Sawtooth electric field drift region structure for power semiconductor devices

Abstract

This invention discloses a semiconductor power device formed in a semiconductor substrate includes rows of multiple horizontal columns of thin layers of alternate conductivity types in a drift region of the semiconductor substrate where each of the thin layers having a thickness to enable a punch through the thin layers when the semiconductor power device is turned on. In a specific embodiment the thickness of the thin layers satisfying charge balance equation q*N.sub.D*W.sub.N=q*N.sub.A*W.sub.P and a punch through condition of W.sub.P<2*W.sub.D*[N.sub.D/(N.sub.A+N.sub.D)] where N.sub.D and W.sub.N represent the doping concentration and the thickness of the N type layers 160, while N.sub.A and W.sub.P represent the doping concentration and thickness of the P type layers; W.sub.D represents the depletion width; and q represents an electron charge, which cancel out. This device allows for a near ideal rectangular electric field profile at breakdown voltage with sawtooth like ridges.

Claims

1. A method for manufacturing a semiconductor power device in a semiconductor substrate comprising: implanting and forming multiple rows of horizontal slices of continuous layers of alternate P-type and N-type doped layers extending continuously and horizontally over an entire length of the semiconductor substrate by forming each of die continuous layers having a doping concentration and layer thickness to achieve a charge balance and a punch through in said alternated P-type and N-type doped layers during a conduction mode for conducting a current in a vertical direction perpendicular to and cross over the horizontal slices of the alternated P-doped and N-doped layers; forming a gate above and overlaying the alternated P-type and N-type doped layers; and forming a top electrode on a top surface of the semiconductor substrate and forming a bottom electrode on a bottom surface of the semiconductor substrate wherein the gate controls the current conducted in the vertical direction.

2. The method of claim 1 wherein: said step of implanting and forming said multiple rows of the horizontal slices of continuous layers further comprising a step of forming said continuous layers of a first conductivity type and forming said continuous layers of a second conductivity type having a layer thickness of Wp, wherein Wp<2*Wd*[Nd/(Na+Nd)] and Wd represents a depletion width, Nd represents a doping concentration of the continuous layers of said first conductivity type and Na represents a doping concentration of said continuous layers of said second conductivity type.

3. The method of claim 2 wherein: said step of forming said continuous layers of the first conductivity type also complies with a charge balance formula q*ND*WN=q*NA*Wp where q represents an electron charge and Wn represents a width of said continuous layers of a the first conductivity type.

4. The method of claim 1 wherein: said step of implanting and forming said multiple rows of horizontal slices of continuous layers further comprising a step of forming said continuous layers having a first conductivity type and forming said continuous layers of a second conductivity type for undergoing the punch through from built in depletion widths of adjacent layers of said first conductivity type.

5. The method of claim 1 wherein: said step of implanting and forming said multiple rows of horizontal slices of continuous layers further comprising a step of growing epitaxial layers over the entire length of the semiconductor substrate by using an in-situ doping and switching between an n-type and a p-type doping in growing the epitaxial layers.

6. The method of claim 1 wherein: said step of implanting and forming said multiple rows of horizontal slices of continuous layers further comprising a step of growing epitaxial layers over the entire length of the semiconductor substrate using an in-situ doping by one or more cycles of growing an epitaxial layer of a first conductivity type, then performing a surface implant of a second conductivity type.

7. The method of claim 1 further comprising: manufacturing said semiconductor power device as an insulated gate bipolar transistor (IGBT).

8. The method of claim 1 further comprising: manufacturing said semiconductor power device as an insulated gate bipolar transistors (IGBT) supported on a P-type substrate with the bottom electrode formed on the bottom surface of the P-type substrate functioning as a collector and having an N-type drift region supported on said P-type substrate with the top electrode formed on the top surface of the semiconductor substrate functioning as an emitter.

9. The method of claim 1 further comprising: manufacturing said semiconductor power device as a metal oxide semiconductor field effect transistor (MOSFET).

10. The method of claim 1 further comprising: manufacturing said semiconductor power device as an N-channel MOSFET.

11. The method of claim 1 further comprising: manufacturing said semiconductor power device as a P-channel MOSFET.

12. The method of claim 1 further comprising: manufacturing said semiconductor power device as an emitter switching thyristor.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is the cross sectional view for showing a conventional IGBT.

(2) FIG. 2 is a cross sectional view of a sawtooth IGBT of this invention, with alternating P and N type layers in the drift region.

(3) FIGS. 3A and 3B are the cross sectional views of the horizontal P layers with adjacent N layers on top and bottom for two different designs of the drift layers of this invention.

(4) FIG. 4 is an I-V diagram for illustrating a comparison of the forward voltage blocking performances of a 1200 V IGBT of a conventional design with an N drift region, and two different sawtooth IGBT devices of this invention.

(5) FIG. 5 is a diagram for illustrating a comparison of the vertical electric field along the drift region for the conventional IGBT and the two different sawtooth IGBT devices at breakdown voltage.

(6) FIG. 6 is an I-V diagram for illustrating a comparison of the forward conducting performances of the conventional IGBT with the two different sawtooth IGBT devices of this invention.

(7) FIGS. 7A and 7B show the cross section of a conventional power MOSFET, and the sawtooth power MOSFET respectively.

(8) FIGS. 8A and 8B show the cross section of a conventional power diode and the sawtooth power diode respectively.

DETAILED DESCRIPTION OF THE METHOD

(9) This invention discloses a technique to achieve a drift region with electric field that approximates the ideal rectangular shape. In order to illustrate the concept, a vertical planar Insulated Gate Bipolar Transistor will be used. However, as stated earlier, the concept can be applied to all vertical/lateral semiconductor power devices.

(10) Referring to FIG. 2 for a cross-sectional view of the sawtooth IGBT 100, the IGBT may be implemented as a component of an integrated circuit not specifically shown. The IGBT 100 is formed in a semiconductor substrate 105 that has a first conductivity type, e.g., a P type substrate 105. An epitaxial layer 110 of a second conductivity type, an N epitaxial layer 110, is supported on top of the P-type substrate 105. The IGBT 100 is a vertical IGBT device with a collector electrode 120 disposed on a bottom surface of the substrate and an emitter electrode 130 disposed on a top surface. A gate 135 is supported on top of an insulation layer 125. An emitter region 115 is formed underneath the emitter electrode 130 encompassed in a P-body region 140 extended below the emitter N-region 115 from a P+ region 145 to a region underneath the gate insulation layer 125. When a gate voltage exceeding a threshold voltage is applied, the NPN bipolar transistor is turned on. An electrical current is conducted from the emitter region 115 through the P-body region 140 and the P+ region 145 to the drift region as part of the N-epitaxial layer 110 to the substrate 105 and then to the collector 120.

(11) The vertical IGBT device 100 further includes a plurality rows of P-doped horizontal columns 150 formed in the drift epitaxial layer 110 thus forming rows of alternating N-type horizontal columns 160 and P-type horizontal columns 150. Referring to FIGS. 3A and 3B for a more detailed illustration of the doped N-type horizontal column 150 and P-type horizontal columns 160. FIG. 3A shows the dimensions of columns 150 and 160 for one possible configuration, designated IGBT1. For IGBT1, the P-type horizontal columns 150 have a width of 1 m, and the N-type horizontal columns 160 have a width of 1 m. FIG. 3B shows the dimensions of columns 150 and 160 for an alternate configuration, designated IGBT2. For IGBT2, the P-type horizontal columns 150 have a width of 0.05 m, and the N-type horizontal columns 160 have a width of 16 m. The width of the horizontal columns 150 and 160 can be calculated according to the following equations. The charge balance condition can be represented as:
q*N.sub.D*W.sub.N=q*N.sub.A*W.sub.PCharge Balance:
where N.sub.D and W.sub.N represent the doping concentration and the thickness of the N type layers 160, while N.sub.A and W.sub.P represent the doping concentration and thickness of the P type layers 150; q represents the charge of an electron, and cancels out.

(12) Furthermore, in order to assure a punch through of the P horizontal columns 150, the thickness of the P horizontal columns 150 is formed within a limit according to the following equation:
W.sub.P<2*W.sub.D*[N.sub.D/(N.sub.A+N.sub.D)]
where W.sub.D is the depletion width of the junctions of the alternating P and N type horizontal columns 150 and 160.

(13) In accordance with the above two equations, the configurations for IGBT1 and IGBT2, as shown in FIGS. 3A and 3B were calculated. It should be clear to those skilled in the art that the configurations IGBT1 and IGBT2 are just two out of countless possible configurations for this invention, in accordance with the above equations.

(14) The alternating P-type and N-type horizontal columns 150 and 160 can be formed during in situ doping of the epitaxial layer 110. In situ doping dopes the epitaxial layer as it is being grown, thus the doping process can be switched between N-type and P-type doping as epitaxial layer 110 is being grown to form the alternating P-type and N-type columns 150 and 160.

(15) An alternative process of forming the horizontal columns 150 and 160 is to grow an N-type epitaxial first (to form a single N-type column 160), then perform a shallow P-type implant to form column 150. This is repeated as many times as needed to form the alternating P-type and N-type horizontal columns 150 and 160. The latter process is appropriate for situations in which there are not many horizontal columns 150 and 160 and the P-type columns 150 are thin, such as in configuration IGBT2.

(16) A major advantage of this invention is that no masks are required to make the P-type and N-type horizontal columns 150 and 160.

(17) The alternating P-type horizontal columns 150 and N-type horizontal columns 160 in the epitaxial layer 110 generate a horizontal electric field in the epitaxial layer 110 functioning as a drift region for the semiconductor power device with sawtooth profile. The sawtooth electric field enables an optimum utilization of the drift region, and requires smaller epitaxial layer thickness for the same forward blocking voltage.

(18) FIG. 4 illustrates a comparison of the forward blocking I-V curves of a conventional 1200V IGBT device with two different configurations of sawtooth IGBT devices, IGBT1 and IGBT2. The I-V curves of the conventional IGBT and the sawtooth IGBTs, IGBT1 and IGBT2, are labeled 401, 402 and 403, respectively. As can be seen in the figure, all the devices have comparable breakdown voltages, all in excess of 1200 volts. However, the sawtooth IGBT devices of this invention have an epitaxial layer that is 20% less thick than that of the conventional IGBT device.

(19) FIG. 5 compares the electric field profile at forward blocking breakdown voltage of the conventional IGBT device with two configurations of the sawtooth IGBT devices, IGBT1 and IGBT2. The horizontal axis, Distance (m) is measured from the top of the epitaxial layer to bottom. The electric field curves of the conventional IGBT, and the sawtooth IGBTs, IGBT1 and IGBT2, are labeled 501, 502 and 503, respectively. The electric field profile of the conventional IGBT 501 is trapezoidal. The field peak at the beginning occurs at the P-body 140 and N-epi 110 junction due to the curvature of the P-body 140. The electric field drops along the drift region with a slope proportional to the doping of the drift region. At the junction of the N-epi 110 and the P+ substrate 105, the electric field goes to zero. Since the voltage supported in the epitaxial layer is the area under this curve, it can be concluded that as the electric field decreases along the drift region, the voltage supported across it will also be decreased, resulting in sub-optimal usage of this layer for a conventional semiconductor power device, e.g., a conventional IGBT. As can be further observed in FIG. 5, the electrical field of the conventional IGBT has a trapezoidal field profile that leads to sub-optimal utilization of the epitaxial layer, and thus requires a thicker drift region for the same blocking voltage. The sawtooth IGBT electric field profile (502 and 503) on the other hand is flat, with the sawtooth-like shape resulting from alternating P type and N type columns 150 and 160. A flat electric field results in optimum utilization of the epitaxial layer. The voltage supported across the epitaxial layer of the sawtooth IGBTs, as represented by the area under the sawtooth curves 502 and 503, is the same as the voltage supported by the conventional IGBT, as represented by the area under the curve 501, even though the epitaxial thickness of the sawtooth IGBTs is 20% smaller. Therefore, the limitation is overcome in the sawtooth IGBT of this invention. As can be seen from the electric field profile for the two IGBT devices shown in FIG. 5. The presence of alternating, charge balanced P and N layers causes the electrical field to remain horizontal, with small spikes that make it resemble a sawtooth, This electric field profile ensures that the voltage supported along the drift region remains constant throughout, resulting optimal utilization.

(20) The presence of a thinner epitaxial region causes a lower forward voltage drop during the conduction state of this device, as shown in FIG. 6. FIG. 6 is an I-V diagram for illustrating a comparison of the forward conducting performances of the conventional IGBT with two possible configurations of sawtooth IGBT devices of this invention, IGBT1 and IGBT2. The I-V curves of the conventional IGBT, and sawtooth IGBTs, IGBT1 and IGBT2, are labeled 601, 602 and 603, respectively. As can be seen in the figure, the sawtooth IGBTs have a reduced forward voltage drop as compared to the conventional IGBT, due to the presence of a thinner drift region. Another factor that leads to lower forward voltage drop for the sawtooth IGBT is that the doping of the N type and P type layers in the epitaxial layers are independent of the breakdown voltage. Therefore, it is possible to achieve an order of magnitude higher doping for these layers, as compared to a conventional IGBT N epitaxial layer.

(21) The concept of replacing the N epitaxial layer with alternating P type and N type layers can be applied to other power semiconductor devices as well. FIGS. 7A (prior art) and 7B show the cross section of a conventional power MOSFET, and the sawtooth power MOSFET respectively. FIGS. 8A (prior art) and 8B show the cross section of a conventional power diode and the sawtooth power diode respectively.

(22) Although the present invention has been described in terms of the presently preferred embodiment, it is to be understood that such disclosure is not to be interpreted as limiting. Various alterations and modifications will no doubt become apparent to those skilled in the art after reading the above disclosure. For example, though an N-channel MOSFET device is shown, the invention could easily have been applied to a P-channel MOSFET by reversing the polarities of the regions and layers (in this case, the N-type horizontal columns would need to undergo punch through rather than the P-type ones and the punch through equation would be similarly reversed: W.sub.N<2*W.sub.D*[N.sub.A/(N.sub.A+N.sub.D)]). Accordingly, it is intended that the appended claims be interpreted as covering all alterations and modifications as fall within the true spirit and scope of the invention.