WO1989005039A1

WO1989005039A1 - Blooming control in ccd image sensors

Info

Publication number: WO1989005039A1
Application number: PCT/US1988/003917
Authority: WO
Inventors: David Newell Nichols; Edward Tichenor Nelson; Charles Vincent Stancampiano; Bruce Curtiss Burkey; Win-Chyi Chang
Original assignee: Eastman Kodak Company
Priority date: 1987-11-16
Filing date: 1988-11-04
Publication date: 1989-06-01

Abstract

A CCD image sensor has a plurality of sensing elements. Such sensing elements are disposed over a doped semiconductor substrate of one conductivity type in which a laterally uniform doping of opposite conductivity type is formed near the substrate surface. Each element includes an insulator layer over the semiconductor substrate, and an electrode on the insulator layer which when potential is applied to the electrode creates a well in the bulk of the substrate which collects charge as a result of a photoelectric process. The laterally uniform region of opposite conductivity forms a potential barrier to the substrate over which excess charge is drained away, thus forming an anti-blooming structure. In addition, this structure is scaleable and reduces color crosstalk.

Description

BLOOMING CONTROL IN CCD IMAGE SENSORS Technical Field

This indention relates to CCD image sensors having a blooming control structure. Background ftrt

One organization of an area image sensor is known as the frame transfer organization shouin in Fig. 1. In this arrangement, charge transfer channels run in υertical directions. Separating each transfer channel is a channel stop which confines charge collected to the transfer channels and prevents charge leakage into adjacent channels. Each transfer channel has a plurality of sensing elements or image pixels. Each sensing element is defined by a plurality of closely spaced electrodes. The electrodes can be made from a transparent conductive material such as polysilicon. Potential is applied to at least one of the electrodes of each sensing element and a potential well or depletion region or well is formed under it. Charge is collected in the potential well which is a function of scene brightness. The potential well can either be what is referred to as a surface or a buried channel. In a buried channe] situation, the electrodes are disposed on an insulator such as SiO . The insulator ouerlies the substrate. The substrate is doped and can be of a giυen polarity for example p-type Near the insulator the substrate is a polarity (n-type) opposite to that of the bulk substrate and of such a concentration that when a predetermined potential is applied to the electrode, a potential well or channel is formed within the substrate spaced from the insulator ,

The electrostatic potential as a function of depth under an appropriately biased electrode is shown in Fig. 3a Curue C represents the electrostatic potential for a buried channel with no free, charge. Charge generated by absorption of incident scene light in the buried channel is collected and stored in the channel. This charge is generated by the photoelectric process and the amount of charge collected depends on scene brightness. Curve D represents the electrostatic potential of a buried channel with stored free charge. As charge fills the channel, the electrostatic potential in the buried channeled becomes more negative, moving away from curve C of Fig. 3a. Charge can also be photogenerated in the substrate. - This charge will be stored in the buried channel if it diffuses to the depleted region before being reco bined with majority carriers in the bulk. Charge is prevented from moving under adjacent electrodes by applying a more negative voltage to adjacent electrodes. The electrostatic potential as a function of depth for adjacent electrodes is represented by curve E. One problem with this arrangement is that when the array is illuminated by a scene in which certain regions are considerably brighter than other scene regions, the array portions receiving intense radiation may become overloaded and produce excess charge, which tends to spread out throughout the channel. The. excess free charge, causes the electrostatic potential in the buried channel to become, so negative, that it comes within 0.5 volts of curve E. Then charge can move from beneath the collecting electrode to adjacent electrodes. This spreading of charge will manifest itself as blooming of the image. Blooming causes the increase in size of a bright portion of an image and is a major limitation to the application of CCD i agers for use as wide dynamic range sensors. A second problem with the arrangement of Fig. 3a is crosstalk. Crosstalk occurs when charge photogeπerated in the undepleted substrate diffuses in the. substrate to, and is collected, in an adjacent pixel. In one—chip color imagers, each pixel has its incident scene light filtered so that the pixel absorbs light of a particular range of wavelengths. When one pixel is designed to absorb long wavelength light (such as red) and light to an adjacent pixel is filtered to prevent absorption of long wavelength light, the presence of crosstalk causes photogenerated charge due to long wavelength light to be stored in the adjacent pixel, creating color mixing and an objectionable reproduction of the image. A common approach to blooming control consists of providing an overflow drain in the channel stops for collecting excess photogenerated charge, This lateral overflow drain can consist of a region doped to the polarity opposite from the semiconductor substrate.

Another common approach for frame transfer and interline transfer image sensors is to build the image sensor on a laterally diffused p—well . In this approach, dopants implanted adjacent to the photodiode (for example for an interline device) are laterally diffused under the photodiode to form a potential barrier over which excess charge flows into the substrate. Neither of the two approaches mentioned above are easily scaled; i.e. , when pixel size is reduced, significant changes in the process must occur and/or dimensions of the structure cannot be reduced in a scaled fashion. Also, the. first approach may not significantly reduce crosstalk for small pixel sizes , DISCLOSURE OF THE TNUENTION The object of this invention is to provide a CCD image, sensor which effectively drains off excess charge carriers from a potential well to prevent blooming.

This object is achieved by apparatus for controlling blooming in a CCD image sensor having a plurality of elements with each such element including a semiconductor substrate of one conductivity type, with its near surface region doped to form a laterally uniform region of opposite conductivity type, an insulator layer over the substrate, an electrode on the insulator layer which when potentials are applied to the substrate, to the semiconductor layer with conductivity type opposite to the substrate, and to the electrodes, creates a channel in the. substrate which collects charge produced by a photoelectric process, characterized by: the. near surface region being doped so that excess charge collected in the channel is drained into the substrate and thereby prevent blooming. This structure also suppresses crosstalk.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is schematic overall view of a frame. transfer CCD area image sensor in accordance with the present invention showing in an exploded format elements of several transfer channels;

Fig. 2 is a cross—sectional view taken along the lines 2—2 of Fig. 1 showing a sensor element in a transfer channel of the sensor; and

Figs. 3a and 3b show electrostatic potential curves of the channel structure taken along lines 3a—3a for devices without, and with, respectively, incorporation of the blooming suppression structure. Curve C of Fig. 3a is for a region ready for collecting photogenerated charge and as yet empty of that free charge; curves D and B of Figs. 3a and 3b respectively, are for regions collecting photogenerated charge, with some of the charge already collected; and curves E and of Figs. 3a and 3b, respectively, are for adjacent electrodes corresponding to curves D and B .

MODES OF CARRYING OUT THE INUENTION The overall configuration of a CCD area image sensor 10 which embodies the present invention is shown in Fig. 1. The image sensor 10 comprises a frame transfer four-phase (υoltage lines Φl/4—Φ4/4) buried channel CCD area image sensor. - The image is sensed by sensing elements or pixels in each transfer channe] 4-2. As will be described, each element is defined by a plurality of transparent electrodes .

The elements define a staggered two—dimensional array A which for illustrative purposes only, is shown to have 740 columns and 485 rows of sensing elements. Each transparent electrode is connected to one line or phase of a four—phase clock signal After exposure, the four clock signals on the voltage lines are sequentially applied in the well known manner to the image sensing array A to move the imagewise charge pattern formed in the sensing elements one row at a time to an output register shown as block H.

Between adjacent transfer channels there are provided channel stops 40. These channel stops 40 may be provided by a thick field oxide and also by diffusion or implants . High frequency clock pulses drive the polysilicon electrodes to read out each of the rows of the image sensor at the standard television video rate. The output register H is shown schematically as a block since it can be provided by a conventional four-phase CCD shift register positioned between a transfer gate 30 and a horizontal channe] stop Each cell of the. output register H has four electrodes aligned with a transfer channel 42. These electrodes are. actuated by signals on voltage lines Φl/4 through Φ4/4 in the conventional manner. The transfer gate electrode 30 is actuated by a first transfer signal T. and transfers a row of photocharges to the. output register H.

After a row of photocharges has been transferred to the. output register H, the transfer gate 30 is closed. By being closed is meant that a potential barrier is formed under this transfer electrode. At this time, the output register is operated in a four—phase, manner, clocking the photocharge out to an output diode 32, one pixel at a time, which converts the photocharges into a voltage in the well known manner.

Turning now to Fig. 2, the CCD sensor is shown to be constructed on a silicon semiconductor substrate 34 of one conductivity type n. The near surface region is doped to form a layer 35 with conductivity type p opposite that of the substrate. The term near surface region means that the region is spaced from the surface of the substrate. As will be described, in the preferred embodiment a buried channel region 35a is disposed between the near surface region and the surface of the substrate. The. near surface region is spaced 0.1 to 1.5 micrometers from the. surface and is 3 to 4 micrometers thick. If a surface channel CCD is fabricated with this anti—blooming structure, then the near surface p—region extends to the surface of the semiconductor.

A silicon dioxide (SiO ) insulating layer 36 overlies the substrate semiconductive layer 35. Silicon dioxide has the fundamental property of preventing the diffusion of impurities through it and is an excellent insulator. Polysilicon conductive electrodes 38 are deposited on top of the silicon dioxide layer 36. Metal conductive layers can also be used. The silicon dioxide layer 36 and the electrodes 38 are transparent and permit photons of visible light to pass through them into the bulk substrate 34 and 35. A thin protective layer 37 of silicon dioxide overcoats the electrodes 38. The bulk of the substrate 34 has been doped to be a n—type substrate. Although the substrate is described as n-type and the near surface region is described as p—type and the buried channel is described as n—type, the device can also be fabricated with a p—type substrate, an n-type near surface region and a p-type buried channel,

A suitable n-type dopant is phosphorous,

Layer 35 can be formed by implantation and diffusion of a suitable p—type dopant such as boron into the substrate to form what is called a p-well . In addition, an n—type dopant is diffused into layer 35 to form an n-type layer 35a, called the buried channel, A suitable dopant for this layer is arsenic or phosphorous .

With layers constructed as above, Fig 3b shows the electrostatic potential as a function of depth beneath adjacent polysilicon electrodes (curves

A and B ) where curve B corresponds to an electrode (called here electrode B ) with a more positive applied potential than curve A . In addition, curve B represents the electrostatic potential when the buried channel contains free charge. The free charge is produced by a photoelectric process as a result of exposure of light from a scene. The exact shape of the electrostatic potential depends primarily upon the doping levels of the various layers, the substrate doping, the dielectric constant and the thickness of the insulator 36, the voltages applied to the electrode 38, the p—well 35, and the substrate 34, and the amount of free charge stored in the buried channel. Once, these parameters are known, the electrostatic potential can be obtained by solution of Poisson's equation. A buried channel is discussed here, and is often preferable over a surface channel since it is not subject to charge, transfer inefficiencies caused by charge interacting with surface states. It should be understood that although the present invention is- particularly suitable with buried channel devices, it is also applicable to surface channel CCD area image sensors. In Fig. 3b, the electrode corresponding to curve (here called electrode A ) is held more negative to prevent charge from moving from beneath electrode. B to the region beneath electrode A . Thus potential U must be sufficient (at least 1 volt) to prevent charge from transferring from electrode B to A.. Blooming occurs when excess free charge is generated by the photoelectric process in a buried channel which already contains its maximum free charge. Then this excess charge causes the potential V DH to be reduced, and the excess charge then spills into the adjacent buried channel of electrode . However, in this invention, the doping of the layers and the voltages applied to the layers are such that when the buried channel contains its maximum charge, excess charge flows to the substrate rather than to adjacent buried channels. This occurs when the voltage barrier to the substrate U DO is about 0.5 volts. It is noted that while this example is for a four phase CCD, it has been applied to a two phase CCD and can be applied to a virtual or uniphase CCD. This anti—blooming structure also reduces crosstalk caused by charge generated below one pixel diffusing to and being collected by the buried channel of a nearby pixel Crosstalk is reduced because charge generated by light absorbed deeper than point Φ__D cannot rD diffuse to any buried channel region because the electrostatic potential of the substrate is more positive than the p—well. Thus, photogenerated charge cannot diffuse to the buried channel of the adjacent pixels.

After the scene has been imaged on the sensor, the potential for each sensing element can now be changed by varying the voltage on the lines Φl/4—Φ4/4 so as to shift the charge down into the horizontal readout register H. Thereafter, the charge in register H is sequentially shifted to the diode 32 which converts the charge into a voltage,

The charge transfer channel 40 and the shift register H are standard charge transfer arrangements wherein by properly applying periodic voltages to their electrodes a moving array of potential wells can be formed in which charge s shifted from one position to another. The operation of these types of charge transfer devices is well known in the art and described fully in the textbook "CHARGE TRANSFER DEUICES" by Sequin et al , pages 23-25, which describe the operation of a conventional four—phase CCD electrode structure. EXAMPLE A CCD image sensor with 1320 (H) x 1035 (U) pixels with pixel size of 6.8 μm x 6 8 μ incorporating this laterally uniform, depleted p—well has been made as follows :

N—type silicon wafers with resistivity of 20-25 ohm-cm were used as substrates and were gettered Then, 2000 Λ thermal oxide was grown on the. wafers. The wafers were then- masked with photoresist such that the image area could receive a uniform implant while the outer edges of the. die close to where, the individual die would be cut from one another would receive no implant. Wafers were then implanted with boron

2 at energy of 120 keU and dose of 2.5E+ll/cm .

Wafers were then annealed at 1200°C for 180 min. in nitrogen to diffuse the implant. The implanted and annealed boron forms a p—well in the n—type substrate. Using procedures outlined in U.S. Patent No. 4,613,402,- and ^"normal silicon semiconductor device processing steps, a full frame, two phase CCD image sensor with anti—blooming was fabricated. The blooming suppression was measured by determining the amount of light necessary to cause a small illuminated square to double in length when the output was viewed on a monitor. Light 60 times more intense was than that required for saturating a pixel with charge was necessary to double the size of the square, thus showing good anti—blooming characteristics. In addition, crosstalk du to carriers photogenerated by red light (wavelength 620—680 n ) was reduced by a factor of about 3.7 when comparing this sensor with one without the uniform, flat p-well.

Industrial Applicability and Advantages

A feature of this invention is that an anti—blooming structure can be manufactured with a minimum of additional processing steps.

A further feature is that blooming is prevented without the. loss of sensor image area.

A still further feature, is that blooming is prevented in CCD area image sensors without degradation in M.T.F.

A still further feature is that blooming is prevented with a design which can be scaled.

Claims

We claim :

1 , A CCD image area sensor including a plurality of transfer electrodes arranged in sequence and overlying pixels, means for applying phase—related voltages to said plurality of transfer electrodes to cause charge to be transferred sequentially between potential wells underlying said transfer electrodes, said image sensor having a plurality of sensing elements, each such sensing element including a semiconductor substrate of one conductivity type, with its near—surface region doped to form a laterally uniform semi conductive region of opposite conductivity type, an insulator layer over the substrate, and an electrode on the insulator layer which when potentials are applied to the electrodes, the se iconductive substrate, and the laterally uniform region of opposite conductivity type, a potential well is created in the substrate which collects charge produced by a photoelectric process, characterized by: the laterally uniform, doped semiconductive region forming a potential barrier over which excess charge flows to the substrate and not to adjacent pixels, thus forming an anti-blooming structure, 2, A CCD image area sensor including a plurality of transfer electrodes arranged in sequence and overlying pixels, means for applying phase-related voltages to said plurality of transfer electrodes to cause charge to be transferred sequentially between potential wells underlying said transfer electrodes, said image sensor having a plurality of sensing elements, each such sensing element including a semiconductor substrate of one conductivity type, the near surface of this substrate is doped laterally uniform to form a well of opposite conductivity type, and within the well a further doping of the substrate to form a buried channel of conductivity type the same as the substrate, an insulator layer over the substrate, and an electrode on the insulator layer which when potentials are applied to the electrodes, the. semiconductive substrate, and the laterally uniform region of opposite conductivity type, a potential well is created in the substrate which collects charge produced by a photoelectric process, characterized by: the laterally uniform, doped semiconductive region forming a potential barrier over which excess charge flows to the substrate and not to adjacent pixels, thus forming an anti—blooming structure.