WO2019218002A1

WO2019218002A1 - A photodetector

Info

Publication number: WO2019218002A1
Application number: PCT/AU2019/050449
Authority: WO
Inventors: Jasper James CADUSCH; Kenneth Brian CROZIER; Jiajun Meng
Original assignee: The University Of Melbourne
Priority date: 2018-05-14
Filing date: 2019-05-14
Publication date: 2019-11-21

Abstract

A photodetector comprises a substrate comprising a first p ⁺ doped region having a first side on which light is intended to be incident when the photodetector is in use, a second p ⁺ doped region, an n ⁺ doped region, a first intrinsic region between the first p ⁺ doped region and the n ⁺ doped region, a second intrinsic region between the n ⁺ doped region and the second p ⁺ doped region, whereby a first P-I-N diode is formed by the first p ⁺ doped region, the first intrinsic region, and the n ⁺ doped region, and a second P-I-N diode is formed by the second p ⁺ doped region, the second intrinsic region, and the n ⁺ doped region. An array of blind holes extends from the first side of the first p ⁺ doped region, through the first intrinsic region at least to an interface of the first intrinsic region and the n ⁺ doped region and no further than an interface between the n+ doped region and the second intrinsic region. The array of blind holes is arranged such that a lattice pattern is formed on the first side of the first p ⁺ doped region. A first electrical contact is formed on the first p ⁺ doped region, and a second electrical contact formed on the second p ⁺ doped region. A first measure of photocurrent can be made with the first P-I-N diode by biasing the first contact negatively relative to the second contact, and a second measure of photocurrent can be made with the second P-I-N diode by biasing the first contact positively relative to the second contact.

Description

A PHOTODETECTOR

FIELD

The present application relates to a photodetector and a method of forming a photodetector. In an example, a plurality of the photodetectors are used to form a m icrospectrometer.

BACKGROUND

Conventional spectrometers generally combine broadband photodetectors with a device for spectral selection such as a grating or a Michelson interferometer. Another example of a device for spectral selection is a filter array.

Recent efforts to miniaturize spectrometers have involved replacing the use of diffractive elements and beam propagation by ultrathin filter devices with tailored transmission spectra. Examples have included quantum dots, thin films, and plasmonic resonators at visible and infrared wavelengths. In each of these devices, a number of distinct wavelength-filtering elements are coupled with either photodetector arrays such as a CCD camera or an external photodetector. The generated photocurrents at each pixel thus depend on the transmission of each filter element, the photodetector responsivity and the incident spectrum. Knowledge of the transmission functions and detector responsivity can then be used to computationally estimate the incident spectrum based on the measured photocurrents. Drawbacks associated with these designs include the use of toxic or CMOS incompatible materials such as cadmiuml or noble metals or the use of fabrication techniques which do not lend themselves to high throughput wafer- scale manufacturing. Furthermore, the materials comprising the spectral filters can introduce their own limitations to the microspectrometer operating range. For example, QD based absorption filters on a silicon CCD matrix will only operate above the band gap of the constituent QDs, typically above 1.5 eV1. On the other hand, thin film bandpass filter arrays and plasmonic nanoantenna arrays both suffer from angle-dependent transmission functions, as well as potential for stray light pixel cross-talk, which if unaccounted for can reduce the utility of a m icrospectrometer. Other approaches have employed etched nanowire photodetectors in order to reduce these effects by combining the spectral filtering and photodetection functions into one all-nanophotonic device which allows for the reduction in optical pixel crosstalk caused by stray light in filter-on-CCD/CMOS sensor architectures. The nanowires behave like cylindrical dielectric waveguides with electric field distributions and charge carrier generation rates determined by the wavelength of light and the radii of the nanowires. However, there are a number of challenges in order to produce a functioning nanowire-based device with the large number of detectors required for a microspectrometer, including the need to planarize the nanowire device to establish electrical contact to the tops of the nanowires, and efficiency issues arising from large surface-area-to-volume ratio of nanowires and due to the nanowires being fragile because they are very thin.

There is a need for alternative photodetectors, in particular photodetectors suitable for microspectrometry.

SUMMARY

Embodiments of the invention provide a photodetector which has a lattice, or “fishnet” pattern formed in an upper portion of a multilayer doped substrate. The lattice pattern provides two sets of interleaved dielectric slab waveguide arrays. A plurality of the photodetectors having waveguide different wavelengths are used to form a spectrometer.

In an embodiment, the invention provides a photodetector comprising:

a substrate comprising

a) a first p⁺ doped region having a first side on which light is intended to be incident when the photodetector is in use,

a second p⁺ doped region,

an n⁺ doped region,

a first intrinsic region between the first p⁺ doped region and the n⁺ doped region,

a second intrinsic region between the n⁺ doped region and the second p⁺ doped region, whereby a first P-l-N diode is formed by the first p⁺ doped region, the first intrinsic region, and the n⁺ doped region, and a second P-l-N diode is formed by the second p⁺ doped region, the second intrinsic region, and the n⁺ doped region, and

an array of blind holes, each blind hole extending from the first side of the p⁺ doped region, through the first intrinsic region at least to an interface of the first intrinsic region and the n⁺ doped region and no further than an interface between the n⁺ doped region and the second intrinsic region, the array of blind holes arranged such that a lattice pattern is formed on the first side of the first p⁺ doped region ;

b) a first electrical contact formed on the first p⁺ doped region; and

c) a second electrical contact formed on the second p⁺ doped region.

whereby a first measure of photocurrent can be made with the first P-l- N diode by biasing the first contact negatively relative to the second contact, and a second measure of photocurrent can be made with the second P-l-N diode by biasing the first contact positively relative to the second contact.

In an embodiment, the array of blind holes is a rectangular array.

In an embodiment, an aperture of each of the blind holes is rectangular.

In an embodiment, each of the blind holes extends partially into the n⁺ doped region.

In an embodiment, each of the blind holes extends to the interface between the n⁺ doped region and the second intrinsic region.

In an embodiment, each intrinsic region is an rr doped region.

In another embodiment, the invention provides a spectrometer comprising a plurality of photodetectors, each photodetector comprising:

a) a substrate comprising

a first p⁺ doped region having a first side on which light is intended to be incident when the photodetector is in use, a second p⁺ doped region,

an n⁺ doped region,

a second intrinsic region between the n⁺ doped region and the second p⁺ doped region,

whereby a first P-l-N diode is formed by the first p⁺ doped region, the first intrinsic region, and the n⁺ doped region, and a second P-l-N diode is formed by the second p⁺ doped region, the second intrinsic region, and the n⁺ doped region, and

an array of blind holes, each blind hole extending from the first side of the first p⁺ doped region, through the first intrinsic region at least to an interface of the first intrinsic region and the n⁺ doped region and no further than an interface between the n⁺ doped region and the second intrinsic region, the array of blind holes arranged such that a lattice pattern is formed on the first side of the first p⁺ doped region;

b) a first electrical contact formed on the first p⁺ doped region; and

c) a second electrical contact formed on the second p⁺ doped region

whereby a first measure of photocurrent can be made with the first P-l- N diode by biasing the first contact negatively relative to the second contact, and a second measure of photocurrent can be made with the second P-l-N diode by biasing the first contact positively relative to the second contact,

wherein the lattice pattern of a first photodetector of the plurality of photodetectors has a first width and the lattice pattern of a second photodetector of the plurality of photodetectors has a second width, different to the first width.

In another embodiment, the invention provides a method of forming a

photodetector, the method comprising:

a) forming a substrate comprising:

a first p⁺ doped region having a first side on which light is intended to be incident when the photodetector is in use,

a second p⁺ doped region,

an n⁺ doped region, a first intrinsic region between the first p⁺ doped region and the n⁺ doped region,

b) forming a lattice pattern on the first side of the first p⁺ doped region by forming an array of blind holes in the substrate such that each blind hole extends from the first side of the first p⁺ doped region, through the first intrinsic region at least to an interface of the first intrinsic region and the n⁺ doped region and no further than an interface between the n⁺ doped region and the second intrinsic region;

c) forming a first electrical contact on the first p⁺ doped region; and d) forming a second electrical contact formed on the second p⁺ doped region.

In another embodiment, the invention provides a photodetector comprising: a substrate comprising:

a) a first n⁺ doped region having a first side on which light is intended to be incident when the photodetector is in use,

a second n⁺ doped region,

an p⁺ doped region,

a first intrinsic region between the first n⁺ doped region and the p⁺ doped region,

a second intrinsic region between the p⁺ doped region and the second n⁺ doped region,

whereby a first NIP diode is formed by the first n⁺ doped region, the first intrinsic region, and the p⁺ doped region, and a second P-l-N diode is formed by the second n⁺ doped region, the second intrinsic region, and the p⁺ doped region, and

an array of blind holes, each blind hole extending from the first side of the first n⁺ doped region, through the first intrinsic region at least to an interface of the first intrinsic region and the p⁺ doped region and no further than an interface between the p⁺ doped region and the second intrinsic region, the array of blind holes arranged such that a lattice pattern is formed on the first side of the first p⁺ doped region;

b) a first electrical contact formed on the first n⁺ doped region; and

c) a second electrical contact formed on the second n⁺ doped region,

whereby a first measure of photocurrent can be made with the first NIP diode by biasing the first contact positively relative to the second contact, and a second measure of photocurrent can be made with the second NIP diode by biasing the first contact negatively relative to the second contact.

In an embodiment, the array of blind holes is a rectangular array.

In an embodiment, an aperture of each of the blind holes is rectangular.

In an embodiment, each of the blind holes extends partially into the p⁺ doped region.

In an embodiment, each of the blind holes extends to the interface between the p⁺ doped region and the second intrinsic region.

In an embodiment, each intrinsic region is an rr doped region.

a) a substrate comprising

a first n⁺ doped region having a first side on which light is intended to be incident when the photodetector is in use,

a second n⁺ doped region,

an p⁺ doped region,

a second intrinsic region between the p⁺ doped region and the second n⁺ doped region, whereby a first NIP diode is formed by the first n⁺ doped region, the first intrinsic region, and the p⁺ doped region, and a second P-l-N diode is formed by the second n⁺ doped region, the second intrinsic region, and the p⁺ doped region, and

b) a first electrical contact formed on the first n⁺ doped region; and

c) a second electrical contact formed on the second n⁺ doped region,

whereby a first measure of photocurrent can be made with the first P-l- N diode by biasing the first contact positively relative to the second contact, and a second measure of photocurrent can be made with the second P-l-N diode by biasing the first contact negatively relative to the second contact,

In another embodiment, the invention provides a method of forming a

photodetector, the method comprising:

a) forming a substrate comprising:

a second n⁺ doped region,

an p⁺ doped region,

b) forming a lattice pattern on the first side of the first n⁺ doped region by forming an array of blind holes in the substrate such that each blind hole extends from the first side of the first n⁺ doped region, through the first intrinsic region at least to an interface of the first intrinsic region and the p⁺ doped region and no further than an interface between the p⁺ doped region and the second intrinsic region;

c) forming a first electrical contact on the first n⁺ doped region; and d) forming a second electrical contact formed on the second n⁺ doped region.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention are shown in the accompanying drawings in which: FIG. 1 is a schematic perspective view of a photodetector of an embodiment.

FIG. 2 is a partial top view of the photodetector of FIG 1.

FIG. 3 is a schematic side view of a photodetector of an embodiment.

FIG. 4 is a scanning electron microscope image of a prototype photodetector. FIG. 5 is a graph of simulated fractional absorption.

FIG. 6 illustrates the responsivity spectra of eight different photodetectors.

FIG. 7 compares a reconstruction of the spectrum to the actual spectrum.

FIG. 8 is a graph showing finite element method calculations of the fractional absorption.

FIG. 9 shows the finite element method calculated absorption spectrum for a wave guide array comprised of 75 nm wide waveguides.

FIG. 10 shows normalized optical absorption at three wavelengths.

FIG. 11 is a graph of responsivity and external quantum efficiency at different wavelengths.

FIG. 12 shows the current-voltage curves for the same detector as FIG. 11.

FIGs. 13A and 13B show normalized responsivities of twenty fishnet and twenty mesa detectors.

FIGs. 14A-14D compare spectra reconstructed by a prototype fishnet

microspectrometer chip with the spectra measured by a commercial

spectrometer. FIG. 15 shows computed color co-ordinates determined from spectra

reconstructed by the prototype fishnet microspectrometer chip and spectra measured by the commercial spectrometer.

FIG. 16 shows finite element method absorption simulations for a 400 nm period waveguide array.

FIGs. 17 and 18 show the effect of array period on the mode cut-off

wavelengths.

FIG. 19 is a schematic layout of a prototype microspectrometer chip.

DETAILED DESCRIPTION

Embodiments of the invention employ a nanostructured, high-refractive index doped-semiconductor substrate to create photodetectors. As explained in further detail below, dimensions of the photodetectors can be varied to respond to different wavelengths, such that a plurality of photodetectors can be used to form a spectrometer.

The doping profile described below provides two vertically-stacked, back-to-back photodiode regions which allows for double the amount of detection to be performed by a given on-chip footprint. By forming the top photosensitive regions with two sets of interleaved vertical slab waveguide arrays of a specific width and period, it is possible to define the absorption spectra (and thus responsivity spectra) of both the upper and lower photodiode regions.

In the embodiments, standard nanolithographic techniques are used to create a an array of vertical dielectric slab waveguides in a lattice pattern, with p⁺ or n⁺ doped regions at the top and bottom of each waveguide, with the mid-section of the waveguide made of intrinsic or lightly doped semiconductor. In this respect, herein any reference to an“intrinsic region” should be understood as a reference to an intrinsic or lightly doped region unless the context expressly requires otherwise.

The waveguide array acts as a first P-l-N photodiode. Beneath the waveguide array is another N-l-P doped region, without any nanostructuring, (referred to as the mesa) so that back to back P-l-N photodiodes are formed - that is the mesa region forms a second P-l-N photodiode. Electrical contacts are formed on the p⁺ doped regions of the waveguide region and the mesa region. In use, this enables bias voltages to be applied to the photodetector. Photocurrent can be generated and collected using the first and second photodiodes depending on the polarity of the applied bias.

An example structure of the photodetector 100 is shown in FIGs. 1 to 3. FIG. 1 is a schematic perspective view of a photodetector 100 of an embodiment. The top (or first) p⁺ doped layer 110 is intended to receive incident light. As can be seen best from FIG. 2, the light receiving surface has a lattice, or“fishnet” pattern 111 such that photodetector 100 comprises interleaved first 114, and second 115 orthogonal sets of vertically oriented silicon slab waveguides. The width 112 and period 113 of the sets of waveguides 114, 115 can be set so as to control the wavelengths of light to which the photodetector 100 responds. The set of waveguides 114, 115 are oriented out of the plane of the substrate, allowing direct coupling of incoming light to the silicon nanofeatures, thereby foregoing the need for momentum matching with on-chip couplers such as gratings or prisms13. The array period 113 and waveguide widths 112 of each photodetector are determined in the lithography step of the fabrication process, thereby enabling control over its target optical absorption.

The interconnected nature of the fishnet pattern 111 avoids the need for planarization involving precise deposition and etch-back steps and provides improved mechanical robustness compared with nanowire designs as well as an extra degree of freedom (the array period) to tailor each detector’s responsivity spectrum.

A first electrical contact 160 is formed on the first p⁺ doped layer 110 (shown here with an electrical take-off lead 175).

A first intrinsic layer 120 is sandwiched between the first p⁺ doped layer 110 and an n⁺ doped layer 130. A second intrinsic layer 140 is sandwiched between the n⁺ doped layer 130, and a second, p⁺ doped bottom layer (the“mesa”) 150 on which a second electrical contact 170 is formed (here shown with an electrical take-off lead 175).

FIG. 3 is a schematic cross-section of the photodetector 100 which incorporates a circuit representation 300 of the photodiode. As can be seen from FIG. 3, the lattice pattern 1 1 1 is formed by an array of blind holes (here blind holes 1 1 1 A,

1 1 1 B, 1 1 1 C, 1 1 1 D, 1 1 1 E can be seen), which extend from the top p⁺ doped layer 1 10 to the interface between the n⁺ doped layer 130 and the second intrinsic layer 140. In other embodiments, the blind holes 1 1 1A-1 1 1 E may terminate at the interface between the first intrinsic layer 120 and the n⁺ doped layer 130 or within the n⁺ doped layer 130.

As can be seen, from FIG.3, the circuit representation shows first 310 and second 320 photodiodes which can be used to measure current with an ammeter 330. These photodiodes 310,320 are also referred to as the“fishnet photodiode” 310 and the mesa photodiode 320 because of their respective relationships to the fishnet and mesa regions of the photodetector.

By tailoring the width 1 12 and period 1 13 of different waveguide arrays it is possible to control the guided-mode dispersion to maximize the absorption of light at a particular wavelength. FIG. 5 is a graph showing simulated fractional absorption using absorption scale 530 relative to wavelength in mm 520 and width in nm 530 of the lattice pattern for a fixed period of 500nm. It will be apparent that other periods can be used to achieve different responsivity.

Thus, embodiments of the invention can be used to make a spectrometer comprising a set of photodetectors on the same substrate where each

photodetector has a peak responsivity at a unique wavelength in the visible spectrum. In order to electrically isolate the photodetectors from one another, a mesa is etched from the embedded n+ layer to the bottom p+ layer below each photodetector.

By carefully measuring the responsivity of each photodetector with a known light source (measure once and store these values) and then measuring the photocurrent from each photodetector when the device is illuminated with an unknown light source it is possible to reconstruct the spectrum of the unknown source. In one example, the reconstruction of spectra is possible by treating the photo currents and measure responsivities as inputs to a linear least-squares problem using Tikhonov regularization as described in U.Kurokawa, B. I. Choi, ad C.C. Chang,“Filter-based miniature spectrometers: spectrum regularization using adaptive regularization”, IEEE Sensors Journal 1 , 1556-1563 (2011 ).

The above embodiment describes a P-l-N-l-P embodiment. Persons skilled in the art will appreciate that an N-l-P-l-N embodiment is also possible.

EXAMPLE 1

FIG. 4 is a scanning electron microscope image 400 of a prototype photodetector formed in accordance, which shows that electrical contact 170 sits on a pad mesa region 430 which is formed from the p⁺ doped layer. The lattice patterned, wave guide array 11 1 sits on mesa region 140. In FIG. 4 scale bar 404 is 50pm. Enlarged region 410 shows the lattice pattern more clearly. In the enlarged region, the scale bar 406 is 2pm.

In the example, a photodetector was formed from a starting p⁺ doped substrate on which a lightly doped rr layer 4 pm thick was formed. A n⁺ doped layer 1 pm thick was formed on the rr layer followed by a further rr layer 2 pm thick and the top p⁺ doped layer 0.2 pm thick. The lattice pattern was defined by e-beam lithography to define an aluminum etch mask (60 nm) for a subsequent ion reactive ion etching step which results in the blind holes being approximately 2.5 deep - i.e. extending partially into the n⁺ doped layer. The mesa 140 was produced by photolithography and reactive ion etching to a depth of

approximately 6 pm. Finally, aluminum contacts, were added by photolithography and evaporation.

Eight photodetectors were produced using this technique where the widths 112 were different. These photodetectors were termed“fishnet” pixels. Responsivities of the individual fishnet pixels (e.g. circuit 300 in FIG.3) were measured by illuminating them with light of known power and wavelength. This was provided by a laser driven white light source (Energetiq EQ-99X) coupled to a monochromator (Princeton Instr Acton SP2150). The photocurrent from each pixel vs wavelength for voltages of +1.5 and -1.5 V were recorded by a

picoammeter (Keithley 2502). The normalized results (FIG.6) show an

appreciable shift in peak responsivity wavelength 610 with increasing waveguide width 1. In this respect, line 631 corresponds to 107nm, line 632 corresponds to 117nm, line 633 corresponds to 122nm, line 634 corresponds to 127nm, line 635 corresponds to 132nm, line 636 corresponds to 137nm, line 637 corresponds to 142nm, and line 638 corresponds to 147nm. The incident spectra was

reconstructed by treating the measured photocurrents and measured

responsivities as inputs to a linear least-squares problem nemploying Tikhonov regularization for solution stability. The device was illuminated with light from a supercontinuum source (NKT SuperK Compact) that was passed through two shortpass filters and a dichroic mirror. The spectrum was reconstructed using the photocurrents from 20 pixels (i.e. 40 photocurrent measurements). The results shown in FIG.7 show the reconstruction 732 is in good agreement with the spectrum 731 measured by a reference spectrometer (Ocean Optics QE-Pro). This is especially true for the wavelength and linewidth of the main peak, while the secondary peak is only partially resolved. It should be noted that the wavelength-dependence of the reference system (QE-Pro) has not been accounted for.

EXAMPLE 2

In this example, a prototype microspectrometer chip having twenty‘fishnet pixels’ was formed (23 fishnet pixels were made but 3 failed). The example also demonstrates the reconstruction of four test spectra using a two-stage

supervised machine learning based reconstruction algorithm.

In this example, the doped silicon layers were epitaxially grown (layer-by-layer) on degenerately doped p+ 4” Si substrates by the company IQE of St Mellons, Cardiff, United Kingdom. Initially 4 pm of lightly n-doped silicon (2x10¹⁶

atoms/cm³) was grown, followed by a 1 pm n+ layer (6x10¹⁸ atoms/cm³), thereby forming the first p-i-n junction which will act as the mesa detector.

Another 2 pm of n- Si was grown and capped with a 200 nm layer of p+ Si, forming the second p-i-n junction that will form the fishnet photodiode region. Electrical contacts were added to the bottom p+ substrate and the top p+ layer to bias the structure and to form two counter-facing PIN photodiodes. In the absence of light the current through this device under forward bias (top electrode positively biased with respect to the bottom electrode) should only be equal to the dark (reverse bias) current of the bottom (mesa) photodiode. When the entire device is reverse biased however (top electrode negatively biased with respect to the bottom electrode), the current is equal to the dark current of the top diode.

To fabricate the 23 fishnet pixels, standard silicon processes were used. First, an Al etch mask (65 nm thick) was created via lift-off using electron beam lithography (100 kV Vistec EPBG5000+) and electron beam evaporation (IntIVac NanoChrome II). Inductively coupled plasma reactive ion etching (ICP- RIE, Oxford Instruments PlasmaLab 100) was used to create the Si

nanofeatures comprising each fishnet photodetector. Here 40 seem of SFe and 90 seem of C4F8 were used in a carefully calibrated pseudo-Bosch etch recipe to create silicon nanofeatures with smooth, vertical sidewalls with an aspect ratio greater than 30. Each fishnet detector region was etched to a depth greater than 2.2 pm to ensure the intrinsic region of each upper p-i-n junction is fully patterned. Failure to etch through the top intrinsic region would result in large contributions to the photocurrent from light absorbed in the remaining unpatterned intrinsic silicon. This fishnet design is also amenable to fabrication via metal-assisted chemical etching as an alternative to ICP-RIE. After etching, the Al mask was chemically removed (Transene Al etchant) and photoresist (AZ4562) was spin-coated to a thickness 8 pm. Direct-write UV lithography (IMP SF100) was used to define an etch mask for the mesa etching step, protecting both the fishnet and pad regions of each pixel. Mesas with thickness 5.5 pm were then etched using the same ICP-RIE tool. Direct-write UV lithography and evaporation were again used to define a lift-off mask for the final metallization step in which Al pads (200 nm thick) were added to the pad mesa of each pixel. Finally, the silicon die was mounted and wire-bonded (FS BondTec, 20 pm AlSi wire) in a ceramic DIP-24 chip carrier, with one pin bonded to a common contact on the substrate p+ layer.

FIG. 19 is a schematic layout of the resultant prototype microspectrometer chip 1900 showing 23 photodetector devices 1901 -1923 arranged in a square configuration. Pad 1931 is the common bottom contact all the photodetector devices 1901 -1923 share, which is on the bottom p+ layer.

Electrical characterization of the photodetector devices 1901 -1923 was carried out using a picoammeter (Keithley 6485). To measure the responsivity of each detector, its photocurrent in response to illumination by monochromatic light of a known wavelength and power was recorded. The illumination was provided by a laser driven light source (Energetiq EQ-99X) filtered by a monochromator (Princeton Instruments Acton SP2150) and focused onto each device with a microscope objective (Nikon LU Plan Fluor, 0.30 NA, 10x). The wavelength range of the measurements was 400 nm to 1000 nm, in steps of 5 nm. At each wavelength, the photocurrent from the fishnet photodiode was measured by supplying a bias to the top electrode of -1.0 V. Similarly, the photocurrent from the mesa photodiode was measured by biasing the device at +1.0 V.

After measuring the responsivities of each fishnet and mesa photodiode, it is possible to use the device to reconstruct unknown spectra. To demonstrate this a white LED lamp (Thorlabs MCWHLP1 ) was used to illuminate the entire chip and the photocurrents from each pixel, generated with +1V and -1V bias, were recorded. These measured photocurrents were then combined with the known detector responsivities to reconstruct an estimate of the unknown spectrum. In the example, a two-stage reconstruction method was used. The first stage employed a modified Tikhonov regularization method described in a paper by Kurokawa, U.; Choi, B. I.; Chang, C.-C. in IEEE Sensors Journal 2011 , 11 , (7), 1556-1563 and also known as weight decay or ridge regression. The output from this calculation was then refined using a simulated annealing algorithm to improve the final reconstructed spectrum using a method outlined in a paper by Redding, B.; Popoff, S. M.; Cao, H. in Optics Express 2013, 21 , (5), 6584-6600. Simulated annealing can be used with an y initial guess spectrum, however using the regularized spectrum as the initial guess resulted in substantial reduction to the required computational time.

More specifically, in order to reconstruct the spectrum, an approach is to consider N detectors illuminated with an unknown spectrum 5(A). The responsivities of the detectors are known and denoted /^(A). The generated photocurrent for the i^th detector is

Pi = f Ri(A)S(A)dA, i E {1,2, ... , N} (S1 )

In order to reconstruct the spectrum, equation S1 is solved for 5(A). As responsivities are measured in a discrete fashion (5 nm steps in this case) it is necessary to discretize S1 into M wavelengths. Doing so yields

P = RS (S2)

Where P and 5 are JV x 1 and M x 1 vectors, respectively, and R is an N x M matrix with element ij corresponding to the i^th detector and j^th wavelength, A,·.

To solve for 5 by inverting R in this case is only possible when N = M and other methods will fail when N < M, since there are more unknown values than simultaneous equations to solve. However, it is possible to apply a

transformation matrix, T, to R and 5 to reduce the number of unknowns in the system. Here following the method described in Kurokawa et at., T is set to be comprised of a set of Gaussian vectors centered on A,· such that R_r = RT and 5 = TS_r, where R_r and S_r are the reduced responsivity matrix and unknown spectrum vector, respectively. Clearly T must be a M x N matrix so that R_r is an N x N matrix and S_r is a N x 1 vector. Equation (S2) then becomes

P = R_rS_r (S3) which could be solved for _r and (mapped back to S, the unknown spectrum, using 5 = TS_r), with the method of least squares to find min

or

S

perhaps direct inversion if det(R_r) ¹ 0. This example uses two sets of

Gaussian basis vectors to form T, with one set used to represent the fishnet photodiode responsivities and the other set the mesa photodiode responsivities. The twenty fishnet Gaussian vectors were centred at wavelengths from 400 nm to 550 nm, with 5 nm steps and assign an FWHM of 85 nm. The 20 mesa vectors, span a range of 550 nm to 800 nm with 5 nm steps and a width of 550 nm. These values roughly correspond to the measured responsivity curves but were found via optimization.

Since both responsivities and photocurrent measurements will contain random observation noise, a Tikhonov regularization is used (also known as weight decay or ridge regression). The regularization parameter is found using the L- curve method, and the values of S_r are restricted to be non-negative since negative spectral values are unphysical.

To improve the reconstruction further, simulated annealing was implemented step with the regularized output as an initial guess. The algorithm of Redding et al was used to achieve improved reconstructions. In this example, using a linear cooling schedule with an initial temperature of 10,500, with a random multiplier between 0.4 and 1.3 and stop the algorithm either when the system energy, that is the difference between the measured photocurrents and calculated

photocurrents using equation (S2) with our reconstructed spectrum, drops below 0.0005 or after 1200 iterations, when the temperature reaches 0. To improve the reliability and reproducibility of the simulated annealing algorithm used by Redding et al, the algorithm is run 1000 times and the reconstructed spectrum with the final system energy closest to the median final system energy of the 1000 trials is used.

As indicated above, tailoring the optical absorption of a given fishnet pixel (photodetector) can be achieved by appropriate choice of the array period and widths of the interleaved silicon fins. (Or put another way by creating blind holes of an appropriate size in the substrate.) The dispersion relation for each pixel can be found by modelling the fishnet region as two interleaved sets of orthogonal, strongly absorbing silicon high contrast gratings (HCG). This differs from the usual applications of HCGs, e.g. very high reflectance mirrors for semiconductor lasers, for which the HCG material has very low absorption losses due to sub-bandgap operation. It can be shown that the cut-off wavelength, X_c, for transverse magnetic (TM) modes within a silicon high- contrast grating in air must satisfy

where n_si is the refractive index of silicon, s is the width of the silicon

waveguides and L is the period of the array. Transverse electric (TE) mode cut- offs can be found by multiplying the first term in equation (1 ) by n_s ² _i. Mode cut- off wavelengths can therefore be tuned by varying both waveguide width and waveguide array period.

FIG. 8, is a graph showing finite element method (FEM, COMSOL) calculations of the fractional absorption of normally incident light for an array of 2.7 pm tall silicon waveguides in air, with a fixed period of 300 nm. The fractional absorption 830 is only computed for the part of each waveguide (2 pm tall) that would correspond to the low-doped region of our device. Absorption of light in the heavily doped p+ and n+ layers is not expected to contribute significantly to the measured photocurrent in each detector as the photogenerated charge carriers would rapidly recombine in these regions. The widths 810 of the vertical waveguides in the array are varied from 40 nm to 260 nm and the wavelength 820 of light is varied from 400 nm to 900 nm, encompassing most of the silicon detection window. The curves 841 , 842, 843 show the cut-off wavelengths for TM-modes TM2, TM4 and TM6 as a function of waveguide width for a silicon WGA calculated using equation (1 ), with L = 300 nm. TM0 is the fundamental mode and has no cut-off while odd TM or TE modes can only be excited with off-normal incidence illumination. It can be seen that the fractional absorption in each waveguide peaks near unity when illuminated with a wavelength around that of the mode cut-off wavelength. It can also be seen that the cut-off wavelength and absorption peak wavelength both redshift, in a near-linear fashion, with increasing waveguide width. This can be exploited to tailor the responsivity the fishnet photodetectors, as the fishnet is merely comprised of two orthogonal sets of interleaved WGAs. The fishnet structures consist of two arrays of vertical dielectric slab

waveguides (orthogonal to one another). Thus, the optical response of our fishnet structures can be understood by considering them to be high contrast gratings (FICGs). The work of Chang-Flasnain, C. J.; Yang, W. Advances in Optics and Photonics 2012, 4, (3), 379-440 provides explicit details on the physics of such structures, including a derivation of dispersion relation. The primary difference between a stand-alone vertical slab waveguide and a near- wavelength waveguide array (WGA) is that for the stand-alone case near the cut-off wavelength of a mode, energy from that mode is mostly directed out of the waveguide sidewalls that is, the fields close to cut-off become spread out, well beyond the core of the waveguide. For the WGA case near cut-off the fields from each waveguide mode extend well beyond the sidewalls of one waveguide and into the neighboring waveguide through its sidewalls and so on. Therefore, near a mode cut-off wavelength the guided light interacts with multiple neighboring waveguides in the array.

When operating the waveguide where the materials optical properties lead to strong absorption of light, such as visible light in silicon, this process can lead to very large fractional absorption coefficients. For example, FIG. 16 shows finite element method (FEM) absorption 1620 simulations for a 400 nm period WGA illuminated with normally incident TE polarized light. The curves

161 1 , 1612, 1613 show the cut-off wavelengths for the TE2 161 1 , TE4 1612 and TE6 1613 modes, as given by FICG theory. It can be seen that similar to the TM mode behavior, the fractional absorption peaks near the calculated mode cut-off wavelengths.

FIG.17 shows the TM2 mode wavelength cut-off 1720 as a function of waveguide thickness 1710 for six different array periods, namely 200nm 1732, 300nm 1733, 400nm 1734, 500nm 1735, 600nm 1736 and 700nm 1737. It is clear that as period increases, the TM2 cut-off wavelengths red-shift. This is a useful feature, since it is much easier to control the array period during fabrication than controlling the waveguide widths that result after etching.

FIG. 18 shows the cut-off wavelengths for several (even) TM 1731 -1734 and TE 1741 -1743 modes as a function of period 1810, L, for a set of 135 nm wide waveguides. Flere the modesl 731 -1734, 1741 -1743 are mostly non-degenerate except at certain values of L and when moving to higher order modes, the mode spacing reduces. This leads to wider absorption bands at these wavelengths. Fishnet pixels, in contrast to WGAs, should simultaneously support both TM and TE modes under normally incident linearly polarized illumination. It can also be seen that as the WGA modes become closely spaced, the absorption peaks associated with each mode begin to overlap. This is especially evident for higher order modes. FIG. 9 shows the FEM calculated absorption spectrum 910 for a WGA comprised of 75 nm wide waveguides (indicated by circle 850 in FIG. 8. The dashed line 920 indicates the first order TM mode cut-off wavelength,

A_c = 570 nm, calculated using equation (1 ), which coincides with the peak wavelength of the simulated absorption spectrum. The spatial distribution of the normalized optical absorption at three wavelengths: 430 nm 931 , 570 nm 932 and 800nm 913, are shown in FIG. 10. These wavelengths are respectively below, at, and above the cut-off wavelength for the 75 nm wide WGA. The calculated power absorption density is normalized to the total power of the exciting plane wave. One can find the total fractional absorption, A, by

integrating this quantity (with units (pm) ²) over the device cross-section. Note that quantities plotted in for the 430nm wavelength 931 and the 800mm wavelength 933 are multiplied by 2 and 20 respectively so that the same scale 1010 as used for the 570nm wavelength 932 can be used. The 430nm

wavelength 931 , is well below the cut-off wavelength of modes TM0 and TM2. The waveguides are thus dual-moded (TM0 and TM2 both propagate) and most of the absorption of these modes is associated with the shorter penetration depth and greater absorption of short wavelength visible light in silicon, as would be the case for an unpatterned Si photodiode. The 570nm wavelength 932 is near the cut-off wavelength of TM2. The absorption thus mostly occurs within the top half of the waveguides. This is because near the TM2 cut-off wavelength, the guided light has a large internal angle of incidence, increasing the effective path length traversed and boosting the total absorption. Whereas the 800nm wavelength 933 is well above the cut-off, the waveguides are single- moded and freely propagate the incident light to the mesa below with little loss in the waveguide region. It is expected that the photocurrent generated in the waveguide regions of the upper detector to exhibit responsivity peaks for shorter wavelengths and the mesa responsivities to peak in the red part of the visible spectrum, where the WGAs are single-moded. Also, of practical note is that the absorption density is larger at the silicon-air interfaces than in the core of the silicon waveguides. This is due to the boundary conditions imposed by arraying the slab waveguides, restricting the dispersion relation for the WGA, in effect forcing the modes to propagate part in air and part in silicon. The result of this is that a large proportion of the photogenerated carries will be near the surface of the waveguides, making them susceptible to surface trap states, which are a natural consequence of the fabrication process and will reduce the detector’s external quantum efficiency.

FIG. 11 is a graph of responsivity 1110 and external quantum efficiency (EQE)

1120 at different wavelengths 1130 for a fishnet WGA having width and period of 105 nm and 375 nm. FIG. 11 shows the measured responsivity 1141 and EQE 1142 for the fishnet detector, and the measured responsivity 1 141 and EQE 1142 for the fishnet detector. It will be apparent from FIG. 11 that the peak responsivity of the fishnet detector is 88 mA/W and occurs at a wavelength of 480 nm. For the mesa detector, the peak responsivity is 50 mA/W, and occurs at a wavelength of 650 nm. It should also be noted that a wavelength for which the fishnet responsivity is high, the mesa responsivity is low and vice-versa.

This is because light that reaches the mesa detector must first pass through the fishnet region, which in this case acts as a passive filter, akin to conventional dye-based or thin film color filters. The maximum EQE for the fishnet detector (shown in FIG. 11 ) is 0.25, whereas the maximum measured mesa EQE from all fabricated detectors is 0.28 at 705 nm (not shown here). These values could be improved by passivating the sidewalls of the waveguide, to reduce the interaction of charge carriers with surface trap states and recombination sites.

FIG. 12 shows the current-voltage curves 1210,1220 for the same detector as FIG. 11 with 1210 and without 1220 illumination from the white LED lamp. A negative bias voltage corresponds to the fishnet region being reversed bias and acting as a photodetector and the mesa acting like a forward biased diode, and vice-versa for a positive bias. For this fishnet detector, the current measured at - 1 V with illumination from the LED lamp was ~100 times larger than the dark current. For this mesa detector, the current measured at +1 V with LED lamp illumination was ~6 times larger than the dark current. Electrical characterization of all 23 pixels revealed that three (3) were defective, leaving 20 viable fishnet and mesa detectors for use in the microspectrometer chip. The inset 1230 of FIG. 12 is the response of the fishnet detector to light from the monochromator at a wavelength of 560 nm that has been optically chopped at 83 Hz. It can be seen that the response is a square-wave. Some earlier prototypes showed very slow transient response both for the case of modulated illumination and for the case when the illumination was fixed, but the bias voltage switched. In addition, this temporal behavior was not stable. This behavior was attributed to electron beam damage of the native oxide layer during electron beam lithography introducing charge trap states between the contact pad and the upper p+ layer. This issue was corrected by a brief RIE etch of the native oxide layer prior to forming the top contact pads.

The normalized responsivities of the twenty fishnet 1310 and twenty mesa 1320 detectors are shown as FIGs. 13A and 13B, respectively. As discussed, each detector contains a fishnet structure with unique geometric parameters, that is, waveguide width and array period, (s, A). For convenience, the term“fishnet number” is used to identify each detector (in FIGs. 13A and 13B). The geometric parameters of each fishnet, i.e. of each of the twenty pixels of the prototype microspectrometer chip, are set out in Table 1.

The peak responsivity wavelength for the fishnet detectors, that is, when the bias applied to each pixel is negative, shifts approximately linearly from 400 nm to 580 nm with increasing fishnet number. When a positive bias is applied, the measured photocurrent originates from the mesa detectors. From FIG. 13B, it can be seen that the responsivities of these detectors span the wavelength range 580 nm to 850 nm, with the center wavelengths shifting from 730 nm down to 660 nm with increasing fishnet number. The complementary nature of the responsivities of each fishnet and mesa detector pair allows us to collect spectral information from the visible to the near-infrared. It should also be noted that the linewidths (full-widths-at-half-maximum, FWHMs) of the responsivity spectra of the fishnet and mesa photodiodes of each pixel are different. For example, for the pixel with fishnet number 9, the FWHMSs are 160 nm and 280 nm for the fishnet and mesa photodiodes, respectively. This is an important issue to consider when implementing the reconstruction algorithm.

Table 1

After measuring all 40 responsivity spectra of the photodiodes of the

microspectrometer chip (20 fishnet and 20 mesa devices), the chip was illuminated with four different test spectra. The first was provided by a white light LED. The other three were generated by passing the output of the white light LED through colored glass filters. Test spectra were generated in this way because these spectra contain both narrow and broad spectral features. Such spectra are likely to be more representative of the spectra of an application of the prototype device, namely measuring the reflection spectra of materials (e.g. pigments and vegetation) to identify them, as opposed to very narrow (e.g. from laser) or very broad (e.g. from blackbody) test spectra. The photocurrents generated in each fishnet and mesa photodiode were then collected and input (with the responsivity spectra) to a two-stage reconstruction algorithm. For comparison the test spectra were also measured with a commercial

spectrometer (Ocean Optics QEPro). The results are shown in FiGs. 14A-D.

It can be seen that the spectra 1410A-D reconstructed by the prototype fishnet microspectrometer chip are in reasonable agreement with the spectra 1420A-D measured by the commercial spectrometer. In particular, the center

wavelengths of the peaks in the test spectra are accurately determined. In addition, for the spectra with two peaks (FIGs. 14A-B)), the relative intensities of the peaks are reconstructed correctly. Furthermore, the FWFIM linewidths of most spectral peaks reconstructed by the prototype fishnet microspectrometer are in agreement with those measured by the commercial spectrometer.

To quantify the accuracy of a reconstructed spectrum 1410 the mean-square- error (MSE) was computed between it and the spectrum measured with the commercial spectrometer. The MSE values are 0.011 , 0.067, 0.018, and 0.014 for FIGs 14A-14D respectively. This indicates that the commercial and fishnet- based spectrometer are in good agreement across the visible spectrum. The red test spectrum (FIG. 14B) exhibited the largest error, due in part to the wider responsivity peaks and therefore lower spectral resolution associated with the mesa photodetectors sensitive in this region.

FIG. 15 shows the computed CIE 1931 xy color co-ordinates determined from the spectra measured by the commercial spectrometer 1420A-D and

reconstructed by the prototype fishnet microspectrometer chip 1410A-D. A D65 standard ilium inant was used in the spectral conversion, with the white point located at (0.3127,0.329). For the white LED lamp (FIG. 14A), the measured and reconstructed spectra give color co-ordinates of (0.33,0.34) and (0.32,0.31 ) respectively. Similarly, for the red, green and blue spectra (FIGs. 14B-D), calculated color values are also in agreement. The red spectrum (FIG. 14B) gives the largest discrepancy, which is consistent with the fact that its MSE is the largest. The mean Euclidean distance between measured and

reconstructed color value pairs is 0.056. In other examples, the reconstruction accuracy could be improved by increasing the number of fishnet detectors in some examples, in excess of 100 fishnet detectors may be appropriate.

In other embodiments, the spectral operating range can be increased to reach longer wavelengths. This can be achieved by increasing the widths and periods of the WGA to shift the mode cut-offs to longer wavelengths. A silicon process as described herein could extend the spectral range up to a wavelength of ~1100 nm. Beyond this wavelength the base material, Si, will need to be replaced with appropriate semiconductors, such as Ge (for a spectrometer operating over the 800-1800 nm range).

However, in the configuration presented here, with two stacked photodiode regions, the upper intrinsic silicon layer thickness would need to be increased, because the absorption coefficient of silicon decreases at longer wavelengths. We also note that as the WGA width is increased, it will begin to support an increasing number of guided modes, meaning that each responsivity spectrum will exhibit multiple peaks. The reconstruction algorithm itself can work with multiple responsivity modes and in principle for any R(A) as long as equation S2 is valid. Here Tikhonov regularization is to solve S2 but there are many related alternatives for solving S2 which may give marginal improvements to the final estimated spectrum under certain conditions, but have their own drawbacks. Further increases in the spectral operating range could be achieved via applying our approach to other semiconductors such as germanium, GaN, and InGaAs.

EXAMPLE EMBODIMENTS

An example embodiment provides a photodetector comprising:

a substrate comprising two vertically stacked, back-to-back photodiode regions comprising:

a) a first p⁺ doped region having a first side on which light is intended to be incident when the photodetector is in use, a second p⁺ doped region,

an n⁺ doped region,

whereby a first photodiode is formed by the first p⁺ doped region, the first intrinsic region, and the n⁺ doped region, and a second photodiode is formed by the second p⁺ doped region, the second intrinsic region, and the n⁺ doped region, and

wherein the first photodiode region comprises two sets of interleaved vertical dielectric waveguides having a width and period corresponding to a target optical absorption;

b) a first electrical contact formed on the first p⁺ doped region; and c) a second electrical contact formed on the second p⁺ doped region.

whereby a first measure of photocurrent can be made with the first photodiode by biasing the first contact negatively relative to the second contact, and a second measure of photocurrent can be made with the second photodiode by biasing the first contact positively relative to the second contact.

In an embodiment, the two sets of interleaved vertical dielectric waveguides define a lattice pattern.

In an embodiment, the two sets of interleaved vertical dielectric waveguides define an array of blind holes is a rectangular array.

In an embodiment, each intrinsic region is an rr doped region. Another example embodiment provides a spectrometer comprising a plurality of photodetectors,

wherein the two sets of interleaved vertical dielectric waveguides of a first photodetector of the plurality of photodetectors corresponds to a first target optical absorption and the two sets of interleaved vertical dielectric waveguides of a second photodetector of the plurality of photodetectors corresponds to a second target optical absorption, different to the first target optical absorption.

In an embodiment, the width of the lattice pattern of the first photodetector is different to the width of the lattice pattern of the second photodetector.

In an embodiment, the period of the lattice pattern of the first photodetector is different to the period of the lattice pattern of the second photodetector.

Another example embodiment, there is provided a method of forming a

photodetector, the method comprising:

a) forming a substrate comprising two vertically stacked, back-to- back photodiode regions comprising:

a second p⁺ doped region,

an n⁺ doped region,

b) forming the first photodiode to comprise two sets of interleaved vertical dielectric waveguides having a width and period corresponding to a target optical absorption;

c) forming a first electrical contact on the first p⁺ doped region; and forming a second electrical contact formed on the second p⁺ doped region.

Another example, embodiment provides a photodetector comprising:

a second n⁺ doped region,

an p⁺ doped region,

whereby a first photodiode is formed by the first n⁺ doped region, the first intrinsic region, and the p⁺ doped region, and a second photo diode is formed by the second n⁺ doped region, the second intrinsic region, and the p⁺ doped region, and

b) a first electrical contact formed on the first n⁺ doped region; and c) a second electrical contact formed on the second n⁺ doped region,

whereby a first measure of photocurrent can be made with the first photodiode by biasing the first contact positively relative to the second contact, and a second measure of photocurrent can be made with the second photodiode by biasing the first contact negatively relative to the second contact.

In an embodiment, the two sets of interleaved vertical dielectric waveguides define an array of blind holes is a rectangular array. In an embodiment, each of the blind holes extends partially into the p⁺ doped region.

In an embodiment, each intrinsic region is an rr doped region.

In another example embodiment, there is provided a spectrometer comprising a plurality of photodetectors, wherein the two sets of interleaved vertical dielectric waveguides of a first photodetector of the plurality of photodetectors corresponds to a first target optical absorption and the two sets of interleaved vertical dielectric waveguides of a second photodetector of the plurality of photodetectors corresponds to a second target optical absorption , different to the first target optical absorption

In an emdbodiment, the width of the lattice pattern of the first photodetector is different to the width of the lattice pattern of the second photodetector.

A further example embodiment provides a method of forming a photodetector, the method comprising:

a) forming a substrate comprising:

a second n⁺ doped region,

an p⁺ doped region,

UTILITY

Commercial applications of a compact, low-cost and reliable spectrometers formed using the on-chip design described above included use in unmanned aerial vehicles, machine vision applications, agricultural applications such as gauging plant health, enhanced colour matching (important for construction, interior design, and insurance industries) and material identification (plastics, food and drink) with a hand held device, or even integrated into a smart phone. Measuring the spectrum of a product at several points on a production line can allow for automated quality control and help determine where and when faults and defects occur. Current spectrometer systems are far too heavy, bulky and expensive for these applications.

An advantage of photodetectors of the embodiment of the invention is that it combines both filter and detector. Another advantage of such photodetectors is that they enable the production of a more compact, lightweight, alignment-free spectrometer chip. A further advantage of certain embodiments of the invention is that they are formed of structured silicon and one layer of metallisation, which should lead to reduced fabrication costs compared to typical grating based spectrometers or other proposed microspectrometer designs.

The realization of on-chip microspectrometers would allow spectroscopy and colorimetry measurement systems to be readily incorporated into platforms for which size and weight are critical, such as consumer grade electronics, smartphones and unmanned aerial vehicles. This would allow them to find use in diverse fields such as interior design, agriculture and in machine vision applications. While the invention has been described with respect to the figures, it will be appreciated that many modifications and changes may be made by those skilled in the art without departing from the spirit of the invention. Any variation and derivation from the above description and figures are included in the scope of the present invention as defined by the claims.

In the claims which follow and in the preceding description of the invention, except where the context requires otherwise due to express language or necessary implication, the word“comprise” or variations such as“comprises” or “comprising” is used in an inclusive sense, i.e. to specify the presence of the stated features but not to preclude the presence or addition of further features in various embodiments of the invention.

Claims

1 . A photodetector comprising:

a substrate comprising:

a second p⁺ doped region,

an n⁺ doped region,

an array of blind holes, each blind hole extending from the first side of the first p⁺ doped region, through the first intrinsic region at least to an interface of the first intrinsic region and the n⁺ doped region and no further than an interface between the n⁺ doped region and the second intrinsic region, the array of blind holes arranged such that a lattice pattern is formed on the first side of the first p⁺ doped region ;

b) a first electrical contact formed on the first p⁺ doped region; and c) a second electrical contact formed on the second p⁺ doped region whereby a first measure of photocurrent can be made with the first P-l-N diode by biasing the first contact negatively relative to the second contact, and a second measure of photocurrent can be made with the second P-l-N diode by biasing the first contact positively relative to the second contact.

2. The photodetector as claimed in claim 1 , wherein the array of blind holes is a rectangular array.

3. The photodetector as claimed in claim 1 or claim 2, wherein an aperture of each of the blind holes is rectangular.

4. The photodetector as claimed in any one of claims 1 to 3, wherein each of the blind holes extends partially into the n⁺ doped region.

5. The photodetector as claimed in any one of claims 1 to 3, wherein each of the blind holes extends to the interface between the n⁺ doped region and the second intrinsic region.

6. The photodetector as claimed in any one of claims 1 to 5, wherein each intrinsic region is an rr doped region.

7. A spectrometer comprising a plurality of photodetectors, each

photodetector comprising:

a) a substrate comprising

a second p⁺ doped region,

an n⁺ doped region,

b) a first electrical contact formed on the first p⁺ doped region; and

c) a second electrical contact formed on the second p⁺ doped region, whereby a first measure of photocurrent can be made with the first P-l-N diode by biasing the first contact negatively relative to the second contact, and a second measure of photocurrent can be made with the second P-l-N diode by biasing the first contact positively relative to the second contact, and

wherein the lattice pattern of a first photodetector of the plurality of photodetectors corresponds to a first wavelength and the lattice pattern of a second photodetector of the plurality of photodetectors corresponds to a second wavelength, different to the first wavelength.

8. The spectrometer as claimed in claim 7, wherein the width of the lattice pattern of the first photodetector is different to the width of the lattice pattern of the second photodetector.

9. The spectrometer as claimed in claim 7 or claim 8, wherein the period of the lattice pattern of the first photodetector is different to the period of the lattice pattern of the second photodetector.

10. A method of forming a photodetector, the method comprising:

a) forming a substrate comprising:

a second p⁺ doped region,

an n⁺ doped region,

c) forming a first electrical contact on the first p⁺ doped region; and a) forming a second electrical contact formed on the second p⁺ doped region.

11. A photodetector comprising:

a) a substrate comprising

a second n⁺ doped region,

an p⁺ doped region,

an array of blind holes, each blind hole extending from the first side of the first n⁺ doped region, through the first intrinsic region at least to an interface of the first intrinsic region and the p⁺ doped region and no further than an interface between the p⁺ doped region and the second intrinsic region, the array of blind holes arranged such that a lattice pattern is formed on the first side of the first p⁺ doped region; b) a first electrical contact formed on the first n⁺ doped region; and c) a second electrical contact formed on the second n⁺ doped region, whereby a first measure of photocurrent can be made with the first NIP diode by biasing the first contact positively relative to the second contact, and a second measure of photocurrent can be made with the second NIP diode by biasing the first contact negatively relative to the second contact.

12. The photodetector as claimed in claim 11 , wherein the array of blind holes is a rectangular array.

13. A photodetector as claimed in claim 11 or claim 12, wherein an aperture of each of the blind holes is rectangular.

14. A photodetector as claimed in any one of claims 11 to 13, wherein each of the blind holes extends partially into the p⁺ doped region.

15. A photodetector as claimed in any one of claims 11 to 14, wherein each of the blind holes extends to the interface between the p⁺ doped region and the second intrinsic region.

16. A photodetector as claimed in any one of claims 11 to 15, wherein each intrinsic region is an rr doped region.

17. A spectrometer comprising a plurality of photodetectors, each

photodetector comprising:

a) a substrate comprising

a second n⁺ doped region,

an p⁺ doped region,

an array of blind holes, each blind hole extending from the first side of the first n⁺ doped region, through the first intrinsic region at least to an interface of the first intrinsic region and the p⁺ doped region and no further than an interface between the p⁺ doped region and the second intrinsic region, the array of blind holes arranged such that a lattice pattern is formed on the first side of the first p⁺ doped region; b) a first electrical contact formed on the first n⁺ doped region; and c) a second electrical contact formed on the second n⁺ doped region, whereby a first measure of photocurrent can be made with the first P-l-N diode by biasing the first contact positively relative to the second contact, and a second measure of photocurrent can be made with the second P-l-N diode by biasing the first contact negatively relative to the second contact,

18. The spectrometer as claimed in claim 17, wherein the width of the lattice pattern of the first photodetector is different to the width of the lattice pattern of the second photodetector.

19. The spectrometer as claimed in claim 17 or claim 18, wherein the period of the lattice pattern of the first photodetector is different to the period of the lattice pattern of the second photodetector.

20. A method of forming a photodetector, the method comprising:

a) forming a substrate comprising:

a second n⁺ doped region,

an p⁺ doped region,

c) forming a first electrical contact on the first n⁺ doped region; and e) forming a second electrical contact formed on the second n⁺ doped region.