JPH03257870A - Schottky barrier diode - Google Patents

Abstract

PURPOSE:To make a recombination current value approximately constant in a voltage region higher than a pinch off voltage (Vp) by providing metal having high phiR to a bottom section and a side of a recessed section of an irregular semiconductor surface and by providing metal having low phi to an approximately planar surface of the upper part of a projecting section. CONSTITUTION:When a backward voltage is applied, a recombination current in a depletion layer extending from a junction formed by a metal 2 occupies a large part of a backward current until a depletion layer (i) spreading from a metal 1 fills a width (a) of a projecting section. Electric field E applied to a Schottky barrier junction formed by the metal 2 is almost fixed from VP and a depletion layer (ii) wherein the width (a) of the projecting section is buried by a depletion layer which extends from a junction of the metal 1. If a voltage is further applied after VP, a recombination current between a depletion layer (iii) which extends from the metal 1 until a breakdown voltage VR and a delpletion layer formed by a metal having a large phiB value becomes a relatively small leak current value.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a Schottky barrier diode which has better rectification characteristics and less loss than conventional ones.

A Schottky barrier diode has a structure as shown in Figure 1, which rectifies using the potential barrier created by contact between metal and semiconductor (in the figure, (6) is a silicon substrate, (4) is an epitaxial layer, and (1) is a barrier metal). These diodes are often used for power applications because they are faster, have a lower positive rising voltage, and have lower losses than other diodes. In particular, the recent demand for lower driving voltages for integrated circuits has been increasing, and there is a strong demand for Schottky barrier diodes with even lower losses.

In order to achieve this, it is necessary to realize a rectifier with good rectification characteristics, which has smaller forward voltage drop and reverse current than the current ones, and has small Guyaode loss, that is, the sum of forward and reverse direction losses. However, the forward voltage drop and reverse current of the Schottky barrier diode are determined by the material of the barrier metal forming the Schottky junction, as shown in the qualitative relationship diagram shown in FIG. In general, a device with a small forward voltage drop has a large reverse current, and a device with a small reverse current has a large forward voltage drop, which are contradictory characteristics.

For example, if we look at the positive direction voltage drop in conventionally known metals, as shown in Figure 2, titanium (Ti), lithium-ion (Cr)
, molybdenum (Mo
), which tends to increase the forward loss, while the reverse current tends to increase in the order of
, chromium (Cr), and titanium (Ti), which tend to increase the reverse direction loss.

Therefore, for diodes whose loss is given by the sum of forward and reverse losses, it is necessary to choose a material that achieves the lowest loss by balancing both forward and reverse losses, and currently molybdenum (Mo) is the most commonly used material. It is being However, in order to realize a Schottky barrier diode with even lower loss using the current structure, difficult issues such as the development of new materials must be solved.

Therefore, the applicant of the present application previously proposed a positive diode made of titanium (Ti) or the like. (Japanese Patent Application No. 1-146573) The present invention relates to an improvement of such a diode, and in a semiconductor device having a metal forming a Schottky barrier on the surface of a semiconductor substrate, a plurality of metals having different barrier heights (φ■) are used. It is provided on an uneven semiconductor surface, and a metal having a high φ is provided on the bottom and side surfaces of the concave portion, and a metal having a low φ is provided on a substantially flat surface of the upper portion of the convex portion. do.

3(a) and 3(b) are a plan view and a sectional view showing an embodiment of the present invention, and FIG. 4 is an explanatory diagram of its operation. 2 is the barrier height φ. 3 is a surface protective film, 4 is a semiconductor (for example, N-type silicon crystal), and 5 is a layer of conductivity type opposite to that of 4 formed in the semiconductor 4, which is radiationally referred to as a guard ring. This is an area called 6 is a low resistance semiconductor layer of the same conductivity type as 4, and 7 is an electrode metal.

Note that the surface of the semiconductor 4 is formed in an uneven shape, and the bottom and side surfaces of the recesses are formed with a barrier height of φ1. A bond is formed with a metal having a large barrier height of φ8, and a bond is formed with a metal having a small barrier height of φ8 on the substantially flat surface of the upper portion of the convex portion.

Incidentally, the barrier height φ is Ti (0,5ev), Cr
(0,61ev), Mo (0,88ev) Al (0
, 74ev), Au (0,8ev).Next, the operating principle of the Schottky barrier diode having the structure of the present invention will be explained below.

This structure is expressed as an equivalent circuit in which a Schottky barrier diode with a small barrier height and a Schottky barrier diode with a large barrier height are connected in parallel.

Therefore, when a forward electric field is applied, with the A electrode being positive and the B electrode being negative, electrons first flow from the semiconductor 4 to the metal 2 in the junction with a small barrier height than in the junction with a large barrier height. When the metal 2 is made of Ti and the metal 1 is made of AI, the JFVP characteristics of a Schottky barrier diode with a small barrier height become dominant up to a current density of about 200 A/-. Then, at a current density of 300 A/car or more, the overlap of the forward current across the junction area of the large barrier height Schottky diode and the barrier height φ■ helps the difficulty in flowing a large forward current with a small barrier junction area. It works like this.

Therefore, in the structure of the present invention, the forward direction characteristics are slightly inferior compared to the SBD characteristics of the conventional structure.

On the other hand, in the reverse direction characteristic, when a negative voltage is applied to the A electrode and a positive voltage is applied to the B electrode, a depletion layer is formed on the semiconductor side due to the metal/semiconductor junction, and expands as the reverse applied voltage increases.

However, the width of the depletion layer W at a junction with a large Schottky barrier height φ■ is large, and the recombination current within the depletion layer is smaller as φ■ becomes larger.Of course, if the junction area is small, reverse leakage occurs. The current becomes smaller.

When a reverse voltage is applied, the metal l
The recombination current within the depletion layer extending from the junction formed by the metal 2 accounts for most of the reverse current until the depletion layer (a) expanding from the junction fills the width a of the convex portion.

Since the width a of the convex portion is filled with a depletion layer extending from the metal 1 junction (mouth) and the pinch-off voltage ■, the electric field E applied to the Schottky barrier junction formed by metal 2 is almost fixed, and then reversed. Even if the directional voltage increases, the electric field E at the two-metal junction does not increase, so the recombination currents JgI and J2 that cross the two-metal Schottky barrier have approximately constant values in the voltage region higher than (2).

In other words, the leakage current across the junction with a small Schottky barrier height φ■ can be suppressed to a small value after the voltage ■.

■If voltage is applied again after that, the depletion layer extending from metal 1 will extend until the voltage breakdown Vn (c), but the recombination current JIIIJI of the depletion layer formed by the metal with a large φ8 value will be a relatively small leakage current value. .

<Example> N is deposited on a silicon epitaxial wafer with a Ti film (φn”
0. 5 e v ) and ALIf4 (φn=0.74e
An example of forming V) will be described below.

On a silicon substrate doped with arsenic impurity atoms with a specific resistance of 0.003 Ω and a thickness of 400 μm (u), an epitaxial silicon layer with a specific resistance of 0.5 Ω and a thickness of 6 μm (i) with phosphorus as an impurity atom is deposited.

A SiO* film with a thickness of approximately 1 μm is formed by steam oxidation treatment, and a first photo process is performed to remove the oxide film only on the card ring area.Then, the guard ring is removed using a buffer etchant. The window opening P″″″diffusion m(c)
form.

Next, a second photograph is applied to remove the oxide film in the area where Ti is to be deposited, and prior to the Ti deposition, the surface Si containing various crystal lattice strains is
Etch people and remove them. Immediately after the pretreatment, Ti
is deposited by electron beam evaporation. The film thickness will be approximately 3,000 to 3,500 people.

A tertiary photograph is applied, and the Ti film is wet-etched using a photographic pattern that leaves a 2.5×2.5 μm square Ti pattern. Ti etching solution composition: H2O, 1 volume, ED
When a mixture of 2 volumes of TA is heated to 65° C.±2° C., the Ti film can be etched in about 3 minutes.

In addition, the area inside the guard ring is 0.01cn+, and as a 4th order photo, the Ti film part is covered with photoresist from above the Ti pattern, and the silicon part other than the Ti part is etched into a square of 2 x 2 μm. The pattern is photoprocessed using stepper alignment that allows microfabrication of about 0.75 μm. At this time, a 6,000-layer thick positive resist is applied.

RI E (Reactive Ion Etcher)
) to introduce CC12F2 gas, approximately 5 mL
After adjusting to orr, apply about 2KW of power and about 3μ
Etch a depth (h) of m and an opening of approximately 3 μm (f) in approximately 7 to 8 minutes.

Etch into a U-shape. Thus a# 24m,
5-3μ, h=3μ uneven shape (8) is almost completed, T
The Si side etch extends to the bottom of the i film, causing the Ti film to protrude, but this protruding Ti is removed by etching with the Ti etching solution for about 30 seconds. In addition, Si
The etching shape can be controlled by selecting the type of Si Etch Gas such as CF, NF, F series such as 1, CC1, CC1, etc.
By adjusting the ratio of CI gas such as Fm, it is also possible to adjust the distance of the upper opening with respect to the depth.

Next, in order to remove the electron beam damage layer, approximately 20 to 5
00 Si was removed using a known Si etching solution.

At this time, the electron beam damaged layer is usually removed by thermal annealing, but in order to realize the structure of the present invention, extreme care must be taken not to thermally destroy the Ti barrier.

1. Immediately after the Si etching process for removing the damaged layer, AL
Deposit. In order to make the AL sufficiently rotate around the bottom side of the U-shape Sj, the deposition incident angle, the rotation and revolution angles, and rotational speed of the wafer were adjusted, and the improvements were made to a level that poses no problem in practical use. A
The film thickness of L was about 5 mm, and Cr and Ni were subsequently deposited in the same vacuum deposition chamber by a normal method. Cr is A
It serves as a diffusion barrier metal for L and Ni. Also, Si
The back side of the wafer is also subsequently treated with Cr/Ni deposition.

Next, apply AL/Cr only to the required area of the A electrode on the pattern surface.
The fifth photograph is taken so that /N i exists.

The Ni etching solution is a ferric chloride based etching solution.
Cr is a ceric nitrate-based etching solution, and A
L was etched using a known H2PO4-based etching solution.

Thereafter, Pb-8n solder was melted on the Ni surface of the wafer, the chip was diced, and a Schottky barrier diode chip was completed through the usual process.

Through the above manufacturing process, the inner area of the guard ring is 0.01
cnf, Ti Schottky area 7.48X10-3
d, a=24mAL Schottky region f・3μm was completed. Figures 5(a) and 5(b) are characteristic diagrams of the diode of the present invention in comparison with a conventional example, in which (a) is a forward characteristic diagram, and b is a reverse characteristic diagram. ) shows the characteristics of the embodiment of the present invention, that is, the forward characteristic (b) of this embodiment is V, = 0, 33 volt (a t 200A+
eP/cm), and Ti 5BD with conventional structure (a)
V, = 0.294 volt. On the other hand, in the reverse direction characteristic, in this example, as shown in characteristic (b), a reverse direction leakage current (IR) of about IR=1.0 mA was obtained at breakdown voltage V + and #52 volt point. With a T1 barrier diode of conventional structure, I R = 13 mA, which can be reduced by a wide margin. Therefore, when applied to a rectifier circuit, the power loss could be reduced to about 27'3 compared to the conventional structure Ti barrier SBD.

According to the principles of the present invention, the effects of the small barrier bonding area, the large barrier bonding area, and the uneven shape dimensions a and f are such that the larger the dimension a is, the more desirable it is from the forward direction characteristics of the small barrier metal. There are contradictory factors that make it desirable for the effect to be smaller, and the experimental results show that 2 = T for barrier metals.
M on i and 1.

When using the above φ■ metals, approximately 1 to 2 μm is optimal, and f is approximately φ1. The larger f is, the less it contributes to the forward characteristics in the practical current range, and the forward characteristics are mostly determined by the area occupied by the metal 2. Therefore, it is desirable that f be as small as possible. but,
Considering manufacturing technology, manufacturing yield, etc., f=2~5μ
As is clear from the explanation that m is 0 or more as the optimal point, the present invention provides a low-loss Schottky barrier diode suitable for power use.

[Brief explanation of drawings]

Figure 1 is an explanatory diagram of a conventional Schottky barrier diode.
FIG. 2 is a rectification characteristic diagram of a conventional Schottky barrier diode using various barrier metals, FIGS. 3(a) and 3(b) are a plan view and a sectional view of an embodiment of the present invention, FIG. FIG. 5 is a characteristic diagram of the embodiment of the present invention in comparison with the conventional example.

Claims (1)

[Claims] In a semiconductor device having a metal that forms a Schottky barrier on the surface of a semiconductor substrate, a plurality of metals having different barrier heights (φ_■) are provided on the uneven semiconductor surface, and a high φ_■ is provided at the bottom and side surfaces of the recess. 2. A Schottky barrier diode characterized in that a metal having a low φ_■ is provided on a substantially flat surface of the upper part of the convex portion.