JP2015079987A - Semiconductor device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make voltage-withstanding of a silicon carbide power device high.SOLUTION: The problem of variation in production is solved by making acceptor concentration in floating guard rings, which are formed at a plurality of intervals, non-uniform, and making the maximum acceptor concentration in a surface-side higher than substrate-side acceptor concentration, making it difficult to influence positive charge that is present between silicon carbide and an insulating film.

Description

  The present invention relates to a semiconductor device using a floating guard ring suitable for a silicon carbide power device. In particular, the present invention relates to a termination structure of a semiconductor device.

In order to reduce the influence of the positive charge of about 1 × 10 ¹² to 2 × 10 ¹² cm ⁻² existing at the interface between silicon carbide and the insulating film on the floating guard ring termination structure, the surface charge compensation region is provided with a plurality of floating guard rings. A technique for providing the n-type silicon carbide surface between them is disclosed in Patent Document 1. Non-patent document 1 discloses a technique for simultaneously forming a silicon carbide main junction and a floating guard ring.

JP-T-2006-516815

Solid State Electronics, vol. 44 (2000), pages 303-308 (Solid State Electronics vol. 44 (2000) pp. 303-308)

  The first problem is manufacturing variation due to the density of positive charges. As a result of variations in the density of positive charges existing at the interface between the p-type silicon carbide region constituting the floating guard ring and the insulating film at each manufacturing, the manufacturing variation of the breakdown voltage in the silicon carbide power device increases.

  The second problem is the breakdown voltage due to the structure of the floating guard ring. In Patent Document 1, the floating guard ring is formed of a p-type silicon carbide region having a uniform acceptor concentration distribution. Therefore, when a reverse bias is applied to the pn junction between the floating guard ring and the n-type silicon carbide region, the depletion layer does not extend into the floating guard ring and the electric field strength is increased. As a result, the silicon carbide power device It is difficult to increase the breakdown voltage.

  A third problem is a breakdown voltage caused by forming the main junction and the floating guard ring in the same process. Non-Patent Document 1 discloses that the main junction and the floating guard ring are made of p-type silicon carbide having substantially the same acceptor concentration and substantially the same depth, and the breakdown voltage between the main junction and the floating guard ring is disclosed. Is not controlled separately, it is difficult to increase the breakdown voltage of the silicon carbide power device.

  The representative inventions according to the present application are as follows.

One aspect of the present invention is a silicon carbide substrate, an n-type silicon carbide layer formed on the silicon carbide substrate, and a first impurity concentration (N1: cm ⁻³ ) formed in the silicon carbide layer. And a p-type second semiconductor region having a second impurity concentration (N2: cm ⁻³ ) greater than the first impurity concentration between the first semiconductor region and the surface of the silicon carbide layer, An insulating film formed on the surface of the silicon carbide substrate, the first and second semiconductor regions are floating guard rings, the first and second semiconductor regions contain Al as an impurity, and silicon carbide The semiconductor device is such that the layer has an impurity concentration of 5 × 10 ¹⁵ cm ⁻³ or less and the depth (d1: μm) of the second semiconductor region is smaller than 7 × 10 ⁻¹⁹ × N2−0.14.

Another invention of the present application has a silicon carbide substrate, an n-type silicon carbide layer formed on the silicon carbide substrate, and a first impurity concentration (N1: cm ⁻³ ) formed in the silicon carbide layer. A p-type first semiconductor region, and a p-type second semiconductor region having a second impurity concentration (N2: cm ⁻³ ) greater than the first impurity concentration between the first semiconductor region and the surface of the silicon carbide substrate. And an insulating film formed on the surface of the silicon carbide layer, the first and second semiconductor regions are floating guard rings, the first and second semiconductor regions contain Al as an impurity, and are carbonized. The silicon layer has an impurity concentration greater than 5 × 10 ¹⁵ cm ^{−3 and} less than 2 × 10 ¹⁶ cm ⁻³ , and the depth (d1: μm) of the second semiconductor region is 8 × 10 ⁻¹⁹ × N 2-0. The semiconductor device is smaller than 24 (μm).

Another invention of the present application has a silicon carbide substrate, an n-type silicon carbide layer formed on the silicon carbide substrate, and a first impurity concentration (N1: cm ⁻³ ) formed in the silicon carbide layer. A p-type first semiconductor region, and a p-type second semiconductor region having a second impurity concentration (N2: cm ⁻³ ) greater than the first impurity concentration between the first semiconductor region and the surface of the silicon carbide layer. And an insulating film formed on the surface of the silicon carbide substrate, the first and second semiconductor regions are floating guard rings, the first and second semiconductor regions contain Al as an impurity, This is a semiconductor device in which the distance (d2: μm) between the surface edge of one semiconductor region and the surface edge of the second semiconductor region is smaller than −5 × 10 ⁻¹⁸ × N2 + 3.9.

  Another invention of the present application includes a silicon carbide substrate, an n-type silicon carbide layer formed on the silicon carbide substrate, a p-type first semiconductor region formed in the silicon carbide layer, and a first semiconductor region A plurality of p-type second semiconductor regions surrounding the first semiconductor region, the first and second semiconductor regions contain Al as an impurity, and the depth of the first semiconductor region is shallower than the depth of the second semiconductor region, The second semiconductor region at the innermost periphery of the second semiconductor region is a semiconductor device that contacts or partially overlaps the first semiconductor region.

  According to the configuration of the present invention, the silicon carbide power device is hardly affected by the positive charge existing between the silicon carbide and the insulating film. Further, by making the acceptor density in the floating guard ring non-uniform, electric field concentration can be suppressed and the breakdown voltage of the silicon carbide power device can be increased. Furthermore, the withstand voltage can be increased by optimizing the main junction and the floating guard ring independently.

1 is a longitudinal sectional view of a pn diode according to a first embodiment of the present invention. It is an example of the Al concentration profile of the depth direction in a floating guard ring. It is a simulation result which shows the effect of the 1st Example of this invention. It is a simulation result which shows the effect of the 1st Example of this invention. It is a simulation result which shows the effect of the 1st Example of this invention. This is a result of measuring the breakdown voltage of the pn diode by changing the Al concentration of the main junction without providing the termination structure. It is a longitudinal cross-sectional structure figure of the pn diode by a prior art. It is a simulation result which shows the withstand voltage in the pn diode shown in FIG. 1 and the insulating film / floating guard ring interface charge density dependence. It is a simulation result which shows N2 and Nd dependence of allowable maximum d1. It is a figure explaining allowable maximum d1 and allowable maximum d2 in the 1st example of the present invention, and the 2nd example. It is a simulation result which shows the dependence of N2 of the permissible maximum depth d1 in case Nd in the 1st Example of this invention is 5 * 10 < ¹⁵ > cm <^-3> or less. It is a simulation result which shows the dependence of N2 of the permissible maximum depth d1 in case Nd in the 1st Example of this invention is larger than 5 * 10 < ¹⁵ > cm < ^-3 >. It is a longitudinal cross-sectional structure figure which shows the manufacturing process of the pn diode which is the 1st Example of this invention. It is a longitudinal cross-sectional structure figure which shows the manufacturing process of the pn diode which is the 1st Example of this invention. It is a longitudinal cross-sectional structure figure which shows the manufacturing process of the pn diode which is the 1st Example of this invention. It is a longitudinal cross-sectional structure figure of the pn diode which is the 2nd Example of this invention. It is a simulation result which shows the dependence of N2 of the permissible maximum width d2 in the 2nd Example of this invention. (A) is a longitudinal cross-sectional structure diagram in the case where the p-type silicon carbide region 4 forming the main junction is deeper than the floating guard ring 5, and (b) of the p-type silicon carbide region 4 forming the main junction. FIG. 6 is a vertical cross-sectional structure diagram when the direction is shallower than the floating guard ring 5. It is a longitudinal cross-section structure figure of the JBS diode which is the 3rd Example of this invention.

  Hereinafter, an embodiment of a silicon carbide pn diode which is a first example of the present invention will be described.

  FIG. 1 is a cross-sectional view of a semiconductor device according to the present invention. As shown in FIG. 1, the semiconductor device of Example 1 according to the present invention is formed in a drift layer 3 made of n-type silicon carbide formed on an n-type silicon carbide substrate 2 and in the drift layer. Floating guard ring 9, p-type silicon carbide region 4 forming a main junction formed inside innermost floating guard ring 9, anode electrode 7 formed on p-type silicon carbide region, n-type It consists of a cathode electrode 1 formed on the back surface of the silicon carbide substrate 2 and an insulating film 8 formed on the floating guard ring 9. The insulating film 8 is provided with an opening for drawing the anode electrode 7 to the outside, and the anode electrode 7 is electrically connected to the p-type silicon carbide region 4 through the opening. Note that the planar layout itself is well known, and thus a top view is omitted, but p-type silicon carbide region 4 is surrounded by a plurality of floating guard rings 9. The p-type silicon carbide region 4 and the drift layer 3 made of n-type silicon carbide constitute a pn diode.

  Here, the characteristics of the semiconductor device of Example 1 will be described. One of the features is that a floating guard ring 9 is formed between a p-type first semiconductor region 5 having a predetermined impurity concentration, and between the first semiconductor region 5 and the surface of the n-type silicon carbide substrate 2. The second semiconductor region 6 is a p-type second semiconductor region 6 having an impurity concentration higher than that of one semiconductor region 5.

Next, details of the impurity concentrations of the first semiconductor region 5 and the second semiconductor region 6 will be described with reference to FIG. FIG. 2 is an impurity concentration profile in which the horizontal axis represents depth and the vertical axis represents the concentration of Al, which is an impurity of the p-type semiconductor region. This profile is shown along the vertical depth of the region including the floating guard ring 9 of FIG. As described above, the concentration peak in the first semiconductor region is about 9.5 × 10 ¹⁷ cm ⁻³ and the concentration peak in the second semiconductor region is about 3.5 × 10 ¹⁷ cm ^−3. It can be seen that the impurity concentration of the first semiconductor region is higher than the concentration of the second semiconductor region. Here, the concentration peak in the first semiconductor region is 1.0 × 10 ¹⁷ cm ⁻³ or more and 5.0 × 10 ¹⁷ cm ⁻³ or less, and the concentration peak in the second semiconductor region is 5.0 × 10 × It is desirable that it is greater than 10 ¹⁷ cm ⁻³ and 1.0 × 10 ¹⁸ cm ⁻³ or less.

Next, the result of studying the relationship between the impurity concentration and the breakdown voltage will be described with reference to FIG. In FIG. 3, the upper part shows a distribution diagram of the Al impurity concentration in the depth direction, and the lower part shows a two-dimensional simulation result of the potential distribution at the time of avalanche breakdown in the impurity concentration distribution corresponding to the upper part. In this simulation, the floating guard ring is actually substituted for the main junction, but it is the same in terms of the p-type semiconductor region, and the result itself can be understood by replacing it with that of the floating guard ring. The left diagram in FIG. 3 shows the Al concentration of the second semiconductor region at 10 ¹⁸ cm ⁻³ , the middle diagram at 10 ¹⁷ cm ⁻³ , and the right diagram at 10 ¹⁶ cm ⁻³ . Each of them is implanted with seven kinds of acceleration energy of 25 to 380 keV.

Next, the breakdown voltage simulation results will be described. The lower left figure shows a 600V equipotential line on the substrate surface. Considering the relationship between the equipotential lines of 400V and 200V, it can be seen that the substrate surface has a potential of about 700V at the time of avalanche breakdown. . That is, the breakdown voltage is about 700V. On the other hand, the breakdown voltage in the middle figure is about 1000V, and the breakdown voltage in the right figure is slightly over 400V. For these reasons, if an attempt is made to manufacture a pn junction diode having a breakdown voltage of 600 V, the second semiconductor region of the floating guard ring preferably has an impurity concentration of 10 18 or 17 cm ⁻³ , It can be seen that 17 17 cm ⁻³ is preferable.

Next, the result of examining the reason why such a result was obtained will be described with reference to FIGS. 4 and 5 show the results of two-dimensional simulation of the hole concentration distribution at the time of avalanche breakdown in the case of having the box-like Al concentration distribution in the upper left portion of FIG. 3 and the upper middle portion of FIG. In each figure, since a reverse voltage is applied, it can be seen that the depletion layer spreads to the higher-order hole concentration side and the electric field gradient is increased. And as shown in the circle shown with the broken line of FIG. 4, hole density distribution with a comparatively small curvature radius is formed. For this reason, it is considered that a high electric field is concentrated in this region and a relatively low breakdown voltage is obtained. On the other hand, in FIG. 5, a hole concentration distribution having a relatively large radius of curvature is formed as shown in a circle indicated by a broken line. For this reason, compared with FIG. 4, it is considered that the electric field concentration is relaxed and a relatively high breakdown voltage is obtained. There has been no report in the past on such a breakdown voltage improvement measure considering depletion of the p-type silicon carbide region. In the case of the Al concentration distribution on the upper right side of FIG. 3, although not shown, the hole concentration at the time of avalanche breakdown is 10 ¹⁵ cm ⁻³ or less on the substrate surface, so that the breakdown voltage is slightly over 400V. Conceivable. In this simulation, the donor density Nd of the drift layer was set to 2 × 10 ¹⁵ cm ⁻³ , which is a condition suitable for a power device having a withstand voltage of several kV. Further, in this simulation, the middle diagram is performed using three semiconductor regions having different impurity concentrations, and the left diagram is performed using four semiconductor regions having different impurity concentrations. As described above, as shown in FIG. Similar effects can be obtained by at least two types of semiconductor regions (first and second semiconductor regions).

Next, as a comparative example, a pn diode having a withstand voltage of several kV in which the floating guard ring 5 is used as one type of semiconductor region instead of the two types of semiconductor regions as in the present invention will be described. First, in determining the impurity concentration of Al in the floating guard ring 5, the Al concentration of the main junction is changed in the range of 3.0 × 10 ¹⁷ to 7.6 × 10 ¹⁷ cm ⁻³ , and the donor concentration in the drift layer is determined. The relationship with the reverse breakdown voltage of the pn diode was measured. FIG. 6 shows the result. For the measurement, a structure in which a terminal structure such as a floating guard ring was not formed was used. As shown in FIG. 6, it was found that the withstand voltage significantly increased in the impurity concentration of Al in the range of 3.8 × 10 ¹⁷ to 5.7 × 10 ¹⁷ cm ⁻³ .

Accordingly, the floating guard ring 5 of the silicon carbide pn diode shown in FIG. 7 is made of p-type silicon carbide (one type of semiconductor region) having an Al concentration of 3.8 × 10 ¹⁷ cm ⁻³ , and 19 to 23 pieces are arranged. When applied to a drift layer having a donor concentration of 2 × 10 ¹⁵ cm ⁻³ and a film thickness of 30 μm, the breakdown voltage hardly depended on the number of floating guard rings 5 and remained at about 3.3 kV.

In the pn diode as a comparative example (FIG. 7), it is another experiment that a positive charge of 3 × 10 ¹² cm ⁻² exists between the SiO ₂ film used as the insulating film 8 and the p-type silicon carbide. It became clear from. When positive charges are present, negatively charged acceptor ions in the floating guard ring 5 are compensated near the interface with the SiO ₂ film 8, so that the p-type concentration is reduced or the surface of the floating guard ring 5 is n-type. It was considered that the breakdown voltage remained at 3.3 kV because the function of the floating guard ring 5, which extended the depletion layer in n-type silicon carbide and relaxed the electric field, was impaired.

Therefore, as shown in FIGS. 1 and 2, the floating guard ring 9 is composed of a semiconductor region 6 having a relatively high impurity concentration and a semiconductor region 5 having a relatively low impurity concentration. In the manufacturing process, the floating guard ring 6 was formed by increasing the injection amount at the minimum injection energy of 25 keV from 5 × 10 ¹¹ cm ⁻² to 5 × 10 ¹² cm ⁻² . As a result, the pn diode breakdown voltage improved to 3.8 kV.

FIG. 8 is a diagram showing the relationship between the interface charge density and the breakdown voltage. In the figure, it is a two-dimensional simulation result when the number of floating guard rings is 21. The broken line is a structure in which the implantation amount at a minimum implantation energy of 25 eV is 5 × 10 ¹¹ cm ⁻² as a comparative example, and the solid line is a structure in which the implantation amount according to the present invention is 5 × 10 ¹² cm ⁻² . As can be seen from the figure, the structure according to the present invention has little dependence on the withstand voltage with respect to the interface charge density. That is, it was confirmed that the breakdown voltage did not decrease even if the interface charge density increased to 1 × 10 ¹³ cm ⁻² due to manufacturing variations.

Next, the maximum depth d1 at which the floating guard ring 6 shown in FIG. 1 can be formed will be described with reference to FIGS. Since this d1 depends on the maximum acceptor concentration N2 of the floating guard ring 6 and the donor density Nd of the drift layer 3, how to obtain d1 will be described using N2 and Nd. First, FIG. 10 will be described. FIG. 10 shows the hole concentration distribution during the avalanche breakdown shown in FIG. The depth at the hole concentration end of 1 × 10 ¹⁷ cm ⁻³ where the depletion layer hardly penetrates even when a reverse voltage is applied is defined as d1. FIG. 9 is a two-dimensional diagram in which d1 is obtained by further reducing the hole concentration of 1 × 10 ¹⁷ cm ⁻³ to 4 × 10 ¹⁷ cm ⁻³ , 5 × 10 ¹⁷ cm ⁻³ , and 6 × 10 ¹⁷ cm ^−3. It is a simulation figure. The horizontal axis Nd is the donor density of the drift layer 3, and the vertical axis d1 is the allowable maximum depth. As is clear from FIG. 9, it was found that d1 hardly depends on Nd when Nd is 5 × 10 ¹⁵ cm ⁻³ or less. On the other hand, when Nd is larger than 5 × 10 ¹⁵ cm ⁻³ , it is dependent on Nd, and it has been found that the maximum allowable depth d1 is significantly reduced as Nd increases.

FIG. 11 is a graph in the case where Nd is 5 × 10 ¹⁵ cm ⁻³ or less and the maximum acceptor concentration N2 is the horizontal axis and the allowable maximum depth d1. Circles, triangles, and squares correspond to the densities in FIG. The straight line drawn in the figure is an approximate straight line obtained by approximation by three points, and d1 = 7 × 10 ⁻¹⁹ × N2−0.14 (μm). That is, if d1 satisfying d1 <7 × 10 ⁻¹⁹ × N2−0.14 (μm), a region having a small radius of curvature as shown in FIG. 4 is unlikely to occur, and an element with high withstand voltage can be obtained. The circle is 0.15 μm, the triangle is 0.21 μm, and the square is 0.28 μm.

On the other hand, FIG. 12 shows a similar graph when Nd is greater than 5 × 10 ¹⁵ cm ⁻³ . Here, since there are almost no application examples when Nd> 2 × 10 ¹⁶ cm ^{−3 is} greater than Nd = 1 × 10 ¹⁶ cm ⁻³ , an approximate straight line as shown in FIG. I can draw. The straight line drawn in the figure is an approximate straight line obtained by approximation by three points, and d1 = 8 × 10 ⁻¹⁹ × N2−0.24 (μm). That is, when Nd is greater than 5 × 10 ¹⁵ cm ^{−3 and} less than 2 × 10 ¹⁶ cm ⁻³ , if d1 satisfies d1 <8 × 10 ⁻¹⁹ × N2−0.24 (μm), A region having a small curvature radius such as 4 is unlikely to occur, and an element with a high breakdown voltage can be obtained. The circle is 0.09 μm, the triangle is 0.16 μm, and the square is 0.24 μm.

Next, the manufacturing method of an element is demonstrated using FIGS. As shown in FIG. 13, after growing an n-type silicon carbide drift layer (film thickness 30 μm, nitrogen concentration 2 × 10 ¹⁵ cm ⁻³ ) 3 on the n-type silicon carbide substrate 2 by vapor phase epitaxy, SiO (not shown) _Two films were deposited, and an ion implantation mask was formed by photolithography and dry etching. Then, Al ions having a concentration distribution indicated by a solid line in FIG. 2 were implanted in the depth direction using seven energies between 25 keV and 380 keV. At this time, at an energy of 25 keV, the Al ion implantation amount was increased compared to other energies, and the Al concentration on the substrate surface side was increased so that the depth d1 was 0.1 μm or less. This Al concentration distribution corresponds to the floating guard rings 5 and 6 in FIG. The number of floating guard rings 5 and 6 can be changed according to the desired breakdown voltage and the nitrogen concentration in n-type silicon carbide drift layer 3. Thereafter, the mask for ion implantation was removed with hydrofluoric acid (FIG. 13). Note that in the case of Nd = 2 × 10 ¹⁵ cm ⁻³ and N2 = 4 × 10 ¹⁷ cm ⁻³ , d1 is essential to be 0.15 μm or less from FIG. Is satisfied.

Subsequently, an SiO ₂ film (not shown) was again deposited, and an ion implantation mask was formed for the p-type silicon carbide region 4 where the main junction was to be formed by photolithography and dry etching. Then, Al ions having a concentration of 10 ¹⁹ cm ⁻³ were implanted using four energies between 25 keV and 130 keV. Thereafter, the ion implantation mask was removed with hydrofluoric acid, and activation annealing of the ion implanted Al was performed at 1700 ° C. (FIG. 14).

Thereafter, a SiO ₂ film 8 (film thickness 0.2 μm) was deposited, and an opening was provided on p-type silicon carbide region 4 where a main junction was formed by photolithography and dry etching (FIG. 15). Finally, a pn diode was manufactured by forming the anode electrode 7 and the cathode electrode 1 (FIG. 1), and the reverse breakdown voltage was measured. As a result, 3.8 kV was obtained.

According to this embodiment, as a result of increasing the concentration only on the outermost surface of the floating guard ring using the floating guard ring having the depth distribution of the acceptor concentration, which is optimal for improving the breakdown voltage of the pn diode, the floating guard ring is obtained. In addition, a silicon carbide pn diode that can ignore the influence of positive charges of about 3 × 10 ¹² cm ⁻² existing at the interface between the insulating film and the breakdown voltage can be manufactured.

In the case of a pn diode having a high Nd of about 1 × 10 ¹⁶ cm ⁻³ and a withstand voltage of about 1 kV, the implantation energy on the outermost surface is lowered to, for example, 15 keV, and d1 <8 × 10 ⁻¹⁹ × N2−0.24 (μm). By satisfying the above, a similar silicon carbide pn diode can be manufactured.

  Further, not only a diode but also a transistor using silicon carbide, such as a field effect transistor or a bipolar transistor, can similarly realize a termination structure.

  Next, an embodiment of a silicon carbide pn diode which is a second example will be described. The contents described in the first embodiment can also be applied to the present embodiment unless there are special circumstances. In Example 1, a preferable depth of the p-type second semiconductor region 6 having an impurity concentration higher than that of the first semiconductor region 5 will be described, and a preferable horizontal distance between the first semiconductor region 5 and the second semiconductor region 6 will be described. In Example 2, a constant distance was provided in the horizontal direction between the first semiconductor region 5 and the second semiconductor region 6.

  FIG. 16 shows an element in which the second semiconductor region 6 is wider than the first semiconductor region 5 and a horizontal distance d2 is provided. However, the present embodiment is not limited to the depth d1 as described in the first embodiment. Other configurations are the same as those in FIG. By doing so, as shown in FIG. 5, the depletion layer at the time of avalanche breakdown also extends in the horizontal direction in the floating guard ring 5, so that the radius of curvature of the hole concentration curve at the time of avalanche breakdown is obtained. Can be increased. Therefore, the electric field there can be relaxed, and the breakdown voltage of the silicon carbide power device can be increased. Moreover, the silicon carbide pn diode which can ignore the influence which a positive charge has on a proof pressure can be provided.

Next, necessary conditions for d2 will be described. FIG. 10 illustrates d2 using the hole concentration distribution during avalanche breakdown. Here, d2 is a distance in the horizontal direction from the hole concentration end of 1 × 10 ¹⁷ cm ⁻³ where the depletion layer hardly penetrates and the side end of the floating guard ring 5, that is, the end of the Al ion implantation region. D2 is substantially equal to the distance between the surface end of the first semiconductor region 5 and the surface end of the second semiconductor region 6. As a result of the simulation, it was found that d2 hardly depends on the donor concentration Nd of the drift layer, but remarkably depends on the maximum acceptor concentration N2 of the floating guard ring (FIG. 17).

Next, FIG. 17 will be described. FIG. 17 is a graph in which the horizontal axis is the maximum acceptor concentration N2 of the floating guard ring, and the vertical axis is the allowable maximum width of d2 at which a predetermined breakdown voltage is obtained. From the fitting straight line of FIG. 17 obtained by the least square method, this straight line was determined to be d2 = −5 × 10 ⁻¹⁸ × N2 + 3.9 (μm). That is, under the conditions satisfying the relationship d2 <−5 × 10 ⁻¹⁸ × N2 + 3.9 (μm), the curvature radius of the hole concentration curve at the time of avalanche breakdown can be increased and the electric field can be relaxed. From the left, d2 is 1.9 μm, 1.4 μm, and 0.9 μm.

  With respect to this manufacturing method, when forming the floating guard ring 6 of FIG. 13, a mask having a smaller opening that satisfies the condition d2 than that for forming the floating guard ring 5 is separately formed, and Al is implanted. That's fine. By doing so, the element of FIG. 16 can be obtained.

  In the second embodiment, the depth d1 is not limited. However, in addition to the second embodiment, a good effect can be obtained by adding the condition of the depth of the first embodiment to the depth d1.

  Further, not only a diode but also a transistor using silicon carbide, such as a field effect transistor or a bipolar transistor, can similarly realize a termination structure.

  Next, an embodiment of a silicon carbide pn diode which is a third example will be described. As for Examples 1 and 2 so far, the embodiment has been described in which the withstand voltage is ensured by changing the concentration distribution of the floating guard rings 5 and 6. On the other hand, the third embodiment is an example in which the withstand voltage is secured by controlling the depth of the impurity concentration region between the floating guard ring 5 and the main junction 4.

  First, the structure examined using FIG. 18 will be described. 18A shows the case where the p-type silicon carbide region 4 forming the main junction is deeper than the floating guard ring 5, and FIG. 18B shows the direction of the p-type silicon carbide region 4 forming the main junction. Is shallower than the floating guard ring 5. In both cases, the innermost floating guard ring closest to the p-type silicon carbide region 4 and the p-type silicon carbide region 4 are in contact with or partially overlapping. 18 (a) and 18 (b), the p-type silicon carbide region 4 and the floating guard ring 5 that form the main junction are formed using different masks, so that the impurity concentration of the p-type silicon carbide region is made independent. Can be controlled. However, since a different mask is used on the other hand, a problem of misalignment between p-type silicon carbide region 4 and floating guard ring 5 may occur. This is particularly the case with a power device termination structure with a withstand voltage of several kV in which the withstand voltage changes by several hundred volts when the distance between the p-type silicon carbide region 4 forming the main junction and the innermost peripheral floating ring is shifted by 0.1 μm. It will be a big challenge. Therefore, in order to avoid the influence of misalignment in such a high breakdown voltage power device, it is effective that the p-type silicon carbide region 4 and the innermost peripheral floating ring are intentionally contacted or partially overlapped. . In this case, two embodiments shown in FIGS. 18A and 18B are conceivable. Normally, the p-type silicon carbide region 4 has a higher impurity concentration than the floating guard ring 5 because the anode electrode needs to be formed with a low contact resistance. For this reason, in the structure of FIG. 18A, an end where a high electric field tends to concentrate is formed at a location indicated by symbol A. On the other hand, in the structure of FIG. 18B, the portion indicated by symbol A is in contact with or partially overlaps the floating guard ring 5 and is an integral p-type silicon carbide region, so that it is p-type silicon carbide. The substantial end of region 4 is symbol B. Therefore, the avalanche breakdown voltage is determined at the end of the floating guard ring 5 having the maximum acceptor concentration N2. In this way, in the structure of FIG. 18B, the portion of the symbol A is not exposed to the drift layer 3, so that the breakdown voltage can be increased, and further, the breakdown voltage fluctuation caused by misalignment of the mask caused by the manufacturing. It is possible to manufacture a silicon carbide power device with a high breakdown voltage that suppresses the above.

  Next, an embodiment of a silicon carbide JBS (Junction Barrier Schottky) diode will be described with reference to FIG. The contents described in Examples 1 and 2 can also be applied to this example unless there are special circumstances.

  The difference between FIG. 1 and FIG. 19 is that the JBS diode has an arrangement in which Schottky junctions and pn junctions are alternately repeated. Therefore, the configuration of the p-type silicon carbide region 4, the impurity concentration distribution of the floating guard ring 5, The p-type silicon carbide region 4 to be joined and the floating guard ring are in contact with or partly overlapped, and the p-type silicon carbide region 4 to be the main junction is shallower than the floating guard ring 5. The impurity concentration of the main junction 4 is higher than the impurity concentration of the floating guard ring 5.

  According to this embodiment, the JBS main junction and the innermost peripheral floating guard ring are brought into contact with each other, and the main junction is shallower than the innermost peripheral floating guard ring. As a result, the JBS main junction and the floating guard ring are independently optimized. However, there is an effect that it is possible to realize a high breakdown voltage silicon carbide JBS diode that does not have a breakdown voltage variation for each manufacture.

  As for the manufacturing method, the floating guard ring and the p-type silicon carbide region to be the main junction of the JBS diode are formed by ion implantation of Al using separate masks, and the p-type silicon carbide region of the floating guard ring is the main junction. This can be realized by contacting or overlapping with the p-type silicon carbide region and deepening the ion implantation on the floating guard ring side. Either the main junction or the floating guard ring may be formed first.

  In addition, you may combine this Embodiment and the form of Example 1, or the form of Example 2. FIG.

  Further, not only the JBS diode but also a transistor using silicon carbide, such as a field effect transistor, a bipolar transistor, or a pn diode, can similarly realize a termination structure.

1: cathode electrode, 2: n-type silicon carbide substrate, 3: n-type silicon carbide drift layer, 4: p-type silicon carbide region, 5, 6, 9: floating guard ring, 7: anode electrode, 8: insulating film

Claims (1)

A silicon carbide substrate;
An n-type silicon carbide layer formed on the silicon carbide substrate;
A termination region comprising a p-type first semiconductor region having a first impurity concentration;
A p-type second semiconductor region formed in the silicon carbide layer surrounded by the termination region and having a second impurity concentration higher than the first impurity concentration;
An insulating film formed on the surface of the first semiconductor region,
The first semiconductor region and the second semiconductor region are in contact with each other;
The semiconductor device, wherein the second semiconductor region is shallower than the first semiconductor region.

Images