Silicon carbide schottky diode

Abstract

A SiC Schottky diode which includes a Schottky barrier formed on a silicon face 4H—SiC body.

Description

FIELD OF INVENTION

The present invention relates to Schottky diodes and in particular to SiC Schottky diodes.

BACKGROUND ART

Although the main intrinsic parameters in Silicon Carbide material have not been exhaustively studied, several experimental and theoretical studies have been performed in recent years in order to better describe the current transport in ohmic and rectifying contact on SiC.

It has been known that in Schottky diodes the metal semiconductor interface (MST) between the Schottky barrier metal and the semiconductor plays a crucial role in the electrical performance of electronic devices. Many factors can worsen the performance of the MSI in a Schottky diode. For example, the quality of the semiconductor surface prior to the deposition of the Schottky barrier metal can cause the device to exhibit characteristics that are different from the ideal characteristics.

Current-voltage (I-V) and capacitance-voltage (C-V) characterizations are useful methods for determining the Schottky barrier height from which the MSI uniformity can be obtained. It has been known that the barrier height determined from C-V measurements often exceeds the value extracted from forward I-V measurements. The discrepancy is also seen to be greater for Schottky contacts with high ideality factor, i.e. poor Schottky MSI quality. Sullivan et al., J. Appl. Physics, 70, No. 12, (1991), 7403-7424, have explained that such a discrepancy is due to barrier height inhomogeneities, and have demonstrated that if the barrier height is assumed to follow some statistical distribution, the barrier height value extracted from C-V measurements is the arithmetic mean of the barrier height distribution (the capacitance of the Schottky contact comes from the modulation of the total space-charge as a function of bias). On the other hand, the value from I-V characterization is well below the value extracted from capacitance analysis, and is dominated by the current that flows through regions of lower barrier height. Also, the inhomogeneity of the barrier height is believed to be a primary reason for the observation of large ideality factors (n>1.1). For intimate contact and uniform MSI, the barrier height values extracted from C-V and I-V characterization are expected to be in close correlation, and ideality factors very close to unity.

SUMMARY OF THE INVENTION

A semiconductor device according to the present invention is a SiC Schottky diode which includes a SiC substrate of one conductivity, a silicon face SiC epitaxial body of the one conductivity formed on a first surface of a the SiC substrate, a Schottky metal barrier formed on the silicon face of the SiC epitaxial body, a back power electrode on a second surface of the SiC substrate opposite said first surface of the SiC substrate.

In one embodiment of the present invention the Schottky metal barrier is formed of Titanium, and the silicon face epitaxial SiC is 4H—SiC. It has been found that this combination results in a very uniform MSI, which allows for accurate electrical characterization of the device.

Other features and advantages of the present invention will become apparent from the following description of the invention which refers to the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a cross-sectional view of a portion of a SiC semiconductor device according to the present invention.

FIG. 2 graphically illustrates the I-V characteristics of a device according to the present invention as a function of temperature.

FIG. 3 illustrates a Richardson's plot of In(Js/T**2) versus 1/kT, experimental data and linear fit.

DETAILED DESCRIPTION OF THE FIGURES

A power semiconductor device according to the preferred embodiment of the present invention is a discrete Schottky diode that includes SiC substrate 10 (preferably 4H—SiC bulk SiC) of N-type conductivity, silicon face 4H—SiC epitaxial body 12 of N-type conductivity formed on one surface of substrate 10, Schottky metal body (Schottky barrier) 14 formed over and in Schottky contact with epitaxial body 12, and a back power electrode 16 formed on another opposite surface of substrate 10. A region of P-type conductivity 18 serving as a guard ring is diffused into epitaxial body 12, is in contact with Schottky body 14, and defines the other boundary of the active region (i.e. region defined by the Schottky contact between Schottky body 14 and epitaxial body 12) of the device. A device according to the preferred embodiment further includes a field insulation body 20 disposed on epitaxial body 12 and surrounding the active region of the device. Note that field insulation body 20 includes sidewalls 22 that taper toward the active region of the device, and that Schottky body 14 extends over sidewalls 22 and a portion of the top surface of field insulation body 20. A device according to the preferred embodiment further includes a metallic front power electrode 24 which is disposed over at least Schottky body 14.

In the preferred embodiment, Schottky body 14 is composed of titanium (which has been observed to make a uniform MSI with silicon face 4H—SiC), front power electrode 24 is composed of aluminum, back power electrode 16 is composed of a trimetal stack of titanium (in ohmic contact with substrate 10), nickel (formed on the titanium layer), and silver (formed on the nickel layer). Field insulation body 20 is preferably composed of silicon dioxide.

The behavior of the electrical parameters of a Schottky barrier diode (SBD) fabricated on the Si face of a 4H—SiC epitaxial layer according to the present invention was studied. The study revealed that the devices according to the present invention exhibit an electrical behavior in accordance to thermoionic emission and a good MSI uniformity confirmed by C-V measurements. I-V measurements (I-V) were also performed in a large temperature range which led to an evaluation of the Richardson constant.

Details of the Study

Schottky barrier diodes were fabricated on 3″ 4H—SiC wafers, production grade, produced by Cree Inc. Epitaxial body 12 and substrate 10 of wafers were n-doped (Nd=10¹⁶ cm⁻³, thickness 7 μm and Nd=10¹⁸ cm⁻³, thickness 380 μm, respectively). The junction extermination extension on fabricated devices was formed by a guard ring 18 obtained by P⁺ type implantation and a mesa structure with an additional ring of silicon oxide 20. Schottky barrier 14 was made by thermal evaporation of Titanium and a successive Aluminum layer 24. Ohmic contact 16 formation was made on the back-side of the wafer by a triple evaporation of Titanium, Nickel and Silver.

Fabricated devices were then packaged in a standard TO-220 commercial package.

All step processes were optimized in previous work, leading to obtain a percentage of working diodes with a reverse current<50 μA@600 V of reverse voltage up to 85%.

Electrical measurements were performed by an SMU237 Keithley Source Measure Unit and an SMU238 Keithley Source Measure Unit. Doping concentration was controlled by C-V measurements at the standard frequency of 1 MHz (HP 4192A LF).

Electrical characterization versus temperature was performed by using an Oxford cryostat. The measurements were performed on selected devices which passed reliability tests working 1000 hours in stress conditions.

Results

In order to avoid any problems related to fabrication process and surface preparation, devices were selected with behavior close to ideal.

A first electrical characterization was done by standard I-V and C-V measurements fitting data according to thermoionic emission theory (TET) in order to obtain ideality factor, barrier height, doping concentration and reverse current at 600 V. The mean values were 1.02, 1.21 eV, 810¹⁵ cm⁻³ and 30 μA. Break down voltage was higher than 1100 V, close to the ideal value. Twenty selected devices were tested further by voltage measurement versus time up to 1000 hour. The results demonstrated an optimal stability with unchanged characteristic.

Electrical measurements were performed in the range 77-300 K with a temperature step of 20 K by means of forward and reverse current voltage analysis. FIG. 2 reports the semi logarithmic plot of the I-V curves of SBD, showing a linear behavior over seven orders of magnitude.

Table 1 reports barrier height and ideality factor versus temperature extracted from experimental data, according to TET and by using the classical value of 146 A K⁻² cm⁻². It is worth noting that the ideality factor is close to 1.1 up to 77 K, and in the range 200-300 K is almost constant. This leads to the conclusion that the barrier/metal interface is very homogeneous which is confirmed also by the Schottky height barrier value obtained by C-V measurements (1.2 eV).

TABLE 1
EXPERIMENTAL VALUES OF SCHOTTKY BARRIER
HEIGHT AND IDEALITY FACTOR.
TEMP	BARRIER HEIGHT
(K)	(eV)	IDEALITY FACTOR
77	1.11 ± 0.2	1.13 ± 0.2
100	1.14 ± 0.2	1.1 ± 0.2
120	1.15 ± 0.2	1.09 ± 0.2
140	1.17 ± 0.1	1.07 ± 0.1
160	1.17 ± 0.1	1.07 ± 0.1
180	1.19 ± 0.1	1.05 ± 0.1
200	1.20 ± 0.1	1.04 ± 0.1
220	1.20 ± 0.1	1.04 ± 0.1
240	1.21 ± 0.1	1.03 ± 0.1
260	1.21 ± 0.1	1.03 ± 0.1
280	1.21 ± 0.1	1.03 ± 0.1
300	1.21 ± 0.1	1.03 ± 0.1

In order to obtain additional information, for the determination of the effective barrier height and of the AA product, saturation current density was extracted from experimental data in the range 200-300 K and reported in a Richardson's plot (FIG. 3).

From the slope of the linear fit and from the intercept, one can obtain an effective value of barrier height of 1.16±0.1 eV and an effective Richardson's constant A=17±8 A K⁻² cm⁻². The Richardson's constant is different from the classical theoretical value. Many have tried to explain and model the Schottky contacts on SiC, in order to fit experimental data to the theoretical value. In theory, the value of the effective Richardson constant A* should be calculated for a semiconductor with indirect energy gap and a number M_C of equivalent ellipsoidal constant energy surface in the first Brillouin zone, by considering diagonal components of the effective mass tensor. Such a theoretical calculation leads to Richardson constant values that are dependent on the particular crystallographic direction considered in Thermionic emission. The commonly reported value for the 4H—SiC Richardson constant was firstly calculated by loth et al., ISPSD '95 (1995) 101-106. It can be shown that such a value was obtained by simply averaging the effective mass value in Gotz et al., J. Appl. Phys., 73, No. 7, (1993) 3332-2228. neglecting any crystallographic anisotropy, and considering a number of equivalent conduction band minima equal to 6 (derived from the conduction band minimum location along the M-K edge of the first Brillouin zone reported in Tairov et al., “Electroluminescence”, J. I. Parkov, Ed., Berlin-Heidelberg-New York, Springer-Verlag (1977). The location of the equivalent conduction band minima has been more recently demonstrated to be exactly at the M-point of the first Brillouin zone, and symmetry considerations lead to a number of equivalent minima M_C equal to 3 (the constant energy surfaces are 6 semi-ellipsoids). Such a value for M_C should be employed in the calculation of the effective Richardson constant of 4H—SiC.

Moreover, a modified Richardson constant A** should be derived from A* taking into account quantum mechanical tunneling and reflections at the Schottky interface and interactions of the emitted electrons with optical phonons, as suggested in Crowell et al., Solid-State Electron., 9 (1966) 1035-1048.

Such a complete model has never been applied to hexagonal materials, and to 4H—SiC in particular. Thus, the commonly accepted value of 146 A/cm²V² was employed because it was estimated that an error of 30% in A* would affect the value of Schottky barrier height φ_B extracted from thermionic emission saturation current measurements of about 1%.

Although the present invention has been described in relation to particular embodiments thereof, many other variations and modifications and other uses will become apparent to those skilled in the art. It is preferred, therefore, that the present invention be limited not by the specific disclosure herein, but only by the appended claims.

Claims

1. A semiconductor device, comprising: a SiC substrate;a silicon face SiC epitaxial body formed on a first surface of a said SiC substrate;a Schottky metal barrier formed on said silicon face of said SiC epitaxial body; anda back power electrode on a second surface of said SiC substrate wherein said Schottky metal barrier is configured to exhibit a voltage characterization of the barrier that reflects uniformity between said Schottky metal barrier and said silicon.

2. The semiconductor device of claim 1, wherein said Schottky metal barrier is comprised of Titanium.

3. The semiconductor device of claim 1, wherein said back electrode is comprised of a stack of Titanium, Nickel, and Silver.

4. The semiconductor device of claim 1, a region of another conductivity diffused into said SiC epitaxial body defining an outer boundary of an active region of said device.

5. The semiconductor device of claim 1, further comprising a field insulation body formed on said SiC epitaxial body, wherein said Schottky metal barrier extends over said field insulation body.

6. The semiconductor device of claim 5, wherein said field insulation body is comprised of silicon dioxide.

7. The semiconductor device of claim 5, wherein said field insulation body includes tapered sidewalls tapering toward said SiC epitaxial body.

8. The semiconductor device of claim 1, further comprising a front power electrode formed on said Schottky metal body.

9. The semiconductor device of claim 8, wherein said front power electrode is comprised of aluminum.

10. The semiconductor device of claim 1, wherein said SiC epitaxial body is comprised of 4H—SiC.

11. The semiconductor device of claim 10, wherein said SiC is comprised of 4H—SiC.

12. A Schottky diode comprising: a silicon face SiC epitaxial body; anda Schottky metal barrier formed on said silicon face of said SiC epitaxial body,wherein said Schottky metal barrier is configured to exhibit a voltage characterization of the barrier that reflects uniformity between said Schottky metal barrier and said silicon.

13. The Schottky diode of claim 12 further comprising: a back electrode comprising Titanium, Nickel, and Silver.

14. The Schottky diode of claim 12 further comprising: a region of another conductivity diffused into said SiC epitaxial body defining an outer boundary of an active region of said Schottky diode.

15. The Schottky diode of claim 14 wherein said region of another conductivity forms a guard ring around said active region.

16. The Schottky diode of claim 12 wherein said SiC comprises 4H SiC.

17. The Schottky diode of claim 12 further comprising a ring of silicon oxide disposed between a portion of said Schottky metal and said SiC epitaxial body.

18. The Schottky diode of claim 12 characterized as having an ideality factor of less than 1.1 for temperatures above 140 K.

19. The Schottky diode of claim 12 characterized as having an ideality factor of less than or equal to 1.05 for temperatures above 200 K.

20. The Schottky diode of claim 12 characterized as having a breakdown voltage of greater than 1100 volts.