This invention relates to the field of semiconductor devices, more specifically but not restricted to, high voltage semiconductor devices, such as lateral high voltage devices in thick-film SOI power integrated circuit technologies.

High voltage semiconductor devices such as diodes, transistors and insulated gate bipolar transistors usually include heavily doped semiconductor areas so as to define a p-n junction. A p-n junction is one of the basic building blocks of semiconductor technology. For high voltage devices such a junction comprises three regions: a p- doped region which has a high concentration of acceptors, acting as 'holes'; an n- doped region which has a high concentration of donors, and a drift region separating the n- and p-doped regions. The drift region is also known as a lightly-doped region or voltage-supporting region. The transport of charge carriers through the drift region indicates whether the device is switched on or off. The p-n junction depletes in the off- state (i.e. charge carriers are swept out) and the drift region insulates the p-doped region from the n-doped region, thus supporting the off-state voltage. Many high voltage devices can carry voltages of 60 V and higher.

Figure 1 shows a high voltage junction 7 represented by a PIN junction. However, other configurations are possible such as RESURF and Super Junctions. The junction 7 shown in Figure 1 has three regions: the p-doped region 7a, the drift region 7b and the n-doped region 7c. The drift region separates the p- and n-doped portions. Figure 1 shows a theoretical illustration, since in this example the junction is suspended in air or vacuum 4.

Figure 2 shows the junction 7 of Figure 1 , wherein a voltage difference V1 is applied across the p-n junction. The electrical potential lines 8 in air or vacuum are homogeneous in the drift region 7b and spread out slowly in vacuum 4. There are no regions in which the density of the electrical field lines increases significantly with respect to the drift region 7b.

However, in any practical application, in particular in ICs, the junction will not be suspended in air or vacuum, but will be surrounded by other semiconductor devices. Figure 3 shows the junction of Figure 1 embedded in a dielectric region 5 and surrounded by a semiconductor region 2 instead of being suspended in vacuum or air. The semiconductor region 2 is normally held at a low voltage. The p- and n-doped regions are separated from each other by a width 7w. The p-n junction is at a distance 5w from the semiconductor region 2 in the direction indicated in the figure.

In the horizontal plane, the junction is surrounded by the dielectric region 5 as shown in Figure 3. In vertical direction, the junction is mounted on an insulating layer (not shown). This configuration is called silicon on insulator (SOI).

Figure 4 shows the junction and surrounding regions of Figure 3 including the electrical potential lines. The potential lines wrap around the higher potential region 7c causing an electric field peak at the critical n+/drift/dielectric interfaces, indicated by 'A' in Figure 4. The electric field peak causes charge carriers inside the doped portion to accelerate towards the interface. If the acceleration of the carriers is large, they will leave the p-n junction and enter the dielectric material from which they will not be able to return. This effect is called carrier injection into the dielectric material and is undesirable, since it will alter the device's electrical characteristics and parameters such as breakdown voltage and on-resistance. Depending on the electric field and current level this may lead to short-term failure or long-term parameter drift, which degrades reliability.

The distance 5w between the sides of the p-n junction where the potential lines leave the drift region and the surrounding semiconductor material 2 is preferably made to be about the same as, or larger than, the width of the drift portion 7b, in order to reduce the electric field peak at the interface. However, even when the distance 5w is large compared to the dimensions of the p-n junction, there will be some bending in the potential lines and an electric field peak because the surrounding region 2 is held at a low voltage, for example OV. A large width 5w causes the combined p-n junction and dielectric material to become larger and take up more space.

It has been appreciated by the inventors that the junction not only needs to be insulated from other devices, but also needs to be terminated such that it does not disturb the electric field in the drift region. Both insulation and termination can be achieved by p-n junctions or by embedding the junction in an insulating material such as a dielectric material.

At least some embodiments of the present invention take measures to actively shape the electric field at an interface of an n-doped portion, a drift portion and a termination portion in a semiconductor device. It is an aim of at least some embodiments to move points of a high electric field away from the interface into a dielectric termination region where it will not cause any damage or at least reduce or eliminate deterioration of the device. Advantages of some embodiments are an increased breakdown voltage, better device stability, improved reliability such as hot-carrier integrity and reduction of the overall size of the device.

Accordingly, in one aspect the present invention provides a semiconductor device comprising: a p or p+ doped portion; an n or n+ doped portion separated from the p or p+ doped portion by a semiconductor drift portion; an insulating portion provided adjacent the drift portion and at least one of the doped portions in a region where the drift portion and said at least one doped portion meet; and at least one additional portion which is arranged for significantly reducing the variation of the electric field strength in said region when a voltage difference is applied between the doped portions.

Some preferred embodiments of the invention will now be described by way of example only and with reference to the accompanying drawings, in which:

Figure 1 is a plan view of a semiconductor p-n junction suspended in vacuum;

Figure 2 is a plan view of the semiconductor device of Figure 1 showing electrical potential lines; Figure 3 is a plan view of a semiconductor p-n junction surrounded by an insulating region and semiconductor material;

Figure 4 is a plan view of the semiconductor device of Figure 3 in which electrical potential lines are shown;

Figure 5 is a plan view of a semiconductor p-n junction with conducting islands according to an embodiment of the present invention; Figure 6 is a plan view of the semiconductor p-n junction of Figure 5 showing electrical potential lines;

Figure 7 is a plan view of a semiconductor p-n junction with further conducting islands according to an embodiment of the present invention; Figure 8 is a plan view of the semiconductor p-n junction of Figure 7 showing electrical potential lines;

Figure 9 is a plan view of the semiconductor p-n junction of Figure 7 showing electrical potential lines; Figure 10 is a plan view of a semiconductor p-n junction with an extended n-doped portion according to an embodiment of the present invention;

Figure 11 is a plan view of the semiconductor p-n junction of Figure 10 showing electrical potential lines;

Figure 12 is a plan view of a semiconductor p-n junction with extended doped portions according to an embodiment of the present invention.

Figure 5 shows an embodiment of the present invention. A p-n junction 7 consists of a heavily p-doped region 7a and a heavily n-doped region 7c, which regions are separated by a drift region 7b. The p-n junction is embedded in a dielectric termination region 5, which insulates the junction from a semiconductor region 2 formed around the dielectric termination region 5 in the horizontal plane. The p- and n-doped portions have an oblong shape in the horizontal plane. Two long sides of the doped portions 7a and 7c are facing each other. The two long sides of the doped portions 7a and 7c are parallel such that when a voltage difference is applied between the doped portions, the field between the doped portions is substantially homogeneous. Two islands 2c of an at least partially conducting material are positioned in the dielectric termination region 5 separated from the semiconductor material 2 and the doped portions 7a and 7b by the dielectric termination region. The two islands are juxtaposed with respect to the short sides of the n-doped portion 7c. The sides of the two islands facing the n-doped portion 7c have the same length as the short side of the n-doped portion, but this is not essential. The other two sides of the two islands (generally facing the p-doped portion) lie in the plane defined by the long side of the n-doped portion. The interface between the dielectric termination region 5 and the semiconductor region 2 is substantially parallel to the nearest side of the p-n junction in order to substantially avoid inhomogeneities in the electric field. Figure 6 shows the configuration of Figure 5 with electrical potential lines. The islands 2c have a connection (not shown) for being biased to the same voltage V1 as the n- doped region 7c. The surrounding semiconductor region 2 has a connection for being biased to a low voltage, such as OV. In the example shown in Figure 6, the semiconductor region 2 is biased to the same voltage of OV as the p-doped portion. As there is no voltage drop across the n-doped portion 7c and the conducting island 2c, the potential lines exit the drift region uniformly on the interface between the drift portion 7b and the dielectric termination portion 5. Since the sides of the islands generally facing the p-doped portion lie in the plane defined by the side of the n-doped portion facing the p-doped portion the electrical potential lines exit the interface between the drift region 7b and the dielectric termination region 5 homogeneously. As a result, the regions with a high concentration of electrical field lines have moved away from the interface between the n-doped portion 7c, the dielectric termination portion 5 and the drift region 7b to the interface between the islands and the dielectric termination region. The areas indicated by 'A' in Figure 6 are the regions with large electric field densities. The islands 2c are not directly involved in the device's operation, i.e. substantially no current is flowing through the islands, and therefore there is substantially no degradation effect due to carrier injection.

The distance 5w between the semiconductor region 2 and the p-n junction 7 can now be made smaller than in the case where there are no islands, as is shown in Figure 4. Without islands, a reduction of the distance 5w would significantly increase the density of electric field lines at the interface between the n-doped portion 7c, the dielectric termination portion 5 and the drift region 7b.

More islands can be formed and connected for being biased in order to be able to manipulate the shape of the electric field. Figure 7 shows an embodiment in which additional islands 2a are formed near to p-doped region 7a inside the dielectric termination region 5. The positioning of the additional islands 2a with respect to p- doped region 7a is substantially the same as the positioning of the islands 2c discussed above with respect to n-doped region 7c. The general positioning and shape of the islands is preferably symmetrical around the axes of symmetry of the p-n junction, because asymmetrical arrangements could result in electrical field lines being asymmetrical as well or would even result in further electric field peaks. The material of the islands 2a, 2c can be the same as that of the doped regions 7a or 7c, or it can be a metal which is connected to be biased to the same voltage as the corresponding doped regions 7a or 7c. The islands should be at least partially conductive.

Figure 8 and 9 show the device of Figure 7 including potential lines. The islands and the dielectric region 2 are connected to be biased to a desired voltage. Figure 8 shows the p-doped region 7a, the adjacent islands 2a and the dielectric region 2 all being biased to the same voltage of OV. The n-doped region 7c and the adjacent islands 2c are biased to a second voltage V1. The potential lines wrap around the n-doped region 7c and islands 2c.

In Figure 9 the n-doped region 7c, the adjacent islands 2c and the semiconductor region 2 are all biased to the same voltage of V1. The p-doped region 7a and the adjacent islands 2a are biased to a second voltage of OV. The potential lines wrap around the p-doped region 7a and islands 2a.

Further embodiments are shown in Figures 10 to 12. Instead of islands being formed which are separate and insulated from the doped portions 7a, 7c, one or both of the doped portions are extended into the dielectric termination region in the direction of the short sides of the doped portions.

In Figure 10, n-doped region 7c is extended laterally. While now the extended region 7c is part of the active device 7, the impact of dielectric degradation to the HV device's characteristics is small. The highly doped region 7c does not significantly deplete outside the region contacting the drift region 7b. The part of the highly doped region 7c which does not deplete (i.e. the portion of region 7c which extends beyond drift region 7b) does not carry out the HV blocking function. Therefore, no (significant) carrier injection occurs at the outer corners of the extension portions despite the electric field being strong.

Figure 1 1 shows the device of Figure 10 in which a voltage OV is applied to the semiconductor region 2 and the p-doped portion 7a while a voltage V1 is applied to the n-doped portion 7c. The electrical potential lines 8 wrap around portion 7c and the regions with a high density of the electric field have moved away from the interface between the drift region 7b and the dielectric termination region 5. There are also no regions with high electric field density around the p-doped portion when the device is biased in this way. One could also extend the p-doped portion 7a instead of the n- doped portion 7c, to suit an arrangement in which the semiconductor portion 2 is biased to the same voltage as the n-doped portion so that there are no regions with high electric field density at the interface between doped region 7a, the drift region 7b and the dielectric termination region 5.

One may also extend both the p-doped region 7a and the n-doped region 7c into the termination region 5 as described above, so that, independent from the bias of surrounding dielectric portion 2, there are substantially no regions with high electric field density at any interface between doped region 7a or 7c, the drift region 7b and the dielectric termination region 5. Such a structure is shown in Figure 12. The electrical potential lines distribution will be similar to that shown in Figures 8 or 9.

In some embodiments, the islands 2a, 2c measure between 0.2μm by 0.2μm and 10μm by 10μm and the distance between the islands and the doped portions is between 0.2μm and 10μm. In some embodiments, the length of the extension of regions 7a and/or 7c is between 0.2μm and 10μm. In the above embodiments, the distance 5w (between the drift portion 7b and the surrounding semiconductor portion 2) is between 1 μm and 20μm.

Sections 7a and 7c in the above embodiments are heavily doped p+ and n+ regions, which is typical for a diode. However, these regions may also be p and n doped regions, as may be typical in case of a HV MOSFET, for example.

The drift region 7b can be made of a single basis material (e.g. silicon), but the doping may either be uniform (e.g. a PIN diode) or non-uniform. For example, it can have a graded doping or it can consist of one pair or several layers of n- and p-type material (e.g. super junctions).

Although the invention has been described in terms of preferred embodiments as set forth above, it should be understood that these embodiments are illustrative only and that the claims are not limited to those embodiments. Those skilled in the art will be able to make modifications and alternatives in view of the disclosure which are contemplated as falling within the scope of the appended claims. Each feature disclosed or illustrated in the present specification may be incorporated in the invention, whether alone or in any appropriate combination with any other feature disclosed or illustrated herein.