WO2007091031A2

WO2007091031A2 - Image sensor comprising a photosensitive dendrimer

Info

Publication number: WO2007091031A2
Application number: PCT/GB2007/000386
Authority: WO
Inventors: Paul Leslie Burn; Stephen Collins
Original assignee: Oxford University Innovation Ltd
Current assignee: Oxford University Innovation Ltd
Priority date: 2006-02-06
Filing date: 2007-02-05
Publication date: 2007-08-16
Anticipated expiration: 2008-08-06
Also published as: WO2007091031A3; GB0602347D0

Abstract

An image sensor comprises an array of photodiodes, each photodiode comprising a photo-sensitive material including a dendrimer arranged to receive incident light. The photodiodes may be separate or may be regions of a layer of photo-sensitive material. In respect of each photodiode, a first electrode and a second electrode is disposed on opposite sides of the respective photodiode to receive a photo-current from the photodiode. A MOS circuit layer is disposed under the photodiodes of photo-sensitive material. The MOS circuit provides detection circuits connected to the electrodes to detect the photo-currents. Use of a dendrimer provides the benefits of availability of desired properties to sensors. By provision of the dendrimer over the MOS circuit layer, the area of the pixel can be increased as compared to MOS pixel for a given resolution.

Description

IMAGE SENSOR

The present invention relates to the design and manufacture of image sensors using a dendrimer.

Dendrimers are macromolecules comprising a core and a number of tree-like branches extending from the core and forming a generational structure. The branches extend from the core in a highly predictable fashion to form structures with highly ordered architectures. The nature of the core and the branches can be precisely controlled in order to achieve the desired properties.

Other advantages of dendrimers include the ability to attach a number of different branches to a core, resulting in multifunctional macromolecules while retaining a high degree of control over the structure and properties of the dendrimer. Dendrimers can also bear functional groups, for example in the outer shell of the structure, which can be used to tune the properties of the dendrimer.

As a result of their unique architecture and construction, dendrimers possess inherently valuable physical, chemical and biological properties, and many applications have been identified. For example, dendrimers are known as solubilising agents, catalysts, drug delivery agents, polymeric additives and MRI contrast agents.

One particular area of interest has been the luminescent properties of dendrimers. The primary focus in this work has been on the photo-emissive properties of the dendrimer. One driver for this has been the desire to develop materials for use in displays. The photo-electric properties of the dendrimer are in general terms reversible. That is to say, the emissive property that an applied current is converted into light is accompanied by the reverse, photo-sensitive property that incident light is converted into a current, commonly referred to as a photo-current. Thus WO-99/21935 mentions the possibility of using the particular dendrimers disclosed therein in a photodiode but contain no further disclosure of the use of the photo-sensitive property.

The present invention is concerned with the use of a photo-sensitive dendrimer to produce a pixellated image sensor. In particular it is concerned with the structure and manufacture of an image sensor which allows advantage to be taken of the photo-sensitive property of the dendrimer.

According to the present invention, there is provided an image sensor comprising: an array of photodetectors, each photodetector comprising a photo-sensitive material including a dendrimer arranged to receive incident light; in respect of each photodetectors, a first electrode and a second electrode disposed on opposite sides of the respective photodetector to receive a photo-current from the photodetector, the first electrodes of the photodetectors of the array being separate from one another; and a MOS circuit layer extending across all the photodetectors on the same side of the photodetectors as the first electrodes and having formed therein a plurality of detection circuits connected to the first electrodes of respective photodetectors, the detection circuits including MOS devices and being arranged to detect the photo-current received by the respective first electrodes from the photodetectors.

This provides a pixellated image sensor because each photodetector in the array provides a pixel of an image. The use of dendrimers in the image sensor provides a number of advantages. First, the architecture of the dendrimers can be accurately designed in order to control the intermolecular interactions for optimising light absorption and charge transport. In addition, the conjugation length of chromophores within the dendrimers can be more accurately controlled than with conjugated polymers, resulting in properties which are easier to predict and control. A further advantage is that dendrimers can be designed to have a large number of chromophores per molecule, particularly when compared with fully delocalised conjugated polymers. This provides the potential for a single dendrimer to be responsive to different sources of light. A further advantage of dendrimers is that they can be solution processed, resulting in a device which is easier to produce.

As well as obtaining the intrinsic advantages of the use of a dendrimer, the particular structure of the image sensor involving a MOS circuit layer extending across all the photodetectors on one side of the photodetectors provides particular advantages over existing image sensors. Typically, such existing image sensors use photodetectors in the form of photodiodes formed in silicon by a p-n junction. Early commercial image sensors were manufactured using specialist CCD (charge-coupled device) processes, but more recently the drive to reduce cost has led to image sensors made using CMOS (complementary metal-oxide-semiconductor) processes that incorporate detection circuits into the same silicon chip as the photodiodes themselves.

Throughout the development of image sensors for electronic cameras, there has been a drive to reduce the size of the pixels by reducing the size of the MOS devices within the detection circuit. A critical aim of reducing the size has been to reduce the size of the pixels below the diffraction limit of the optics so that the image sensor does not limit the resolution of the camera. For typical lens systems, this means ensuring that each pixel has dimensions of the order of 5μm. For commonly used designs of detection circuit in which the circuit for each photodiode has a geometry of 14 times the geometry of MOS devices, this objective was achieved using manufacturing processes with a minimum feature size of the MOS devices of 0.35μm.

A further limitation of existing image sensors is generated by the need to produce a full colour image. In the simplest type, the CMOS circuit has a uniform structure and colour information is obtained by arranging the colour filter array over the photodiodes. The colour filter array consists of plural types of filters, typically red, green and blue filters, that limit the light incident on the photodiode to a limited band of the visible spectrum so that each photodiode captures information about that band of the spectrum. A colour de-mosaicing algorithm is then applied to the output of the image sensor to estimate the missing colours at each photodiode location which correspond to pixels, and create an image including data for each of the three colours at each pixel. The use of a colour filter array and de-mosaicing algorithm is very common, but does in fact involve a compromise between spatial information and accuracy of colour representation. In principle, an alternative approach to overcome this compromise would be to reduce the size of the photodiodes to include photodiodes which respond to each of the colours within the diffraction spot of the optics. However, this produces technical problems in the manufacture of the CMOS circuit. For example, with a typical arrangement of groups of four colour pixels, this would mean placing four photodiodes in an area of dimension of the order of 5μm. For the commonly used design of detection circuit mentioned above, this suggests that the minimum geometry of the MOS devices is 0.18μm. Unfortunately, many problems occur when scaling the design of the MOS devices to such sizes. These problems include (i) a need to include a silicidation processing step, (ii) the reduction in depletion depth causing a loss of sensitivity, (iii) a reduction in the maximum voltage usable, (iv) an increased risk of light divergence and spatial waveguiding between the filters and the semiconductor, and (v) an increase in electrical cross-talk. Whilst these problems can be alleviated to some extend by modifications of the circuit design or manufacturing process, such modifications will themselves increase the cost of the final product.

Compared to such existing image sensors, the particular structure of the image sensor and consequently the present invention provides advantages. As a result of the MOS circuit layer being disposed on one side of the photodetectors, both the photo-sensitive dendrimer and the components of each detection circuit which are separate for respective photodetectors may be arranged in generally the same area of the image sensor. Therefore, for a given resolution, the actual photo-sensitive area of the photodetector may be increased as compared to a conventional CMOS image sensor, because it is not necessary to include the photodetectors in the same silicon layer as the MOS circuits. Typically, this means the photo-sensitive area of the photodetectors is approximately doubled. This improves the sensitivity of a single pixel in proportion to the increase in the area.

Conversely, for a given size of photodetector, it is possible to use MOS devices which are larger. For example, for a given size of photodetector which would require MOS devices having a minimum dimension of 0.18μm in a conventional CMOS sensor, the present invention allows the same detection circuit to be formed from MOS devices having a minimum dimension of 0.25μm. This considerably reduces the problems discussed above which occur when the size of the MOS circuit is scaled down.

Thus, as compared to conventional CMOS circuits, the image sensor of the present invention provides the advantage of either increasing the sensitivity or facilitating reduction of the size of the photodetector or both, depending on the trade-off between reducing the size of the photodetectors and reducing the size of the MOS devices. For example, the image sensor of the present invention opens the possibility of arranging four colour photodetectors within an area of dimensions in the order of 5μm which corresponds to the diffraction limit of a typical optical system, which using MOS devices having a minimum dimension of the order of 0.25μm.

The performance advantages are achieved whilst maintaining the advantages in design and manufacture of using a MOS circuit to form the detection circuits. In particular, the MOS circuit layer and the first electrodes of the respective photodetectors may be manufactured using conventional techniques in the field of MOS circuits. The subsequent deposition of the dendrimer may be performed by a number of different techniques, some examples of which are described in detail below.

Thus, in accordance with the present invention there is also provided a method of manufacturing an image sensor comprising: forming a MOS circuit layer having formed therein a plurality of detection circuits including MOS devices, each detection circuit being arranged to detect a photo-current generated by a photodetector comprising a photo-sensitive material including a dendrimer; forming an array of separate, first electrodes over the MOS circuit layer, each connected to a respective detection circuit; forming respective photodetectors comprising a photo-sensitive material including a dendrimer on the respective first electrodes; and in respect of each photodetector, forming a second electrode on the photodetector.

Advantageously, each photodetector comprises a separate piece of photo-sensitive material. This reduces cross-talk between the pixels. Furthermore it facilitates the formation of an image sensor including respective photodetectors in the array comprising photo-sensitive dendrimers of a plurality of different types which are photo-sensitive to light of different wavelengths.

However, the use of pixels comprising separate pieces of photo-sensitive material is not essential. In an alternative configuration, the image sensor comprises a layer of photo-sensitive material extending over all the first electrodes, the photo-sensitive material of the photodetectors being respective regions of the layer of photo-sensitive material.

In terms of the nature of the dendrimer, a huge range of dendrimers could potentially be used provided at least a portion of the dendrimer is capable of absorbing energy in the form of light and convert at least a portion of this into electrical energy. In effect, at least a portion of the dendrimer must be photo-sensitive. Exemplary dendrimers include those comprising at least partially conjugated dendritic molecular structures such as those described in WO-A-02/66552.

The entire dendrimer need not be photo-sensitive. For example, the core of the dendrimer can bear other dendritic molecular structures such as polyamidoamine (PAMAM) dendrimers or Frechet dendrimers (thus, such DENDRON¹ groups may include ether-type aryl dendrons).

A general structure for suitable dendrimers cart be defined by the formula (I): [DENDRON¹I_1n-CORE-[DENDRON]_n (I) wherein n is an integer of one or more, m is zero or an integer of one or more, CORE is a central group to which the DENDRON and DENDRON¹ groups are attached, the DENDRON groups, which are the same or different, are at least partially conjugated dendritic molecular structures, and DENDRON¹ groups are dendritic molecular structures.

To allow better understanding, embodiments of the present invention will now be described by way of non-limitative example with reference to the accompanying drawings, in which:

Fig. 1 is a cross-sectional side view of a pixel of a first image sensor;

Fig. 2 is a cross-sectional side view of a pixel of a second image sensor;

Fig. 3 is a cross-sectional side view of a pixel of a third image sensor;

Fig. 4 is a circuit diagram of a linear detection circuit of a photodiode;

Fig. 5 is a circuit diagram of a logarithmic detection circuit of a photodiode;

Fig. 6 is a flow chart illustrating the manufacturing method of a image sensor;

Fig. 7 is a cross-sectional view of a MOS circuit after a first stage of manufacture;

Fig. 8 is a cross-sectional view of the MOS circuit of Fig. 7 having a deposited electrode after a second stage manufacture;

Fig. 9 is a cross-sectional view of the MOS circuit of Fig. 8 after deposition of a protective overlayer;

Fig. 10 is a cross-sectional view of the MOS circuit of Fig. 9 after formation of a well in the protective overlayer;

Fig. 11 is a diagram of a first arrangement of coloured pixels in an image centre;

Fig. 12 is a diagram of a second arrangement of coloured pixels in an image centre; and

Fig. 13 is a plan view of the arrangement of detection circuits in a MOS circuit layer of an image sensor.

There will first be described three image sensors which embody the present invention. These image sensors have many elements in common. For brevity, the common elements will be given same reference numerals and a description thereof will not be repeated. The three image sensors are shown in Figs. 1 to 3, respectively. These drawings are cross-sectional views in which the thickness of the individual layers is exaggerated for clarity. In Figs. 1 to 3 the direction of incident light is from above. References in the following text to relative locations as being above or below are to given relative to the incident light and do not imply any orientation of the image sensor.

The first image sensor shown in Fig. 1 comprises a semiconductor chip 1 on which there is formed a MOS circuit layer 2 in which are formed detection circuits implemented by MOS devices 3 which are CMOS devices. All the MOS devices 3 are MOSFETs (metal-oxide-semiconductor field effect transistors) and have a construction which is conventional for a MOSFET. For clarity, Fig. 1 shows only a single MOS device 3, although the detection circuits in fact include plural MOS devices 3 as will be described further below.

The MOS circuit layer 2 comprises a silicon substrate 4 having formed thereon an insulating layer 5 made of dielectric material. The individual MOS devices 3 are each formed as shown in Fig. 1. At two spaced locations, the semiconductor layer 4 has a source 6 and a drain 7, each formed as a doped region extending from the upper surface of the semiconductor layer 4. A channel 8 is formed by the region of the semiconductor layer 4 between the source 6 and the drain 7. A gate 9 is formed in the insulating layer 5 above the channel 8. The gate 9 is a conductor which may be a metal but is more typically poly-silicon. The gate 9 is separated from the channel 8 by a gate oxide layer 10 formed in the insulating layer 5.

The MOS device 3 operates in a conventional manner by means of the voltage on the gate 9 generating an electric field in the channel 8 which controls the conductivity of the channel 8 and hence the current flow between the drain 6 and the source 7.

The source 6 and gate 7 are electrically contacted by a source contact 11 and a drain contact 12, respectively. The contacts 11 and 12 are formed from a conductive material, typically metal. The contacts 11 and 12 extend through vias in the insulating layer 5. The contacts 11 and 12 extend onto the upper surface of the insulating layer 5 where they are patterned to contact other electrical components such as other MOS devices 3.

On top of the insulating layer 5 is formed a protective layer 13 also made of a dielectric material. The protective layer 13 provides electrical insulation and also physical protection of the. contacts 11 and 12.

On top of the protective layer 13, the semiconductor chip 1 is provided with an array of first electrodes 14. Fig. 1 shows a single complete first electrode 14 and part of an adjacent first electrode

14 but the image sensor in fact comprises many first electrodes 14 arranged in a two-dimensional array. The first electrodes 14 are typically square and arranged in a regular square array but in principle may be of any shape and in an array of any shape. The first electrodes 14 are separated by gaps 15 which are sufficient to provide electrical insulation between the separate first electrodes 14.

Each first electrode 14 is connected to a respective detection circuit formed in the MOS circuit layer 2. As shown in Fig. 1, the connection is formed by a conductive arm 16 extending through a via in the protective layer 13 from the first electrode 14 to the source contact 11 of a MOS device 3 which forms part of the detection circuit. The first electrode 14 and the conductive arm 16 are formed from a conductive material, typically metal. The first electrodes 14 extend over the detection circuits formed in the MOS circuit layer 2. The first electrodes 14 thus occupy an area of the semiconductor chip 1 of generally the same size as a single detection circuit, that is ignoring the gaps

15 which will be of minimal size and ignoring the spaces between the individual components of the detection circuit which again will be relatively small in view of the desire to maximise the size of the individual MOS devices 3.

On top of the protective layer 13 and the first electrodes 14, the semiconductor chip 1 has deposited thereon a layer 17 of photo-sensitive material comprising a photo-sensitive dendrimer as described in more detail below. The layer 17 of photo-sensitive material extends continuously over all the first electrodes 14. Deposited on top of the layer 17 of photo-sensitive material is a common electrode 19 which extends in common over all the pixels 20 formed in the layer 17 of photo-sensitive material. The common electrode 19 is formed from a conductive material which is transparent to the incident light, typically ITO.

In operation, a reverse bias is applied between the first electrodes 14 and the common electrode 19. When the layer 17 of photo-sensitive material receives incident light, the photo-sensitive dendrimer in the layer 17 of photo-sensitive material, due to its photo-sensitive nature, absorbs incident photons to form excitons which are subsequently separated into separate charges, that is electrons and holes. In operation, the electrons and holes flow to the first electrodes 14 and the common electrode 19 as photo-current between the first electrodes 14 and the common electrode 19. The photo-current from respective regions 20 of the layer 17 of photo-sensitive material is collected by different first electrodes 14. In Fig. 1, the boundaries between the respective regions 20 of the layer 17 of photo-sensitive material are shown by dotted lines 18 which are imaginary, although in reality the boundaries will not be precise, which may lead to a degree of cross-talk between the pixels. Consequently, each of the respective regions 20 of the layer 17 of photo-sensitive material constitute a respective photodiode, and the common electrode 19 constitutes a second electrode in respect of every photodiode.

As the dendrimer in the layer 17 of photo-sensitive material is the same across the entire layer 17 of photo-sensitive material, the photo-sensitivity with wavelengths is the same across the entire image sensor. As shown in Fig. 1, the first image sensor is a mono-chrome image sensor. Alternatively, the first image sensor shown in Fig. 1 may be used in combination with a colour filter array so that different regions 20 of the layer 17 of photo-sensitive material which constitute different photodiodes receive light of different colours, in the same manner as a conventional CMOS image sensor, although the use of such a colour filter array reduces the effective spatial resolution.

The second image sensor shown in Fig. 2 has the same construction as the Erst image sensor shown in Fig. 1 except that the layer 17 of photo-sensitive material is absent and instead a separate piece 21 of material comprising a photo-sensitive dendrimer is formed on top of each respective first electrode 14. The common electrode 19 extends continuously over all the pieces 21 of material, as well extending down the sides of the pieces 21 of material and across the upper surface of the protective layer 13 therebetween.

The separate pieces 21 of material in operation respond to incident light in the same manner as the regions 20 of the layer 17 of photo-sensitive material in the first image sensor. Accordingly, each separate piece 21 of material constitutes a respective photodiode which generates photo-current from incident light.

The third image sensor shown in Fig. 3 has the same construction as the first image sensor show in Fig. 1 except that the layer 17 of photo-sensitive material is replaced by the following elements. An overlayer 22 of dielectric material is formed on top of the protective layer 13. The overlayer 22 has wells formed over each respective first electrode 14 and extending entirely through the overlayer 22. Thus the remaining overlayer 22 extends around each first electrode 14 in the gaps 15 between the first electrodes 14. In each of the wells formed in the overlayer 22, there is deposited a respective piece 23 of photo-sensitive material comprising a photo-sensitive dendrimer. The separate pieces 23 of material each fill the respective well and the common electrode 19 extends continuously over all the separate pieces 23 of material and the protective overlayer 22.

The separate pieces 23 of material in operation respond to incident light in the same manner as the regions 20 of the layer 17 of photo-sensitive material in the first image sensor. Accordingly, each separate piece 23 of material constitutes a respective photodiode which generates photo-current from incident light.

In the case of each of the second image sensor of Fig. 2 and the third image sensor of Fig. 3, as the individual pieces 21 and 23 of material are separate from each other, it is straightforward for the pieces 21 and 23 to comprise different photo-sensitive dendrimers which are photo-sensitive to light of different wavelengths. This facilitates the second and third image sensors producing a full colour image by use of dendrimers which are photo-sensitive to light of different bands of the visible spectrum.

The detection circuits formed in the MOS circuit layer 2 will now be described. A linear detection circuit is shown in Fig. 4 and a logarithmic detection circuit is shown in Fig. 5. Either of the linear detection circuit of Fig. 4 or the logarithmic detection circuit of Fig. 5 may be used in any of the first to third image sensors.

Each of Figs 4 and 5 show the detection circuit for a single photodiode P which corresponds to the regions 20 of the layer 17 of photo-sensitive material in the case of the first image sensor and to the pieces 21 and 23 of material in the case of the second and third image sensors, respectively.

The detection circuits comprise MOS devices Ml to M7 (which are MOS devices 3 as shown in Figs. 1 to 3) and a capacitor C_pj_Mi. The MOS devices Ml to M7 are shown as single devices, but they may in fact be implemented by plural devices such as a pair of devices or more complicated MOS circuit elements. The capacitor C_p^_ei may also be implemented as a MOS device.

The components 50 of the detection circuit shown in the dotted outline including MOS devices Ml, M2 and M3 and the capacitor C_pixei are formed separately for each photodiode P. These separate components 50 from all the photodiodes P in the array are physically disposed adjacent the first electrodes 14, that is under the first electrodes 14 in Figs. 1 to 3. Most conveniently, all the separate components 50 of a single detection circuit are disposed adjacent the first electrode 14 to which the detection circuit is connected. This is not essential and alternatively there can be some displacement so that the separate components 50 of respective detection circuits are adjacent first electrodes 14 connected to other detection circuits. However even in this case, viewing all the detection circuits together, it remains the case that the detection circuits separate components 50 are physically disposed adjacent the first electrodes 14.

The components of the detection circuit shown in dotted outline 51 including MOS devices M4, M5 and M6 are shared in common by all the detection circuits of a column (or row) of photodiodes P and may be physically arranged outside the array of first electrodes 14. Similarly, the component of the detection circuit shown in dotted outline 52 including MOS device M7 is shared in common by all the detection circuits of the entire array of photodiodes P and may be physically arranged outside the array of first electrodes 14.

One possible physical arrangement of the MOS devices Ml to M7 in the MOS circuit layer 2 is shown in Fig. 13 which illustrates the positions in the MOS circuit layer 2 of the components 50, 51 and 52 of the detection circuits with respect to the position of an array of first electrodes 14 (of small size for clarity, real image sensors typically being larger). As can be seen, the shared components 50 are adjacent the first electrodes 14 and the common components 51 and 52 are outside the first electrodes 14.

Fig. 13 also shows the vertical scanning circuit 53 and the horizontal scanning circuit 54 of the detection circuits. Optionally the circuit layer 2 could include further circuit elements to perform other functions.

The linear detection circuit of Fig. 4 is arranged as follows. The photodiode P and a capacitor C_pi_xei are connected to ground and in parallel to the drain of a control device Ml which is a PMOS device whose source is connected to the rail at voltage VDD and whose gate is controlled by a reset voltage V_reset. The capacitor C_pixei may be an actual capacitor formed in the circuit layer 2 or may be the capacitance of the MOS devices connected to the photodiode P. The control device Ml controls charging and discharging of the capacitor C_pjχ_ei by the photo-current from the photodiode P. When the control device Ml is switched on by application of a low reset voltage V_reset, the voltage V_pjχ_ei across the capacitor C_pj_xeι is clamped to the high voltage V_DD of the rail. This resets the voltage V_Pjχ_ei. When the control device Ml is switched off by application of a high reset voltage V_reSet, the capacitor C_p^i is discharged by the photo-current from the photodiode P.

The gates of the control devices Ml of each row (or column) of photodiodes P are all connected to a respective track extending across the chip 1 so that the reset voltages V_reset of each row (or column) of photodiodes P are controlled in common by the vertical scanning circuit 53.

In operation, a photodiode P is operated by first setting the reset voltage V_reset low to reset the voltage V_pi_xei across the capacitor C_pj_xei and subsequently setting the reset voltage V_reset high to allow the capacitor C_pbie\ to be discharged by the photodiode P. After an integration time t, the voltage V_pixei across the capacitor C_piχ_ei is sampled as a voltage sample V_sampi_e by a sampling circuit described further below. At the end of the integration time t, the voltage sample V_sampi_e is related to the photo- current I_ph by the equation:

Vsampie ⁼ VDD - I_ph ■ t / C

In this equation C represents the capacitance on the node, that is the capacitance of the capacitor C_pixei and any stray capacitance. This equation assumes a constant photo-current I_ph. More generally the term (I_ph . t) may be replaced by the integral of the photo-current I_ph over the discharge time t. in any event it is evident that the voltage sample V_sampi_e varies linearly with the photo-current I_ph (or average photo-current I_ph if the photo-current I_ph is not constant over the integration time t).

The sampling circuit may selectively sample the voltage V_pixci across the capacitor C_p^ for any one of the pixels P and is arranged as follows. The photodiode P is connected to a column (or row) readout circuit formed by devices M2, M3 and M4 which are NMOS devices connected in series with each other between the rail of voltage V_DD and ground. The current source device M4 operates as a current source and has a bias voltage V_bi_asi applied to its gate. The photodiode P is connected to the gate of the follower device M2 and thus the follower device M2 and the current source device M4 together form a source follower circuit controlled by the switch device M3 which acts as a switch controlled by a select voltage V_sei_eotl applied to its gate. When the switch device M3 is switched on by application of a high select voltage Vsei_ecth ^ source of the current source device M4 is an output voltage V_outi which follows the voltage V_pixel on the photodiode P. When the switch device M3 is switched off by application of a low select voltage V_sei_ecti, the source follower circuit is opened.

The current source device M4 is shared in common by a column (or row) of photodiodes P whereas the detection circuit of each photodiode P has its own switch device M3. The gates of the switch devices M3 of each row (or column) of photodiodes P are all connected to a respective track extending across the chip 1 so that the select voltages V_sei_ecti of each row (or column) of photodiodes P are controlled in common by the vertical scanning circuit 53.

The column (or row) readout circuit is connected to a row (or column) readout circuit formed by devices M5, M6 and M7 which are PMOS devices connected in series with each other between ground and the rail of voltage VQD- The column (or row) readout circuit has a similar construction and operation to the row (or column) readout circuit. The current source device M7 operates as a current source and has a bias voltage V_bi_aS2 applied to its gate. The output voltage V_outi from the row (or column) readout circuit is connected to the gate of the follower device M5 and thus the follower device M5 and the current source device M7 together form a source follower circuit controlled by the switch device M6 which acts as a switch controlled by a select voltage V_sei_ect2 applied to its gate. When the switch device M6 is switched on by application of a low select voltage V_Sei_ect2j ^me source of the current source device M7 is an output voltage V_out2 which follows the output voltage V_outi from the row (or column) readout circuit and hence also the voltage V_pjχ_ei on the photodiode P. When the switch device M6 is switched off by application of a high select voltage V_sei_ect2! the source follower circuit is opened.

The current source device M7 is shared in common by the entire array of photodiodes P whereas the detection circuit of each row (or column) of photodiodes P has its own switch device M6. The gates of the switch devices M6 of each column (or row) of photodiodes P are all connected to a track extending across the chip 1 so that the select voltages V_sei_ect2 of each column (or row) of photodiodes P are controlled in common by the horizontal scanning circuit 54.

By choice of the select voltages V_sei_ecti of each row (or column) and the select voltages V_sei_ect2 of each column (or row), the output voltage V_out2 may selectively sample the voltage V_p^-i on any single photodiode P at a time. The select voltages V_sei_ecti and V_sei_ecQ are timed by the vertical scanning circuit 53 and by the horizontal scanning circuit 54 to sample the voltage sample V_samp|_e, that is the voltage V_pi_xei across the capacitor C_pj_xei at the end of the integration time t. The output voltages V_out2 are the sample voltages V_sampi_e less an offset dropped by the source follower circuits, and hence vary linearly with the photo-current I_ph. In operation, each photodiode P is scanned in turn by: (1) controlling the reset voltage V_reset of the photodiode P as described above and (2) subsequently at the end of the integration time t, applying select voltages V_sei_ecti and V_sei_ect2 to sample the voltage sample

"sample- In principle the order in which the photodiodes P are scanned does not matter, but conveniently the rows of photodiodes P are scanned in turn by the vertical scanning circuit 53 and, for each row, the photodiodes P of the row are scanned in turn by the horizontal scanning circuit 54.

The linear detection circuit of Fig. 4 has the same form as is known for a detection circuit in a conventional CMOS image sensor. It is commonly used in a consumer camera used to take photographs.

One problem with the linear detection circuit of Fig. 4 is that its output dynamic range is limited to approximately three decades of illumination. This is adequate for scenes with controlled illumination, but is often inadequate for external scenes with a wide range of illuminations caused by shadows. For applications that require imaging of external scenes wide dynamic range photodiodes have been proposed and are being developed. The logarithmic detection circuit of Fig. 5 is one such circuit with a wide dynamic range and is arranged as follows.

The logarithmic detection circuit of Fig. 5 is superficially similar to the linear detection circuit of Fig. 4 in that the photodiode P is connected between ground and the drain of a load device M8 whose source is connected to the rail at voltage V_DD. However, in this logarithmic detection circuit the gate of the load device M8 is also connected to the photodiode P. This the load device M8 is designed with respect to the properties of the photodiode P so that under typical illumination conditions the load device M8 is operating in a regime known as weak inversion or subthreshold. In this operating regime, the gate-source voltage of the load device M8 is proportional to the logarithm of the photo-current I_ph flowing through the load device M8 from the photodiode P. Hence the voltage V_pi_xei °ⁿ the photodiode P varies with the logarithm of the photo-current I_ph in accordance with the equation:

V_pi_xei = V_offset - ( G . ln( I_ph / I_ref) ) where V_offset, G, and I_ref are parameters of the photodiode P and the load device M8 that characterize the response of the photodiode P.

The voltage V_pixei is sampled by a sampling circuit which is identical to that of the linear detection circuit of Fig. 4 and so for brevity a description thereof will not be repeated.

The operation is generally the same as that of the linear detection circuit of Fig. 4 except that the photo-current flows continuously through the load device M8 and tiiere is no need to reset the voltage V_pi_xei on a photodiode P and perform an integration. Thus, one option in operation is that each row (or column) of photodiodes P is scanned in turn by applying select voltages V_sei_ect to that scanned row so that the output voltages V_out of the photodiodes P of the scanned row (or column) are selected to appear on the source of the current source device M4. These output voltages V_out are the voltages V_pi_xei less an offset dropped by the source follower circuits, and hence vary logarithmically with the photo-current I_ph.

Alternatively, the absence of need to perform an integration can be exploited in various ways to change the readout scheme. In some applications, this may allow the selective readout of a subset of photodiodes P at a higher sampling (or frame) rate. In other applications, particularly when the camera is imaging a scene containing moving objects, the voltages V_p^i on the photodiodes P in the entire array can be sampled onto the source follower circuit in order to avoid motion smearing of the objects in the scene.

The manufacture of the first to third image sensors will now be described. As shown in Fig. 6, the manufacture involves five stages Sl to S5.

The first stage Sl is formation of the MOS circuit layer 2. This involves formation of the detection circuits on the silicon substrate 4 and deposition of the protective layer 13. The resultant semiconductor chip 1 is shown in Fig. 7. The first stage Sl may be performed using conventional MOS processing techniques, for example as used to produce CMOS image sensors.

The second stage S2 is deposition of the first electrode 14 which results in the semiconductor chip 1 shown in Fig. 8. The second stage may include a further step of depositing an overlayer 22 which results in the semiconductor chip 1 shown in Fig. 9. In subsequent stages, the overlayer 22 may be removed or may form the overlayer 22 of the third image sensor. This second stage S2 may again be performed using conventional techniques for depositing material on a MOS chip.

The first and second stages Sl and S2 are typically performed in an existing silicon fabrication facility.

The third stage S3 is deposition of the photo-sensitive material of the layer 17 or of the separate pieces 21 and 23. The processing in the third stage S3 may differ for the first to third image sensors.

To produce the first and second image sensors, the third stage S3 comprises a step of etching away the entire overlayer 22 and subsequently a step of depositing the layer 17 of photo-sensitive material or the pieces 21 of material. This produces a sensor as shown in Figs. 1 or 2, but without the common electrode 19.

To produce the third image sensor, the third stage S3 comprises a step of selectively etching away wells 40 in the overlayer 22, resulting in a chip 1 as shown in Fig. 10, and subsequently a step of depositing the pieces 23 of material in the wells 40. This produces an image sensor as shown in Fig. 3, but without the common electrode 19.

The step of selectively etching away wells 40 in the overlayer 22 may be performed using the same etching techniques as are currently performed to selectively remove portions of the overlayer to create contact windows that lie over contact pads made of the topmost metal layer. Normally these contact pads are created on the edge of the substrate and are relatively large to enable wire connections to be made to the contact pads using a process known as bonding, and similar techniques may be applied to etch the wells 40.

The dendrimer itself is prepared using conventional techniques and the step of depositing the dendrimer may be performed in a variety of ways with the choice dependent on the need for pixel resolution and absorption colour.

To produce the first image sensor, the layer 17 of photo-sensitive material can be simply deposited over the whole of the surface, for example by spin-coating or ink-jet printing the material.

To produce the second and third image sensors, it is necessary to deposit the pieces 21 or 23 of material of different types separately. This requires more controlled techniques but there are a number of options such as photo-patterning, ink-jet printing, and thermal transfer. Whilst any of these options may be applied to either of the second and third image sensors, photo-patterning and thermal transfer are more easily applied to produce the second image sensor than the third image sensor because these techniques are facilitated by the flat surface of the protective layer 13 as opposed the uneven surface formed by the wells 40 in the overlayer 22.

Photo-patterning may be carried out in a conventional manner, including several individual processing steps. First, a dendrimer of a particular colour is deposited from solution (an activating agent is sometimes included as disclosed for example in C. D. Mϋller, A. Falcou, N. Reckefuss, M. Rojahn, V. Wiederhirn, P. Rudati, H. Frohne, O. Nuyken, H. Becker, K. Meerholz, Nature, 2003, 421, 829). A mask is applied and the film is exposed to irridiation to cause crosslinking in the exposed regions. The mask is then removed and the uncrosslinked material is removed by washing with a suitable solvent. The process is repeated until all the pixels of the required colours are deposited.

Thermal transfer may be used in the manner described in US 2005/0123850.

Ink-jet printing is more easily applied to produce the third image sensor than the second image sensor because the wells 40 facilitate printing and reduce the spreading of the dendrimer into unwanted areas as the solvent evaporates. This concept has been used for ink-jet printed conjugated polymers where polyimide walled wells are produced for organic light-emitting diodes.

The fourth stage S4 is the deposition of the common electrode 19 to produce an image sensor as shown in Figs. 1 to 3. This fourth stage S4 may again be performed using conventional techniques for depositing material on a MOS chip. The common electrode 19 may be connected to a pad that provides the required voltage bias signal.

The fifth stage S5 is to encapsulate the semiconductor chip 1 shown in Figs. 1 to 3, for example between two plates (not shown). This stage is not essential in terms of the operation of the image sensor but is beneficial in producing a robust device.

Of course various modifications to the manufacturing method are possible. One alternative is for the second stage S2 of depositing the first electrode 14 to be performed after removal of the overlayer 22. The photo-sensitive dendrimers which may be used in image sensors in accordance with the present invention will now be described.

As discussed earlier, there is a huge range of dendrimers which could be used in the invention. Generally the dendrimers will be defined by formula (I). There must be at least one moiety present in the dendrimer which is capable of at least partially absorbing light energy to form an exciton that can be separated to form positive and negative charges, i.e. the dendrimer must be photosensitive. The photo-sensitive moiety or moieties may be in any part of the dendrimer but it is preferably within the CORE or DENDRON groups or both.

As used herein, the phrase "at least partially conjugated" means that at least a portion of DENDRON is made up of alternating multiple (including double and/or triple) bonds and single bonds or lone pairs, apart from the surface groups. In some embodiments, all the dendrons or branching structures can be made up of alternating single or multiple bonds or lone pairs; such a structure being termed a conjugated dendron. However this does not mean that the p system is fully delocalised. The delocalisation of the p system is dependent on the regiochemistry of the attachments.

As used herein the term Ci_-I5 alkyl is a linear or branched alkyl group or moiety containing from 1 to 15 carbon atoms such as a Ci_-8 alkyl group or moiety or a Ci_-4 alkyl group or moiety. Examples of Ci_-4 alkyl groups and moieties include methyl, ethyl, n-propyl, i-propyl, n-butyl, i-butyl and t-butyl. Similarly, a C₂.₁₅ alkenyl group or moiety is a linear or branched alkenyl group or moiety containing from 2 to 15 carbon atoms respectively such as a C_2-8 alkenyl group or moiety or a C_2-4 alkenyl group or moiety. For the avoidance of doubt, where two or more alkyl or alkenyl moieties are present in a group, the alkyl or alkenyl moieties may be the same or different.

As used herein, a halogen is typically chlorine, fluorine, bromine or iodine. It is preferably chlorine, fluorine or bromine.

As used herein the term amino represents a group of formula NH₂. The term Ci_-I5 alkylamino represents a group of formula NHR' wherein R' is a Ci_-I5 alkyl group, preferably a Ci_-I5 alkyl group, as defined previously. The term Ci_-15 dialkylamino represents a group of formula NR 'R" wherein R' and R" are the same or different and represent CM₅ alkyl groups, preferably Ci_-6 alkyl groups, as defined previously.

As used herein the term aryl refers to C_6-I4 aryl groups which may be mono-or polycyclic, such as phenyl, naphthyl and fluorenyl. An aryl group may be unsubstituted or substituted at any position. Unless otherwise stated, it carries 0, 1, 2 or 3 substituents. Preferred substituents on an aryl group include halogen, Q_-I5 alkyl, C_2-I5 alkenyl, C(O)R wherein R is hydrogen or Ci_-I5 alkyl, CO₂R wherein R is hydrogen or Ci_-I5 alkyl, hydroxy, Ci_-I5 alkoxy, C_2-I5 alkenyloxy, Ci_-J5 alkylthio, C_2-I5 alkenylthio, Ci_-6 haloalkyl, C_2-I5 haloalkenyl, Ci_-I5 haloalkoxy, C_2-I5 haloalkenyloxy, amino, CM₅ alkylamino, di(Ci.i₅)alkylamino, C₆._H aryloxy, O₂SR wherein each R is the same or different and represents C 1-15 alkyl or C2-15 alkenyl, SiR3 wherein each R is the same or different and represents hydrogen, Cl-15 alkyl or C2-15 alkenyl, C₆-_I4 arylthio, C_6-I4 aryl and 5- to 10-membered heteroaryl, and wherein the substituents are themselves unsubstitated or substituted. When the substituents are themselves substituted, suitable substituents on the substituents include 1, 2, 3 or 4 groups selected from C_1-15 alkyl, C_2-I5 alkenyl, Ci_-I5 alkoxy, C_2-is alkenyloxy, hydroxy and halogen. Particularly suitable are 1 or 2 groups selected from C_1-8 alkyl, C_2-S alkenyl, C_1-8 alkoxy and C_2-8 alkenyloxy. In particular, when an aryl group is substituted by a Ce_-I4 aryl group or by a 5- to 10-membered heteroaryl group, these substituents are themselves unsubstituted or substituted by one or more substituents selected from Cj_-15 alkyl, C_2-I5 alkenyl, Ci_-15 alkoxy and C_2-15 alkenyloxy. When an aryl group is substituted by groups other than C_6-14 aryl groups or 5- to 10-membered heteroaryl groups, the substituents are themselves preferably unsubstituted.

As used herein, a heteroaryl group is typically a 5- to 14-membered aromatic ring, such as a 5- to 10-membered ring, more preferably a 5- or 6-membered ring, containing at least one heteroatom, for example 1, 2 or 3 heteroatoms, selected from O, S and N. Examples include pyridyl, pyrazinyl, pyrimidinyl, pyridazinyl, furanyl, thienyl, pyrazolidinyl, pyrrolyl, oxadiazolyl, isoxazolyl, thiadiazolyl, triazinyl, thiazolyl, imidazolyl, pyrazolyl, oxazolyl, isothiazolyl, benzofuranyl, isobenzofuranyl, benzothiophenyl, indolyl, indazolyl, carbazolyl, acridinyl, purinyl, cinnolinyl, quinoxalinyl, naphthyridinyl, benzimidazolyl, benzoxazolyl, quinolinyl, quinazolinyl and isoquinolinyl. When the heteroaryl group is a monocyclic heteroaryl group, preferred groups include thiophenyl, pyrrolyl, pyridyl, imidazolyl, triazinyl and triazolyl.

As used herein, references to a heteroaryl group include fused ring systems in which a heteroaryl group is fused to an aryl group. When the heteroaryl group is such a fused heteroaryl group, preferred examples are fused ring systems wherein a 5- to 6-membered heteroaryl group is fused to one or two phenyl groups. Examples of such fused ring systems are benzofuranyl, isobenzofuranyl, benzopyranyl, cinnolinyl, carbazolyl, benzotriazolyl, phenanthridinyl, benzothiophenyl, indolyl, indazolyl, benzimidazolyl, benzoxazolyl, quinolinyl, quinazolinyl and isoquinolinyl moieties.

A heteroaryl group may be unsubstituted or substituted at any position. Unless otherwise stated, it carries 0, 1, 2 or 3 substituents. Preferred substituents on a heteroaryl group include those listed above in relation to aryl groups. When a heteroaryl group is substituted by a C_6-I4 aryl group or by a 5- to 10-membered heteroaryl group, these substituents are themselves unsubstituted or substituted by one or more substituents selected from C_1-15 alkyl, C_2-15 alkenyl, C_1-I5 alkoxy and C_2-I5 alkenyloxy. When a heteroaryl group is substituted by groups other than Cg_-14 aryl groups or 5- to 10-membered heteroaryl groups, the substituents are themselves preferably unsubstituted.

As used herein, an alkoxy group is typically a said alkyl group attached to an oxygen atom. Similarly, alkenyloxy groups and aryloxy groups are typically a said alkenyl group or aryl group respectively attached to an oxygen atom. An alkylthio group is typically a said alkyl group attached to a thio group. Similarly, alkenylthio groups and arylthio groups are typically a said alkenyl group or aryl group respectively attached to a thio group.

Considering in turn each of the components of the dendrimer of formula (I), the function of the CORE is primarily to provide a central group to which the DENDRON groups can bond. However, the nature of the CORE can also be controlled in order to change the photo-sensitive properties of the dendrimer.

Particularly suitable COREs in formula (I) include metal ions or groups containing a metal ion, and non-polymeric organic groups. Exemplary non-polymeric organic groups include aryl and heteroaryl groups such as fluorenes, naphthalenes and porphyrin and perylene rings. As used herein, "non-polymeric" means that the core is not a polymeric group, although it may be in the form or a dimer, trimer or oligomer, or may be macrocyclic. When the core is in the form of an oligomer consisting of a number of units, it will preferably contain four or fewer units. Suitable units are single aryl or heteroaryl groups (e.g. a single fluorene unit). When it is a dimer, trimer or oligomer, it may comprise more than one such aryl or heteroaryl group, which are the same or different, bonded together and/or connected via alkenyl and/or acetylenyl groups, and optionally substituted. For instance, suitable cores include difluorene, trifluorene and biphenyl groups as well as other combinations of single aryl and/or heteroaryl groups such as phenyl and thiophenyl. Other suitable organic cores include aryl-substituted alkyl groups, such as tetraphenylmethane.

When CORE is a metal ion or a group containing a metal ion, it typically comprises a metal cation and attached ligands; i.e. the ligands form part of the core itself. The metal is typically near the centre of the core. It is preferred that the metal ion chromophore is sited at the core of the molecule, because then it will be relatively isolated from the core chromophores of adjacent molecules. The atoms or groups coordinating/binding to the metal typically form part of the core itself, e.g. the 2- phenylpyridine ligands of a fac-tris(2-phenylpyridyl)iridium (III) complex to which the DENDRONs can then bond.

Integer n is one or more, for example from 2 to 10. The number of DENDRONs surrounding the CORE can vary significantly, and in particular can be controlled by altering the relative sizes of the DENDRONs and the CORE. Preferred values of n are from two to six. Each DENDRON present in a dendrimer is independently chosen. Thus, a single dendrimer may contain a number of different DENDRONs and the DENDRONs themselves may contain different branching and/or linking groups, if present. However, the DENDRON groups are preferably the same.

The subscript m is zero or one or more. Thus, when m is zero, there are no DENDRON¹ groups present. As with DENDRON, the DENDRON¹ groups are independently chosen. Thus, if there are two or more DENDRON¹ groups present, these can be the same or different, although they are preferably the same. Preferred values of m are zero and from 1 to 6, more preferably zero.

The DENDRON groups can vary considerably. They comprise a number of branching groups and optionally linking groups. Moving outwards along the DENDRON group and away from the CORE, each branching group adds another generation to the dendrimer. The dendrimers used in the invention can be first generation or higher, for example first to tenth generation.

Particularly suitable DENDRON groups include dendrons such as those described in WO-A-02/66552. The DENDRON groups are at least partially conjugated, and are preferably completely conjugated.

DENDRON can be bound to the CORE via either a branching or a linking group, although it is preferably bonded via a single bond from the CORE which terminates in the first branching group ofDENDRON.

The branching groups form a bond to the previous generation ofDENDRON (to a branching or linking group from the previous generation) or to the CORE. The branching groups also bond to at least two other groups. These at least two groups may be linking groups, branching groups or groups which terminate the DENDRON.

The branching groups can be any groups known in the field which are capable of forming a branching point within the DENDRON. Particularly suitable branching groups being include aryl and heteroaryl groups and nitrogen atoms.

When a branching group is an aryl group, suitable groups include phenyl, naphthalene, anthracene and, where appropriate, substituted variations. When a branching group is a heteroaryl group, suitable groups include pyridine, carbazolyl, triazole, triazine and, where appropriate, substituted variations.

The branching groups are unsubstituted or substituted. Suitable substituents include those listed below as solubilising groups, and also those listed below as cross-linkable groups. Preferably the branching groups are not substituted.

The linking groups are chosen because they are able to form divalent moieties which are capable of bonding to two groups. The groups to which they are bonded include other linking groups, branching groups and groups which terminate the DENDRON. For example, linking groups may link any combination of branching groups, the CORE, other linking groups, and the groups which terminate the DENDRON. Linking groups may also be substituted with other small groups, but preferably are unsubstituted.

Typical linking groups include aryl and heteroaryl groups, alkyleneoxy groups, vinyl and acetylenyl groups. However, if the branching group is a nitrogen atom then the linking groups are not vinyl or acetylenyl groups.

When a linking group is an aryl group, suitable groups include C_6-^ aryl groups such as phenyl, naphthalenyl, anthracenyl, fluorenyl and, where appropriate, substituted variations. Preferably, when a linking group is an aryl group it is a phenyl or fluorenyl group. When the linking group is a phenyl ring, it is preferably coupled at ring positions 1 and 4. When the linking group is a fluorenyl ring, it is preferably coupled at ring positions 2 and 7. When a linking group is a heteroaryl group, suitable groups include pyridine, oxadiazole, thiophene and, where appropriate, substituted variations. Preferred heteroaryl linking groups include thiophene and pyridine.

The linking groups are unsubstituted or substituted. Suitable substituents include those listed below as surface groups, and also those listed below as cross-linkable groups. More than one of the groups described above as linking groups can couple together to form larger linking groups. For example, a phenyl ring and a further phenyl ring can couple to form a biphenyl group which can itself be a linking group.

The groups which terminate DENDRON can be any of the groups listed previously as branching or linking groups. These groups can then be further substituted, for example by the groups described below as surface groups. Preferred groups which terminate DENDRON include aryl and heteroaryl groups.

As an example of a suitable DENDRON structure, this may be an at least partially conjugated dendritic molecular structure comprising at least one branching group and optionally at least one linking group, the branching groups being selected from aryl and heteroaryl groups and nitrogen atoms, and the linking groups being selected from aryl, heteroaryl, vinyl and acetylenyl groups, said at least one branching group being bonded to three or more groups, and said at least one linking group being bonded to two groups, said dendritic molecular structure terminating at its distal points in aryl and/or heteroaryl groups.

As noted previously, DENDRON may also comprise surface groups. Generally the substituents are chosen such that the dendrimers have increased solubility in the solvent in which they will be processed. Such groups are therefore sometime called "solubilising groups". When the solubilising groups are attached to the distal aryl or heteroaryl groups or the dendrons they are termed "surface groups". Preferred surface groups are those which are capable of improving the solubility of the claimed dendrimers in solvents suitable for solution processing. Accordingly, suitable surface groups include those which result in the dendrimers having increased solubility in solvents such as tetrahydrofuran, toluene, chloroform, chlorobenzene, xylenes and alcoholic solvents such as methanol.

The surface groups are capable of changing the electronic properties of the groups to which they are attached. The groups preferably impart good solubility to the dendrimers and may also contain moieties that allow patterning. The attachment position and number of the surface groups attached to DENDRON is dependent on their structure and well known to those skilled in the art of organic chemistry. Suitable surface groups include those disclosed in PCT/GB02/00750, to which reference should be made for further details. Suitable surface groups therefore include hydroxy, Ci_-]5 alkyl, C_2-is alkenyl, amino, Ci_i₅ alkylamino, di(Ci.i₅)alkylamino, -COOR wherein R is hydrogen or Ci.₁₅ alkyl, C]_-15 alkoxy, C_2-i₅ alkenyloxy, C_6-10 aryloxy, -O₂SR wherein R is C_1-I5 alkyl or C_2-15 alkenyl, SiR₃ wherein each R is the same or different and represents hydrogen, C_1-15 alkyl or C_2-15 alkenyl, C_1-15 alkylthio, C_2-I5 alkenylthio, C_6-10 arylthio, C₆-_M aryl and 5- to 10-membered heteroaryl, wherein the groups C_6-10 aryl and 5- to 10-membered heteroaryl, when present, are substituted with from one to five substituents which are themselves unsubstituted and are selected from Ci_-1S alkyl, C_2-I5 alkenyl, C_1-15 alkoxy and C_2-15 alkenyloxy.

Different surface groups may be present on different dendrons or different distal groups of a dendron. The surface groups can also be chosen such that the dendrimer can be patterned. For example, a crosslinkable group can be chosen, which can be crosslinked upon irradiation or by chemical reaction. Alternatively, the surface groups can comprise protecting groups that can be removed to leave crosslinkable groups. Accordingly, the dendrimers of the invention may also comprise one or more readable groups which can be reacted in order to cross-link. Suitable cross-linking groups include oxetanes.

Preferred DENDRON¹ groups include PANAM-type dendrons, Frechet-type dendrons. There is no real restriction on the nature of the DENDRON¹ groups, other than that they should not be the same as the DENDRON groups. The DENDRON¹ groups will, as with DENDRON groups, comprise branching and linking groups, but these will be chosen such that the DENDRON¹ groups are preferably non-conjugated.

DENDRON¹ can in general be any type of dendron other than those defined by DENDRON. For example, each DENDRON¹, which may be the same or different if m is greater than one, can be a PANAM-type dendron, or can be a Frechet-type dendron as shown for example in Chem. Rev., 2001, 101, 3819-3867. It can be an at least partially conjugated or a non-conjugated dendron. For at least partially conjugated dendrons for example, the links between the branching points in DENDRON¹ may be non-conjugated yet the DENDRON¹ group may still contain chromophores, e.g. in the form of conjugated branching groups. Furthermore, DENDRON¹ can be substituted as for DENDRON. As with DENDRON, DENDRON¹ can provide at least one branching group and optionally linking groups. It can be of any generation, although it is preferably a first to sixth generation dendrimer. In one embodiment, DENDRON¹ is an at least partially conjugated dendritic molecular structure comprising at least one branching group and optionally at least one linking group, the branching groups being selected from aryl and heteroaryl groups and nitrogen atoms, and the linking groups being selected from aryl, heteroaryl, alkyleneoxy, vinyl and acetylenyl groups, said at least one branching group being bonded to three or more groups, and said at least one linking group being bonded to two groups, said dendritic molecular structure terminating at its distal points in aryl and/or heteroaryl groups. As with the DENDRON groups, the DENDRON¹ groups can be substituted by the surface groups defined earlier.

The dendrimer may in general be sensitive to any band of wavelengths which it is desired to detect. This includes photons of wavelengths in the visible spectrum as will be the case for consumer cameras, but more generally includes wavelengths outside the visible spectrum. The response may in general be of any width depending on what it is desired to detect.

To provide a monochrome image sensor (for example the first image sensor having a layer

17 of photo-sensitive material or the second or third image sensors in the case that the separate pieces

, 21 or 23 of material are the same material), the dendrimer may in general have a response for a specific wavelength range or be chosen to have sensitivity to a broad range of visible light, for example substantially the entire visible spectrum. To provide a colour image sensor (for example the second or third image sensors in the case that the separate pieces 21 or 23 of material comprise different dendrimers), the dendrimers will be selected to have responses sensitive to light of different wavelengths. . The design of chromophores that absorb light of different wavelengths is well known to those skilled in the art. For example, a dendrimer that has a porphyrin at its core has a large absorption at 431 nm whilst a dendrimer containing a distyrylanthracene chromophore has an absorption maximum at 413 nm as discussed in J.N.G. Pillow, M. Halim, J.M. Lupton, P.L. Burn, LD. W. Samuel, Macromolecules, 1999, 32, 5985.

The dendrimers used in the invention can be synthesized in a convergent or divergent route, but a convergent route is preferred. With particular regard to organometallic dendrimers, the dendrons are attached to the appropriate ligands and these are subsequently attached to the metal cation to form the dendritic metal complex. Optionally other non-dendritic ligands can subsequently be attached to said complex. Alternatively a ligand with a suitably reactive functional group can be complexed to the metal ion, and then reacted with appropriately functionalised dendrons. In this latter method, not all ligands have to have the reactive functional groups, and thus this method allows the attachment of dendrons to some but not all of the ligands complexed to the metal.

The photo-sensitive material of the layer 17 or the pieces 21 or 23 may include a single photo-sensitive dendrimer or plural photo-sensitive dendrimers. The photo-sensitive material of the layer 17 or the pieces 21 or 23 typically comprises the dendrimer mixed with one or more further substances. For example, the further substance or substances may be any one or combination of a small molecule, another dendrimer, or a polymer. The further substance may include one which enhances charge separation of electrons and holes generated in the dendrimer, for example a soluble form of C60. The further substance may include one which increases charge transport, for example a hole or electron transporting material. The further substance may include one which improves the processing properties of the photo-sensitive material.

The photo-sensitive material could in principle be a pure dendrimer. The use of dendrimers potentially allows the design of a suitable material. However, it is expected that in practice a further substance will be used to enhance the charge separation.

Although in the above image sensors, the photodiodes 3 are formed merely by a region 20 of the layer 17 of photo-sensitive material or by the pieces 21 or 23 of photo-sensitive material, the photodiodes 3 could have a more complicated structure including the photo-senstive material.

For example, the photodiodes 3 could include further layers comprising one or more further substances. The further substance may include one which enhances charge separation of electrons and holes generated in the dendrimer. There may be respective additional layers adjacent the first electrode 14 and the second electrode 19 with the further substance of the additional layers including one which enhances the transfer of photo-current to the respective one of the first electrode 14 and the second electrode 19. For example, a layer comprising l,3,5-tris(2-N-phenylbenzimidazolyl)benzene could be used to enhance electron transport to an electrode and/or a layer comprising 4,4'-bis(N-3- methylρhenyl-N-phenyl)biρhenyl could be used to enhance hole transport to an electrode. Where the photodiode 3 has such a layered structure, the individual layers typically have thicknesses in the range from IOnm tolOOOnm, more preferably in the range from 50nm to 200nm.

As described above, the photodiodes 3 are typically operated under reverse bias applied between the first electrode 14 and the second electrode 19. Alternatively the photodetector could be a phototransistor or a material whose conductivity is photodependent.

For a full colour image sensor, the dendrimers may be selected to be sensitive to light of red, green and blue wavelengths. In this case, the different dendrimers may be arranged in groups of adjacent pixels as shown in Fig. 11 in which the pixels 30 are shown as boxes and the letters R, G and B indicate the red, green and blue sensitivity of the pixels. To improve the representation of colour additional dendrimers sensitive to other colours may be included, for example, the dendrimers may be selected to be sensitive to light of red, green, blue and yellow wavelengths. In this case, the different dendrimers may be arranged in groups of adjacent pixels as shown in Fig. 12 in which the letters R, G, B and Y indicate the red, green, blue and yellow sensitivity of the pixels.

Where dendrimers are selected to have responses sensitive to light of different wavelengths, the image signal output from the image sensor may be processed in the same manner as a conventional CMOS image sensor, using a de-mosaicing algorithm and/or a transformation matrix which are implemented by a microprocessor which processes the output image signal from the image sensor.

The de-mosaicing algorithm compensates for the fact that each pixel is only sensitive to light in one part of the spectrum. Information from neighbouring pixels with one of the two different spectral responses is therefore used to estimate the response of each pixel in their part of the spectrum, by interpolation. The result is an image signal containing either the measured or estimated response of each pixel in each of the three spectral regions. This image contains information concerning the pixel response in three spectral bands that are specific to the particular filters that have been used.

The transformation matrix creates an image that can be displayed on a display device, usually in a standard format such as standard RGB, sRGB. The transformation matrix is derived based on the responses of the dendrimers used to minimize the errors introduced when transforming a wide range of colours from a camera specific RGB image into sRGB. This is done using the same techniques as for a conventional CMOS image sensor by imaging test samples and processing the output image signals. This transformation using a transformation matrix means that image sensors using different combinations of dendrimers will give a similar response to each other and to conventional CMOS sensors.

In recent years, it has been suggested that the accuracy with which the colour in a scene can be rendered is improved using an image sensor responsive to four different bands of light, that is red, green, blue and yellow. However, this may be unnecessary if the dendrimers are chosen to be similar to the responses of the long, medium and short cones (red, green and blue) in the human visual system. The use of dendrimers opens up the possibility of designing dendrimers with appropriate responses.

With a logarithmic detection circuit as shown in Fig. 5, to transform between spectral responses using a transformation matrix it is necessary first to convert the logarithmic response to the corresponding linear response. To avoid the costs of performing these transformations and the errors that will be introduced it is better to avoid the need to transform the spectral response of the camera into any other response. This means that it is more important to have pixels with the correct spectral response in a logarithmic camera than it is with a linear detection circuit.

Using a logarithmic detection circuit with a specific spectral response is particularly important when developing systems that use colour information to automatically identify objects or regions within a scene. The problem with using colour information in these applications is that the spectrum of light arising from an object depends upon both the reflectivity of the object at different wavelengths and the spectrum of the illumination source. Humans are able to compensate for the variations in the illuminant so that the appearance of an object is almost independent of the illuminant. It has shown using a mathematical model that this "colour constancy" can be achieved with digital cameras that combine logarithmic pixels and three or four colour filters with a narrow spectral response.

The original mathematical model assumed that the illuminant could be represented by a black body at a particular temperature and that the spectral responses could be modeled with an infinitely narrow peak. Further work has confirmed that both these conditions can be relaxed significantly if sensors with four or more spectral responses are used. Some work suggests that the optimum peak responses for these sensors are near the values of 450nm, 540nm, 620nm and 690am. The four signals from these sensors can then be processed to create a coordinate in a two-dimensional space that is characteristic of the colour of the object that is almost independent of the illuminant. The spread of positions for each colour at different illuminations depends upon the width of the response peaks. Simulation results suggest that ideally the peak in the spectral response should be narrower than 40nm. Based on this work, it will be seen that there may be advantage in selecting four or more dendrimers with responses which are narrow, eg no more than 40nm FWHM. Again the use of dendrimers opens up the possibility of designing dendrimers with appropriate responses.

Claims

1. An image sensor comprising: an array of photodetectors, each photodetector comprising a photo-sensitive material including a photo-sensitive dendrimer arranged to receive incident light; in respect of each photodetector, a first electrode and a second electrode disposed on opposite sides of the respective photodetector to receive a photo-current from the photodetector, the first electrodes of the photodetectors of the array being separate from one another; and a MOS circuit layer extending across all the photodetectors on the same side of the photodetectors as the first electrodes and having formed therein a plurality of detection circuits connected to the first electrodes of respective photodetectors, the detection circuits including MOS devices and being arranged to detect the photo-current received by the respective first electrodes from the photodetectors.

2. An image sensor according to claim 1, wherein each photodetector comprises a separate piece of photo-sensitive material.

3. An image sensor according to claim 2, wherein the separate pieces of photo-sensitive material are each arranged in apertures in an overlayer of dielectric material.

4. An image sensor according to claim 3, wherein the image sensor comprises an insulation layer of dielectric material disposed over the MOS circuit layer, the first electrodes of the photodetectors being disposed on the insulation layer.

5. An image sensor according to any one of claims 2 to 4, wherein respective photodetectors in the array comprise photo-sensitive dendrimers of a plurality of different types which are photosensitive to light of different wavelengths.

6. An image sensor according to claim 5, wherein the array of photodetectors are arranged in groups of adjacent photodetectors consisting of photodetectors comprising photo-sensitive dendrimers of each of said plurality of different types.

7. An image sensor according to claim 1, wherein the image sensor comprises a layer of photosensitive material extending over all the first electrodes, the photo-sensitive material of the photodetectors being respective regions of the layer of photo-sensitive material.

8. An image sensor according to any one of the preceding claims, wherein the second electrode in respect of each photodetector is formed by a common electrode extending over all the photodetectors.

9. An image sensor according to any one of the preceding claims, wherein the detection circuits include components which are separate for each detection circuit, the components of the detection circuits which are separate for each detection circuit being disposed adjacent the first electrodes of the photodetectors of the array.

10. An image sensor according to claim 9, wherein the detection circuits include further components which are common for plural detection circuits, the common components of the detection circuits being disposed outside the first electrodes of the photodetectors of the array.

11. An image sensor according to any one of the preceding claims, wherein the detection circuits are each arranged to output a voltage which is representative of the photo-current received by the respective first electrode.

12. An image sensor according to claim 11, wherein the detection circuits are each arranged to output a voltage which varies linearly with the photo-current received by the respective first electrode.

13. An image sensor according to claim 11, wherein the detection circuits are each arranged to output a voltage which varies logarithmically with the photo-current received by the respective first electrode.

14. An image sensor according to any one of the preceding claims, wherein MOS circuit layer is a CMOS circuit layer, the detection circuits including both PMOS devices and NMOS devices.

15. An image sensor according to any one of the preceding claims, wherein the photodetectors are photodiodes.

16. An image sensor according to any one of the preceding claims, wherein the photo-sensitive material of the photodetectors has a thickness in the range from 50nm to 200nm.

17. An image sensor according to any one of the preceding claims, wherein the photo-sensitive material comprises said photo-sensitive dendrimer mixed with a further substance.

18. An image sensor according to claim 17, wherein the further substance enhances separation of electrons and holes generated in the photo-sensitive dendrimer.

19. An image sensor according to any one of the preceding claims, wherein the photodetector comprises a layer or piece of photo-sensitive material and at least one further layer of a further substance.

20. An image sensor according to claim 19, wherein the further substance enhances separation of electrons and holes generated in the photo-sensitive material.

21. An image sensor according to claim 19, wherein the at least one further layer comprises two layers adjacent the first electrode and the second electrode, respectively, the further substance of the two layers enhances the transfer of photo-current to the first electrode and the second electrode.

22. An image sensor according to any one of the preceding claims, wherein the dendrimer comprises a dendrimer defined by formula (I): [DENDRON']_m-CORE-[DENDRON]_n (I) wherein n is an integer of one or more, m is zero or an integer of one or more, CORE is a central group to which the DENDRON and DENDRON¹ groups are attached, the DENDRON groups, which are the same or different, are at least partially conjugated dendritic molecular structures, and DENDRON¹ groups are dendritic molecular structures.

23. An image sensor according to claim 22, wherein m is zero and n is an integer of two or more.

24. An image sensor according to claim 22 or 23, wherein the DENDRON groups are substituted by surface groups selected from hydroxy, C_M5 alkyl, C_2-I5 alkenyl, amino, C_w5 alkylamino, di(Ci.is)alkylamino, -COOR wherein R is hydrogen or Ci_-I5 alkyl, Ci_-15 alkoxy, C_2-I5 alkenyloxy, Cβ-io aryloxy, -O₂SR wherein R is Cj_-I5 alkyl or C_2-I5 alkenyl, SiR₃ wherein each R is the same or different and represents hydrogen, Ci._!5 alkyl or C_2-I5 alkenyl, Ci_-I5 alkylthio, C_2-I5 alkenylthio, Ce_-I0 arylthio, C₆.i₄ aryl and 5- to 10-membered heteroaryl, wherein the groups C₆_io aryl and 5- to 10-rαembered heteroaryl, when present, are substituted with from one to five substituents which are themselves unsubstituted and are selected from Ci_-J5 alkyl, C_2-I5 alkenyl, Ci_-I5 alkoxy and C_2-I5 alkenyloxy.

25. An image sensor according to any one of claims 22 to 24, wherein at least one DENDRON, is an at least partially conjugated dendritic molecular structure comprising at least one branching group and optionally at least one linking group, the branching groups being selected from aryl and heteroaryl groups and nitrogen atoms, and the linking groups being selected from aryl, heteroaryl, vinyl and acetylenyl groups, said at least one branching group being bonded to three or more groups, and said at least one linking group being bonded to two groups, said dendritic molecular structure terminating at its distal points in aryl and/or heteroaryl groups

26. An image sensor according to any one of claims 22 to 25, wherein the DENDRON¹ groups are polyamidoamine (PAMAM) dendrons or Frechet dendrons

27. An image sensor according to any one of the preceding claims, wherein the photo-sensitive material is solution processable.

28. An image sensor according to claim 27, wherein the photo-sensitive material is capable of being photo-patterning.

29. An image sensor according to any one of claims 1 to 27, wherein the photo-sensitive material is capable of deposition by thermal transfer.

30. A dendrimer as claimed in any one of the preceding claims, wherein the photo-sensitive material is photo-sensitive in the solid state.

31. A method of manufacturing an image sensor comprising: forming a MOS circuit layer having formed therein a plurality of detection circuits including MOS devices, each detection circuit being arranged to detect a photo-current generated by a photodetector comprising a photo-sensitive material including a dendrimer; forming an array of separate, first electrodes over the MOS circuit layer, each connected to a respective detection circuit; forming respective photodetectors comprising a photo-sensitive material including a dendrimer on the respective first electrodes; and in respect of each photodetector, forming a second electrode on the photodetector.