GB2307304A

GB2307304A - Optical semiconductor device

Info

Publication number: GB2307304A
Application number: GB9523464A
Authority: GB
Inventors: Andrew James Shields
Original assignee: Toshiba Cambridge Research Centre Ltd
Current assignee: Toshiba Europe Ltd
Priority date: 1995-11-16
Filing date: 1995-11-16
Publication date: 1997-05-21
Anticipated expiration: 2015-11-16
Also published as: GB2307304B; JPH09230295A; GB9523464D0

Abstract

An optical semiconductor device 11 comprises an optical path region 13 for transmission of an optical beam 24 therethrough, along an optical path 23. The optical path region 13 comprises mutually separated first and second quantum well layers 35, 37. If the device is electrically operable, at least two electrical contacts 19, 25, 31 are provided for altering an optical property of the optical path 23. Alternatively, the device may be optically operable. The device may be configured e.g. as an amplitude or phase modulator or an optical switch.

Description

OPTICAL DEVICE The present invention relates to an optical device in which the optical properties of that device are changed by application of an optical or electrical signal to the device. The invention especially, although not exclusively, manifests itself as an electrically operable optical switch or an optical modulator capable of producing amplitude modulation or phase modulation of an optical beam.

Optical devices normally achieve amplitude modulation by changing the absorption of an optical beam, and phase modulation, by changing the refractive index of the optical path through the device. A typical optical modulator is described in US-A4 872744. This kind of device requires carriers to flow in and out of a quantum well layer via ohmic contacts over a distance which typically is at least several microns. This inherently limits the speed of operation of the device.

A new form of device has now been devised which considerably improves the speed of operation relative to the conventional kind of device.

Thus, the present invention now provides an optical semiconductor device comprising an optical path region for transmission of an optical beam therethrough, along an optical path, the optical path region comprising mutually separated first and second quantum well layers.

In those devices which are electrically operable, at least two electrical contacts are provided for altering an optical property of the optical path.

For the avoidance of doubt, the devices according to the present invention are not limited to operation with electromagnetic radiation of only visible wavelengths but may, for example, operate at ultra-violet, near-, mid- or far-infrared wavelengths or even upper microwave wavelengths. Therefore, references herein to an "optical" beam or to "light" are to be interpreted accordingly.

The device of the present invention can be used as a discrete device or integrated into a larger optical and/or electrical circuit. It can form an optical logic element when combined with other components.

As indicated above, the present invention relates to a wide range of optically or electrically operable optical devices, including modulators and switches. These operate at very high speeds compared with the conventional form of modulator or switch.

Any modulator according to the present invention offers the ability to produce amplitude modulation or phase modulation or switching of the transmitted optical beam and operates at very high speed compared to the conventional device. Although not wishing to be bound by any particular explanation or theory, the applicants believe that the operation of the device according to the present invention is as described as follows.

Figure 1 of the accompanying drawings shows the spatial variation of the conduction and valence band edges along the direction normal to the layers within the optical path region, for one example embodiment of the device. For this example, the upper quantum well layer 1 has been chosen to be narrower than the lower quantum well layer 3. Dopants are arranged only in an upper barrier. An undoped lower barrier is situated below the lower quantum well layer 3.

In state 1 (on the left hand side of Figure 1), the front and back gate biases are such that the electron density in the upper quantum well layer 1 (no1) is finite, while that in the lower quantum well layer 3 is negligible (N1 14). Raising the front gate bias and simultaneously lowering the back gate bias, transfers electrons from the upper to the lower quantum well layer, so that they then have densities of NU2 and N12, respectively. The front and back gate biases are changed so as to keep the total electron density constant, i.e. Nul + N11 =Nu2+N12.

The light propagating in the optical path region is chosen to have a photon energy near the bandedge transition of the wider quantum well. Since the absorption threshold of the narrower quantum well layer lies to higher photon energy, it will be essentially transparent to the light. The absorption of the wider well layer on the other hand will change considerably when electrons are transferred to it from the narrower well by changing the voltages applied to the front and the back gates of the structure. This in turn causes the desired change in the amount of light transmitted along the optical path.

A preferred embodiment of a device of the present invention described hereinbelow works on the principle of intensity modulation, where the absorption of the waveguide is changed to modulate the optical beam. This intensity modulation can be achieved in two different modes of operation, either using a photon energy close to that of the bandedge exciton of the wider quantum well, or, alternatively, at a lower photon energy.

The preferred embodiment is actually an electrically operable device in which the two quantum well layers are separated by a middle barrier region.

The optical region may then comprise a first barrier region including a doped layer. The at least two electrical contacts may comprise an ohmic contact in contact with the optical path region. Respective upper and lower gates may be provided above and below the optical path region.

In use of such an electrically operable device, voltages are applied between the ohmic contact and the front gate (Vfront), as well as between the ohmic contact and the back gate (Vback). Changing Vfront and Vback causes a change in the intensity of the light beam transmitted by the waveguide, as discussed below. To operate as a modulator, the values of Vfront and Vback are changed so that the optical path is switched between more transparent and more opaque states.

Varying the applied voltages (Vfront and Vback) causes excess electrons to be transferred from one quantum well to the other, which changes the transparency of the waveguide. For high speed operation, the front and back gate voltages are varied so as to maintain the same total electron density. In one example embodiment the two quantum well layers in the optical path region are chosen to have different well widths, so that their absorption thresholds occur at different photon energies. The transition threshold of either quantum well layer is very sensitive to the density of excess electrons in the quantum well layer. Therefore, if the light propagating through the structure is chosen to have a photon energy near the energy of the absorption threshold of one well layer, then transferring electrons to it from the other well layer will radically alter the amount of light it absorbs.

The fast operation of such an electrically operable device according to the present invention is probably due to the fact that electrons are transferred from one quantum well layer to the other upon changing the biases and do not flow in and out of the ohmic contact. Hence the change in Vfront and Vback should be chosen so that the overall electron density in the two quantum wells remains constant. It is also important that the two biases are changed simultaneously. The high speed of the device derives, therefore, from the fact that the electrons must only tunnel across a relatively narrow barrier layer between the quantum well layers, which is preferably less than 100 nm thick.

Figure 2 of the accompanying drawings shows the absorption profile of the wider quantum well layer with zero and finite electron density in the quantum well. In the first mode, when the photon energy is chosen near that of the bandedge neutral exciton of the wider quantum well layer, it will be highly absorbing when depleted of electrons, and hence the transmission of the waveguide will be relatively small. Adding excess electrons to the wider quantum well layer reduces its absorption due to phase space filling and screening effects, turning the device more transparent.

The device can alternatively be operated at a lower photon energy. The applicants have found that adding excess electrons to a quantum well layer increases its absorption at an energy just below the energy of the bandedge exciton, as shown schematically in Figure 3 of the accompanying drawings.

This is probably due to formation of the negatively-charged exciton in the presence of an excess electron density in the quantum well.

The negatively-charged exciton can be thought of as an excess electron bound to a neutral exciton and has a slightly lower transition energy than the neutral exciton. Hence with a negligible number of excess electrons in the quantum well, the transition due to the negatively-charged exciton is very weak.

However, the negatively-charged exciton transition strengthens considerably when excess electrons are added to the quantum well.

For the second mode of operation one should choose a photon energy less than that of the bandedge neutral exciton of the wider well and close to the photon energy of its negatively-charged exciton transition. In this case the absorption of the quantum well will be relatively small when it is depleted of electrons (solid line in Figure 3), since the photon energy lies below that of the bandedge neutral exciton. In contrast to the first mode of operation, the state with the wider quantum well depleted of electrons now corresponds to the transparent state of the device. When excess electrons are added to the quantum well, absorption due to the negatively-charged exciton increases (dashed line in Figure 3) and the waveguide turns more opaque to the waveguided slight.

For the first mode of operation, the applied front and back gate voltages are chosen to give the maximum reduction in the neutral exciton absorption.

On the other hand, for the second mode of operation, the applied front and back gate biases are chosen so that there is maximum increase in the absorption of the charged exciton transition. This is achieved with a finite electron density typically smaller than for the first mode of operation. Although the optical path region comprises first and second mutually separated quantum well layers, the present invention includes devices having repeat units of such pairs of quantum well layers. Moreover, the term "quantum well layer" includes the situation where carriers are induced adjacent to an interface (heterojunction) in a heterostructure, rather than providing a discrete (separate) layer especially for confinement of a quantum well.

Maximum change in the absorption of the wider quantum well will be achieved by switching its electron density between a negligible and finite value, although of course the device could also operate by switching its density between two finite values.

In the example of Figure 1, the upper quantum well is chosen to be narrower than the lower one and the dopants are arranged in the upper barrier only. However, several other arrangements are possible, some of which are shown in Figure 4 of the accompanying drawings. Figure 4A corresponds to the arrangement shown in Figure 1. In Figure 4B the upper well is chosen as the wider and the dopants are placed in the lower barrier. In Figure 4C the upper well is the narrower and the dopants have been placed in both barriers.

Other arrangements are also possible. For instance the lower barrier region could be omitted entirely. In this case electrons will be confined at the interface between the middle barrier and the lower-bandgap layer immediately below it. They can then be transferred from this interface to the upper quantum well.

In Figure 4 the two quantum well layers are chosen to have different thicknesses so that they have different bandgap energies. However, this can also be achieved by forming the two layers out of different materials. For instance, for the example presented below, where the quantum well layers are GaAs and the barriers AlxGa1 xAs, one of the quantum well layers could alternatively be made of AlwGai.wAs where 0 < w < x (which has a larger bandgap than GaAs), or InvGal vAs (which has a lower bandgap than GaAs).

Again, there are several arrangements for the order of the lower bandgap quantum well layer, higher bandgap quantum well layer and doped barrier layer or layers.

The device could even operate if the quantum well layers were chosen to have the same thickness and be made of the same material. In this case, with reference to Fig. 1, the carrier density of the lower quantum well layer in state 1 should not equal that of the upper quantum well layer in state 2. For cases where the quantum well layers have the same thickness, it is nevertheless preferred to form them of semiconductor materials of respective different bandgaps.

A thin layer of wider bandgap material could be added immediately below the lower quantum well and immediately above the upper quantum well so as to provide extra confinement for the electrons. For the example GaAs/AlxGal-xAs heterostructure, these wider bandgap layers could be AlAs, e.g. 3 nm thick.

Some or all of the barrier layers can also be replaced with a short period superlattice or higher bandgap than the quantum well layers. For the example GaAs/AlxGal.xAs heterostructure, such a short period superlattice can be formed from alternating layers of GaAs and AIwGal.wAs e.g, each about 2.5 nm thick. This improves the quality of the layer interfaces.

The barrier between the two quantum wells could be made of a material which has a bandgap intermediate between that of the quantum well and the upper and lower barriers, so as to reduce the tunnelling time between them. For instance, for the example GaAs/AlxGa1 xAs heterostructure, the barrier could be AlyGal yAs where y is smaller than x. The tunnelling time could also be reduced by grading the Al concentration of the barrier.

In one variant of the preferred embodiment, instead of placing the doped layer in the first barrier region, the dopant may be placed in one of the quantum well layers.

The aforementioned preferred embodiment utilises a stripe waveguide structure and the optical path region is located between respective cladding layers. However, conventional waveguide structures could also be employed.

Alternatively, the cladding layers could be dispensed with and the optical beam be incident normal, or at an oblique angle to, the device surface.

In this case the beam reflected or transmitted by the optical path region is modulated by the transfer of charge between quantum well layers. In that case, a window should be etched in the substrate if it is absorbing at the energy of the optical beam, and the transmitted beam is to be detected.

With a device according to the present invention, amplitude modulation involves changing the absorption coefficient of the optical path region.

However, the same principles can be employed to achieve phase modulation, where the real part of the refractive index of the optical path region is modified, which in turn changes the phase of an optical wave travelling through the optical path region. In this case it is better to use a photon energy lower than the bandgap of the quantum well layers, so that the light travelling through the optical path region is not significantly attenuated. The device will also function as a phase modulator, because transferring charge between the quantum well layers will modify the real part of the refractive index at photon energies smaller than the bandgaps of the quantum well layers.

The transfer of charge between the two wells can also be used to achieve amplitude modulation for light of energy much less than the bandgaps of the quantum well layers. In this case the light, which is in the mid-infra-red, farinfra-red or microwave region of the optical spectrum, causes transitions within the conduction or valence bands of the optical region. Transferring the charge carriers between the quantum wells will alter the transparency of the optical region to this radiation.

In the preferred embodiment, one or more of the barrier regions within the optical path region is doped n-type, so as to provide an excess density of electrons in the quantum well layers. However, such a device could also operate by replacing the n-type doped layer(s) in the barrier region(s) with ptype doped layer(s). In this case the quantum well layers would contain excess holes. The device would then operate by transferring excess holes from one quantum well layer to the other. For the second mode of operation discussed above, positively charge excitons are formed in the presence of an excess hole density in the quantum well layer. This can be thought of an excess hole bound to a neutral exciton. Adding excess holes to the quantum well layer strengthens the intensity of the transition due to the positively charged exciton.This is analogous to the strengthening of the negatively charged exciton transition when excess electrons are added to the well, shown in Figure 3.

In the preferred embodiment, the charge is transferred between the quantum well layers by applying a first voltage between a front gate and the ohmic contact and a second voltage between the back gate and the ohmic contact. However, other means of transferring the charge can be envisaged.

For instance, the charge transfer could be effected by a second optical beam incident on the structure.

The present invention is not limited to GaAs/AlxGal xAs systems.

Other quantum well heterostructures may also be used, for example GaAs/AlAs, InxGa1 xAs/AlyGal yAs, InxGal xAs/InP, GaxInl xP/(AlyGal y)zInl zP,Znl ZhlxCdxSeZnSe, ZnSe/Zn 1 xMgxSyse 1 y, Znl CdxSe/ZnSySel y CdTe/Cd1 .xZnxTe, GaN/AlN, GaN/AlGal xN, InxGal .xN/GaN, InxGal xN/AlyGa 1 yN.

The doped layer or layers may be doped to be n-type to provide excess electrons in the quantum well layers, or it may be doped ptype to provide excess holes.

Conveniently, a recess may be formed in the optical path region to provide an electrical contact to the gate below the lower quantum well layer.

Alternatively such a back gate layer can be grown on a highly doped substrate and buffer layers of the same doping type and ohmic contact made to the back face of the substrate.

The preferred embodiment also provides the said at least two electrical contacts with an ohmic contact to the optical path region, however a plurality of such ohmic contacts could be made to the optical path region, e.g. one either side of the optical path.

The present invention will now be explained in more detail by the following non-limiting description of a preferred embodiment and with reference to the accompanying drawings, in which: Figure 1 shows energy band diagrams for explaining the basic principles of operation of modulators according to the present invention; Figure 2 shows absorption spectra for explaining a first mode of operation of modulators according to the present invention; Figure 3 shows absorption spectra for explaining a second mode of operation of modulators according to the present invention; Figure 4 shows energy band diagrams for explaining the principles of operation of different embodiments of modulators according to the present invention; Figure 5 shows a sectional perspective view of a preferred embodiment of a modulator according to the present invention;; Figure 6 shows details of the optical path region of the modulator shown in Figure 5; and Figure 7 shows details of the complete layer structure of the device shown in Figures 5 and 6.

Figures 5-7 show a preferred embodiment of an optical modulator 11 according to the present invention. As seen, in particular, from Figure 5, the modulator 11 comprises an optical path region 13 situated between an upper cladding layer 15 and a lower cladding layer 17. A lower (back) gate layer 19 is situated below the lower cladding layer 17.

The upper cladding layer 15 is etched to form a ridge 21 which constitutes a stripe waveguide for defining an optical path 23 through the optical path region 13, for transmission of a beam 24 underneath the ridge 21 and in the plane of the layers. An upper Schottky metal front gate 25 is formed on top of the ridge 21. The back gate layer 19 is contacted by an ohmic contact 27 through the lower cladding layer 17. An electrical connection pad 29 is provided for this ohmic contact 27, within a recess formed in the optical path region 13 and the upper cladding layer 15.

Another ohmic contact 31 is also provided to make contact to the optical path region 13, and in particular to quantum well layers thereof (to be described further hereinbelow). This ohmic contact 31 has a contact pad 33 to one side of the ridge 21. Optionally, an ohmic contact can be made on both sides of the ridge.

Details of the optical path region 13 are shown in Figure 6. It comprises an upper quantum well layer 35 and a lower quantum well layer 37 separated by a middle barrier region 39. An upper barrier region 41 is situated above the upper quantum well layer 35. A lower barrier region 43 is situated below the lower quantum well layer 37.

The lower barrier region 43 is undoped. The upper barrier region 41 comprises an undoped spacer layer 45 in contact with the upper quantum well layer 35. Above the undoped spacer layer is situated a doped upper barrier layer 47, and above this, an undoped upper barrier layer 49, thus completing the upper barrier region 41.

Complete details of the layer structure of this embodiment are shown in Figure 7, based on a GaAs/AlxGa1 xAs/AlyGal yAs heterostructure. The layers, grown by molecular beam epitaxy on a (100)-oriented GaAs substrate, comprise (in order of growth): GaAs 1 Sun buffer and substrate layers 51 Al0.5Ga0.5As 1CLm n-type (1018an~3) back gate 19 Al0.5Ga0.5As 1.0 llm lower cladding layer 17 Ab0.33Ga0.67As 0.3 llm lower barrier 43 GaAs 20 nm lower quantum well layer 37 Al0.33Ga0.67As 2 nm middle barrier 39 GaAs 15 nm upper quantum well layer 35 A%.33Gao.67As 60 nm upper barrier (spacer) 45 Ab0.33Ga0.67As 200 nm n-type (1017cm.3) upper barrier (doped) 47 Al0.33Ga0.67As 23 nm upper barrier 49 Al0.5Ga0.5As 0.5 Rm upper cladding layer 15 GaAs 1 nm cap layer53 The stripe 21 is etched on the top surface of the structure using conventional techniques to have a width of 1 - 5 Cun and to leave around 0.2 Stm of the upper cladding 15 in the etched region.

The front gate 25 is formed by evaporating 20 nm NiCr, followed by 40 nm Au on top of the stripe region 21. Alternatively, the metal gate can be replaced with a highly doped GaAs or AlxGal xAs layer to which an ohmic contact is made. The upper cladding layer can optionally be replaced with a dielectric material of lower refractive index than the guided region.

The "shallow" ohmic contact 31 made to the electrons in the quantum well layers 35, 37 avoids shorting to the highly doped back gate layer 19. This is done by etching the contact region to within around 10 nm of the quantum well layers 35, 37 (not shown in Figure 5) and then annealing a Pd-Ge contact at 350 C for 5 minutes, as described in Patel et al, Appl. Phys. Lett. 65, 851 (1994).

An alternative method of avoiding a short circuit between the ohmic contact 31 and the back gate 19 is to damage the region of the back gate 19 underneath the ohmic contact using an in-situ ion-gun so that it is insulating, as described by Linfield et al, Semicond. Sci. Technol. 8,415 (1993).

The reason why ohmic contact 27 to the back gate layer 19 is recessed below the depth of the quantum well layers 35, 37 is to avoid a shorting contact to the quantum wells. A suitable ohmic contact metal to the n-type back gate is Ni-Pd-Au.

In the light of this disclosure, modifications of the described embodiment as well as other embodiments, all within the scope of the present invention as defined by the appended claims, will now become apparent to persons skilled in the art.

Claims

1. An optical semiconductor device comprising an optical path region for transmission of an optical beam therethrough, along an optical path, the optical path region comprising mutually separated first and second quantum well layers.

2. A device according to claim 1, wherein the first and second quantum well layers are separated by a middle barrier region.

3. A device according to either preceding claim, wherein the first and second quantum well layers have substantially the same thickness as each other.

4. A device according to any preceding claim, wherein the first and second quantum well layers are made from semiconductor materials having respective different forbidden bandgaps.

5. A device according to claim 1 or claim 2, wherein the first and second quantum well layers have respective different thicknesses.

6. A device according to any preceding claim, wherein at least two electrical contacts are provided for altering an optical property of the optical path.

7. A device according to claim 6, the optical path region further comprises a first barrier region including a doped layer, the at least two electrical contacts comprising an ohmic contact in contact with the optical path region and respective upper and lower gates above and below the optical path region.

8. A device according to claim 7, wherein the quantum well layer nearest to the doped layer is thinner than the other quantum well layer.

9. A device according to claim 7 or claim 8, the optical path region further comprising a second barrier region so that both quantum well layers are disposed between the first and second barrier regions.

10. A device according to claim 9 when dependent upon claim 7, wherein the second barrier region also contains a doped layer.

11. A device according to any of claims 7 to 10, wherein a respective layer of semiconductor material having a higher bandgap than the bandgap of the first and (if present) second barrier region is provided immediately above the upper of the first and second quantum well layers and immediately below the lower of the first and second quantum well layers.

12. A device according to any of claims 7 to 11, wherein the optical path region is disposed between respective cladding layers and the optical path is defined by a stripe waveguide structure.

13. A device according to any of claims 7 to 12, wherein a recess is formed in the optical path region to provide access to an electrical contact to one of the gates.

14. An optical semiconductor device substantially as hereinbefore described with reference to any of the accompanying drawings.

14. A device according to any of claims 7 to 12, wherein the lower gate is formed on a highly doped substrate, buffer layers being provided with the same conductivity type as said substrate, and an ohmic contact being provided to the reverse side of said substrate.

15. A device according to any of claims 1 to 5, further comprising means for irradiating the device with an optical control beam for altering an optical property of the optical path.

16. An optical semiconductor device substantially as hereinbefore described with reference to any of the accompanying drawings.

Amendments to the claims have been filed as follows 1. An optical semiconductor device comprising an optical path region for transmission of an optical beam therethrough, along an optical path, the optical path region comprising mutually separated first and second quantum well layers and a first barrier region including a doped layer, at least two electrical contacts are provided for altering an optical property of the optical path, the at least two electrical contacts comprising an ohmic contact in contact with the optical path region and respective upper and lower gates above and below the optical path region.

6. A device according to claim 1, wherein the quantum well layer nearest to the doped layer is thinner than the other quantum well layer.

7. A device according to claim 1 or claim 6, the optical path region further comprising a second barrier region so that both quantum well layers are disposed between the first and second barrier regions.

8. A device according to claim 7 when dependent upon claim 1, wherein the second barrier region also contains a doped layer.

9. A device according to any of claims 1 to 8, wherein a respective layer of semiconductor material having a higher bandgap than the bandgap of the first and (if present) second barrier region is provided immediately above the upper of the first and second quantum well layers and immediately below the lower of the first and second quantum well layers.

10. A device according to any of claims 1 to 9, wherein the optical path region is disposed between respective cladding layers and the optical path is defined by a stripe waveguide structure.

11. A device according to any of claims 1 to 10, wherein a recess is formed in the optical path region to provide access to an electrical contact to one of the gates.

12. A device according to any of claims 1 to 10, wherein the lower gate is formed on a highly doped substrate, buffer layers being provided with the same conductivity type as said substrate, and an ohmic contact being provided to the reverse side of said substrate.

13. A device according to any of claims 1 to 5, further comprising means for irradiating the device with an optical control beam for altering an optical property of the optical path.