GB2561811A

GB2561811A - Optical device

Info

Publication number: GB2561811A
Application number: GB1701240.2A
Authority: GB
Inventors: Zilkie Aaron; Yannick Gardes Frederic; Mastronardi Lorenzo
Original assignee: University of Southampton; Rockley Photonics Ltd
Current assignee: University of Southampton; Rockley Photonics Ltd
Priority date: 2017-01-25
Filing date: 2017-01-25
Publication date: 2018-10-31
Also published as: GB201701240D0

Abstract

A method of producing an optical device includes growing an optically active region (OAR) 4 (e.g. comprising e.g. Ge, SiGe, GeSn or SiGeSn) and providing a capping layer on top of the OAR. A first resist (e.g. photoresist 6a, figure 1) is disposed along a part of the capping layer above the OAR. Part of the capping layer not covered by the resist is etched to create a raised portion 8a of the capping layer below the resist, the raised portion having a width less than the OAR, a first part of the OAR extending from a first side of the raised portion and away therefrom, a second part of the OAR extending from the second side of the raised portion and away therefrom, and a third part of the OAR, disposed between the first and second parts of the OAR and beneath the raised portion of the capping layer. The first resist is then removed, followed by the step (a) of disposing a second resist over the first part of the OAR, leaving the second part of the OAR exposed and implanting a first dopant in the second part of the OAR to produce a first doped region 14, removing the second resist; and/or the step (b) of disposing a third resist over the second part of the OAR, leaving the first part of the OAR exposed, and implanting a second dopant in the first part of the OAR to produce a second doped region 16. The third resist is subsequently removed. The first and second parts of the optically active material are etched such that the height of the third part is greater than the heights of the first part and second part. The optical device is annealed so that the implanted dopants diffuse (see arrows) into at least a part of the third part of the OAR, thereby producing at least one diffusively doped region. The method produces an optoelectronic device which is operable as an Electro-absorption modulator (EAM) or a photodiode.

Description

(56) Documents Cited:

GB 2543122 A

H01L 21/225 (2006.01)

GB 2477131 A (71) Applicant(s):

Rockley Photonics Limited

Cooley (UK) LLP, 10th Floor Dashwood,

Old Broad Street, London, Greater London, EC2 1QS, United Kingdom (58) Field of Search:

INT CL G02F, H01L Other: WPI,EPODOC

University of Southampton (Incorporated in the United Kingdom) Highfield, SOUTHAMPTON, SO17 1BJ, United Kingdom (72) Inventor(s):

Aaron Zilkie

Frederic Yannick Gardes Lorenzo Mastronardi (74) Agent and/or Address for Service:

Mewburn Ellis LLP

City Tower, 40 Basinghall Street, LONDON, Greater London, EC2V 5DE, United Kingdom (54) Title of the Invention: Optical device

Abstract Title: Optical device and method of production thereof (57) A method of producing an optical device includes growing an optically active region (OAR) 4 (e.g. comprising e.g. Ge, SiGe, GeSn or SiGeSn) and providing a capping layer on top of the OAR. A first resist (e.g. photoresist 6a, figure 1) is disposed along a part of the capping layer above the OAR. Part of the capping layer not covered by the resist is etched to create a raised portion 8a of the capping layer below the resist, the raised portion having a width less than the OAR, a first part of the OAR extending from a first side of the raised portion and away therefrom, a second part of the OAR extending from the second side of the raised portion and away therefrom, and a third part of the OAR, disposed between the first and second parts of the OAR and beneath the raised portion of the capping layer. The first resist is then removed, followed by the step (a) of disposing a second resist over the first part of the OAR, leaving the second part of the OAR exposed and implanting a first dopant in the second part of the OAR to produce a first doped region 14, removing the second resist; and/or the step (b) of disposing a third resist over the second part of the OAR, leaving the first part of the OAR exposed, and implanting a second dopant in the first part of the OAR to produce a second doped region 16. The third resist is subsequently removed. The first and second parts of the optically active material are etched such that the height of the third part is greater than the heights of the first part and second part. The optical device is annealed so that the implanted dopants diffuse (see arrows) into at least a part of the third part of the OAR, thereby producing at least one diffusively doped region. The method produces an optoelectronic device which is operable as an Electro-absorption modulator (EAM) or a photodiode.

Fig. 7

1/12

Fig. 1

2/12

Fig. 2

3/12

Fig. 3

4/12

Fig. 4

5/12

Fig. 5a

6/12

Fig. 5b co

7/12

Fig. 6 co cr

8/12

Fig. 7

9/12

Fig. 8

ΙΌ

Ο

10/12

	<·<<<.<< <<?<?<?<?.</<? <?<?<.·<.<?<.Γ·Λ·<·Λ·ί·Λ· ίϊΛΐ/ί^ίλϊίϊΛϊίϊΛΐίΐΛ-ίνίΛϊ^ΛϊίϊΛΐΛ-'Λ-ϊΛΐΛΐΛΪΛϊΛ’ a··.··.··.··.··,··.·· a··,··.··-··.··-·· a··.··.··.··,·· .-	i	k H
-F LU	I ‘ Hw T	k			jlillilil ΙΟΥ «fiis
	WO		Β\\·\^^	hi	§Ww^w NsvNSN.fiX’^'W^Ss'sV'

Fig. 9

11/12

²³⁰⁹ 2304

Fig. 10

12/12

Fig. 11

OPTICAL DEVICE

Field of the Invention

The present invention relates to an optical device including a ridged waveguide.

Background of the Invention

Silicon photonics technology is a very fast moving field, which has the potential to substantially improve board-to-board chip-to-chip, intra-datacentre, and inter-datacentre optical interconnects. However, a number of issues related to conventional optical interconnect technology should be solved before widespread adoption is likely. These include: power dissipation, physical size, and cost. As information processing systems are increasingly constrained by power usage, optical interconnects are unlikely to be implemented if they require more power than their electrical counterparts. Photonic interconnect solutions are advantageous over electrical interconnects because they can have lower energies per bit and higher power efficiency at higher data rates.

Optical interconnects may also alleviate the issues associated with the shrinkage of integrated circuits (i.e. the increase in signal delay) by provision of wavelength division multiplexing. By using wavelength division multiplexing, multiple channels are combined to travel in a single optical link, therefore optical interconnect cost and fiber count can be significantly reduced.

Compact silicon photonic elements that can be integrated with a high density are needed to enable pitch-matching of optical interconnects to high-bandwidth, high-density ASICs for minimization of electrical connection length. An issue with electrical interconnects relates to the fact that scaling compute capacity requires increasing the number in microprocessor cores. As the number of cores increases, so does the level of interconnection required. Increasing the number and length of electrical interconnects between an increasing number of cores means increasing the latency, power consumption, parasitic capacitance, and heat load. Therefore, it is desirable to provide micro-scale optical devices (devices with < 1 mm^A2 area and < 250 micron pitch) which enable high density of devices, low capacitance, and low power usage so as to provide a viable solution to the bottleneck that exists in current electronic technology. It is inevitable that silicon photonics will be utilised in systems where the number of microprocessor cores increases beyond the capabilities of the existing on-chip interconnect technology.

The challenge faced by the silicon photonics industry is to fabricate micro-scale electrooptical components (namely, modulators) that meet the above specified energy and integration density requirements, whilst also offering a high bandwidth density by exploiting WDM capabilities. An example of a good practical solution for micro-scale silicon photonics modulators is electro absorption modulators integrated in silicon ridge waveguides that have SiGe as an optically active material in the ridge waveguide.

Summary of the Invention

Thus, at its most general, the invention provides a method of controlling the level of sidewall doping in a ridged waveguide by diffusion of dopants from the slab regions.

Accordingly, the present invention aims to solve the above problems by providing, in a first aspect, a method of producing an optical device having the steps of: growing an optically active region providing a capping layer on top of the optically active region; disposing a first resist along a part of the capping layer above the optically active region; etching a part of the capping layer not covered by the resist to create: a raised portion of the capping layer below the resist, the raised portion having a width (W) less than the optically active region; a first part of the optically active region extending from a first side of the raised portion away from the raised portion; a second part of the optically active region extending from the second side of the raised portion away from the raised portion; and a third part of the optically active region, disposed between the first and second parts of the optically active region and beneath the raised portion of the capping layer; removing the first resist; performing a step (a) of: disposing a second resist over the first part of the optically active region, leaving the second part of the optically active region exposed; and implanting a first species of dopant in the second part of the optically active region, and subsequently removing the second resist; and/or performing a step (b) of: disposing a third resist over the second part of the optically active region, leaving the first part of the optically active region exposed; and implanting a second species of dopant in the first part of the optically active region, and removing the third resist; etching the first and second parts of the optically active material such that the height of the third part is greater than the respective heights of the first and second part; and annealing the optical device so that the implanted dopants diffuse into at least a part of the third part of the optically active region which is below the raised capping layer, thereby producing at least one diffusively doped region.

Advantageously, the implantation stages are self-aligned using the raised portion of the capping layer (e.g. as a hard mask). Therefore, the implantation of the dopants can be performed at any angle.

The modulators or other silicon photonics modulators address the challenges set out above, at least in part by having doped regions in the sidewalls of the ridge waveguide that can be controlled in manufacturing with practical process steps to enable good performance with good device yields. This can enable these modulators to become a mainstay of all silicon photonic optical interconnect devices, addressing a huge market need.

In a second aspect, the invention provides an optical device comprising: an optically active region; a raised capping layer disposed on top of the optically active region, the raised capping layer having a width less than the optically active region; and a ridged waveguide comprising: a first part of the optically active region which extends from a first side of the raised portion away from the raised portion; a second part of the optically active region which extends from a second side of the raised portion away from the raised portion; a third part of the optically active region, disposed between the first and second parts of the optically active region and beneath the raised capping layer; wherein the third part of the optically active region has a height which is greater than respective heights of the first and second parts of the optically active region, thereby providing sidewalls of the ridged waveguide; a first doped region, within the first part of the optically active region; a second doped region, within the second part of the optically active region; and at least one diffusively doped region located on a respective sidewall of the ridged waveguide, the diffusively doped region extending beneath or only part of the way up the height of its respective sidewall. By height, it may be meant that the third part of the optically active region extends further from a substrate than the first and second parts of the optically active region.

In some embodiments, each diffusively doped region is located at a respective sidewall of the ridges waveguide and extends only a part of the way up the height of the respective sidewall.

Optional features of the invention will now be set out. These are applicable singly or in any combination with any aspect of the invention.

When only step (a) has been performed, the method may include, after the step of annealing the optical device, a step of: performing an angled implant of dopants into the first part of the optically active region.

When only step (b) has been performed, the method may include, after the step of annealing the optical device, a step of: performing an angled implant of dopants into the second part of the optically active region.

Step (a) and step (b) may both be performed.

The device may include two diffusively doped regions located on opposing sidewalls of the ridged waveguide, wherein at least one of the diffusively doped regions extends beneath or only part of the way up the height of its respective sidewall.

The annealing step may comprise multiple sub-annealing steps. For example, the first species of dopant may have a higher diffusion temperature than the second species of dopant (or vice versa). Therefore, by annealing first at a lower temperature, and then at a higher temperature, varying degrees of diffusion can be obtained.

Alternatively, the first species of dopant may be implanted and annealed before the second species of dopant is implanted. In this example, the first species of dopant may be chosen to have a higher diffusion temperature than the second species of dopant. As such, the second species of dopant can be annealed without affecting the diffusion of the first species of dopant.

The method may further include: disposing a fourth resist over: a portion of the first part of the optically active region and the second part of the optically active region; and the raised portion of the capping layer, and further implanting dopants of the first species in the exposed portion of the first part of the optically active region thereby providing a heavily doped region. A subsequent step includes removing the fourth resist and disposing a fifth resist over a portion of the second part of the optically active region and the first part of the optically active region; and the raised portion of the capping layer, further implanting dopants of the second species in the exposed portion of the second part of the optically active region thereby providing a heavily doped region; and removing the fifth resist. Therefore, the first and second doped regions may further comprise respective heavily doped regions. These process steps may be performed before or after the device has been annealed. In some embodiments, the process steps are performed after annealing the device. These process steps may be performed after the angled implant of dopants described above.

The portion of the first part and the portion of the second part may be adjacent to the raised portion of the capping layer. Therefore, the respective heavily doped regions may be distal from the third part of the optically active region. This can allow the heavily doped regions to be distal from the optical mode.

The amount of first and second species dopant may be controlled such that, when the optical device is annealed, the dopants diffuse an equal distance into the third part of the optically active region.

The optically active region may be grown within a channel or cavity of a silicon on insulator (SOI) chip. The sidewalls of the channel and/or a bottom layer of the channel may be lined with an insulating layer before the optically active material is grown. The insulating layer may be, for example, a silicon dioxide layer. The insulating layer can promote a more uniform growth to the optically active material.

Alternatively, the optically active region can be formed directly on oxide layer, this oxide layer being either formed at the bottom of a channel or cavity, or being a buried oxide layer in an SOI. It is also possible to form a doped region partially below the raised capping layer or waveguide ridge using the method described herein, thereby forming a diode with a vertical, diagonal, or other non-horizontal electric field.

Optionally, the optically active region is formed from Ge, SiGe, GeSn or SiGeSn.

The SOI chip may be provided on top of a buried oxide layer.

The first species of dopant may be a p-type dopant and the second species of dopant may be an n-type dopant. Alternatively, the first species of dopant may be an n-type dopant and the second species of dopant may be a p-type dopant. When there is an intrinsic semiconductor between the first and second doped regions, e.g. the third part of the optically active region is formed of an intrinsic semiconductor, the resulting structure will form a PIN junction. The PIN junction may operate as a modulator (involving application of a forward bias across the junction, although it should be noted that this would only work in an injection mode) or a photodetector (involving application of a reverse bias across the junction). A PN junction could be created by further diffusing the dopants such that they meet at a PN interface within the waveguide ridge.

When the capping layer is etched, a protective layer of capping material can be left such that the optically active region is protected, e.g. from oxidisation, during subsequent steps. The protective layer should be very thin so as to not interfere with the implantation of dopants. Therefore, the exposed regions are exposed in so far as they may be doped by implantation. In such cases the dopants are implanted through the protective layer into the optically active region. The protective layer may be between 50 nm and 500 nm in thickness or between 20nm and 50 nm.

Alternatively, if the optically active region is formed of Ge or SiGe, the processing may take place in an oxygen-free atmosphere such that oxidisation does not readily occur.

The optical device may be an electro-absorption modulator (EAM) which operates using the Franz-Keldysh effect. Electrical contacts may be provided connected to the doped or heavily doped regions, such that an E-field can be provided across the third optically active region which may be made from a bulk semiconductor material. The electric field lines may be horizontally or substantially horizontally orientated i.e. along the plane of the SOI layer.

The optical device may be connected at one end to an input waveguide, and at another end to an output waveguide. As such, light passing from the input waveguide through the device and into the output waveguide may be modulated or detected depending on the optical device’s function.

The method may include the step of further etching the first and second parts of the optically active material to create a waveguide ridge, the diffusively doped regions each being located at a respective sidewall of the ridge but extending only part of the way up the length of the respective sidewalls. As will be understood by a person skilled in the art, the etching step to produce a ridged waveguide may be performed before the implantation of dopants, as can etching the first and second parts of the optically active region such that the height of the third part is greater than the respective heights of the first part and second part. However etching the first and second parts of the optically active region such that the height of the third part is greater than the respective heights of the first part and second part must be done before the annealing step. By height it is meant the distance from the SOI chip to the furthest surface of the relevant optically active region.

The optically active region may be grown on a silicon on insulator chip which is provided on top of a buried oxide layer. Alternatively, the optically active region may be grown directly on the buried oxide layer with no intervening silicon on insulator chip. For example, a Rapid Melting Growth (RMG) method may be used.

Further optional features of the invention are set out below.

Brief Description of the Drawings

Embodiments of the invention will now be described by way of example with reference to the accompanying drawings in which:

Figure 1 is a schematic diagram showing a manufacturing step;

Figure 2 is a schematic diagram showing another manufacturing step;

Figure 3 is a schematic diagram showing the doping of an optically active region;

Figure 4 is another schematic diagram showing doping of another optically active region;

Figure 5a is another schematic diagram showing doping of optically active regions;

Figure 5b is another schematic diagram showing doping of optically active regions

Figure 6 is a schematic diagram showing another manufacturing step;

Figure 7 is a schematic diagram showing the results of annealing the device;

Figure 8 is a schematic diagram showing an optical device as produced by including optional manufacturing steps;

Figure 9 is an enlargement of Figure 8;

Figure 10 is an example of a device produced by the method disclosed herein; and

Figure 11 is a further example of device produced by the method disclosed herein

Detailed Description and Further Optional Features of the Invention

Figure 1 illustrates a starting point in a manufacturing process resulting in an optical device and is a cross-section through a semiconductor device. A buried oxide (BOX) layer 12 has disposed immediately on top a silicon-on-insulator (SOI) layer 2. Previous manufacturing steps in this example may include etching a part of the SOI layer 2 to provide a channel, and lining 10 the sidewalls 2a, 2b of the channel with silicon dioxide. An optically active region (4), in this case formed from SiGe, is epitaxially grown in the channel of the SOI layer.

Next, a capping layer 8 is disposed over the optically active region and (optionally) the sidewalls 2a, 2b of the channel. The capping layer is preferably an oxide disposed by plasma-enhanced chemical vapour deposition, however any hard mask can be provided. A resist (6a) is then disposed over a part of the capping layer above the optically active region e.g. above the channel of the SOI layer 2. The resist has a width across the channel which is less than the optically active region. Preferably, the resist is positioned centrally with respect to a width of the channel i.e. equidistant from the sidewalls 2a, 2b. The resist may be, for example, an S1813 photoresist or a ZEP e-beam resist. The resist 6a can then be provided by patterning, for example by UV or e-beam lithography. The resists may be removed by an ashing process.

Next, the device is etched to remove most (if not all) of the capping layer not directly underneath the resist. This can be performed by, for example, inductively coupled plasma (ICP) or reactive-ion etching (RIE) etching. The resulting device is shown in Figure 2. The capping layer directly beneath the resist remains as a raised capping layer 8a, and has a width which is less than the width of the optically active region 4. In this example, a thin protective layer of capping material 8b is retained. The protective layer may have a thickness in the range of 20 nm - 50 nm. The thin protective layer of capping material provides protection for the SiGe layer from oxidisation. As discussed above, the thin layer should not interfere with the subsequent implantation of dopants.

At this stage, it is useful to discuss the parts of the optically active region which are present. A first part 4a of the optically active region, can be considered to comprise a portion of the optically active region which is not beneath the raised capping layer 8a, i.e. is not between the raised capping layer and the SOI chip 21 BOX layer 12, and is on one side thereof i.e. it is displaced horizontally (along the plane of the BOX layer) in a first direction. Similarly, a second part 4b of the optically active region can be considered to comprise a portion of the optically active region which is not beneath the raised capping layer and is on another side thereof i.e. it is displaced horizontally (along the plane of the BOX layer) in a second direction). A third part 4c of the optically active region can be considered to comprise a portion of the optically active region which is beneath the raised capping layer.

After etching the capping layer, a second resist 6b is then applied to the device as shown in Figure 3. The resist masks all of the optically active region on one side of the raised capping layer (e.g. the first part of the optically active region), and does not mask the optically active region on the other side of the raised capping layer (e.g. the second part of the optically active region). As a result, the implantation of dopants (indicated here by the downwardspointing arrows) affects only the unmasked part of the optically active region. The result is a part 14 of the optically active region which is doped by the species of dopant chosen. In some examples, this dopant is a p-type dopant e.g. Boron and may have a dopant concentration of between 5 χ 10¹⁶ - 5 χ 10¹⁹ cm·³. As the raised capping layer 8 is present, the implantation is self-aligned on this side with the capping layer. This allows the implantation to occur at substantially any angle.

The second resist 6b is then removed and, as shown in Figure 4, a third resist 6c is applied. This resist masks the portion of the optically active region that has just been doped, but does not mask the (undoped) optically active region on the opposing side ofthe raised capping layer. Therefore, the implantation of dopants (again indicate by the downwards pointing arrows) affects only the unmasked portion ofthe optically active region, and a portion 16 of the optically active region is doped. In some examples the dopant chosen is an n-type dopant e.g. phosphorus or arsenic and may have a dopant concentration of between 5 χ 10¹⁶-5 χ 10¹⁸cnr³.

An optional step is shown in Figure 5a. Here, a fourth resist 6d is applied over a part of the previous doped region 14 and further dopants are implanted in these regions. This results in heavily doped regions 18 adjacent to the previously doped regions and distal from the raised capping layer 8a. This optional step can be performed before or after the annealing step described below. A further optional step is shown in Figure 5b. Here, a fifth resist 6e is applied over a part ofthe previously doped region 16, and further dopants are implanted in the exposed region. This results in a heavily doped region 20, adjacent to the previously doped region and distal from the raised capping layer 8a.

The heavily doped regions may have a dopant concentration of 10¹⁹ - 10²⁰ cm’³. When electrical contacts are provided, they may be substantially adjacent to the heavily doped regions.

Either before or after the implantation of dopants, the first and second parts 4a, 4b of the optically active region must be initially etched so that they are thinner than the third part 4c of the optically active region e.g. the third part has a greater height from the SOI chip than the first and second parts. By thinner, it is meant that they do not extend from the floor of the channel as far as the third region i.e. their height is reduced in comparison thereto.

After these steps (including the optional steps shown in Figures 5a and 5B, and the etching described in the preceding paragraph) a device as shown in Fig. 6 is provided. A SiGe optically active region 4 is epitaxially grown in the channel of a SOI layer 2. The optically active region comprises at least a first doped region 14 and a second doped region 16, and may optionally include a first heavily doped region 18 and a second heavily doped region 20. On top of the optically active region is a raised capping layer 8a, this capping layer having a width which is less than the optically active region. An optional protective layer 8b is provided on top of the doped and heavily doped regions.

In this example, the device is first annealed as shown in Figure 7. The annealing is generally a flash-annealing, e.g. one performed at around 650°C for between 1 and 20 seconds. The result of this annealing is the diffusion of dopants from the doped regions 14 and 16 from their respective regions into at least the optically active region as indicated by the black horizontal arrows.

The annealing can either be done as one step or two steps, depending on whether the regions have both been doped beforehand or if one is doped before the annealing and one after. It is possible to implant the first dopant and anneal it before implanting and annealing the second dopant. To achieve this, a first dopant should be chosen with a diffusion temperature (i.e. a temperature at which the dopants will diffuse) which is greater than the diffusion temperature ofthe second dopant. In doing so, the second dopant can be diffused without causing further diffusion of the first dopant. This can be particularly helpful if the first and second dopants diffuse at different rates, as it enables the dopants to be diffused a similar or substantially equal distance into the optically active region.

After annealing, the device may be etched so as to produce a proud ridged waveguide 30 as shown in Figure 8. The ridged waveguide is formed of three parts of the optically active region 4: the first part 4a of the optically active region which is disposed on one side of the raised capping layer 8a; the second part 4b ofthe optically active region which is disposed on an opposing side of the raised capping layer; and the third part 4c of the optically active region which is disposed between the first and second parts and is below the raised capping layer. After the etching, the third part of the optically active region has a greater height, i.e.

extends further from the SOI chip/BOX layer, than the first and second parts, thereby providing the ridge and sidewalls of the ridged waveguide. The first part forms a slab region on one side of the ridge of the waveguide, and the second part forms another slab region on an opposing side of the ridge of the waveguide.

After the annealing and etching, the device 100 now also includes diffusively doped regions 16b and 14b which vertically extend (i.e. away from the SOI chip/BOX layer) relative to the doped regions 16a and 14a. The diffusively doped regions extend only along a portion of the sidewalls of the ridged waveguide, that is to say they do not extend the entire height H_w (see Figure 9) of the third optically active region, instead extending only to a height hi and h₂respectively. Therefore, the third optically active region can be described as having a T shape, with a width W and height Hw. The total height H of the ridge is the height of the third optically active region H_w plus the height of the raised capping layer. The depth to which the first and second optically active regions are etched is shown in Figure 9 as dimension E. The final dimensions of note are the depth to which the dopants have diffused into the third optically active region - Di and D₂. The depths Di and D₂ may each have a value in the range 50 nm - 500nm. The depth will depend on the anneal temperature and time over which annealing is performed. The doping may be graduated with the highest concentration of dopants being located near the sidewall.

In this configuration, the device 100 is operable as either an electro-absorption modulator (EAM) through use of the Franz-Keldysh effect or as a photodiode. The methods described above allow the capacitance and E field strength and optical insertion loss to be controlled.

The capacitance, C, is proportional to the surface, separation, and concentration of the doping regions 14 and 16. Therefore the capability to tune the intrinsic region width described as W-(Di+D₂) - and heights hi and h₂ enable the doping position to be chosen accurately, and to decrease the capacitance whilst retaining a uniform field strength across the waveguide core (i.e. the ridge of the ridged waveguide).

Similarly, the E field strength should be optimised such that there is a maximal overlap of the field with the optical mode. The capability to modify Di, D₂, hi, and h₂ separately facilitates the maximisation of the field strength overlap with the core of the optical mode, whilst also reducing the amount of doping present in the waveguide (and therefore the overlap with the optical mode). In Ge or SiGe optically active regions, p-type doping represents a much higher optical loss compared to n-type doping. Therefore, the capability to reduce the overlap of the field with p-type doping by modifying D2 and h2 is very desirable.

Figures 10 and 11 show, respectively, example devices which have been produced using the method described above.

In Figure 10, an optical device has been provided on top of a buried oxide (BOX) layer and within a trench 2304 formed of silicon. Electrical contacts 2312 and 2313 respectively connect to heavily doped regions 2307 and 2311. The heavily doped regions are located adjacent to and within doped regions 2309 and 2314. These doped regions have been provided via the method described above i.e. they have diffused into a ridge waveguide 2301. Unlike previous embodiments, there is therefore no subsequent etching step (such as that shown in Fig 8).

In the embodiment shown in Figure 10, a first diffusively doped region 2309 extends only beneath a first sidewall; it does not extend up the sidewall. The second doped region 2314 is located at the opposite sidewall and may have been diffused into the region beneath the second sidewall. Alternatively, the second doped region 2314 may have been implanted after an annealing step via an angled implant (for example at 45°). As a further variant, the first doped region 2309 may have been provided by an angled implant and the second doped region 2314 may be provided by diffusion.

The doped region then extends upwards from the laterally diffused region, up the second sidewall. In fact, in the embodiment shown, the second doped region (in the form of an N doped region) extends along the entire second sidewall and onto the roof of the waveguide.

A similar structure is shown in Figure 11. However, in this instance the device was formed directly on top of a buried oxide layer, with no intervening Si slab (as is present 2304 in the device shown in Figure 10).

The method allows either of the doped regions to be in the SiGe slab, instead of in the Si slab. This avoids the need to have dopants of one species in the Si slab, and dopants of another species in the SiGe (which may lead to a heterojunction existing in the device).

While the invention has been described in conjunction with the exemplary embodiments described above, many equivalent modifications and variations will be apparent to those skilled in the art when given this disclosure. Accordingly, the exemplary embodiments of the invention set forth above are considered to be illustrative and not limiting. Various changes to the described embodiments may be made without departing from the spirit and scope of the invention.

REFERENCE NUMERALS FROM FIGS

- Si wafer/substrate

2a, 2b - Side walls of SOI channel

- SiGe

4a - First portion of the optically active region

4b - Second portion of the optically active region

4c - Third portion of the optically active region

6, 6a, 6b, 6d - Resist

- PECVD oxide

8a - Raised portion of the capping layer

- S1O2 thermal layer

- BOX

- P-doping

14a - non-diffusively doped region

14b - diffusively doped region

- N-doping

16a - non-diffusively doped region

16b - diffusively doped region

- P++ doping

- N++ doping

100 - Optical device

W - Width of raised portion of the capping layer

H - Height of the raised portion of the capping layer from the SOI

H_w - Height of the ridged waveguide from the SOI

E - Depth of the etch into the optically active region to form the ridged waveguide D1, D2 - Depth of diffusion into the sidewalls of the ridged waveguide hi, h2 - Height of diffusively doped regions from the SOI

Claims

1. A method of producing an optical device (100) having the steps of: growing an optically active region (4) providing a capping layer (8) on top of the optically active region; disposing a first resist (6a) along a part of the capping layer above the optically active region;

etching a part of the capping layer not covered by the resist to create:

a raised portion (8a) of the capping layer below the resist, the raised portion having a width less than the optically active region;

a first part (4a) of the optically active region extending from a first side of the raised portion away from the raised portion;

a second part (4b) of the optically active region extending from the second side of the raised portion away from the raised portion; and a third part (4c) of the optically active region, disposed between the first and second parts of the optically active region and beneath the raised portion of the capping layer;

removing the first resist (6a); performing a step (a) of:

disposing a second resist (6b) over the first part of the optically active region, leaving the second part of the optically active region exposed; and implanting a first species of dopant in the second part of the optically active region, and subsequently removing the second resist;

and/or performing a step (b) of:

disposing a third resist (6c) over the second part of the optically active region, leaving the first part of the optically active region exposed; and implanting a second species of dopant in the first part of the optically active region (16), and removing the third resist;

etching the first and second parts of the optically active material such that the height of the third part is greater than the respective heights of the first part and second part; and annealing the optical device so that the implanted dopants diffuse into at least a part of the third part of the optically active region which is below the raised capping layer, thereby producing at least one diffusively doped region (14b, 16b).

2. The method of claim 1, wherein when only step (a) has been performed, the method may include, after the step of annealing the optical device, a step of:

performing an angled implant of dopants into the first part of the optically active region (16)

3. The method of claim 1, wherein when only step (b) has been performed, the method may include, after the step of annealing the optical device, a step of:

performing an angled implant of dopants into the second part of the optically active region (14).

4. The method of claim 1, wherein step (a) and step (b) are performed.

5. The method of any preceding claim, further including:

disposing a fourth resist (6d) over:

a portion of the first part of the optically active region and/or a portion of the second part of the optically active region; and the raised portion of the capping layer, and further implanting dopants of the first species in the exposed portion of the second part of the optically active region thereby providing a heavily doped region, and/or further implanting dopants ofthe second species in the exposed portion ofthe first part of the optically active region thereby providing a heavily doped region; and removing the fourth resist.

6. The method of claim 2, wherein steps of disposing the fourth resist, implanting dopants, and removing the fourth resist are performed after annealing the optical device.

7. The method of claim 2, wherein the portion of the first part and the portion of the second part are adjacent to the raised portion of the capping layer.

8. The method of claims 6 or 7, wherein the amount of first and second species dopant is controlled such that, when the optical device is annealed, the dopants diffuse an equal distance into the third part of the optically active region.

9. The method of any preceding claim, wherein the optically active region is grown within a channel of a silicon on insulator (SOI) chip.

10. The method of claim 9, wherein the sidewalls and/or a bottom layer of the channel are lined with an insulating layer.

11. The method of any preceding claim, wherein the SOI chip is provided on top of a buried oxide layer.

12. The method of any preceding claim, including a step of further etching the first and second parts of the optically active region to create a proud waveguide ridge (30), the diffusively doped regions each being located at a respective sidewall of the ridge but extending only part of the way up the length of the respective sidewalls.

13. An optical device comprising:

an optically active region (4);

a raised capping layer (8a) disposed on top of the optically active region, the raised capping layer having a width less than the optically active region; and a ridged waveguide (4a, 4b, 4c) comprising:

a first part (4a) of the optically active region which extends from a first side of the raised portion away from the raised portion;

a second part (4b) of the optically active region which extends from a second side of the raised portion away from the raised portion;

a third part (4c) of the optically active region, disposed between the first and second parts of the optically active region and beneath the raised capping layer;

wherein the third part of the optically active region has a height which is greater than respective heights of the first and second parts of the optically active region, thereby providing sidewalls of the ridged waveguide;

a first doped region, within the first part of the optically active region; a second doped region, within the second part of the optically active region; and at least one diffusively doped region (14b, 16b) located on a sidewall of the ridged waveguide, the diffusively doped region extending beneath or only part of the way up the height of its respective sidewall.

14. The optical device of claim 13, including two diffusively doped regions (14b, 16b) located on opposing sidewalls of the ridged waveguide, wherein at least one of the diffusively doped regions extends beneath or only part of the way up the height of its respective sidewall.

15. The optical device of claim 14, wherein each diffusively doped region is located at a respective sidewall of the ridge waveguide and extends only a part of the way up the height of the respective sidewall.

16. The optical device of claim 14, wherein one diffusively doped region is located beneath a first sidewall of the ridge waveguide, and the other diffusively doped region extends along a second sidewall.

17. The optical device of any one of claims 13 to 16, wherein the first and second doped regions further comprise respective heavily doped portions.

18. The optical device of claim 17, wherein the respective heavily doped portions are distal from the third part of the optically active region.

19. The optical device of any of claims 13 to 18, wherein the diffusively doped regions extend an equal distance into each respective sidewall.

20. The optical device of any of claims 13 to 19, wherein the optically active region is disposed within a channel of an SOI chip.

21. The optical device of claim 20, wherein the sidewalls of the channel are lined with a silicon dioxide thermal layer.

22. The optical device of either claim 20 or 21, wherein the SOI chip is provided on top of a buried oxide layer.

Intellectual

Property

Office

Application No: GB1701240.2 Examiner: Mr Jeremy Cowen