TECHNICAL FIELD

[0001] This specification relates to semiconductor devices, in particular nitride- based devices such as transistors and diodes, which include one or more ohmic contacts.

BACKGROUND

[0002] Transistors used in power electronic applications have typically been fabricated with silicon (Si) semiconductor materials. Common transistor devices for power applications include Si CoolMOS, Si Power MOSFETs, and Si Insulated Gate Bipolar Transistors (IGBTs). While Si power devices are inexpensive, they can suffer from a number of disadvantages, including relatively low switching speeds and high levels of electrical noise. More recently, silicon carbide (SiC) power devices have been considered due to their superior properties. Ill-Nitride or III-N semiconductor devices, such as gallium nitride (GaN) devices, are now emerging as attractive candidates to carry large currents, support high voltages, and to provide very low on-resistance and fast switching times.

[0003] Forming ohmic contacts to III-N material structures, which are utilized in many devices, is often achieved by depositing one or more metal layers on the III-N material and then annealing the structure, causing the metals and underlying III-N material to intermix and form an alloy. While such an alloying process for forming ohmic contacts has been shown to successfully result in low-resistivity ohmic contacts, the reliability and yield of this process has typically been lower than required for commercial scale production.

SUMMARY

[0004] In a first aspect, a device includes a III-N layer having an upper side and a lower side, the lower side being opposite the upper side, and at least one conductive contact on the upper side of the III-N layer, the conductive contact extending into the III-

N layer. The conductive contact comprises a top side facing away from the lower side of the III-N layer, and a bottom side facing towards the lower side of the III-N layer. The bottom side includes a first end and a second end opposite the first end, a first side rising from the first end to an intermediate point closer to the top side than the first end, and a second side falling from the intermediate point to the second end, the second end being further from the top side than the intermediate point.

[0005] In a second aspect, a method for fabricating a device includes forming a

III-N layer having an upper side and a lower side, the lower side being opposite the upper side. The method further includes forming a recess in a surface on the upper side of the III-N layer, the forming of the recess including etching the surface of the III-N layer using a resist pattern, and forming a conductive contact over the recess in the surface of the III-N layer. The conductive contact includes a top side facing away from the lower side of the III-N layer, and a bottom side facing towards the lower side of the III-N layer. The bottom side comprises a first end and a second end opposite the first end, a first side rising from the first end to an intermediate point closer to the top side than the first end, and a second side falling from the intermediate point to the second end, the second end being further from the top side than the intermediate point.

[0006] In a third aspect, a method for fabricating a device includes forming a recess in a surface of a III-N layer having a conductive channel therein, where forming the recess includes etching the surface of the III-N layer using a resist pattern, and forming a conductive contact over the recess in the surface of the III-N layer, the conductive contact being in electrical contact with the conductive channel, the conductive contact having a bottom side in contact with a bottom surface of the recess and a top side opposite the bottom. The etching of the surface causes the bottom surface of the recess to have a first end and a second end opposite the first end, a first side rising monotonically from the first end to an intermediate point closer to the top side of the conductive contact than the first end, and a second side falling monotonically from the intermediate point to the second end, the second end being further from the top side of the conductive contact than the intermediate point.

[0007] In a fourth aspect, a device includes a III-N layer having an upper side and a lower side, the lower side being opposite the upper side, and a conductive contact on the upper side of the III-N layer. The conductive contact includes a top side facing away from the lower side of the III-N layer and a bottom side facing towards the lower side of the III-N layer, the bottom side including a first end, a second end opposite the first end, and an intermediate point between the first end and the second end. The device further includes a 2DEG channel in the III-N layer, wherein the 2DEG channel includes a first portion below the intermediate point and second portions below the first and second ends, the second portions having a higher electron concentration than the first portion.

[0008] Devices and methods described herein may include one or more of the following features. The first side may rise monotonically from the first end to the intermediate point and the second side may fall monotonically from the intermediate point to the second end. The III-N layer may include a III-N channel layer and a III-N barrier layer, wherein a compositional difference between the III-N channel layer and the III-N barrier layer causes a 2DEG channel to be induced in the III-N channel layer adjacent to the III-N barrier layer. The first side can curve from the first end to the intermediate point and the second side can curve from the second end to the intermediate point, forming a substantially rounded trench shape in the bottom side of the conductive contact. The III-N layer may include a GaN layer, a III-N spacer layer on the GaN layer, and a III-N barrier layer on the III-N spacer layer. Additionally, the III-N spacer layer may have a larger bandgap than the III-N barrier layer. In the device, at least one of the first end and the second end may extend through the III-N spacer layer to contact the GaN layer. The III-N spacer layer can include A1N and the III-N barrier layer can include AlGaN.

[0009] The conductive contact may be a source contact, the device further including a drain contact and a gate contact, forming a transistor. The drain contact can include a drain top side facing away from the lower side of the III-N layer, and a drain bottom side facing towards the lower side of the III-N layer, wherein the bottom layer comprises a drain first end and a drain second end opposite the drain first end, a drain first side rising from the drain first end to a drain intermediate point closer to the top side than the drain first end, and a drain second side falling from the drain intermediate point to a drain second end further from the top side than the drain intermediate point. The transistor can be a lateral transistor having the drain, source, and gate on a same side. The drain first side may rise monotonically from the drain first end to the drain intermediate point and the drain second side may fall monotonically from the drain intermediate point to the drain second end. The conductive contact can have a normalized contact resistance of 0.3 Ohm-mm or less. In the device, the first end and/or the second end may have a width less than 300 nanometers. The III-N layer may be on a substrate.

[0010] Forming the conductive contact can include forming the conductive contact so that the first side rises monotonically from the first end to the intermediate point and the second side falls monotonically from the intermediate point to the second end. The method of forming the device may include heating the device to a temperature between 300° C and 600° C. Heating the device may also include heating the device for one to three minutes. Etching the surface of the III-N layer can include performing a dry etch using a chlorine based gas. Furthermore, etching the surface of the III-N layer can include performing a plasma etch in Ch plasma at an RF bias of 25W or less. Forming the III-N layer may comprise forming a GaN layer, an A1N spacer layer on the GaN layer, and an AlGaN layer on the A1N spacer layer. Additionally, forming the recess can include forming the recess through the AlGaN layer up to the A1N spacer layer. Forming the recess may include forming the recess through the AlGaN layer and into the A1N spacer layer. Forming the recess can include forming the recess through the AlGaN layer and the A1N spacer layer and into the GaN layer.

[0011] The III-N layer may include a recess, and the conductive contact is in the recess.

A separation between the 2DEG channel and the first end of the bottom side of the conductive contact may be less than a separation between the 2DEG channel and the intermediate point. The conductive contact can be in ohmic contact with the 2DEG channel.

[0012] The details of one or more embodiments of the subject matter described in this specification are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages of the subject matter will become apparent from the description, the drawings, and the claims.

DESCRIPTION OF DRAWINGS

[0013] Figure 1 A is a plan view (top view) of an example Ill-Nitride (III-N) transistor.

[0014] Figure IB is a cross-sectional view of the transistor.

[0015] Figure 2 is cross-sectional view of three example contacts. [0016] Figures 3-5 are cross-sectional views of example devices where the source and drain contacts 110 and 112 are recessed to different depths.

[0017] Figure 6 is a flow diagram of an example process for fabricating a semiconductor device.

[0018] Figures 7A-7F are cross-sectional views of an example semiconductor device during fabrication of the device.

[0019] Figure 8 is a cross-sectional view of a semiconductor material structure including a recess.

[0020] Figures 9A-9B are band diagrams along the vertical dashed lines in Figure

8.

[0021] Like reference symbols in the various drawings indicate like elements.

DETAILED DESCRIPTION

[0022] Figure 1 A is a plan view (top view) of an example Ill-Nitride (III-N) transistor 100. Figure IB is a cross-sectional view of the transistor 100. As illustrated in Figure 1 A, the transistor includes a substrate 102 (which may optionally be omitted) and a III-N layer 122.

[0023] The substrate can be silicon, SiC, A1N, GaN, sapphire, or any other suitable growth substrate for the growth of III-N materials. The III-N layer 122 includes a III-N channel layer 104 and a III-N spacer layer 106, and a III-N barrier layer 108, where the compositions of layers 104, 106, and 108 are selected to induce the 2DEG 116 in the III-N channel layer 104 near the interface between the III-N channel layer 104 and the III-N spacer layer 106. In some implementations, the spacer layer 106 may be omitted. In this case, the III-N barrier layer 108 is formed directly on the III-N channel layer 104, and the compositions of layers 104 and 108 are selected to induce the 2DEG 116 in the III-N channel layer 104 near the interface between the III-N channel layer 104 and the III-N barrier layer 108.

[0024] A gate contact 118 is deposited on the III-N layer 122. In some implementations, the gate contact 118 directly contacts the underlying III-N layer 122 (not shown). In other implementations, as illustrated in Figure IB, an insulator layer 120 is included between the gate 118 and the underlying III-N layer 122. Source and drain ohmic contacts 110 and 112 are deposited on opposite sides of the gate 118 on the III-N layer 122. The source and drain contacts 110 and 112 form ohmic (or substantially ohmic) contacts to the device channel 116. The transistor 100 can be a lateral transistor having the gate 118, drain 110, and source 112 on a same side. A voltage applied to the gate 118, relative to the source contact 112, modulates the 2DEG charge density in the gate region of the device (i.e., the region directly under the gate).

[0025] As used in this document, the terms Ill-Nitride or III-N materials, layers, devices, and structures refer to a material, device, or structure comprised of a compound semiconductor material according to the stoichiometric formula BwAlxIn _y Ga _z N, where w+x+y+z is about 1, and w, x, y, and z are each greater than or equal to zero and less than or equal to 1. In a Ill-Nitride or III-N device, the conductive channel can be partially or entirely contained within a III-N material layer.

[0026] The ohmic metals, of which source and drain contacts 110 and 112 are formed, typically include at least one metal with a relatively low metal work function, for example aluminum (Al) or titanium (Ti). In some implementations, the ohmic contacts 110 and 112 include titanium, aluminum, nickel, or gold, or combinations thereof. In some implementations, the III-N layer 122 includes a GaN layer 104 on the substrate 102, an A1N spacer layer 106 on the GaN layer 104, and an AlGaN layer 108 on the A1N spacer layer 106. The A1N spacer layer 106 can be useful, e.g., to increase mobility and lower an on-state resistance of the transistor 100. The A1N spacer layer 106 may be formed of A1N, or alternatively may be formed of another material having a larger bandgap than the AlGaN layer 108. For example, layer 106 may be formed of AlInGaN, where the compositions of Al, In, and Ga are selected such that layer 106 has a larger bandgap than layer 108. Or, layer 106 may be formed of AlGaN, where the fractional composition of Al in layer 106 is larger than that in layer 108.

[0027] The transistor 100 can include an insulator layer 120 between the gate 118 and the AlGaN layer 108. The insulator layer 120 can also serve as a passivation layer, preventing or suppressing dispersion by preventing or suppressing voltage fluctuations at the uppermost III-N surface on either sides of the gate 118. The insulator layer 120 can be made of SixNy, AI2O3, S1O2, AlxSiyN, or the like, and can be prepared by metal organic chemical vapor deposition (MOCVD), low pressure chemical vapor deposition (LPCVD), plasma enhanced chemical vapor deposition (PECVD), chemical vapor deposition (CVD), sputtering, atomic layer deposition (ALD), high density chemical vapor deposition, or any suitable deposition process. In a particular example, the insulator layer 120 is a Silicon Nitride (SixNy) layer formed by MOCVD.

[0028] The source and drain contacts 110 and 112 exhibit a normalized contact resistance that can be measured, in Ohm-mm, with respect the width W of the transistor 100. The total resistance of each contact, measured in Ohms, is then equal to the normalized contact resistance (measured in Ohm-mm) multiplied by the width W

(measured in millimeters) of the transistor. It can be useful, e.g., to improve device performance, to have ohmic contacts with low normalized resistances. In some conventional transistors, the ohmic contacts exhibit normalized resistances in the range of 0.8 - 2.0 Ohm-mm. The example transistor 100 of Figures 1A-B can have ohmic contacts that exhibit normalized resistances in the range of 0.2 - 0.3 Ohm-mm, e.g., by virtue of the shape and the depth of the ohmic contacts.

[0029] Figure 2 is cross-sectional view of three example contacts 200a-c that can be used, e.g., for any one or more of the source and/or drain contacts 110 and 112 of the example transistor 100 of Figures 1A-B. Each of the example contacts 200a-c includes a top side 208a-c facing away from the substrate 102 and a bottom side facing in the opposite direction.

[0030] The bottom side of each contact 200a-c includes a first end 202a-c and a second end 206a-c opposite the first end 202a-c. Each contact 200a-c includes a first side that rises, e.g., monotonically, from the first end 202a-c to an intermediate point 204a-c closer to the top side 208a-c than the first end 202a-c and a second side that falls, e.g., monotonically, from the intermediate point 204a-c to the second end 206a-c. The second end 206a-c is further from the top side 208a than the intermediate point 204a-c.

[0031] The first ends 202a and/or the second ends 206a of the contacts 200a-c can have a width ("w") where the end is flat or substantially flat. The width can be, e.g., between 20 and 300 nanometers, for example between 50 and 200 nanometers.

[0032] The first example contact 200a has a first side that curves from the first end 202a to the intermediate point 204a and a second side that curves from the intermediate point 204a to the second end 206a, forming a substantially rounded trench shape in the bottom side of the contact 200a. The second example contact 200b has a first side that rises in three line segments from the first end 202b to the intermediate point 204b and a second side that falls in three line segments from the intermediate point 204b to the second end 206b. The third example contact 200c has a first side that rises in three line segments from the first end 202c to the intermediate point 204c and a second side that falls in three line segments from the intermediate point 204c to the second end 206c.

[0033] Figures 3-5 are cross-sectional views of example devices 300, 400, and

500 where the source and drain contacts 110 and 112 are recessed to different depths than they are in the example device 100 of Figure 1.

[0034] Figure 3 shows an example device 300 where the source and drain contacts 110 and 112 are recessed through the spacer layer 106 to a top of a III-N channel layer 104 (e.g., so that the first and/or second end of the bottom side of the contact reach the top of the III-N channel layer 104). Figure 4 shows an example device 400 where the source and drain contacts 110 and 112 are recessed through the channel 116 of the III-N channel layer 104. Figure 5 shows an example device 500 where the source and drain contacts 110 and 112 are recessed into the III-N barrier layer 108 but do not reach the spacer layer 106.

[0035] The recess depth can affect the resistance of the contacts 110 and 112. In some implementations, increasing the recess depth can decrease the resistance up to a point where further increasing the recess depth does not decrease the resistance. Beyond a certain depth, the resistance may increase as the recess depth is increased.

[0036] Figure 6 is a flow diagram of an example process 600 for fabricating a semiconductor device. The process 600 can be used, e.g., to fabricate the devices 100, 300, 400, and 500 of Figures 1A-B and 3-5.

[0037] A III-N layer is formed on a substrate (602). The substrate can be a silicon wafer. The III-N layer can include a GaN layer on the substrate; an A1N spacer layer on the GaN layer; and an AlGaN layer on the A1N spacer layer. The III-N layer can be formed by either directly growing the III-N layer on the substrate, or alternatively by growing the III-N layer or parts of the III-N layer on a first substrate, detaching the buffer layer from the first substrate, and bonding the buffer layer to the substrate. Forming the III-N layer can include using any appropriate deposition process. [0038] One or more recesses are formed in a surface of the III-N layer (604). For example, the surface of the III-N layer can be etched using a patterned resist. Etching the surface can include performing a dry etch using a chlorine based gas. In some implementations, dry etching techniques, e.g., plasma etching, digital plasma etching, or reactive ion etching (RIE), are used to form the recesses.

[0039] The parameters of the etching can be adjusted to control the shape and depth of the recesses. For example, the ion energy used during the etch, the pressure applied during the etch, and the resist pattern used can be varied to achieve a target depth and shape. As an example, the following process may achieve a recess having the profile of the recesses in Figure IB in which source and drain contacts 110 and 112, respectively, are deposited. After formation of III-N layer structure 122 and insulator layer 120, a single layer photoresist layer having a thickness of between 1.1 and 1.2 microns is deposited over the surface and patterned to expose the underlying material in the regions is which the recess is formed. The photoresist acts as an etch mask, preventing material beneath it from being etched during a subsequent etch process. The recess is then etched in an ECR plasma etcher using Ch as an etchant gas. 50 seem of CI2 is injected into the chamber, the RF bias power is maintained at 20W, and the chamber pressure is maintained at 2.5 mTorr. Following the etch, the device is removed from the etch chamber, and the photoresist mask is removed.

[0040] Conductive contacts are formed over the one or more recesses (606). Due to the shape of the recesses, the conductive contacts have a bottom side including a first side rising, e.g., rising monotonically, from a first end to an intermediate point and a second side falling, e.g., falling monotonically, from the intermediate point to a second end opposite the first end.

[0041] The device is heated at a certain temperature for a certain amount of time

(608). For example, the device can be heated to a temperature between 300° C and 600° C. The device can be heated for one to three minutes. Typically, the heat application is performed at a temperature that is not sufficiently high to cause the contacts to alloy with the III-N layers. [0042] Figures 7A-7F are cross-sectional views of an example semiconductor device 700 during fabrication of the device 700. The device 700 can be fabricated, e.g., using the process 600 illustrated in Figure 6.

[0043] Figure 7 A shows the device 700 after III-N layers 104, 106, and 108 have been formed on a substrate 102, and insulator layer 120 has been formed on III-N layer 108. Figure 7B shows the device 700 after a patterned photoresist 122 (or alternatively another masking layer) has been placed on the device 700 so that the device 700 can be etched. As shown, the photoresist layer 122 includes one or more apertures, so that some the III-N materials below the one or more apertures may be subsequently etched. Figure 7C shows the device 700 after the device 700 has been etched to leave recesses in one or more of the III-N layers (e.g., layer 108) below the apertures in the photoresist layer. For example, the recesses can be etched to the top of the spacer layer 106, as shown. As also seen in Figure 7C, the etch is performed under conditions that cause III-N material below the outer portions of the aperture to etch at a higher rate than III-N material below the inner portion of the aperture, thereby resulting in the desired profile for the recess.

[0044] Figure 7D shows the device 700 after the patterned resist 122 has been removed from the device 700. Figure 7E shows the device 700 after source and drain contacts 110 and 112 have been formed over the recesses. Figure 7F shows the device 700 after a gate 118 has been formed between the source and drain contacts 110 and 112, thereby forming a transistor.

[0045] Figure 8 shows a III-N semiconductor structure used in simulations to calculate band structures and electron carrier concentrations in regions below the source and drain electrodes of device 100. Figure 9A is a band diagram of the semiconductor structure along line 82 (i.e., in layers 106 and 104), and Figure 9B is a band diagram of the semiconductor structure along line 84 (i.e., in the remaining portion of layer 108 and in layers 106 and 104). In Figure 9A, line 92 is the conduction band energy Ec

(measured in eV, corresponding to the axis on the left), line 93 is the valence band energy Ev (measured in eV, corresponding to the axis on the left), and line 91 is the electron carrier concentration n _e (measured in cm ^"3 , corresponding to the axis on the right). In Figure 9B, line 95 is the conduction band energy Ec (measured in eV, corresponding to the axis on the left), line 96 is the valence band energy Ev (measured in eV, corresponding to the axis on the left), and line 94 is the electron carrier concentration n _e (measured in cm ^"3 , corresponding to the axis on the right).

[0046] The spike in electron carrier concentration 91 in Figure 9 A corresponds to the position of the 2DEG channel 116 beneath the ends of the contact which extend downwards and contact the A1N spacer layer (i.e., ends 202a-c and 206a-c in Figure 2), and the spike in electron carrier concentration 94 in Figure 9A corresponds to the position of the 2DEG channel 116 beneath the center portion of the contact (i.e., portions 204a-c in Figure 2). As seen in Figure 8, there is no AlGaN barrier layer material directly above the A1N spacer layer below the downward-extending ends of the contact (along line 82), while in the center portion of the contact (along line 84) there is approximately 7nm of AlGaN barrier layer material remaining over the A1N spacer layer. The simulations indicate that the 2DEG carrier concentration is substantially higher in the regions beneath the ends of the contact (i.e., the regions labeled 'High 2DEG concentration' in Figure 8) than in the region beneath the center portion of the contact (the 2DEG carrier

concentration in the regions beneath the ends of the contact was substantially higher than in any other portions of the 2DEG). This enhanced 2DEG carrier concentration beneath the downward extending ends of the contact is believed to be a reason why the recessed ohmic contacts described herein have a lower contact resistance than other types of ohmic contacts.

[0047] A number of implementations have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the techniques and devices described herein. For example, the processes described herein for forming alloyed ohmic contacts with reduced metal spitting can be used in the fabrication of other devices that require alloyed or annealed ohmic contacts, for example diodes, lasers, and LEDs. Accordingly, other implementations are within the scope of the following claims.