GB2469961A

GB2469961A - A Photon Detector

Info

Publication number: GB2469961A
Application number: GB1013236A
Authority: GB
Inventors: Beata Ewa Kardynal; Andrew James Shields
Original assignee: Toshiba Research Europe Ltd
Current assignee: Toshiba Europe Ltd
Priority date: 2007-02-27
Filing date: 2007-02-27
Publication date: 2010-11-03
Anticipated expiration: 2027-02-27
Also published as: GB201013236D0; GB2447054B; WO2008104761A3; GB2447054A; GB0703848D0; WO2008104761A2; GB2469961B

Abstract

A photon detector comprisies a plurality of quantum dots 207, a sensing region and measurement means for measuring an electrical characteristic of the sensing region. The sensing region is located such that a change in the charged state of said quantum dots 207 causes a change in the electrical characteristics of said sensing region. The detector for example a resonant tunnel structure, further comprises photon counting means for determining the number of photons detected in a pulse of radiation from changes in the electrical characteristics of the sensing region. The counting means is configured to determine the number of photons from changes in the electrical characteristics of the sensing region as the detector returns to its non-illuminated state. The device may comprise a High mobility Electron Transistor (HEMT) formed on a GaAs substrate and include a graded absorber layer of AlGaAs.

Description

A Photon Detector The present invention relates to the field of photon detectors and methods for detecting photons which are configured to measure a single photon, more specifically, the present invention relates to the field of photon detectors and methods for detecting photons which can determine the exact number of photons received at a photon detector.

Detectors which are capable of detecting a single photon so-called single photon detectors are an important component of any system which operates using the principles of quantum cryptography. Such systems rely upon the transmission of data bits as single particles, in this case, photons, which are indivisible. The data may be encoded using polarisation of the electric field vector of the photons, the phase of the photons etc. However, as well as detecting a single photon, there is also a need to produce detectors which are capable of resolving the number of photons in a pulse of radiation. Such detectors are useful for characterising non-classical light sources such as single photon generators to determine if they are genuine and reliable single photon sources. Another potential application is for determining if one or more photons are received per pulse in a quantum communication system. If two or more photons are present in a pulse then the pulse may be subject to a pulse splitting attack where just one photon from the pulse is read by a eavesdropper. This seriously degrades the security of the system. Thus, it is desirable to develop a detector which allows the number of photons in a pulse to be accurately determined.

Single photon detection is also useful as a low level light detection means for spectroscopy, medical imaging or astronomy. Both in medical and astronomical applications the high energy photons (X-ray etc) or high energy particles are converted in scintillators into many (10-100) low energy photons. These low energy photons are then detected by avalanche photodiodes or photomultiplier tubes. As the low energy photons that are produced are scattered in space there is a need for large area detectors which are very sensitive. Also arrays of such detectors allow the spatial distribution of low energy photons to be obtained in order to gain information about the original photon. These applications would also benefit from the ability to count the number of photons incident on a single photon detector.

Previous attempts to develop a single photons detector which is capable of determining the number of photons detected from a pulse include single photon avalanche photodiodes (APDs). These detectors are binary ("click counting") detectors, several schemes have been proposed involving either multiple devices or time multiplexing to allow photon number counting. However such schemes cannot resolve two photons incident on the same detector at the same time. Furthermore some a priori information about the light source is required to reconstruct the light distribution from a non-classical light source. A variant of the APD, the visible light photon counter allows detection of the multiple avalanches within a single detector with an efficiency> 90% at 532 nm wavelength at temperatures around 6.5 K. A transition edge sensor (TES) combines high detection efficiency (up to 89% at lS5Onm wavelength) with photon number resolving capabilities. However it operates at 1 OOmK making TES unsuitable for many applications. Commercial PIN diodes cooled down to 4.2 K were also shown to be photon number resolving although with limited time resolution (1Hz bandwidth).

The present invention has been developed to address at least some of the above problems and in a first aspect provides a photon detector comprising a plurality of quantum dots, a sensing region and measurement means for measuring an electrical characteristic of the sensing region, said sensing region being located such that a change in the charged state of said quantum dots causes a change in the electrical characteristics of said sensing region, the detector further comprising photon counting means for determining the number of photons detected in a pulse of radiation from changes in the electrical characterisitcs of the sensing region.

Single photon detectors have been proposed which operate using quantum dots to capture photo-carriers. Photo-carriers trapped within the quantum dots can affect the electric current in a transport region of the device which allows the presence of photons to be detected.

In one such detector, the sensing region comprises an active region configured to support a narrow channel and said measurement means comprises source and drain contacts to said narrow channel. The narrow channel has a width, perpendicular to the direction of carrier transport through the channel which is less than 1 OOnm. More preferably, the narrow channel is a two dimensional carrier gas. The device may be configured such that the two dimensional carrier gas is a two dimensional electron gas or a two dimensional hole gas.

Another such detector is configured as a resonant tunnel structure wherein absorption of a pulse of radiation affects the tunnelling conditions. The detector may be configured such that an incident radiation pulse increases the flow of a tunnelling current or reduces the flow of a tunnelling current.

On both types of structure, the sensing region may be monitored by measuring the current, voltage, resistance or any quantity related thereto.

Structures of this type are described in GB 2 365 210 which describes a photon detector based on a resonant tunnelling diode structure and GB 2 341 722 which describes a transistor structure.

Other types of detectors based on quantum dots may be used where a signal due to detection of one photon: is finite and/or which can discriminate number of photons in an illumination pulse. In such a case the amplitude of the signal on the output of the detector scales with the number of photons.

The counting means may be configured to count the number of sharp changes measured by the measurement means due to the quantum dots returning to their non-illuminated state after excitation with a pulse of radiation.

Alternatively or additionally, the counting means are configured to measure the magnitude of a sharp change in the electrical characteristics due to absorption of a pulse of radiation by the quantum dots.

The counting means may also be configured to measure the magnitude of the sharp changes in the electrical characteristics due to the quantum dots returning to their non-illuminated state.

The sharp change in the electrical signal which is measured may be a peak, trough, change in gradient, step etc. Preferably, the sharp changes in the signal (if not originally manifested as peaks) will be converted into peaks by amplifiers or other signal conditioning components. The magnitude of the sharp change in the signal may then be determined by measuring the amplitude of the peak.

When counting changes in the signal due to the device returning to its state prior to illumination or "resetting", it is desirable to allow enough time between incident pulses in order to allow a high probability that all reset events will be performed before the next radiation pulse is received. The bit rate of a system will be reduced by increasing the spacing between incident pulses. However, it is also possible to reduce the time over which the device is expected to reset itself by increasing the sensing current density, which resets the dots, or increasing the probability of the device resetting by decreasing the distance between the quantum dots and the narrow channel in the embodiment with the narrow channel and increasing scattering of electrons in the RTD embodiment of the device.

Also, it is desirable to fabricate a detector with a suitable time resolution so that the probability of two reset events not being resolved is small. The time resolution is a property of the detector and the electronic circuit that is used to "condition" the signal to be compatible with the counting electronics. In fact the time resolution is inversely proportional to the bandwidth of the measurement.

To accommodate the above two conflicting requirements, a reset time which is preferably iOns or more and! or lps or less is used.

It is possible to enhance the time resolution of the system by reducing the capacitance of the detector, and increasing the bandwidth of the amplifier used with the detector.

Preferably, the conditioning circuit has a bandwidth of 2MHz or less and/or 200MHz or more.

In addition or as an alternative to counting reset signals, it is possible to perform a count of the number of photons received by the detector by measuring the amplitude of a sharp change in the electrical characteristics due to absorption of a pulse of radiation by the quantum dots.

The amplitude may be compared with one or more predetermined levels which correspond to different numbers of photons being received by the detector.

The amplitude of the reset signals may also be measured to determine the number of photons. If the time resolution of the device results in some reset events contributing to the same pulse, then these pulses will have greater amplitude than those reset pulses arising from single reset events. Therefore, preferably, the counting means is configured to measure the magnitude of a reset signal to produce a weighting for the signal and count the number of reset signals. For example, if two reset peaks are observed and one peak has an amplitude which indicates the presence of one photon and the other peak has an amplitude which indicates the presence of two photons then the counting means will determine that three photons were absorbed by the device even though only two reset peaks were observed.

The magnitude of the change in the signal due to the absorption of the pulse of radiation may also be used as this will be proportional to the number of photons absorbed by the detector. This can be used in combination with the reset signals as a method of providing a confidence measure of the accuracy of the result. If the results from the reset signal and the initial absorption signal agree then the used of the detector can be confident that the measured number of photons is accurate. This method would reduce the effect of dark counts on the measurement. For example, if the dark count error probability in time t is x, then by accepting only events when the amplitude of the main peak equals number of reset pulses, the dark count errorprobability is reduced to xt1xt2, where t1 is the time of measurement of absorption peak, t2 is the total time of measuring the reset pulses In a preferred embodiment, the detector further comprises an absorption region. The cross section of said sensing region, which is preferably substantially perpendicular to the incident direction of photons, is preferably smaller than the cross sectional area of the absorption region in this plane. The absorption region may be the same as the sensing region.

In a second aspect, the present invention provides a method of determining the number of photons received by a photon detector, comprising: providing a plurality of quantum dots, providing a sensing region, said sensing region being located such that a change in the charged state of said quantum dots causes a change in the electrical characteristics of said sensing region; measuring electrical characteristics of said sensing region; and determining the number of photons from the changes in the electrical characteristics of said sensing region.

The present invention will now be described with reference to the following non-limiting embodiment in which: Figure la is a schematic of a device in accordance with an embodiment of the present invention and Figure lb is a schematic band diagram for the device of figure 1 a; Figure 2a is a plot of the voltage on the output of the amplifier used to convert * detector's current steps due to photon detection into voltage peaks against time for the detector of figure 1 a when the detector absorbs a single photon and figure 2b is a plot of the detector voltage against time for the detector of figure 1 a when the detector absorbs two photons; Figure 3 is a plot of probability of detecting N reset events of the detector of figure 1 a against the pulse separation for N 1, 2, 3 and 6, the probability being expressed as a percentage and the pulse separation time being expressed in units of the theoretical reset time; Figure 4 is a plot of the probability of two reset pulses not being resolved against the time resolution capabilities of the detector of figure Ia with its measurement circuit, time resolution is expressed in terms of the theoretical reset time, the data is shown for N expected reset events where N = 2, 3 and 6; Figure 5 is a plot of the average detected number of photons for a pulse of radiation using the detector of figure 1 a against the amplitude of the detection signal; Figure 6 is a plot of the number of detected photons against incident flux; Figure 7a is a schematic band diagram of a device in accordance with a further embodiment of the invention prior to illumination and figure 7b is a schematic band diagram of the device of figure 7a after illumination; and Figure 8 is a schematic of a device in accordance with a further embodiment of the present invention.

Figure 1 schematically illustrates a device in accordance with an embodiment of the present invention. Figure la shows the actual layer structure of the device and figure lB shows the corresponding band diagram. The device is a GaAs/AlGaAs High Electron Mobility Transistor (HEMT).

The structure is formed on GaAs substrate (not shown). First, an nGaAs layer 201 is formed overlying and in contact with the substrate. The nGaAs doped layer 201 will form a back gate 201 for the device. Overlying and in contact with the back gate layer 201 is a thick Al0 33Gao 67As back-gate barrier layer 203. This back gate barrier layer 203 serves to separate the back gate layer 201 from the remainder of the device.

Overlying and in contact with the back gate barrier 203 is absorber layer 205. Absorber layer 205 comprises 200nm of graded AlGa1..As where the composition of x is 0.25 at the lower interface with the back gate barrier layer 203 and where x0 at the top of the layer 205. This graded absorber layer composition creates an internal electric field which directs the holes towards the top of absorber layer 205 and improves the detection capability of the detector. However GaAs may also be used for an absorber.

Dot layer 207 is then formed overlying and in contact with the absorber layer 205. The quantum dot layer 207 is formed by growing a thin layer of InAs (2-3 monolayers) which is then capped with 1 Onm GaAs 209. Overlying GaAs layer 209 is an undoped barrier layer 211 of lOnm ofAl033Ga067As.

Quantum well layer 213 is then formed overlying in contact with barrier layer 211.

Quantum well layer 213 comprises 2Onm GaAs. Overlying and in contact with said quantum well layer 213 is an undoped barrier layer 215 with remote n-type doping barrier layer 217 formed overlying and in contact with said barrier layer 215. Doped barrier layer 217 comprises Al0 33Ga067As. A two dimensional electron gas (2DEG) forms in the quantum well layer 213 at the heteroj unction with the undoped barrier layer 215.

The structure is then finished with GaAs cap layer 219.

Source and drain contacts 223 and 225 which are ohmic contacts are then formed to the quantum well layer 213. A front gate which comprises 8nm of semi-transparent NiCr 221 is also provided overlying the cap layer 219.

The above structure was shown to provide quantum dots in the quantum dot layer 207 which are relatively large and which form at least three excited states. This was determined using PL characterization.

The doped barrier layer 217 provided carriers to the 2DEG such that the 2DEG has a density of 2x10'5m2 in an ungated region of the device. The electrically active area of the device was defmed by the overlap between a 1 tm wide mesa between the source and drain context and a 2p.m wide front gate. The additional back gate 201 provides means of tuning the electric field with an absorber layer 205.

Figure lb shows the corresponding band diagram for the device of figure 1 a.

The above detector operators by the trapping of photo excited carriers within the quantum dots of quantum dot layer 207. The presence or absence of photo excited carriers within the quantum dots of the quantum dot layer 207 results in different transport characteristics of the two dimensional electron gas formed in quantum well layer 213 measured via source drain contracts 223 and 225.

When a photo carrier generated in the device is trapped in a quantum dot, this changes the current flowing in the quantum well which is measured by source drain contacts 223 and 225. After a certain amount of time (reset time), the action of the photo carrier is "neutralized" by capturing a carrier of the opposite charge and current of the device returns to original value.

The reset time can be expressed statistically through a reset time constant t. The reset time constant depends on the exact configuration of the device and can be varied by.

altering, for example, the current through the device, the coupling between quantum dots in the dot layer and the charge reservoirs. In this type of device, the reset time may change from several tens of nanoseconds to up to ims or more. It is also possible to achieve shorter and longer reset times.

If N photo carriers from a light pulse were trapped on the dots, there will be N reset events following this event. Thus, it is possible to determine how many photons have been captured by the quantum dots in a single pulse of radiation by counting the number of reset pulses. This is shown in figures 2a and b.

Figures 2a and 2b show the results of measurements performed by cooling the detector down to 4.2 K and applying a constant source-drain voltage of VSD 30 mV. The front (VFG) and back gate (V50) voltages are fixed at -60 mV and 150 mV, respectively. The detector was illuminated via an optical fibre from a pulsed laser diode; with a wavelength of 684 nm and a repetition rate of 200 kHz. Laser light is focused on the active area of the detector using an objective lens with the spot size of 835 nm FWHM.

The wavelength of the laser was chosen so that the absorption of light in Al033Ga067As layers of the detector is negligible and photo-carriers are generated only in a channel and absorber of the detector. It is estimated that 25% of photons are transmitted through the gate of the device (and 32% of those photons are absorbed in the active area of the device). The built-in electric field in the structure and negative charge stored on the dots sweeps holes towards the dots. Capturing of the hole on the dot results in an increase in the source-drain current which can be as large as 5 nA.

Figure 2a shows a measurement of the voltage from the output of the amplifier, which is proportional to the changes in current between the sOurce and drain contacts of the device of figure la over time. A first peak 301 is seen, peak 301 is positive in voltage and corresponds to a change in current due to a dot capturing a carrier during illumination. After a time of around 2ms, a reset peak 303 can be seen due to the dot which has trapped the hole being reset to its pre-illumination state.

Looking at figure 2b, it can be seen that absorption positive peak 305 is larger than that of 301 of figure 2a. This time, there are two reset peaks 307, 309 which suggest that two photons were absorbed by the device and hence there are two reset events. The device may detect two photons by a single dot having two photo excited carriers or two dots may each have a single photo excited carrier. In either event, two reset peaks will be observed.

In order to count all N reset peaks, the separation between light pulses of the irradiating light should be considered.

Figure 3 shows a plot of the probability of measuring N reset events against pulse separation time T of the incident light pulses. The pulse separation time T is expressed in terms of the reset time t. The data is shown for N 1, 2, 3 or 6. The lines indicate the probability of detecting all N reset events for N = 1, 2, 3 or 6.

As expected, as the time between pulses increases, the probability of the detector detecting all N reset events increases. For example, if the reset time is 1 ms and three reset pulses are to be measured with a 99.999% probability, the light pulses should be separated by 12.6ms. If the probability of 99.9% is sufficient, then the separation time of 8.Oms is sufficient.

It is not only beneficial to consider the time between pulses when considering the sensitivity of the detector. In order to distinguish between reset peaks, the time resolution of measurements needs to be considered.

Figure 4 shows a plot of the probability of coincidence reset pulses as a percentage (i.e. the chances of two or more reset pulses coinciding) as a function of the time resolution of the system dt. The time resolution is expressed in terms of the reset time r. The figure shows the probability of not being able to resolve two or more reset peaks for N=2,3and6.

As expected, as N increases, the time resolution must increase (i.e. dt must decrease) in order to avoid coincident reset peaks. For example, if the reset time is lps and N=3, a time resolution of 5.8ns is required to guarantee that the reset peaks are to be resolved with a 99.99% degree of accuracy. If six reset pulses are required, the time resolution must be increased in order to detect all six measurements with the same probability.

In addition, to measuring the reset peaks, the number of photons detected by a detector may be determined more accurately if the number of reset peaks is used in combination with measuring the height of the reset peaks. This is because the height of the reset peaks will indicate the number of reset events which have contributed to the reset peak.

In other words, if two reset events contribute to the same peak, that peak will be larger than a peak which results from just one reset event.

If the relative variation between signals from each reset event is small (y%) then it is possible to have (100% / y%) reset pulses overlapping in time because the amplitude of the resulting peak will uniquely inidcate the number of reset events which contributed to the peak. It is also possible to measure up to (100% / y%) photons directly from the photon induced peak. Thus the number of reset peaks may be counted and weighted by determining the number of reset events each peak contains.

Figures 2a and 2b showed a peak in voltage due to the initial absorption of the pulse, 301, 305. Peak 305 of figure 2b is twice the size of that of peak 301 of figure 2a indicating that more than one photon was absorbed during the event measured in figure 2b.

Figure 5 shows a plot of the measured count rate per pulse against the discriminator level measured for the absorption pulse. If a peak has amplitude higher than the discriminator level, it will result in a count, if it is smaller it will not. The data is shown for pulses of incident radiation which have average photon concentrations (.t) of 0.34 and 1.9 photons per pulse. Also, the total dark count and the contribution from the amplifier are shown as open circles and black lines respectively on the figure.

In figure 5, each point is an average of 2x106 measurements. At the lowest voltage levels, the signal is dominated by the noise of the amplifier which has a standard deviation of 6.7mv. As the voltage increases, a clear signal due to photons being detected by the detector can be seen. At a peak height of about 2OmV, almost all the counts originate from at least one dot losing at least one electron. When the devices are illuminated with laser pulses of =0.34, there is a clear shoulder extending up to 6OmV, followed by another shoulder with an order of magnitude lower number of counts and extending to 9OmV. Similar data can be observed for the situation with l.9 although the drop in adjacent shoulders is smaller in this case.

The features of figure 5 can be explained by detection of different photon numbers in the incident laser pulses. At amplitude levels of below 6OmV, but above the amplifier noise, the voltage peaks result from the detection of a single photon. At peak heights between 7OmV up to 1 OOmV, it is believed that the peaks are due to detecting 2 photons, and for peaks heights greater than 11 OmV, it is believed that three or more photons are being detected. Since the horizontal axis of figure 5 is a discriminator level, it represents all the peaks with amplitudes larger than the discriminator level and for discriminator levels between the amplifier noise and 6OmV we detect at least one photon, between 7OmV up to 1 OOmV we detect at least 2 photons, above 11 OmV at least 3 photons.

From the results above, it is possible to determine that a peak corresponds to noise, detection of one photon, detection of 2 photons etc by comparing the amplitude of the peak with a voltage level which is termed a "discriminator level". For this device, the discriminator levels are 2OmV (noise), 7OmV (below this level indicates just 1 photon, above this level indicates 2 or more photons) and 110 mV (below this level indicates less than 3 photons, above this level indicates more than 3 photons). The discriminator levels for N (no of photons) 1, 2 and 3 are indicated on figure 5.

The probability of detecting at least N photons is a single laser pulse can be expressed 1->exp(-iip) . (iiii)' Ii!, where i is detection efficiency. Such probability, as a function of p is shown in Figure 6 for N= 1, 2 and 3 (lines) together with measured number of counts at three discriminator levels 30, 70 and 110 mV.

The experimental data in the model agree very well with only one fitting parameter common for all three sets, the detection efficiency. The best agreement between experimental data and the model is obtained with an efficiency of 1.3%. At low fluxes, the probability of detecting at least N photons is proportional to ()N as shown with the dotted lines in figure.

The low value of the efficiency is partly a result of small probability of generating a photo-hole in the well and the absorber, which is 8% at 684 nm at 4.2 K. This figure may be further enhanced by tuning of the dot size and their interaction with the channel.

A useful photon number resolving detector may then be realised by enhancing absorption in the well, reducing a probability of electron-hole recombination (due to compositional grading in the absorber) and reducing reflections from the top surface.

The efficiency of this device can be improved by replacing the semi-transparent metal gate with transparent (transmitting >98% photons ITO gate), replacing graded A1GaAs absorber with GaAs absorber and increasing GaAs absorber thickness.

Figure 7a and Figure 7b are schematic band structures which are used to illustrate a mode of operation for a photon detector in accordance with a further embodiment of the present invention. This detector operates using resonant tunnelling.

A conduction band I and a valence band 3 are shown. Figure 7a shows the detector prior to illumination. Figure 7b shows the detector after illumination.

An emitter 5 and a collector 7 are provided at either end of the detector. An emitter-collector bias Yce is applied across the detector such that the potential of the collector 7 is more positive than that of the emitter 5, thus inducing the flow of electrons from the emitter to the collector. In this example, the carriers will be electrons. However, it will be appreciated by those skilled in the art that detector could be configured with holes as the majority carriers.

The detector comprises a first low dimensional system 9 which is located between the emitter 5 and the first barrier layer 11. Electrons in the low dimensional layer 9 have energy of the first energy level 13 (this level can be seen more clearly on Figure 7b).

Adjacent the barrier layer 11 and on the opposing side to the first low dimensional system 9 is a second low dimensional system 15. The second low dimensional system is capable of confining electrons with a second energy level 17. In the detector shown in Figure 7a (before illumination), the first energy level 13 and the second energy level 17 align.

Adjacent the second low dimensional system 15 is a second barrier layer 19. The second barrier layer 19 is thin enough so that when the first 13 and the second 17 energy levels align, resonant tunnelling takes place through the first barrier layer 11 and the second barrier layer 19. An absorption layer 23 is then provided between the second barrier layer 19 and the collector 7. A quantum dot layer 21 is then provided in said absorption layer 23, on the opposing side of the second barrier layer to that of the second low dimensional system 15.

Due to the alignment between the first and second energy levels 13, 17, charge flows freely from the emitter 5 to the collector 7 when a bias V is applied. The alignment of the first and second energy levels will be dependent on the magnitude of the applied bias. The magnitude Vce is chosen such that the energy levels 13 and 17 align.

Figure 7b shows the same device as that of Figure 7a. However, the device has been illuminated. To avoid unnecessary repetition, like numerals have been used to denote like features. On absorption of a photon, an electron-hole pair is excited, here shown as electron 25 in the conduction band and hole 27 in the valence band. The bias V causes the electron 25 to be swept towards the collector. However, the hole 27 is swept in the opposite direction and is swept into dot 21 where it is trapped. The change in the charging state of dot 21 causes a change in the alignment of the first energy level 13 and the second energy level 17 near the dot. As these two levels do not now align, the detector is brought "off-resonance" and tunnelling through the first barrier layer is suppressed locally near the dot. Therefore, the charge cannot flow freely from the emitter 5 to the collector 7. This change in current can easily be detected and signifies the absorption of a single photon.

In the above device, the photon detector is switched on by applying collector/emitter bias V across the detector. When a single photon is absorbed, the emitter/collector current is reduced.

In figure 7, the device is configured so that resonant tunnelling occurs unless a photon has been absorbed. However, the device may also be configured in the reverse manner where current flow from the emitter to the collector is blocked unless a photon is absorbed. In Other words, a photon brings the device onto resonance.

Further variations on a resonant tunnelling diode single photon detector are described in GB 2365 210.

Figure 8 shows a schematic cross-section of a device in accordance with an embodiment of the present invention.

In figure 8, an emitter contact layer 108 is formed overlying and in contact with a substrate 114. The emitter contact layer 108 may comprise n-doped GaAs if the system is built around the GaAs/Al(Ga)As material system or n-doped InGaAs if the device is fabricated around the InGaAs/InAIAs system.

Overlying and in contact with said emitter contact layer 108 is the emitter layer 107.

Emitter layer 107 comprises n-type GaAs or graded AlGaAs in a GaAs/A1(Ga)As system or n-doped InGaAs in an InGaAs/InAlAs system. The emitter layer 107 is formed as a narrow pillar. Typically, the pillar will have a cross-sectional diameter measured in the plane of the substrate of approximately I tm or less. How this pillar is defined will be described later. The emitter pillar 107 is then surrounded by an insulator 109 which provides insulation and mechanical support. A suitable material for insulating layer 109 is polyimide.

Overlying the emitter pillar 107 and the polyimide surrounding layer 109 is first tunnel barrier 106. First tunnel barrier 106 comprises AlAs or AlGaAs in GaAs/Al(Ga)As systems or AlAs or InAlAs in InGaAs/InAlAs systems.

Quantum well layer 105 is then provided overlying and in contact with said first tunnel barrier 106. Said quantum well layer comprises GaAs in GaAs/Al(Ga)As systems or InGaAs in InGaAsTlnAlAs systems.

Second tunnel barrier 104 (equivalent to second tunnel barrier 19 in figures 7a and 7b) is then formed overlying and in contact with said quantum well layer 105. Said second tunnel barrier layer 104 comprises the same material as the first tunnel barrier layer 106.

Next, photon absorption region 102 is formed overlying and in contact with said second tunnel barrier layer 104.

The photon absorption region 102 comprises either insulating GaAs or insulating InGaAs dependent on whether the material system of the photon detector is GaAs/Al(Ga)As based or InGaAs/InAlAs based respectively. The photon absorption region acts to collect photons and for capture by quantum dots.

A layer of quantum dots for example InAs quantum dots 103 is provided in the absorption region close to the second tunnel barrier layer 104. Quantum dot layer 103 may be formed by a variety of techniques. For example, it may be formed using strained layer growth where 1 to 10 monolayers of a material having a different lattice constant of that of a photon absorption region 102. InAs quantum dots in GaAs or InGaAs will provide a strained growth system. Due to the strained layer growth, the InAs layer will form as quantum dots. Alternatively, the quantum dots may be formed by interface fluctuations where a few monolayers of a material having a narrow band gap are embedded in-between layers of semiconductors with a wide band-gap. This forms quantum well, fluctuations in the thickness of the well act as quantum dots.

Layer 103 forms quantum dot 21 of figures 7a and 7b.

Collector contact layer 101 is then formed overlying and in contact with said photon absorption region 102. Collector contact layer comprises n-doped GaAs in a GaAs/Al(Ga)As system or n-doped InGaAs in InGaAs/InAlAs system.

In this particular embodiment, the layers from the first barrier layer 106 to the collector contact layer 101 are patterned together to form a structure having a circular cross section with an approximate cross-sectional diameter of between 10jm and lOOj.tm. It should be noted that this diameter is larger than the diameter of the emitter pillar 107.

The device of figure 8 is intended to be illuminated from the top so that top of device provides the surface of the device which is exposed to incident radiation.

An ohmic contact 110 is made to the emitter contact layer 108 and ohmic contact 111 is made to the collector contact layer 101. In order to allow illumination of the top of the device 115, collector ohmic contact 111 is provided close to the edge of collector contact layer 101. It is difficult to electrically bond to contact 111. Therefore, a further insulating layer 116 is provided overlying the side of the device. A contact layer 112 is then provided overlying insulating layer 116. Contact layer 112 extends to bonding pad 113 which allows electrical bonding away from the absorption region of the device. By electrically bonding to area 113, an electrical contact can be made to ohmic collector contact 117.

The device of figure 8 may be fabricated in two main ways. In both ways, it is important that there is a separate patterning step for the photon absorber layer 102 and for the emitter 107.

In a first fabrication method, the emitter contact layer 108 and the emitter layer 107 are grown using a epitaxial growth technique, for example, molecular beam epitaxy (MBE).

The device is then removed from the growth chamber and emitter pillar 107 is defined using a standard technique such as an electron beam lithography coupled with either wet etching or a dry etching technique e.g. reactive ion etching.

An insulator is then spun over the layers and etched back to expose the top of emitter pillar 107 such that there is a flat growth surface formed by the top of emitter pillar 107 and insulating layer 109. The device is then placed back into the growth chamber for growth of the layers from the first tunnel barrier layer 106 to the collector layer 101.

The layers grown on an insulator are not going to be epitaxial but rather amorphous.

This may work but there may also be many traps for the photo-carriers. The layers 101 to 106 are then patterned using standard photolithography to define the cross section of the photon absorption region.

In an alternative fabrication technique, the layers from the emitter contact layer 108 to the collector contact layer 101 are formed one after each other in a growth chamber.

The device is then patterned to form the patterned photon absorption layer 102 and layers 101 to 106.

Emitter pillar 107 is then patterned during a second selective etching step where the etch is used to undercut layer 106. Polyimide or another insulator 109 is then spun onto the structure in order to give the overhanging layers 106 to 101 some structural support.

Ohmic emitter contact 110 and ohmic collector contact 111 are then patterned and defined in the standard manner. Next, bonding pad 113 is formed to the substrate 114 well away from the part of the device which is to be illuminated. A further insulator 116 is then spun over the device. The insulator is patterned and etched so that it provides an insulator 116 for contact metal 112. Contact metal 112 is then provided between the bonding pad 113 and the ohmic contact 111.

In the previous structure, the emitter 107 is patterned on its own. The first tunnel barrier layer 106, the quantum well layer 105, the second tunnel barrier layer 104 are patterned together with a photon absorber layer 102. However, it is possible to pattern the emitter 107 together with the first tunnel barrier layer 106, the quantum well layer and the second tunnel barrier layer 104 so that layers 107, 106, 105 and 104 together form a narrow pillar with a cross section of lp.m or less. Patterning of the emitter together with both barriers and the well provides better defmition of the tunnelling area. It may also result in electrically less noisy devices due to removal ofa source of possible charge traps on oxidised large exposed surface of AlAs or A1GaAs in GaAs/A1GaAs devices or AlAs or InAlAs in InGaAs/InAlAs devices.

Claims

CLAIMS1. A photon detector comprising a plurality of quantum dots, a sensing region and measurement means for measuring an electrical characteristic of the sensing region, said sensing region being located such that a change in the charged state of said quantum dots causes a change in the electrical characteristics of said sensing region, the detector further comprising photon counting means for determining the number of photons detected in a pulse of radiation from changes in the electrical characteristics of the sensing region wherein the counting means is configured to determine the number of photons from changes in the electrical characteristics of the sensing region as the detector returns to its non-illuminated state.
2. A photon detector according to claim 1, wherein the counting means is configured to count the number of sharp changes in the electrical characteristics measured by the measurement means due to the quantum dots returning to their non-illuminated state after excitation with a pulse of radiation.
3. A photon detector according to either of claims 1 or 2, wherein the counting means are configured to measure the magnitude of sharp changes in the electrical characteristics due to the quantum dots returning to their non-illuminated state.
4. A photon detector according to claim 3 when dependent on claim 2, wherein the counting means counts the number of sharp changes weighted by their magnitude.
5. A photon detector according to either of claims 3 or 4, further comprising means for comparing the magnitude of the sharp change with one or more predetermined levels.
6. A photon detector according to any of claims 3 to 5, wherein the magnitude of the signal is the amplitude of the signal.
7. A photon detector according to any preceding claim wherein the counting means further comprising a conditioning circuit to condition the output of the measurement means, wherein the conditioning circuit converts sharp changes in electrical signals into peaks.
8. A photon detector according to claim 7, wherein said conditioning circuit has a bandwidth of 2MHz or less.
9. A photon detector according to either of claims 7 or 8, wherein said conditioning circuit has a bandwidth of 200MHz or more.
10. A photon detector according to any preceding claim, wherein the sensing region comprises an active region having a channel and said measurement means comprises source and drain contacts to said channel.
11. A photon detector according to any of claims I to 10, wherein said detector is configured as a resonant tunnel structure wherein absorption of a pulse of radiation affects the tunnelling conditions.
12. A photon detector according to any preceding claim, wherein the reset time of the quantum dots is configured to be IOns or more.
13. A photon detector according to any preceding claim, wherein the reset time of the quantum dots is configured to be lp.s or less.
14. A photon detector according to any preceding claim, further comprising an absorption region.
15. A photon detector according to claim 14, wherein the cross section of said sensing region substantially perpendicular to the incident direction of photons is smaller than the cross sectional area of the absorption region in this plane.
16. A method of determining the number of photons received by a photon detector, comprising: providing a plurality of quantum dots, providing a sensing region, said sensing region being located such that a change in the charged state of said quantum dots causes a change in the electrical characteristics of said sensing region; measuring electrical characteristics of said sensing region; and determining the number of photons from the changes in the electrical characteristics of said sensing region wherein the number of photons are determined by counting the number of sharp changes in the electrical characteristics as the detector returns to its non-illuminated state.
17. A method according to claim 16, wherein the number of photons are determined from the magnitude of sharp changes in the electrical characteristics.