FR2616990A1 - Charge-transfer matrix device tolerant to high vertical-transfer frequencies, and use of such a device in an image pick-up - Google Patents

Abstract

The charge-transfer matrix device according to the invention includes, in addition to the horizontal lines, made of polysilicon for example, at the ends of which are applied the transfer control potentials, one or more sets of additional vertical metal electrodes arranged above certain judiciously chosen vertical gates, connection points being made between these vertical electrodes and the horizontal lines raised to the same potentials, in order to reduce the constant distributed over the lines and lower the time for establishing the control potentials; this makes it possible to increase the vertical-transfer frequency. Application, in particular, to image pick-ups with frame transfer, for lowering the trailing effect.

Description

Matrix device t charge transfer
tolerant of high vertical transfer frequencies
and use of such a device in an image sensor.

 The invention relates to matrix devices with charge transfer, and in particular to photosensors for image sensors.

 One of the problems encountered in charge transfer photosensors by frames is that during the transfer of charges from the photosensitive zone to the memory zone, the different lines are shifted while the incident illumination continues to generate information in the photosensitive zone. . In con # sequence we obtain a mixture of information proportional to the vertical transfer speed.

By way of example for a frame transfer structure, the parasitic illumination can be highlighted in the following way: an image consisting of N lines lit on a dark background results in the rendering by a trail of constant level above and below the lighted lines. This effect, known as the dragging effect, or "smesring" in Anglo-Saxon technical literature, is very annoying both in video and in detection: target tracking, tracking.

 To reduce the dragging effect, an effective solution consists in placing a shutter in front of the photosensitive zone which eliminates any lighting during the transfer, but with all the constraints of space and opacity that that implies another solution consists in carrying out the electrodes by the photolithographic technique with high resolution, of the order of a micrometer. The latter solution does not however solve the problem of matrices having a very large number of points, and requiring a large surface. Furthermore, reducing the size of the elementary photosensitive point using “fine” technology also has the disadvantage of reducing the level of saturation, and therefore the dynamics of the photosensor. This effect can also be reduced by increased vertical transfer frequency.

 Simultaneously, an increase in the vertical transfer frequency makes it possible to increase the bit rate of information transmitted, and this whatever the type of charge transfer matrix device, as soon as the transfer is carried out in a vertical charge transfer register. .

 However, this increase is today limited by the time constants distributed along the lines, in particular for devices with a large number of points in line, and therefore of large surface area: an excessively high frequency results in a modification of the timing of the signals. along the lines, which results in a degradation of the performance of the photosensor device.

 The subject of the invention is a matrix device with charge transfer which solves this problem by a particular arrangement of additional metal electrodes which reduces the distributed time constants. It is then possible to increase the vertical transfer frequency without disturbing the control signals, and to decrease accordingly when there is the drag effect, without reducing the size of the elementary point.

According to the invention a matrix device for charge transfer tolerating high vertical transfer frequencies, comprising a substrate, a network of vertical grids
and a network of horizontal line electrodes, the ends of which are brought to the control potentials, is characterized in that, the horizontal line electrodes having a non-negligible resistance, at least one set of vertical metal electrodes is arranged on the surface matrix, these electrodes being brought to the same control potentials as the horizontal lines of the connection points being made at the crossing points of these vertical electrodes and line electrodes brought to the same potentials.

 The invention also relates to the use of such a device in a frame transfer image sensor, in particular with a view to reducing the dragging effect.

 The invention will be better understood and other characteristics will appear from the following description with reference to the appended figures.

Figure 1 is a diagram of a photosensor with charge transfer by frame transfer, known
FIG. 2 is the diagram of a frame transfer photosensor according to the invention in which the time constant distributed over a frame is divided by 4
FIG. 3 is the diagram of a frame transfer photosensor according to the invention in which the time constant distributed over a line is divided by N
FIG. 4 is a more detailed partial diagram showing the horizontal electrodes making the interconnection with the sets of additional electrodes, vertical
Figure 5 is a partial detailed plan view of the frame transfer photosensor according to the invention.

 Figures 6a and 6b show a cross-sectional view along a line of photosensitive dots, and the corresponding potential diagram, of a photosensor according to the invention.

 As indicated above, increasing the frequency of vertical transfer, in a matrix device with charge transfer, is of great interest, in order to increase the bit rate of information transmitted, in particular in transfer structures. of frame but also in all the other matrix structures where the transport of loads is carried out by registers with transfer of charge (DTC or "CCD" technical English): interline transfer, CSD (for "Charge Sweep Currency"), TDI (for "Time Deiay Integration"); direct read frame transfer.

 For frame transfer image sensors from the photosensitive area to a memory area, increasing the vertical transfer frequency also makes it possible, as indicated above, to reduce the effect of vertical drag in the image.

 A frame transfer photosensor of known type, like that shown in FIG. 1, comprises a photosensitive zone 10, a memory zone 20 and a read charge transfer register 30, the charges of which are shifted horizontally towards a video output amplifier. 40. The control potentials are applied to the different points of the matrix by transparent, generally polysilicon line electrodes, brought to the control potentials.

In a transfer system with 4 control phases ~ 2P ~ 3P 'Q) 3P' - the line electrodes are successively connected to these 4 phases. To reduce the time constant distributed along the line formed by each control electrode and the ground plane, the ad hoc potentials are applied simultaneously to the two ends of the line electrodes, for each line. Thus the maximum distributed time constant is equal to half of the time constant which would correspond to the application of the potential to only one end of a line of the matrix.

 The same stucco re is used for the vertical offset control in the memory area 20, the horizontal electrodes or lines being successively connected to the potentials of four phases of control of the memory area ~ 1MS ~ 2M # 3Mand ~ 4M by electrodes arranged on either side of the matrix.

In a charge transfer device in which it is desired to substantially increase the vertical transfer frequency, the distributed constant is still too large. Indeed, this constant is proportional to the square resistance of the line, of the order of 30or2 for a transparent poiysillclum electrode
FIG. 2 schematically represents a first embodiment of the invention, applied to a frame transfer photosensor to a frame memory area. In this embodiment, the matrix has the same structure: photosensitive area 10, memory area 20 and output read register 30 with output amplifier 40, but the control potentials are also applied to the lines by a set J, of additional vertical metal electrodes E ~ l, E # 2, E #, E # 4> located in the central area of the matrix.

These electrodes are deposited on non "useful" zones of the photosensitive zone, for example on the drains of anti-glare transistors located in this central zone or on "walls1, of isolation between points. Thus in addition to the electrodes associated with the lines on either side of the matrix, these vertical connections also bring the control potentials approximately to the midpoints of the horizontal lines of the photosensitive zone. Connection points are provided between the vertical electrodes and the horizontal line electrodes to make the necessary connections, as will be explained in more detail below. Thus the distributed time constant is divided by 4 compared to that which would correspond to the total length of the line.

FIG. 3 schematically represents a second embodiment of the invention in which the control potentials are applied at 5 points on the horizontal lines, in addition to the two ends of the lines, by 5 sets J1 J5 of vertical metal electrodes E ~ 11 ... E # 41, ..., E # 15 ... E # 45, interconnected and connected to these control potentials outside the photosensitive zone
As in the embodiment described with reference to FIG. 2, in the photosensitive zone the sets of vertical metal connections are preferably deposited above non-photosensitive zones, for example above drains of anti-glare transistors for the dies with anti-glare, or above isolation walls for dies without anti-glare structure. Thus these sets of additional electrodes do not affect the sensitivity of the device. In the memory area, as in non-photosensitive charge transfer matrix devices, such as memories, the location of the sets of additional metal electrodes can be arbitrary. In the example shown, they have been arranged in the same way as in the photosensitive zone for the sake of simplification, especially during the manufacturing process. So the distributed constant is divided by 12, i.e. by N = 2 xn + 2, if n is the number of additional sets of vertical connections, the same arrangement is provided in the memory area
Figure 4, detail of Figure 3, is a diagram showing the horizontal interconnections of the sets of additional vertical connections to be connected to the pads to which the different control phases are applied. This structure being applicable to the photosensitive zone as well as to the memory zone, the indices P and M of these control phases have not been specified in the diagram in FIG. 4. In this figure we distinguish the 5 sets of 4 metallic connections vertical, in aluminum for example, connected outside the transfer area to the four control phases #i '~ 2) ~ 3, and In fact, the strl. > matrix matrix of polysilicon line electrodes and columns extends outside the strictly useful zones, photosensitive zone and memory zone, in particular in the image zone in an optically masked zone which constitutes the darkness reference. structure, the lines are used to interconnect the electrode sets in the following manner: as in the useful area (photosensitive area or memory area), contacts are made between the additional lines outside the useful area, ~ 1 , ~ 2, ~ 3 'and the corresponding vertical metal electrodes, for example at the points noted P1, P2, P3, P4. In addition, horizontal metal electrodes E '~ l, E' E 'and ET ~ 4 represented by bands; are also deposited above the horizontal lines represented by lines, and interrupted only at intersections with the vertical networks. At each interruption, contact is made between the horizontal metal electrode E '~ and the underlying line, for example at the points A, B, C, D for line ~ 2.

In the central area of the matrix, these horizontal metal electrodes can also be interrupted because these electrodes join the pads for applying the control potentials on both sides of the matrix. The dotted lines represent those sections of horizontal metal electrodes which can be deleted. In the memory area, these horizontal interconnections of the sets of vertical metal electrodes can be located anywhere because of the total optical masking of this region.

The access of the vertical networks to the ends of the matrix is balanced, resistances of equal access, to preserve a correct timing of control between the various stages of the same column
FIG. 5 is an even more detailed diagram of an area comprising these horizontal interconnection lines of sets of additional vertical electrodes which shows that, since the consecutive lines partially overlap it is not possible to produce electrodes metal bars for interconnecting sets of vertical electrodes above consecutive lines. Consequently, the horizontal metal electrodes E '# for two consecutive phases are located above lines spaced 4 lines apart.

 This figure shows the successive horizontal lines connected to # 1, # 2, # 3, ~ 4, C # g ~~~ 1 the vertical isolation grids Gg, and vertical electrodes, drain D and grids G1, G2 forming devices anti-glare. As stated above, in such a structure. the vertical metal electrodes are arranged above the drains of the anti-glare trammstors. They are hatched in FIG. 5, like the horizontal metal electrodes intended for the interconnection of the sets of vertical electrodes. The points P1, P2, A, indicated in this figure are the points analogous to the points Pl, P2, A marked on Figure 4, even if they are not arranged in exactly the same way because the details shown in Figures 4 and 5 are oriented differently because they respectively detail an area above the photosensitive area of the figure 3 and an area under the memory area. As shown in Figure 5, all the lines connected to the same potential, either ~ 1 or # 2 in the figure have contacts with the vertical metal electrodes, which gives the contact points marked P'1 and P'2 on the figure. The photosensitive zone and the memory zone have the same structure as that detailed in FIG. 5 with the exception of the horizontal metal electrodes.

This structure is shown in section along a line of the photonsensitive zone in FIG. 6a, and the corresponding potential diagram V is represented in FIG. 6b. In the P-type substrate 1 are formed the drains of the transistors
Anti-glare MOS, by N diffusion. Then after depositing an oxide layer, the different vertical grids are formed, at a level noted level "O". To simplify the figure, the oxide separating the different levels has not been shown.

These gates are on the one hand the isolation gates Gg, on the other hand the gates G1 and G2 of the MOS transistors. After depositing another oxide layer, the horizontal line electrodes at level "1", ie those connected to ~ l and are formed, for example of polysilicon. Then after depositing another layer it oxidizes the horizontal line electrodes at level "2", ie those connected to ~ 2 and ~ 4 for example. By way of example, it has been assumed that the section in FIG. 6a was made along a line connected to
Finally, the vertical metal electrodes are deposited above the anti-glare structures while providing in the oxide deposited on the line electrodes of the contact zones. It has been assumed in the figure that the vertical electrodes represented were electrodes E ~ 1 and E ~ 2 to be connected to ~ 1 and # 2. We therefore notice on the section the line contact ~ 1 vertical metal electrode ~ 1, that is to say a P1 type contact point. This arrangement (deposit above the drains) was chosen because the MOS transistors separate two points of a line, ie i-1 and i on the one hand, and 1 + 1 and 1 + 2 on the other hand in FIG. 5 of the isolation grids separating points 1-2 and 1-1, i and 1 + 1 etc ...; the electrodes therefore do not encroach on the photosensitive zone. 2 consecutive electrodes, E1 and
E1 # 2 are separated by two columns. However, the invention is not limited to this arrangement. In particular, the consecutive electrodes of a set of vertical electrodes could be separated by a single column, the electrodes being able to be placed above the isolation grids Go A l 'Conversely, the consecutive electrodes of a set of vertical electrodes could be more distant, and separated by a number of columns greater than 2.

 FIG. 6b represents the diagram of the corresponding potentials and the charges discharged via the drain of a transistor for a point 1-2 at saturation, the other points being for example unlit (no charges).

 As stated above, the invention is not. limited to frame transfer image sensors. It applies to all charge transfer structures in which it is desired to increase the vertical transfer frequency.

The use of such a matrix device is particularly advantageous in a frame transfer image sensor, in order to reduce the dragging effect.
Indeed, a measure of this effect can be as follows if: Vs is the signal level of the lit lines
VR the stray signal level due to the drag effect,
N the number of lines lit, the rest being unlit (case of a spot on a dark background),
Tv the passage time under an illuminated line during vertical transfer
Ti the integration time, then S the drag coefficient or "smearing" can be estimated by S = TV / Ti and VR = NS V5. For Ti = 20 ms and TV = llls, i.e. FV (vertical transfer frequency) = 1MHz, then S = 5 x 10-5 With FV = 10 MHz, S is no more than 5 x 106 and with FV = 20 MHz, S is no longer equal to 2.5 x 106.

 The drag effect is therefore considerably reduced.

Claims (10)

 1. Matrix charge transfer device tolerating high vertical transfer frequencies, comprising a substrate, a network of vertical grids (G0, G1, G2) and a network of electrodes of horizontal lines whose ends are brought to the control potentials # 2. ~ 3, ~ 4), characterized in that, the horizontal line electrodes having a non-negligible resistance, at least one set of vertical metal electrodes (En1, E # 2, En3, E ~ 4) is arranged on the surface matrix, these electrodes being brought to the same control potentials as the horizontal lines, connection points being made at the crossing points of these vertical electrodes and line electrodes brought to the same potentials.  2 Device according to claim 1, characterized in that, for a photosensitive matrix, the vertical metal electrodes are arranged on areas which do not participate in the sensitivity of the device.  3. Device according to re-endication 2, characterized in that, anti-glare transistors being provided in the matrix, the vertical metal electrodes are arranged above drains of these transistors.  4. Device according to claim 2, characterized in that the vertical metal electrodes are arranged above vertical grids (Go) isolating the elementary photosites and delimiting the columns of the matrix.  5. Device according to one of the preceding claims, characterized in that these electrodes are made of aluminum.  6. Device according to 11 one of claims 1 to 3, characterized in that the matrix structure extending above or below the zone of the matrix useful for the transfer of charges, a set of horizontal metal electrodes is also provided extending above the electrodes of corresponding horizontal lines, interrupted at the intersections with the sets of vertical electrodes, for the interconnection of vertical metal electrodes respectively connected to the same potentials, connection points being provided with the corresponding lines on the and other of these sets of electrodes.  7. Device according to claim 5, characterized in that the resistances of access to the ends of the electrodes of horizontal lines and to the points of contact of the same line via the sets of vertical metal electrodes are balanced  8. Use of a matrix device according to any one of the preceding claims in a frame transfer image sensor comprising a matrix photosensitive zone and a matrix memory zone.  9. Use according to claim 8, characterized in that, in the photosensitive zone, the interconnections of the sets of vertical metal electrodes are made under an optical masking zone of the matrix constituting the darkness reference.  10. Use according to one of claims 8 and 9, with a high transfer frequency to reduce the effect of vertical drag.