Insulated gate field effect transistor and manufacturing method of the same

[0002] This application is based upon and claims the benefit of priority of Japanese Patent Application No. 9-32265filed on Feb. 17, 1997, the contents of which are incorporated herein by reference.

[0003] 1. Field of the Invention:

[0004] The present invention relates to an insulated gate field effect transistor having a DMOS structure and utilized as a power switching device which can withstand high voltage and allow large current to flow, and to a manufacturing method of the same. Such an insulated gate field effect transistor is suitably used for a vehicule ignition apparatus and a motor driving inverter.

[0005] 2. Discussion of Related Art:

[0006] It is known that an insulated gate bipolar transistor (“IGBT”)enables high withstand voltage compatible with low ON resistance. Although the IGBT has a similar structure to a power MOSFET, the IGBT has a layer of an opposite electrical conductivity type from a source layer as a drain region.

[0007] FIG. 19A shows a top schematic view of the main part of an unit cell of the IGBT and FIG. 19B shows a sectional view of the IGBT taken along a line 19B—19B in FIG. 19A. The structure of the IGBT will be explained below based on FIGS. 19A and 19B. What an N⁻ type epitaxial layer 2 is grown on the surface of a P⁺ type substrate 1 is used as a substrate of the IGBT. Then, a deep P⁺ type diffusion layer (a “deep P well”) 3 of which concentration is relatively high and a P type diffusion layer (a “channel P well”) 4 of which concentration is lower than that of the deep P well 3 are formed in the surface of the N⁻ type epitaxial layer 2.

[0008] An emitter region 5 composed of an N⁺ type diffusion layer (corresponding to a source region of a power MOSFET) is formed in the surfaces of the deep P well 3 and the channel P well 4. Further, a gate electrode 7 is provided on the channel P well 4 with a gate oxide film 6ainterposed. The gate electrode 7 is formed to have a pattern indicated by the hatched portion in FIG. 19A. The gate electrode 7 is covered by an interlayer insulating film 6b. A contact hole (a part surrounded by a dotted line in FIG. 19A) is formed through the gate oxide film 6aand the interlayer insulating film 6b. An emitter electrode 8 is disposed to contact the emitter region via this contact hole. A collector electrode 9 is disposed on the back of the P⁺ type substrate.

[0009] FIG. 20 is a partially sectional view of the IGBT to describe the operation thereof when the IGBT thus constructed is actually operated. The operation of the IGBT will be explained based on FIG. 20.

[0010] When a certain threshold voltage is applied to the gate electrode 7, the surface of the channel P well 4 below the gate electrode 7 is inverted, thus creating an electron channel. Then, electrons flow into the N⁻ type epitaxial layer 2 through this channel. The electrons which have flown into the N⁻ type epitaxial layer 2 drop the potential of the N⁻ type epitaxial layer 2 so as to forwardly bias the PN junction between the P⁺ type substrate 1 and the N⁻ type epitaxial layer 2. As a result, holes which are a minority carrier flow from the P⁺ type substrate 1 to the N⁻ type epitaxial layer 2. Due to the flow of the holes, the N⁻ type epitaxial layer 2 is subjected to modulation of the electrical conductivity, thereby considerably reducing its resistivity. As a result, a large hole current flows from the collector electrode 9 to the emitter electrode 8.

[0011] Because the horizontal resistance of the channel P well 4 (the part indicated by a mark of resistance in FIG. 20) is large, a large voltage drop occurs in the channel P well 4. When the internal resistance of the channel P well 4 is taken into account, the electrical connection in the IGBT may be expressed by a circuit diagram as shown in FIG. 21.

[0012] That is, an FET shown in FIG. 21 is composed of the gate electrode 7, the channel P well 4, the emitter region 5 and the N⁻ type epitaxial layer 2, a PNP transistor is composed of the P type substrate 1, the N⁻ type epitaxial layer 2, and the channel P well 4 and the deep P well 3, and an NPN transistor (a “parasitic transistor”) is composed of the emitter region 5, the channel P well 4 and the N⁻ type epitaxial layer 2. A resistance R denotes the internal resistance of the channel P well 4.

[0013] When the quantity of the hole current flowing through the resistance R increases as shown in FIG. 21, the large voltage drop occurs at the resistance R. Then, when this voltage drop reaches a value which can forwardly bias the PN junction of the emitter region 5 and the channel P well 4, the parasitic transistor turns on. Thereby, a so-called latch-up phenomenon occurs, by which the IGBT is kept continuously conductive state.

[0014] FIG. 22 shows the state of the flow of the holes in the IGBT on the top surface of a wafer. Most of the holes injected from the P⁺ type substrate 1 to the N⁻ type epitaxial layer 2 flow upward in the N⁻ type epitaxial layer 2 and reach the emitter electrode 8 passing through the channel P well 4 which is located at the upper part thereof. However, the holes injected to the N⁻ type epitaxial layer 2 at the periphery of a cell region in which the current flows via the channel flow heading toward the nearby cell region and reach the emitter electrode 8 because no cell region is formed thereabove. Because the flows of holes from the periphery are thus concentrated at the edge portion of the cell region as a result, its density becomes higher than that of the hole current flowing through the inner part of the cell region. That is, the flow rate of the holes increases in the vicinity of the edge portion of the cell region. Therefore, there has been a problem that the voltage drop due to the horizontal resistance of the channel P well 4 in the vicinity of the cell edge portion becomes large and the above-mentioned latch-up phenomenon readily occurs.

[0015] The present invention is intended to solve the aforementioned problem and its object is to improve the ruggedness against the latch-up phenomenon by relieving the concentration of the holes at the cell region in the vicinity of the cell edge portion.

[0016] In order to achieve the aforementioned object, according to the present invention, in addition to a first region which is formed in the cell region and transmits hole current to a emitter electrode as the deep well, a second region having a high concentration is formed at the outside of the cell region so that its width is made wider than the first region to overlap with all or a part of the channel well. As a result, the flows of holes at the cell edge portion are drawn to the second region by forming the high concentration, i.e. low resistant, second region at the outside of the cell region. Therefore, the holes do not go around from the cell edge portion into the first region of the cell region but flow via the second region. Thereby, the flow rate of the holes will not increase in the cell region in the vicinity of the cell edge portion due to the holes flowing from the cell edge portion. Accordingly, it becomes possible to prevent the increase of the voltage drop in the cell region in the vicinity of the cell edge portion and to improve the ruggedness against the latch-up phenomenon.

[0017] The second region may be formed to be linked with another second region formed for the adjacent cell region. The holes coming from the outside of the cell region may be drawn almost completely into the second region by thus connecting the adjacent second regions to each other. Thereby, the ruggedness against the latch-up phenomenon may be improved further.

[0018] It is noted that the concentration of the second region is not limited to only a concentration equal to that of the first region. That is, the second region may have the same concentration as the first region or the second region may have a concentration higher than that of the first region.

[0019] The second region can be formed together with the first region by. using a mask having an opening at the region where the first region and the second region are to be formed. As a result, a step required for forming only the second region may be eliminated and a number of steps for manufacturing the semiconductor device according to the present invention may be reduced by forming the first region together with the second region in the same time by one mask as described above.

[0020] These and other objects, features and characteristics of the present invention will be appreciated from a study of the following detailed description, the appended claims, and drawings, all of which form a part of this application. In the drawings:

[0021] FIG. 1 is a schematic top view of an IGBT unit cell according to a first embodiment;

[0022] FIG. 2A is a sectional view taken along a line 2A—2A in FIG. 1, FIG. 2B is a sectional view taken along a line 2B—2B in FIG. 1, and FIG. 2C is a sectional view taken along a line 2C—2C in FIG. 1, and FIG 2D is a sectional view taken along a line 2D—2D in FIG. 1;

[0023] FIG. 3 is a schematic view showing flows of holes in the IGBT;

[0024] FIGS. 4A through 10B are views showing a manufacturing process of the IGBT in FIG. 1;

[0025] FIG. 11 is a schematic top view of an IGBT according to a second embodiment;

[0026] FIGS. 12A through 18B are views showing a manufacturing process of the IGBT in FIG. 11;

[0027] FIG. 19A is a schematic top view of a conventional IGBT and FIG. 19B is a sectional view taken along a line 19B—19B in FIG. 19A;

[0028] FIG. 20 is a sectional view to describe an operation of the IGBT shown in FIGS. 19A and 19B;

[0029] FIG. 21 is a circuit diagram showing an electrical connection in the IGBT shown in FIGS. 19A and 19B; and

[0030] FIG. 22 is a schematic top view showing flows of holes in the IGBT shown in FIGS. 19A and 19B.

[0031] The present invention will be explained below based on the preferred embodiments shown in the drawings.

[0032] An IGBT is composed of a plurality of unit cells provided in parallel. FIG. 1 is a schematic top view of the unit cell. FIGS. 2A through 2C are sectional views taken along lines 2A—2A, 2B—2B and 2C—2C in FIG. 1, respectively. The structure of the IGBT will be explained below based on FIGS. 1 and 2A-2C.

[0033] What an N⁻ type epitaxial layer 2 has grown on the As surface of a P type substrate 1 is used as a substrate of the IGBT. A rectangular patterned deep P well 3 whose concentration is relatively high is formed on the surface of the N⁻ type epitaxial layer 2. Further, a high concentration P⁺ type region 10 for lowering the resistance of the deep P well 3 is formed on the surface of the deep P well 3. Further, a channel P well 4 whose concentration is lower and whose configuration is wider and shallower than those of the deep P well 3 is disposed so as to interpose the deep P well 3 at the center thereof.

[0034] P⁺ type regions 20, having a concentration higher, a depth deeper and a width wider than those of the channel P well 4, are formed in the vicinity of cell edge portions while being connected to the edge of the deep P well 3 in its longitudinal direction (cell edge portion) to improve the ruggedness against the latch-up phenomenon. Specifically, the P⁺ type regions 20 are formed approximately into a shape which diverges from the cell edge portion with a predetermined angle. It is noted that although the P⁺ type region 20 is shown separately from the deep P well 3 in FIG. 1 in order to clarify their roles, actually they are formed by the P⁺ type layer having the same concentration in the present embodiment.

[0035] Further, an emitter region 5 composed of an N⁺ type diffusion layer is formed on the deep P well 3 and the channel P well 4. A gate electrode 7 is disposed above the channel P well 4 and provided commonly for the respective unit cells, interposing a gate oxide film 6atherebetween. The gate electrode 7 is covered further by an interlayer insulating film 6b. The gate electrode 7 is patterned into a shape indicated by the hatched portion in FIG. 1. A contact hole is formed in the interlayer insulating film 6b. An emitter electrode 8 is formed to contact the deep P well 3 and the emitter regions 5 through this contact hole as a contact window. A collector electrode 9 is formed on the back of the P type substrate 1.

[0036] When a certain voltage is applied to the gate electrode 7 in the IGBT thus constructed, holes (current) flow from the collector electrode 9 to the emitter electrode 8 as described before. At this time, holes injected outside a cell region are drawn to the cell edge portion due to the flow of the holes in the cell region. As a result, a flow of holes is generated at the cell edge portion.

[0037] However, because the P⁺ type region 20 having a low resistance and a large width is formed in the vicinity of the cell edge portion, the flow of holes at the cell edge portion are mostly drawn into the P⁺ type region 20. FIG. 3 shows the flows of holes at the cell edge portion. The holes reach the emitter electrode 8 while flowing through the P⁺ type region 20 as shown in FIG. 3 without going around from the cell edge portion to the channel P well 4 having a higher resistance in the cell region because the resistance of the P⁺ type region 20 (slanting line part in FIG. 3) at the cell edge portion is low. Especially, because the P⁺ type region 20 is electrically connected to the deep P well 3 having a low resistance, the holes injected in the vicinity of the cell edge portion can reach the emitter electrode 8 only through the P⁺ type region 20 and the deep P well 3. Therefore, the voltage drop at the channel P well 4 is not increased due to the hole current flowing in the vicinity of the cell region.

[0038] In this way, the flow rate of the holes in the cell region in the vicinity of the cell edge portion will not increase due to the holes flowing from the cell edge portion. Therefore, the voltage drop will not increase at the cell region in the vicinity of the cell edge portion and the ruggedness against the latch-up phenomenon is improved as a result. In other words, the voltage drop in the cell region in the vicinity of the cell edge portion can be suppressed and the ruggedness against the latch-up phenomenon can be improved by forming the low resistant and wide P⁺ type region 20 in the vicinity of the cell edge portion.

[0039] FIGS. 4A through 10B schematically show a manufacturing process for manufacturing the IGBT thus constructed. It is noted that FIGS. 4A through 10A are schematic top views of the IGBT and FIGS. 4B through 10B are sectional views taken along a line X—X in FIGS. 4A through 10A. The manufacturing process for manufacturing the IGBT will be explained below based on FIGS. 4A through l0B.

[0040] At first, the N⁻ type epitaxial layer 2 is grown on the P type substrate 1. Then, it is used as the substrate. Next, a photoresist having an opening pattern at the region where the deep P well 3 and the P⁺ type region 20 are to be formed is formed on the upper surface of the N⁻ type epitaxial layer 2 by photolithography. Then, the deep P well 3 and the P⁺ type region 20 are formed selectively by implanting boron ions using the photoresist as a mask.

[0041] At this time, the P⁺ type region 20 is formed such that it is connected with the edge of the deep P well 3 and such that it has a concentration higher (resistance lower) and a width wider than those of the channel P well 4 to be formed in the later step. It is noted that a step required only for forming the P⁺ type region 20 may be eliminated by forming the P⁺ type region 20 together with the deep P well 3 .

[0042] Next, the surface of the N⁻ type epitaxial layer 2 is oxidized to form the gate oxide film 6a. Then, poly-silicon doped in high concentration is deposited on the gate oxide film 6ato form the gate electrode 7 having a plurality of rectangular hollow patterns. At this time, the P⁺ type region 20 is located under the gate electrode 7 as shown in FIGS. 5A and 5B. It is noted that the gate electrode 7 when the wafer is viewed from the top is indicated by slanting lines in FIG. 5A.

[0043] After that, the channel P well 4 is formed by implanting boron ions using the gate electrode 7 as a mask and by diffusing the boron ions. It is noted that because the rectangular hollow pattern of the gate electrode has a width wider than the opening pattern of the photoresist used for forming the deep P well 3, the channel P well 4 is formed at both sides of the deep P well 3 as shown in FIG. 2B. Further, the channel P well 4 is formed shallower than the deep P well 3 by adjusting the energy for implanting boron ions. Thereby, the channel P well 4 is aligned with the gate electrode 7 as shown in a schematic plan view in FIG. 6A.

[0044] Then, the photoresist having an opening pattern at the region where the high concentration P⁺ type region 10 is to be formed is formed by photolithography. Boron ions are implanted using this photoresist as a mask. Then, the boron ions are diffused to form the high concentration P⁺ type region 10 for lowering the resistance of the deep P well 3. It is noted that the opening pattern of the photoresist for forming the high concentration P⁺ type region 10 is positioned at the center of the deep P well 3 and has a width narrower than that of the deep P well 3. Therefore, the high concentration P⁺ type region 10 is formed at the center of the deep P well 3 to make contact with the emitter electrode 8.

[0045] Next, a photoresist having an opening pattern at the region where the emitter region 5 is to be formed is formed by photolithography. The emitter region 5 is formed so that one end portion thereof makes contact with the emitter electrode 8 and the channel P well 4 remains between the other end portion thereof and the N⁻ type epitaxial layer 2. Phosphorus or arsenic ions are implanted using the photoresist as a mask. Then, these implanted ions are diffused to form the emitter region 5 composed of the N⁺ type diffusion layer.

[0046] Then, after forming the interlayer insulating film 6bon the whole surface of the wafer, a contact hole for obtaining an ohmic contact with the emitter region 5 and the channel P well 4 (via the high concentration P⁺ type region 10 and the deep P well 3) is formed selectively in the interlayer insulating film 6band the gate oxide film 6aformed before.

[0047] After that, a metal film of several microns is deposited by sputtering to form the emitter electrode 8. Then, a metal film is deposited by an evaporation or spattering on the back of the P type substrate 1 to form the collector electrode 9. Thereby, the IGBT is completed.

[0048] It is noted that while the P⁺ type region 20 has been formed together with the deep P well 3 as described above, the P⁺ type region 20 may be formed separately from the deep P well 3. Further, although the P⁺ type region 20 has been formed approximately into the shape of a home base in the present embodiment, the same effect may be obtained when it is formed into a shape wider than the width of the channel P wells 4 (or wider than the opening width of the rectangular hollow pattern of the gate electrode). For instance, it may be formed approximately into a circular or rectangular shape. However, when the P⁺ type region 20 is formed so as to cross the emitter region 5, there may be a case when the voltage drop increases at the region where it crosses. Therefore, it is preferable to form the P⁺ type region 20 so that the part where it crosses the emitter region 5 is less.

[0049] FIG. 11 is a schematic view of the IGBT of the second embodiment. It is noted that because the basic structure of the IGBT is almost the same as that of the first embodiment, only the part which is different from the first embodiment will be explained below.

[0050] While the second embodiment is the same as the first embodiment in that the P⁺ type region 20 whose concentration is higher and whose width is wider than the channel P well 4 is formed at the edge portion of the deep P well 3 formed in each unit cell, it is different from the first embodiment in that the P⁺ type regions 20 are linked to each other in the neighboring unit cells. That is, the cell region is completely separated from the outside thereof by linking the P⁺ type regions 20 to each other in the neighboring unit cells in the second embodiment.

[0051] The second embodiment allows the flows of holes from the outside of the cell region to be almost completely drawn into the P⁺ type region 20. As a result, it can be more perfectly prevented that the holes at the cell edge portion flow into the channel P well 4 of the cell region in the vicinity of the cell edge portion. In other words, the second embodiment allows the flow rate of the holes to be lessened at the cell edge portion and the ruggedness against the latch-up phenomenon to be improved.

[0052] It is noted that although the manufacturing steps of the present embodiment are the same as those in the first embodiment, the mask used in the step shown in FIGS. 4A and 4B must be changed to what conforms to the P⁺ type region 20 shown in FIG. 11 because the shape of the P⁺ type region 20 is different from that of the first embodiment.

[0053] FIGS. 12A through 18B schematically show steps for manufacturing the IGBT of the third embodiment. The manufacturing process of the IGBT of the third embodiment is different from the first embodiment mainly in that although the P⁺ type region 20 has been formed before forming the gate electrode 7 to improve the ruggedness against the latch-up phenomenon in the first embodiment, the P⁺ type region 20 is formed after forming the gate electrode 7 in the third embodiment. It is noted that FIGS. 12A through 18A are schematic top views of the IGBT and FIGS. 12B through 18B are sectional views taken along lines Y—Y in FIGS. 12A through 18A. The steps for manufacturing the IGBT will be explained below based on FIGS. 12A through 18B.

[0054] At first, the substrate in which the N⁻ type epitaxial layer 2 has been grown on the P type substrate 1 is prepared. Then, a photoresist having an opening pattern at the region, where the deep P well 3 is to be formed, is formed on the upper surface of the N type epitaxial layer 2. The deep P well 3 is formed selectively by implanting boron ions using the photoresist as a mask.

[0055] Next, the surface of the N⁻ type epitaxial layer 2 is oxidized to form the gate oxide film 6a. Then, poly-silicon doped in high concentration is deposited on the gate oxide film 6ato form the gate electrode 7 having a nearly I-shaped hollow pattern (shape in which edges in the longitudinal direction of a rectangular shape are widened).

[0056] After that, the channel P well 4 and a peripheral region 30 are formed by implanting boron ions using the gate electrode 7 as a mask and by diffusion. It is noted that the peripheral region 30 is formed because the shape of the gate electrode 7 has been selected so as to be able to form the P⁺ type region 20 in the later steps. The peripheral region 30 has nothing to do with the role of the channel P well 4.

[0057] Then, a photoresist having an opening pattern at the region where the P⁺ type region 20 is to be formed and at a part of the deep P well 3 is formed by photolithography. Boron ions are implanted using this photoresist as a mask. Then, the boron ion are diffused to form the high concentration P⁺ type region 20. The concentration of the part of the deep P well 3 is increased to form the high concentration P⁺ type region 10. It is noted that the P⁺ type region 20 continues to the high concentration P⁺ type region 10. When the IGBT is actually operated, the positive holes drawn into the P⁺ type region 20 flow to the emitter electrode 8 via this high concentration P⁺ type region 10.

[0058] Next, a photoresist having an opening pattern at the region where the emitter region 5 is to be formed is formed by photolithography. Phosphorus or arsenic ions are implanted using the photoresist as a mask. Then, the implanted ions are diffused to form the emitter region 5 composed of the N⁺ type diffusion layer.

[0059] Then, after forming the interlayer insulating film 6bon the whole surface of the wafer, a contact hole for obtaining an ohmic contact with the emitter region 5, the channel P well 4 and a P⁺ type region 10 is formed selectively in the interlayer insulating film 6band the gate oxide film 6a. It is noted that because the P⁺ type region 20 is formed after forming the gate electrode 7, the gate electrode 7 is not formed above the P⁺ type region 20. Therefore, the ohmic contact is made also in the P⁺ type region 20 by removing the interlayer insulating film 6b and the gate oxide film 6aabove the P⁺ type region 20.

[0060] After that, a metal film of several microns is deposited by sputtering to form the emitter electrode 8. Then, a metal film is deposited by evaporation or spattering on the back of the P type substrate 1 to form the collector electrode 9. Thereby, the IGBT is completed.

[0061] Thus, the P⁺ type region 20 may be formed independently from the deep P well 3. Because the P⁺ type region 20 is formed in the separate steps from the deep P well 3, the concentration of the P⁺ type region 20 may be enhanced without depending on the concentration of the deep P well 3. When the concentration of the P⁺ type region is enhanced, this P⁺ type region facilitates the holes to be drawn therethrough.

[0062] Next, a modification of the third embodiment will be described.

[0063] Although the rectangular deep P well 3 is formed in the step shown in FIGS. 12A and 12B in the third embodiment, the I-shaped deep P well composed of a rectangular portion i-lo and a width-spreading portion extending from the edges of the rectangular portion may be formed instead. In this case, when the hollow pattern of the gate electrode is a nearly I-shaped pattern as shown in FIG. 13A, the high concentration P⁺ region having a higher impurity concentration can be obtained by overlapping with the above-described the width-spreading portion of the deep P well at the time of formation of the high concentration P⁺ region in a later manufacturing step.

[0064] The width-spreading portion of the deep P well can be formed to have a width wider than that of the channel P well of when the gate electrode has been disposed. For achieving this, the width of the width-spreading portion of the I-shaped hollow pattern of the gate electrode may be made narrower than that of the width-spreading portion of the deep P well.

[0065] According to this construction, because the P type region at the cell edge portion can be made higher in concentration than the case of the third embodiment, it is possible to facilitate drawing of holes.

[0066] It is noted that although the present invention has been applied to the IGBT in the embodiments described above, it is applicable to a DMOS transistor, instead of the IGBT. For instance, when a DMOS transistor to which the present invention is applied breaks down, the P⁺ type region allows to suppress the density of current at the cell edge portion from increasing and to suppress the parasitic operation from occurring. Furthermore, the present invention is also applicable to an up-drain type DMOSFET or IGBT.

[0067] Further, in the above embodiments, examples having a planar type DMOS structure are disclosed. However, the present invention is also applicable to an example having a trench gate type DMOS structure.

[0068] Although preferred embodiments of the present invention has been described in detail, it is understood that the present invention is not limited to the specific embodiments but rather be construed broadly within its spirit and scope as set out in the accompanying claims.