GB2432716A

GB2432716A - Growth of p-type nitride semiconductor structures

Info

Publication number: GB2432716A
Application number: GB0524016A
Authority: GB
Inventors: Stewart Edward Hooper
Original assignee: Sharp Corp
Current assignee: Sharp Corp
Priority date: 2005-11-25
Filing date: 2005-11-25
Publication date: 2007-05-30
Also published as: GB0524016D0

Abstract

A supply of p-type dopant material is interrupted at regular intervals during epitaxial growth to allow excess dopant material to evaporate. This prevents excessive accumulation of p-type dopant material which can segregate and lead to the formation of crystal defects. The supply of non-dopant material may be constant or periodic. The method is applicable to MBE and CVD growth methods using Mg dopant material.

Description

Growth of a p-type doped Nitride Semiconductor Structure This invention

relates to a method for the epitaxial growth of a nitride semiconductor structure such as, for example, a layer of a nitride semiconductor material (for example a layer of a member of the (Al,In,Ga) N material family) or a structure including layers of two or more different nitride semiconductor materials. It particularly relates to the growth of p-type doped nitride semiconductor material.

The (Al,Ga,ln)N material system includes materials having the general formula AlGaIni..N where 0 = x = I and 0 = y = 1. In this application, a member of the (Al,Ga,ln)N material system that has non- zero mole fractions of aluminium, gallium and indium will be referred to as AlGaInN, a member that has a zero aluminium mole fraction but that has non-zero mole fractions of gallium and indium will be referred to as InGaN, a member that has a zero indium mole fraction but that has non- zero mole fractions of gallium and aluminium will be referred to as AIGaN, and so on. There is currently considerable interest in fabricating semiconductor light-emitting devices in the (Al,Ga,In)N material system since devices fabricated in this system can emit light in the blue-violet wavelength range of the spectrum (corresponding to wavelengths in the range of approximately 380-450nm).

Semiconductor light-emitting devices fabricated in the (Al,Ga,ln)N materials system are described, for example, by S. Nakamura et al in "Jap. J. AppI. Phys." Vol. 35, ppL74-L76 (1996). They are also described in US-A-S 777 350, which teaches use of the metal-organic chemical vapour deposition (MOCVD) growth technique to fabricate light-emitting devices in the (Al,Ga,In)N materials system. MOCVD (also known as metal-organic vapour phase epitaxy or MOVPE) takes place in an apparatus which is commonly at atmospheric pressure but sometimes at a slightly reduced pressure of typically about 10 kPa. Ammonia and the species providing one or more Group Ill elements to be used in epitaxial growth are supplied substantially parallel to the surface of a substrate upon which epitaxial growth is to take place, thus forming a boundary layer adjacent to and flowing across the substrate surface. It is in this gaseous boundary M 2 layer that decomposition to form nitrogen and the other elements to be epitaxially deposited takes place so that the epitaxial growth is driven by gas phase equilibria.

In the growth of p-type GaN or AIGaN by MOCVD the p-type dopant source is typically bis(cyclopentadienyl)magnesium or bis(methylcylopentadienyl)magnesium; these are also known as Cp2Mg and MCp2Mg respectively.

Another known semiconductor growth technique is molecular beam epitaxy (MBE). In contrast to MOCVD, MBE is carried out in a high vacuum environment. In the case of MBE as applied to the (Al,ln,Ga)N system, a high or ultra-high vacuum (UHV) environment, typically around 1 x 1 0 Pa, is used. A nitrogen precursor is supplied to the MBE chamber by means of a supply conduit and species providing aluminium, gallium and/or indium, and possibly also a suitable dopant species, are supplied from appropriate sources within heated effusion cells fitted with controllable shutters to control the amounts of the species supplied into the MBE chamber during the epitaxial growth period. The shutter-control outlets from the effusion cells and the nitrogen supply conduit face the surface of the substrate upon which epitaxial growth is to take place. The nitrogen precursor and the species supplied from the effusion cells travel across the MBE chamber and reach the substrate where epitaxial growth takes place in a manner which is driven by the deposition kinetics.

At present, the majority of growth of high quality nitride semiconductor layers is carried out using the MOCVD process. The MOCVD process allows growth to occur at a V/Ill ratio well in excess of 1000:1. The V/Ill ratio is the molar ratio of the group V element to the Group III element during the growth process. A high V/Ill ratio is preferable during the growth of a nitride semiconductor material, since this allows a higher substrate temperature to be used which in turn leads to a higher quality semiconductor layer.

Many electronic devices and optoelectronic devices require at least one interface between an n-type doped material and a p-type doped material, in order to form a p:n junction and/or allow injection of electrical carriers into the device. GaN and AIGaN are both naturally an n-type doped semiconductor material, and p-type doped GaN or AIGaN is obtained by introducing a suitable dopant species during the GaN or AIGaN growth process. Magnesium is often used as a p-type dopant for GaN and AIGaN.

Many devices require a free carrier concentration in the p-type doped GaN or A1GaN of at least 1018 cm3, however, and there have been difficulties in obtaining such carrier concentrations in magnesium-doped GaN or AIGaN. It is relatively straightforward to incorporate magnesium atoms into GaN or AIGaN, but only a few per cent of magnesium dopant atoms are electrically activated and the un-activated dopant atoms do not give rise to free charge carriers.

European patent application No. 02738708.3 discloses a method of growing a p-type nitride semiconductor material by molecular beam epitaxy (MBE), in which magnesium is used as a p-type dopant. The beam equivalent pressure of magnesium during the growth process is controlled such that high temperature growth of the nitride semiconductor material is made possible, while at the same time providing increased concentration of activated magnesium atoms in the material. A free carrier concentration of up to 1 x l0'8cm3 was obtained.

UK patent application No. 0228019.6 discloses a method of growing a p-type nitride semiconductor material by molecular beam epitaxy, in which bis(cyclopentadienyl)magnesium (Cp2Mg) is used a source of magnesium during the growth process. The method can obtain a p-type free carrier concentration of up to around 2 x l0'7cm3.

A further problem encountered with growth of a p-type doped (Al,Ga,ln)N layer is that high doping levels of magnesium give rise to the formation of defects in the semiconductor layer. Applied Physics Letters Vol. 82, p3041 (2003) and Applied Physics Letters Vol. 81, p4748 (2002) report how a high level of magnesium doping in a semiconductor layer generates crystal defects in the semiconductor layer, and also leads to segregation of magnesium on the growth surface of the semiconductor layer. A model is proposed in the latter paper to describe this phenomenon. The model suggests that thin semiconductor layers (with a thickness up to approximately 5Onm) can be -4 heavily doped with magnesium without generating defects in the semiconductor layer, but that thicker layers will contain crystal defects arising from the magnesium doping.

Phys. Stat. Sol. Vol. 10, p2537 (2004) paper reports how thin layers of GaN (with a thickness of less than 2Onm) can be grown with heavy magnesium doping. As the layer thickness is increased, however, the magnesium doping level must be reduced in order to prevent the formation of defects.

A first aspect of the present invention provides a method of growing a p-type nitride semiconductor structure, the method comprising the steps of: a) supplying constituents, or precursors thereof, of a nitride semiconductor material and a p-type dopant, or a precursor thereof, to a growth chamber to grow a first thickness of the p-type nitride semiconductor material; b) interrupting the growth of the p-type nitride semiconductor material; and c) repeating steps (a) and (b) thereby to obtain a desired p-type nitride semiconductor structure.

In growth of a semiconductor material it is necessary to provide the constituents of the material, and any desired dopant, to the growth chamber -to grow AIGaN, as an example, it is necessary to provide the constituents Al, Ga and N and any desired dopant to the growth chamber. This may be done by supplying the constituents as elements; for example, in MBE growth beams of elemental Ga or elemental Al may be supplied direct to the growth chamber. Alternatively, the growth may be effected by supplying a precursors of a constituent to the growth chamber, whereby the precursor decomposes in the growth chamber to provide the constituent. As an example, in MOCVD growth trimethylgallium is used as a precursor of Ga -trimethylgallium is supplied to the growth chamber, where it decomposes to provide Ga. The term "precursor" of a constituent of the semiconductor material being grown (or of a dopant) as used herein means any chemical species that can be supplied to a growth chamber to provide the constituent (or the dopant).

The invention may be applied to growth of a layer of a p-type nitride semiconductor material, and the term "p-type nitride semiconductor structure" as used herein is accordingly intended to cover a layer of a single p-type nitride semiconductor material.

Alternatively, the invention may be applied to growth of a structure that includes layers of two or more different p-type nitride semiconductor materials.

Step (b) may comprise maintaining the supply of nitrogen or a nitrogen precursor to the growth chamber while stopping the supply of all other constituents of the semiconductor material and the p-type dopant, or the precursors thereof, to the growth chamber.

Since the supply of all constituents (or precursors thereof) of the semiconductor material apart from nitrogen (or a precursor thereof) is stopped in this embodiment, no material is grown in step (b). When the growth is restarted, the semiconductor material grown in each repetition of step (a) is the same as the semiconductor material grown in the initial occurrence of step (a) (that is, the constituents (or precursors thereof) supplied in each repetition of step (a) are the same as the constituents (or precursors thereof) supplied in initial occurrence of step (a). According to this embodiment of the invention, a p-type nitride semiconductor layer of a desired thickness is grown as a series of thin layers, one thin layer being grown in each repetition of step (a), with the growth process being interrupted after each thin layer has been grown. The growth interruptions allow any surface layer of dopant material, such as a surface layer of metallic magnesium in the case of a magnesium doped layer, that has accumulated as a result of segregation of the p-type dopant during the growth of the thin layer, to evaporate. This prevents the accumulation and incorporation of p-type dopant above a critical level which would degrade the crystal quality of the semiconductor layer and increase its resistance. As a result, the thin layer grown in each repetition of step (a) can be heavily doped p-type while still having good crystal quality and a low electrical resistance. The method of the invention thus allows a thick (>5Onm thick) heavily p-type doped nitride semiconductor layer with good crystal quality and a low electrical resistance to be grown, by using a repeated sequence of depositing a thin, highly p-type doped nitride semiconductor layer followed by a growth interruption.

Alternatively, step (b) may comprise stopping the supply of the p-type dopant, or the precursor thereof, to the growth chamber while maintaining the growth of a semiconductor material. This is done by maintaining the supply of the constituents of a semiconductor material, or the precursors thereof, to the growth chamber while the supply of the p-type dopant, or the precursor thereof, is stopped.

In this embodiment, the constituents, or the precursors thereof, supplied to the growth chamber may be unchanged when the supply of the p-type dopant, or the precursor thereof, is stopped. In this case, the semiconductor material grown while the supply of the p-type dopant, or the precursor thereof, is stopped will be the same material as the material grown in step (a) (except that, as described below, its overall doping level will be lower). This embodiment again allows a thick layer of a p-type nitride semiconductor material to be grown as a series of thin layers, and this again prevents the accumulation and incorporation of p-type dopant above a critical level which would degrade the crystal quality of the semiconductor layer and increase its resistance. In this embodiment, one thin layer of the material is grown in step (a), another thin layer is grown in step (b), the next thin layer is grown in the first repetition of step (a), and so on.

Alternatively in this embodiment, the constituents, or the precursors thereof, supplied to the growth chamber may be changed when the supply of the p-type dopant, or the precursor thereof, is stopped. In this case, the semiconductor material grown while the supply of the p-type dopant, or the precursor thereof, is stopped will be different from the semiconductor material grown in step (a). This allows a structure consisting of layers of two or more different nitride semiconductor materials to be grown. As an example, assume that aluminium, gallium and nitrogen, and a p-type dopant, (or precursors thereof) are supplied in step (a) so that a layer of p-type AIGaN is grown in step (a). If the supply of aluminium (or precursor thereof) is stopped at the same time as the supply of the p-type dopant (or precursor thereof) is stopped, the result is that a layer of GaN will be grown in step (b). The GaN will be doped p-type although, as explained below, its overall doping level will be lower than the doping level of the AIGaN. Provided that the supply of aluminium (or precursor thereof) is restarted when the supply of the p-type dopant (or precursor thereof) is restarted, a further layer of p-type AIGaN will be grown in the first repetition of step (a), and so. This allows an AIGaN/GaN structure to be grown. Other structures may be grown by supplying the appropriate constituents (or precursors thereof).

The duration of step (b) may be at least 5 seconds. It may be less than 300 seconds.

Step (a) may comprise growing the p-type nitride semiconductor material to a thickness of at least 3nm. It may comprise growing the p-type nitride semiconductor material to a thickness of less than 5Onm. It may comprise growing the p-type nitride semiconductor material to a thickness of approximately 2Onm or to a thickness of approximately 3Onm.

The p-type nitride semiconductor material may be p-type (Al,Ga,In)N.

The p-type dopant may be magnesium.

Step (a) may comprise growing the p-type nitride semiconductor material by molecular beam epitaxy. Alternatively, it may comprise growing the p-type nitride semiconductor material by metal-organic chemical vapour deposition.

A second aspect of the present invention provides a layer of p-type nitride semiconductor material grown by a method of the first aspect.

The semiconductor layer may have a thickness of at least 5Onm.

Preferred embodiments of the present invention will now be described by way of illustrative example with reference to the accompanying figures in which: Figure 1(a) illustrates the supply sequence of Group III materials and a dopant for a

prior art growth method;

Figure 1(b) shows supply sequences of group III materials and a dopant for a growth method of the present invention; Figure 2 shows current versus voltage characteristic curves for a device including a semiconductor layer of the invention and for a conventional device; and Figure 3 shows supply sequences of group III materials and a dopant for another growth method of the present invention.

The present invention will be described with reference to the MBE growth of a p-type doped (AlGaln)N semiconductor layer, in particular a magnesium-doped AIGaN layer.

The invention is not, however, limited to a MBE growth process, nor to growth of an AIGaN layer.

Initially in the method, a thin layer of p-type doped AIGaN is grown by a conventional MBE growth process. A suitable substrate is prepared and cleaned, and is introduced into the growth chamber of an MBE growth apparatus. The steps of cleaning and preparing the substrate may be conventional, and they will not be described further.

The p-type doped AIGaN layer is grown over the substrate by MBE. One or more intervening semiconductor layers may be grown over the substrate before the p-type doped AIGaN layer is grown.

To grow the p-type doped AIGaN layer, nitrogen or a nitrogen precursor is supplied to a MBE growth chamber by means of a supply conduit, and species providing aluminium and gallium are supplied to the growth chamber, for example from appropriate sources within heated effusion cells fitted with controllable shutters to control the amounts of the species supplied into the growth chamber during the epitaxial growth process. A species providing a p-dopant is also supplied to the grown chamber. Where the p-dopant is magnesium, a suitable species could be elemental magnesium, Cp2Mg or MCp2Mg. The layer of AIGaN is grown to a thickness below the critical thickness at which the magnesium will generate crystal defects and at which segregation of magnesium will occur so as to leave a metallic magnesium layer on the growth surface and, therefore, it is generally necessary to keep the thickness of the layer to below 5Onm, in accordance with the teaching of Applied Physics Letters Vol. 8!, p4748 (2002) (above). (It should be noted that the critical thickness at which crystal defects are generated is dependent on the dopant concentration, and decreases as the dopant concentration increases.) It is also preferable that the layer of AIGaN is grown to a thickness of at least 3nm. To achieve peak magnesium incorporation without generating defects in a layer with a thickness less than 3nm would require a very high magnesium flux. Also, the magnesium incorporation rate does not reach its peak until approximately 5-10 seconds after the onset of the magnesium flux, so if the layer is too thin it will not be heavily doped and control of the doping level will be difficult. The layer is preferably grown to a layer thickness of, for example, 20 or 30 nm.

Once the AIGaN layer has been grown to the chosen thickness, for example 2Onm or 3Onm, the growth process is interrupted. This may be done, for example, by completely stopping the supply of the species providing the Group Ill materials, and also stopping the supply of the species providing the magnesium dopant. It is, however, preferable for the supply of the species providing the Group V material to be continued during the interrupt period, to prevent decomposition of the upper surface of the semiconductor layer.

This embodiment of the invention is shown schematically in figure 1(b). In figure 1(b), trace (i) indicates the time-dependence of the supply of the dopant species, and trace (ii) indicates the time dependence of the species providing the Group III materials. The species providing the Group V material is preferably supplied continuously from time TI to time TI I. Thus, during the time period from TI to T2, the species providing the dopant and the species providing the Group III and Group V materials are supplied, and MBE growth of a p-type doped AIGaN layer occurs. The time period from T2 to T3 is an interrupt period, in which the supply of the dopant species and the species providing the Group Ill materials is stopped, so that no growth occurs between the time T2 and time T3.

The effect of interrupting the MBE growth at time T2 is that any metallic magnesium surface layer that has accumulated during the growth period from time Ti to time T2, as a result of segregation of magnesium, is able to evaporate. This prevents the accumulation of magnesium at the upper surface of the AIGaN layer, and thus prevents the incorporation of magnesium at a concentration which would generate crystal defects, degrade the crystal quality of AIGaN layer, and increase the electrical resistance of the AIGaN layer.

The duration of the interrupt period (T3 -T2) should be made sufficiently large to ensure that any metallic magnesium surface layer can completely evaporate during the interrupt period. It has been found that a suitable duration for the interrupt period is between 5 seconds and 300 seconds. If the duration of the interrupt period is below 5 seconds it is possible that the surface layer of metallic magnesium will not completely evaporate during the interrupt period. If the duration of the interrupt period is greater than 300 seconds, the total growth time required to grow a thick layer becomes very long. For a temperature of 950 C, an interrupt period with a duration of around 30 seconds has been found suitable, since this duration is long enough to ensure that the surface layer of metallic magnesium completely evaporates during the interrupt period but is not so long that the total growth time becomes inconveniently long.

The rate of evaporation of magnesium increases as the temperature increases. The higher the growth temperature, therefore, the shorter can be the interrupt period.

At time T3, the supply of the dopant species and the supply of the species providing the Group 111 materials are restarted. MBE growth occurs in the next growth period, which lasts until time T4. The duration of the second growth period (namely, T4 -T3) is again chosen so that the thickness of the AIGaN layer that is grown in the second growth period is below the critical level at which crystal defects are generated in the AIGaN layer as a result of accumulation of magnesium. That is, the duration of the second growth period is preferably chosen such that less that 5Onm thickness of AIGaN is grown, and is preferably chosen such that 20-3Onm thickness of A1GaN is grown in the second growth period. In general, each growth period may have the same duration, so that T2 -Ti = T4 -T3, but in principle the growth periods may have different durations from one another (provided that each growth period is sufficiently short such that the thickness of material grown in the growth period does not exceed the critical level at which crystal defects are generated in the A1GaN layer as a result of accumulation of magnesium).

At time T4 the supply of the dopant species and the supply of the species providing the Group III materials are again stopped, to provide a second interrupt period. The second interrupt period lasts until time T5, and allows any metallic magnesium that has accumulated at the upper surface of the semiconductor layer during the second growth period to evaporate.

The cycle of a growth period followed by an interrupt period is then repeated. The second interrupt period is followed by a third growth period from time T5 to T6, a third interrupt period from 16 to T7, a fourth growth period from time T7 to time T8, a fourth interrupt period from time T8 to T9, a fifth growth period from time T9 to T1O, a fifth interrupt period from time T1O to 111, and so on. The cycle of a growth period followed by an interrupt period shown in figure 1(b) may be repeated as many times as necessary to provide a magnesium-doped AIGaN layer having any desired thickness.

An interrupt period is provided after the final growth period, to ensure that any surface magnesium that accumulated during the final growth period can evaporate. In many cases, the growth process would be halted after the final growth period anyway, for example if the AIGaN layer was the final layer of a layer structure or to allow the growth temperature to be reduced before growth of another layer -in such cases, the cooldown period would effectively act as an interrupt period in which any surface magnesium could evaporate.

As an example, if the duration of the growth periods are chosen such that 2Onm thickness of AIGaN is grown in each growth period, then repeating the sequence of a growth period followed by an interrupt period ten times would produce a 200 nm thick, continuous AIGaN layer that is doped p-type. Although examination of the final material, for example using examination techniques such as Transmission Electron Microscopy (TEM) or Secondary Ion Mass Spectroscopy (SIMS), could determine the thickness of material that had been grown in each growth period, the A1GaN layer would behave as if it had been grown in a single growth step.

Once the p-type doped AIGaN layer has been grown to its desired thickness, further semiconductor layers may be grown over the p-type doped AIGaN layer if desired. In the example of figure 1(b), a further semiconductor layer could be grown over the p-type doped AIGaN layer after time Ti 1.

Figure 1(a) shows, as a contrast, the time-dependence of the supply of a dopant species (trace (i)) and the species for providing Group III materials (trace (ii)). As can be seen, in a conventional growth method the species providing the Group III materials and the dopant species are supplied continuously, so that a thick AIGaN layer is grown in a single growth step. As a result, any excess magnesium introduced into the AIGaN layer is liable to cause crystal defects, and also give rise to a layer of metallic magnesium at the upper surface of the layer structure. This places a limit on the maximum concentration of free-charge carriers that can be obtained.

The maximum free carrier concentration that can be obtained by a method of the invention depends on the aluminium composition of the layer that is grown, since the maximum free carrier concentration decreases with aluminium content. It is believed that a method of the invention could obtain a free carrier concentration of 5x10'8cm3 or greater for a GaN layer. For a layer with a thickness that exceeds the critical thickness, a growth method of the invention should give a higher maximum achievable carrier concentration than a conventional method.

Figure 2 illustrates the effect of the present invention. Figure 2 shows the current-voltage characteristic curves for two light-emitting diodes fabricated in the (AIGaN)N materials system. The device structure includes an upper cladding layer which is constituted by a 500nm thick layer of p-type doped A1GaN. Trace (a) in figure 2 shows the current-voltage characteristic for an LED in which the 500nm thick ptyped dope AIGaN layer was grown by a method of the present invention, whereas trace (b) of figure 2 shows the current voltage characteristic for a LED in which the 500nm thick p-type doped AIGaN layer was grown by a conventional, continuous growth process such as that shown in figure 1 (a). Apart from the different growth process used to grow the 500nm thick p-type AIGaN layer, the two LEDs were identical to one another in both their structure and their fabrication process.

It can be seen that, for a given applied voltage across the LED, trace (a) in figure 2 has a higher current than trace (b). This indicates that the method of the present invention has reduced the electrical resistance of the 500nm thick p-type doped AIGaN layer of the LED structure, and has thereby reduced the overall resistance of the LED structure. The reduction in resistance of the 500nm thick p-typed doped A1GaN layer arises from the better crystal quality obtained by a growth process of the present invention.

In the embodiment of figure 1(b), the growth process is stopped completely during each interrupt period, by stopping the supply of the doping species and the species for providing the Group Ill material at the start of each interrupt period. In an alternative embodiment of the invention, the growth process is continued during an interrupt period, and only the supply of the dopant species is turned off during an interrupt period. This embodiment is shown schematically in figure 3, which shows the time dependence of the supply of the dopant species (trace (i)) and the time dependence of the supply of species providing the Group 111materials (trace (ii)). As can be seen in figure 3, the rate of supply of the species providing the Group III materials is maintained constant from time Tl to time TI I. An interrupt period is defined by stopping the supply of the dopant species.

In the embodiment of figure 3, the growth of the AIGaN layer continues during each interrupt period, with nominally undoped material being grown since the supply of the dopant species is stopped during an interrupt period. The nominally undoped material which is grown in an interrupt period can incorporate the excess magnesium supplied during the previous growth period, and thereby prevents excess magnesium supplied during a growth period from giving rise to crystal defects or magnesium segregation.

The doping profile of material grown in the interrupt period will be graded, with material grown at the start of the interrupt period having the highest dopant concentration and with material grown at the end of the interrupt period having the lowest dopant concentration. Overall, the dopant concentration of material grown in the interrupt period is likely to be lower than the dopant concentration of material grown in the growth period.

In a modification of the embodiment of figure 3, the constituents, or the precursors thereof, supplied to the growth chamber are changed when the supply of the p-type dopant, or the precursor thereof, is stopped at time 12. In this case, the semiconductor material grown between time T2 and time T3 while the supply of the p-type dopant, or the precursor thereof, is stopped will be different from the semiconductor material grown from time Ti to time T2. This allows growth of a structure consisting of layers of two or more different nitride semiconductor materials, in particular growth of a layer structure in which the composition of the layers alternates -for example an AIGaN-GaN-AIGaN-GaN etc. structure -and in which the layers of different composition require different doping levels.

As an example, assume that a layer of p-type AIGaN is grown in the period from time Ti to time T2. If the supply of aluminium (or precursor thereof) to the growth chamber, is stopped at time T2 when the supply of the p-type dopant (or precursor thereof) is stopped, the result is that a layer of GaN will be grown in the period from time T2 to time T3. The GaN will be doped p-type as a result of incorporation of any excess magnesium supplied between time Tl and time T2 although, as explained above, its overall doping level will be lower than the doping level of the AIGaN layer grown between time Ti and time T2. Provided that the supply of aluminium (or precursor thereof) is restarted at time T3 when the supply of the p-type dopant (or precursor thereof) is restarted, a further layer of p-type AIGaN will be grown from time T3 to time T4, and so on, so that an AIGaN/GaN structure is grown.

Where the method of figure 3 is used to grow a semiconductor structure including layers of two different materials arranged in an alternating manner, it possible in principle for either one of the materials to be grown in the interrupt periods of T2 to T3, T4 to T5 etc. En practice, however, it may be preferable for a particular one of the materials to be grown in the interrupt periods. For example, in the case of growth of an AIGaN/GaN structure it may be preferable to grow GaN in the interrupt periods, since GaN is easier to dope p-type than A1GaN and requires lower magnesium flux than AIGaN.

The embodiment of figure 1(b) is preferred for the growth of a thick layer that is desired to have a uniform doping concentration over its thickness, since it can provide a more uniform doping profile than the method of figure 3.

If the method of figure 3 is applied to the growth of a layer of a single semiconductor material the overall duration of the growth process is lower than if the method of figure 1(b) is used.

In the method of figure 3, the supply of the species supplying nitrogen is maintained continuously during the growth process.

In the case where the invention is applied to MBE growth, examples of suitable species for providing nitrogen are atomic nitrogen and ammonia gas (NH3).

Although the invention has been described above with reference to MBE growth the invention is not limited to an MBE growth process. The invention may be effected using other growth processes such as, for example, an MOCVD growth process. As with embodiments which use MBE growth, where the method of growth is effected using MOCVD or another growth process, interrupt periods are defined in the growth process. An interrupt period may be defined by stopping supply of the species that provide the dopant and the group III materials, or an interrupt period may be defined by stopping only supply of the dopant species. As in the case of MBE growth, the supply of nitrogen is preferably maintained through an interrupt period in MOCVD growth The present invention has been described above with reference to the growth of p-type doped AIGaN. The invention is not, however, limited to growth of this particular material but may be applied to growth of other nitride semiconductor materials. In

I

particular, the invention may be applied to growth of other members of the (AlGaIn)N materials systems such as, for example, GaN, InN, InGaN and AIGaTnN. In all cases, interrupt periods are defined in the growth process by stopping the supply of all Group III species and the dopant species or by stopping only the supply of the dopant species.

The present invention may be applied in the growth of a semiconductor electronic or opto-electronic device (such as a light-emitting device) in a nitride materials system. It may be used to grow a thick p-type nitride semiconductor layer in a device structure - for example, it may be used to grow a cladding layer in a device structure, where the cladding layer is formed of a thick p-type doped nitride semiconductor layer (for example a thick p-doped AIGaN layer), in a laser diode structure or an LED structure.

The invention may be used to provide a cladding layer that has a high doping concentration but a low electrical resistance, so that the overall resistance of the device is kept low thereby reducing the threshold current of the device. The advantages of providing a low-resistance p-type doped nitride semiconductor layer are not limited to light-emitting devices, but apply to all devices fabricated in a nitride material system which require a thick p-type doped nitride semiconductor layer. As an example, the invention may also be applied to the manufacture of a photodetector or an A1GaN Heterojunction Bipolar Transistor (HBT) which also contains a relatively thick p-doped AIGaN layer.

It should be noted that, where the invention is applied to manufacture of an opto-electronic or electronic device, the lower electrical resistance of the p-type nitride semiconductor layer that can be obtained by the invention also has the effect of reducing the amount of heat generated in the device during its operation. This increases the lifetime of the device. /-

Claims

CLAIMS: 1. A method of growing a p-type nitride semiconductor

structure, the method comprising the steps of: a) supplying constituents, or precursors thereof, of a nitride semiconductor material and a p-type dopant, or a precursor thereof, to a growth chamber to grow a first thickness of the p-type nitride semiconductor material; b) interrupting the growth of the p-type nitride semiconductor material; and c) repeating steps (a) and (b) thereby to obtain a desired p-type nitride semiconductor structure.

2. A method as claimed in claim 1 wherein step (b) comprises maintaining the supply of nitrogen or a nitrogen precursor to the growth chamber while stopping the supply of all other constituents of the semiconductor material and the p-type dopant, or the precursors thereof, to the growth chamber.

3. A method as claimed in claim I wherein step (b) comprises stopping the supply of the p-type dopant, or the precursor thereof, to the growth chamber while maintaining growth of a semiconductor material.

4. A method as claimed in claim 1, 2 or 3 wherein the duration of step (b) is at least seconds.

5. A method as claimed in any preceding claim wherein the duration of step (b) is less than 300 seconds.

6. A method as claimed in any preceding claim wherein step (a) comprises growing the p-type nitride semiconductor material to a thickness of at least 3nm.

7. A method as claimed in any preceding claim wherein step (a) comprises growing the p-type nitride semiconductor material to a thickness of less than 5Onm.

8. A method as claimed in claim 7 wherein step (a) comprises growing the p-type nitride semiconductor material to a thickness of approximately 2Onm.

9. A method as claimed in claim 7 wherein step (a) comprises growing the p-type nitride semiconductor material to a thickness of approximately 3Onm.

10. A method as claimed in any preceding claim wherein the p-type nitride semiconductor material is p-type (Al,Ga,In)N.

11. A method as claimed in any preceding claim wherein the p-type dopant is magnesium.

12. A method as claimed in any preceding claim wherein step (a) comprises growing the p-type nitride semiconductor material by molecular beam epitaxy.

13. A method as claimed in any of claims I to 10 wherein step (a) comprises growing the p-type nitride semiconductor material by metal-organic chemical vapour deposition.

14. A layer of a p-type nitride semiconductor material grown by a method as defined in any of claims I to 13.

15. A layer of p-type nitride semiconductor material as claimed in claim 14 and having a thickness of at least 5Onm.