GB2306772A - Radiation detector - Google Patents

Description

2306772 RADIATION DETECTOR The present invention relates to a

semiconductor detector for electromagnetic radiation.

Many different semiconductor structures are known which are capable of detecting electromagnetic radiation. However, the detector or mixer according to the present invention is particularly adapted to detection of electromagnetic radiation in the submillimetre, far-infrared or midinfrared wavelength ranges. However, in is broadest sense, the present invention is not limited to operation at any particular microwave, infrared or visible wavelength or range of wavelengths and the terms "opticaP and "light" are to be interpreted as encompassing all of these.

It is well known that quantum-effect devices can be made by arranging layers of semiconductor layers of different band gaps, such that it is possible to induce a quantum well adjacent an interface. Carriers can be confined in the well layer so that the current flowing in the well layer, between source and drain regions can be modulated by means of a control potential applied to any overlying gate electrode. The carriers, usually of high mobility, may exist in two dimensions to behave as a "twodimensional electron gas" (2DEG) or they may be influenced by means of applied electrode potentials to exist substantially in only one dimension, i.e. as a "onedimensional electron gas" (1DEG) otherwise sometimes referred to as a "quantum wire".

Of course, the majority carriers can be electrons or holes so equivalent devices utilising a two-dimensional hole gas (2DHG) or one-dimensional hole gas (I DHG) can also be realised. However, for simplicity, the generic terms 2DEG or I DEG or simply 1 fcarrier gas" will be used herein and should be understood as encompassing both possibilities, unless specifically indicated to the contrary.

As an extrapolation of such confinement, it is possible to arrange barrier potentials in three dimensions to confine a puddle of 100 or so electrons. llis puddle is commonly referred to as a "quantum doC or quantum box". In this structure, electron motion is confined in all three dimensions.

This kind of quantum dot confinement has previously been realised by placing four or more depleting Schottky gates in a turnstile arrangement on the surface of a semiconductor heterostructure which itself contains a 2DEG. The gates squeeze the electron gas such that the remaining two degrees of freedom are impaired. This additional squeezing or confinement produces tunnelling barriers around the dot through which electrons must pass if they are to enter or exit the dot. Such transport through the dot is then affected by applying an external current or voltage to the dot. This external bias raises the energy of electrons, allowing them to pass over andlor tunnel through the tunnelling barriers.

Figures 1 and I A of the accompanying drawings show the heterostructure 111, in which a substrate 113 has formed thereon, a GaAs layer 115, above which is formed an A1GaAs layer 117. A plurality of gates 119, 121, 123, 125, 127, 129 and 131 are situated above the A1GaAs layer. These are shown in more detail in Figure 1A.

Electrons in a 2DEG induced in the active layer 115 are squeezed by a gate voltage (negative) applied to the gate electrodes. A negative voltage removes electrons and can electrostatically shape the electron gas to a desirable pattern, in the present case, that of a quantum dot.

3 Referring specifically to Figure IA of the accompanying drawings, the black squares and rectangles 119-131 are front gates which deplete the electron gas underneath, as mentioned above. Barriers are formed between all adjacent gates. The arrows show two possible conduction paths through the device. A quantum dot is formed at the centre of the gates.

An arrangement more common than the turnstile arrangement of Figure 1 A is shown in Figure 1B of the accompanying drawings. A metal gate 135 with a two dimensional array of holes/openings 137 etc. is lithographically formed in the gate to leave individual gate regions 139 etc. When a negative bias is applied to the gate, electrons are depleted under the gate metal but not in the regions under the holes 137 etc. For holes in some semiconductor systems other than GaAs/A1GaAs, the opposite bias may be required to confine carriers zero-dimensionally.

Another manner in which arrays of dots are frequently formed is by selectively etching away the 2D electron gas in certain areas of the heterostructure, leaving behind a 2D array of puddles containing the charge carriers. One way of implementing this scheme is dispense with the patterned gate in Fig 1B, but instead etch away the electrons in areas where the gate would have gone.

Single electron charging, commonly referred to as Coulomb Blockade, manifests itself in the current-voltage characteristics of quantum dots when the confinement length is sufficiently small (<-30Onin) and the number of electrons is sufficiently small (several hundred or less). The capacitance (C) of such a dot is Qe'-10-15F. The single electron charging energy, commonly referred to as the Coulomb energy, is the energy penalty incurred by having to add a whole electron to the dot when electrostatics require only a fractional amount of charge to produce neutrality between the dot and the surrounding reservoir of electrons. This charging energy is &/2C, where C is the capacitance of the dot, The performance of the detector according to the present 4 invention as defined hereinbelow is enhanced if arranged to operate in a Coulomb blockade regime. However, this is not a pre-requisite.

For quantum dots of sufficiently small diameter (i.e. <-70Onm for GaAs/AsGaAs systems) motion of electrons is described by the well-known 'particle in a box' model used in quantum physics. This model is depicted in Fig 2 of the accompanying drawings. For wider diameters, the particle in a box analogy still holds, but the electron (or hole) motion is not truly quantum because many zero-dimensional (OD) energy states in the box are occupied.

As shown in Figure 2, the infrared radiation excites an electron from the ground state or lowest occupied state in a quantum dot to the last state near the top of the barriers used to confine the dot, below a continuum of energy states just above the dot barriers. The confinement width of the dot may be varied by gate biases. Electrons in the continuum can be accelerated to a drain electrode under the influence of source-drain bias or electrons can tunnel out of the dot under the influence of sourcedrain bias to be accelerated to the drain.

For a wide variety of high electron mobility transistor (HEMT) structures and semiconductor systems, the energy separation between the OD states in such a dot will lie in the microwave or infrared region of the electromagnetic spectrum. If the polarisation state of incoming radiation has a non-zero component in the xy plane (parallel to the wafer) then the radiation may used to excite electrons (or holes) from the occupied to the unoccupied state. This is depicted a process 1. in Figure 2. The applicants have discovered that for an array of the GaAs/A1GaAs or other regrowth dots, excitation of the sort described above does take place. The existence of this excitation mode was detected by measuring the absorption spectrum of infrared radiation by the dot array. Figure 3 shows the result. The absorption peak at high frequency peak (peak P2) is due to the excitation mode.

1 Thus, the present inventid provides a detector for electromagnetic radiation, the detector comprising a plurality of elements each comprising an active region for induction of a two dimensional carrier gas therein and confinement means for confining a pool of carriers from the carrier gas within a confinement region of the active layer, the device further comprising a source region and a drain region, the plurality of elements being located in an electrical path between the source and drain regions.

Conventionally, arrays of quantum dots have either been fabricated by selectively etching away portions of a 2DEG ina HEMT or else fabricating a patterned front gate on a HEMT and applying a bias to the gate to deplete the electrons in the underlying 2DEG, in the manner described above with reference to Figures 1 and 1A. In the broadest definition, the present invention includes a detector in which the elements are formed by either of the aforementioned known means. However, it is particularly advantageous to form the elements by a regrowth technique whereby the confined pool of carriers of each element is located within a recessed region of the device.

Arrays fabricated via regrowth have the advantage that both side gates and backgates (as will be explained hereinbelow) can be easily fabricated. The sidegate can be used to adjust the confinement width and thereby tune the detector operating frequency over a wide range of frequencies (microwave to infrared). A reduction in width is affected by placing an increasingly negative bias on the side gates. Ibis also results in some reduction in electron density in the dot. Maintaining electron density is important because the detected signal will be proportional to the density of electrons in the dot. In regrowth structures, any change in density could be compensated for by applying a more positive backgate bias, which primarily raises the density without significantly affecting the confinement width. These biases are for GaAs/A1Ga.As electron systems. For hole gases in Ga.As/A1GaAs and for Si based dots, the polarity of the biases may be different but the principle is the same 6 Part of the versatility of using elements made by the regrowth technique is that it is frequency tuneable, so that radiation of a certain frequency may be detected and other frequencies rejected, and the detection frequency may be tuned. The detection frequency is given as the energy difference between the last occupied OD subband and the photoexcited level as shown in Figure 2. Thus by tuning the OD subband spacing, the frequency of the radiation which is detected may be easily altered. The OD subband spacing may be altered by changing the confinement width; smaller widths generally mean larger OD subband spacings and hence larger operating frequencies. For the Ga.As/A1Ga.As regrowth systems considered here, the confinement width of the dot may be tuned by changing the bias side gates and/or backgates.

The ability to tune the response of such detectors is particularly in the mid-infrared range. Such a device can in principle be tuned (via side gate bias) to operate in both of the commercially important 8pm-12pm and 5pm-6pm wavelength ranges. Moreover, by using the backgate voltage, it is possible to re-adjust the electron density and thereby optimise the noise equivalent power (NEP) of the detector. Optimising the NEP involves increasing the electron number so that the responsivity of the device rises, whilst the dark current (which flows when there is no irradiation) is maintained at an acceptable level (this current represents the background or noise)

On the other hand, the relative disadvantages of using the more conventional arrays fabricated as mentioned above as infrared detectors are that:- (a) It is much more difficult to fabricate them with side gates or backgates, and thus it is more difficult to optimise the electron carrier density for maximum photosignal. To change electron density, front gates are often used. This is a major disadvantage because all front gate material absorbs microwavelinfrared photons to a significant degree, thus reducing the sensitivity of the detector. Backgates are ideal for 7 use in these detectors because the incident radiation does not pass through the backgate. Thus a regrowth dot array with a backgate and no front gate might be expected to have significant sensitivity advantages over a conventional array with a frontgate only.

(b) In addition, using the patterned front gates does not result in as wide a change in confinement widths as are realisable using side gates. Thus the frequency of the device is less tuneable and the resultant detector less versatileluseful. With structures based only on etched dots, no tuneability at all is possible.

Devices according to the present invention having elements made by the regrowth technique also have distinct advantages over infrared detectors based on excitations of electrons from 2D subbands and out of quantum wells in that the latter cause electrons to be confined not to quantum dots but to two-dimensional (M) quantum wells. The use of regrowth dot arrays has several advantages over this system. First of all, a major disadvantage of such 21) detectors is that the polarisation vector of normally incident radiation is in the wrong direction to photoexcite electrons out of the 21) wells. The polarisation vector of normally incident radiation will be in the xy plane parallel to the plane of the wafer. For a 2D detector in the same orientation, the polarisation vector would need to be in the z direction orthogonal to photoexcited electrons. To account for this fact, 21) detectors often employ either surface gratings or bevelling of the edges to effectively rotate the polarisation vector. Such systems are very lossy, dramatically reducing the sensitivity of the detector. In addition, these coupling systems are more complicated and thus costly to implement. OD dot arrays are superior in this respect' the polarisation vector in x-y is already optimally oriented for photoexciting electrons.

Another advantage of OD arrays over 2D detectors is that the additional confinement in the remaining two dimensions theoretically yields larger oscillator strengths for 8 photoexcitations that in 2D. Better detector sensitivity may therefore be achieved for a given number of electrons in a unit area.

Also, most 2D detectors are based on quantum wells. The confinement widths of such wells cannot be easily or significantly varied, which means that such detectors are usually fabricated to work in one frequency range, and are not frequency tuneable. Regrowth dot detectors should, by contrast, be frequency tuneable over wide ranges in the microwavelinfrared region.

Lastly, the carrier density in most 2D detectors cannot be varied. Regrowth dot arrays by contrast allow the carrier density in the dots to be varied over a wide range. Precise control over the carrier density may be important if the responsivity of the detector is to be maximised whilst keeping the dark current in the detector at an acceptable level.

The preferred recessed kind of structure fabricated by regrowth techniques permits the confinement means of each element to be formed as a plurality of "side gates". However, it is also preferred for the confinement means to comprise at least one secondary gate for restricting the width of the carrier gas.

With the recessed form of construction, an antenna means can be provided to extend into the recessed region. However, antenna coupling can be provided for other kinds of quantum-dot inducing structure, such as those based on Figures 1 and 1 A. Antenna means for any of these structures may comprise a pair of substantially triangular members for impedance matching (in the manner of microwave microstrip). For recessed type structures, the apex of the triangular members can extend into the recessed region. These triangular features can be, e.g. in the form of equilateral or isosceles triangles. However, antennae are optional and may become of marginal or no benefit at higher frequencies.

9 Other forms of matched antenna are possible, for example a log-periodic structure (i.e. in the manner of a low-frequency Yagi antenna).

In the described preferred embodiments, the active region of each element is formed on or adjacent an angled facet of an etched layer structure, (the carrier gas is not necessarily always confined actually on the angeld facet). The confinement means of the elements comprises a front gate overlying the active layer and at least one back gate constituted by a layer or layers or the etched layer structure.

In any event, it is convenient to form the active region or layer as part of a high electron mobility transistor structure.

For quantum dots elements formed via regrowth in GaAS/A1GaAs, GaInAs/A1InAs or any other semiconductor systems such as silicon/silicon dioxide, then the barriers can be larger, and therefore radiation in the mid-infrared (MIR) or near-infrared (NIR) can be used. Other suitable heterostructures are CdTe, InSb, InAlA&W, SiGe. However, InGaAS-based systems will have low electron effective mass and therefore intrinsically large OD subband separations, making such arrays suitable for mid/nearinfrared detectors.

The detector of the present invention employs a plurality of the elements i.e. two or more, although a practical device might contain anything from 1,000 to 1,000,000 elements.

The present invention will now be explained in more detail by reference to the following description of a preferred embodiment and with reference to the accompanying drawings in which:

Figure 1 shows a cross section through a known device which is used to form a quantum dot or confine carriers in three dimensions; Figure 1A shows a plan view of the device of Figure 1, showing the turnstile arrangement of gate electrodes; Figure I B shows a plan view of an alternative gate arrangement for confining a quantum dot; Figure 2 shows an energy band diagram for explaining the operation of a quantum dot element in a device according to the present invention; Figure 3 shows the absorption spectrum for a quantum dot element made by the regrowth technique and according to the present invention; Figure 4 shows a radiation detector according to the present invention; Figure 5 shows a plan view of a first kind of quantum dot element for a radiation detector according to the present invention; Figure 6 shows a cross-section through the device shown in Figure 4; and Figure 7 shows a cross-section through a second kind of quantum dot element for a radiation detector according to the present invention.

Referring now to Figure 4, there is shown a radiation detector 201 according to the present invention. This is arranged to detect radiation from a source 203.

11 The detector 201 comprises a single wafer 205 on which are formed, a plurality of elements 207 etc. for confining a respective quantum dot in each. These are lithographically formed on the single wafer 205. On one side 209 of the wafer 205, is arranged a source region/contact 211. On the opposite side 213 of the wafer 205, is arranged a drain region/contact 215. The elements 207 are arranged in the wafer, between the respective sides on which the source and drain 211, 215 are respectively arranged. In use, a source-drain bias Vsd lies between source and drain. Each element has a gate arrangement 217 to be biased with a gate voltage Vgate. Details of different possible structures for the elements 207 are described in more details hereinbelow.

Between the radiation source 203 and the wafer 205, is arranged an optical system 219 for ensuring that the radiation is correctly transmitted to the elements 207 etc.

As mentioned above, in the broadest sense of the present invention, it is possible to fabricate such a device as shown in Figure 4 with the elements 207 etc. with the flat overlying gate turnstile arrangement of Figures 1 and IA. The grid-type arrangement of Figure IB on the aforementioned selective etching of the heterostructure may also be employed. However, it is especially preferred to use regrowth-type structures for these elements and examples of the latter wW now be described in detail- Referring now to Figures 5 and 6, there is shown a single quantum dot element 1 for use in the radiation detector shown in Figure 4 according to the present invention. As seen in plan view (Fig. 5), a first antenna member 3 and a second antenna member 5 each are in the shape of an equilateral triangle. One apex 7 of the first antenna member 3 and one apex 9 of the second antenna member 5 point towards one another either side of a horizontal axis of symmetry 10. Thus, the edge 11 of first antenna member 3, remote from its apex 7 and the edge 13 of the second antenna member 5, remote from its apex 9, are parallel to each other and each is disposed outwardly. The aforesaid apexes 7, 9 are separated by a gap 15.

12 First, second, third and fourth side gates 21, 23, 25, 27 are buried in the device structure and are each provided with a respective electrical connection 29, 31, 33, 35.

The first and second side gates 21, 23 are both disposed one side of the imaginary line joining the source and drain regions 17, 19 but respectively either side of the horizontal axis of symmetry 10. Similarly the third and fourth side gates 25, 27 are both disposed at the other side of the horizontal axis of symmetry, 10, opposite to the first and second side gates 21, 23. However, the third and fourth side gates 25, 27 are also disposed on opposite sides of the apexes 7, 9. Thus, the side gates 21, 23, 25, 27 are in positions (as seen in plan view) roughly corresponding to the corners of an imaginary square bounding the gap 15 between the apexes 7, 9.

Also, within the gap 15 and the aforementioned imaginary square but either side of the horizontal axis of symmetry, are first and second secondary gates 37, 39 or "plungers", also buried in the structure. Each secondary gate 37, 39 is provided with a respective electrical contact 41, 43.

The first and second antenna members 3, 5, are each provided with a respective electrical connection 45, 47, extending from the respective outwardly-facing sides 13 thereof. In use, these may be left open circuit or connected to ground. However, depending on the intended response frequency, the antenna members 3, 5 may be omitted.

11, Figure 6 shows how, along one axis (the source-drain axis), a recess 49 is etched in a layer structure 51 grown on a GaAs substrate 53. The antenna members 3, 5 are formed on the structure 51 and extend into the recess 49 so that their facing apexes 7, 9 are lower-most. Between the apexes is an uncovered region 55 at the bottom of the recess 49.

13 The layer 51 is comprised of the following layers formed in turn, upon the substrate 53, prior to selective etching of the recess 49, namely an n+ GaAs back gate layer 57 of 20Orim thickness, an AlAs/Ga.As superlattice 59, an undoped GaAs layer 60 of 100run thickness, a p GaAs side gate layer 61 of 10Onm thickness, and a conventional HEMT structure 63 (which of course includes an active layer). The triangular antenna members 3, 5 are deposited on top of the structure, after selective etching to form the recess 49 and formation of the HEMT structure 63 by regrowth.

The HEMT structure 63 extends into the side gate layer 61 to divide it into the first and second secondary side gates 37, 39. The primary side gates 21, 23, 25, 27 are not visible in this cross-section.

In use, a bias voltage between the source 17 and drain 19 includes a twodimensional electron gas in the active layer of the HEMT 63. The back gate and secondary gates 37, 39 are used to deplete-out all except a narrow conduction region in the centre region 55. Then, bias voltages applied to the primary side gates 21, 23, 25, 27 confine carriers in a quantum dot. Current flowing through the quantum dot is interrupted until radiation is received via the antenna members 3, 5 to provide the additional energy required to allow conduction across or above the potential barriers which confine the quantum dot.

Using dot elements of this device, the absorption spectrum of Figure 3 demonstrated a OD subband separation of 1.5meV, with a dot width of 1.8gm. A 2D array of 160,000 of the dots depicted in Figure 5 was used to boost the strength of the absorption signal. The dots were spaced 2.5gm apart, and had a lithographic width of 1.8gm.

If a source drain bias Vsd of sufficient magnitude is applied across the dot array (Figure 4) then an electron excited by infrared or microwave radiation to a OD state 14 located near the top of the dot (i.e. a state with energy just below that of the confining barriers which define the dot) can tunnel out of the dot and be swept away to the drain, where it is detected as a change in current through the device. This is illustrated as Processes 1. and 2a. in Figure 2. Alternatively, the infrared or microwave radiation may excite the electrons completely out of the dot and into the continuum of states located just above the electrostatic barriers defining the dot. In this case, the electron is swept away to the drain under the influence Of VA, and no tunnelling is required for a photosignal to be detected. This is illustrated as Processes 1. and 2b. in Figure 2.

As shown in Figure 7, a second kind of quantum dot element 151 for use in the device of Figure 4 in accordance with the present invention is somewhat analogous to that of the first kind.

A recess 153 is etched in a layer structure 155 consisting of a 2000A (100) n+Ga.As back gate layer 157 formed on a 500 gm SI substrate 159. The recess divides the n+GaAs layer into respective side gates 161, 163 and to reveal a surface 165 which acts as a regrowth interface.

On the regrowth interface surface 165 are grown as a regrowth structure 167, in order, a 300A GaAs layer 169, a 1500A AlAs/GaAs superlattice 17 1, a 300A active GaAs layer 173, and a 650A top structure 175, consisting of, in order, and A1GaAs spacer layer 177, an n+AlGaAs layer 179 and a GaAs cap layer 18 1.

Optionally, a double -triangular antenna could be formed over the cap layer 18 1, as in the first kind of quantum dot element, extending apexwise into the recess 153. In use, a 2DEG 183 within the active GaAs layer 173 adjacent the A1GaAs spacer layer is restricted by the applied potentials so that a quantum dot 185 is confined in the 2DEG plane, of the bottom of the recess 153.

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

16

Claims (1)

1. A detector for electromagnetic radiation, the detector comprising a plurality of elements each comprising an active region for induction of a two dimensional carrier gas therein and confinement means for confining a pool of carriers from the carrier gas within a confinement region of the active layer, the device further comprising a source region and a drain region, the plurality of elements being located in an electrical path between the source and drain regions.

   2. A detector according to claim 1, wherein the elements are formed on a single wafer so that the active region of each element is a discrete area of an active layer of the device.

   3. A detector according to claim 1 or claim 2, further comprising optical means for coupling radiation to the elements.

   4. A detector according to any preceding claim, wherein the confinement means of each element comprises a plurality of gates above the active region.

   5. A detector according to claim 4, wherein each plurality of gates is in a turnstile configuration.

   6. A detector according to any of claims 1 to 3, wherein the confinement means comprises a metal layer having an array of openings formed thereon.

   7. A detector according to any of claims 1 to 3, wherein the confinement means comprises etched away portions of a heterostructure of the device.

   17 8. A detector according to any of claims 1 to 3, wherein the confinement region of the active region of each element is located within a recessed region of the device.

   9. A detector according to claim 8, wherein said confinement means of each element comprises a plurality of side gates.

   10. A detector according to claim 9, wherein the confinement means of each element further comprises at least one secondary gate for restricting the width of the carrier gas.

   A detector according to any preceding claim, wherein each element is provided with antenna means for coupling electromagnetic radiation to the confinement region of that element.

   12. A detector according to claim 11, wherein said antenna means of each element extends into the recessed region of that element.

   13. A detector according to claim 11 or claim 12, wherein each said antenna means comprises a pair of substantially triangular members.

   14. A detector according to claim 13, wherein an apex of each substantially triangular member extends into the recessed region.

   15. A detector according to claim 13 or claim 14, wherein said substantially triangular members are substantially equilateral.

   16. A detector according to any of claims 11 to 15, wherein each said antenna means is configured as a log-periodic antenna.

   18 17. A detector according to claim 8, wherein the active region of each element is formed on an angled facet of an etched layer structure.

   18. A detector according to claim 17, wherein each confinement means comprises a front gate overlying the active region and at least one back gate constituted by a layer or layers of said etched layer structure.

   19. A detector according to any preceding claim, wherein each active region is part of a high electron mobility transistor structure.

   20. A detector for electromagnetic radiation, the detector being substantially as hereinbefore described with reference to any of Figures 4-6 of the accompanying drawings.