GB2380059A

GB2380059A - Semiconductor optical source and modulator

Info

Publication number: GB2380059A
Application number: GB0123045A
Authority: GB
Inventors: Andrew James Shields
Original assignee: Toshiba Research Europe Ltd
Current assignee: Toshiba Europe Ltd
Priority date: 2001-09-25
Filing date: 2001-09-25
Publication date: 2003-03-26
Anticipated expiration: 2021-09-25
Also published as: GB2380059B; GB0123045D0

Abstract

An electro-optical device formed on a GaAs substrate comprises a reservoir 13 of excess carriers (eg quantum well, quantum wire), n-type 21 and p-type 23 ohmic contacts, source 25 and drain 27 contacts, and a stripe waveguide 22. The p-type contact 23 and n-type contact 21 are biased with respect to the drain contact 27 to inject holes and an excess of electrons, thus generating charged excitons in the reservoir. These charged excitons decay to produce photons which are guided out of the device by the waveguide 22. By applying an electric field between the source 25 and drain 27, charged excitons can be caused to drift out of the waveguide 22 before decay, thus the output intensity can be controlled or switched off. In an alternative embodiment, output wavelength may be modulated. Furthermore, the generation of excitons may be optically activated by illumination from a light source, rather than electrically activated by electrical contacts 21,23.

Description

An Optical Device and an Opticallv Activated Device The present invention relates to the field of devices which emit or modulate light and devices which are activated by light. More specifically. the present invention relates to devices whose operation requires the formation of charged excitons.

An exciton is a semiconductor complex of electrons and holes. Due to their opposite charge. electrons and holes experience a mutual attraction due to the Coulomb interaction and can bind to form quasi-particles. The simplest exciton comprises a single electron and a single hole and consequently has no net charge. However, more exotic. so-called. charged excitons may also be formed. which contain different numbers of electrons and holes. For example. the single negatively charged exciton comprises a single hole and two electrons. similarly. the single positively charged exciton comprises a single electron and two holes. Charged excitons which are multiply charged or which comprise more than one electron/hole pair also exist.

It has been discovered recently that. surprisingly. charged excitons drift in an electric field. Although the existence of charged excitons was speculated in 1958 by Lampert (Phys. Rev. Lett. 1. 450-453 (1958) ). evidence for their mobility has only recently been observed.

As charged excitons can drift in an electric field prior to their radiative recombination, it is possible to use applied electric fields to control the spatial position at which a photon is emitted.

In a first aspect. the present invention provides a device comprising : a reservoir of excess carriers having mobility in at least one direction : said reservoir having a variation in at least one of its characteristics in said at least one direction ;

generating means for generating electron/hole pairs in the reservoir: and means for applying an electric field across at least a part of the reservoir for moving recombining carriers in said at least one direction.

The recombining carriers are an electron and a hole. Generally an electron and hole may bind. due to their mutual attraction. to form a simple exciton. Since this simple exciton has no net charge. it does not drift in an applied electric fie) d. However. in the presence of excess carriers. one or more excess carriers can interact with the simple exciton to form a charged exciton. The charged exciton has a net charge and thus drifts in an applied electric field. The excess carriers ma) be either electrons or holes and the charged exciton may therefore have a net negative or positive charge. The charged exciton. and thus the recombining carriers which comprise an electron and a hole of the charged exciton. experience a Coulomb force in the presence of an electric field and can hence be shifted into a part of the reservoir. which has a different characteristic. For example. part of the reservoir may form part of a waveguided region.

The device may be configured as an electrically controllable optical source when a part of the reservoir lies in a waveguided region. When charged excitons decay. they emit a photon. If this photon is emitted from an exciton in the waveguide. then the photon can be guided out of the device and into a suitable collector such as an optical fibre or the like. If a photon is emitted by an exciton which is outside the waveguided region. then the photon is less likely to be guided out of the device into a collector and will be lost. Thus. it is possible to vary the intensity of the collectable output of the device by varying the number of charged excitons within the waveguide using an applied electric field.

Since the intensity of the device is varied by moving the charged excitons over short distances of a few microns or less. the device has a very high switching speed.

Preferably, the width of the waveguide in said at least one direction is of the order of. or less than, the drift length of the charged excitons in the said at least one direction. Typically. the waveguide will be at most 54m wide in this direction.

The waveguide is preferably provided by a so-called stripe waveguide which channels photons emitted in the stripe along said stripe and out of the device.

Alternatively. or in addition to the provision of the above waveguide, further characteristics of the reservoir may also vary in the said direction. For example. there may be a variation in the width or composition of the reservoir in the said direction which may allow the electrical field to modulate the output wavelength of the dex ice h ! moving the charged excitons to the appropriate part of the reservoir.

A variation in the characteristics of the reservoir may also be prox ided b\ varying the radiative or non radiative lifetime of the material in said at least one direction. Variations in the radiative and non-radiative lifetimes may be achie\ed b\ implanting the reservoir with ions or dopant ions since regions implanted with ions or dopant ions will have a different radiative or non-radiative lifetime.

Preferably. if a waveguide region is provided. the reservoir is configured such that the non-radiative lifetime is shorter. and/or the radiative lifetime longer. outside the waveguided region.

Either electrical or optical generating means may be provided. If electrical generation means are used. the device essentially acts as an electrically-injected optical source. The device forms a three or four terminal light source. with separate contacts and voltages for carrier injection and output modulation. Electrical generation means may comprises means to inject carriers having an opposing conductivity type to that of the excess carriers into the reservoir or may comprise means to inject both electrons and holes into the reservoir.

Where optical generation means are used. the device acts as an optically pumped light source or optical modulator. as optically excited electron-hole pairs combine with excess carriers to form charged excitons. The output of the device can be modulated as the position of these charged excitons can be controlled with the electric field.

Such a device may preferably comprise a first and a second waveguide in said reservoir. The first and second waveguides being positioned such that charged excitons may be moved between the two waveguides by the applied electric field. In order to be able to move the charged excitons from one waveguide to the other before the\ recombine. the waveguides should typically have a separation in the range 0. 5 11m to 20 urn and generally be about I nm wide. Preferably the separation is of the order of or

less than the drift length of the charged excitons. c In a preferred mode of operation. an optical beam is incident on the first waveguide and excites electron/hole pairs within this waveguide which combine with the excess carriers to form charged excitons. When the electric field is applied. at least some of the charged excitons are swept into the second waveguide so that they can recombine and light may be outputted from the second waveguide.

Thus. the light distribution between the two waveguides may be controlled by the electric field so that light may be switched from one waveguide to the other or split as required between the two waveguides. The output from both waveguides may be collected independently by suitable elements such as an optical fibres or the like.

The reservoir may preferably be provided by a quantum well. More preferably. the excess carriers are provided in a two dimensional electron or hole gas where the excess carriers will have a high mobility in two dimensions. However. the reservoir may also comprise a so-called quantum wire. where the carriers have mobility in just one dimension or a bulk structure where the carriers have mobility in three dimensions.

The means to apply an electric field preferably comprise a source and drain ohmic contact provided to said reservoir.

It is important to maintain excess carriers in the reservoir region. otherwise neutral excitons will be formed as opposed to charged excitons. Thus. the device preferably comprises means to provide excess carriers to the reservoir. These means may be provided by a doped barrier layer.

More preferably. the device comprises means to vary the excess carrier concentration within the reservoir. These means may comprise an electrode for varying the electric field across a doped layer configured to provide carriers to said reservoir.

The same electrode may also be used to control injection of carriers into the resern oir.

For example. the device may comprise a p-i-n structure where the p and n doped regions are biased to inject carriers into the reservoir which is provided in the intrinsic region. A doped barrier layer is provided between one of the doped regions and the reservoir. By applying a bias between the doped regions with respect to the rester\ or region. the number of excess carriers in the reservoir may be controlled.

In a second aspect. the present invention provides a method of operating an optical device. the device comprising: a reservoir of excess carriers having mobility in at least one direction. said

reservoir having a variation in at least one of its characteristics in said at least one direction. the method comprising : generating electron/hole pairs in the reservoir and applying an electric field across at least a part of the reservoir for moving recombining carriers in said at least one direction.

In a third aspect. the present invention provides a method for fabricating a device. the method comprising: forming a reservoir region configured to support excess carriers such that they have mobility in at least one direction: forming a doped region configured to supply excess carriers to said reservoir region ; providing means to create electron hole pairs in said reservoir region: forming ohmic contacts to said reservoir region : and forming a variation in the characteristics of said reservoir region in said at least one direction.

The step of providing means to create electronlhole pairs in the reservoir may comprise forming a heavily doped layer having an electrical contact. the layer being

positioned within the device such that upon electrical activation. it can inject carriers having an opposing conductivity type to the excess carriers. into the reservoir region.

Alternatively. this step may comprise providing a p-type region on one side of

the reservoir region and an n-type region on the opposing side of the reservoir region to -, ion to the p-type region. The n and p-type regions being electrically contactable such that upon electrical activation. electrons and holes are injected into the reservoir region. t The region for providing excess carriers to the reservoir region preferahl\ comprises a doped barrier layer. More preferably. the doped barrier layer will he a modulation doped barrier laver comprising a doped barrier layer and an undoped spacer laver. the spacer layer being provided between the reservoir region and the doped layer.

The step of providing a variation in the reservoir region may be performed as the reservoir region is formed. For example. a variation in the composition or thickness of the region may achieved by stopping the rotation of the substrate during growth and depositing the material of the reservoir region in a spatially non-uniform manner.

Alternatively. it may be achieved by modifying the layers around the reservoir region. For example. dielectric cladding regions may be provided on either side of the reservoir region such that the reservoir region is located in a one dimensional optical cavity. By forming a thicker stripe of dielectric cladding layer on one side of the optical cavity. a stripe waveguide is formed which guides radiation out of the device along this stripe.

There may also be a spatial variation in the radiative or non-radiative lifetime in the reservoir layer. which can be achieved by ion beam damage of part of the reservoir.

In a further example. radiation emitted from the device is collected normal to the plane of the layers. By masking parts of the output surface of the device. in areas which are complementary to the positions where the recombining electron/hole pairs are generated. the intensity of the device may be modulated. Thus by applying an electric field in the plane of the reservoir. the recombining carriers can be moved to areas under

the mask and thus be prevented from contributing to the emission normal to the plane of the layers. The output normal to the layers may also be modulated using a spatial variation in a characteristic of the reservoir. such as its composition or width.

Thus. in a fourth aspect. the present invention provides a device comprising : a reservoir of excess carriers having mobility in at least one direction: generating means for generating electron/hole pairs in the reservoir : means for applying an electric field across at least a part of the reservoir for moving recombining carriers in said at least one direction. wherein the device has a variation in at least one of its characteristics in said at least one direction.

The variation in the characteristic may be in regions or layers other than the reservoir. Thus. photons emitted from the reservoir will experience variations in the electrical and optical characteristics of the layers through which they are transmitted depending on their generation position within the reservoir. This generation position being moveable by the applied electric field.

Preferably. the device has a variation in its optical characteristics in said at least one direction such that photons emitted due to recombination of carriers in one region of said reservoir are transmitted through layers having different optical characteristics to the layers through which photons are transmitted which are emitted from another region of the reservoir. For example. photons emitted from one region of the reservoir may be blocked by a mask layer, whereas photons emitted from the other regions of the reservoir may be freely transmitted out of the structure.

The above description has concentrated on the use of charged excitons for electrical control of an optical device. However. the converse is also possible where light is used to control an electrical device.

In a fifth aspect, the present invention provides a device comprising: a reservoir having an excess of carriers which have mobility in at least one direction;

a source and a drain contact provided to said reservoir such that a current can be measured along said at least one direction : and a switchable optical source configured to illuminate said reservoir with radiation sufficient to optically excite a number of electron/hole pairs within the de\ ice. \\ herein the number of electron/hole pairs is of the order of the number of excess carriers.

In the absence of illumination. a current of excess carriers Hows fred from the source to the drain. When the device is illuminated. electron-hole pairs are collected in the reservoir of excess carriers. These pairs bind to excess carriers to form charged excitons. The charged excitons have a much louer mobility than that of the excess carriers and thus. the current falls.

Alternatively. just minority carriers are collected in the reservoir. If the concentration of photo excited minority carriers is approximately half or twice that of the excess carriers the dominant population of charge carriers are charged excitons. If the carrier concentration of the "minority" carriers is substantial I) equal to that of the excess carriers. neutral excitons may be formed. In either case. the current substantially falls upon illumination.

Preferably. the number of excited electron-hole pairs is from 0. 1 to 10 times the number of excess carriers. more preferably 0. 25 to 2 times the number of excess carriers.

For example. if the device is configured so that the reservoir region comprises a two dimensional carrier gas having a concentration of approximately 4xi0'"cm. the illumination intensity in then set to excite approximately 4xIO"cm'electron/hole pairs.

The wavelength of the incident radiation may be set to just excite electron/hole pairs in the reservoir region or may be set to excite electron/hole pairs in the other layers of the device such that one or both of the photo-excited carriers may be swept into the reservoir region.

As the change in current is not due to a change in carrier concentration. but rather in the nature of the charge carriers themselves and thereby their mobility. the device should have very high switching speeds.

The optical source is preferably a pulsed source. More preferably the pulse width is less than the radiative lifetime of recombining carriers in the reservoir.

Preferably. means are provided to control the excess carrier concentration within the reservoir. These means may comprise a doped barrier layer which is controlled N ia an electric field applied with respect to the reservoir.

In a sixth aspect. the present invention provides a method of optically operating an electrical switch. the switch comprising a reservoir having an excess of carriers which have mobility in at least one direction and source and drain contacts provided to said reservoir. such that a current can be measured along said at least one direction. the method comprising varying the illumination level of the reservoir up to a level which results in the generation of a number of electron/hole pairs of the order of the number of excess carriers.

The present invention will now be described with reference to the preferred nonlimiting embodiments in which: Figure I illustrates a cross-section of a device in accordance with a first embodiment of the present invention, Figure 2 illustrates a plan view of the upper surface of the device of Figure ! : Figure 3 illustrates a device in accordance with a further embodiment of the present invention Figure 4 illustrates a plan view of the upper surface of the device of Figure 3 :

Figure 5 illustrates a cross section of a further device in accordance with an embodiment of the present invention : Figure 6 illustrates a plan \ ie\\ of the upper surface of the device in Figure 5 : Figure 7 illustrates a cross-section of the device in accordance with a further embodiment of the present invention which comprises two wave guides: Figure 8 illustrates a plan view of the upper surface of the de\ ice of Figure 7: Figure 9 illustrates an optically switched electrical device in accordance with an embodiment of the present invention : and Figure 10 illustrates an optically switched electrical device in accordance with a yet further embodiment of the present invention.

Figure I illustrates an optical source. Charged excitons created within the optical source recombine to emit light. which is channeled out of the structure by a

waveguide. Bs applying an electric field across the waveguide. charged excitons can be swept out of the waveguide before they recombine. thus the number of charged excitons in the waveguide can be varied and hence the intensity of the source can be varied or even switched off.

The device is formed on an n+ doped GaAs substrate. A buffer layer 1 is provided overlying and in contact with said substrate 1. The buffer layer comprises 200 nm of GaAs n+ doped Si 5 x 1018 cm-3. A lower doped cladding layer 3 is then provided overlying and in contact with said buffer layer 1. The lower doped cladding layer comprises 1700 nm of Al, Gal~, As. where y=0.6. which is n'doped with Si 5x lO" ! cm. A lower undoped cladding layer 5 is then formed overlying and in contact with said lower doped cladding layer 3. The undoped lower cladding layer 5 comprises 300 nm of AIyGal yAs.

A first lower undoped barrier layer 7 is then provided overlying and in contact with said lower undoped cladding layer 5. The lower undoped barrier layer comprises 40 nm of Al, Gal., As. where x=O. 3. A lower doped barrier layer 9 is then formed overlying and in contact with said lower undoped barrier layer 7. The lower doped barrier layer 9 comprises 40 nm of AI, Gal~, As doped n-type Si 1 x 1018 cm--'.

An undoped spacer layer 11 is then formed overlying and in contact w ith said I L- doped barrier layer 9. The undoped spacer layer 11 comprises 60 nm of AI, Gal~, As.

A quantum well layer 13 is then formed overlying and in contact with said spacer la : er 11. The quantum well layer 13 comprises 20 nm of undoped GaAs.

An undoped upper barrier layer 15 is then provided overlying and in contact with said quantum well layer 13. The undoped upper barrier layer comprises 140 nm of undoped Al, Gal., As. Upper cladding layer 17 is then formed overlying and in contact with said upper undoped barrier layer. Said upper cladding layer comprises 300 nm of undoped AI, Gal., As.

Upper doped layer 19 is then formed overlying and in contact with said upper cladding layer 17. Said upper doped layer comprises 200 nm of p-type Al@ Ga1-@ As. It is doped with 5 x 1018 cm' Be. A GaAs capping layer (not shown) is then formed overlying and in contact with said upper doped layer 19. The capping layer comprises 10 nm of p-type GaAs doped with 5 x 10' cm'3 Be.

The layers are first processed by forming a n'ohmic contact 21 on the lower side of substrate/buffer layer I. As both the substrate and buffer layer 1 are doped with an excess of n-type carriers. it is possible to make electric contact to the device using a substrate side contact.

The n-type contact 21 is formed from Au: Ni: Ge alloy in the weight composition ratio of 88: 12: 1. This is applied to the back surface of the wafer using evaporation and is annealed for l minute at a temperature of 430 C in a reducing N2-H2 atmosphere.

Next. a stripe having a width of I pm is etched along the top surface of the structure. In the figure. the stripe is oriented along a direction normal to the plane of the page and only its cross-section is seen. The etch proceeds down and into upper cladding layer 17. to within 250 nm of the waveguide region. i. e. through at least 50nm of the upper cladding region 17. The stripe can have a width in the range 0. 5 to 20 um.

The stripe is defined by spinning photoresist. for example Shipley 1830. onto the chip and baking the photoresist at 90 C for 10 minutes to harden the resist. The resist is exposed and developed such that the photoresist remains only on the top of the stripe which is to be etched after developing. The stripe is then etched with a solution of hydrogen peroxide: sulphuric acid: water in a ratio of 1: 8: 1 00.

A p-type ohmic contact 23 is then made to the upper surface of stripe 22. The ptype contact is made using AuBe contacts which are annealed at 480 C for 3 minutes.

A source ohmic contact 25 and a drain ohmic contact 27 are then made to the quantum well layer 13. Ohmic contacts 25 and 27 are disposed on opposing sides of stripe 22 to allow an electric field to be applied. in a direction Iving in the plane of the quantum well 13. and in the region which lies directly underneath stripe 22.

These n-type ohmics 25 and 27 should contact the quantum well laver 13.

However. it is important that they do not short to lower doped layer 3 or lower doped cladding 5. To achieve this. these contacts may either be made from AuNiGe and the annealing temperature and time carefully controlled so as to not allow the contact to penetrate the underlying doped layers.

Alternatively. a"shallow"ohmic contact can be used using Pd-Ge as the ohmic metal. These types of contacts have been described in Patel et al. Applied Physics Letters. volume 65. no 7. pages 851 to 853. 15 August 1994. To make shallow ohmic contacts. a window (not shown) is etched through the semiconductor layers to within 20 nm of the quantum well layer. Pd-Ge is then evaporated into this window and annealed at 350 C for 5 minutes.

Once these contacts have been made. a laver of insulator. such as polyimide 29 is then formed on the surface of the device. Windows are defined in the polyimide layer to allow contact to be made to upper p-type ohmic 23 and source and drain n-type ohmics 25 and 27. The insulating layer is deposited by spinning pohimide. which can be processed in a similar manner to photoresist. The polyimide is cured b\ baking it on a hot plate at 1150C for 900 seconds. The laver is then exposed to UV light through a chrome-on-glass mask. UV light cross-links the polyimide so that when the laxer is exposed and developed. only the unexposed areas of the polyimide are dissohed. The insulating layer was 500 nm thick.

Figure 2 illustrates a plan view of the upper layer of the device of Figure 1. This figure allows the contact arrangement to be viewed. As described in relation to Figure 1. windows 31. 33. 37 and 41 are defined in the insulator by photolithography in order to allow contact to each of the ohmic contacts.

Contact metal which may be NiCr/Au 35 is then evaporated in order to make contact to the areas of the p-type ohmic 23 exposed by windows 3) and 33. Two contacts are made to the ohmic contact 23 as a precaution in case one of the contacts does not work or becomes detached during processing. Simultaneously contact metal 39 is also defined to make contact to the source n-type contact 25 in window 37 and contact metal 43 to electrically connect to drain ohmic 27 in window 41.

In order to produce an output surface. the device is cleaved at edge 45.

Alternatively. the device could be etched to expose edge 45.

The above structure forms an optical wave guide. the optical field being confined mainly to the optical cavity. The optical cavity comprises lower barrier layer 7. doped barrier layer 9. spacer layer 11. quantum well layer 13 and upper barrier layer 15. This optical cavity 47 is sandwiched between lower cladding region of the lower undoped cladding layer 5 and lower doped cladding layer 7 and upper cladding layer 17.

If the cladding layers have a lower average refractive index than the optical cavity 47, the optical mode is confined to the layers which comprise the optical cavity.

In the above example. this condition is achieved ify > x. for example y = 0.6 and x = 0. 3 have been found to work particularly well. It is important that the lower cladding region which comprises both the lower doped cladding layer 3 and lower undoped cladding layer 5 is sufficiently thick to prevent coupling to the substrate modes.

Stripe 22 further defines the waveguided region in one of the directions in the plane of the cavity 47. The stripe-wave guide concentrates the optical field under stripe 22 and directs photons emitted here out of the device at output surface 45. The stripe should be etched down to within 250 nm of the cavity region. so as to affect the optical field in the cavity region.

Optionall) one or more of the AI, Gal-, As or AI, Gal-, As lavers can be replaced with a GaAs/AlAs or GaAs/AlGaAs short period superlattice with the same average AI concentration. The individual layer thicknesses in the superlattice are a few nanometers. For example. an Al0 6Ga0 4As layer can be replaced with a GaAs (2nm)/AlAs (3nm) superlattice. while an AInGaopAs layer can be replaced with a GaAs (2nm) l AIAs (1. 5nm) superiattice. Replacing a bulk alloy layer with a short period superiattice helps to smooth the layer interface and thereby increase the charged exciton mobility.

In this example the remote dopants were arranged below the quantum well.

However. they may also be placed in the barrier layer above the quantum well. or even in both the lower and upper barrier layers.

Optionally the areas outside the stripe region may be implanted with n-type dopants. so as to reduce the resistance between the source or drain contacts and the active region of the reservoir. Donors are not implanted in the active region so as to maintain a high charged exciton mobility. The implanted donors also reduce the exciton lifetime in the area outside the waveguide region, thereby enhancing the modulation of the light output intensity.

To operate the device. the n'ohmic 21 is biased with respect to the drain contact 27 to fix an excess electron density within quantum well layer 13. The electron density

is set to be in the region ! x] 0'"- 3x10"cm'. but more preferably may be 5x10'"cm'''.

I ni The upper p-type contact 23 is then also biased with respect to the drain contact 27 so as to inject holes into the quantum well from the p-type contact 23 and upper doped layer

19. This occurs when the p-type contact 23 is positive)) biased \uth respect to drain contact 27.

In this configuration. charged excitons are formed in the quantum well layer 13.

Due to the excess of electrons. a single hole will form a quasi-particle \\ith two t electrons. This state will then decay by recombination of an electron and a hole to produce a photon. leaving behind an excess electron.

These photons will be guided by stripe wave guide 22 out of the structure.

However. if an electric field is applied within the plane of the quantum sell la ! eu the charged excitons can be swept out of the stripe waveguided region. Thus. it is impossible to modulate the intensity output of the source and even switch the source oil As this switching action occurs due to the excitons moving over a distance of around 1 lam. this device will have a fast switching speed. An applied source-drain voltage of 0. 3V. applied between source and drain contacts separated b\ 3u. will develop an electric field of 1 kVcm-1 in the plane of the quantum well. This electric field of I kVcm'sweeps charged excitons over a distance of I 11m in 33ps. assuming a charged exciton mobility of 3000 cm2V-1s-1. Within this time there will be negligible radiative recombination. This switching time can be reduced by applying a larger source-drain voltage or by cooling the device to increase the charged exciton mobility.

For a mobility of 50 000 cm2V-1s-1 and a field of I kVcm''the switching time is reduced to 2ps.

The emission wavelength of the device may be extended by replacing the GaAs quantum well 13 with an In\Gal. In@ Ga1-yNy. GaAs1-ySb@ layer. The emission wavelength can also be decreased by using an AI, Gal-\, As layer. instead of the GaAs quantum well 13.

In the above example the device was formed from a GaAs/AlGaAs heterostructure. however. other combinations of semiconductor materials are also possible. For example the structure could be formed from a In0 53Ga0 47As/

Ino A. to4 ! As on an InP substrate. Another possibility is a GaN/AGaN heterostructure on a Sapphire. SiC or GaN substrate. Yet another possibility is a Si/Si02 structure on Silicon substrate.

In a further example. the composition and/or width of the quantum well la\ er 1) may be varied so that the output wavelength of photons generated in different parts of the quantum well layer 13 will be different. Thus. the in-plane electric field may be used to modulate the output wavelength of the device.

Figures 3 and 4 illustrate a further embodiment of the present invention. The device is based on that of Figure 1. To avoid unnecessary repetition. like reference numerals will be used to denote like features.

The device is formed on an undoped substrate. An undoped GaAs buffer layer is then formed overlying and in contact with the substrate 1. The remainder of the layer structure of the device remains identical to that described with reference to Figure 1.

As the substrate/buffer layer 1 is undoped. it is not possible to make contact to lower doped layer 3 without performing substantial etching of the substrate/buffer layer I. Therefore. contact to doped layer 3 is made from the top of the device.

A mesa 51 is etched in order to define the area of the device. The mesa etch proceeds down as far as the substrate/buffer layer 1. In order to make contact to lower doped layer 3. a window 59 is etched through mesa 51. The contact to the n-doped layer 3 may be formed using an n-type ohmic contact comprising AuNiGe annealed at 430 C for I minute.

In order to avoid the source ohmic contact 25 and drain ohmic contact 27 shorting to doped layer 3, these contacts are formed using shallow ohmic contact metallisation such as PdGe which does not penetrate deeply into the structure.

Windows are etched in the same manner as described in relation to the device for Figures 1 and 2 and then shallow ohmic contacts are formed in order to make contact to the source 25 and drain 27 ohmics.

An insulating layer 29 is then formed with windows 37. 41. 59. 31 defined in said insulating layer in order to expose the source and drain contacts 25 and 27. the upper ohmic contact 23 and the contact to the lower doped layer. Contact metal 61. 39.

35. 43 is then defined to allow external electrical connection to the device.

Figures 5 and 6 show a yet further variation on the device of Figure 1. The device is fabricated using a rye, growth technique. In the same manner as described in relation to Figure 3. the substrate 1 is undoped and a 500 nm undoped GaAs buffer layer is grown on the substrate. A back-gate layer 71 is then formed overlying and in contact with the substrate/buffer layer. The back-gate layer 71 comprises 500 nm of n doped GaAs. This layer is doped with Si at a concentration of5 x 10's cm--'. This layer 71 is then etched with a suitable hydrogen peroxide/buffered hydrofluoric acid based etch such that only a small stripe 73 of the layer remains. This stripe has shallow sloping sides which allows layers to be formed overlying this relief without discontinuities.

In order to etch back gate 71. the structure has to be removed from the growth chamber. The structure is cleaned prior to reinsertion back into the growth chamber.

However. there will inevitably be some defects on the surface of the structure due to irregularities in the etch and spurious contamination.

A 5 nm growth initiation layer (not shown) is then formed from undoped GaAs.

This layer overlies and it is in contact with the substrate/buffer 1 and the backgate 71.

Lower cladding layer 5 is then formed overlying and in contact with the growth limitation layer. The lower cladding layer comprises 2000nm of AI, Gal-, As with y=0.6.

A lower undoped barrier layer 7 is then formed overlying and in contact with lower cladding layer 5. The lower and undoped barrier layer comprises 35nm of undoped AI, Gal-xAs with x=0.3.

Doped barrier layer 9 is then formed overlying and in contact with undoped barrier layer 7. Doped barrier layer 9 comprises 40 nm of n-type doped Al, Gal~, As doped with Si with a concentration of I x 1018 cm-3.

Undoped spacer I I is then formed overlying and in contact with doped barrier layer 9. Undoped spacer 11 comprises 60 nm of undoped Al@ Ga1-@ As. Quantum well layer 13 is then provided overlying and in contact with said spacer layer 11. Quantum well layer 13 comprises 30 nm of undoped GaAs.

Upper barrier la) er 15 is then provided overlying and in contact with said Quantum well layer 13. Upper barrier layer comprises 135 nm of nm of undoped AI, Gal~, As. An upper cladding layer 17 is then provided overlying and in contact with said upper barrier layer 15. Upper cladding layer 17 comprises 300 nm of undoped AIxGaAs.

Upper doped layer 19 is then formed overlying and in contact with said upper cladding layer 17. Upper doped layer 19 comprises 200 nm ofp-type AGaAs doped with Be at a concentration of5 x 10'8 cm-3. The structure is finished with a cap layer (not shown) comprising 10 nm of p-type GaAs doped with Be at a concentration 5 x 10'cm'.

A mesa is then formed by etching the layers down to lower undoped barrier layer 7.

In the same manner as described with reference to the devices of Figures]-4. contact is made to the source and drain and upper doped layers. These are formed in identical manner to those previously described. However. it should be noted. that because the back-gate 71 does not lie directly underneath contact 25 and 27. the depth of these contacts does not need to be controlled since they cannot short to the back gate.

Window 81 is etched through the mesa down to back-gate layer 71 in order to provide a back-gate contact.

Insulating layer 29 is then provided overlying the structure. Figure 6 shows the

contact arrangement for the device of Figure 5. Windows are defined in the insulating t layer 29 in order to contact source and drain ohmic 25 and 27. upper ohmic contact 13 and back gate contact.

In a variation of the above device. part of the lower cladding layer may be highly doped and grown and etched with the back gate. It is important to remove any doped layers from underneath the positions of the ohmic contacts to the upper lax ers. to avoid a shorting contact.

It is also possible to use other techniques such as ion beam damage of the layer 71 in the vicinity of the source and drain ohmic contacts to ensure that these contacts cannot short to back gate 71.

Figure 7 shows a further device in accordance with an embodiment of the present invention. The device of Figure 7 has two stripe waveguides. By applying a bias between the source and drain contacts. the excitons can be shifted between the two waveguides.

The layer structure of the device is virtually identical to that described with reference to Figure 3. Therefore. its description will not be repeated. Like reference numerals are used to denote like features.

The device is processed slightly differently from that of the device of figure 3.

Instead of forming a single stripe. two stripes 101 and 103 are formed. A source contact 25 is provided on one side of the stripes and a drain contact 27 is provided to the quantum well layer 13 on the opposing side of the parallel ridges. Window 81 is etched through the mesa down to back-gate layer 71 in order to allow a AuNiGe ohmic contact to be formed to the back-gate contact.

The contact arrangement of the device is illustrated in Figure 8 which shows a plan view of the upper surface. In this particular example, no contacts are made to waveguides 101 and 103. Insulator 115 is spun onto the device. exposed and developed

so as to provide contact windows for connecting to the source. drain and back gate contacts. Contact metal 11 1 is provided to contact to the top of source ohmic contact 25 and contact metal 113 is provided to contact to drain ohmic contact 27. Contact metal 117 is evaporated onto the device to allow external contact to be made to the back-gate 71. through window 81.

The ridges have separation of between 0.5 to 20 um. ! n operation, the excess carrier density of the quantum well is set in the same manner as described with reference to the device of Figure !. The de\ ice is irradiated along the waveguide defined b) one of the ridges. sa\. for example. ridge 101. This radiation causes electron/hole pairs to be excited within the quantum well 13. Due to the excess of electrons within the quantum sell 13. a charged exciton complex is formed.

Bs applying a bias between source and drain contacts 25 and 27. the charged excitons can be moved from underneath stripe 101. where they are generated. to an area of the quantum well underneath stripe 03. There. they recombine to emit radiation.

Thus. it is possible to control optical output from waveguide 101 and waveguide 103 by varying the source-drain bias. This provides an electrical device where light can he switched between one or two output arms on picosecond timescales.

Figure 9 illustrates an optically activated electrical switch in accordance with a further embodiment of the present invention.

The structure of the device of Figure 9 is different to that of the preceding devices. As for the preceding device. the structure is formed on an undoped GaAs substrate. Buffer layer 201 is then provided overlying and in contact with said substrate. Buffer layer 201 comprises 200 nm of undoped GaAs. Lower doped layer 203 is then formed overlying and in contact with buffer layer 20 I. Lower doped layer 203 comprises 500nm of n + -doped GaAs. doped with Si to a concentration of 5 x 10"cm-'.

A lower undoped barrier la) er 205 is then formed overlying and in contact with lower doped layer 203. Lower undoped barrier layer 205 comprises 500 nm of undoped AI, Gal~, As.

Lower doped barrier layer 207 is then provided overlying and in contact with lower undoped barrier 205. Lower doped barrier layer 207 comprises 40 nm of n-t) pc

AI, Gal~, As. with x=0. 3. doped with Si at a concentration of 1 x 10'8 cm'. Spacer la) er 209 is then formed overlying and in contact with said lower doped la) er 207. Spacer layer 209 comprises 40 nm of un doped Al, Gal~, As. Quantum well layer 211 is then provided overlying and in contact with said spacer layer 209. Quantum wet) : er 211 comprises 30 nm of undoped GaAs.

Upper undoped barrier layer 213 is then formed overlying and in contact with said Quantum well layer 211. Upper undoped barrier layer comprises 200 nm of undoped AI, Gal~, As. Cap la) er 215 is formed overlying and in contact with said undoped barrier layer. Cap layer comprises 10 nm of undoped GaAs.

A mesa is defined to limit the area of the device. In addition. a contact (not shown) is made to lower doped layer 203. after etching a window through the mesa down to the lower doped layer 203.

Source ohmic contact 217 and drain ohmic contact 219 are then made to Quantum well layer 211. Source and drain ohmic contact can be any type of contact which can make a good ohmic contact to n-type GaAs. It is important to ensure that these contacts do not short to lower doped layer 203. Therefore. these contacts are either formed from NiAuGe and annealed at a temperature and for a time to ensure that they do not diffuse as far as lower doped layer 203 or they are formed using shallow ohmic contact technology. for example. using Pd-Ge shallow ohmic contacts as described with reference to Figure 1.

External contact is then made to the source and drain contacts 217.219 using contact metal layer 221. Contact metal is also formed to the lower doped ohmic

contact. To avoid the contact metal layer 221 shorting to the other layers of the device. insulating layer 223 is formed underlying metal layer 221.

The area on the upper surface of the device between source ohmic contact 217 and drain ohmic contact 219 is kept free of insulator.

In the same manner as described with reference to Figure 1. 100'ver doped layer 203 is biased with respect to drain contact 219 to fix the number of excess carriers in quantum well layer 211.

In the absence of illumination and upon the application of a bias between source contact 217 and drain contact 219. a current flows between the two contacts.

Upon illumination. electronlhole pairs are created within quantum well 211.

The electron/hole pairs bind to the excess electrons to form charged excitons. This results in a reduction in the mobility of the charge carriers and hence in the conductivity of the device. This change in conductivity is achieved without needing to change the net charge density within the quantum well laver. Instead. the mobility of the charge carriers is altered. as electronlhole pairs bind to the electrons to form charged excitons.

Thus. this device can have fast switching speeds.

To see a definite change in the conductivity, it is necessary to illuminate the device with enough radiation sufficient to attach excitons to a substantial fraction of the excess carriers in the quantum well.

In this example. the is operated at an excess carrier density (n) in the range 0. i/ (7t r) to 1. 0/(7rr). where r is the exciton Bohr radius in the reservoir material. The laser intensity is set to excite a density of electron-hole pairs in the range 0.25 n to 2 n.

On the other hand. if the wavelength of the laser is such that only minority carriers are collected in the reservoir and it is desired to produce charged excitons as opposed to neutral excitons. the excess carrier density in the reservoir may be doubled.

If the device is operated with an excess electron density of 4x101 cm'2. a value suitable for a GaAs quantum well. the illumination intensity should be set to excite around 4x10'0 cm-2 electron-hole pairs within the quantum well. For the device shown. taking an absorption coefficient 4x 104 cm-'and an illumination wa\ elength of 690nm.

2. 1-lit's reflected the laser energy should be 0. 15 Jcm-2. assuming 30% of the incident tight is reflected by the front surface. Focused to a spot of 2 um diameter, the laser energ\ should be 4. 7 fJ. The incident intensity required can be reduced by coating the surface of the dc\ice with an anti-reflection coating.

The laser illumination can be supplied in a short pulse with a width which is much less than the radiative lifetime of the exciton (r) which is of order Ins.

For pulse durations longer than the radiative lifetime. the illumination power

should be 0. 15 ujcm'/i : or 15 mWcm'. Focused to a spot of 2 urn diameter. the laser power should be 0.47 nW.

More intense lasers can be used by reducing the laser power incident on the device to the appropriate value using an attenuator. or increasing the diameter of the beam incident on the device so as to create the required excitation density.

Figure 10 shows a further variation on the device of Figure 9. However, in Figure 10. the excess carrier density in the quantum well is controlled by biasing a topgate as opposed to a buried doped layer.

A device is formed on an undoped GaAs substrate (not shown). First. a GaAs buffer layer 251 is formed overlying and in contact with said substrate. The buffer layer comprises 500 nm of undoped GaAs.

A lower undoped barrier layer 253 is then provided overlying and in contact with said buffer layer 251. Lower undoped barrier layer 253 comprises 1000 nm of undoped Al@ Ga1-x As. Superlattice 255 is then formed overlying and in contact with said

undoped barrier ayer 253. Superlattice 255 comprises 500 nm of undoped GaAs (2 nm)/AI, Gal-As (2 nm). where x=0. 3.

Superlattice 255 is provided in order to prevent contaminants from the growth surface from diffusing through the structure and affecting the quality of the quantum 'Well. Quantum we) ! layer 257 is then provided overlying and in contact with said superiattice 255. Quantum well layer 257 comprises 30 nm of undoped GaAs.

Spacer layer 259 is then provided overlying and in contact with said quantum well layer 257. Spacer layer 259 comprises 60 nm of undoped Al@ Ga1-@ As. Doped barrier layer 261 is then provided overlying and in contact with said spacer layer 259.

Doped barrier layer 261 comprises 2000 nm of n-type Al@ Ga1-@ As. doped Si 1x1017 cm@ 2. A cap laver 263 is then formed overlying and in contact with said doped barrier layer 261. Cap layer 261 comprises 10 nm of undoped GaAs.

The device is processed in a similar manner to that described with reference to Figure 9. However. there is no need to make contact to the buried doped layer. Instead. a transparent gate comprising indium tin oxide is evaporated from the top surface of the device between the source and drain contact 217 and 219.

This gate 265 forms a Schottky gate. Application of a bias to this gate with respect to the drain contact causes the carrier concentration of Quantum well ayer 257 to change. The gate 265 can be contacted by Au contact metal.

The device operates in the same manner as described with reference to Figure 9 except for the fact that Schottky 265 is used to induce carriers into quantum well layer 257.

Claims

CLAIMS: 1. A device comprising: a reservoir of excess carriers having mobility in at least one direction: said reservoir ha\ ing a variation in at least one of its characteristics in said at least one direction: generating means for generating electron/hole pairs in the reservoir : and means for applying an electric field across at least a part of the reservoir for moving recombining carriers in said at least one direction.

2. A device according to claim 1. wherein the generating means comprises means c to electrically inject carriers of an opposing conductivity type to the excess carriers into the reservoir.

3. A device according to any preceding claim. wherein the generating means comprises means to inject both electrons and holes into the reser\oir.

4. A device according to claim 3. comprising a p-i-n structure and said reservoir being located in the intrinsic region.

5. A device according to any preceding claim. wherein the generating means comprises means to illuminate the device such that electron/hole pairs are excited in the reservoir.

6. A device according to any preceding claim. wherein the generating means comprises means to illuminate the device such that minority charge carriers are collected in the reservoir.

7. A device according to any preceding claim, wherein a first waveguided region is provided in a part of the reservoir such that photons generated in this part of the reservoir are guided out of the device by the first waveguided region.

<Desc/Clms Page number 26>

8. A device according to claim 7. wherein the size of the waveguide in said at least one direction is. of the order of. or less than the drift length for recombining carriers in the reservoir in said at least one direction.

9. A device according to either of claims 7 or 8. wherein the size of the va\ guide in said at least one direction is at most 5 u. m.

10. A device according to any of claims 7 to 9. comprising a plurality of

waveguided regions. the means to apply an electric field being configured such that L-r~l recombining carriers may be moved between the waveguided regions.

11. A device according to claim 10. wherein the waveguided regions ha\ e a separation of from 0.5 um to 20 um in the at least one direction.

12. A device according to either of claims 10 or 11. wherein the waveguided regions have a separation. of the order of. or less than the drift length for recombining carriers in the reservoir in the said at ieast one direction.

13. A device according to any of claims 7 to 12. wherein the waveguided region is provided by a stripe-type waveguide.

14. A device according to any preceding claim. wherein there is a variation in the composition of the reservoir in said at least one direction.

15. A device according to any preceding claim. wherein there is a variation in the size of the reservoir in said at least one direction.

16. A device according to any preceding claim. wherein there is a variation in the radiative or non-radiative lifetime of the material of the reservoir in said at least one direction.

<Desc/Clms Page number 27>

17. A device according to claim 16. wherein the said variation in the radiative or non-radiative lifetime is produced by implanting parts of the reservoir with ions or dopant ions.

18. A device according to either of claims 16 or 17 when dependent on an : - of claims 7 to 13. wherein ions or dopant ions are provided in the reservoir outside the waveguided region.

19. A device according to any preceding claim. wherein the means for applying an electric field comprises a source and a drain ohmic contact to said reservoir.

20. A device comprising: a reservoir having an excess of carriers which have mobility in at least one direction: a source and a drain contact provided to said reservoir such that a current can be measured along said at least one direction: and a switchable optical source configured to illuminate said reservoir with radiation sufficient to optically excite a number of electron/hole pairs within the device. wherein the number of electron/hole pairs is of the order of the number of excess carriers.

21. A device according to claim 20. wherein the number of excited electron/hole pairs is from 0. 1 to 10 times the number of excess carriers.

21. A device according to either of claims 19 or 20. wherein the device is configured such that the optically excited electron/hole pairs are collected in the reservoir.

23. A device according to either of claims 19 or 20. wherein the device is configured such that only the minority carriers from the optically activated eiectron/hole pairs are collected in the reservoir.

24. A device according to any of claims 20 to 23. wherein the excess carriers are located in two dimensional carrier gas and the areal carrier density is from 0. 1/ (7t r2) to

<Desc/Clms Page number 28>

10/ (7r r2). where r is the Bohr radius of the bound state formed between recombining carriers.

25. A device according to an\ : of claims 20 to 24. wherein said source is a pulsed source.

26. A device according to claim 25. wherein the pulse \\idth of the pulsed source is less than the radiative lifetime of the recombining carriers in the reservoir.

27. A device according to an : preceding claim. wherein the reservoir is a quantum we) ! or heterointerface configured to support a two dimensional carrier gas.

28. A device according to an : preceding claim. further comprising means to adjust the number of excess carriers in the reservoir.

29. A device according to claim 28. comprising a doped barrier layer pro\ ided to supply excess carriers to the reservoir and means to control a field applied across said barrier layer with respect to the reservoir region.

30. A device comprising: a reservoir of excess carriers having mobility in at least one direction: generating means for generating electron/hole pairs in the reservoir: means for applying an electric field across at least a part of the reservoir for moving recombining carriers in said at least one direction. wherein the device has a variation in at least one of its characteristics in said at least one direction.

31. A method of operating an optical device. the device comprising: a reservoir of excess carriers having mobility in at least one direction. said reservoir having a variation in at least one of its characteristics in said at least one direction. the method comprising: generating electron/hole pairs in the reservoir and applying an electric field across at least a part of the reservoir for moving recombining carriers in said at least one direction.

<Desc/Clms Page number 29>

32. A method of operating a device according to claim 31. the device further comprising first and second waveguided regions in said reservoir. the method comprising irradiating the device so that optically excited carriers are collected in the first waveguided region and moving at least some of the recombining carriers in the first

, A, ave,, u ion. waveguided region to the second waveguided region.

33. A method of optically operating an electrical switch. the switch comprising a A method of opt'catIN I- reservoir having an excess of carriers which have mobility in at least one direction and source and drain contacts provided to said reservoir, such that a current can be measured along said at least one direction. the method comprising varying the illumination level of the reservoir up to a level which results in the generation of a number ofclectron/hole pairs of the order of the number of excess carriers.

34. A method for fabricating a device. the method comprising : forming a reservoir region configured to support excess carriers such that they have mobility in at least one direction: forming a doped region configured to supply excess carriers to said reservoir region : providing means to create electron hole pairs in said reservoir region : forming ohmic contacts to said reservoir region : and forming a variation in the characteristics of said reservoir region in said at least one direction.

35. A method according to claim 34. wherein the step of providing means to create electron/hole pairs in the reservoir comprises forming a heavily doped laver having an electrical contact. the layer being configured such that upon electrical activation. carriers having an opposing conductivity type to those of the excess carriers are injected into the reservoir region.

36. A method according to claim 34. wherein the step of providing means to create electron/hole pairs comprises providing a p-type region on one side of the reservoir region and an n-type region on the opposing side of the reservoir region to the p-type

<Desc/Clms Page number 30>

region. the n and p-type regions being configured such that upon electrical activation.

electrons and holes are injected into the reservoir region. t 37. A method according to any ; of claims 34 to 36. wherein the step of presiding a variation in the reservoir region comprises a step of varying the composition or thickness of the reservoir region as the reservoir region is formed.

18.

38. A method according to any of claims 34 to 37. wherein the step ofpro\ iding a variation in the reservoir region comprises providing a local variation in the thickness of the layers surrounding the reservoir region.

19.

39. A device as substantially hereinbefore described with reference to any of the accompanying figures.

40. A method of operating a device as substantial hereinbefore described with reference to any of the accompanying figures.

41. A method of fabricating a device as substantial hereinbefore described with reference to any of the accompanying figures.