Analysis of Reverse Leakage Current and Breakdown Voltage in GaN and InGaN/GaN Schottky Barriers

1986 IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 58, NO. 7, JULY 2011

Analysis of Reverse Leakage Current and BreakdownVoltage in GaN and InGaN/GaN Schottky Barriers

Wei Lu, Student Member, IEEE, Lingquan Wang, Member, IEEE, Siyuan Gu, Student Member, IEEE,David P. R. Aplin, Daniel M. Estrada, Paul K. L. Yu, Fellow, IEEE, and Peter M. Asbeck, Fellow, IEEE

Abstract—A study of the reverse-leakage-current mechanismsin metal–organic-chemical-vapor-deposition (MOCVD)-grownGaN Schottky-barrier diodes is presented. An analysis is carriedout of the characteristics of GaN Schottky diodes as well as ofdiodes with an InGaN surface layer to suppress the reverse leakagecurrent and increase the breakdown voltage. The experimentalresults of the diodes with InGaN surface layers showed a ∼40-Vbreakdown voltage increase and a significant leakage-currentreduction under high reverse bias, in comparison with the designwith GaN only. Such improvements are attributed to the reducedsurface electric field and the increased electron tunneling distanceinduced by the polarization charges at the InGaN/GaN interface.We also report the effect of a high-pressure (near atmosphericpressure) MOCVD growth technique of the GaN buffer layer tofurther improve the leakage current and breakdown voltage.

Index Terms—Diode breakdown voltage, high-pressure (HP)metal–organic-chemical-vapor-deposited (MOCVD) buffer,InGaN/GaN heterojunction, polarization charge, Schottky-diodeleakage current.

I. INTRODUCTION

III–nitride-based Schottky diodes have achieved many appli-cations, such as ultraviolet photodetectors [1], gas sensors [2],high-voltage rectifiers [3], and varactors [4]. One key factor toimprove the performance of these devices is to minimize re-verse leakage current, particularly under high reverse voltages.Many efforts have been reported to reduce the leakage cur-rent in GaN-based Schottky barriers, such as electrochemicalsurface treatment [5], SiO2 dielectric surface passivation [6],capping with low-temperature GaN layers [7], and oppositelydoped surface layers to increase the Schottky-barrier height

Manuscript received December 8, 2010; revised March 13, 2011; acceptedApril 12, 2011. Date of publication May 19, 2011; date of current versionJune 22, 2011. This work was supported in part by the University ofCalifornia San Diego Center for Wireless Communications, by the FutureWeiTechnologies Inc., and by the UC Discovery Grant Program. The work ofD. M. Estrada was supported by the U.S. Department of Homeland SecurityFellowship administered by the Oakridge Institute for Science and Educationthrough an interagency agreement between the U.S. Department of Energy andthe Department of Homeland Security. The review of this paper was arrangedby Editor S. Bandyopadhyay.

W. Lu, S. Gu, D. P. R. Aplin, P. K. L. Yu, and P. M. Asbeck are with theDepartment of Electrical and Computer Engineering, University of CaliforniaSan Diego, San Diego, CA 92093 USA (e-mail: [email protected]).

L. Wang is with the SuVolta Inc., Los Gatos, CA 95032 USA, and alsowith the Department of Electrical and Computer Engineering, University ofCalifornia San Diego, San Diego, CA 92093 USA.

D. M. Estrada is with the Materials Science and Engineering Program,University of California San Diego, San Diego, CA 92093 USA.

Color versions of one or more of the figures in this letter are available onlineat http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/TED.2011.2146254

(SBH) [8]. However, these studies do not discuss in detailthe leakage-current-suppression mechanisms particularly underhigh reverse voltages, which is important in high-power appli-cations. We have recently reported that, for the developmentof high-Q high-voltage varactors, using a thin in situ grownInGaN surface layer, the reverse leakage current of GaN-basedSchottky barriers under high reverse voltages is significantlysuppressed and the breakdown voltage is increased by ∼40 V[4]. In this paper, we present a detailed analysis of the char-acteristics of the resultant devices and discuss the mechanismsfor leakage-current reduction. We also report that the leakagecurrent can be further significantly suppressed by using ahigh-pressure (HP, near atmospheric pressure) metal–organic-chemical-vapor-deposition (MOCVD) growth technique [9] forthe GaN buffer layers (i.e., an HP GaN buffer).

In this paper, the details of our experimental work are presentfirst, followed by the discussions on the reverse-leakage-currentmechanisms in GaN Schottky barriers and the InGaN-surface-layer design. Capacitance–voltage (C–V ) and current–voltage(I–V ) characteristics of these diodes are studied. Lastly, theimprovements stemming from the HP GaN buffer are discussed.

II. EXPERIMENTAL PROCEDURE

The samples used in this paper were grown using a ThomasSwan close-coupled showerhead 3 × 2′′ MOCVD system withan adjustable gap between the showerhead and the susceptor.The substrates used were 2-in c-sapphire. For comparison pur-poses, five samples have been grown using a conventional LPMOCVD growth technique [10], and their key growth detailsare shown in Table I. Thick InGaN samples (> 300 nm toensure full strain relaxation [11]) were grown for estimatingthe indium concentration and optimizing the InGaN growth byX-ray-diffraction (XRD) measurements. To fabricate the Schot-tky diode, a 120-nm-thick Ni Schottky contact was depositedimmediately after the sample growth and surface cleaning.The Ni film was then patterned, and circular mesa diodeswith diameters in the range of 80–300 μm were formed usingBCl3/Cl2 in a Trion reactive-ion-etching/inductively-coupled-plasma dry-etch system, which was followed by an ∼90 ◦CKOH (0.1 mol/L) treatment to reduce the dry-etch residuals.A Ti/Al/Pd/Au (20 nm/80 nm/50 nm/100 nm) metal stack wasused to form ohmic contacts to n+ GaN [4]. Fig. 1 shows theschematic diagram of the cross section of the GaN/InGaN/GaNsample structure and the corresponding scanning-electron-microscopy (SEM) top view.

0018-9383/$26.00 © 2011 IEEE

LU et al.: REVERSE LEAKAGE CURRENT AND BREAKDOWN VOLTAGE IN SCHOTTKY BARRIERS 1987

TABLE ISAMPLE LABELS AND KEY GROWTH DETAILS USED IN THIS PAPERa

Fig. 1. (a) Schematic cross-sectional diagram of a GaN/InGaN/GaN samplestructure. (b) A SEM top view of a fabricated diode with an 80-μm diameter.

III. ANALYSIS OF REVERSE LEAKAGE CURRENT

IN GAN SCHOTTKY BARRIERS

The reverse leakage current of the Schottky barrier hasbeen attributed to the electron tunneling directly through theSchottky barrier [12], [13]. Fig. 2 shows a schematic energyband diagram of a Schottky barrier under reverse bias. Accord-ing to conventional analysis [14], for reverse biases greater than3kT/q, the current density Jr from a semiconductor to a metalthrough the Schottky barrier can be expressed by the following:

Jr =A∗T

kexp

[−qφb

kT

]×

∞∫0

P (ζ) exp(− ζ

kT

)dζ

+A∗T

k×

q(Vb−Δφ)∫0

FmP (η) [1 − Fs(Vr)] dη

=JA + JB (1)

where A∗ is the effective Richardson constant, T is the temper-ature in kelvin, k is the Boltzmann constant, q is the charge of

an electron, qφb is the effective barrier height, Vb is the built-inpotential at the reverse bias Vr ignoring image force lowering,qΔφ is the image force lowering of the barrier height, Fm andFs are the Fermi–Dirac distribution functions of electrons inthe metal and the semiconductor, respectively, ζ and η are theenergy of the electrons above and below the top of the effectivebarrier, respectively, and P (ζ) and P (η) are the tunneling prob-abilities of the electrons above and below the top of the effectivebarrier, respectively. Assuming P (ζ) = 1 and P (η) calculatedusing the Wentzel–Kramers–Brillouin approximation [15],we find

P (E) = exp

⎡⎣−2

√2m∗

�

x2∫x1

√qφ(x) − Edx

⎤⎦ (2)

where m∗ is the electron effective mass, � is the reducedPlanck constant, x is the distance measured from the metal-semiconductor interface, and x1 and x2 are the classical turningpoints, which determine the electron tunneling distance as L =x2 − x1. Measured from the Fermi level of metal, φ(x) is theelectrostatic potential distribution, and E = qφb − η is the en-ergy of tunneling electrons. In (1), the first term JA correspondsto the TE component, and the second term JB corresponds tothe tunneling component, which dominates at high dopings andlow temperatures. We need only to consider the second termfor studying n-type GaN (n-GaN) Schottky barriers at roomtemperature. Under low reverse bias, P (E) sharply decreasesas E decreases based on (2); most of electrons tunnel nearthe top of the effective barrier, so that 1 − Fs ≈ 1, and Fm isgiven by

Fm =1

1 + exp(

E−Efm

kT

) ≈ exp(−E

kT

). (3)


Fig. 2. Schematic energy band diagram of a Schottky barrier under reversebias. qφb0 is the barrier height ignoring image force lowering (denoted here asqΔφ), qVn = Ec − Ef , Em is the energy where the electron tunneling peaks,Vb is the built-in potential at the reverse bias Vr ignoring image force lowering,and L = x2 − x1 is the tunneling distance at Em.

Assuming that the barrier has a triangular shape near the top ofthe effective barrier, the electrostatic potential distribution canbe written as

qφ(x) = −qξx + qφb0 (4)

where ξ is the constant electric field near the top of the effectivebarrier. Then, the integral of (2) can be evaluated exactly, andby ignoring JA, Jr can be approximated as

Jr ≈ JB =A∗T

k

×qφb0∫

−(Vr−Vn)

exp

[−4

3

√2m∗

�

(qφb0 − E)32

qξ− E

kT

]dE. (5)

Taking a derivative of the exponential factor in (5), we can findenergy Em where the electron tunneling peaks as follows:

Em = qφb0 −�

2q2ξ2

8m∗k2T 2. (6)

Assuming that the semiconductor has the uniform doping con-centration Nd, (6) becomes

Em = qφb0 −�

2q3Nd(Vb0 + Vr)4m∗εk2T 2

(7)

where Vb0 is the built-in potential at zero bias and ε is thedielectric constant of the semiconductor. Equation (7) is adecreasing function with the reverse bias Vr and is valid aslong as (3) is a good approximation (Em > 3kT ). With furtherincreasing Vr, Em eventually becomes almost constant whenit reaches a few kT value below the Fermi level of metalwhere Fm is almost constant at 1, but P (E) decreases sharplyas E decreases further. Therefore, under high reverse bias,most of the electrons tunnel near the Fermi level of metal.For a Ni/n-GaN Schottky barrier with Nd = 1 × 1017 cm−3,T = 300 K, and qφb0 = 1 eV, Em = 1 − 0.03 × (0.92 + Vr)in electronvolts. When Vr > 33 V, most of the electrons tunnelnear the Fermi level of metal. This finding suggests that theleakage-current suppression under high reverse bias can beachieved by reducing the electron tunneling probability aroundthe Fermi level of metal.

Fig. 3. Experimental and calculated reverse leakage current of a controlsample with an area of 5 × 10−5 cm2.

Anticipating the measurement results reported in the fol-lowing, we find that the theoretical analysis outlined aboveagrees well with our experimental measurements. Fig. 3 showsthe measured reverse leakage current for a control sample(i.e., a Ni/n-GaN Schottky diode with Nd ≈ 1 × 1017 cm−3

confirmed by C–V measurement) with an area of 5 × 10−5 cm2

and, for comparison, the calculated result based on (1) withparameters of Nd = 1.1 × 1017 cm−3, qφb0 = 0.95 eV, T =300 K, A∗ = 26.4 A · cm−2 · K−2 [16], and m∗ = 0.2m0 [17],where m0 is the electron mass. Departures from the tunneling-based curve that occur at low bias (< 30 V), where thedirect tunneling current is very small, are possibly related totrap/defect center-assisted tunneling and/or generation currentwithin the depletion region and at high bias (> 80 V). ForVr = 80 V, the calculated electric field at the surface reachesξ = 1.9 MV/cm, where impact ionization effects are expectedto become significant.

IV. InGaN-SURFACE-LAYER DESIGN FOR

REDUCED LEAKAGE CURRENT

Based on the above analysis, we introduced a design forreducing the electron tunneling probability around the Fermilevel of metal by Schottky-barrier engineering. It is knownthat large polarization-charge densities present at III–nitridesemiconductor heterojunction interfaces significantly affect theelectric-field distributions in the semiconductors [18], provid-ing opportunities for Schottky-barrier engineering. Thus, wedesigned an InGaN surface layer and utilized the polarizationcharges at the InGaN/GaN interface to engineer the tunnelingbarrier. As shown in Fig. 4, the surface electric field is sig-nificantly reduced for the InGaN-surface-layer design (InGaN/GaN case), leading to an increased electron tunneling distancenear the Fermi level of metal where most of the electronstunnel under high reverse bias as aforementioned. The leakagecurrent is suppressed because the tunneling probability P (E)given by (2) sharply decreases with the increase in the tunnel-ing distance. A 2-nm-thick GaN cap layer was also designed(GaN/InGaN/GaN) in case that there was indium out diffusionduring the cooling down process of the MOCVD growth, whichcould weaken the effectiveness of the InGaN surface layers


Fig. 4. (a) Schematic diagram of the electric field under the same reverse biasfor the designs with InGaN/GaN, GaN/InGaN/GaN, and GaN only. Δξ is thesurface electric-field reduction of the InGaN/GaN design relative to GaN-onlydesign. (b) Representative energy band diagrams (Ec only) at Vr = 0 V and(c) Vr = 80 V by solving the 1-D Poisson equation.

(i.e., a less tunneling distance increase compared with thatwithout a GaN cap layer, as shown in Fig. 4(c)). In order toobtain high-quality InGaN layers and long tunneling distances,we designed the indium concentration and the thickness ofthe InGaN surface layer to be 10% and 15 nm, respectively.Higher indium concentration and InGaN thickness might resultin InGaN-layer-quality degradation and strain relaxation (lesspolarization effects). The corresponding energy band diagramsat 0 V and a reverse bias of 80 V shown in Fig. 4(b) and (c),respectively, were obtained by solving the 1-D Poisson equa-tion. In this calculation, a 4.9 × 1012 cm−2 sheet polarization-charge density and a 0.21-eV conduction-band discontinuitywere assumed at the In0.1Ga0.9N/GaN and GaN/In0.1Ga0.9Ninterfaces [19]; the SBH (ignoring the image force lowering)of In0.1Ga0.9N was also assumed to be 0.21 eV lower thanthat of GaN, which corresponds with the expected conduction-band energy difference ΔEc between In0.1Ga0.9N and GaN.It is seen that the tunneling distance for the In0.1Ga0.9N/GaNdesign at a reverse bias of 80 V is doubled compared with theGaN-only design.

V. RESULTS AND DISCUSSION

Characterization of the Schottky diodes was carried out withC–V and I–V measurements.

Fig. 5. C–V results of the five samples with area of 7.85 × 10−5 cm2, whichare measured in the dark at 1 MHz. The colored dots are the measurement data,and the dash lines are the fitting curves.

A. C–V Measurements and Analysis

Capacitance measurements were carried out using 1-MHzsignals. Fig. 5 shows the plots of 1/C2 versus the reverse volt-age V . Curves for different samples exhibit slight differencesof slopes due to small variations (< 1%) of the fabricated diodeareas and the doping concentration Nd of the n-GaN layers,which is approximately 1 × 1017 cm−3 (assuming completeionization), extracted from the C–V results. It is noted thatthe curves of the InGaN-containing samples consistently showa voltage shift when compared with the control sample. Thisshift is due to the change in electric-field distribution and elec-trostatic potential in the InGaN layers caused by the inducedpolarization charges, as shown in Fig. 4(a). The voltage shiftcan be estimated by considering the integral of the electric-fieldreduction between the GaN and InGaN layers, which is over thewidth of the InGaN layer. This estimate results in the followingexpression:

ΔV =tσq

εInGaN+

t2Ndq

2εInGaN(8)

where ΔV is the voltage shift, t is the InGaN layer thickness,σ is the sheet polarization-charge density at the InGaN/GaNinterface, and εInGaN is the dielectric constant of InGaN. Asshown in Fig. 5, ΔV are ∼0.6 and ∼0.8 V for In0.06Ga0.94N-and In0.1Ga0.9N-containing samples, respectively. Using (8),the polarization-charge density at the In0.06Ga0.94N/GaN andIn0.1Ga0.9N/GaN interfaces is estimated to be 1.9 × 1012 and2.7 × 1012 cm−2, respectively. The estimated numbers are lessthan the theoretical expectations (i.e., 0.9 × 1012 and 4.0 ×1012 cm−2 for the spontaneous and piezoelectric polarizations,respectively, in the In0.06Ga0.94N/GaN samples and 1.5 × 1012

and 6.4 × 1012 cm−2 for the spontaneous and piezoelectriccomponents, respectively, for In0.1Ga0.9N/GaN) [18], [19]probably due to partial strain relaxation.

B. Forward I–V Measurements and Analysis

The polarization charges at the InGaN/GaN interfaceschange the Schottky-barrier shape and, correspondingly, theforward I–V characteristics. Fig. 6 illustrates schematicallythe energy band diagram of an InGaN/GaN structure. The


Fig. 6. Schematic energy band diagram of an InGaN/GaN Schottky barrierunder forward bias. ΔEc is the conduction-band discontinuity, qφb0 is thebarrier height ignoring image force lowering, qφb is the effective barrier height,Vf is the forward bias, t is the InGaN thickness, and Wd is the forward-bias-dependent depletion thickness measured from the InGaN/GaN interface.

effective barrier height qφb is forward-bias (Vf ) dependent andis given by

qφb = qφb0 + qσq − WdNdq

εInGaNt + ΔEc (9)

where ΔEc is the conduction-band discontinuity at the InGaN/GaN interface and Wd is the forward-bias-dependent depletionwidth measured from the InGaN/GaN interface. It may becalculated by

Wd =

√2εGaN(φb − Vn − Vf )

Ndq(10)

where εGaN is the dielectric constant of GaN. With (9) and (10),the following equation holds:

∂φb

∂Vf=

tεGaN

tεGaN + WdεInGaN. (11)

Using the thermionic emission (TE) theory, the forward I–Vcharacteristics may be described as

Jf ≈ A∗T 2 exp(

qVf − qφb

kT

). (12)

Therefore, the ideality factor n (or the inverse subthresholdslope) may be obtained as

n =q

2.3kT

(∂(log10 Jf )

∂Vf

)−1

=(

1 − ∂φb

∂Vf

)−1

= 1 +tεGaN

WdεInGaN. (13)

As shown in (13), the ideality factors for the indium-containing samples are expected to be bias dependent andgreater than unity.

Fig. 7(a) shows the experimental I–V curves of the controland In0.06Ga0.94N-containing samples and the correspondinglytheoretical calculations in the semilogarithmic scale. The ide-ality factors are 1.08 for the control sample, and 1.33 and 1.3for the 0 nm/In0.06Ga0.94N and 2 nm/In0.06Ga0.94N samples,respectively. These results agree well with the values computedin (13). The turn-on voltages for the same samples are ∼0.8,

Fig. 7. Forward I–V curves of the five samples with areas of 5 ×10−5 cm2. (a) Control and In0.06Ga0.94N-containing samples. (b) Control andIn0.1Ga0.9N-containing samples. The colored dots are the measurement data,and the solid lines are the calculated TE currents.

∼1.1, and ∼1 V, respectively, as a result of the increasedeffective barrier height. The forward I–V calculations are basedon (12), and the image force lowering and the bias-dependenteffective barrier height are considered as well. The values ofthe parameters used for all calculations (including forward andreverse I–V curves) in this paper are summarized in Table II.

The experimental I–V curves of the In0.1Ga0.9N-containingsamples appear to consistently exhibit excess current in thelow-forward-bias region, as shown in Fig. 7(b). The distinctivetwo-region-like I–V characteristics may indicate that two cur-rent mechanisms exist in the In0.1Ga0.9N-containing samples.One component is the TE current, which dominates in thehigh-forward-bias region, and the other component (i.e., thedifference between the measurement data and the calculatedTE current) may be defect-related current, which dominatesin the low-forward-bias region. Lower growth temperature forthe higher-In-concentration InGaN layers may have inducedmore defects. The calculated TE current for the In0.1Ga0.9N-containing samples fits the experimental curves in the high-forward-bias region reasonably well.

C. Reverse I–V Measurements and Analysis

The diodes with different areas for the layer stacks specifiedin Table I were fabricated and measured. The measurementresults showed that the leakage current is approximately pro-portional to the corresponding area of the diodes (instead of thediameters), indicating that the leakage current is not dominatedby edge effects.


TABLE IIVALUES OF THE PARAMETERS USED FOR ALL CALCULATIONS (INCLUDING FORWARD AND REVERSE I–V CURVES) IN THIS PAPERa

Fig. 8. Reverse I–V curves of the five samples with area of 5 × 10−5 cm2.(a) Experimental curves. (b) Experimental and calculated curves of the con-trol and In0.06Ga0.94N-containing samples. (c) Experimental and calculatedcurves of the control and In0.1Ga0.9N-containing samples. The colored dotsare measurement data and the solid lines are calculated curves.

The experimental and calculated reverse I–V curves areshown in Fig. 8. For the experimental results shown in Fig. 8(a),the InGaN-surface-layer designs significantly reduce the leak-age current under high reverse bias. Furthermore, it is seenthat the suppression of leakage current is more effective forlayer designs without a GaN cap (i.e., 0 nm/In0.06Ga0.94N and0 nm/In0.1Ga0.9N), as predicted in Fig. 4(c). However, it is alsonoted that the leakage-current reduction of the In0.1Ga0.9N-containing samples is less than expected. This corresponds wellwith the less-than-expected polarization-charge density at theIn0.1Ga0.9N/GaN interface, as extracted from the C–V results.The calculated curves based on (1) fit the experimental curveswell in the high-reverse-bias regime, as shown in Fig. 8(b)and (c). In the cases of control, 2 nm/In0.06Ga0.94N, and0 nm/In0.06Ga0.94N samples, the last several experimental datapoints are higher than the calculated values, which is probablydue to impact ionization effects at very-high-surface-electric-field conditions, as aforementioned. In the lower reverse-bias regime, the leakage current is expected to be dominatedby mechanisms other than direct tunneling described by (1),such as trap/defect center-assisted tunneling and/or generationcurrent.

A significant improvement is also observed in the breakdownvoltages for the InGaN-surface-layer designs, which exhibitan increase in ∼40 V with respect to the control sample.A histogram of the breakdown voltages for the control andInGaN-containing samples is shown in Fig. 9. To ensure thevalidity of this statistics, the data of the control and InGaN-containing diodes shown here come from different fabricatedsamples and different grown wafers.

VI. LEAKAGE-CURRENT REDUCTION

USING THE HP GAN BUFFER

MOCVD-grown III–N epitaxial layers suffer from highdislocation density due to the highly lattice-mismatchedgrowth substrates, such as the sapphire used in this paper [21].The excess reverse leakage current of the III–N-based Schottkydiodes has been reported to be dislocation related [5], [22]. It


Fig. 9. Histogram of the breakdown voltages for the control and InGaN-containing diodes.

has also been reported that, compared with conventional LPMOCVD growth, the HP-MOCVD-growth technique favorsthe reduction of the dislocation density and the growth of high-quality GaN layers, although there will be an increase in thegas-phase parasitic reactions [9], [23]. In this paper, we appliedthe HP-MOCVD-growth technique to the GaN buffer layer toreduce the dislocation density but used the LP MOCVD growthfor other layers to minimize the parasitic gas-phase reactions.For comparison purposes, control and 2 nm/In0.1Ga0.9Nsamples using the HP GaN buffer were grown and labeled asControl_HP and 2 nm/In0.1Ga0.9N_HP, respectively. Thegrowth and fabrication details are the same as for the samplesaforementioned, except the higher growth pressure of theGaN buffer layer, as shown in Table I. The full-width at half-maximum of the (0002) XRD rocking curve of the samplesusing the HP GaN buffer was ∼284 arcsec, in comparison withthe ∼337 arcsec of the samples using an LP GaN buffer, indi-cating lower dislocation density and better layer quality [24].As shown in Fig. 10, the corresponding reverse I–V curves ofthe samples using the HP GaN buffer exhibit significantly lessleakage current and higher breakdown voltages compared withthat of the samples using an LP GaN buffer.

We verified that the improved characteristics were not theresult of the changed doping level by C–V measurements sincewe measured Nd ≈ 1 × 1017 cm−3 for the n-GaN layer forthe samples with both types of GaN buffers. As shown inFigs. 10(b) and (c) and 11, the significantly reduced leakagecurrent and the forward current for the samples using the HPGaN buffer can be fit with the expressions outlined above,with only increasing SBH values by 0.1 and 0.05 eV for theControl_HP and 2 nm/In0.1Ga0.9N_HP samples compared withthe control and 2 nm//In0.1Ga0.9N samples, respectively. InFig. 11, the defect-related current is not observed in the low-forward-bias region for the samples using the HP GaN buffer,which indicates lower defect densities in these samples. This isconsistent with our XRD measurement results. In Fig. 12, thehistogram of the extracted SBH values from the experimentalforward I–V curves also exhibits an average of ∼0.1-eV higherSBH for the Control_HP samples relative to the control sam-ples. The significant SBH increase in the samples using the HPGaN buffer can be attributed to the different strain distributions(i.e., the dislocation density) in the semiconductors caused by

Fig. 10. Reverse I–V curves of the samples using HP and LP GaN bufferlayers with area of 5 × 10−5 cm2. (a) Experimental curves. (b) Experimentaland calculated curves of the control and Control_HP samples. (c) Experimentaland calculated curves of the 2 nm/In0.1Ga0.9N and 2 nm/In0.1Ga0.9N_HPsamples. The colored dots are the experimental data, and the solid lines arecalculated curves.

Fig. 11. Forward I–V curves of the Control_HP and 2 nm/In0.1Ga0.9N_HPsamples with an area of 5 × 10−5 cm2. The colored dots are the experimentaldata, and the solid lines are the calculated TE currents.


Fig. 12. Histogram of SBH values extracted from the experimental forwardI–V curves of the control and Control_HP samples.

the different buffer layer growth pressures. The dependence ofresidual strain on growth pressures in MOCVD-grown GaN onc-sapphire substrates has been reported in [24]. Based on thestudy, we estimated the differences of Δa ≈ −0.009 Å in thelattice constant a and Δc ≈ 0.006 Å in the lattice constant cbetween the samples grown in this paper with HP and LP GaNbuffers. The corresponding differences of compressive strainalong a-axis and tensile strain along c-axis can be calculatedto be εxx = εyy = Δa/a0 = −0.009 Å/3.189 Å [17], which is≈ −0.0028, and εzz = Δc/c0 = 0.006 Å/5.185 Å [17], whichis ≈0.0011, where a0 and c0 are the theoretical lattice constantsof GaN. It is known from the linear deformation potentialtheory that, when the semiconductor is under strain, the edgeof the conduction band shifts [25] and also the SBH. The shiftcan be described by ΔEc = aczεzz + act(εxx + εyy), whereΔEc is the conduction-band edge shift and acz and act are theconduction-band deformation potentials. We assume the changeof the conduction-band energy results in an equal change ofSBH. Then, if we consider the values of acz and act in therange of −10 to −20 eV [26], [27], the change in the SBHvalues induced by the strain values inferred from our samplesare in the range of 0.05 to 0.1 eV, in reasonable accord withour analysis.

VII. CONCLUSION

Leakage-current reduction and breakdown voltage increasein GaN Schottky barriers were achieved by the incorporationof suitable InGaN surface layers. Analysis shows that theInGaN surface layers change the electric-field distribution nearthe surface due to polarization-induced charges. Experimen-tal C–V and I–V measurements confirm this analysis. TheHP-MOCVD-growth technique of the GaN buffer layer wasalso found to be effective in reducing the leakage current andincreasing the breakdown voltage of GaN Schottky barriers.Based on our experimental results and analysis, a further im-provement in leakage current and breakdown voltage may beobtained with an InGaN/GaN design with a higher indiumconcentration and smaller InGaN layer thickness to minimizestrain relaxation and maximize the polarization-charge densitywhen combined with a HP GaN buffer.

REFERENCES

[1] Q. Chen, J. W. Yang, A. Osinsky, S. Gangopadhyay, B. Lim, M. Z. Anwar,M. A. Khan, D. Kuksenkov, and H. Temkin, “Schottky barrier detectorson GaN for visible-blind ultraviolet detection,” Appl. Phys. Lett., vol. 70,no. 17, pp. 2277–2279, Apr. 1997.

[2] M. Ali, V. Cimalla, V. Lebedev, H. Romanus, V. Tilak, D. Merfeld,P. Sandvik, and O. Ambacher, “Pt/GaN Schottky diodes for hydrogengas sensors,” Sens. Actuators B, Chem., vol. 113, no. 2, pp. 797–804,Feb. 2006.

[3] Z. Z. Bandic, P. M. Bridger, E. C. Piquette, T. C. McGill, R. P.Vaudo, V. M. Phanse, and J. M. Redwing, “High voltage (450 V) GaNSchottky rectifiers,” Appl. Phys. Lett., vol. 74, no. 9, pp. 1266–1268,Mar. 1999.

[4] W. Lu, L. Wang, S. Gu, D. P. R. Aplin, D. M. Estrada, P. K. L. Yu,and P. M. Asbeck, “InGaN/GaN Schottky diodes with enhanced voltagehandling capability for varactor applications,” IEEE Electron Device Lett.,vol. 31, no. 10, pp. 1119–1121, Oct. 2010.

[5] E. J. Miller, D. M. Schaadt, E. T. Yu, P. Waltereit, C. Poblenz, andJ. S. Speck, “Reverse-bias leakage current reduction in GaN Schottkydiodes by electrochemical surface treatment,” Appl. Phys. Lett., vol. 82,no. 8, pp. 1293–1295, Feb. 2003.

[6] V. Adivarahan, G. Simin, J. W. Yang, A. Lunev, M. A. Khan, N. Pala,M. Shur, and R. Gaska, “ SiO2-passivated lateral-geometry GaN transpar-ent Schottky-barrier detectors,” Appl. Phys. Lett., vol. 77, no. 6, pp. 863–865, Aug. 2000.

[7] M. L. Lee, J. K. Sheu, W. C. Lai, S. J. Chang, Y. K. Su, M. G. Chen,C. J. Kao, G. C. Chi, and J. M. Tsai, “GaN Schottky barrier photodetectorswith a low-temperature GaN cap layer,” Appl. Phys. Lett., vol. 82, no. 17,pp. 2913–2915, Apr. 2003.

[8] H. Jiang, T. Egawa, H. Ishikawa, Y. B. Dou, C. L. Shao, and T. Jimbo,“Barrier-height-enhanced n-GaN Schottky photodiodes using a thinp-GaN surface layer,” Jpn. J. Appl. Phys., vol. 43, no. 7A, pp. 4101–4104,Jul. 2004.

[9] K. Matsumoto and A. Tachibana, “Growth mechanism of atmosphericpressure MOVPE of GaN and its alloys: Gas phase chemistry and itsimpact on reactor design,” J. Cryst. Growth, vol. 272, no. 1–4, pp. 360–369, Dec. 2004.

[10] S. Nakamura, “GaN growth using GaN buffer layer,” Jpn. J. Appl. Phys.,vol. 30, no. 10A, pp. L1 705–L1 707, Oct. 1991.

[11] M. Leyer, J. Stellmach, C. Meissner, M. Pristovsek, and M. Kneissl, “Thecritical thickness of InGaN on (0001)GaN,” J. Cryst. Growth, vol. 310,no. 23, pp. 4913–4915, Nov. 2008.

[12] F. A. Padovani and R. Stratton, “Field and thermionic-field emissionin Schottky barriers,” Solid State Electron., vol. 9, no. 7, pp. 695–707,Jul. 1966.

[13] T. Hashizume, J. Kotani, and H. Hasegawa, “Leakage mechanism inGaN and AlGaN Schottky interfaces,” Appl. Phys. Lett., vol. 84, no. 24,pp. 4884–4886, Jun. 2004.

[14] C. Y. Chang and S. M. Sze, “Carrier transport across metal-semiconductor barriers,” Solid State Electron., vol. 13, no. 6, pp. 727–740,Jun. 1970.

[15] C. R. Crowell and V. L. Rideout, “Normalized thermionic-field (T-F)emission in metal-semiconductor (Schottky) barriers,” Solid StateElectron., vol. 12, no. 2, pp. 89–105, Feb. 1969.

[16] P. Hacke, T. Detchprohm, K. Hiramatsu, and N. Sawaki, “Schottky barrieron n-type GaN grown by hydride vapor phase epitaxy,” Appl. Phys. Lett.,vol. 63, no. 19, pp. 2676–2678, Nov. 1993.

[17] M. E. Levinshtein, S. L. Rumyantsev, and M. S. Shur, Properties ofAdvanced Semiconductor Materials: GaN, AIN, InN, BN, SiC, SiGe.New York: Wiley, 2001.

[18] O. Ambacher, J. Majewski, C. Miskys, A. Link, M. Hermann,M. Eickhoff, M. Stutzmann, F. Bernardini, V. Fiorentini, V. Tilak,B. Schaff, and L. F. Eastman, “Pyroelectric properties of Al(In)GaN/GaNhetero- and quantum well structures,” J. Phys.: Condens. Matter, vol. 14,no. 13, pp. 3399–3434, Apr. 2002.

[19] H. Zhang, E. J. Miller, E. T. Yu, C. Poblenz, and J. S. Speck,“Measurement of polarization charge and conduction-band offset atInxGa1−xN/GaN heterojunction interfaces,” Appl. Phys. Lett., vol. 84,no. 23, pp. 4644–4646, Jun. 2004.

[20] J. K. Sheu and G. C. Chi, “The doping process and dopant characteristicsof GaN,” J. Phys.: Condens. Matter, vol. 14, no. 22, pp. R657–R702,Jun. 2002.

[21] B. Heying, X. H. Wu, S. Keller, Y. Li, D. Kapolnek, B. P. Keller,S. P. DenBaars, and J. S. Speck, “Role of threading dislocation structureon the X-ray diffraction peak widths in epitaxial GaN films,” Appl. Phys.Lett., vol. 68, no. 5, pp. 643–645, Jan. 1996.


[22] J. W. P. Hsu, M. J. Manfra, R. J. Molnar, B. Heying, and J. S. Speck,“Direct imaging of reverse-bias leakage through pure screw dislocationsin GaN films grown by molecular beam epitaxy on GaN templates,” Appl.Phys. Lett., vol. 81, no. 1, pp. 79–81, Jul. 2002.

[23] P. Fini, X. Wu, E. J. Tarsa, Y. Golan, V. Srikant, S. Keller, S. P. Denbaars,and J. S. Speck, “The effect of growth environment on the morphologicaland extended defect evolution in GaN grown by metalorganic chemicalvapor deposition,” Jpn. J. Appl. Phys., vol. 37, no. 8, pp. 4460–4466,Aug. 1998.

[24] S. N. Lee, J. K. Son, H. S. Paek, T. Sakong, W. Lee, K. H. Kim,S. S. Kim, Y. J. Lee, D. Y. Noh, E. Yoon, O. H. Nam, and Y. Park, “Growthpressure dependence of residual strain and threading dislocations in theGaN layer,” Phys. Stat. Sol. (C), vol. 1, no. 10, pp. 2458–2461, Jul. 2004.

[25] J. Bardeen and W. Shockley, “Deformation potentials and mobilities innon-polar crystals,” Phys. Rev., vol. 80, no. 1, pp. 72–80, Oct. 1950.

[26] H. Y. Peng, M. D. McCluskey, and Y. M. Gupta, “Shock-induced band-gap shift in GaN: Anisotropy of the deformation potentials,” Phys. Rev. B,Condens. Matter, vol. 71, no. 11, p. 115 207, Mar. 2005.

[27] S. L. Chuang and C. S. Chang, “K.p method for strained wurtzite semicon-ductors,” Phys. Rev. B, Condens. Matter, vol. 54, no. 4, pp. 2491–2504,Jul. 1996.

Wei Lu (S’10) received the B.S. degree in opticaland electrical engineering from Zhejiang University,Hangzhou, China, in 2005. He is currently work-ing toward the Ph.D. degree in electrical and com-puter engineering with the University of CaliforniaSan Diego, San Diego.

His current research interests include III–nitride-semiconductor-material metal–organic-chemical-vapor-deposition growth and related high-powerhigh-speed devices.

Lingquan (Dennis) Wang (S’04–M’10) receivedthe B.E. degree in electrical engineering fromTsinghua University, Beijing, China, in 2002 andthe Ph.D. degree in electrical and computer engi-neering from the University of California San Diego,San Diego.

He is currently with Suvolta Inc., Los Gatos, CA,for the meantime, holding a Visiting Scholar positionwith the Department of Electrical and ComputerEngineering, University of California San Diego. Hisresearch interests include GaN-based high-power

high-speed devices, III–V metal–oxide–semiconductor field-effect transistors(FETs), and tunneling FETs.

Siyuan Gu (S’10) received the B.S. degree inphysics from Fudan University, Shanghai, China,in 2005. He is currently working toward the Ph.D.degree in electrical and computer engineering withthe University of California San Diego, San Diego.

His current research interests include nitride elec-tronics for radio-frequency applications and novelnanometer-scaled metal–oxide–semiconductor field-effect transistors based on III–V compounds as wellas silicon.

David P. R. Aplin, photograph and biography not available at the time ofpublication.

Daniel M. Estrada received the B.S. degree in en-gineering physics from Cornell University, Ithaca,NY, in 2008 and the M.S. degree in materials scienceand engineering from the University of California,San Diego, La Jolla, in 2009 and continued studyingthere until the end of 2010. He received the Depart-ment of Homeland Security Fellowship, in 2009.

His fields of interest include GaN-based tech-nologies for national security and renewable energy,specifically radiation detection and photovoltaics.

Paul K. L. Yu (M’83–SM’91–F’08) received theB.S., M.S., and Ph.D. degrees from California In-stitute of Technology, Pasadena, in 1979, 1979, and1983, respectively.

Currently, he is a Professor with the Departmentof Electrical and Computer Engineering, Universityof California San Diego (UCSD), San Diego. Atthe UCSD, he conducts research on semiconductormaterials and devices for various photonics and mi-crowave photonics applications. His current researchinterests include high-speed and high-power optical

detectors and modulator devices for analog fiber links and high-power semi-conductor optical switches for microwave generation.

Prof. Yu is a Fellow of the Optical Society of America and the InternationalSociety for Optics and Photonics.

Peter M. Asbeck (M’75–SM’97–F’00) received theB.S. and Ph.D. degrees from Massachusetts Instituteof Technology (MIT), Cambridge, in 1969 and 1975.

He worked with the Sarnoff Research Center,Princeton, NJ, and with Philips Laboratory, Briar-cliff Manor, NY, in the areas of quantum electron-ics and GaAlAs/GaAs laser physics. In 1978, hejoined the Rockwell International Science Center,where he was involved in the development of high-speed devices and circuits using III–V compoundsand heterojunctions. He pioneered efforts to develop

heterojunction bipolar transistors based on GaAlAs/GaAs and InAlAs/InGaAsmaterials. In 1991, he joined the University of California San Diego, San Diego,and became the Skyworks Chair Professor with the Department of Electricaland Computer Engineering. He is the author of more than 350 publications.Hisresearch interests include the development of high-performance transistortechnologies and their circuit applications.

Dr. Asbeck is a member of the National Academy of Engineering and hasbeen a Distinguished Lecturer of the IEEE Electron Devices Society and of theMicrowave Theory and Techniques Society. He was the recipient of the 2003IEEE David Sarnoff Award for his work on heterojunction bipolar transistors.

Documents

Analysis of Reverse Leakage Current and Breakdown Voltage in GaN and InGaN/GaN Schottky Barriers