JPH0267766A - Bipolar-type semiconductor switching device - Google Patents

Abstract

PURPOSE:To enhance an ON characteristic and an OFF characteristic by a method wherein a separation groove is formed between an anode region on the anode side and a short drain region and these regions are wired independently so that a sufficiently high forward bias and a sufficiently high reverse bias can be applied during an ON operation and an OFF operation. CONSTITUTION:Anode wiring parts 9 and short region wiring parts 10 are formed individually in anode regions 2 and short drain regions 8; the anode regions 2 and the short drain regions 8 are separated by separation grooves 22. Thereby, a high forward bias and a high reverse bias can be applied between the anode regions 2 and the short drain regions 8 independently of a gate bias. When the high forward bias is applied and flowe 15 of electrons are injected toward the anode regions 2 from the short drain regions 8, holes are injected to a substrate 1 from the anode regions 2, and an ON operation is completed more quickly.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a bipolar semiconductor switching device.

[Conventional technology]

As an example of the prior art, the structure of a drain short type IGBT will be explained as an example.

FIG. 8 shows a schematic cross-sectional view of a conventional short drain type IGBT. In the figure, (1) is a high resistivity semiconductor substrate (usually N-based, i), and this substrate (1) has a P-type impurity diffusion region (hereinafter referred to as an anode region) that becomes a drain.
21, an N-type impurity diffusion region (hereinafter referred to as a cathode region) (3) serving as an N+ source region is formed. Anode area (
2) P-type impurity diffusion where the gate (5), which controls the on/off state of the main current flowing between the cathode region (3) and the cathode region (3), becomes a path for the main current (2) through the gate insulating film (6) Area (4
) It is formed on the N channel region (6) of the instep (hereinafter referred to as P well). The cathode region (3) is short-circuited with the P-well (4) and is metal wired like a cathode wire (7). The anode region (2) (drain) also has an N-type impurity region (8) (which is formed short-circuited with the substrate (1)).
The anode wiring (9) is connected in a short-circuited manner to a region (hereinafter referred to as a short drain region). In this way, the anode region (2
) (drain) and the short drain region (8) is formed in a short-circuited manner, so it is generally called a drain short type.

Next, the operation of the drain short type IGBT will be explained.

First, in the on operation, by applying a positive voltage to the gate, the portion of the P well (4) close to the surface of the substrate (1) becomes an N channel region (b), and turn-on begins.

Through the N channel region qυ, electrons are injected from the cathode region (3) (source) to the substrate (1), which becomes the N base, and along with this, holes are injected from the anode region (2) (drain). be done. When the ON operation is completed, carriers accumulate in the substrate +11, which is the N base layer, and the substrate (1), which was originally a high resistivity layer, undergoes conductivity modulation.
The on-resistance is about an order of magnitude smaller than that of a normal power MOS FET.

Next, the off operation is performed by removing the voltage applied to the gate (5) and making it zero. Gate (5)
f [When the pressure becomes zero, the N channel region (b) disappears and returns to the original P well (4), so the cathode region (3)
The injection of electrons into the base region of the substrate (1) is cut off and the state shifts to an off state, and in the final stage, electrons are formed in the P well (4) of the P-type impurity region and the N- region of the substrate (1). The depletion layer spreads over the entire substrate (1) up to this side of the anode region (2), and the OFF operation is completed. Until then, in the transient state, holes remained in the substrate (1), and the holes formed in the anode region (2), the substrate (1), and the P-well (4) were faint.
The hole current flows to the PNP transistor formed of the P-type impurity region (P-s). After this N channel disappears until the off operation is completed;
The time during which current flows through the NP transistor is called the tail time (ital).When considering switching loss, the main voltage recovers during this ttiH period.
The longer ttiH, the greater the power loss when off (
In the worst case, this may lead to destruction of the element. Therefore, j
The smaller the tall, the better. The N1 impurity region of the short drain region (8) is exactly the same as the emitter short circuit structure of a GTO (gate turn-off thyristor), and is a π portion provided for the purpose of reducing jtall. During tllll, electrons among the carriers remaining in the substrate (1) A flow into the N+ short drain region (8) and the electron concentration decreases, so that the electrically neutral condition in the substrate (1) is reduced. In order to maintain the short drain region (8), the hole concentration also decreases. Therefore, when the short drain region (8) is provided, Lt& is smaller (becomes smaller) than when the short drain region (8) is provided. By, −
The smaller proportion of the anode region (2) in the element also contributes to the tail reduction effect.

[Problem to be solved by the invention]

Since the conventional drain short type IGBT was constructed as described above, it had the following problems.

First, although the conventional type of short drain region functions during turn-off, it is not only useless for the on operation (in fact, it has the function of inhibiting the on characteristics. The initial process of turn-on is as described above. The turn-on is dominated by the formation of the N channel region due to gate forward bias, but turn-on is completed when the holes injected from the anode region of the drain region reach the cathode region.
This is controlled by the hole injection efficiency from the anode region side. Since the hole injection efficiency is considered to be a function of the concentration and area of the P 2 impurity region, the larger the area of the anode region per element, the better. However, by providing the short drain region, the area of the anode region is significantly reduced, and the area near the interface between the anode region and the N2 short drain region loses its function as a drain or short drain, and Some of the holes injected into the substrate flow into the N+ short drain region and disappear without reaching the cathode region.

Second, conventionally shaped short drain regions also have a negative impact on on-resistance. In other words, even when the on-state is completed and carry γ is accumulated in the substrate, the hole density decreases near the N+ short drain region, so the low resistance region is smaller than when there is no short drain region. (
Become.

Thirdly, the conventional type of short drain region cannot be said to exhibit sufficient functionality when turned off.

During ttiH, the hole concentration in the immediate vicinity of the short drain region decreases to maintain electrically neutral conditions by flowing into the short drain region of N2, but the hole concentration between the anode region (drain) and the cathode region (source) decreases. When the potential gradient is large, most of the remaining holes flow to the P-well on the cathode side. It can be said that it is extremely limited and insufficient.

In this way, conventional drain short type rGBTs have problems in both on and off characteristics, and this is due to the problems faced by single-gate Sl thyristors, GTO thyristors, etc., which have a short-circuited N+ region on the anode side. This is a problem common to all bipolar semiconductor switching devices.

In order to solve the above problems, Figures 9 and 10
As shown in the schematic cross-sectional view of the drain short type IGBT in the figure, ri! There is one in which the I-pole area, color area and drain area are independently wired (patent pending). In the figure (1) ~
(9) and (6) are equivalent to those shown in the conventional example of FIG. OG is a short-circuit region wiring, and OG is a high resistivity region. However, in what is shown in Fig. 9 and Fig. 1θ,
There was a problem in that the reverse bias between the anode region (2) and the short drain (8) could not be set sufficiently high due to device design.

In other words, as shown in FIG. 9, since the anode region is provided adjacent to the short drain region in the same plane, the voltage that can be applied during reverse bias is as follows: It is determined by the distance between this anode region and the short drain region, and the larger this distance is as shown in FIG. 10, the higher the voltage can be applied. This means that when operating with a high reverse bias, the A element area becomes large (or, conversely, when the element area is constant,
This has resulted in limitations such as reducing the area of the anode region and the short drain region, which has led to the problem of moving away from the original objectives (high-speed switching and low on-resistance characteristics).

The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide a bipolar semiconductor switching device that can improve both on-characteristics and off-characteristics.

[Means to solve the problem]

In the semiconductor device according to the present invention, the N+ short region on the anode side is separated from the anode region am, and each region is insulated and wired.

[Effect]

In the semiconductor switching device of the present invention, the N2 short drain region on the anode side is wired to separate the riIJ pole region, so that forward and reverse biases are applied between the anode region and the short drain region during on and off operations. can be applied, reducing turn-on time and turn-off time. Furthermore, since the N+ short drain region and the anode region are separated by the isolation trench, a sufficiently high forward and reverse bias can be applied, further improving the switching characteristics.

〔Example〕

Hereinafter, one embodiment of the present invention will be explained as shown in FIG.
This will be explained using a schematic cross-sectional view of a short drain type IGBT. FIG. 1 is largely the same as the conventional example shown in FIG. 8, and since (1) to @ in the figure are equivalent to those shown in FIG. 8 and FIG. 9, their explanation will be omitted. The difference is that the anode region (2) and the short drain region (8) are each provided with an anode wiring (9) and a short region wiring QQ, and the anode region (2) and the short drain region (8) are separated by a trench. (
(b)).

This separation groove (2) makes it possible to apply a high IA 111i and reverse bias between the anode region (2) and the short drain region (8) independently of the gate bias.

Next, regarding the operation, the ON operation will be explained using FIG. 2, and the OFF operation will be explained using FIG. 3.

FIG. 2 is a cross-sectional view showing the initial process of turning on the IGBT shown in FIG.
-Forward bias *S and α4011 for applying a forward bias between the short drain region (8) are electron flows schematically represented. The ON operation is started by applying a positive voltage to the gate (5), as described in the description of the prior art. Simply applying a positive voltage to the gate (6) does not cause hole injection from the anode region (2), and electrons flow from the cathode region (3) through the channel into the substrate (1), but the anode region ( Only after reaching the illusion,
Hole injection starts from the anode region (2). Therefore, time is required for the electrons leaving the cathode region (3) to travel through the substrate (1) to the anode region (2). Therefore, by applying a high forward bias between the anode region (2) and the short drain region (8), a flow of electrons is rapidly injected from the short drain region (8) toward the anode region (2). Then, without waiting for the arrival of electrons from the cathode region (3), holes are injected from the anode region (2) to the substrate (1) and from the in region (8) to the wh polar depletion region 2). They are attracted to the flow of electrons (a) and are injected. As a result, the ON operation is completed more quickly than when the forward bias between the anode region (2) and the short drain region (8) is zero or small.

FIG. 3 is a cross-sectional view showing the first and second half of the off operation of the IGBT shown in FIG. ) represents the depletion layer starting to extend from the P-well (4) side to the substrate (1) A. (to) represents the flow of holes, αη
schematically represents the flow of electrons. In the off-state, as in the explanation of the prior art, if the positive voltage of the gate (5) is removed, the N-channel region (6) disappears and electron injection from the cathode region (3) ceases. , P-well (
The depletion ah begins to expand from the P region including 4) as shown in the figure. The holes left on the board (1) are PN
P) P well (4) corresponding to the transistor collector
) etc. flow into the P region as a flow of holes αQ. At this time, anode region (2) - short drain region (
8) If a sufficiently high reverse bias power supply (12b) is applied during this period, a considerable portion of the remaining electrons in the substrate (1) will flow into the short drain region (8). for that,
The turn-off time, particularly jtall, can be shortened not by a passive method such as neutralizing holes using the short drain structure of the prior art, but by an active method of sweeping out excess electrons. This tail
The effect of shortening is the extraction of electrons from the short drain region (8), which can be expected to have the same effect as the second gate of a double-sided gate SI thyristor.

Furthermore, the short drain region (gJ) in the present invention works much more effectively for the turn-off operation than the short drain region (8) in the conventional short drain structure (1).
It is possible to reduce the ratio of the area occupied by the - element to the - element. As a result, it is possible to increase the area of the anode region (2), improving the efficiency of hole injection from the anode region (2), and in addition to the forward bias effect during on operation, further on characteristics and on The resistance can be improved.

In this way, the forward bias voltage i (12,)
By applying , the short drain region (8), which conventionally had a negative effect on on operation, actively improves the on characteristics, and also during off operation, instead of the conventional passive method, the short drain region (8) can be actively improved. The off-state characteristics can be improved primarily, and as a side effect, element design with good on-off trade-off can be achieved.

In the embodiments shown in FIGS. 1, 2, and 3, the anode region (2) and the short drain region (8) are separated by a separation groove. When applying a large reverse bias (12b) in a stable state while keeping the leakage current small, as shown in FIG. ) It is better to compensate for the separation # @ with an essential material such as polysilicon. Fig. 4 is a cross-sectional view of an IGBT according to another embodiment of the present invention, and (1) to (6) are the first It is equivalent to the one shown in the figure.The slope is the filler insulator.The isolation groove (material) is filled with the filler insulator (2); by doing so, the mechanical breakdown strength of the device is also increased. In terms of meaning, a substance 'R having similar physical properties to the substrate (1) (for example, if the substrate (1) is Si
However, if the size of the element is relatively small (about a few square meters), it is more convenient to apply an organic polymer liquid such as polyimide and bake it. It is easy to use. Because it has the same meaning, SOG
(5 pin on grass) method can also be used.

Further, in the above embodiment, the drain short type IGBT
However, the present invention is applicable to all bipolar switching elements having a short structure on the anode side.

The case of a Sl thyristor and a GTO thyrisk diode will be explained below.

FIG. 5 is a cut-away diagram of an Sl thyristor according to an embodiment of the present invention. In the figure, (1) to (6) * (7
) t(9), Q(J is the same as shown in Figure 1. (2a) is the anode (P+ impurity region), (3
a) is the cathode (N+ impurity region), (8a) is the short emitter (N+ impurity region), and is the gate rp”
Ql represents an N-epitaxial layer, (1) represents a gate electrode interconnection, and N- represents a high resistivity channel region. The basic on/off operations are gates (
G) Apply a forward/pure bias between the force sword (3a),
This is done by opening and closing the channel region υ using a depletion layer. As in the case of IGBT, as the channel opens and closes, the anode (2a) - short emitter (8a)
) to assist in on/off operation. Further, FIG. 6 shows an Sl according to another embodiment of the present invention.
FIG. 3 is a cross-sectional view showing a case where the isolation trench is filled with a thyristor and an insulator (2).

FIG. 7 is a sectional view showing a GTO thyristor according to another embodiment of the invention. In the figure, the same reference numerals correspond to the Sl thyristors shown in FIG. 5. In the case of a GTO thyristor, the main current flows between the cathode (3a) and the anode (2a) through the gate.

〔Effect of the invention〕

As described above, according to the present invention, the anode region on the anode side and the short drain region are wired independently with a separation groove, and are configured so that sufficiently high forward and reverse biases can be applied during on/off, so that the on-state characteristics - There is an effect that a semiconductor switching device with improved off-characteristics and an easy trade-off between on-off characteristics can be obtained.

[Brief explanation of the drawing]

FIG. 1 is a cross-sectional view of a drain short type IGBT semiconductor switching device which is an embodiment of the present invention, FIGS. 2 and 3 are cross-sectional views explaining the operation of the semiconductor switching device of FIG. 1, and FIG. is a cross-sectional view of a drain short type I GET semiconductor switching device which is another embodiment of the present invention, FIGS. 5 and 6 are cross-sectional views of an Sl thyristor which is another embodiment of the present invention, and FIG. A sectional view of a GTO thyristor according to another embodiment of the invention, FIGS. 8, 9, and 10 show a sectional view of a conventional drain short type IGBT semiconductor device. In the figure, (1) is the substrate, (2a) is the anode region, (2a is
) is the anode (P+ impurity region), (3) is the cathode region, (3a) is the cathode (N impurity region), (April P well, (5) is the gate, (6) is the gate insulating film, (
7) is the cathode wiring, (August short drain region, (8)
a) is the short emitter (N+ impurity region), (9) is the anode wiring, α[ is the short region wiring, αυ is the N channel region, (b) is the main current, α0 is the flow of electrons, α is the hole Flow, (g) is gate CP + impurity region), Ol
e is the N-epitaxial layer, gradient is the gate electrode wiring, (2
) is the channel region; In addition, in the figures, the same reference numerals indicate the same or equivalent parts.

Claims (1)

[Claims] A bipolar semiconductor having a first conductivity type region serving as a first main electrode on the back side of the substrate across a semiconductor substrate, and a second conductivity type region serving as a second main electrode on the front surface of the semiconductor substrate. In the switching device, a second conductivity type region serving as a control electrode is provided on the back surface via a separation groove, and the second conductivity type region serving as a control electrode is wired independently from the first conductivity type region serving as the first main electrode on the back surface, and the above separation groove is deeper than the depth at which the first conductivity type region serving as the first main electrode extends and the depth at which the second conductivity type region serving as the control electrode extends.