ELECTRONIC DEVICE WITH TRANSISTORS

Description

Domaine technique

[0001] This description relates generally to electronic devices. This description relates more specifically to high-electron-mobility transistors (HEMTs), and in particular to normally blocked gallium nitride (GaN) HEMTs and the processes for making such transistors. Technique antérieure

[0002] Normally blocked HEMTs based on gallium nitride are commonly used, for example, in electrical power conversion applications, typically for power levels ranging from a few milliwatts to several tens of watts. There are different types of normally blocked HEMTs, including metal-insulator-semiconductor (MIS) HEMTs, also known as MIS-HEMTs. Such transistors generally exhibit a low threshold voltage and high on-state resistance, which negatively impacts their performance.

[0003] Document WO 2017 / 213644 describes a monolithic integration of a background P-channel transistor with a type III-N N-channel transistor. Résumé de l'invention

[0004] There is a need to improve existing high-electron-mobility transistors that are normally blocked and based on gallium nitride.

[0005] One embodiment overcomes all or part of the disadvantages of existing normally blocked high electronic mobility gallium nitride-based transistors.

[0006] The invention is defined in the attached claims.

[0007] One embodiment provides for an electronic device comprising: a semiconductor region located on a gallium nitride layer; two electrodes, located on either side and insulated from the semiconductor region, the electrodes partially penetrating the gallium nitride layer; and two lateral MOS transistors formed in and on the semiconductor region.

[0008] According to one embodiment, each transistor is configured to control the passage of an electric current between one of said electrodes and one or more second electrodes carried by the semiconductor region.

[0009] According to one embodiment, each transistor includes a gate region located above a channel region comprised between source and drain regions formed in the semiconductor region.

[0010] According to one embodiment, the source regions of the transistors are doped with a first type of conductivity and separated by a well formed in the semiconductor region and doped with a second type of conductivity, opposite to the first type of conductivity.

[0011] According to one embodiment, the drain region of each transistor is connected, via a two-dimensional electron gas, to one of said electrodes.

[0012] According to one embodiment, the device further comprises one or more conductive regions contacting the drain regions of the transistors and partially penetrating the gallium nitride layer.

[0013] According to one embodiment, the device comprises exactly two conductive regions covering opposite sides of the semiconductor region located opposite said electrodes and a single second electrode located between the transistors.

[0014] According to one embodiment, the device comprises exactly two conductive regions including conductive vias located on either side of the semiconductor region and a single second electrode located between the transistors.

[0015] According to an embodiment not covered by the claims, the device comprises a single conducting region including a conducting via through the semiconductor region and exactly two second electrodes located on either side of the transistors.

[0016] According to one embodiment, the semiconductor region partially penetrates the interior of the gallium nitride layer.

[0017] According to one embodiment, a lower part of the semiconductor region is isolated from the gallium nitride layer.

[0018] According to one embodiment, the first electrodes are intended to be brought to the same potential.

[0019] According to one embodiment, the semiconductor region is made of a different material than the gallium nitride layer, preferably silicon or silicon carbide.

[0020] An embodiment provides a method for making a device as described, the method comprising the following steps: a) form a trench inside the gallium nitride layer; b) form the first electrodes on either side of the trench; c) fill the trench with the semiconductor region; and d) form the MOS transistors.

[0021] According to one embodiment, the process further comprises, after step d), the step of forming the second electrode(s).

[0022] According to one embodiment, the conducting region(s) are formed in step b). Brève description des dessins

[0023] These features and advantages, as well as others, will be described in detail in the following description of specific embodiments and implementation methods, provided by way of non-limiting example, in relation to the attached figures, among which: there figure 1 is a schematic, partial, cross-sectional view of an example transistor; the figure 2 is an electrical circuit equivalent to the transistor of the figure 1 ; there figure 3 is a schematic and partial cross-sectional view of an electronic device according to a first embodiment; the figure 4 is an electrical diagram equivalent to the electronic device of the figure 3 ; there figure 5 is a schematic and partial cross-sectional view of an electronic device according to a second embodiment; the figure 6 is a schematic and partial cross-sectional view of an electronic device according to a third embodiment; the figure 7 is a schematic and partial cross-sectional view of an electronic device according to a fourth embodiment not covered by the claims; the figure 8 is a schematic and partial cross-sectional view illustrating a step in the implementation of a process for manufacturing the electronic device of the figure 3 ; there figure 9 is a schematic and partial cross-sectional view illustrating a later stage in the implementation of the process for manufacturing the electronic device of the figure 3 ; there figure 10 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing the electronic device of the figure 3 ; there figure 11 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing the electronic device of the figure 3 ; there figure 12 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing the electronic device of the figure 3 ; there figure 13 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing the electronic device of the figure 3 ; there figure 14 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing the electronic device of the figure 3 ; and the figure 15 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing the electronic device of the figure 3 . Description des modes de réalisation

[0024] In the following description, various examples, embodiments, and variants are presented. Notwithstanding their names, only those covered by the claims form part of the present invention.

[0025] The same elements have been designated by the same references in the different figures. In particular, structural and / or functional elements common to the different embodiments and implementation methods may have the same references and may have identical structural, dimensional and material properties.

[0026] For the sake of clarity, only the steps and elements necessary for understanding the described embodiments and implementations have been shown and are detailed. In particular, applications and systems that could benefit from the described electronic devices are not detailed, as the embodiments and implementations of this description are compatible with common applications and systems featuring normally blocked, high-electron-mobility gallium nitride transistors.

[0027] Unless otherwise specified, when referring to two connected elements, this means directly connected without any intermediate elements other than conductors, and when referring to two coupled elements, this means that these two elements can be connected or linked through one or more other elements.

[0028] In the description that follows, when referring to absolute positional qualifiers, such as the terms "front", "back", "top", "bottom", "left", "right", etc., or relative positional qualifiers, such as the terms "above", "below", "superior", "inferior", etc., or to orientational qualifiers, such as the terms "horizontal", "vertical", etc., unless otherwise specified, it refers to the orientation of the figures.

[0029] Unless otherwise specified, the expressions "approximately", "roughly", "about", and "in the order of" mean to within 10%, preferably to within 5%, or, when referring to angular values, to within 10°, preferably to within 5°.

[0030] In the description that follows, unless otherwise specified, the terms insulator and conductor mean electrically insulating and electrically conductive respectively.

[0031] There figure 1 This is a schematic, partial cross-sectional view of an example of a High-Electron-Mobility Transistor (HEMT) 100. In this example, the HEMT 100 is normally off.

[0032] In the example shown, the HEMT 100 transistor is formed on a substrate 102, for example a wafer or a piece of wafer, only a part of which is shown in figure 1 For example, substrate 102 is made of a semiconductor material, for example silicon, sapphire, etc.

[0033] In the orientation of the figure 1 , a layer 104 covers a top face 102T of the substrate 102. As an example, the layer 104 is made of gallium nitride (GaN), for example in intrinsic gallium nitride i.e. not intentionally doped.

[0034] In the example shown, another layer 106, for example in aluminum gallium nitride (AlGaN), covers a top face 104T of the layer 104 and yet another layer 108, for example in silicon nitride (SiN), covers a top face 106T of the layer 106.

[0035] Transistor 100 has a gate region 110G. This gate region 110G is recessed within the gallium nitride layer 104. More precisely, in the example shown, the gate region 110G extends vertically from an upper face 108T of layer 108 towards the upper face 102T of the substrate 102. Each layer 106, 108 thus comprises two disjoint parts, located on either side of the gate region 110G.

[0036] In the example shown, the gate region 110G of transistor 100 includes a gate electrode 112G (G). The 112G electrode has a T-shaped cross-section, a vertical portion of which passes through layers 108 and 106 and partially penetrates the thickness of layer 104, and a horizontal portion of which extends above layer 108.

[0037] In the example shown, the sides of the vertical part of the T formed by the grid electrode 112G are inclined so that they approach each other at the bottom of the electrode 112G. This facilitates the passage of electrons from one side to the other of the electrode 112G.

[0038] The 112G grid electrode is made of a conductive material. For example, the 112G electrode is made of a metal or a metal alloy, for example a titanium nitride (TiN) and tungsten (W) alloy.

[0039] In the example shown, yet another layer 114 isolates electrode 112G from layers 104, 106 and 108. Layer 114 is made of an insulating material, for example alumina (Al 2 O 3 ).

[0040] The insulating layer 114 covers the lateral faces 116L, 116R, or walls, and a lower face 116B, or base, of the vertical part of the T formed by the electrode 112G. Furthermore, the insulating layer 114 extends laterally on either side of the vertical part and under the horizontal part of the T formed by the electrode 112G. The layer 114 extends over and in contact with the upper face 108T of the layer 108.

[0041] In the example shown, the horizontal part of the T formed by the electrode 112G extends over and is in contact with parts of the layer 114. These parts of the layer 114 are thus intercalated vertically between the layer 108 and the horizontal part of the T formed by the electrode 112G.

[0042] The portions of layer 114 located directly above gate electrode 112G can be considered as part of gate region 110G of transistor 100.

[0043] In the example shown, the insulating layer 114 has a virtually constant thickness. For example, the thickness of layer 114 is approximately 30 nm, plus or minus 20%. This allows for adequate isolation of the gate electrode 112G from the gallium nitride layer 104 while maintaining a relatively low threshold voltage Vth, for example, on the order of 1 V.

[0044] Transistor 100 also includes two other electrodes, 112S and 112D. Electrodes 112S and 112D are located on either side of the gate electrode 112G of transistor 100. Electrode 112S, for example, is closer to the gate electrode 112G than electrode 112D. Electrodes 112S, 112G, and 112D of transistor 100 extend perpendicularly to the cutting plane of the figure 1 , along a direction commonly referred to as the width (W) of transistor 100. The cutting plane of the figure 1 is thus oriented parallel to another direction commonly called the length (L) of transistor 100 and perpendicular to the top face 108T of layer 108.

[0045] For example: The 112S electrode is separated from the 112G electrode by a distance of between 1 and 2 µm, for example equal to about 1 µm; and the 112D electrode is separated from the 112G electrode by a distance of between 10 and 20 µm, for example equal to about 10 µm, in order to ensure voltage withstand for example up to 1000 V (the gallium nitride of the 104 layer can theoretically withstand a maximum electric field of the order of 2 MV / cm; however, this field is in practice limited to about 1 MV / cm due to defects present in the 104 layer).

[0046] In the example shown, electrodes 112S and 112D each extend vertically from the top face of the insulating layer 114 towards the top face 102T of the substrate 102. More specifically, in the example shown, electrodes 112S and 112D each pass through layers 114, 108 and 106 and partially penetrate the thickness of layer 104.

[0047] Electrodes 112S and 112D are each made of a conductive material, for example, a metal or a metal alloy. Electrodes 112S and 112D are, for example, made of the same material as electrode 112G.

[0048] For example, when transistor 100 is operating, electrode 112S is a source electrode (S) and electrode 112D is a drain electrode (D). Since electrode 112D is further from electrode 112G than electrode 112S, this allows a high potential, for example on the order of 650 V, to be applied to electrode 112D without risk of breaking down transistor 100, as electrodes 112S and 112G are generally subjected to potentials on the order of a few volts.

[0049] In the HEMT 100 transistor, a two-dimensional 2DEG electron gas, symbolized in figure 1 by a dotted line, forms inside layer 104 near an interface between layers 104 and 106. The electrodes 112S and 112D of the HEMT 100 transistor are each in contact with the two-dimensional 2DEG electron gas.

[0050] In the example shown, where the HEMT 100 transistor is normally off, the 2DEG two-dimensional electron gas is interrupted by the 110G gate region. More precisely, in this example, the 2DEG two-dimensional electron gas is discontinuous and has two parts located on either side of the 110G gate region of transistor 100 (to the left and right of the 110G region, in the orientation of the figure 1 ).

[0051] When a voltage Vgs lower than the threshold voltage Vth of transistor 100, for example, substantially zero, is applied between the gate electrode 112G and the source electrode 112S, the gate region 110G embedded in the gallium nitride layer 104 prevents electrons from flowing between the source electrode 112S and the drain electrode 112D. Transistor 100 is then in a blocked state.

[0052] Conversely, when the voltage Vgs applied between the gate electrode 112G and the source electrode 112S exceeds the threshold voltage Vth of transistor 100, electrons can flow between the source electrode 112S and the drain electrode 112D. Transistor 100 is then in a conducting state.

[0053] In the conducting state, electrons flow from one part of the two-dimensional electron gas 2DEG to the other, bypassing the gate region 110G. More precisely, when the transistor 100 is conducting and subjected to a bias voltage applied between its drain D and its source S, the electrons then use a conduction path located within the 104 layer along an interface between the 104 layer and the insulating layer 114 to flow from one side of the gate G to the other.

[0054] The conduction path taken by the electrons to bypass the 110G gate region is symbolized, in figure 1 , by arrows 118L, 118B and 118R. More precisely, in the orientation of the figure 1 : Arrow 118L symbolizes a path taken by electrons to move down the side 116L of the grid electrode 112G located on the source electrode 112S side; arrow 118B symbolizes a path taken by electrons to move horizontally under the bottom 116B of the grid electrode 112G; and arrow 118R symbolizes a path taken by electrons to move up the other side 116R of the grid electrode 112G opposite the side 116L, located on the drain electrode 112D side.

[0055] Although this was not represented in figure 1 , the transistor 100 may also include other elements such as insulating layers covering layer 114 and / or part of the gate electrode 112G, field plates, contact resumption elements, etc.

[0056] One drawback of HEMT transistors similar to transistor 100 is their high drain-source resistance (Ron) in the on-state, which significantly degrades their electrical performance. This is primarily due to resistances associated with the conduction paths 118L, 118B, and 118R, which electrons use to bypass the gate region 110G, resulting in reduced electron mobility.

[0057] Another drawback of these transistors stems from defects within the gallium nitride layer 104, which induce electron trapping on the conduction paths 118L, 118B, and 118R. This undesirable hysteresis phenomenon can be observed by measuring variations in drain current Id as a function of gate-source voltage Vgs (Id(Vgs) curves), as well as a drain current Id attenuation phenomenon after transistor biasing, also known as current collapse.

[0058] To reduce the on-state resistance Ron of transistor 100, one could consider a structure in which the insulating layer 114 is omitted and the gate electrode 112G does not penetrate the gallium nitride layer 104. This would be equivalent, for example, to having electrode 112G stop within the thickness of layer 106, before the interface between layers 106 and 104. This would create a Schottky-type gate, which would locally interrupt or attenuate the two-dimensional electron gas 2DEG directly above the gate electrode 112G, resulting in a normally off transistor. However, this would not allow a threshold voltage Vth greater than approximately 1 V, which is problematic for most applications using HEMT transistors.

[0059] One could also consider creating a structure in which the 110G gate region is not embedded in layers 108, 106, and 104. This would amount, for example, to forming the 112G gate electrode on and in contact with layer 108. As an example, one could notably implant fluorine ions (F+) in layer 106, directly above the 112G gate electrode, which would have the effect of attenuating or interrupting the two-dimensional 2DEG electron gas under the gate of transistor 100. However, this would tend to complicate control of the threshold voltage Vth.

[0060] Alternatively, one could omit layer 114, make the 112G gate electrode above layer 108 and insert a P-type doped gallium nitride (p-GaN) layer between the gate electrode and layer 108. However, this would not allow a sufficiently high threshold voltage Vth to be achieved for the intended applications.

[0061] Alternatively, one could consider reducing the width of the lower face 116B of the vertical part of the T formed by the gate region 110G, so as to reduce the length of the horizontal conduction path 118B. However, this would not significantly reduce the on-state resistance Ron of the transistor 100.

[0062] There figure 2 is an electrical circuit equivalent to transistor 100 of the figure 1 .

[0063] As illustrated in figure 2 The on-state resistance Ron of transistor 100, between its source terminal 112S and its drain terminal 112D, can be broken down into: a resistance Rgs, corresponding to a conduction path using the part of the two-dimensional electron gas 2DEG located between the source electrode 112S and the grid electrode 112G; another resistance Rt1, corresponding to the conduction path 118L taken by the electrons to go down the flank 116L of the electrode 112G; yet another resistance Rg, corresponding to the conduction path 118B taken by the electrons to move horizontally under the bottom 116B of the electrode 112G; yet another resistance Rt2 corresponding to the conduction path 118R taken by the electrons to go up the flank 116R of the electrode 112G; and yet another resistance Rgd, corresponding to a conduction path using the part of the two-dimensional electron gas 2DEG located between the gate electrode 112G and the drain electrode 112D.

[0064] The resistors Rgs, Rt1, Rg, Rt2, and Rgd, whose series combination is equivalent to the resistance Ron (neglecting the source and drain ohmic contact resistances), impair the operation of transistor 100, as previously explained in relation to the figure 1 In particular, the resistances Rt1 and Rt2 generally have a large value, which strongly impacts the resistance Ron in the on state of transistor 100.

[0065] There figure 3 is a schematic and partial cross-sectional view of an electronic device 300 according to a first embodiment.

[0066] The device 300 is, for example, formed on a substrate 302 similar to the substrate 102 previously described in relation to the figure 1 .

[0067] In the orientation of the figure 3 , a 304 layer covers a 302T top face of the 302 substrate. As an example, the 304 layer is made of gallium nitride (GaN), for example in intrinsic gallium nitride i.e. not intentionally doped.

[0068] In the example shown: another layer 306, for example of aluminum gallium nitride (AlGaN), covers a top face 304T of the layer 304; yet another layer 308, for example a passivation layer of silicon nitride (SiN), covers a top face 306T of the layer 306; and yet another layer 310, for example of alumina (A2O3), covers a top face 308T of the layer 308.

[0069] According to one embodiment, the device 300 comprises a semiconductor region 312, located on the gallium nitride layer 304, and two electrodes 314L (D) and 314R (D), located on either side of the semiconductor region 312. In the example shown, the semiconductor region 312 and the electrodes 314L and 314R partially penetrate the interior of the gallium nitride layer 304. The electrodes 314L and 314R are isolated from the region 312, in other words the electrodes 314L and 314R are not in contact with the region 312. Each electrode 314L, 314R is for example separated from the region 312 by a distance between 10 and 20 µm, for example equal to about 10 µm, in order to ensure a voltage withstand for example up to 1000 V.

[0070] The semiconductor region 312 is made of a different material than the layer 304. The region 312 is preferably made of silicon (Si) or silicon carbide (SiC).

[0071] In device 300, a two-dimensional electron gas 2DEG, symbolized in figure 3 indicated by a dashed line, is present in layer 304 near an interface between layers 304 and 306. Electrodes 314L and 314R of device 300 are both in contact with the 2DEG two-dimensional electron gas. In the example shown, the 2DEG two-dimensional electron gas is interrupted by region 312 and thus comprises two parts located on either side of region 312 (to the left and right of region 312, in the orientation of the figure 3 ).

[0072] In the example shown, flanks 312L and 312R of the semiconductor region 312, located respectively opposite electrodes 314L and 314R, are coated with conductive regions 316L and 316R. The conductive regions 316L and 316R extend, for example, over the entire height of the flanks 312L and 312R of the semiconductor region 312 and continue onto and in contact with an upper face 310T of the layer 310. In particular, in this example, the conductive regions 316L and 316R penetrate into the gallium nitride layer 304 and each contact a portion of the two-dimensional electron gas 2DEG. Each part of the two-dimensional electron gas 2DEG thus extends laterally between one of the conducting regions 316L, 316R and the electrode 314L, 314R facing it.

[0073] In one embodiment, the device 300 further comprises two insulated-gate metal-oxide-semiconductor field-effect transistors (MOSFETs), also known as MOSFETs or, more simply, MOS transistors. More precisely, T1 and T2 are lateral MOS transistors formed in and on the semiconductor region 312. In the example shown, each MOS transistor T1, T2 includes a gate region 318G located above a channel region 318C situated between a source region 318S and a drain region 318D. The 318S and 318D regions are formed in the semiconductor region 312. The 318C region is also located in the semiconductor region 312.The gate region 318G of each MOS transistor T1, T2 includes, for example, an insulating region 320, located on and in contact with a top face of the region 312 above the channel region 318C, and a conducting region 322 covering a top face of the insulating region 320. The regions 320 and 322 form, for example, respectively, a gate insulator, or oxide, and a gate electrode of the MOS transistor.

[0074] The drain regions 318D of transistors T1 and T2 are respectively located on and in contact with portions of the conductive regions 316L and 316R. In this way, the conductive regions 316L and 316R establish electrical contact between, on the one hand, the drain regions 318D of transistors T1 and T2 and, on the other hand, the portions of the two-dimensional electron gas located on the electrode side 314L and 314R of device 300. The conductive regions 316L and 316R each exhibit substantially zero resistance, thus advantageously allowing electrons to flow easily between the portions of the two-dimensional electron gas 2DEG and the drain regions 318D of transistors T1 and T2.

[0075] In the example shown, the device 300 further includes another electrode 314C (S) located between the two MOS transistors T1 and T2. The electrode 314C is carried by the semiconductor region 312. More precisely, in this example, the electrode 314C is located on and in contact with the upper face of the region 312 directly above a well 324 formed in the region 312.

[0076] The source regions 318S and drain regions 318D of MOS transistors T1 and T2 are, for example, doped with a first type of conductivity, for example, n-type, while well 324 and the other parts of semiconductor region 312 are, for example, doped with a second type of conductivity opposite to the first type of conductivity, p-type in this example. More precisely, regions 318S and 318D are, for example, heavily doped with n-type (n+), while well 324 is heavily doped with p-type (p+), and the remaining parts of region 312 are lightly doped with p-type (p-). In this case, the channel regions 318C of MOS transistors T1 and T2 are lightly doped with p-type. As an alternative, well 324 can however be omitted, with electrode 314C then being in direct contact with a part of the lightly doped p-type region 312.

[0077] Electrodes 314L and 314R of device 300 are intended to be held at the same potential, for example, approximately 650 V, while electrode 314C is held at a reference potential, for example, ground (0 V). Electrodes 314L and 314R can, for example, be considered as drain electrodes of device 300, while electrode 314C can be considered as a source electrode of device 300.

[0078] Each MOSFET T1, T2 is configured to control the flow of an electric current between one of the electrodes 314L, 314R and the electrode 314C, for example, based on a control potential applied to the gate electrode 322 of the transistor. Transistors T1 and T2 are, for example, controlled simultaneously by applying a common potential (relative to the source electrode 314C) to their respective gate electrodes 322, which, depending on its value, either interrupts or establishes the current between the electrode 314L, 314R and the electrode 314C.

[0079] When transistor T1, T2 is in a conducting state, an electric current flows between electrode 314L, 314R and electrode 314C. More precisely, this electric current flows through: by the portion of the two-dimensional electron gas 2DEG located between the electrode 314L, 314R and the conductive region 316L, 316R opposite this electrode; by the conductive region 316L, 316R; by the drain region 318D, channel region 318C and source region 318S of transistor T1, T2; and by the well 324.

[0080] One advantage of device 300 is that it allows the voltage withstand function, achieved by the distance separating each electrode 314L, 314R from the semiconductor region 312, from the switching function, achieved by the lateral MOS transistors T1 and T2. This allows device 300 to operate similarly to a HEMT transistor whose source electrode corresponds to electrode 314C, whose drain electrode corresponds to electrodes 314L and 314R, and whose control is achieved by biasing the gate electrodes 322 of MOS transistors T1 and T2.

[0081] Because the switching function is performed by MOS transistors T1 and T2, each with a threshold voltage of approximately 5 V, device 300 advantageously has a higher threshold voltage than HEMT transistor 100. figure 1 .

[0082] In addition, the p-type (p+) doping of well 324 advantageously allows for the formation of a good ohmic contact, and therefore a low contact resistance, between electrode 314C and region 312. This improves the energy efficiency of device 300 by reducing its on-state resistance Ron.

[0083] Device 300 is also more compact than a device in which at least one transistor analogous to transistors T1 and T2 would be located outside the region delimited by electrodes 314L and 314R, for example on a separate chip.

[0084] There figure 4 is an electrical diagram equivalent to the electronic device 300 of the figure 3 .

[0085] The on-state resistance Ron of device 300, between its electrodes 314L, 314R and its electrode 314C, can be decomposed into two resistances RL and RR connected in parallel. Each resistance RL, RR more precisely comprises a series combination: of a resistance RdsL, RdsR corresponding to a conduction path using the part of the two-dimensional electron gas 2DEG located between the electrode 314L, 314R and the conductive region 316L, 316R; and of another resistance RcL, RcR corresponding to a conduction path through the channel region 318C of transistor T1, T2.

[0086] The resistance RdsL, for example, is approximately equal to the resistance RdsR, within manufacturing variations. Similarly, the resistance RcL is approximately equal to the resistance RcR, within manufacturing variations.

[0087] In the scheme of the figure 4 Contact resistances and resistances corresponding to conduction paths through the conductive regions 316L, 316R, through the drain regions 318D and source regions 318S and through the well 324 are not shown, these resistances being able for example to be considered negligible compared to the resistances RdsL, RdsR, RcL and RcR.

[0088] Each resistance RdsL, RdsR of device 300 is, for example, approximately equivalent to the resistance Rgd of transistor 100. Each resistance RcL, RcR is lower than the resistances Rt1, Rg, and Rt2. This is due in particular to the fact that the conduction paths through region 318C have a shorter length and lower resistivity than the conduction paths 118L, 118B, and 118R within the gallium nitride layer 304. Device 300 thus advantageously exhibits a lower on-state resistance Ron than transistor 100. Furthermore, device 300 advantageously avoids the current trapping and collapse phenomena previously mentioned in relation to the figure 1 .

[0089] There figure 5 is a schematic and partial cross-sectional view of an electronic device 500 according to a second embodiment. The device 500 of the figure 5 includes common elements with the 300 device of the figure 3 These common elements will not be detailed again below. The 500 device of the figure 5 differs from the 300 device of the figure 3 principally in that the conductive regions 316L and 316R of the device 500 do not cover the flanks 312L and 312R of the semiconductor region 312.

[0090] More specifically, in device 500, each conducting region 316L, 316R includes, for example, a conducting via 502L, 502R that partially penetrates the gallium nitride layer 304. The via 502L, 502R contacts the portion of the two-dimensional electron gas 2DEG located on the electrode side 314L, 314R. As illustrated in figure 5 , each conductive region 316L, 316R further includes a portion extending over and in contact with the drain region 318D of transistor T1, T2.

[0091] The conductive regions 316L and 316R of device 500 advantageously allow, as previously described in relation to the figure 3 , to facilitate the flow of electrons between the parts of the two-dimensional 2DEG electron gas and the 318D drain regions of transistors T1 and T2.

[0092] In the example shown, the device 500 optionally includes a layer 504 formed inside the gallium nitride layer 304. The layer 504 is, for example, a doped insulating layer of the second type of conductivity, in this example type p, and is called a back-barrier layer. As an example, the layer 504 is made of gallium aluminum nitride (AlGaN) and has an aluminum content of, for example, between 10 and 15%. The layer 504 has, for example, a thickness of between 50 and 100 nm and is located at a distance of approximately 50 nm from a lower face 312B of the region 312.

[0093] The presence of layer 504 in device 500 advantageously confines electrons to a thin portion of layer 304, located between layer 504 and the upper face 304T of layer 304. This avoids, or limits, the volume trapping of electrons within layer 304. Furthermore, layer 504 advantageously isolates the semiconductor region 312 from the substrate 302, thus avoiding or limiting, for example, the appearance of a vertical current from the substrate 302 resulting from an interruption of the current flowing between electrodes 314L, 314R and electrode 314C when device 500 is switched from the conducting state to the blocking state.

[0094] Another advantage of this embodiment is that the conductive vias 502L and 502R can be fabricated independently of the semiconductor region 312. This allows, in particular, greater flexibility in adjusting the dimensions of these vias. Furthermore, the vias 502L and 502R and the electrodes 314L and 314R can advantageously be fabricated in a single step, for example, at the end of the fabrication of the device 500.

[0095] There figure 6 is a schematic and partial cross-sectional view of an electronic device 600 according to a third embodiment. The device 600 of the figure 6 includes common elements with the 500 device of the figure 5 These common elements will not be detailed again below. The 600 device of the figure 6 differs from the 500 device of the figure 5 mainly because the semiconductor region 312 does not penetrate inside the gallium nitride layer 304.

[0096] More specifically, in device 600, the lower face 312B of the semiconductor region 312 is located at the level of layer 308, for example, on and in contact with the upper face 308T of layer 308 or within the thickness of layer 308. The lower face 312B of region 312 is thus isolated from the gallium nitride layer 304 by all or part of layer 308 and by layer 306. In device 600, region 312 has a thinner profile compared to devices 300 and 500. This advantageously limits or prevents lateral leakage in region 312 and provides better electrostatic control. The thickness of region 312 is, for example, adapted according to the intensity of a current flowing in device 600 in the conducting state, that is, when transistors T1 and T2 are controlled in the conducting state.

[0097] Compared to device 500, device 600 also lacks layer 504 in the example illustrated in figure 6 As an alternative, layer 504 may, however, be provided in device 600 to achieve advantages similar to those described in relation to the figure 5 .

[0098] As an example, the semiconductor region 312 is fabricated using a silicon-on-insulator (SOI) wafer manufacturing process, such as the process known commercially as "SMART CUT" by SOITEC. In this case, a thin silicon layer is deposited, for example, on top of the 304 layer, for example, by bonding it to the 308 layer. This thin silicon layer is then ground down to a thickness of, for example, between 50 and 100 nm, and then etched to form the 312 region of each 600 device. The semiconductor region 312 can thus advantageously be made of a crystalline material, such as silicon or crystalline silicon carbide, within which electron mobility is higher due to a low presence of defects. This further reduces the on-state resistance Ron of the 600 device.

[0099] Another advantage of the 600 device lies in the fact that part of the two-dimensional electron gas 2DEG is present beneath the 312 semiconductor region, between the two conducting vias 502L and 502R. This allows the p-type doped region 312 to be depleted electrostatically. As a result, the threshold voltage of the 600 device is increased, particularly compared to the 500 device.

[0100] As an alternative, the semiconductor region 312 can in this case be undoped, with the exception of regions 318S and 318D and possibly well 324, due to the presence of the two-dimensional 2DEG electron gas under region 312. The channel regions 318C of transistors T1 and T2 are in this case undoped.

[0101] There figure 7 is a schematic and partial cross-sectional view of an electronic device 700 according to a fourth embodiment not covered by the claims. The device 700 of the figure 7 includes common elements with the 600 device of the figure 6 These common elements will not be detailed again below.

[0102] The 700 device of the figure 7 differs from the 600 device of the figure 6 principally in that the device 700 comprises two electrodes 702L (S) and 702R (S) carried by the semiconductor region 312 and located on either side of transistors T1 and T2, on and in contact with the source regions 318S of these transistors. Furthermore, in the device 700, the two-dimensional electron gas 2DEG is electrically connected to the drain regions 318D of transistors T1 and T2 by a single conductive region 704. The region 704 includes, for example, a conductive via 706 that passes completely through the region 312, the layer 308, and the layer 306, and partially penetrates the gallium nitride layer 304. In order to obtain low contact resistances, the conductive region 704 may, as illustrated in figure 7 , include a portion extending laterally over and in contact with the 318D drain regions of transistors T1 and T2, transistors T1 and T2 being located on either side of the via conductor 706.

[0103] In the example shown, the source regions 318S of transistors T1 and T2 are separated from the top face 310T of layer 310 by heavily doped regions 708 of the second type of conductivity, here p(p+) type. The regions 708 located on either side of the conductive via 706 contact electrodes 702L and 702R, respectively. The regions 708 help maintain the potential of the lightly doped p(p-) type portion of layer 312 at the source potential. Alternatively, the regions 708 are omitted.

[0104] One advantage of device 700 is that electrodes 702L and 702R ensure better current distribution. This helps to limit or prevent overheating of device 700 when it is in the conducting state.

[0105] THE figures 8 à 15 The following illustrate successive steps in an example of a method for implementing a process for manufacturing device 300 of the figure 3 .

[0106] There figure 8 is a schematic and partial cross-sectional view illustrating a step in the implementation of the manufacturing process for device 300 of the figure 3 .

[0107] During this step, layers 304, 306, 308, and 310 are successively produced, in that order, on the 302T face of substrate 302. More precisely: layer 304 extends continuously and over the entire upper face 302T of substrate 302; layer 306 extends continuously and over the entire upper face 304T of layer 304; layer 308 extends continuously and over the entire upper face 306T of layer 306; and layer 310 extends continuously and over the entire upper face 308T of layer 308.

[0108] Layers 304, 306, 308, and 310, for example, are created by epitaxy. Although this is not shown in figure 8 An intermediate layer of aluminium nitride (AlN) can be intercalated between layers 304 and 306. If necessary, this allows the epitaxy of the aluminium-gallium nitride layer 306 to be promoted on the gallium nitride layer 304 by lattice parameter adaptation.

[0109] In figure 8 , the two-dimensional electron gas 2DEG extends laterally and continuously in the 304 layer, under the top face 304T of the 304 layer.

[0110] There figure 9 is a schematic and partial cross-sectional view illustrating a later stage in the implementation of the process for manufacturing device 300 of the figure 3 .

[0111] During this step, a trench 902 is created extending vertically towards the substrate 302 from the upper face 310T of the layer 310. More precisely, in the example shown, the trench 902 crosses the layers 310, 308 and 306 and partially penetrates the thickness of the layer 304. The trench 902 stops in the layer 304 and does not reach the upper face 302T of the substrate 302.

[0112] As an example, trench 902 is made by atomic layer etching ("Atomic Layer Etching" - ALE).

[0113] Trench 902 has side walls 902L and 902R and a bottom 902B which correspond respectively to the sides 312L and 312R and the bottom 312B of the future semiconductor region 312 of device 300.

[0114] In the example shown, the lateral walls 902L and 902R of trench 902 are oblique. More precisely, trench 902 is, for example, cut so that its lateral walls 902L and 902R each form an angle α with respect to a normal to the upper face 310T of layer 310. As an example, the angle α formed by walls 902L and 902R is between 5° and 45°, for example, approximately 10°. Trench 902 thus has a flared profile with an opening (at the top, in the orientation of the figure 9 ) wider than its base 902B (at the bottom, in the orientation of the figure 9 The inclination of the lateral walls 902L and 902R of trench 902 results, for example, from an isotropic chemical etching step. Alternatively, trench 902 may have substantially vertical lateral walls 902L and 902R, i.e., walls with an angle α of approximately 0°.

[0115] For example, trench 902 features: a maximum width Lmax between 0.4 and 1 µm, for example equal to about 0.5 µm; and a minimum width Lmin between 0.2 and 0.4 µm, for example equal to about 0.3 µm.

[0116] Alternatively, the side walls 902L and 902R of trench 902 are substantially vertical, the angle α being in this case equal to approximately 0°.

[0117] Electrodes 314L and 314R are then formed on either side of trench 902. As an example, electrodes 314L and 314R are obtained by deposition and then local etching of a metallic layer (not shown) covering the upper face of layer 310.

[0118] At the end of this step, the two-dimensional electron gas 2DEG is discontinuous. More precisely, trench 902 separates the two-dimensional electron gas 2DEG into two parts located on either side of trench 902. Layers 306, 308, and 310 are further separated, each into two disjoint layers located on either side of trench 902.

[0119] There figure 10 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing device 300 of the figure 3 .

[0120] During this step, a conductive layer 1002 is deposited on the structure. The conductive layer 1002 covers the upper face 310T of the parts of layer 310 as well as the side walls 902L, 902R and the bottom 902B of trench 902.

[0121] As an example, the conductive layer 1002 is produced by a conformal deposition technique, for example by chemical vapor deposition (CVD). This avoids, or limits, the presence of crystalline defects at the interface between the 304 and 1002 layers, as these defects can form trap states for electrons.

[0122] There figure 11 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing device 300 of the figure 3 .

[0123] During this step, the conductive layer 1002 is etched to expose the bottom 902B of the trench 902. This is achieved by removing a portion of the layer 1002 covering the bottom 902B of the trench 902. The layer 1002 is thus separated into two distinct parts. More precisely, after the etching, the conductive regions 316L and 316R are obtained, covering the walls 902L and 902R of the trench 902, respectively. Since the electrodes 314L and 314R are at the same potential, it is alternatively possible to provide a continuous conductive region, that is, one that continuously covers the sides 902L, 902R and the bottom 902B of the trench 902.

[0124] In the example shown, layer 1002 is also etched on the side of electrodes 314L and 314R, that is, on the left and right in the orientation of the figure 11 , so as to expose parts of the top face 310T of the layer 310 located between the electrodes 314L, 314R and the future semiconductor region 312.

[0125] As an example, the conductive regions 316L and 316R are produced by a wet etching process of layer 1002. This makes it possible to obtain, in particular, a low roughness at the bottom 902B of trench 902.

[0126] Wet etching can be preceded by a dry etching step, for example, plasma-assisted. In this case, wet etching improves the surface finish of the bottom of trench 902B, which is achieved after the dry etching step.

[0127] There figure 12 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing device 300 of the figure 3 .

[0128] During this step, the semiconductor region 312 is made inside the trench 902. In the example shown, the region 312 completely fills the trench 902 and extends laterally on and in contact with a top face of the parts of the conductive regions 316L and 316R located above the parts of the layer 310.

[0129] The semiconductor region 312 is for example produced by a conformal deposition technique, for example by chemical vapor deposition (CVD) of lightly doped p-type polycrystalline silicon (p-).

[0130] There figure 13 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing device 300 of the figure 3 .

[0131] During this step, the semiconductor region 312 is doped to form the 318D and 318S regions of the future MOS transistors T1 and T2 and the 324 well. As an example, the doping of the 312 region is carried out by ion implantation.

[0132] There figure 14 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing device 300 of the figure 3 .

[0133] During this step, the 320 gate oxide of each future MOS transistor T1, T2 is fabricated on and in contact with the 318D drain, 318C channel, and 318S source regions of the transistor. For example, an oxide layer, such as alumina (Al₂O₃), with a thickness between 10 and 30 nm is deposited onto the structure and then etched to form the 320 gate oxides of the future MOS transistors T1 and T2.

[0134] There figure 15 is a schematic and partial cross-sectional view, illustrating yet another later stage in the implementation of the process for manufacturing device 300 of the figure 3 .

[0135] During this step, the gate electrode 322 of each MOS transistor T1, T2 is fabricated on and in contact with the corresponding gate oxide 320. As an example, a conductive layer, for example made of a titanium nitride (TiN) and tungsten (W) alloy, is deposited on the structure and then etched to form the gate electrodes 322 of the MOS transistors T1 and T2.

[0136] Based on the structure described in relation to the figure 15 , then electrode 314C is fabricated to obtain device 300 shown in relation to the figure 3 As an example, electrode 314C is obtained by depositing and then locally etching a metallic layer (not shown) coating the structure. Electrode 314C of device 300 ( figure 3 ) is more precisely formed on and in contact with an upper face of the well 324 previously formed inside the semiconducting region 312.

[0137] Various embodiments, implementation methods, and variations have been described. Those skilled in the art will understand that certain features of these various embodiments, implementation methods, and variations could be combined, and other variations will become apparent to them.

[0138] Based on the above indications, a person skilled in the art is notably able to foresee devices similar to devices 300, 600 and 700 but having an insulating layer similar to layer 504 of device 500 inside layer 304 in gallium nitride.

[0139] Furthermore, the adaptation of the manufacturing process for device 300 described in relation to the figures 8 à 15in order to obtain devices 500, 600 and 700 of figures 6, 7 and 8 respectively, as well as devices of other embodiments and variants mentioned in this description, is also within the reach of a person skilled in the art.

[0140] Finally, the practical implementation of the described embodiments, implementation methods, and variants is within the grasp of a person skilled in the art, based on the functional specifications given above. In particular, a person skilled in the art is able to select the materials for the conductive regions 316L, 316R, and 704, as well as the doping levels for each of the regions 312, 318S, 318D, 708, and the well 324, according to the intended application.

Claims

1. Electronic device (300; 500; 600; 700) comprising: - a semiconductor region (312) located on a gallium nitride layer (304) and partially penetrating into said layer; - two electrodes (314L, 314R), located on either side of and not in contact with the semiconductor region, the electrodes partially penetrating into the gallium nitride layer and being connected to the semiconductor region via a two-dimensional electron gas, 2DEG; and - two lateral MOS transistors (T1, T2) formed inside and on top of the semiconductor region (312) and connected in series between the two electrodes (314L, 314R).

2. Device according to claim 1, wherein each transistor (T1, T2) is configured to control the flowing of an electric current between one of said electrodes (314L, 314R) and one or a plurality of second electrodes (314C; 702L, 702R) supported by the semiconductor region (312) and wherein a drain region (318D) of each transistor (T1, T2) is coupled, via the two-dimensional electron gas, 2DEG, to one of said electrodes (314L, 314R), the device further comprising one or a plurality of conductive regions (316L, 316R; 704) contacting the drain regions of the transistors and partially penetrating into the gallium nitride layer.

3. Device according to claim 2, wherein each transistor (T1, T2) comprises a gate region (318G) located vertically in line with a channel region (318C) located between source regions (318S) and the drain regions (318D) formed in the semiconductor region (312).

4. Device according to claim 3, wherein the source regions (318S) of the transistors (T1, T2) are doped with a first conductivity type and separated by a well (324) formed in the semiconductor region (312) and doped with a second conductivity type, opposite to the first conductivity type.

5. Device according to any one of claims 2 to 4, comprising exactly two conductive regions (316L, 316R) coating opposite sides (312L, 312R) of the semiconductor region (312) located in front of said electrodes (314L, 314R) and a single second electrode (314C) located between the transistors (T1, T2).

6. Device according to any one of claims 2 to 4, comprising exactly two conductive regions (316L, 316R) comprising conductive vias (502L, 502R) located on either side of the semiconductor region (312) and a single second electrode (314C) located between the transistors (T1, T2) .

7. Device according to any one of claims 1 to 6, wherein a lower portion (312B) of the semiconductor region (312) is insulated from the gallium nitride layer (304).

8. Device according to any one of claims 1 to 7, wherein the first electrodes (314L, 314R) are intended to be taken to a same potential.

9. Device according to any one of claims 1 to 8, wherein the semiconductor region (312) is made of a material different from that of the gallium nitride layer (304), preferably of silicon or of silicon carbide.

10. Method of forming a device (300; 500; 600; 700) according to any one of claims 1 to 9, the method comprising the steps of: a) forming a trench (902) inside of the gallium nitride layer (304); b) forming the first electrodes (314L, 314R) on either side of the trench; c) filling the trench with the semiconductor region (312); d) forming the MOS transistors (T1, T2).

11. Method according to claim 10 of forming a device according to claim 2, further comprising, after step d), the step of forming the second electrode(s) (314C; 702L, 702R).

12. Method according to claim 10 or 11 of forming a device according to claim 2, wherein the conductive region(s) (316L, 316R; 704) are formed at step b).