–1000800Withoutannealing10001000

Contactresistance,Ω cm21.3 × 10–5

Note

The use of a Si layer

Refe-rences[304][305][306][309][310][320][311][303][318][301][307][308][313][327][314][315][317][316][322][323][325][319][324][320][312][326][328][321]

Formation of the Ni31Si12 and Ni2Si layers and formation of donor-like C vacancies; the barrier height 0.38 eV

10–610–4

Formation of Ni2Si or a Cr3C2 layer

8001000–1050

1.5 × 10–610–410–43.8 × 10–59.6 × 10–59.5 × 10–59 × 10–5

Formation of a CoSi2 layer

Formation of the Ni2Si and TiC layersThermionic emission

Formation of Al4C3

Thermal–field emission through the 0.097-eV barrier

750700–950950

7001000850–9501000600800750600–7001100800

(1.4–8.3) × 10–5(2–7) × 10–5Formation of Ti3SiC2

10–5Formation of Ti3SiC210–5Formation of Al3Ti, Ti3SiC2, Al4C3,

and Ti3Si3

10–410–35.52 × 10–5

Formation of PdSi25.5 × 10–5

10–4Diffusion of C into metal layers10–4Formation of Ti3SiC2

to 0.37 eV (Fig. 14a). Temperature dependences of the

contact resistance [334] showed that the main currentflow mechanism in p-4H-SiC and p-6H-SiC (p ≈1019 cm–3) is thermal field emission (Fig. 14b).

5. MAIN CONCLUSIONS CONCERNING THE CURRENT FLOW MECHANISMS5.1. From the standpoint of the formation of ohmiccontacts, all main semiconductors can be separated intotwo groups:(i) Semiconductors with a low density of surfacestates located deep in the band gap (for example, ZnSe,GaN, and SiC) or with the surface states located in theconduction band. Ohmic contacts to these semiconduc-tors can be obtained by choosing the work function forelectrons Φm lower than the electron affinity Xs for then-type semiconductors (Φm < Xs) or by choosing thecontact metals with the electrons’ work function Φmlarger than the sum of the band gap Eg and the electronaffinity Xs for the p-type semiconductors (Φm > Eg + Xs).If such metals are available, fabrication of the ohmic

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

Table 21. Ohmic contacts to 6H-SiC

Metal

Ni, Ni/W, Ni/Ti/W, Ni/Cr/W, Cr/MoWNiSiTiSixRe

W/WC/TaCAl/TiAl/TiPt, W

W/Ti, Al/TiTiN

Carrier concen-Annealing tem-tration, cm–3perature, °Cn

n

n = 5 × 1019n = 1.28 × 1018nppppp

1000–1050600900–11001000100010001000No annealing

10–5

(4–7) × 10–67 × 10–52 × 10–42.8 × 10–44.4 × 10–5Contact resis-tance, Ω cm2

Note

Formation of Ni2Si or a Cr3C2 layer

1285

References[333][330][332][329][331][336][335][338][337][339]

contacts amounts to the removal of the near-surface dam-aged or oxide layer (for example, Ti/Pt/Au–n-ZnSe).If such metals are not available, fabrication of these con-tacts consists in formation of a compound with a low workfunction between the metal and semiconductor (for exam-ple, the Si/Ti–n-GaN structure with formation of the tita-nium oxide in the course of heat treatment).

(ii) Semiconductors with a high density of surfacestates located near the midgap (for example, Si, Ge,GaAs, GaP, and InP).

In these materials, the Fermi level is pinned at thesurface and the work function for electrons leaving thecontacting metal affects only slightly the contact’sproperties.

Fabrication of ohmic contacts to these semiconduc-tors amounts to either heavy doping of the near-surfaceregion, which ensures the tunneling-like passage ofelectrons through the interface (for example, the levelof doping of GaAs with Ge is as high as 1021 cm–3 in theNi/Au + Ge/Ni/Au–n-GaAs contact); or formation ofchemical compounds in the near-surface region thatreduce appreciably the height of the metal–semicon-ductor potential barrier, which makes it possible forelectrons to pass the interface due to thermionic emis-sion (for example, in the In–GaAs structure); or passi-vation of the semiconductor surface, which leads to adecrease in the density of surface states (for example,the treatment of the GaAs surface in (NH4)2Sx bringsabout a decrease in the density of surface states by anorder of magnitude).

5.2. Nonalloyed ohmic contacts (thin-film contactsand the contacts formed at low temperatures) can beconsidered as Schottky barriers with a low or thinpotential barrier. Therefore, the following current flowmechanisms are characteristic of the Schottky diodes:(I) thermionic emission (the contact resistancedecreases exponentially with increasing temperatureand the height of the metal–semiconductor potentialbarrier);

SEMICONDUCTORS Vol. 41 No. 11 2007

(II) field emission and tunneling (the contact resis-tance decreases exponentially as the doping level isincreased and is virtually independent of temperature);and

Rc, Ω × cm210–110–210–310–410–510–6151010–410161017101810191020Na, cm–3(a)(b)4H-SiC6H-SiC10–52110–6050100150200250300T, °C

Fig. 14. Dependences of the ohmic-contact resistance Rc on(a) the concentration Na of uncompensated acceptors at 300 Kfor Al/Ti–p-4H-SiC [340] and (b) temperature T for theAl/Ti–p-6H-SiC (1) and Al/Ti–p-4H-SiC contacts (2). Themechanism of the current flow is thermionic emission (qϕb =0.53 eV for 6H-SiC and qϕb = 0.82 eV in 4H-SiC) [334].

1286BLANK, GOL’DBERG

decreases as temperature is increased; however, this(III) thermal field emission (the contact resistancedecrease is much slighter than in the case of thermionicemission).

lightly doped Thermionic emission manifests itself in contacts top-n- and p-GaAs, n-changed in the near-contact region. This mechanismInGaN. p-InP, and0.53Ga0.47As, in which chemical composition waswas identified on the basis of exponential decrease inthe contact resistance as temperature was increased.The height of the metal–semiconductor potential bar-rier determined from comparison of theoretical resultswith experimental data was found to be equal to 0.068–0.09 eV for GaAs, 0.13 eV for GaN, lower than 0.2 eVfor InN, and 0.13 eV for InGaAs. These values aremuch smaller than the height of the Shottky potentialbarrier at the metal–semiconductor interface or theheight of the potential barrier at the free semiconduc-tor’s surface.

tures based on GaAs, InP, GaN, AlGaN, and SiC, inTunneling (field emission) manifests itself in struc-which the surface region was subjected to heavy dop-ing. This mechanism was identified on the basis ofindependence of the contact resistance on temperatureand exponential decrease in the contact resistance asthe charge-carrier concentration was increased. Thedoping level in the near-contact region was as high as1020rowing of the metal–semiconductor potential barrier. In–1021 cm–3, which resulted in an appreciable nar-this case, the potential barrier height was fairly high:0.3–0.5 eV for for SiC structures. In this case, the contact resistancep-GaAs, ~0.5 eV for GaN, and ~0.4 eVwas fairly low due to a small thickness of the potentialbarrier.

case of semiconductors with a high density of imper-5.3. For alloyed ohmic contacts (especially in thefections), another mechanism, noncharacteristic of theSchottky diodes, can manifest itself; this mechanismconsists in the current flow over the metal shunts. Basedon experimental data on the increase in the resistance ofohmic contacts with increasing temperature, this mech-anism was verified for In contacts to lightly doped GaPand GaN. The number of shunts (per unit contact area)determined from the temperature dependence of thecontact resistance correlated well with the dislocationdensity in the near-surface region of the semiconductor(the density of shunts was 107–108 cm–2 for GaN and(4.5–8) × 107 cm–2mechanism becomes unimportant at low dislocation for GaP). At the same time, thisdensity (for example, for GaAs); as a result, other (tra-ditional) mechanisms manifest themselves).

ACKNOWLEDGMENTS

sion of this study.

We thank Professor O.V. Konstantinov for discus-REFERENCES

1.W. Schottky, Z. Phys. B 118, 539 (1942).2.J. Bardeen, Phys. Rev. 71, 717 (1947).

3.W. E. Spicer, I. Lindau, P. Skeath, and C. YSci. Technol. 17, 1019 (1980).

. Si, J. Vac.4.A. K. Henish, don, Oxford, 1957).

Rectifying Semiconductor Contacts (Clare-5.A. L. Fahrenbruch and R. H. Bube, Solar Cells (Academic, New York, 1987; Énergoatom-Fundamentals ofmash, Moscow, 1987).

6.Yu. A. Gol’dberg, Fiz. Tekh. Poluprovodn. (St. Peters-burg) 28, 1681 (1994) [Semiconductors 28, 935 (1994)].7.R. H. Cox and H. Strack, Solid-State Electron. (1969).

12, 8.G. K. Reeves, Solid-State Electron. 23, 477 (1980).9.A. G. Milnes and J. J. Feucht, Metal–Semiconductor Junctions (Academic, New York,Heterojunctions and1972; Mir, Moscow, 1975).

10.E. H. Rhoderick, endon, Oxford, 1978; Radio i Svyaz’, Moscow, 1982).Metal–Semiconductor Contacts (Clar-11.C. A. Mead, Solid-State Electron. 9, 1023 (1966).

12.S. Sze, (Wiley, New York, 1981; Mir, Moscow, 1984).

Physics of Semiconductor Devices, 2nd ed.

13.A. Peotrowska, Thin Solid Films 193–194, 511 (1990).14.K. Ip, G. T. Thaler, Hyucksoo Yang, et al., J. Cryst.

Growth 287, 149 (2006).

15.W. Faschinger, J. Cryst. Growth 197, 557 (1999).

16.M. Murakami and YMater. Sci. 23, 1 (1998).

. Koide, Crit. Rev. Solid State

17.L. M. Porter and R. F. Davis, Mater. Sci. Eng. B (1995).

34, 83

18.K. Das, VFilms 212, 19 (1992).

. Venkatesan, K. Miyata, et al., Thin Solid

19.Wu Dingfen and K. Heime, Electron. Lett. (1982).

18, 940

20.Wu Dingfen, W. Dening, and Kl. Heime, Solid-State

Electron. 29, 4 (1987).

21.R. K. Kupka and W. A. Anderson, J. Appl. Phys. 3623 (1991).

69,

22.F. A. Padovani and R. Stratton, Solid-State Electron. 695 (1966).

9,

23.YPhys. .-C. Shic, M. Murakami, and W. H. Price, J. Appl.

65, 3539 (19).

24.C.-F. Lin, D. B. Ingerly, and YLett. 69, 33 (1996).

. A. Chang, Appl. Phys.

25.R. K. Kupka and W. A. Anderson, J. Appl. Phys. (1991).

69, 362

26.M. Shur, York, 1987; Mir, Moscow, 1991).

GaAs Devices and Circuits (Plenum, New

27.A. Y. C. Yu, Solid-State Electron. 13, 239 (1970).

28.C. YElectron. . Chang, Y14, 1 (1971).

. K. Fang, and S. M. Sze, Solid-State

29.S. M. Cho, J. D. Lee, and H. H. Lee, J. Appl. Phys. 282 (1991).

70,

30.F. A. Padovani and R. Stratton, Solid-State Electron. 695 (1966).

9,

31.Nien-Po Chen, H. J. Ueng, D. B. Janes, et al., J. Appl.

Phys. 88, 309 (2000).

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

1287

32.D. Qiao, L. S. Yu, L. Jia, et al., Appl. Phys. Lett. (2002).

80, 992

33.T. VPis’ma Zh. Tekh. Fiz. . Blank, Yu. A. Gol’dberg, O. V. Konstantinov, et al.,

Lett. 30, 806 (2004)].

30 (19), 17 (2004) [Tech. Phys.34.T. VFiz. Tekh. Poluprovodn. (St. Petersburg) . Blank, Yu. A. Gol’dberg, O. V. Konstantinov, et al.,

(2006) [Semiconductors 40, 1173 (2006)].

40, 120435.J. Narayan, P. Tiwari, X. Chen, et al., Appl. Phys. Lett.

61, 1290 (1992).

36.E. J. Miller, D. M. Schaadt, E. T. 94, 7611 (2003).

Yu, et al., J. Appl. Phys.

37.E. J. Miller, E. T. Yu, P. Waltereit, and J. S. Speck, Appl.

Phys. Lett. 84, 535 (2004).

38.Chin-Yuan Hsu, Wen-How Lan, and Yew Chung Ser-mon Wu, Jpn. J. Appl. Phys., Part 1 44, 7424 (2005).39.L. Wang, F. M. Mohammad, and J. Adeside, Appl. Phys.

Lett. 87, 14915 (2005).

40.Yu. A. Gol’dberg, E. A. Posse, and B. VTekh. Poluprovodn. (Leningrad) Phys. Semicond. 20, 947 (1986)].

20, 1510 (1986) [Sov.. Tsarenkov, Fiz.

41.B. VE. . Tsarenkov, Yu. A. Gol’dberg, M. I. Il’ina, and

555 (1988) [Sov. Phys. Semicond. A. Posse, Fiz. Tekh. Poluprovodn. (Leningrad) 22, 342 (1988)].22,42.Yu. A. Gol’dberg and E. A. Posse, Zh. Tekh. Fiz. 61 (2001) [Tech. Phys. 46, 1128 (2001)].

71 (9),

43.P. Visconti, K. M. Jones, M. A. Reshchikov, et al., Appl.

Phys. Lett. 77, 3532 (2000).

44.Handbook Series on Semiconductor ParametersM. Levinshtein, S. Rumyantsev, and M. Shur (World, Ed. by

Sci., Singapore, 1996 and 1999), Vols. 1 and 2.

45.Properties of Advanced Semiconductor Materials:

GaN, AlN, InN, BN, SiC, SiGe, Ed. by M. E. Levinsh-tein, S. L. Rumyantsev, and M. S. Shur (Wiley, New Y2001).

ork,46.J. D. Guo, M. S. Feng, R. J. Guo, et al., Appl. Phys. Lett.

67, 2657 (1995).

47.A. N. Andreev, A. A. Lebedev, M. G. Rastegaeva, et al.,

Fiz. Tekh. Poluprovodn. (St. Petersburg) (1995) [Semiconductors 29, 957 (1995)].

29, 183348.M. Sawada, T. Sawada, Yings of Second International Conference on Nitride. Yanagata, et al., in Proceed-Semiconductors (Tokushino, Japan, 1997), p. 706.

49.Vprovodn. (St. Petersburg) . Yu. Davydov and A. A. Klochikhin, Fiz. Tekh. Polu-tors 38, 861 (2004)].

38, 7 (2004) [Semiconduc-50.Lu-Min Liu, Gr. Lindauer, W. Brock Alexander, and

P. H. Holloway, J. Vac. Sci. Technol. B 13, 2236 (1995).51.J. J. Fijol, J. T. Trexier, L. Calhoun, et al., J. Vac. Sci.

Technol. B 14, 159 (1996).

52.T. Miyajima, H. Okuyama, and K. Akimoto, Jpn. J.

Appl. Phys. 31 (12B), L1743 (1992).

53.K. Akimoto, T. Miyajima, H. Okuyama, and YJ. Cryst. Growth 115, 683 (1991).

. Mori,

.Y(1992).

. X. Wang and P. H. Holloway, Vacuum 43, 1149

55.M. R. Park, W. A. Anderson, M. Jeon, and H. Luo,

Solid-State Electron. 43, 113 (1999).

56.R. G. Dandrea and C. B. Duke, Appl. Phys. Lett. 2145 (1994).

,

SEMICONDUCTORS Vol. 41 No. 11 2007

57.P. M. Mensz, Appl. Phys. Lett. , 2148 (1994).

58.F. Vigué, P. Brunet, P. Lorenzini, et al., Appl. Phys. Lett.

75, 3345 (1999).

59.R. Schwarz, Th. Studnitzky, Fr. Goesmann, and R. Schmid-Fetzer, Solid-State Electron. 42, 139 (1998).

60.S. Einfeldt, H. Heinke, M. Behringer, et al., J. Cryst.

Growth 138, 471 (1994).

61.Y25 (1992).

. Lansari, J. Ren, B. Sneed, et al., Appl. Phys. Lett. 61,

62.F. Hiei, M. Ikeda, M. Ozawa, et al., Electron. Lett. 878 (1993).

29,

63.K. Akimoto, T. Miyajima, H. Okuyama, and YJ. Cryst. Growth 115, 683 (1991).

. Mori,

.J. J. Chen, T. J. Anderson, S. Jang, et al., J. Electrochem.

Soc. 153, G462 (2006).

65.B. Ghosh, N. K. Mondal, P. Banerjee, et al., IEEE Trans.

Electron Devices 49, 2352 (2002).

66.Z. Yang and J. F. Schetzina, J. Electron. Mater. (1994).

23, 1071

67.S. Miller and P. H. Holloway, J. Electron. Mater. 1709 (1996).

25,

68.J. Rennie, M. Onomura, S. Nunoue, et al., in ings of Second International Conference on NitrideProceed-Semiconductors (Tokushino, Japan, 1997), p. 711.

69.D.-W. Kim and H. K. Baik, Appl. Phys. Lett. (2000).

77, 1011

70.H. Morkoç, S. Strike, G. B. Gao, et al., J. Appl. Phys.

76, 1363 (1994).

71.M. E. Lin, Z. Ma, F. Y, 1003 (1994).

. Huang, et al., Appl. Phys. Lett.

72.L. Wang, F. M. Mochammed, and I. Adesida, J. Appl.

Phys, 101, 013702 (2007).

73.T. Gessmann, YLett. 80, 986 (2002).

. L. Li, E. L. Waldron, et al., Appl. Phys.

74.P. H. Holloway, T. J. Kim, J. T. Trexler, et al., Appl. Surf.

Sci. 117–118, 362 (1997).

75.C. R. Abernathy, J. D. MacKenzie, S. J. Pearton, and

W. S. Hobson, Appl. Phys. Lett. 66, 1969 (1995).

76.H. W. Jang, W. Urbanek, M. C. Yoo, and J.-L. Lee, Appl.

Phys. Lett. 80, 2937 (2002).

77.H. Morkoç, S. Strike, G. B. Gao, et al., J. Appl. Phys.

76, 1363 (1994).

78.S. Prakashs, L. S. Tan, K. M. Ng, et al., in International Conference on SiC and Related MaterialsAbstracts of

(Sheraton, 1999), p. 48.

79.J. K. Kim, J. H. Je, J.-L. Lee, et al., J. Electrochem. Soc.

148, 45 (2000).

80.Z. Fan, S. N. Mohammad, W. Kim, et al., Appl. Phys.

Lett. 68, 1672 (1996).

81.C.-YElectron. . Chang, T.-W. Lan, G.-C. Chi, et al., Solid-State

9, G155 (2006).

82.R. Khanna, S. J. Pearton, F. Ren, et al., Appl. Surf. Sci.

252, 1826 (2005).

83.D.-W. Kim, J. C. Bae, W. J. Kim, et al., J. Electron.

Mater. 30, 855 (2001).

84.D.-W. Kim and H. K. Baik, J. Appl. Phys. (2000).

77, 1011

85.M. L. Schuette and Wo Lu, J. Vac. Sci. Technol. 3143 (2005).

23,

1288BLANK, GOL’DBERG

86.J. K. Sheu, Ytron. 43, 2081 (1999).

. K. Su, G. C. Chi, et al., Solid-State Elec-87.J.-L. Lee and J. K. Kim, J. Electrochem. Soc. (2000).

147, 2297

88.J.-S. Jang, C.-W. Lee, S.-J. Park, et al., J. Electron.

Mater. 31, 903 (2002).

.A. Lunev, Vof International Conference on SiC and Related Mate-. Chaturvedi, A. Chitmis, et al., in Abstracts

rials (Sheraton, 1999), p. 159.

90.K. Kumakura, T. Makimoto, and N. Kobayashi, Appl.

Phys. Lett. 79, 2588 (2001).

91.L. S. Yu, L. Jia, D. Qiao, et al., IEEE Trans. Electron

Devices 50, 292 (2003).

92.Th. Gessmann, Ytron. Mater. 31, 416 (2002).

.-L. Li, E. L. Waldron, et al., J. Elec-93.D. B. Ingerly, YJ. Nitride Semicond. Res. . A. Chang, and Y4S1 (G6), 49 (1999).

. Chen, MRS Internet

94.J. S. Jang, H. G. Kim, K. H. Park, et al., Mater. Res. Soc.

Symp. Proc. 482, 1653 (1998).

95.J.-K. Ho, C.-S. Jong, C. C. Chiu, et al., Appl. Phys. Lett.

74, 1275 (1999).

96.D. B. Ingerly, YLett. 74, 2480 (1999).

. A. Chang, and Y. Chen, Appl. Phys.

97.L.-C. Chen, F.-R. Chen, J.-J. Kai, et al., J. Appl. Phys.

86, 3826 (1999).

98.J.-K. Ho, C.-S. Jong, C. C. Chiu, et al., J. Appl. Phys.

86, 4481 (1999).

99.K.-M. Chang, J.-YElectron. 49, 1381 (2005).

. Chu, and C.-C. Cheng, Solid-State

100.J.-S. Jang, S. J. Park, and T.-Y76, 28 (2000).

. Seong, Appl. Phys. Lett.

101.M. Suzuki, T. Arai, T. Kawakami, et al., J. Appl. Phys.

86, 5079 (1999).

102.R. W. Chuang, A. Q. Zou, and H. P. Lee, MRS Internet

J. Nitride Semicond. Res. 4S1 (G6), 42 (1999).

103.C.-F. Chu, C. C. Yu, Y77, 3423 (2000).

. K. Wang, et al., Appl. Phys. Lett.

104.K. Kumakura, T. Makimoto, and N. Kobayashi, Phys.

Status Solidi A 188, 363 (2001).

105.J.-O. Song, J. S. Kwak, and T.-YTechnol. 21, L7 (2006).

. Seong, Semicond. Sci.

106.J. D. Guo, C. I. Lin, M. S. Feng, et al., Appl. Phys. Lett.

68, 235 (1996).

107.C. Lu, H. Chen, X. Lv, et al., J. Appl. Phys. (2002).

91, 9218

108.S. N. Mohammad, J. Appl. Phys. 95, 7940 (2004).

109.K. Suzue, S. N. Mohammad, Z. F. Fan, et al., J. Appl.

Phys. 80, 4467 (1996).

110.J.-S. Jang and T.-Y(2000).

. Seong, Appl. Phys. Lett. 76, 2743

111.A. Katz, S. Nakahara, W. Savin, and B. E. Weir, J. Appl.

Phys. 68, 4133 (1990).

112.J. S. Kwak, O.-H. Nam, and Y5917 (2004).

. Park, J. Appl. Phys. 95,

113.J. Yun and K. Choi, IEEE Electron Device Lett. (2006).

27, 22

114.D. Qiao, Z. F. Guan, J. Carlton, et al., Appl. Phys. Lett.

74, 2652 (1999).

115.M. K. Guse(1996) [Tech. Phys. Lett. œnov, Pis’ma Zh. Tekh. Fiz. 22, 514 (1996)].

22 (12), 75

116.H.-Y406 (2006).

. Cha, X. Chen, H. Wu, et al., J. Electron. Mater. 35,

117.F. M. Mohammed and I. Adesida, Appl. Phys. Lett. 141915 (2005).

87,

118.C. H. Kuo, J. K. Sheu, G. C. Chi, et al., Solid-State Elec-tron. 45, 717 (2001).

119.YLett. .-L. Li, E. F. Schubert, J. W. Graff, et al., Appl. Phys.

76, 2728 (2000).

120.T. VTekh. Poluprovodn. (St. Petersburg) . Blank, Yu. A. Gol’dberg, E. V. Kalinina, et al., Fiz.

[Semiconductors 35, 529 (2001)].

35, 550 (2001)121.Yu. A. Gol’dberg, T. VB. 22, 1712 (1988) [Sov. Phys. Semicond. V. Tsarenkov, Fiz. Tekh. Poluprovodn. (Leningrad). L’vova, R. V. Khasimova, and

22, 1081 (1988)].122.F. Ren, C. R. Abernathy, S. J. Pearton, and P. W. Wisk,

Appl. Phys. Lett. , 1508 (1994).

123.C. R. Abernathy, S. J. Pearton, F. Ren, and P. W. Wisk,

J. Vac. Sci. Technol. B 11, 179 (1993).

124.H.-K. Kim, J.-S. Jang, S.-J. Park, and T.-YJ. Electrochem. Soc. 147, 1573 (2000).

. Seong,

125.F. Ren, J. Vac. Sci. Technol. A 15, 802 (1997).

126.C. B. Vartuli, S. J. Pearton, C. R. Abernathy, et al.,

Solid-State Electron. 41, 531 (1997).

127.C. B. Vartuli, S. J. Pearton, C. R. Abernathy, et al.,

J. Vac. Sci. Technol. B 14, 3520 (1996).

128.M. S. Carpenter, M. R. Melloch, M. S. Lundstrom, and

S. P. Tobin, Appl. Phys. Lett. 52, 2157 (1988).

129.M. S. Carpenter, M. R. Melloch, and T. E. Dungan,

Appl. Phys. Lett. 53, 66 (1988).

130.YAppl. Phys. . Nannichi, J.-F. Fan, H. Oigawa, and A. Koma, Jpn. J.

27, L2367 (1988).

131.J.-F. Fan, H. Oigawa, and YPhys. 27, L2125 (1988).

. Nannichi, Jpn. J. Appl.

132.S. Logha, D. B. Janes, and N. P. Chen, Appl. Phys. Lett.

80, 4452 (2002).

133.C.-T. Lee, Y2573 (2001).

.-J. Lin, and D.-S. Liu, Appl. Phys. Lett. 79,

134.J. O. Song, S.-J. Park, and T.-Y80, 3129 (2002).

. Seong, Appl. Phys. Lett.

135.G. YAbstracts of E-MRS Spring Meeting, Symposium F:. Chung, C. C. Tin, J. R. Williams, et al., in

Amorphous and Crystalline Silicon Carbide, Materialand Applications (Strasbourg, France, 2001), F7.

136.J. Massies, J. Chaplart, M. Lavirou, and N. T. Linh,

Appl. Phys. Lett. 38, 693 (1988).

137.T. J. Kim, P. H. Holloway, and E. A. Kenik, Appl. Phys.

Lett. 71, 3835 (1997).

138.Ylegari, J. Appl. Phys. .-C. Shih, M. Murakami, E. L. Wilkie, and A. C. Cal-62, 582 (1987).

139.M. Otsudo, H. Kumabe, and H. Miki, Solid-State Elec-tron. 20, 617 (1977).

140.A. Christou, Solid-State Electron. 22, 141 (1979).

141.C. L. Chen, L. J. Mahoney, M. C. Finn, et al., Appl.

Phys. Lett. 48, 535 (1986).

142.Yu. A. Gol’dberg and B. Vprovodn. (Leningrad) cond. 3, 1447 (1969)].

3, 1718 (1969) [Sov. Phys. Semi-. Tsarenkov, Fiz. Tekh. Polu-SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

12

143.H. J. Ueng, N.-P. Chen, D. B. Janes, et al., J. Appl. Phys.

90, 5637 (2001).

144.N. Okamoto, T. Takahashi, H. Tanaka, and M. Taki-kawa, Jpn. J. Appl. Phys. 37 (6A), 3248 (1998).

145.M. S. Islam, Pbert, J. Mater. Process. Technol. . J. McNally, D. C. Cameron, and P77, 42 (1998).

. A. F. Her-146.J. Zhou, G. Xia, B. Li, et al., Jpn. J. Appl. Phys. 2609 (2003).

42 (5A),

147.H.-C. Lin, S. Senenayake, K.-YTrans. Electron Devices 50, 880 (2003).

. Cheng, et al., IEEE

148.J. Zhou, G. Xia, B. Li, and W. Liu, Appl. Phys. A 939 (2003).

76,

149.J.-W. Lim, J.-K. Mun, S.-J. An, et al., Jpn. J. Appl. Phys.

39 (5A), 26 (2000).

150.Y76 (2002).

. Tsunoda and M. Murakami, J. Electron. Mater. 31,

151.H. J. Ueng, N.-P. Chen, D. B. Janes, et al., J. Appl. Phys.

90, 5637 (2001).

152.J.-W. Lim, J.-K. Mun, S. Nam, et al., J. Phys. D 1611 (2000).

33,

153.M. S. Islam, P. J. McNally, A. H. M. Zahirul Alam, and

M. Q. Huda, Solid-State Electron. 44, 655 (2000).

1.F. A. Amin, A. A. Rezazadeh, and S. W. Bland, Mater.

Sci. Eng. B 66, 194 (1999).

155.J.-W. Lim, J.-K. Mun, M.-H. Kwak, and J.-J. Lee,

Solid-State Electron. 43, 13 (1999).

156.N. E. Lumpkin, G. R. Lumpkin, and M. G. Blackford,

J. Mater. Res. 14, 1261 (1999).

157.N. Okamoto, T. Takahashi, and H. Tanaka, Appl. Phys.

Lett. 73, 794 (1998).

158.D. W. Davies, D. VSci. Technol. 14, 615 (1999).

. Morgan, and H. Thomas, Semicond.

159.J.-W. Lim, J.-K. Mun, and J.-J. Lee, Appl. Surf. Sci. 208 (1999).

85,

160.J. S. Kwak, H. K. Baik, J.-L. Lee, et al., Thin Solid

Films 291, 497 (1996).

161.J. S. Kwak, J.-L. Lee, and H. K. Baik, IEEE Electron

Device Lett. 19, 481 (1998).

162.Y1310 (1998).

. G. Wang, D. Wang, and D. G. Ivey, J. Appl. Phys. 84,

163.M. S. Islam and P. J. McNally, Thin Solid Films 253 (1998).

320,

1.M. S. Islam, Pbert, J. Mater. Sci. Technol. . J. McNally, D. C. Cameron, and P77, 42 (1998).

. A. F. Her-165.M. S. Islam and P. J. McNally, Mictroelectron. Eng. 35 (1998).

40,

166.M. S. Islam, Pbert, Thin Solid Films . J. McNally, D. C. Cameron, and P291, 417 (1996).

. A. F. Her-167.M. O. Aboelfotoh, S. Oktyabrsky, J. Narayan, and

J. M. Woodall, J. Mater. Res. 12, 2325 (1997).

168.T. Oku, M. Furumai, C. J. Uchibori, and M. Murakami,

Thin Solid Films 300, 218 (1997).

169.D. G. Ivey, S. Eicher, S. Wingar, and T. P. Lester,

J. Mater. Sci. 8, 63 (1997).

170.C. J. Uchibori, Y117–118, 347 (1997).

. Ohtani, T. Oku, et al., Appl. Surf. Sci.

171.C. J. Uchibori, YMater. 26, 410 (1997).

. Ohtani, T. Oku, et al., J. Electron.

SEMICONDUCTORS Vol. 41 No. 11 2007

172.D. Y81, 297 (1997).

. Chen, Y. A. Chang, and D. Swenson, J. Appl. Phys.

173.T. Lee, N.-P. Chen, J. Liu, et al., Appl. Phys. Lett. 212 (2000).

76,

174.I. G. Akdogan and M. A. Parker, Electrochem. Solid-State Lett. 8, G106 (2005).

175.C.-H. Wu, S.-M. Liao, and K.-C. Chang, Mater. Sci.

Eng. B 117, 205 (2005).

176.Y166. Shiraishi and H. Shimawali, Int. Phys. Conf. Ser.

, 301 (2000).

177.M. O. Aboelfotoh, M. A. Borek, and J. Narayan, Appl.

Phys. Lett. 75, 3953 (1999).

178.M. Ogura and M. Murakami, J. Electron. Mater. (1998).

27, L

179.K. A. Jones, M. W. Cole, W. Y82, 1723 (1997).

. Han, et al., J. Appl. Phys.

180.B. M. Henry, A. E. Staton-Bevan, VM. A. Crouch, Appl. Surf. Sci. 108, 485 (1997).

. K. M. Sharma, and

181.R. Vtron. Adv. Mater. . Ghita, C. Logofatu, C. Negrila, et al., J. Optoelec-7, 3033 (2005).

182.VFiz. Tekh. Poluprovodn. (St. Petersburg) . Kazlauskene, V. Kazukauskas, Yu. Miskinis, et al.,

[Semiconductors 38, 78 (2004)].

38, 78 (2004)183.I. H. Kim, Mater. Lett. 57, 2769 (2003).184.I. H. Kim, Mater. Lett. 56, 775 (2002).

185.J.-M. Lee, I.-H. Choi, S. H. Park, et al., J. Korean Phys.

Soc. 37, 43 (2000).

186.I.-H. Kim, S. H. Park, J. Kim, et al., Jpn. J. Appl. Phys.

37 (3B), 1348 (1998).

187.S. H. Park, I. H. Kim, T. W. Lee, and M. P. Park, Int.

Phys. Conf. Ser. 156, 459 (1998).

188.S. H. Park, T.-W. Lee, J. M. Lee, et al., J. Korean Phys.

Soc. 34 (Suppl. 3), S273 (1999).

1.I.-H. Kim, S. H. Park, T.-W. Lee, and M.-P. Park, Appl.

Phys. Lett. 71, 18 (1997).

190.E. Nebauer, M. Mai, E. Richter, and J. Würfl, J. Elec-tron. Mater. 27, 1372 (1998).

191.I. H. Kim, S. H. Park, T. W. Lee, et al., Int. Phys. Conf.

Ser. 156, 455 (1998).

192.D. G. Ivey, R. Zhang, Z. Abid, et al., J. Mater. Sci. (1997).

8, 281

193.B. Luo, G. Dang, A.-P. Zhang, et al., J. Electrochem.

Soc. 148, G676 (2001).

194.F. Ren, S. J. Pearton, W. S. Hobson, et al., Appl. Phys.

Lett. 58, 1158 (1991).

195.YPhys. Lett. .-T. Kim, J.-L. Lee, J. K. Mun, and H. Kim, Appl.

71, 2656 (1997).

196.YPhys. . Shiraishi, N. Furuhata, and A. Okamoto, J. Appl.

76, 5099 (1994).

197.S. Vand J. Würfl, Appl. Phys. Lett. . Morozov, Yu. V. Dubrovskii, V72, 2882 (1998).

. N. Abrosimova,

198.J. W. Lim, J. K. Mun, M. H. Kwak, and J. J. Lee, Solid-State Electron. 43, 13 (1999).

199.S. Oktyabrsky, M. O. Aboelfotoh, and J. Narayan,

J. Electron. Mater. 25, 1673 (1996).

200.H. S. Lee, M. W. Cole, R. T. Lareau, et al., J. Appl. Phys.

72, 4773 (1992).

201.R. Vtron. Adv. Mater. . Ghita, C. Logofatu, C. Negrila, et al., J. Optoelec-7, 3033 (2005).

1290BLANK, GOL’DBERG

202.S. E. Aleksandrov, VPis’ma Zh. Tekh. Fiz. . VLett. 31, 581 (2005)].

31. V (14), 1 (2005) [Tech. Phys.olkov, V. P. Ivanova, et al.,

203.T. VZh. Tekh. Fiz. . Blank, Yu. A. Gol’dberg, O. V(2007)].

77 (2), 140 (2007) [Tech. Phys. . Konstantinov, et al.,

52, 285204.T. J. Kim and P. H. Holloway, Crit. Rev. Solid State

Mater. Sci. 22, 239 (1997).

205.YPhys. . Shiraishi, N. Furuhata, and A. Okamoto, J. Appl.

76, 5099 (1994).

206.F. Radulescu and J. M. McCarthy, J. Vac. Sci. Technol.

17, 2294 (1999).

207.S. N. G. Chu, A. Katz, T. Boone, et al., J. Appl. Phys.

67, 37 (1990).

208.K. A. Jones, E. H. Linfield, and J. E. F. Frost, Appl.

Phys. Lett. 69, 4197 (1996).

209.H. Shimawaki, N. Furuhata, and K. Honjo, J. Appl.

Phys. 69, 7939 (1991).

210.N.-P. Chen, H. J. Ueng, D. B. Janes, et al., J. Appl. Phys.

88, 309 (2000).

211.C. J. Uchibori, T. Oku, N. Kameya, et al., Mater. Trans.,

JIM 37, 670 (1996).

212.A. A. Bergh and P. J. Dean, (Clarendon, Oxford, 1976; Mir, Moscow, 1979).

Light Emitting Diodes

213.T. Vvodn. (St. Petersburg) . Blank and Yu. A. Gol’dberg, Fiz. Tekh. Polupro-37, 999 (2003)].

37, 1025 (2003) [Semiconductors214.M. H. Park, L. C. Wang, D. C. Dufner, et al., Int. Phys.

Conf. Ser. 155, 369 (1997).

215.M.-H. Park, L. C. Wang, D. C. Dufner, et al., J. Appl.

Phys. 81, 3138 (1997).

216.L. C. Wang, Moon-Ho Park, H. A. Jorge, et al., Elec-tron. Lett. 82, 409 (1996).

217.F. Zhang, B. Li, and B. Sang, Mater. Sci. Eng. B (1994).

25, 24

218.F. Zhang, Z. Song, and J. Peng, Appl. Surf. Sci. (1992).

62, 83

219.YPart 2: Electron. . Koide and M. Murakami, Electron. Commun. Jpn.,

82, 43 (1999).

220.L. Baojun, L. Enke, and Z. Fujia, Solid-State Electron.

41, 917 (1997).

221.M. VLett. . Tagare, T. P. Chin, and J. M. Woodall, Appl. Phys.

68, 3485 (1996).

222.F. Ren, S. J. Pearton, J. R. Lothian, et al., Appl. Phys.

Lett. 65, 2165 (1994).

223.R. Dutta, M. A. Shahid, and P. J. Sakach, J. Appl. Phys.

69, 3968 (1991).

224.J. S. Park, H. J. Park, Ynol. B 21, 795 (2003).

. B. Hahn, et al., J. Vac. Sci. Tech-225.W. C. Huang, Appl. Surf. Sci. 245, 141 (2005).

226.T. Clausen, A. S. Pedersen, and O. Leistiko, Microelec-tron. Eng. 15, 157 (1991).

227.R. Dutta, M. A. Shahid, and P. J. Sakach, J. Appl. Phys.

69, 3968 (1991).

228.M. B. Takeyama, A. Noya, T. Hashizume, and H. Hase-gawa, Jpn. J. Appl. Phys. 65 (2B), 1115 (1999).

229.F. Ren, M. J. Antonell, C. R. Abernathy, et al., Appl.

Phys. Lett. 74, 1845 (1999).

230.D. G. Ivey, D. Wang, D. Yang, et al., J. Electron. Mater.

23, 441 (1994).

231.P. A. Leigh, R. M. Cox, and P. J. Dobson, Solid-State

Electron. 37, 1353 (1994).

232.W. C. Huang, T. F. Lei, and C. L. Lee, J. Appl. Phys. 9200 (1996).

79,

233.P. Jian, D. G. Ivey, R. Bruce, and G. Knight, J. Electron.

Mater. 23, 953 (1994).

234.J. Morais, T. A. Fazan, R. Landers, and E. A. S. Sato,

J. Appl. Phys. 79, 7058 (1996).

235.W.-C. Huang, T.-F. Lei, and C.-L. Lee, J. Appl. Phys.

78, 6108 (1995).

236.W. O. Barnard, G. Myburg, F. D. Auret, et al., Vacuum

46, 3 (1995).

237.D. G. Ivey, P. Jian, L. Wan, et al., J. Electron. Mater. 237 (1991).

20,

238.S. Hwang and J. Shim, J. Korean Phys. Soc. (2005).

46, L751

239.M.-H. Park, L. C. Wang, J. Ystrom, Appl. Phys. Lett. 70, 99 (1997).

. Cheng, and C. J. Palm-240.A. Yamaguchi, T. Okada, YConf. Ser. 162, 162 (1999).

. Iguchi, et al., Int. Phys.

241.A. Yamaguchi, H. Asamizu, T. Okada, et al., J. Appl.

Phys. 85, 7792 (1999).

242.A. Yamaguchi, I. Tonai, N. Yamabayashi, and M. Shi-bata, Bunseki Kagaki 40, 741 (1991).

243.H. Asamizu, A. Yamaguchi, YTrans., JIM 43, 1352 (2002).

. Iguchi, et al., Mater.

244.A. Yamaguchi, H. Asamizu, T. Okada, et al., J. Vac. Sci.

Technol. 18, 1957 (2000).

245.N. Asamizu, A. Yamaguchi, YSci. 159, 174 (2000).

. Iguchi, et al., Appl. Surf.

246.Vcond. Sci. Technol. . Malina, L. Moro, V11, 1121 (1996).

. Micheli, and I. Mojzes, Semi-247.M. H. Park, L. C. Wang, and D. M. Hwang, J. Electron.

Mater. 25, 721 (1996).

248.M. H. Park, L. C. Wang, J. YLett. 68, 952 (1996).

. Cheng, et al., Appl. Phys.

249.W. Zhao, S. Kim, J. Zhang, and I. Adesida, IEEE Elec-tron Device Lett. 27, 1 (2006).

250.S.-W. Chang, E. YPhys. Lett. 44, L9 (2005).

. Chang, S.-S. Lee, et al., Jpn. J. Appl.

251.Z. Tang, H. Hsia, H. C. Kuo, et al., Int. Phys. Conf. Ser.

166, 321 (2000).

252.J. Herniman, J. S. Yu, and A. E. Staton-Bevan, Appl.

Surf. Sci. 52, 2 (1991).

253.A. Katz, A. Feingold, and S. J. Pearton, Semicond. Sci.

Technol. 7, 436 (1992).

2.A. Katz, A. El-Roy, A. Feingold, et al., Appl. Phys. Lett.

62, 2652 (1993).

255.YPhys. Lett. .-H. Yeh, J.-T. Lai, and J. Ya-min Lee, Jpn. J. Appl.

35 (12A), L1569 (1996).

256.P. W. Leech, G. K. Reeves, and M. H. Kibel, J. Appl.

Phys. 76, 4713 (1994).

257.P. Jian, D. G. Ivey, R. Bruce, and G. Knight, J. Mater.

Sci. 7, 7 (1996).

258.T. Clausen and O. Leistiko, Appl. Phys. Lett. (1993).

62, 1108

SEMICONDUCTORS Vol. 41 No. 11 2007

MECHANISMS OF CURRENT FLOW IN METAL–SEMICONDUCTOR OHMIC CONTACTS

1291

259.T. Clausen, O. L. Chorkendorff, and J. Larsen, Thin

Solid Films 232, 215 (1993).

260.A. N. Imenkov, E. A. Grebenshchikova, B. E. Zhur-tanov, et al., Fiz. Tekh. Poluprovodn. (St. Petersburg)38, 1399 (2004) [Semiconductors 38, 1356 (2004)].261.K. Ikossi, M. Goldenberg, and J. Mittereder, Solid-State

Electron. 46, 1627 (2002).

262.R. K. Huang, C. A. Wang, C. T. Harris, et al., J. Elec-tron. Mater. 33, 1406 (2004).

263.C. A. Wang, D. A. Shiau, R. K. Huang, et al., J. Cryst.

Growth 261, 379 (2004).

2.J. A. Robinson and S. E. Mohney, J. Appl. Phys. 033703 (2005).

98,

265.J. Sigmund, M. Saglam, A. V227, 625 (2001).

ogt, et al., J. Cryst. Growth

266.K. Varblianska, K. Tzenev, and T. Kotsinov, Phys. Sta-tus Solidi A 163, 387 (1997).

267.A. VTechnol. B ogt, H. L. Hartnagel, G. Miehe, et al., J. Vac. Sci.

14, 3514 (1996).

268.E. Villemain, S. Gaillard, M. Rolland, and A. Joullie,

Mater. Sci. Eng. B 20, 162 (1993).

269.YPhys. . K. Su, F. S. Juang, and J. G. Kuang, Jpn. J. Appl.

30, 914 (1991).

270.K. G. Merkel, C. L. A. Cerny, VState Electron. 39, 179 (1996).

. M. Bright, et al., Solid-271.A. V66, 199 (1999).

ogt, A. Simon, J. Weber, et al., Mater. Sci. Eng. B

272.A. G. Milnes, M. Ye, and M. Stam, Solid-State Electron.

37, 37 (1994).

273.C.-T. Lee, K.-L. Jaw, and C.-D. Tsai, Solid-State Elec-tron. 42, 871 (1998).

274.YSoc. . Zhao, M. J. Jurkovic, and W. I. Wang, J. Electrochem.

144, 1067 (1997).

275.YPhys. . Shiraishi, N. Furuhata, and A. Okamoto, J. Appl.

76, 5099 (1994).

276.VMater. . Srivastav, R. Pal, B. L. Sharma, et al., J. Electron.

34, 225 (2005).

277.E. M. Lysczek, S. E. Mohney, and T. N. Wittberg, Elec-tron. Lett. 39, 1866 (2003).

278.J. B. Boos, B. R. Bennett, W. Kruppa, et al., J. Vac. Sci.

Technol. 17, 1022 (1999).

279.T. Teraji, S. Koizumi, and H. Kanda, Phys. Status Solidi

A 181, 129 (2000).

280.Tokuyuki Terai, Masayuki Katagiri, Satoshi Koizumi,

et al., Jpn. J. Appl. Phys. Lett. 42 (8A), L882 (2003).281.Han Gao, Cheng Mu, Fan Wang, et al., J. Appl. Phys.

93, 5602 (2003).

282.M. Werner, R. Job, A. Denisenko, et al., Diamond Relat.

Mater. 5, 723 (1996).

283.V13. Venkatesan and K. Das, IEEE Electron Device Lett.

, 126 (1992).

284.VJ. Appl. Phys. . Venkatesan, D. M. Malta, K. Das, and A. M. Belu,

74, 1179 (1993).

285.T. Iwasaki, K. Okano, YRelat. Mater. 3, 30 (1994).

. Matsumae, et al., Diamond

286.Ycond. Sci. Technol. . Chen, M. Ogura, S. Yamasaki, and H. Okushi, Semi-20, 860 (2005).

287.YDiamond Relat. Mater. . G. Chen, M. Ogura, S. Yamasaki, and H. Okushi,

13, 2121 (2004).

SEMICONDUCTORS Vol. 41 No. 11 2007

288.C. Zhen, Y(2003).

. Wang, S. He, et al., Opt. Mater. 23, 117

2.C. M. Zhen, YRelat. Mater. 11. Y, 1709 (2002).

. Wang, Q. F. Guo, et al., Diamond.

290.YPhys. B . Y. Wang, C. M. Zhen, Z. J. Yan, et al., Int. J. Mod.

16, 927 (2002).

291.C. Zhen, X. Liu, Z. Yan, et al., Surf. Interface Sci. 106 (2001).

32,

292.C. M. Zhen, X. Q. Liu, Z. J. Yan, et al., Chin. Phys. Lett.

17, 827 (2000).

293.H. A. Hoff, G. L. Waytena, C. L. VRelat. Mater. 5, 1450 (1996).

old, et al., Diamond.

294.M. Werner, C. Johnston, P. R. Chalker, et al., J. Appl.

Phys. 79, 2535 (1996).

295.K. Das, VPhys. 76, 2208 (1994).

. Venkatesan, and T. P. Humphreys, J. Appl.

296.J. Nakanishi, A. Otsuki, T. Oku, et al., J. Appl. Phys. 2293 (1994).

76,

297.M. Werner, O. Dorsch, H. U. Baerwind, et al., Diamond.

Relat. Mater. 3, 983 (1994).

298.YSin. . Y. Wang, C. M. Zhen, H. X. Gong, et al., Acta Phys.

49, 1348 (2000).

299.J. R. Waldrop, J. Appl. Phys. 75, 48 (1994).

300.E. Vet al., Diamond Relat. Mater. . Kalinina, G. F. Kholuyanov, A. V8, 1114 (1999).

. Shchukarev,

301.M. W. Cole, P. C. Joshi, C. Hubbard, et al., J. Appl.

Phys. 91, 38 (2002).

302.H. Na, H. Kim, K. Adachi, et al., J. Electron. Mater. (2004).

33,

303.T. VInstrum. Methods Phys. Res. A . Blank, Y. A. Goldberg, and O. V509, 109 (2003).

. Goldberg, Nucl.

304.O. J. Guy, G. Pope, I. Blackwood, et al., Surf. Sci. 253 (2004).

573,

305.O. J. Guy, G. Pope, I. Blackwood, et al., Mater. Sci.

Forum 457–460, 853 (2004).

306.G. Pope, O. Guy, and P. A. Mawby, Mater. Sci. Forum

457–460, 833 (2004).

307.S. J. Yang, C. K. Kim, I. H. Noh, et al., Diamond Relat.

Mater. 13, 1149 (2004).

308.Nam-Ihn Cho, Kyung-Hwa-Jung, and Yong Choi,

Semicond. Sci. Technol. 19, 306 (2004).

309.S. YSci. Forum . Han, N. K. Kim, E. D. Kim, and J. L. Lee, Mater.

3–393, 7 (2002).

310.L. G. Fursin, J. H. Zhao, and M. Weiner, Electron. Lett.

37, 1092 (2001).

311.S. Liu, K. Reinhardt, C. Severt, and J. Scofield, Int.

Phys. Conf. Ser. 142, 5 (1996).

312.L. Kassamakova, R. D. Kakanakov, I. Vet al., IEEE Trans. Electron Devices 46, 605 (1999).. Kassamakov,

313.K. Vassilevski, K. Zekentes, G. Constantinidis, et al.,

Mater. Sci. Forum 338–343, 1017 (2000).

314.L. Kassamakova, R. Kakanakov, I. Kassamakov, et al.,

Mater. Sci. Forum 338–343, 1009 (2000).

315.R. Kakanakov, L. Kassamakova, I. Kassamakov, et al.,

Mater. Sci. Eng. B 80, 374 (2001).

316.K. Vassilevski, K. Zekentes, K. Tsagaraki, et al., Mater.

Sci. Eng. B 80, 370 (2001).

317.L. Kassamakova, R. Kakanakov, I. Kassamakov, et al.,

Mater. Sci. Forum 353–363, 251 (2000).

1292BLANK, GOL’DBERG

318.T. VPis’ma Zh. Tekh. Fiz. . Blank, Yu. A. Gol’dberg, O. V. Konstantinov, et al.,

Lett. 27, 776 (2001)].

27 (18), 43 (2001) [Tech. Phys.319.O. Nakatsuka, T. Takei, Ya. Koide, and M. Murakami,

Mater. Trans. 43, 1984 (2002).

320.S. Tsukimoto, T. Sakai, T. Onishi, et al., J. Electron.

Mater. 34, 1310 (2005).

321.S. Tsukimoto, T. Sakai, and M. Murakami, J. Appl.

Phys. 96, 4976 (2004).

322.R. Kakanakov, L. Kazamakova-Kolaklieva, N. Hristeva,

et al., Mater. Sci. Forum 457–460, 877 (2004).

323.S. Tsukimoto, K. Nitta, T. Sakai, et al., J. Electron.

Mater. 33, 460 (2004).

324.T. Sakai, K. Nitta, S. Tsukimoto, et al., J. Appl. Phys.

95, 2187 (2004).

325.R. Kakanakov, L. Kassamakova, N. Hristeva, et al.,

Mater. Sci. Forum 3–393, 917 (2002).

326.L. Kassamakova, R. Kakanakov, N. Nordell, et al.,

Mater. Sci. Eng. B 61–62, 291 (1999).

327.K. Vet al., Mater. Sci. Eng. B . Vassilevski, G. Constantinidis, N. Papanicolaou,

61–62, 296 (1999).

328.N. A. Papanicolaou, A. Edwards, M. VW. T. Anderson, Appl. Phys. Lett. 73, 2009 (1998).. Rao, and

329.G. YP. H. Holloway, J. Appl. Phys. . McDaniel, S. T. Fenstermaker, W. V96, 5357 (2004).

. Lampert, and

330.C. Deeb and A. H. Heuer, Appl. Phys. Lett. (2004).

84, 1117

331.T. Jang, B. Odekirk, L. D. Madsen, and L. M. Porter,

J. Appl. Phys. 90, 4555 (2001).

332.U. Schmid, R. Getto, S. T. Sheppard, and W. Wondrak,

J. Appl. Phys. 85, 2681 (1999).

333.S. Liu, K. Reinhardt, C. Severt, and J. Scofield, Int.

Phys. Conf. Ser. 142, 5 (1996).

334.A. Scorzoni, F. Moscatelli, A. Poggi, et al., Mater. Sci.

Forum 457–460, 881 (2004).

335.F. Moscatelli, A. Scorzoni, A. Poggi, et al., Mater. Sci.

Forum 433–434, 673 (2002).

336.F. Moscatelli, A. Scorzoni, A. Poggi, et al., Semicond.

Sci. Technol. 18, 5 (2003).

337.K. Nakashima, O. Eryo, O. Kais, et al., Mater. Sci.

Forum 338, 1005 (2000).

338.A. A. Iliadis, S. N. Andronescu, W. Yang, et al., J. Elec-tron. Mater. 28, 136 (1999).

339.A. A. Iliadis, S. N. Andronescu, K. Edinger, et al., Appl.

Phys. Lett. 73, 35 (1998).

340.J. Crofton, PAppl. Phys. Lett. . A. Barnes, J. R. Williams, and J. A. Edmond,

62, 384 (1993).

Translated by A. Spitsyn

SEMICONDUCTORS Vol. 41 No. 11 2007

因篇幅问题不能全部显示,请点此查看更多更全内容

Copyright © 2019- igat.cn 版权所有 赣ICP备2024042791号-1

违法及侵权请联系:TEL:199 1889 7713 E-MAIL:2724546146@qq.com

本站由北京市万商天勤律师事务所王兴未律师提供法律服务