OPTICAL POWER CONVERTER

Abstract

An optical power converter device (200, 300) comprises a semiconductor waveguide structure having a first end facet (F1) configured to receive an incident light beam, and one or more light absorbing layers (210a) with a total thickness of substantially less than 100 nm and configured to absorb light guided by the waveguide structure. The device further comprises a cathode (202) and an anode (204) in contact with substantially the entire length of the waveguide structure in the direction of propagation of light from the first end facet for outputting generated electrical power. An optical power converting system (1000) comprising the optical power converter device is also provided, as is a method of operating the optical power converter.

Claims

1. An optical power converter device comprising: a semiconductor waveguide structure having a first end facet configured to receive an incident light beam, and one or more light absorbing layers with a total thickness of substantially less than 100 nm and configured to absorb light guided by the waveguide structure; and a cathode and an anode in contact with substantially the entire length of the waveguide structure in the direction of propagation of light from the first end facet for outputting generated electrical power, wherein the optical power converter device operates in a photovoltaic mode such that a positive voltage is generated between the anode and cathode in use.

2. The device of claim 1, wherein the waveguide structure comprises an n-type cladding region, a p-type cladding region, and a core region between the n-type and p-type cladding regions for confining and guiding light; and wherein the cathode and anode are electrically connected to the respective n-type and p-type cladding regions of the waveguide structure.

3-5. (canceled)

6. The device of claim 1, wherein the device is configured to generate an open circuit voltage of at least 90% of the band-gap energy/voltage of the absorbing layer(s).

7. The device of claim 1, wherein the one or more light absorbing layers spatially overlap a guided mode of the waveguide structure.

8. (canceled)

9. The device of claim 1, wherein the n-type and p-type cladding regions are formed of or comprise one or more semiconductor layers having a band-gap energy substantially higher than that of the core region, the one or more light absorbing layers and the photon energy of the incident light beam.

10. The device of claim 1, wherein the waveguide structure is or comprises a double clad waveguide structure.

11. The device of claim 1, wherein at least one of the n-type and p-type cladding regions comprises an inner and an outer cladding layer, with respect to the core region.

12. The device of claim 1, wherein the outer cladding layer(s) is (are) formed of or comprises a semiconductor material with a band-gap energy substantially higher than that of the respective inner cladding layer(s) and the photon energy of the incident light beam.

13. The device of claim 1, wherein the waveguide structure comprises a series of epitaxially stacked p-n junctions, each p-n junction containing at least one of the one or more light absorbing layers.

14-16. (canceled)

17. The device of claim 1, wherein one of the cathode and anode is a top contact that contacts a top surface of the waveguide structure with respect to the growth direction of the waveguide structure; and wherein: the top contact is provided as a strip having a width that at least partially defines the lateral width of the fundamental mode of the waveguide structure; and/or the top contact is provided as a strip having a width that extends, at least partially, across the lateral width of the waveguide structure.

18. The device of claim 1, wherein the waveguide structure comprises a ridge formed in a top surface of the waveguide structure with respect to the growth direction, the ridge having a width that defines the lateral width of the fundamental mode of the waveguide structure.

19-20. (canceled)

21. The device of claim 1, wherein the top contact and/or ridge has a width that is substantially constant along its length, or has a width that varies or tapers along at least a portion of its length.

22. The device of claim 1, wherein the waveguide structure further comprises a second end facet, optionally arranged substantially opposite and/or perpendicular to the direction of light propagation in the waveguide structure.

23. The device of claim 1, wherein the top contact and/or ridge extends between the first and second end facets.

24. (canceled)

25. The device of claim 1, wherein the first end facet comprises a coating for reducing the reflectance of the first end facet for the wavelength of the incident light beam; and, optionally or preferably, wherein in the coating is a layered dielectric coating configured to reduce the reflectance of the first end facet to less than 10% for the wavelength of the incident light beam.

26-27. (canceled)

28. The device of claim 1, comprising an array of said waveguide structures arranged on a substrate, wherein each waveguide structure has a cathode and anode connected thereto, and the cathodes and anodes of each respective waveguide structure are electrically connected in series or in parallel.

29. An optical power converting system comprising: an optical power converter device as defined in claim 1; and a coupling arrangement configured to couple an incident light beam into the waveguide structure of the power converter device, wherein the anode and cathode of the power converter device are connectable to an electrical component and configured to provide electrical power to the electrical component.

30. The system of claim 29, wherein the coupling arrangement comprises an optical fiber and/or optical waveguide configured to guide the light beam to the power converter device.

31. The system of claim 29, wherein the coupling arrangement comprises a lens configured to receive an incident light beam and couple the incident light beam into the waveguide structure.

32. (canceled)

33. A photonic integrated circuit comprising an optical power converter device as defined in claim 1 for providing electrical power to the photonic integrated circuit.

34. A method of operating the optical power converter of claim 1 operating in a photovoltaic mode, comprising: coupling an incident light beam into the waveguide structure; absorbing the guided light to generate photocarriers and an associated photovoltage; re-distributing the generated photocarriers and photovoltage over the length of the device; and extracting photocarriers at the cathode and anode to generate an electrical power output signal.

35-37. (canceled)

Description

BRIEF DESCRIPTION OF DRAWINGS

[0082] In order that the invention can be well understood, embodiments will now be discussed by way of example only with reference to the accompanying drawings, in which:

[0083] FIG. 1 shows a schematic layer structure of a conventional top illumination optical power converter device;

[0084] FIG. 2 shows a schematic illustration conventional top illumination optical power converter device;

[0085] FIG. 3a shows a schematic layer structure of an optical power converter device according to an embodiment of the invention;

[0086] FIG. 3b shows a layer structure of the absorbing region of FIG. 3a according to an embodiment of the invention;

[0087] FIG. 3c shows a layer structure of the absorbing region of FIG. 3a according to another embodiment of the invention;

[0088] FIG. 4 shows a schematic illustration of an optical power converter device according to an embodiment of the invention;

[0089] FIG. 5 shows an example cross-section of the device of FIG. 4;

[0090] FIG. 6 shows the device of FIG. 4 including a lateral waveguide;

[0091] FIGS. 7 to 9 show a top view of the device of FIG. 6 with different shaped lateral waveguides;

[0092] FIG. 10 shows a schematic layer structure of an optical power converter device according to another embodiment of the invention;

[0093] FIGS. 11a and 11b show multiple devices connected in series and parallel, respectively;

[0094] FIG. 11c shows an example of multiple sub-devices connected in series;

[0095] FIG. 12 shows a schematic illustration of an optical power converter system according to an embodiment of the invention;

[0096] FIG. 13a shows a schematic layer structure of an optical power converter device according to another embodiment of the invention;

[0097] FIG. 13b shows a schematic layer structure of the absorbing region of FIG. 13a;

[0098] FIGS. 14a and 14b show example energy band diagrams of the device of FIG. 13a;

[0099] FIG. 15 shows a schematic layer structure of an optical power converter device according to another embodiment of the invention;

[0100] FIG. 16 shows light induced current and voltage characteristics of the device of FIG. 15;

[0101] FIG. 17 shows a schematic layer structure of an optical power converter device according to another embodiment of the invention; and

[0102] FIGS. 18a and 18b show example energy band diagrams of the device of FIG. 17.

[0103] It should be noted that the figures are diagrammatic and may not be drawn to scale. Relative dimensions and proportions of parts of these figures may have been shown exaggerated or reduced in size, for the sake of clarity and convenience in the drawings. The same reference signs are generally used to refer to corresponding or similar features in modified and/or different embodiments.

DETAILED DESCRIPTION

[0104] FIGS. 1 and 2 show a conventional photovoltaic (PV) optical power converter device 100 configured for top-illumination. The device 100 comprises a sequence of semiconductors layers in the growth direction G comprising, in order, an n-type substrate 101, and an n-type semiconductor layer 112 and a p-type semiconductor layer 116. The n- and p-type layers 112, 116 form a p-n junction (PNJ) and provide the absorbing region 110. The absorbing region 110 is thick to increase the light absorption, with a thickness t typically in the range 2-4 .Math.m. An n-side electrode (cathode) 102 is provided in electrical contact with the n-type layer 112 (via the substrate 101) and a p-side electrode (anode) is provided in electrical contact with the p-type layer 114 for collecting/extracting the photo-generated electrons and holes, respectively. In this example, the anode 104 is the top electrode (with respect to the growth direction G) and takes the form of a metal grid to permit transmission of incident light through to the underlying absorbing region 110. If a p-type substrate is used, the PNJ is reversed and the cathode is the top electrode (not shown). Some variants of the device 100 comprise an intrinsic region between the n-type and p-type layers 112, 114 which may contain a large number of quantum wells (QWs) that act as an additional absorber (not shown).

[0105] In use, a light beam LB with a photon energy (E.sub.ph) at or just above the band-gap energy (E.sub.g) of the n- and p-type layers 112, 116 of the absorbing region 110 is incident normal to the top surface S of the device 100 (i.e. parallel to the growth direction G). Incident light is transmitted through the metal grid 104 to the underlying semiconductor layers where it is absorbed in the absorbing region 110 generating electron-hole pairs (photo-carriers). The electron-hole pairs are separated by the PNJ and move (via drift and diffusion) towards the respective cathode and anode 102, 104 at opposite sides of the PNJ where they are collected, thereby generating electrical power.

[0106] Several factors can affect the optical to electrical energy conversion efficiency (referred to hereafter as conversion efficiency) of an optical power converter device (and PV devices in general). These include, but are not limited to, transmission losses at the light receiving surface, incomplete absorption in the absorbing region, carrier recombination losses, thermal losses, and resistive losses in the electrodes and/or in the semiconductor layers.

[0107] For maximum conversion efficiency, all incident photons should be absorbed and the photo-carriers collected/extracted with minimal resistance, with each photon delivering the photon energy (E.sub.ph). The top-illumination configuration of the device 100 suffers from inherent optical/photon losses due to the shadowing effect of the top electrode 104 and incomplete absorption of light in the absorber 110 due to the limited optical path length in the absorbing region 110, and as well as inefficient photo-carrier extraction at high incident powers due to the lateral transport of photo-generated holes (or electrons if a p-type substrate is used) from the PNJ to the contacting areas of the metal grid electrode 104. Example transport paths of photo-generated electrons (solid circles) and holes (open circles) in the device 100 are indicated by the dashed arrows in FIG. 2. In PV devices, photo-carriers are considered collected once they become majority carriers, i.e. when photo-generated electrons reach the n-type layer 116 and holes reach the p-type layer 112. In the device 100, photo-carriers are generated throughout the thickness of the n-type or p-type layers 112, 116 of the absorbing region 110 and must therefore cross the p-n junction to be collected. This increases the transport length and time for collection, which in turn increases the resistive losses and the probability of carrier recombination before collection. Lateral transport paths to the top electrode 104, indicated by the curved arrows, also increase the device resistance leading to power loss.

[0108] The maximum voltage provided by the device 110 (and PV devices in general) is the open circuit voltage which is determined by the band-gap energy, E.sub.g, of the absorbing region 110 and the incident power density, among other things. Under typical operating conditions of a conventional top-illumination device 100 the open circuit voltage is typically 0.2-0.4 V less than the band-gap energy, E.sub.g (in eV) of the absorbing region 110, demonstrating significant power loss through loss mechanisms such as those described above.

[0109] FIGS. 3a, 3b, 3c and 4 show an optical power converter device 200 according to an embodiment of the invention. The device 200 is configured for side illumination and comprises a semiconductor waveguide structure to guide light along the length of the device 200 with a thin absorbing region 210 arranged to absorb the guided light. The waveguide structure comprises a sequence of semiconductors layers to provide the required index contrast and arrangement n-type and p-type regions for separating and extracted photo-generated carriers. The layer structure comprises an n-type substrate 201, an n-type cladding region 222, a p-type cladding region 226, a core region CR positioned between the n-type and p-type cladding regions 222, 226, and an absorbing region 210 within the core region CR (N.B. if a p-type substrate is used, the position of the n and p-type regions is switched). The absorbing region 210 comprises one or more absorbing layers 210a with a band-gap energy E.sub.g ≤ E.sub.ph and a total thickness t of less than 100 nm in the growth direction G. The core region CR has a higher refractive index than that of the n-type and p-type cladding regions 222, 226 to form a transverse (slab) waveguide WG_t1 for confining light in the growth direction G to one or more guided modes M (indicated by the dashed line in FIGS. 3a and 4). The n-type and p-type cladding regions 222, 226 and the core region CR (excluding the absorbing layer(s) 210a) have a band-gap energy E.sub.g > E.sub.ph so that they are substantially transparent to the incident light beam LB. The n-type and p-type regions 222, 226 form a p-n junction for separating photo-carriers generated in the light absorbing region 210, and are electrically contacted by a low resistance cathode 202 (bottom electrode) and anode 204 (top electrode), respectively, for outputting generated electrical power.

[0110] In the illustrated embodiment, the absorbing region 210 is positioned within, or forms at least part of, the higher index core region CR. However, this is not essential. For example, the absorbing region 210 can be located within the lower index n-type or p-type cladding regions 222, 226 (not shown) provided it is within the envelope or spatial extent of a guided mode M (any mode in general, but preferably the fundamental mode) of the waveguide WG_t1 to thereby absorb guided light. The waveguide WG_t1 can be configured to be single mode for the wavelength of incident light, but can be also advantageously multi-moded, depending on the application. Although shown as separate regions in FIGS. 3a and 4, the cladding region adjacent the substrate 201 can include the substrate 201 itself, if the substrate 201 is transparent to the incident light.

[0111] Each region of the device 200 comprises one or more semiconductor layers. In particular, the absorbing region 210 comprises one or more absorbing layers 210a, such as a sequence of absorbing and non-absorbing/transparent layers 210a, 210b, as shown in FIG. 3b. In an embodiment, the one or more absorbing layers 210a are QWs 210a or layers of quantum dots (QDs) 210a, as is known in the art. For example, the device may comprise 1 to 10 QWs 210a sandwiched between non-absorbing/transparent barrier layers 210b (e.g. see FIG. 13b), where the total thickness of the QWs 210a is less than 100 nm.

[0112] The n-type and p-type cladding regions 222, 226 are doped to have a charge carrier density/concentration in the range of 0.1-5 ×10.sup.18 cm.sup.-3 to minimise resistance without introducing excess free carrier loss. The layer structure may also comprise an n-type contact layer 240 and a p-type contact layer 230 between the respective n-type and p-type cladding region 222, 226 and cathode/anode 202, 204 to reduce the contact resistance of the respective cathode and anode 202, 204 and thereby assist power extraction, as is known in the art (see FIG. 3a). The contact layers 230, 240 are more highly doped than the respective cladding regions 222, 226 and/or are formed of a semiconductor material with a lower band-gap energy than that of the respective cladding regions 222, 226. In this embodiment, the absorbing region 210 and the core region CR are un-doped (i.e. intrinsic).

[0113] In an embodiment, the absorbing region 210 comprises a series of epitaxially stacked (in the growth direction G) p-n junctions PNJ1-PNJ3, three in this example, each containing at least one of the one or more absorbing layers 210a, e.g. one or more quantum wells and/or layers of quantum dots 210a, as shown in FIG. 3c. Each p-n junction PNJ1-PNJ3 is formed by a p-type semiconductor layer 210p and an n-type semiconductor layer 210n and is separated by a low resistance tunnel junction TJ. In this case, the transverse waveguide WG_t1 forms essentially a single multimode waveguide containing all the p-n junctions PNJ1-PNJ3. Alternatively, the layer structure of the n-type cladding layer 222, absorbing region 210 and the p-type cladding region 226 can be repeated in the growth direction, separated by a low resistance tunnel junction, to form the series of stacked p-n junctions (not shown). In this case, each p-n junction may form a separate transverse waveguide WG_t1 (which can be single mode). Under edge or side illumination, the resultant series connection of p-n junction (i.e. diodes) will result in a summation of the voltage of each of the number of p-n junctions and a reduction in current by the corresponding number. This permits the delivery of higher voltages to the system.

[0114] FIG. 5 shows a side view of the device 200. The device 200 comprises a first (front) end facet F1 for receiving a light beam LB in a direction perpendicular to the growth direction G, and a second (rear) end facet F2 substantially opposite the first end facet F2. The distance between front and rear end facets F1, F2 defines the length L of the device 200 and absorbing layer(s) 210a. The anode/top electrode 204 extends along the entire length L of the device 200, as will be discussed in more detail below. The front end facet F1 includes an anti-reflection coating F1_c for the wavelength of incident light to reduce transmission losses by reflection, as is known in the art. For example, the coating F1_c may be a layered dielectric coating configured to reduce the reflectance of the wavelength of incident light to between 0.01 and 10%. The rear end facet F2 can include a highly reflective coating F2_c for the wavelength of incident light to direct/reflect any light not absorbed over the length L of the device 200 back through the device 200 for a second pass. For example, the coating F2_c can be a metal mirror coating, as is known in the art. In this case, the plane of the rear end facet F2 is substantially parallel to the growth direction G and perpendicular to the direction of light propagation in the waveguide structure. The plane of the front end facet F1 can be parallel to the rear end facet F2 as shown, or it can be inclined at an angle between 0 and 8 degrees to the growth direction G to reduce feedback to the light source.

[0115] In use, a light beam LB with a photon energy E.sub.ph at or just above the band-gap energy E.sub.g of the one or more absorbing layers 210a is incident onto the front end facet F1 in a direction substantially normal to the growth direction G and couples into the transverse waveguide WG_t1. Light is guided by and propagates along the waveguide WG_t1 in one or more modes M, where it is absorbed along the length L of the absorbing layer(s) 210a by virtue of the spatial overlap (Γ) with the mode(s) M. Photo-carriers are generated directly in the absorbing layer(s) 210a and are separated and collected by the n-type and p-type cladding regions 222, 226. This generates a forward bias on the device 200. As the light power is increased, the generated current and voltage increases and the device 200 can be used as an electrical power source, e.g. by connecting an electrical component 1400 to the cathode and anode 202, 204 (see FIG. 12). Light can be efficiently coupled into the waveguide WG_t1 using a number of techniques known in the art. In one example, a lens 1201 is used to focus the light beam LB onto the front end facet F1 within the numerical aperture of the waveguide WG_t1 (see FIG. 12). However, other means known in the art, such as a butt coupling can also be used (not shown). In alternative arrangement (not shown), a grating or prism coupling can be used to allow light to be coupled into the waveguide WG_t1 at an oblique angle of incidence.

[0116] The device 200 solves several problems with the top illumination power converter device 100 described above. Because photo-carriers are generated directly in the thin absorbing layer(s) 210a within the p-n junction, there is a minimal transport distance to reach and be collected by the respective n- and p-type cladding regions 222, 226, which reduces recombination losses. Further, because light is incident from the side of the device 200 there is no shadowing of incident light by the top electrode 204 and photon/transmission losses at the light receiving surface substantially reduced. The top electrode 204 extends along the entire length L of the device with a width W to match the width of the incident light beam LB and/or waveguide mode M facilitating efficient carrier extraction with minimum resistive losses (e.g. see FIG. 4). This top electrode configuration allows a low resistance Ohmic metal top contact 204 to be formed with low contact and spreading resistance, and photo-carriers to be extracted/collected without lateral transport. Example transport paths of photo-generated electrons (solid circle) and holes (open circle) are indicated by the dashed arrow in FIG. 4. In addition, because the absorbing layer(s) 210a is relatively thin (< 100 nm), the photo-carrier density (cm.sup.-3) is higher than it is for a conventional device 100 with a 2-4 .Math.m thick absorber 110 under the same incident light conditions, which increases the generated power output of the device 200. As a result of these advantages, open-circuit voltages (V.sub.oc) within 5 or 10% of the band-gap energy E.sub.g of the absorbing layer(s) 210a can be generated by the device 200 (see discussion of FIG. 16).

[0117] Absorption of light in the absorbing layer(s) 210a is distributed over the propagation distance x in the waveguide WG_t1 according to the Beer-Lambert law:

[00001]$\begin{array}{cc}\text{I}\left(\text{x}\right)=\text{I}\left(0\right)\text{exp}\left(\text{-}\Gamma \alpha \left(\lambda \right)\text{x}\right)& \text{­­­(1)}\end{array}$

where I(0) is the initial intensity (W/cm.sup.2) of light at the first end facet F1, I(x) is the remaining intensity of light after propagating a distance x (where 0 ≤ x ≤ L), Γ is the spatial overlap factor of the waveguide mode M with the absorbing layer(s) 210a (ranging from 0-1), and α(λ) is the absorption coefficient of the semiconductor material of the absorbing layer(s) 210a as a function of the wavelength λ of incident light. The absorption length is defined as the propagation distance L.sub.α = 1/Γα(λ)over which the light intensity drops by a factor e = 2.7182.

[0118] The operating principle of the device 200 is based upon having low absorption of light in the thin absorbing layer(s) 210a distributed over a long distance. Low absorption is achieved by having a low overlap factor Γ, by reducing the thickness or number of the absorbing layers 210a and/or by the increasing the transverse size of the mode M (e.g. by engineering the materials or thicknesses of core and cladding regions). In an embodiment, the waveguide WG_t1 is configured to provide an overlap factor Γ less than 0.1 (< 10%) and preferably less than 0.01 (< 1 %). As a result, the absorption per unit length of light propagation (Γα(λ)) in the absorbing layer(s) 210a is relatively low compared to a conventional top illumination device 100 where the overlap is almost 100%. However, the total absorption in the device 200 is high, and complete absorption is achievable, because absorption occurs over a long optical path/propagation length x which is at least equal to the device 200 length L, which in turn is substantially greater than the absorption length L.sub.α. By way of example, near band-edge wavelength light (E.sub.ph ~ E.sub.g) is typically fully absorbed over a distance of about 4 .Math.m in a direct bandgap material such as GaAs with 100% overlap (Γ = 1). This is distance is extended to about 400 .Math.m when Γ = 0.01.

[0119] In an embodiment, the length L of the device 200 is greater than 300 .Math.m. In this case, substantially all the incident light is absorbed. Any light not absorbed over the length L of the device 200 is reflected at the rear end facet F2 (e.g. assisted by the reflective coating F2_c) for a second pass of the waveguide WG_t1 which further increases the path length x, as is known in the art (see FIG. 5).

[0120] The photo-generated charge density in the absorbing layer(s) 210a is proportional to the light intensity and therefore also decays with distance from the front end facet F1 with a functional form similar to that of equation 1. As such, movement of the photo-generated electrons and holes to the respective cathode and anode 202, 204 generates an initially non-uniform voltage on the device 200. However, this voltage is rapidly re-distributed over the length L and width of the device 200, along with the extracted power (I*V), within the resistance-capacitance (RC) time constant of the device 200 due to the low resistance cathode and anode 202, 204 (particularly the anode/top electrode 204) that provide equipotential surfaces extending along the length L of the device 200.

[0121] In the embodiment of FIG. 4, the top contact/anode 204 is provided as a strip having a width W which defines the lateral width of the waveguide mode(s) M. The strip contact 204 modifies the effective refractive index of the transverse/slab waveguide WG_t1 in the region beneath it, which provides lateral confinement (i.e. in a direction perpendicular to the growth direction G and light propagation). The transverse waveguide WG_t1 and the stripe contact 204 together form a channel waveguide that guides light in two-dimensions, as is known in the art

[0122] FIG. 6 shows an alternative embodiment of the device 200 where the waveguide structure comprises a ridge or ridge pattern R to provide lateral confinement and assist in lateral guiding of light propagation. The top electrode/anode 202 covers the top surface S of the ridge R. The transverse waveguide WG_t1 and the ridge R together form a channel or ridge waveguide that guides light in two-dimensions, as is known in the art. The ridge can be formed by an etching process during device fabrication, as is known in the art. The depth of the ridge extends at least partly through the p-type (upper) cladding layer 224 as shown, but can extend through the absorbing layer(s) 210 (not shown). In an embodiment, the length of the ridge R is equal to the length L of the device 200 and the width W is in the range 2 to 500 .Math.m.

[0123] The width W of the top contact 204 and/or ridge R provides control over the width of the waveguide mode M and, together with the device length L, the area (A = L x W) over which light is absorbed in the absorbing layer(s) 210a (see FIGS. 7 to 9). For a given incident optical power (P.sub.i), the spatial overlap Γ, absorption coefficient α, area A and the thickness t of the absorbing layer(s) 210a determine the generated photo-carrier density. Since the device 200 is preferably operated at maximum incident power, the area A can be used to set the operating point of the device 200. The area A can be precisely controlled during device fabrication and configured to achieve an predefined/optimum photo-carrier density at the maximum input power for generating a high power output (as described above) while avoiding non-linear saturation effects. In an embodiment, the optimum photo-carrier density is the transparency density in open circuit conditions, which is approximately 1-2 ×10.sup.18 cm.sup.-3 for GaAs and indium phosphide (InP)-based semiconductors. The transparency density is the density at and beyond which the absorbing layer(s) 210a stops being absorbing to the incident light. Beyond this point, i.e. at greater incident powers, the optical gain occurs and the device 200 can operate in stimulated emission mode, e.g. as a laser.

[0124] By way of example only, consider an incident light beam LB with P.sub.i = 1 W, E.sub.ph = 1 eV, and a spot size of 5 .Math.m.sup.2 (beam width ~ 2.5 .Math.m), an absorbing region 210 having a single t = 10 nm thick QW 210a where Γ = 0.01 and α = 8,000 cm.sup.-1, and a ridge width W = 3 .Math.m. According to equation 1, assuming full transmission through the front end facet F1, 1% of the light remains after a propagation distance x.sub.1 = 570 .Math.m (by equating exp(-Γαx.sub.1) = 0.01). Over this distance, P.sub.i/(E.sub.ph * e * x.sub.1 * W * t) = 0.36 × 10.sup.30 cm.sup.-3s.sup.-1 photo-carriers are generated per second per unit volume, where e = 1.6 × 10.sup.-19 J. Assuming a carrier lifetime of 5 ns, this leads to a nominal photo-carrier density of ~ 1.5 × 10.sup.21 cm.sup.-3 in open circuit conditions. At that photo-carrier density the device 200 would be operating in stimulated emission mode, i.e. not as an efficient optical power converter. To achieve a transparency photo-carrier density of ~ 1.5 × 10.sup.18 cm.sup.-3 at open circuit, the photo-carrier density must be diluted by a factor of about 1000. This can be done by controlling the thickness t or overlap Γ of the absorbing layer(s) 210a, and/or by increasing the area A = L x W of the waveguide or top contact 204. In practice, the waveguide area A operates as the primary buffer to control the generated photo-carrier density, but this can only work if the generated voltage is re-distributed over the length L and width W of the top electrode 204, as described above.

[0125] The thickness t of the transverse waveguide WG_t1 and width W of the top contact/ridge R provides control over the size and shape of the resulting waveguide mode M. In practice, this also determines the required light beam LB profile for efficiently coupling incident light into the waveguide, and incident light beam LB profile can be matched to the waveguide mode M, or vice versa.

[0126] FIGS. 7 to 9 show top views of the device 200 of FIG. 6 with ridges R of different shapes. The width W of the ridge R may be substantially constant over its length L as shown in FIG. 7, or it may vary over at least a portion of its width W as shown in FIGS. 8 and 9. In particular, the width W may taper out towards the front end facet and/or the rear end facet F2 (see FIGS. 8 and 9). A front end taper may help to compensate for the increased light absorption in that region by diluting the photo-carrier density. The width W of the strip contact 204 may be varied in the same way, with or without a ridge R (not shown).

[0127] At the open circuit voltage of the device 200, no power is extracted through the contacts 202, 204 (generated photo-carriers have nowhere to go) and energy can dissipate through radiative (e.g. spontaneous emission) and non-radiative carrier recombination. Emitted light can be re-cycled using appropriate reflective layers and/or coatings. In an embodiment, the front end facet coating F1_c is also configured to be highly reflective for the emission wavelength where this is different to the incident wavelength, and the rear end facet coating F2_c is configured to be highly reflecting for the wavelengths of both the incident and emitted light. In addition, a distributed Bragg reflector (which may be part of the layer structure) and/or a highly reflective metal-based mirror (which may be part of the bottom electrode 202) can be also included to recycle light emitted towards the bottom electrode 202 (not shown), as is known in art. Defect or trap related non-radiative carrier recombination can be minimised by the use of high crystalline quality semiconductor materials (see below). Auger recombination, a non-radiative recombination process well known in the art, can be limited by keeping the photo-carrier density below a suitable level and by using a material where Auger effects are minimal. For example, Auger recombination is proportional to the carrier density cubed and the strength of the effect depends on the material and wavelength (it is minimal below 1350 nm for indium gallium arsenide (InGaAs)-based materials).

[0128] The photo-response of the device 200, and therefore the absorption of the light in absorbing layer(s) 210, should preferably be independent of the polarisation of incident light. Where the absorbing layer(s) 210a is a QW or a QD layer 210a, the absorption is polarisation dependent for near band-edge wavelength light (where E.sub.ph ~ E.sub.g), but becomes polarisation independent at higher photon energies E.sub.ph > E.sub.g. In such cases the photon energy E.sub.ph can be offset above the band edge E.sub.g. The offset should be sufficient for absorption to be polarisation independent while minimising thermalization losses. In an embodiment, the offset is between 50 and 100 meV above the band edge E.sub.g.

[0129] At high enough incident powers, where the photocarrier density and current density exceed a threshold (such as the transparency density of about 1-2 × 10.sup.18 cm.sup.-3), stimulated emission occurs and the device 200 can operate as an optically pumped laser. In practice, the device 200 may switch from an efficient absorber to a laser quite abruptly, once some additional losses are overcome. In an embodiment, the device 200 is operated in the stimulated emission mode while current is extracted and emitted light from the device 200 is modulated to transmit information back towards the source. The emitted light can be modulated by applying a bias to the cathode and anode. Since the emitted wavelength of the optically pumped device 200 is down-converted from the pump/incident wavelength (e.g. since E.sub.ph > E.sub.g), the emitted light (i.e. optical output signal) can be separated from the pump/incident light at the source/receiving end. For example, the device 200 can be used to power a separate electronic component or device, and that device may be configured to generate a modulated data output (voltage) signal that can be applied to the anode and cathode to modulate the emitted light. This allows both remotely powering and communicating of remote devices. In such applications, the optical power converter device 200 can both convert the incident (pump) light beam to power a remote device/electronic component and generate an optical signal (through light emission) to send information back from it. This has advantages in confined spaces, one such application being at the distal end of an endoscope where a camera could be located and powered by the optical power converter device 200. This provides a significant advantage in compactness as a single fibre and a single device can provide the channel for both the powering and the signalling. Operating in this mode is less efficient (compared when the device 200 is not emitting) due to sacrificing the energy needed to make the device 200 lase, but has the significant advantage that the same device (and coupling) can be used to power and signal.

[0130] FIG. 10 shows another embodiment of an optical power converter device 300 according to the invention. The device 300 has the same features as device 200 but the waveguide structure is configured as a double clad waveguide structure with an inner waveguide WG_t1 and an outer waveguide WG_t2. The n-type cladding region 222 includes a first (inner) n-type cladding layer 222a and a second (outer) n-type cladding layer 222b, and the p-type cladding region 226 includes a first (inner) p-type cladding layer 226a and a second (outer) p-type cladding layer 226b, with respect to the core region CR. The outer cladding layers 222b, 226b have a refractive index that is lower than that of the inner cladding regions 222a, 226a so as to provide an outer transverse waveguide WG_t2 (which is typically multi-mode) for confining light to a region including the first n-type and p-type cladding layers 222a, 226a, core region CR and absorbing region 210.

[0131] The double clad waveguide structure allows a light beam LB with greater width (e.g. greater than 10 .Math.m wide) to couple into the device 300 to deliver more light for converting to electrical power. In use, incident light in a wider light beam LB couples first to the outer waveguide WG_t2, and then to the inner waveguide WG_t1 as it propagates, where it can be efficiently absorbed by the light absorbing layer(s) 210a.

[0132] The principle of the double clad waveguide structure is the following. The power density (in terms of W/cm.sup.2) of the light beam LB incident on the front end facet F1 must be kept below the damage threshold value, which is typically around 10 MW/cm.sup.2 for most semiconductor materials. To efficiently couple to a waveguide, the size/beam width of the incident light beam LB at the front end facet F1 should be substantially matched to the size of the waveguide mode M. Thus, the power density at the front end facet F1 for a given maximum incident power is primarily set by the size of the waveguide mode M. Increasing the thickness of the core region CR increases the size of the waveguide mode M and allows wider light beams LB to be used. For a fixed power this reduces the incident power density. Wider light beams LB also increase/relax the alignment tolerances for coupling the incident light beam LB to the waveguide. However, in a single clad waveguide structure as in device 200, increasing the thickness of the core region CR would further reduce the overlap Γ of the collected light with the absorbing layer(s) 210 requiring potentially impractically long device lengths L to achieve total light absorption. The double clad waveguide structure of device 300 allows light to be coupled into the inner waveguide WG_t1 (containing the absorbing layer(s) 210a) using a wider light beam LB than would be possible in an equivalent single clad waveguide structure, and without changing the mode size of the inner waveguide WG_t1. By using a wider light beam LB, more optical power can be delivered to the device 300 at a given power density, or the same optical power can be delivered to the device 300 at a lower power density.

[0133] FIGS. 11a and 11b show multiple devices 200, 300 connected in series and parallel to increase the voltage or current output, receptively. To form a series connection, the cathode 202 of a first device 200, 300 is electrically connected to the anode 204 of the next device 200, and so on, by interconnects 205. To form a parallel connection, the cathode 202 of each device 200, 300 is electrically connected by an interconnect 205 and the anode 204 of each device 200, 300 is electrically connected by an interconnect 205. In this case, the front end facet F1 of each device 200, 300 is arranged to receive a separate light beam LB (not shown), which may come from the same source or different sources, and the generated voltages or currents in each device 200, 300 add depending on the configuration.

[0134] FIG. 11c shows an embodiment of the device 200 comprising an array of sub-devices 200′ electrically connected in series (the same principle can be applied to device 300). In this case, the semiconductor layer structure is grown on a semi-insulating or substantially non-conducting substrate 201, and the array of sub-devices 200′ are formed from the semiconductor layer structure provided on the substrate 201 using standard semiconductor fabrication techniques, such as lithography, etching, deposition and ion implantation. Each sub-device 200′ has a respective cathode and anode 202, 204. Similar to the arrangement in FIG. 11a, the cathode 202 of a first sub-device 200′ in the array is electrically connected to the anode 204 of the next sub-device 200′ and so on by interconnects 205 to form a series connection, as shown. The front end facet F1 of each sub-device 200′ is arranged to receive a separate light beam LB (not shown), which may come from the same source or different sources, and the generated voltages in each sub-device 200′ adds up. Although the illustrated embodiment shows a series connection, it will be appreciated that a similar arrangement of sub-devices 200′ can be implemented for a parallel connection.

[0135] FIG. 12 shows a power-by-light system 1000 comprising an optical power converter device 200, 300 according to an embodiment. The system 1000 comprises a coupling arrangement 1200 for directing the incident light beam LB to the first end facet F1 of the device 200, 300 in a direction substantially perpendicular to the growth direction G. The coupling arrangement 1200 comprises a lens 1201 for receiving an incident light beam LB and coupling the light beam LB to the waveguide structure. This can be achieved by aligning the focused light beam LB within the numerical aperture of the waveguide WG_t1, WG_t2, as is known in the art. The coupling arrangement 1200 also comprises an optical fiber 1202 for guiding the light beam LB from a light source 1100, in this case a laser, to the lens 1201. Alternatively, light emitted from the source 1100 can be delivered to the lens 1201 through free-space. The laser 1100 emits monochromatic or narrow spectral bandwidth light at photon energy E.sub.ph at or just above the band-gap energy E.sub.g of the absorbing region 210, as described above. The cathode and anode 202, 204 of the device 200, 300 are connected/connectable to power terminals A and B of an electrical component or system 1400 for providing converted electrical power to the electrical component or system 1400. Multiple devices 200, 300 or multiple sub-devices 200′, 300′ can be used as described above. In that case, a separate lens 1201 is used for each device 200, 300 or sub-device 200′, e.g. a fiber array and a lens array (lens-ended fiber array) can be used. In an embodiment, the device 200, 300 and the electrical component or system 1400 are provided or fabricated on the same semiconductor chip 1500.

[0136] The system 1000 further comprises a passive cooling arrangement 1300 for removing heat generated from unconverted optical energy from the device 200, 300. The cooling arrangement 1300 comprises a heat sink in thermal contact with the device 200, 300. The heat sink comprises a material with high thermal conductivity, such as copper or ceramic or any other suitable heat sink material known in the art. The heat sink may be coupled directly or indirectly to either the cathode or anode 202, 204. Where the device 200, 300 and the electrical component or system 1400 are provided or fabricated on the same semiconductor chip 1500, the heat sink may be shared between them.

[0137] For optical power transfer by fibers, wavelengths in the range 1300 - 1600 nm are most suitable for delivering optical power over long distances with low attenuation inside the fiber(s). It is also desirable to operate in the 800 - 1300 nm range to take advantage of high power glass and fibre lasers for optical power delivery. Operation of the optical power converter device 200, 300 with different incident wavelengths is achieved through the choice of the semiconductor materials making up the waveguide structure. In particular, by choosing a semiconductor material for the absorbing layer(s) 210 with a band-gap energy E.sub.g matched to the incident photon energy E.sub.ph as described above.

[0138] The semiconductor layers of the device 200, 300 are formed of direct band-gap III-V compound semiconductor materials. III-V semiconductor layers with direct band-gaps in the range of 300 nm to 4000 nm can be epitaxially grown with high crystal quality using known methods such as molecular beam epitaxy (MBE) and metal organic vapour phase epitaxy (MOVPE). GaAs-based materials can be used for efficiently converting light in the 600 nm to 1300 nm wavelength range. Gallium nitride (GaN)-based materials can be used for converting light in the 360 nm to 500 nm range. InP-based materials can be used for converting light in the 1100 nm to 2000 nm range. Gallium antimonide (GaSb)-based materials can be used for converting light in the 1500 nm to 3000 nm wavelength range.

[0139] A particular application of the waveguide power converter device 200, 300 is in powering photonic integrated circuits, e.g. comprising silicon or SiN waveguides. Such circuits have widespread application in data centres, consumer products and in the Internet of Things. The circuits are connected via optical fiber to send and receive information (i.e. optical data signals). The same or an additional optical fiber can be used to deliver optical power to the photonic integrated circuit thereby saving the connection of an external electrical connection. Therefore, in an embodiment, a photonic integrated circuit comprises a power converter device 200, 300 for powering the circuit. In one example, the power-by-light system 1000 comprises a photonic integrated circuit (which can take the place of, or be part of, the semiconductor chip 1500) with a power converter device 200, 300 for powering the circuit. Where the same fiber is used to transmit the optical signal and power, the light delivered through the fiber can be at a separate wavelength to the operation of the signal channel of the photonic integrated circuit, and the power providing wavelength can be separated on the photonic integrated circuit and delivered to a waveguide power converter 200, 300, as described above. Alternatively, a separate fiber containing the powering wavelength can be provided. The light can be delivered to the photonic integrated circuit by grating couplers, butt coupling, or other techniques known in the art. The waveguide power converter device 200, 300 can be integrated on the photonic circuit using wafer bonding, transfer printing or flip chip methods, as are known in the art.

[0140] Specific example embodiments of the devices 200, 300 that operate in the common fiber wavelength ranges are described below.

Example 1

[0141] FIG. 13a shows a first example 200a of the single clad device 200 described above with an absorbing region 210 having an absorption band edge in the range 850 nm to 1300 nm. A layered structure is formed on an n-type GaAs substrate by an epitaxial growth technique, such as MOVPE. In growth order, the layer structure comprises an n-side contact layer 240 formed of n-type GaAs, a lower n-type cladding layer 222 formed of n-type Al.sub.xGa.sub.1-xAs (0<x<1), an In.sub.xGa.sub.1-xAs-based absorbing region 210, an upper cladding p-type cladding layer 226 formed of p-type Al.sub.xGa.sub.1-xAs, a first p-side contact layer formed of p-type GaAs, and a second p-side contact layer 234 formed of heavily doped p-type GaAs. The In.sub.xGa.sub.1-xAs-based absorbing region 210 comprises one or more (between 1 and 10) In.sub.xGa.sub.1- xAs QWs 210a disposed between GaAs barrier layers 210b, as shown in FIG. 13b.

[0142] A low resistance metal anode 204 is provided on the highly doped p-type GaAs contact layer 234 and a low resistance metal cathode 202 is provided on the n-type GaAs substrate to make electrode connections thereto. Suitable materials and fabrication techniques for making low resistance contacts to a given semiconductor material are known in the art. For example, the anode 204 may comprise Ti, Pt, and Au, and the cathode 202 may comprise Au, Ge, and Ni. The cathode and anode 202, 204 can be annealed to improve the electrical contact. The resistance of the cathode and anode 202, 204 can be further reduced by electroplating a thick metal layer such as Au. A lateral waveguide is formed by etching a ridge with a width W in the range 2 to 500 .Math.m. Devices 200a are formed by cleaving the layer structure perpendicular to the lateral waveguide axis to a length L between 300 .Math.m and 3 cm. The resistance of the device 200a is less than 5 Ω and preferably less than 1 mΩ for best performance (low resistive power dissipation) and operation at high incident power e.g. > 1 W (e.g. the device resistance is inversely proportional to the contact area). The front end facet F1 is coated with a layered dielectric coating F1_c configured to reduce the reflectance of the wavelength of incident light to less than 10%, as is known the art. For example, the layered dielectric can be a quarter wave thick layer of dielectric, such as SiN or SiO.sub.2, such that reflections from the surface(s) of the dielectric layer(s) undergo destructive interference for the incident light. Optionally, the coating F1_c can also be configured to be reflecting for the down-converted wavelength of any light emitted from the QW(s) 210a, as is known the art. For example, the layered dielectric can be designed such that reflections from the surfaces undergo constructive interference for the emitted light. The dielectric layers also serve to protect the front end facet F1 from damage at high incident power densities. The rear end facet F2 is coated with a high reflection coating F2_c for the wavelengths of incident and emitted light. Suitable coating materials and fabrication techniques are known in the art. For example, a quarter wave layer of dielectric (e.g. SiO.sub.2) and metal (e.g. Au), or a quarter wave stack comprising several (e.g. 5) alternating layers of low and high refractive index dielectrics (e.g. SiO.sub.2/TiO.sub.2).

[0143] The composition (x) and thickness of the In.sub.xGa.sub.1-xAs QW(s) 210a and the Al.sub.xGa.sub.1-xAs cladding layers 222, 226 can be controlled/configured to exhibit the required absorption band edge (effective band-gap energy) and provide a transverse waveguide WG_t1 for incident wavelengths in the range 850 nm to 1300 nm. Auger non-radiative recombination is reduced in this wavelength range and radiative quantum efficiency can be greater than 90% (typically, a good power converter should also be a good light emitter). The QW(s) 210a may be pseudomorphically strained to take advantage of a reduced density of states at the band edge. The barrier layers 210b may be tensile strained to balance the excess strain. In one example, for an incident wavelength of 920 nm, the In.sub.xGa.sub.1-xAs QW(s) 210a can be a 10 nm thick In(17%)Ga(83%)As QW surrounded by GaAs barriers 210b, and the composition of the Al.sub.xGa.sub.1-xAs cladding layers 222, 226 can be graded from 10% to 25% in the direction away from the QW 210a. The doping of the Al.sub.xGa.sub.1-xAs cladding layers 222, 226 can also be graded.

[0144] FIG. 14a shows a schematic energy band diagram of the device 200a (under flat band conditions) illustrating conduction band E.sub.c and valance band E.sub.v alignment and relative band-gap energies (where E.sub.g = E.sub.c - E.sub.v) of the layers. The composition (x) of the upper and lower cladding layers 222, 226 is varied/graded at the heterointerfaces to provide a gradual change in the band-gap energy E.sub.g, as shown. This may assist photo-carrier extraction (at forward bias). In FIG. 14a, the thickness (in the growth direction G) and composition (x = 0.2) of the upper and lower Al.sub.xGa.sub.1-xAs cladding layers 222, 226 is the same, resulting in a substantially symmetric profile for the fundamental waveguide mode M of the first transverse waveguide WG_t1, as indicated by the dashed curve. FIG. 14b shows an energy band diagram for an alternative configuration where the thickness and composition of the upper and lower Al.sub.xGa.sub.1-xAs cladding layers 222, 226 is different. Here, the lower cladding layer 222 has a lower Al composition (x = 0.1) compared to the upper cladding layer 226 (x=0.2) and is thicker than the upper cladding layer 226, which shifts and broadens the waveguide mode M compared to that in FIG. 14a. This demonstrates that, for a given core region CR, the relative thickness and composition the upper and lower Al.sub.xGa.sub.1-xAs cladding layers 222, 226 can be configured to vary the overlap of the transverse waveguide mode M with the QW(s) 210a. The overlap Γ is in the range between 0.1% and 10% dependent on the layer structure.

Example 2

[0145] FIG. 15 shows a second example 200b of the single clad device 200 described above with an absorbing region 210 having an absorption band edge close to 1540 nm. The features of the first example device 200a apply equally to the second example device 200b apart from the materials of the layers, as described below. A layered structure is formed on an n-type InP substrate by an epitaxial growth technique, such as MOVPE. In growth order, the layer structure comprises an n-side contact layer 242 formed of n-type InP, a lower n-type cladding layer 222 formed of n-type In.sub.1-xGa.sub.xAs.sub.1-yP.sub.y, an In.sub.xGa.sub.1-xAs-based absorbing region 210, an upper p-type cladding layer 226 formed of p-type Al.sub.xIn.sub.1-x- yGa.sub.yAs, a first p-side contact layer formed of p-type InP, and a second p-side contact layer 234 formed of p-type In.sub.xGa.sub.1-xAs. The In.sub.xGa.sub.1-xAs-based absorbing region 210 comprises one or more (between 1 and 10) un-doped In.sub.xGa.sub.1-xAs QWs 210a disposed between Al.sub.xIn.sub.1-x-yGa.sub.yAs barrier layers 210b, similar that shown in FIG. 13b. The p-type Al.sub.xIn.sub.1-x-yGa.sub.yAs provides a lower valance band offset for holes than p-type In.sub.1-xGa.sub.xAs.sub.1-yP.sub.y, while n-type In.sub.1-xGa.sub.xAs.sub.1-yP.sub.y provides a lower conduction band offset for electrons than n-type Al.sub.xIn.sub.1-x-yGa.sub.yAs. In an alternative embodiment, both the upper and lower cladding layers 222, 226 are formed of Al.sub.xIn.sub.1-x-yGa.sub.yAs or In.sub.1-xGa.sub.xAs.sub.1-yP.sub.y. A low resistance metal anode 204 is provided on the p-type InGaAs contact layer 234 and a low resistance metal cathode 202 is provided on the n-type InP substrate 201 to make electrical connections thereto, as described above for device 200a.

[0146] The composition and thickness of the QW(s) 210a and cladding layers 222, 226 can be controlled/configured to exhibit the required absorption band edge (effective band-gap energy) and transverse waveguide WG_t1 for incident wavelengths in the range of substantially between 1000 nm to 2000 nm. The composition of the upper and lower cladding layers 222, 226 can be graded at their heterointerfaces to assist in carrier extraction as described above with reference to FIGS. 14a and 14b. In one example, the absorbing region 210 comprises a series of five 6 nm thick compressively strained In(0.55)Ga(0.45)As QWs 210a with 10 nm thick Al(0.175)In(0.529)Ga(0.295)As barrier layers 210b to provide a bandgap wavelength of 1547 nm for TE polarized light and 1517 nm for TM polarized light. The waveguide is interfaced on the n-side to n-type InP 240 with a doping of 1×10.sup.18 cm.sup.-3 and on the p-side to a 20 nm thick AlInAs layer 232 and InGaAsP graded layer 234 (from a 1100 nm to 950 nm band edge energy) doped at 8×10.sup.17 cm.sup.-3.

[0147] FIG. 16 shows current-voltage measurements of an example device 200b with an absorption band edge of approximately 1540 nm obtained under illumination by an incident laser beam emitting at a wavelength of 1430 nm with varying incident powers. The device 200b has a 3 .Math.m wide ridge R and a length L of 1.1 mm. The incident power was increased from 0 to 80 mW in steps of 3.08 mW, and the resistance of the device 200b was 1.6 Ohms. It should be noted that a power of 80 mW corresponds to an estimated incident power density of 2×10.sup.6 W/cm.sup.2 which, if absorbed across the entire length of the waveguide structure, would correspond to a current density of approximately 2.7 kA/cm.sup.2 (80 mA/( 3 .Math.m x 1.1 mm)). Such current densities are orders of magnitude larger than applicable in current laser power converters or in concentrator solar cells. The photon energy E.sub.ph is approximately 60 meV above the band-gap energy E.sub.g of the InGaAs QW(s) 210a. Light from a laser source 1100 was delivered onto the anti-reflection coated front end facet F1 by an optical fiber 1202, and a lens 1201 was used to illuminate the ridge waveguide region R with high coupling efficiency by alignment. The incident laser light is naturally polarized. The absorption response of the QW 210a is polarisation dependent near the absorption band edge, but the energy offset of E.sub.ph-E.sub.g ≈ 60 meV results in a polarisation independent response of the device 200b. As shown, an open circuit voltage of V.sub.oc ≈ 0.83 V is obtained at the maximum input power, which exceeds the band-gap energy E.sub.g ≈ 0.805 eV of the InGaAs QW 210a, and an open circuit voltage of V.sub.oc ≈ 0.8 V is obtained at relatively low input powers, which is within 5% of the band-gap energy E.sub.g. This is due to the high carrier densities achieved as a result of the incident power being absorbed in a few (1-5) QWs 210a. The threshold photocarrier density for lasing in QWs at this wavelength is approximately 2×10.sup.18 cm.sup.-3, and is reached under the maximum incident power conditions in these this experiment. If we assume an absorption coefficient for the quantum wells 210a of α = 4,000 cm.sup.-1 at the incident wavelength and a mode overlap factor Γ of 3%, an absorption length of L.sub.α ≈ 80 .Math.m can be estimated, which is much shorter than the total length L of the waveguide (1.1 mm). Even at a Γ of 1%, corresponding to a single QW 210a, the absorption length is L.sub.α ≈ 250 .Math.m. As discussed above, the absorption of photons generates a local photovoltage which is then re-distributed along the length of the waveguide through the low resistance metal contacts, resulting in a redistribution of the photocarriers along the entire length of the device 200b within the R-C time constant of the device 200b. This demonstrates that improved performance can be achieved by reducing the mode overlap to < 1 % (by reducing the thickness/number of QWs 210a and/or by increasing the transverse mode size) and increasing the width W and/or length L of the waveguide, which in turn reduces the device resistance and saturation effects. The device 200b is preferably operated at the maximum power point, which is this specific example device 200b is approximately 42 mW (current x voltage) at 0.6 V. The maximum internal conversion efficiency is estimated to be approximately 57 % at an absorbed power of 20 mW.

[0148] The operating current density of the device 200b is between 100 - 3000 A/cm.sup.2. At each forward bias the QW 210a has a certain photo-carrier density which results in radiative recombination. Generated light is recycled by re-absorption, assisted by the waveguide WG_t1 and by reflective structures on the device 200b as described above. If the current density becomes very high (e.g. > 3000 A/cm.sup.2) the device 200b can start operating as a laser in stimulated emission mode and generate a voltage greater than the band-gap energy E.sub.g due to the splitting of the quasi-Fermi levels. This mode of operation is not optimal as an efficient power converter, but is preferred if a voltage reference is sought.

Example 3

[0149] FIG. 17 shows an example 300a of the double clad device 300 described above with an absorbing region 210 having an absorption band edge in the range 850 nm to 1300 nm, by adjusting the composition of the In.sub.xGa.sub.1-xAs quantum well(s) 210a. The device 300a has the same features as device 200a described above, but with an additional n-type cladding layer 222b and p-type cladding layer 226b. In growth order, the layer structure comprises a n-type GaAs substrate 201, an n-side contact layer 240 formed of n-type GaAs, an outer n-type cladding layer 222b formed of n-type Al.sub.xGa.sub.1-xAs, an inner n-type cladding layer 222a formed of n-type Al.sub.xGa.sub.1-xAs with a lower Al composition (x) (i.e. higher index) than cladding layer 222b, an In.sub.xGa.sub.1-xAs-based absorbing region 210, an inner p-type cladding layer 226a formed of p-type Al.sub.xGa.sub.1-xAs, an outer p-type cladding layer 226b formed of p-type Al.sub.xGa.sub.1-xAs with a higher Al composition (x) (i.e. lower index) than cladding layer 226a, a first p-side contact layer 232 formed of p-type GaAs, and a second p-side contact layer 234 formed of heavily doped p-type GaAs. The In.sub.xGa.sub.1-xAs-based absorbing layer 210 comprises one or more (between 1 and 10) un-doped In.sub.xGa.sub.1-xAs QWs 210a disposed between GaAs barrier layers 210b, as shown in FIG. 13b. The second (outer) lower n-type Al.sub.xGa.sub.1-xAs cladding layer 224 has a thickness in the range 5 to 15 .Math.m.

[0150] The composition and thickness of the QW(s) 210a and cladding layers 222a, 222b, 226a, 226b and can be selected/configured to exhibit the required absorption band edge (effective band-gap energy) and transverse waveguide WG_t1 for incident wavelengths in the range 850 nm to 1300 nm. In one example, a 20 nm thick GaAs QW 210a will absorb light at 850 nm while a 6 nm thick In.sub.0.5Ga.sub.0.5As QW will absorb wavelengths up to 1211 nm. The addition of nitrogen to the OW 210a will further extend the wavelength. The use of InP-based structures also enables operation in the 1200 nm to 1650 nm wavelength range.

[0151] FIG. 18a shows a schematic energy band diagram of the device 300a (ignoring band bending from doping) illustrating conduction band E.sub.c and valance band E.sub.v alignment and relative band-gap energies (E.sub.g = E.sub.c - E.sub.v) of the layers. The composition (x) of the inner cladding layers 222a, 226a is the same (x = 0.1 in this example), and the composition (x) of the outer cladding layers 222b, 226b is the same (x = 0.2 in this example). The composition (x) of the cladding layers 222a, 222b, 226a, 226b is varied/graded at the heterointerfaces to provide a gradual change in the band-gap energy E.sub.g, as described for 200a. The profile for the fundamental waveguide mode M of the first transverse waveguide WG_t1 is indicated by the dashed curve. The profile of the fundamental waveguide mode M′ of the second transverse waveguide WG_t2 is indicated by the dot-dashed curve. Note that due to the larger width/thickness of the waveguide WG_t2 (in the growth direction G), it is multi-modal. FIG. 18b shows an energy band diagram for an alternative configuration of the layer structure of the device 300a where the Al.sub.xGa.sub.1-xAs cladding layer 226a is omitted and the Al.sub.xGa.sub.1-xAs cladding layer 226b provides the upper confinement for both the inner and outer transverse waveguides WG_t1, WG_t2. Equivalently, FIG. 18b represents the case where only one of the n-type and p-type cladding regions 222, 226 includes an inner and outer cladding layer.

[0152] It will be appreciated that the design principles in example 3 can be applied to an InP-based layer structure as in example 2, or a GaSb-based structure or a GaN-based structure (not shown).

[0153] From reading the present disclosure, other variations and modifications will be apparent to the skilled person that are within the scope of the claims. Such variations and modifications may involve equivalent and other features which are already known in the art, and which may be used instead of, or in addition to, features already described herein.

[0154] Although the appended claims are directed to particular combinations of features, it should be understood that the scope of the disclosure of the present invention also includes any novel feature or any novel combination of features disclosed herein either explicitly or implicitly or any generalisation thereof, whether or not it relates to the same invention as presently claimed in any claim and whether or not it mitigates any or all of the same technical problems as does the present invention.

[0155] Features which are described in the context of separate embodiments may also be provided in combination in a single embodiment. Conversely, various features which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable sub-combination.