GB2204961A - Quantum well electro-optic device - Google Patents

Abstract

An electro-optic device includes a multiple quantum well region 28 comprising quantum well layers 38, 40, 42, undoped cladding layers of GaAs, 12 and doped cladding layers. These are arranged on a AlAs layer and a GaAs substrate disposed on a GaAs prism. A Schottky contact 32 is arranged on the quantum well region. The device is biasable by means of the Schottky contact and an ohmic contact. The doped layers provides modulation doping of the quantum well layers which are absorptive or transmissive according to whether these layers are electron populated or depleted. The depleted condition is obtained by a bias voltage applied across the contacts. Light passing via the prism to the quantum layer is accordingly modulatable, by varying the bias voltage. Alternatively, using a different prism plasmon generation can be controlled by the bias voltage. incident light being absorbed by the plasmon generation. An alternative detector embodiment which has a coupling mechanism is a diffraction grating disposed on the Schottky contact. <IMAGE>

Description

ELECTRO#PTIC DEVICE This invention relates to an electrorptic device of semiconductor material, and more particularly a device of the kind including a quantum well region and quantum well populating means.

Semiconductor electro-optic devices having populated quantum wells are known, and include modulations and detectors.

A quantum well detector is theoretically described by Goossen et al in Appl Phys Lett 47 (12), (15 December 1985). This comprises a GaAs quantum well layer bounded by a AlxGal~xAs layer and a AlyGai yAs layer, and is situated on a GaAs substrate. A copper diffraction grating is situated above the quantum well layer and substrate. Goossen et al indicate that in operation radiation will be incident substantially normal to the diffraction grating, via the substrate and AIxGa1~xAs layer. This radiation is scattered by the grating to produce an component of the electric vector perpendicular to the quantum well plane, which is consequently absorbed. Goosen et al suggest that the quantum absorption effect might be used to provide an optical detector, they do not however disclose any means for achieving this.Moreover, no details of device composition (x and y) are explicitly given. In addition, because the Goosen device scatters the input radiation, it is not capable of producing a single strong output beam. It is therefore unsuitable for use as a modulator.

A Kastalsky et al in Appl Phys Lett 50 (12) (23 March 1987) describes a prior art electrooptic modulator for operation at a wavelength in the range 0.7 to 0.8 ym. This device is a single quantum well structure comprising a GaAs layer, disposed between two Al0#27GaAs layers, and separated from those layers by respective AlAs layers. The quantum well structure is mounted on an A10.45GaAs layer and is surmounted by a metallised rib acting as a Schottky contact. In addition, the Al0#276aAs layers are designed to support a single optical mode at the operating wavelength, and the A10,45GaAs layer acts as a waveguide cladding layer. Furthermore, the metallised rib serves to confine the optical mode laterally.In effect, the quantum well is within a waveguide.

In operation, applying a potential to the Schottky contact serves to populate the quantum well of electrons, the well being electron depleted at zero bias. When the quantum well is depleted, it becomes receptive to electrons from the valence band. This in turn creates holes in the valence band, and each electron-hole pair forms an exciton. Radiation at the operating wavelength is input into the device and confined within the region of the quantum well by the waveguide action of the quantum well structure. The operating wavelength is equal to the exciton resonance wavelength, and is consequently absorbed. In the absence of a potential on the Schotty contact, the quantum wells are electron populated and cannot accept further electrons from the valence band. Consequently, excitons cannot be formed.Additionally, the electrostatic attraction binding an existing exciton is reduced by virtue of the screening effect of the quantum well electron population. Ultimately, the electron and hole become unbound, destroying the exciton. Accordingly, input radiation absorption is negligible when the quantum well is populated. Input radiation modulation is therefore provided by Schottky contact potential modulation.

The Kastalsky et al device suffers from the disadvantage of requiring a waveguide to couple input radiation to the quantum well. Such a coupling scheme presents difficulties when extended to the Construction of two dimensional arrays of devices, in which each device is addressed individually. It requires an undesirably complex multiple waveguide structure. No other radiation coupling means appears to be viable because of the opacity of GaAs at the operating wavelength.

Furthermore, device operation is restricted to the exciton resonance wavelength.

D S Chemla et al in Appl Phys Lett 50(10), (9 March 1987) describe an electro-optic modulator for operation at a wavelength of 1.3 to 1.6 Clam. This device comprises a heterostructure of AlInAs and GaInAs on a InP substrate. A quantum well layer of In0#53Ga0#47As is located between a In0#52AI0#48As spacer layer and an In0.52A10.48As buffer layer, which surmounts the InP substrate. As before, absorption is achieved by exciton resonance within the quantum well and modulation by applying a potential to a Schottky contact located above the well.

Radiation is coupled to the quantum well by directing radiation at normal incidence to the InP substrate, which is transparent at the operating wavelength.

Moreover, the absorption mechanism is not greatly sensitive to angle of incidence which may therefore vary. This device has a range of operation restricted to 1.3 to 1.6 Fm because of the wavelength dependence of the exciton resonance. It exhibits an absorption change of 2%, which is very small compared with conventional Pockels or Kerr cell modulators.

A different electro-optic modulator is described by Harwit et al in Appl Phys Lett 50(11), (16 March 1987). This has a wavelengths range of 8 to 16 Fm.

This device consists of a GaAs and AlGaAs multilayered structure including fifty quantum wells. Each quantum well has a central GaAs layer bounded by silicon doped Al0#5Ga0#5As layers. The quantum wells are mounted on a GaAs substrate, and are surmounted by a silicon doped GaAs layer, having a NiAuGe ohmic layer above. In operation, polarised radiation is incident on the device at the Brewster angle to achieve an electric vector perpendicular to the quantum well planes. Absorption then occurs by virtue of electric field coupling to inter-sub-band transitions within the quantum wells. Absorption is achieved by applying a potential to the ohmic contact, which produces a Stark shifts in the transition wavelength.

A disadvantage of this device is that a large number of quantum wells are needed to produce significant absorption. This is because the radiation is only weakly coupled at the Brewster angle. The device requires fifty quantum wells to provide only 5% absorption.

It is an object of this invention to provide an electro-optic device which has an alternative coupling, absorption and modulation scheme, which exhibits improved absorption modulation properties and is suitable for array applications.

The present invention provides an electro-optic device including a semiconductor quantum well region containing at least one quantum well, for modulating or detecting incident radiation polarised with an electric vector component perpendicular to the plane of the quantum well, and having: (a) means for populating the quantum well; (b) either (i) means for passing a current through the quantum well region to facilitate detection of photocarriers, or (ii) means for reducing the population of the quantum well to alter absorption and provide optical modulation; and (c) means for coupling incident radiation to the device such that, either (i) when the quantum well is population reduced, plasmon creation provides optical modulation, or (il) when the quantum well is populated, quantum well intersub-band absorption provides a modulation or detection mechanism.

The invention provides the advantage of a greater modulation depth than prior art quantum well devices. Embodiments of the invention have achieved modulation depths above 90%, as opposed to a few % for the prior art.

Moreover, the invention may be used as the basis for a device array in which individual devices are easily addressable.

In a preferred embodiment, the quantum well comprises a GaAs layer bounded by AlxGa1#xAs layers, where x is between 0.25 and 0.35, and particularly 0.3.

Furthermore, the quantum well may be populated by modulation doping means comprising silicon doped regions of the AlxGa1 #xAs layers.

The quantum well region may be population reduced or receive current by means of a Schottky contact or an insulated electrode located adjacent to the quantum well. Preferably, the Schottky contact is gold and the insulated electrode platinum silicide.

Incident radiation coupling means may include a first coupling means comprising first and second regions, the first region being located between the quantum well region and the second region and having a refractive index less than that of the second region. In addition, it is preferred that the coupling means also include second means for coupling the incident radiation to the interface between the first and second region such that, when the quantum well is depleted, a surface plasmon is created at the interface between the Schottky or insulated electrode and the quantum well region. This provides an absorption mechanism for modulation of the radiation.

The coupling means may also include second means for coupling the incident radiation to the first region/second region interface at an angle less than the initial angle and such that, when the qauntum well is populated, quantum well inter-sub-band transitions are excited to provide an absorption mechanism for modulation or detection of the radiation. The second coupling means may be a prism.

In an alternative embodiment, the coupling means may be a diffraction grating adjacent to the quantum well region. The diffraction grating is preferably a two dimensional pattern of dots or squares.

In order that the invention might be more fully understood, embodiments of the invention will now be described with reference to the accompanying drawings, in which: Figure 1 is an exaggerated scale plan view of an electro-optic device of the invention configured as a modulator; Figure 2 is an exaggerated scale plan view of the Figure 1 modulator with the scale expanded vertically in the region of a quantum well layer; Figure 3 is an exaggerated scale plan view of the Figure 1 modulator with the vertical scale reduced in the region of the quantum well layer and an AlAs layer; Figure 4 shows a theoretically calculated graph of power reflection coefficient against angle of incidence for radiation incident on an interface of the Figure 1 modulator; Figure 5 is a curve of the minimum energy of the conduction band in a region of the Figure 2 quantum well layer;; Figure 6 is an exaggerated scale plan view of a modulator similar to that of Figure 3 but including an alternative form of prism; Figure 7 shows a detector of the invention employing diffraction grating coupling; Figure 8 shows a plasmon dispersion curve for the Figure 7 device; and Figure 9 shows a graph of operating wavelength versus the quantum well layer thickness of the embodiments of Figures 3, 6 and 7.

A sectional view (not to scale) of an electro-optic device 10 of the invention is shown in Figure 1, and comprises an optical modulator. The modulator 10 comprises an insulating GaAs substrate 12 mounted on a base 14 of an insulating GaAs triangular prism 16 having an isoscelese cross-section with an apex angle of 59j8 . The sides 18 and 20 of the prism 16 are coated with antireflection layers 22 and 24 respectively, which inhibit reflection at a wavelength of 10.6 ym. A layer 26 of AlAs, 9 pm thick, is grown on top of the GaAs substrate 12, forming a common interface 27. A quantum well region 28 of multilayer construction (not shown) is arranged on the AlAs layer 26.An ohmic contact 30, 500 pm wide and 1 pm thick, is mounted on the region 28 by vapour deposition of nickel-germanium-gold (NiGeAu) alloy. A gold Schottky contact 32 is deposited on the region 28 by vapour deposition, and is spaced by 500 pm from the ohmic contact 30. The contact 32 is 200 pm wide and 1 pm thick.

Referring now also to Figure 2, there is shown a plan view of the upper part of the modulator 10 in more detail. The region 28 consists of eight quantum well structures 34, of which only two are shown in full and a discontinuous region 36 represents the remainder. Each structure 34 has a central layer 38 of undoped GaAs, this being disposed between undoped cladding layers 40 of undoped AlxGa1#xAs (x = 0.3). Layers 38 and 40 are sandwiched by doped cladding layers 42 of n-type AlxGa1#xAS (x = 0.3) containing a silicon dopant concentration of 5 x 1017 atoms cm~3. The central layers 38 undoped cladding layers 40, and doped cladding layers 42 have respective thicknesses of 82 A, 100 and 100 A The quantum well structures 34 are grown using known Molecular Beam Epitaxy (MBE) techniques.

Referring to Figure 3, which is not to scale, there is shown a sectional view of the modulator 10 with the layers 26 and 28 illustrated reduced in thickness compared to Figure 1. A ray 44 represents radiation incident on the device and a ray 46 represents emergent radiation.

Referring now to Figure 4 there is shown a theoretically calculated graph 50 of the power reflection coefficient (R) for radiation at a wavelength of 10.6- pm versus angle of incidence on the interface 27 between the GaAs substrate 12 and AlAs layer 26.

Graph 50 incorporates two curves 52 and 54 depicting the variation of R with angle of incidence for a bias voltage of 30 V applied across the contacts 30 and 32, and no bias voltage, respectively. Curves 52 and 54 have minima at 56 and 58 respectively.

Referring now to Figure 3 once more, polarised radiation with an electric vector perpendicular to the plane of the quantum wells represented by the ray 44 is incident on the device 10 at normal incidence on the prism side 18. This radiation then propagates through the prism 16, and is incident on the interface 27 at an angle of 60.1 to the normal. With a bias voltage of about 30 V applied across the contacts 30 and 32, the quantum wells will be depleted, reflection from the interface 27 follows curve 52. This illustrates that reflection is small for incidence at 60.1 . The modulator 10 absorbs most of the incident radiation (10 pm wavelength radiation).If the bias voltage is switched off, then in accordance with curve 52 substantially all the incident radiation is reflected It emerges from side 20 of the prism 16, as indicated by ray 46. By switching the bias voltage on and off, the ray 46 is intensity modulated.

The principles governing operation of the modulator 10 will now be described with reference to Figure 5, in which there is schematically shown the conduction band 60 in the region of each quantum well structure 34. GaAs has a smaller bandgap than AlxGa1#xAs (x = 0.3). This variation in bandgap between GaAs and AlxGai #xAs produces a quantum well 64 in the central layer 38 of each quantum well structure 34. Within the quantum well there is a sub-band structure 66. Doping of the AlxGal~xAs component in such a structure 34 provides so called 'modulation' doping of the quantum wells, eg doped layers 42 of Figure 2. Accordingly, in the absence of any external influences, the sub-band structure 66 is populated by electrons.

A potential applied across the contacts 30 and 32 (see Figure 2) causes the quantum wells 64 of each quantum well structure 34 to be depopulated of electrons. This is due to the Schottky effect between the Schottky metal contact 32 and the quantum well layer 28.

The critical angle for the interface 27 is 59.9 . When the angle of incidence for radiation incident on the interface 27 is greater than the critical angle, an evanescent wave is formed in the layers 26 and 28. The decay length of the evanescent wave in these layers is dependent on the angle of incidence. At an incident angle of 60.1, the decay length is greater than the combined thickness of layers 26 and 28, and therefore the evanescent wave propagates without significant attenuation to contact 32. Moreover, if there are no free electrons in the quantum wells 64, then this evanescent wave excites a surface electromagnetic wave along the interface between the contact 32 and the quantum well layer 25 This wave is called a surface plasmon.The creation of a plasmon results in the absorption depicted by the minimum 56 of curve 52. when there are free electrons in the quantum wells 64, the evanescent wave cannot be formed; the surface plasmon therefore cannot be excited, and so no absorption occurs; this is depicted by curve 54, where no minimum is apparent at the angle of incidence of 60.1 . The modulator 10 exploits coupling of the incoming radiation to the surface plasmons to provide absorption. It also exploits the Schottky effect and the boundary conditions for plasmon generation to provide modulation of the incoming radiation.

Referring now to Figure 6, there is shown an alternative embodiment of the invention, in which different effects are used to provide absorption and modulation. This embodiment consists of a modulator 100 having similar elements as modulator 10, except that the prism 16 is replaced by a prism 60 with sides 62 and 64 (not to scale). Elements of modulator 100 similar to those of modulator 10 are like referenced. The prism 60 is of the same material as prism 16, but has an apex angle of 65. 2 .

The modulator 100 operates as follows. Polarised radiation with a component perpendicular to the plane of the quantum wells, represented by a ray 70, normal to the prism side 62 is incident on the interface 27 at an angle of 57.4 When no voltage is applied across the contacts 30 and 32, incident radiation zs absorbed by the modulator 100. This is represented by minimum 58 of curve 54 in graph 50 (Figure 3). When a bias voltage of 30 V is applied across the contacts 30 and 32 then insignificant absorption occurs, as shown by curve 52.

Radiation is then emergent normal to the side 64 of prism 60, as represented by a ray 72. By switching the bias voltage on and off, incident radiation 70 is reflected and attenuated respectively.

The principles governing the operation of tha modulator 100 are as follows. The angle of incidence of 57.4. for radiation incident on the interface 27 is within 3 of the critical angle of 59.9.. Radiation 70 incident at an angle of 57.4 is refracted to substantial parallelism with the interface 27, and propagates in the AlAs layer 26. The electric vector of the radiation in the layer 26 is substantially perpendicular to the plane of the quantum wells 64 in the quantum well layer 28, as required to excite inter-sub-band electron transitions. Such transitions are possible when the lowest sub-band is occupied by electrons, as is the case when the bias voltage is off. These transitions give rise to absorption as indicated by the minimum 58 of curve 54.When the lowest sub-band is unoccupied due to the bias voltage being on, then no transitions occur and hence no absorption, as depicted by curve 52 at an angle of incidence of 57.4 As has been mentioned, the curves 52 and 54 of Figure 3 are theoretically calculated. In practice, the minima 56 and 58 will be present, but their size and location may differ from that indicated. An equivalent graph may be obtained experimentally using known techniques. The angles at which the minima 56 and 58 occur as previously described will then be those experimentally determined, and the prisms 16 and 60 will be modified appropriately.

The embodiment 100 of the invention may also be used as detector. A bias current is passed through the quantum well layer 28 by applying an appropriate bias potential to the Schottky contact 32. Incident radiation generates photocarriers in the quantum wells which are detected as a photocurrent superimposed on the bias current.

Referring to Figure 7 there is shown a further embodiment of the invention in the form of a detector 200. The detector 200 includes a GaAs substrate 202 upon which a multiple quantum well layer 204 is grown. An ohmic contact 206 and a Schottky contact 208 are in contact with the layer 204. The parts of the modulator 200 are similar to those of the modulator 10, except that the Schottky contact 208 is made of a semimetal, such as Platinum Silicide (PtSi) and has a top surface 210 configured as a sinusoidal diffraction grating. The surface of grating 210 has a periodicity P, where 2r P g and Kg is the magnitude of the surface plasmon wavevector.

The operation of the device will now be considered using graphical means to determine Kg. Referring now also to Figure 8, there is shown a graph 230 of angular frequency W (1 04 radians sex~1) against wavevector magnitude K (106 m-l) depicting a surface plasmon dispersion curve for plasmons in the interface between contact 208 and quantum well layer 204. The detector 200 is operated at a wavelength of 10.6 pm (~ 1.9 x 1014 radians sex~1) indicated at a point 234. A chain line 236 perpendicular to the angular frequency (cm)) axis intersects the dispersion curve 232 at 238.A chain line 240 through the point 238 perpendicular to the wavevector (K) axis intersects that axis at a point 242, which corresponds to a wavevector magnitude of Kg. A shaded region 244 about line 236 represents the spectral width of the quantum well inter-sub-band transition. The region 244 intersects the curve 232 at A and B, lines drawn from 242 through A and B defined a shaded acceptance region 246. The angle between the two lines drawn from 242 define an acceptance angle. Radiation incident on the grating 210 within this angle is matched in wavevector to the surface plasmon. The acceptance angle calculated from Figure 8 is 180, ie radiation may be incident at any angle on grating 210 to cause wavevector match.If the contact 206 is formed from material with a greater conductivity than PtS, then the acceptance angle will be less than 180'.

The detector 200 operates as follows. Radiation incident on the grating 21Q propagates through it and is matched in wavevector with the surface plasmon.

Plasmons are then created which couple to electron populated quantum wells, causing generation of photocarriers. A potential on the contacts 206 and 210 generates a current through the quantum well layer 204. The photocarriers are detected as a superimposed current on the current between the contacts 206 and 210.

The detector 200 is operative for radiation of one polarisation. Its operation may be extended to two polarisations by replacing the one dimensional sinusoidal grating 210 with a grating having a two dimensional surface pattern such as dots or squares.

The Figure 7 embodiment may also be used to couple incident radiation to the quantum wells. In this case the angle of incidence (8, measured to the grating normal) of the radiation incident on grating 210, and the periodicity of the grating (P), are related by the following theoretical formula: sinS - n - X P med O where nmed is the refractive index of the layer 204 and Xo is the wavelength of the incident light in free space. Moreover, nmed GO is the wavelength of the light in the layer 204, and this is aranged to be equal to the inter-sub-band transition wavelength. when radiation is incident on the grating at an angle which satisfies the above equation, a first order diffraction beam is produced which propagates substantially parallel to the contact 210.The diffracted beam has an electric field component perpendicular to the quantum well plane, and accordingly, coupling occurs and photocarrier production.

In practice, the above formula will not hold perfectly. However, the angle of incidence may be tuned experimentally to achieve coupling.

In the embodiments previously described, the doped cladding 44 of the quantum well structures 34 may be removed and the central layer 38 doped directly.

The Schottky contact of earlier embodiments may be replaced by a metal electrode spaced by an insulating layer from the quantum well region. Such an insulated electrode provides for the required electric field induced depopulation of the quantum wells. The embodiments previously described may operate with radiation with a wavelength in the interval between 8 and 17 microns. The operating wavelength is selectable by adjustment of the width of the GaAs central layer 38 of the quantum well structures 34. In this connection, Figure 9 provides a graph 250 of operating wavelength (pom) against width ( ) of layer 38.

A table below gives the range of possible values for layer widths in the embodiments.

Reference Thickness Layer Number(s) Range AlAs 26 (204) 1-10 ym doped cladding 42 100-300 A undoped cladding 44 0-300 A The number of quantum wells may be in the range 1 to 10, with the modulator absorption being highest with 10 quantum wells.

Claims (20)

1. An electro-optic device including a semiconductor quantum well region, containing at least one quantum well, for modulating or detecting incident radiation polarised with an electric vector component perpendicular to the plane of the quantum well, and having: (a) means for populating the quantum well; (b) either (i) means for passing a current through the quantum well region to facilitate detection of photocarriers, or (ii) means for reducing the population of the quantum well to alter absorption and provide optical modulation; and (c) means for coupling incident radiation to the device such that either (i) when the quantum well is population reduced, surface plasmon creation provides optical modulation, or (ii) when the quantum well is populated, a quantum well inter-sub--band absorption provides a modulation or detection mechanism.

2. A device according to Claim 1 wherein the quantum well population means comprises modulation doping means.

3. A device according to Claim 2 wherein the modulation doping means comprises a n-doped region adjacent to the quantum well region.

4. A device according to any claim wherein the means for reducing the population of passing a current through the quantum well is a Schottky contact associated with the quantum well region and having a first interface therewith.

5. A device according to any one of Claims 1 to 3 wherein the means for reducing the quantum well population is an insulated metal electrode located adjacent to the quantum well layer, having a first interface therewith.

6. A device according to any preceding claim wherein the quantum well region comprises a GaAs layer bounded by AlxGai xAs layers.

7. A device according to Claim 6 wherein x is in the range 0.25 to 0.35.

8. A device according to Claim 7 wherein x is 0.3.

9. A device according to Claims 6, 7 or 8 wherein the quantum well is modulation doped by silicon doping of the AlxGa1 #xAs layers.

10. A device according to anyone of Claims 6 to 9 wherein the means for reducing the population of or passing a current through the quantum well is a Schottky contact of gold located associated with the quantum well region having a first interface therewith.

11. A device according to anyone of Claims 6 to 9 wherein the means for population reducing the quantum well is an insulated electrode of platinum silicide associated with the quantum well region having a first interface therewith.

12. A device according to any preceding claim wherein the means for coupling incident radiation includes a first coupling means comprising first and second regions having a second interface therebetween, wherein the first region is located between the quantum well region and the second region, and has a refractive index less than that of the second region.

13. A device according to Claim 12 wherein the incident radiation coupling means includes second means for coupling the incident radiation to the second interface at an angle greater than the critical angle of the second interface and such that, when the quantum well is depleted, a surface plasmon is created at the first interface to provide an absorption mechanism for modulation of the radiation.

14. A device according to Claim 12 wherein the incident radiation coupling means includes second means for coupling the incident radiation to the second interface at an angle less than the critical angle and such that, when the quantum well is populated, quantum well inter-sub-band transitions are excited to provide an absorption mechanism for modulation or detection of the radiation.

15. A device according to Claims 13 or 14 wherein the second coupling means is a prism.

16. A device according to any one of Claims 1 to 11 wherein the means for coupling incident radiation to facilitate detection comprises a diffraction grating adjacent to the quantum well region.

17. A device according to Claim 16 wherein the diffraction grating is a two dimensional pattern of dots or squares.

18. An electro-optic device substantially as herein described with reference to the accompanying Figures 1 to 3.

19. An electro-optic device substantially as herein described with reference to the accompanying Figures 1 to 3 and 6.

20. An electro-optic device substantially as herein described with reference to the accompanying Figure 7.