DETAILED DESCRIPTION OF THE INVENTION

[0028] In the following detailed description, reference is made to the accompanying drawings, which form a part hereof, and in which is shown by way, of illustration of specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention, and it is to be understood that other embodiments may be utilized, and that structural, logical and electrical changes may be made without departing from the spirit and scope of the present invention.

[0029] The terms “wafer” and “substrate” are to be understood as including silicon, silicon-on-insulator (SOI), or silicon-on-sapphire (SOS) technology, doped and undoped semiconductors, epitaxial layers of silicon supported by a base semiconductor foundation, and other semiconductor structures. Furthermore, when reference is made to a “wafer” or “substrate” in the following description, previous process steps may have been utilized to form regions or junctions in the base semiconductor structure or foundation. In addition, the semiconductor need not be silicon-based, but could be based on silicon-germanium, germanium, or gallium-arsenide.

[0030] The term “pixel” refers to a picture element unit cell containing a photosensor and transistors for converting electromagnetic radiation to an electrical signal. For purposes of illustration, a representative pixel is illustrated in the figures and description herein, and typically fabrication of all pixels in an image sensor will proceed simultaneously in a similar fashion.

[0031] Applicants propose several trench isolation techniques to isolate areas of semiconductor devices and in an exemplary embodiment to minimize dark current and suppress leakage current in CMOS image sensors, as described below with reference to FIGS. 2 - 15 . To better illustrate these techniques, a brief description of an exemplary CMOS image sensor pixel is described first with reference to FIGS. 1A and 1B hereinbelow. However, it should be noted that the invention is not limited to CMOS image sensors and may be used in any suitable device, for example, a DRAM, flash memory, SRAM, microprocessor, DSP or ASIC.

[0032] Referring now to FIGS. 1A and 1B , a semiconductor wafer fragment of an exemplary CMOS image sensor four-transistor (4T) pixel, generally designated by reference numeral 10 , is shown. It should be noted that while FIGS. 1 A- 1 B show the use of a transfer gate 50 and associated transistor, the transfer gate 50 provides advantages, but is not required. Thus, the invention may be used in any CMOS imager including, for example, a three transistor (3T) environment where the transfer gate is omitted and an n-type charge collection region of a photodiode is connected with an n-type diffusion region 21 . The CMOS image sensor 10 generally comprises a charge collection region 21 for collecting charges generated by light incident on the pixel and transfer gate 50 for transferring photoelectric charges from the collection region 21 to a sensing node, typically a floating diffusion region 25 . The floating diffusion region is electrically connected to the gate of an output source follower transistor. The pixel also includes a reset transistor 40 for resetting the sensing node to a predetermined voltage before sensing a signal, a source follower transistor 60 which receives at its gate an electrical signal from the sensing node 25 , and a row select transistor 80 for outputting a signal from the source follower transistor 60 to an output terminal in response to an address signal.

[0033] The exemplary CMOS image sensor uses a pinned photodiode as the charge collection region 21 . The pinned photodiode is termed such since the potential in the photodiode is pinned to a constant value when the photodiode is fully depleted. The pinned photodiode has a photosensitive or p-n-p junction region comprising a p-type surface layer 24 and an n-type photodiode region 26 within a p-type active layer 20 . The pinned photodiode includes two p-type regions 20 , 24 so that the n-type photodiode region is fully depleted at a pinning voltage. Impurity doped source/drain regions 22 , preferably having n-type conductivity, are provided about the transistor gates 40 , 60 , 80 . The floating diffusion region 25 adjacent to transfer gate 50 is also preferable n-type.

[0034] In a typical CMOS image sensor, trench isolation regions 28 formed in the active layer 20 are used to isolate the pixels. FIGS. 1A and 1B illustrate typical STI isolation trenches 28 . The trench isolation regions 28 are formed using a typical STI process and are generally formed by etching a trench in the doped active layer or substrate 20 via a directional etching process, such as Reactive Ion Etching (RIE), or etching with a preferential anisotropic etchant used to etch into the doped active layer 20 to a sufficient depth, generally about 1000 to 3000 Angstroms.

[0035] The trenches are then filled with an insulating material, for example, silicon dioxide, silicon nitride, ON (oxide-nitride), NO (nitride-oxide), or ONO (oxide-nitride-oxide).

[0036] The gate stacks for the pixel transistors are formed before or after the trench is etched. The order of these preliminary process steps may be varied as is required or convenient for a particular process flow, for example, if a known photogate sensor (not shown) which overlaps the transfer gate is desired, the gate stacks must be formed before the photogate, but if a non-overlapping photogate is desired, the gate stacks may be formed after photogate formation.

[0037] A translucent or transparent insulating layer 30 is formed over the CMOS image sensor. Conventional processing methods are then carried out to form, for example, contacts 32 (shown in FIG. 1A ) in the insulating layer 30 to provide an electrical connection to the source/drain regions 22 , the floating diffusion region 25 , and other wiring to connect gate lines and other connections in the sensor 10 . For example, the entire surface may then be covered with a passivation layer of e.g., silicon dioxide, BSG, PSG, or BPSG, which is planarized and etched to provide contact holes, which are then metallized to provide contacts to the photogate (if used), reset gate, and transfer gate.

[0038] In CMOS image sensors depicted in FIGS. 1A and 1B , electrons are generated by light incident externally and stored in the n-type photodiode region 26 . These charges are transferred to the diffusion node 25 by the gate structure 50 of the transfer transistor. The source follower transistor produces an output signal from the transferred charges. A maximum output signal is proportional to the number of electrons extracted from the n-type photodiode region 26 . The maximum output signal increases with increased electron capacitance or acceptability of the photodiode. The electron capacity of pinned photodiodes typically depends on doping levels and the dopants implanted to form regions 24 , 26 , 20 .

[0039] A problem associated with the shallow trench isolation technique is photon diffusion under the shallow trench isolation structure from one pixel to an adjacent pixel. Attempts have been made to enhance isolation by implanting ions beneath the shallow trench isolation structure. However, these implants result in high current leakage. The invention provides a novel technique for improved isolation between adjacent pixels that does not require additional implants beneath the trench, thereby minimizing the generation of dark current in the CMOS image sensor.

[0040] Another consideration in CMOS image sensor fabrication are isolation design rules are constructed to make sure that there is enough margin to prevent punch-through in CMOS circuits. For example, the trench 28 separates the source/drain regions 22 of one pixel from the active layer of an adjacent pixel. Accordingly shallow trenches are generally sufficiently wide to allow a margin adequate enough to prevent punch-through or current leakage. The invention further provides novel techniques for preventing current leakage while allowing tighter design rules in CMOS circuits.

[0041] A first embodiment according to the invention is now described with reference to FIGS. 2 - 6 . Applicant proposes an STI process, which uses an isolation trench filled with a doped conductive material containing silicon. Shallow trench isolation regions for CMOS image sensors generally have a depth of less than about 3000 Angstroms and generally around about 2000 to about 2500 Angstroms. Typically, shallow trench regions are filled with a conventional insulator, such as oxides or high density plasma (HDP) oxides. However, it is difficult to fill trenches having a depth greater than 2500 Angstroms with conventional insulators due to the limited spacing within the trench, for example, undesirable voids or air gaps are formed when oxides are used to fill trenches having a depth greater than about 2500 Angstroms. In accordance with the first embodiment of the invention, Applicants propose filling trenches with conductive materials containing silicon, preferably polysilicon or silicon-germanium. Conductive materials containing silicon may be easily deposited into trenches of various depths, unlike conventional insulation materials, e.g., silicon dioxide, silicon nitride, NO, ON, HDP, and ONO, which are difficult to fill in deep trenches. Thus, using a conductive material containing silicon to fill the trench 328 will allow easy formation of a trench, particularly, a deep trench having a depth greater than about 2000 Angstroms, and preferably about 4000 to about 5000 Angstroms.

[0042] Generally, the deeper the trench the better the isolation. With respect to CMOS image sensors in particular, the deeper the trench the higher the electron storage capacitance of the CMOS image sensor. A trench according to the invention is deeper than a shallow trench, and accordingly has longer sidewalls than a shallow trench. Therefore, the longer sidewalls allow for a larger electrical connection region 323 along the sidewalls of the trench such that electron storage capacitance, e.g., hole accumulation, in the electrical connection region 323 is increased in accordance with the invention.

[0043] In a CMOS image sensor having a trench filled with a conductive material containing silicon in accordance with the present invention, as shown in FIG. 2, a trench 328 is etched into a doped active layer 320 . A resist and mask are applied, and photolithographic techniques are used to define the area to be etched-out. A directional etching process, such as Reactive Ion Etching (RIE), or etching with a preferential anisotropic etchant is used to etch into the doped active layer 320 to form the trench 328 . The resist and mask are removed leaving a structure that appears as shown in FIG. 2 .

[0044] Referring now to FIG. 3 , an oxide, i.e., SiO ₂ or other dielectric liner 327 is grown within the trench 328 . The oxide liner may be formed of NO, ON, or ONO among many other suitable materials. The dielectric liner 327 may be substantially conformal. In other words, the thickness of the liner 327 is substantially the same along the sidewalls 319 and at the bottom of the trench 328 . In general, the thickness of the dielectric liner 327 along the sidewalls should be at least about 100 Angstroms.

[0045] Referring now to FIG. 4, a highly doped (in-situ doped) n-type or p-type conductive material containing silicon 329 is deposited to fill the trench 328 . Suitable conductive materials containing silicon include polysilicon and silicon-germanium. Alternatively, as shown in FIG. 5 , the trench 328 may be filled with a conductive material containing silicon 329 then, a masked ion implant (indicated by arrows) may be performed to dope the conductive material containing silicon. For example, in the case of a p-type active layer 320 , with p-type wells, p-type ions such as boron (B) can be implanted into the conductive material containing silicon using a photoresist mask 326 . Similarly, in the case of an n-type active layer 320 with n-type wells, n-type ions such as phosphorous (P), arsenic (As), or antimony (Sb) can be implanted.

[0046] Conductive materials containing silicon are easily filled into deep trenches. The deeper the trench, the harder it is to fill the trench with conventional insulators. Oxides and other conventional insulators form voids or air gaps when used to fill deep trenches. However, in accordance with the invention, a trench may be filled with a conductive material containing silicon easily and effectively.

[0047] An exemplary CMOS image sensor in accordance with the invention and having a pinned photodiode 321 is shown in FIG. 6 . The pinned photodiode 321 has a p-type surface layer 324 and an n-type photodiode region 326 within a p-type active layer 320 . A junction is formed around the entirety of the n-type region 326 . An impurity doped floating diffusion region 325 , preferably having n-type conductivity, is provided on one side of the channel region of transfer gate 350 , the other side of which has a portion of n-type region 326 . A trench isolation region 328 is formed in the active layer 320 adjacent to but spaced from the n-type region 321 . An electrical connection region 323 for providing hole accumulation is formed adjacent the sidewalls of the trench isolation region 328 . The trench isolation region 328 is formed as described above with respect to FIGS. 2 - 5 .

[0048] The gate stacks, for example the transfer gate 350 , may be formed before or after the trench is etched. The order of these process steps may be varied as is required or convenient for a particular process flow, for example, if a photogate sensor which overlaps the transfer gate is desired, the gate stacks must be formed before the photogate, but if a non-overlapping photogate is desired, the gate stacks may be formed after photogate formation.

[0049] A translucent or transparent insulating layer 330 is formed over the CMOS image sensor 300 . Conventional processing methods are then carried out to form for example, contacts (not shown) in the insulating layer 330 to provide an electrical connection to the source/drain regions 322 , the floating diffusion region 325 , and other wiring to connect gate lines and other connections in the sensor 300 . For example, the entire surface may then be covered with a passivation layer, of e.g., silicon dioxide, BSG, PSG, or BPSG, which is CMP planarized and etched to provide contact holes, which are then metallized to provide contacts to the photogate (if used), reset gate, and transfer gate.

[0050] The use of a trench in accordance with the invention provides improved isolation between pixels. The deeper trench better inhibits electrons from diffusing under the isolation trench to an adjacent pixel thereby preventing cross-talk between neighboring pixels. Accordingly, by enhancing isolation via a deeper trench, additional implants under the trench are not necessary, therefore by reducing the implants needed for isolation, current leakage is also reduced. Another advantage of the invention, is that the use of a deep trench filled with a conductive material containing silicon in accordance with the invention provides a deeper hole accumulation region, thereby increasing electron storage capacity. Also the deeper trench allows for tighter isolation design rules. Deeper trenches may also be narrower than shallow trenches, while still providing effective isolation between neighboring regions. Accordingly, the source/drain regions of one pixel may be brought closer to the active layer of an adjacent pixel, by narrowing the width of the deep trench.

[0051] A second embodiment in accordance with the invention is now described with reference to FIGS. 7 - 13 . Referring now to FIG. 7, a trench 428 is etched into an active layer 420 . The trench is preferably a deep trench having a depth greater than about 2500 Angstroms and preferably between about 4000 to about 5000 Angstroms. A resist and mask are applied, and photolithographic techniques are used to define the area to be etched-out. A directional etching process, such as RIE, or etching with a preferential anisotropic etchant is used to etch into the doped active layer 420 to form the trench 428 . The resist and mask are removed leaving the FIG. 7 structure.

[0052] Referring now to FIG. 8, a nitride liner 432 is formed in the trench 428 via Chemical Vapor Deposition (CVD). This nitride liner 432 may be formed of any suitable nitride including NO, ON, ONO, and is preferably formed of silicon nitride.

[0053] Referring now to FIG. 9 , an oxide, e.g. SiO ₂ or other dielectric liner 427 is formed within the trench 428 and over the silicon nitride liner 432 . The liner 427 may be non-conformal, in that its thickness may vary along the trench sidewalls 429 . A relatively thick liner can be formed near the bottom of the trench and a thinner liner formed near the top of the trench. Non-conforming materials such as the well-known PSG, BPSG, SOG can be used to produce the liner 427 .

[0054] Referring now to FIG. 10, a bottom portion of the oxide liner 427 and nitride liner 432 is stripped away. This can be accomplished by an anisotropic dry etch or a masked wet or dry etch.

[0055] Referring now to FIG. 11, a selective epitaxial layer 433 is grown to fill the trench 428 with silicon. The epitaxial layer 433 may be grown using any suitable technique and may be grown as a single layer or multi-layer. The epitaxial layer 443 is grown in directly on a surface of the active layer 420 so as to provides a direct electrical contact to the doped active layer 420 through the trench while providing improved field isolation between pixels. Providing a direct electrical contact to the active layer in accordance with the invention, eliminates the need for a top contact, therefore saving space and allowing for tighter pixel formation.

[0056] Referring now to FIG. 12 , in accordance with yet another embodiment of the invention, the selective epitaxial layer 433 is grown to partially fill the trench 428 with silicon.

[0057] Referring now to FIG. 13, a deposition process is performed to fill the rest of the trench with a filler material 434 . The filler material is preferably an oxide material and is more preferably an HDP oxide. Alternatively, a conductive material containing silicon, for example polysilicon or silicon-germanium, may also be used to fill the rest of the trench 428 .

[0058] By providing an epitaxial layer 433 , the amount of oxide needed to fill the trench is reduced. Accordingly by using a reduced amount of oxide, or not using oxide in situations where conductive material containing silicon is used to fill the rest of the trench or when the trench is filled with the epitaxial layer 433 (as shown in FIG. 11 ), a deep trench in accordance with the invention may be formed. As discussed above deep trenches provide improved isolation and in the case of CMOS image sensors, prevention of cross-talk between adjacent pixels. And also as discussed above with regard to the first embodiment, the use of a deep trench to provide improved isolation eliminates the need to use excess implants beneath the trench, thereby reducing dark current in CMOS image sensors caused by current leakage. A selective-EPI filled or partially filled trench according to the invention may be used in combination with other aspects of the invention, for example, the selective-EPI-partially filled trench may be used along with a deep trench filled with a conductive material containing silicon.

[0059] An exemplary CMOS image sensor in accordance with the invention and having a pinned photodiode 421 is shown in FIG. 14 . The pinned photodiode 421 has a p-type surface layer 424 and an n-type photodiode region 426 within a p-type active layer 420 . A junction is formed around the entirety of the n-type region 426 . An impurity doped floating diffusion region 425 , preferably having n-type conductivity, is provided on one side of the channel region of transfer gate 450 , the other side of which has a portion of n-type region 426 . A trench isolation region 428 is formed in the active layer 420 adjacent to but spaced from n-type region 421 . An electrical connection region 423 for providing hole accumulation is formed adjacent the sidewalls of the trench isolation region 428 . The trench isolation region 428 is formed as described above with respect to FIGS. 7 - 13 .

[0060] The gate stacks, for example transfer gate 450 , may be formed before or after the trench is etched. The order of these preliminary process steps may be varied as is required or convenient for a particular process flow, for example, if a photogate sensor which overlaps the transfer gate is desired, the gate stacks must be formed before the photogate, but if a non-overlapping photogate is desired, the gate stacks may be formed after photogate formation.

[0061] A translucent or transparent insulating layer 430 is formed over the CMOS image sensor 400 . Conventional processing methods are then carried out to form for example, contacts (not shown) in the insulating layer 430 to provide an electrical connection to the source/drain regions, the floating diffusion region 425 , and other wiring to connect gate lines and other connections in the sensor 400 . For example, the entire surface may then be covered with a passivation layer, of e.g., silicon dioxide, BSG, PSG, or BPSG, which is CMP planarized and etched to provide contact holes, which are then metallized to provide contacts to the photogate (if used), reset gate, and transfer gate.

[0062] Pixel arrays according to the invention, and described with reference to FIGS. 2 - 14 , may be further processed as known in the art to arrive at CMOS image sensors having the functions and features of those discussed with reference to FIGS. 2 - 14 .

[0063] A typical processor based system, which includes a CMOS image sensor according to the invention is illustrated generally at 642 in FIG. 15 . A processor based system is exemplary of a system having digital circuits, which could include CMOS image sensors. Without being limiting, such a system could include a computer system, camera system, scanner, machine vision, vehicle navigation, video phone, surveillance system, auto focus system, star tracker system, motion detection system, image stabilization system and data compression system for high-definition television, all of which can utilize the present invention.

[0064] A processor based system, such as a computer system, for example generally comprises a central processing unit (CPU) 644 , for example, a microprocessor, that communicates with an input/output (I/O) device 646 over a bus 652 . The CMOS image sensor 642 also communicates with the system over bus 652 . The computer system 600 also includes random access memory (RAM) 648 , and, in the case of a computer system may include peripheral devices such as a flash memory card 654 , or a compact disk (CD) ROM drive 656 which also communicate with CPU 644 over the bus 652 . It may also be desirable to integrate the processor 654 , CMOS image sensor 642 and memory 648 on a single IC chip.

[0065] The above description and drawings are only to be considered illustrative of exemplary embodiments, which achieve the features and advantages of the invention. Modification and substitutions to specific process conditions and structures can be made without departing from the spirit and scope of the invention. Accordingly, the invention is not to be considered as being limited by the foregoing description and drawings, but is only limited by the scope of the appended claims.