Method for forming image sensorTechnical Field
The present invention relates to the field of semiconductors, and more particularly, to a method for forming an image sensor.
Background
An image sensor is a semiconductor device that can convert an optical image into an electrical signal and capture the image accordingly. Image sensors may be classified into two types, a Charge Coupled Device (CCD) image sensor and a Complementary Metal Oxide Semiconductor (CMOS) image sensor, according to the difference of image sensor elements. Compared with a CCD image sensor, the CMOS image sensor can be manufactured by using the existing semiconductor equipment, and does not need additional equipment investment. And the quality of the image capturing system can be improved along with the improvement of semiconductor technology, and the requirements of users on the continuous improvement of the quality, such as more flexible image capturing, higher sensitivity, wider dynamic range, higher resolution, lower power consumption and more excellent system integration, can be better met. For the above reasons, the demand growth rate of the current CMOS image sensor has reached seven times the demand growth rate of the CCD sensor.
CMOS image sensors may be classified into Front Side Illumination (FSI) image sensors and Back Side Illumination (BSI) image sensors. In the front-illuminated image sensor, the pixel circuit is located between the light receiving side and the photodiode. Light enters the CMOS image sensor from the light receiving side and passes through the pixel circuit before reaching the photodiode, and the pixel circuit blocks part of the light to reduce the light reaching the photodiode, so that the fill factor (the ratio of the photodiode area to the pixel area) of the CMOS image sensor is reduced.
The back-illuminated image sensor allows light to enter from the back side of the chip, pass through the substrate, and then pass to the photodiode without passing through the pixel circuit; alternatively, the photodiode is formed directly on the back surface of the substrate, and light directly enters the photodiode. The back-illuminated image sensor overcomes the disadvantage of a small fill factor of front-illuminated image sensors.
A method of forming a back-illuminated image sensor, which is commonly used in the art, is as follows:
referring to fig. 1, a first wafer 1 is provided, an n-doped layer 2 and a p-doped layer 3 are formed on the first wafer 1, the n-doped layer 2 and the p-doped layer 3 constituting a photodiode in an image sensor.
Referring to fig. 2, a second wafer 4 is bonded on the p-type doped layer 3, and the first wafer 1 is thinned by polishing or etching.
Referring to fig. 3, the first wafer 1 is turned over so that the back surface of the first wafer 1 faces upward, and a patterned mask layer 5 is formed on the back surface of the first wafer 1, where the patterned mask layer 5 defines the position of a diffusion plug (diffusion plug). And carrying out n-type impurity doping on the first wafer 1 by taking the patterned mask layer 5 as a mask to form a diffusion column 6, wherein the diffusion column 6 is connected with the n-type doping layer 2.
Then, referring to fig. 4, the patterned mask layer 5 is removed, and a pixel circuit is formed on the first wafer 1. The transfer transistor 7 and the source follower transistor 8 in the pixel circuit are schematically shown in fig. 4. Wherein the photodiode is connected to the drain of the transfer transistor 7 (the drain of the transfer transistor 7 in fig. 4, i.e., the diffusion column 6) through the diffusion column 6.
Next, a dielectric layer 9 and an interconnection structure (not numbered) located in the dielectric layer 9 are formed on the first wafer 1, and the interconnection structure connects the transistors in the pixel circuit. The interconnect structure connecting the pass transistor 7 and the source follower transistor 8 is schematically shown in fig. 4.
Referring to fig. 5, a third wafer 10 is bonded on the dielectric layer 9, the first wafer 1 is turned over so that the third wafer 10 is at the bottom, and then the second wafer 4 is etched or polished away.
And finally, forming a filter 11 on the p-type doped layer 3.
The process for forming the back-illuminated image sensor is very complex, and at least three wafers are needed to form the back-illuminated image sensor, so that the process cost is too high; and the resulting back-illuminated image sensor also has a large dark current.
Disclosure of Invention
The invention solves the problems that the process for forming the image sensor is complex, the cost is high and larger dark current exists in the prior art.
To solve the above problems, the present invention provides a method of forming an image sensor, comprising: providing a wafer; forming each transistor of the pixel circuit on the front surface of the wafer; forming a dielectric layer on the wafer and the transistors and an interconnection structure in the dielectric layer, wherein the transistors are connected through the interconnection structure to form a pixel circuit; forming a bearing structure on the dielectric layer after forming the pixel circuit; after the bearing structure is formed, turning over the wafer to enable the back of the wafer to face upwards; forming a graphical mask layer on the back of the wafer, wherein the graphical mask layer defines the position of a diffusion column; doping the wafer by taking the patterned mask layer as a mask, and forming a diffusion column in the wafer, wherein the diffusion column is connected with the pixel circuit; removing the patterned mask layer; and after removing the patterned mask layer, forming a photodiode on the back of the wafer.
Optionally, the bearing structure is a wafer, a polycrystalline silicon wafer or a glass sheet.
Optionally, after the photodiode is formed, the method further includes: forming a filter on the photodiode.
Optionally, after the optical filter is formed, the method further includes: forming an isolation structure between two adjacent pixels, wherein the bottom of the isolation structure is positioned at the bottom of the photodiode; or the bottom of the isolation structure is positioned in the wafer.
Optionally, the method for forming the isolation structure includes: forming a graphical mask layer on the optical filter, wherein the graphical mask layer defines the position of the isolation structure; etching by taking the patterned mask layer as a mask to form an isolation trench; removing the patterned mask layer; filling an isolation material in the isolation groove, wherein the isolation material is level to the upper surface of the optical filter; or, depositing an isolation material in the isolation trench and on the optical filter, wherein the isolation material is higher than the isolation trench.
Optionally, after forming the isolation trench and before filling the isolation material, the method further includes: and forming a p-type doped region on the side wall and the bottom of the isolation trench.
Optionally, the method for forming the p-type doped region on the sidewall and the bottom of the isolation trench includes: carrying out p-type doping on the side wall and the bottom of the isolation groove to form a p-type doped region; or depositing or epitaxially growing a p-type thin film layer on the side wall and the bottom of the isolation trench to serve as a p-type doped region.
Optionally, the thickness of the p-type doped region is
Optionally, the impurity doped in the p-type doped region is B or BF2。
Optionally, the isolation material is one or more of silicon oxide, silicon oxynitride, silicon nitride, or a polymer.
Optionally, the isolation material is flush with the upper surface of the optical filter, and after the isolation structure is formed, the method further includes: and forming a protective layer on the optical filter.
Optionally, the material of the protective layer is one or more of silicon oxide, silicon oxynitride, silicon nitride, or polymer.
Optionally, after the wafer is turned over, before a patterned mask layer is formed on the back side of the wafer, the method further includes: and thinning the wafer.
Optionally, the method for forming the photodiode includes: forming a first thin film layer with n-type doping on the back of the wafer by using an epitaxial growth method or a deposition method; and forming a second thin film layer with p-type doping on the first thin film layer by using an epitaxial growth method or a deposition method, wherein the first thin film layer and the second thin film layer form the photodiode.
Optionally, the impurity doped in the diffusion column is an n-type impurity.
Optionally, the pixel circuit comprises a reset transistor, a transfer transistor, a row-gating transistor and a source follower transistor.
Optionally, the front surface of the wafer is provided with a p-type thin film layer, and the reset transistor, the transmission transistor, the row gating transistor and the source following transistor are located on the p-type thin film layer.
Compared with the prior art, the technical scheme of the invention has the following advantages:
according to the technical scheme, the pixel circuit is formed on the wafer, and the formation of the photodiode is carried out after the pixel circuit is formed. This process adjustment has at least the following advantages:
first, a process of forming the image sensor is simplified. In the prior art, a photodiode is formed first, then a pixel circuit is formed, and when the pixel circuit is formed, a wafer needs to be turned over; after the pixel circuit is formed, the wafer needs to be turned over again so that the photodiode is located at the top. According to the technical scheme, the manufacture of the image sensor can be completed only by turning the wafer once, so that the process is simple.
Secondly, the use of wafers is saved. In the prior art, the image sensor can be manufactured only by turning over the wafer twice, and one wafer is required to be used as a bearing structure at the bottom during each turning over, so that three wafers are required to be used for forming the image sensor. According to the technical scheme, the wafer is turned over only once, and at most, only two wafers are used. Compared with the prior art, the technical scheme saves the use of the wafer, thereby reducing the process cost.
Furthermore, the technical scheme forms an isolation structure in the image sensor, so that dark current in the image sensor can be reduced.
Furthermore, p-type doped regions are formed on the side wall and the bottom of the isolation groove, so that the photodiode is separated from the isolation structure, and the dark current of the image sensor can be effectively reduced.
Drawings
Fig. 1 to 5 are schematic cross-sectional views illustrating a method of forming an image sensor according to the related art;
fig. 6 to 13 are schematic cross-sectional views illustrating a method of forming an image sensor according to a first embodiment of the present invention;
FIG. 14 is a schematic cross-sectional view showing a method of forming an image sensor according to a second embodiment of the present invention;
fig. 15 is a schematic cross-sectional structure diagram of a method of forming an image sensor in a third embodiment of the present invention.
Detailed Description
In the prior art, the photodiode is formed first, and then the pixel circuit is formed, and the first wafer 1 needs to be turned over twice to complete the manufacture of the image sensor, so that the process is complex. Secondly, one wafer is used as a carrying structure for turning over the first wafer 1 each time, so at least three wafers are needed for manufacturing the image sensor, and the process cost is high.
In practice, it has been found that in the prior art, since no isolation structure is formed between the pixels of the image sensor, light incident on a pixel may pass through the pixel and enter an adjacent pixel, and light between adjacent pixels may interfere with each other. The absence of isolation structures between pixels of an image sensor also causes large dark current.
Therefore, the invention provides a method for forming an image sensor, which can effectively solve the problems of complex process, high cost and larger dark current in the prior art for forming the image sensor.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below.
First embodiment
The embodiment provides a method for forming an image sensor, which comprises the following steps:
referring to fig. 6, a wafer 110 is provided.
In an embodiment, the front surface of the wafer 110 includes a p-doped layer 111.
The wafer 110 may be made of conventional semiconductor materials such as silicon, silicon germanium, Silicon On Insulator (SOI).
Referring to fig. 7, transistors of a pixel circuit are formed on the front surface of the wafer 110; forming a dielectric layer 130 on the wafer 110 and the transistor, and forming an interconnection structure 101 in the dielectric layer 130; the individual transistors are connected by the interconnect structure 101 to form a pixel circuit.
In a specific embodiment, the transistors are formed on the p-doped layer 111.
Only the transfer transistor 121 and the source follower transistor 122, and the connection relationship between the transfer transistor 121 and the source follower transistor 122 are schematically shown in fig. 7. The source of the pass transistor 121 is connected to the gate of the source follower transistor 122.
In the existing image sensor, the pixel circuit includes three, four, or five transistors. Any pixel circuit capable of reading and transferring charge from a photodiode can be used with the present solution. The most common of these is a combination of four transistors, respectively a reset transistor, a pass transistor, a row strobe transistor and a source follower transistor.
The method for forming the pixel circuit is a common CMOS (complementary metal oxide semiconductor) process in the prior art, which is a well-known technology and will not be described herein again.
Referring to fig. 8, a carrier structure 112 is formed on the dielectric layer 130; after the carrying structure 112 is formed, the wafer is turned over so that the back side of the wafer faces upward.
In an embodiment, after turning over the wafer, the method further includes: and thinning the wafer. The thinning method can be etching or polishing. Referring to fig. 8, in an embodiment, after thinning the wafer, only p-type doped layer 111 remains.
In an embodiment, the supporting structure 112 is a wafer, a polysilicon wafer, or a glass sheet. Since the poly-silicon chip and the glass chip are generally cheaper than the wafer, if the supporting structure 112 is the poly-silicon chip or the glass chip, the usage of the wafer can be reduced, and the process cost can be reduced.
When the support structure 112 is a wafer, the wafer may be bonded to the dielectric layer 130 by direct oxidation bonding or by an adhesive.
Referring to fig. 9, a patterned mask layer 140 is formed on the back surface of the wafer, and since only the p-type doped layer 111 remains after the wafer is thinned, accordingly, the patterned mask layer 140 is directly formed on the back surface of the p-type doped layer 111. The patterned mask layer 140 defines the locations of the diffusion pillars.
Then, the p-type doping layer 111 is doped by using the patterned mask layer 140 as a mask, and a diffusion column 150 is formed in the p-type doping layer 111, wherein the diffusion column 150 is connected to the pixel circuit.
In a specific embodiment, the diffusion column 150 is connected to the drain of the pass transistor 121.
The impurity doped in the diffusion column 150 is an n-type impurity.
Referring to fig. 10, the patterned mask layer 140 (refer to fig. 9) is removed, and then a photodiode, which is composed of a first thin film layer 161 and a second thin film layer 162, is formed on the p-type doped layer 111. After the photodiode is formed, the method further comprises the following steps: an optical filter 170 is formed on the second thin film layer 162.
In a specific embodiment, the method of forming the photodiode is:
forming a first thin film layer 161 on the p-type doped layer 111 using an epitaxial growth method or a deposition method; n-type in-situ doping the first thin film layer 161 during its formation; alternatively, after the first thin film layer 161 is formed, the first thin film layer 161 is n-doped by an ion implantation method or a thermal diffusion method.
Forming a second thin film layer 162 on the doped first thin film layer 161 using an epitaxial growth method or a deposition method; p-type in situ doping the second thin film layer 162 during its formation; alternatively, after the second thin film layer 162 is formed, the second thin film layer 162 is p-doped using an ion implantation method or a thermal diffusion method.
The doped first thin film layer 161 and the doped second thin film layer 162 constitute a photodiode. In addition, any other type of photodiode may be used with the present teachings.
The filter 170 may be a red filter, a green filter, or a blue filter. The pattern of the filter 170 may be a Bayer pattern (Bayer pattern), a Bayer derivative pattern (Bayer derivative pattern), or any other type of filter pattern known in the art.
In a specific embodiment, after the filter 170 is formed, the method further includes: an isolation structure is formed between two adjacent pixels, and the specific forming method comprises the following steps:
referring to fig. 11, a patterned mask layer 141 is formed on the filter 170, wherein the patterned mask layer 141 defines the location of an isolation structure;
and then, etching is carried out by taking the patterned mask layer 141 as a mask to form an isolation trench 180, wherein the bottom of the isolation trench 180 is positioned at the bottom of the photodiode. The bottom of the photodiode is the lower surface of the first thin film layer 161.
In fig. 11, the bottom of the isolation trench 180 is located at the bottom of the photodiode; in other embodiments, the bottom of the isolation trench 180 is located within the wafer.
Next, referring to fig. 12, the patterned mask layer 141 is removed; and an isolation material is filled in the isolation trench 180, and the isolation material is flush with the upper surface of the optical filter 170, so as to form an isolation structure 181.
In a specific embodiment, the method for depositing the isolation material in the isolation trench 180 includes:
forming an isolation material layer in the isolation trench 180 and on the upper surface of the optical filter 170 by using physical vapor deposition, chemical vapor deposition or atomic layer deposition, wherein the isolation material layer is higher than the isolation trench 180; then, a chemical mechanical polishing method is used to remove a portion of the isolation material layer, so as to expose the upper surface of the optical filter 170, and only the isolation material in the isolation trench 180 remains.
In a specific embodiment, the isolation material is one or more of silicon oxide, silicon oxynitride, silicon nitride, or a polymer. But may be other materials known in the art.
The isolation structure 181 can prevent incident light from entering adjacent pixels, thereby preventing light interference between adjacent pixels; the isolation structure 181 may also reflect light to return the incident light to the photodiode, thereby improving the light capture capability of the photodiode.
In other embodiments, the bottom of the isolation structure 181 is located within the wafer. In this case, the isolation structure 181 may also prevent electrical interference between adjacent pixels. The mutual electrical interference means that when two pixels work, current enters adjacent pixels.
Referring to fig. 13, after forming the isolation structure 181, a protective layer 190 is formed on the optical filter 170. The material of the protection layer 190 is one or more of silicon oxide, silicon oxynitride, silicon nitride or polymer.
The protection layer 190 is used to protect the photodiode from being damaged.
The working principle of the image sensor in the technical scheme is as follows:
referring to fig. 13, incident light passes through the protective layer 190 and enters the optical filter 170, and the optical filter 170 selects light of a specific color to enter the photodiode, which converts the light into electric charges.
The charge is transferred to the transfer transistor 121 through the diffusion column 150, and the charge is stored to the source of the transfer transistor 121 through the transfer transistor 121. Then, the charge transfers the image data taken by the photodiode through the source follower transistor 122.
According to the technical scheme, the pixel circuit is formed firstly, and then the photodiode is formed. The wafer 110 is turned over only once to complete the fabrication of the image sensor, thereby simplifying the process complexity. Secondly, each time the wafer 110 is turned over, one wafer is required to be used as the bottom bearing structure 112, and in the technical scheme, only one wafer 110 needs to be turned over once, and at most two wafers are used. Compared with the prior art, the technical scheme saves the use of the wafer, thereby reducing the process cost.
Second embodiment
The second embodiment differs from the first embodiment in that:
referring to fig. 14, an isolation material is deposited in the isolation trenches and on the optical filter 170, the isolation material being higher than the isolation trenches. And the isolation material higher than the isolation groove is used as a protective layer.
For further information, refer to the first embodiment.
Third embodiment
The third embodiment is different from the first embodiment in that, referring to fig. 15, after forming the isolation trench 180 and before filling the isolation material, the third embodiment further includes: p-type doped regions 102 are formed on the sidewalls and bottom of the isolation trenches 180.
In an embodiment, the method for forming the p-type doped region 102 comprises: and depositing or epitaxially growing a p-type thin film layer on the side wall and the bottom of the isolation trench to serve as a p-type doped region 102.
The presence of the p-type doped region 102 isolates the photodiode from the isolation structure, effectively preventing dark current generation. Dark current is generated by electrons moving from the photodiode to another area when light does not enter the photodiode. Dark current is typically present at the contact of the photodiode with the isolation structure. If the photodiode is in direct contact with the isolation structure, a large dark current may be generated in the photodiode, which may seriously affect the performance of the image sensor and reduce the charge storage capacity of the image sensor.
In a specific embodiment, the p-type doped region 102 has a thickness ofThe impurity doped in the p-type doped region 102 is B or BF2。
In other embodiments, the method for forming the p-type doped region may also be:
the sidewall and the bottom of the isolation trench 180 are doped p-type to form a p-type doped region, and the p-type doping method may be ion implantation or thermal diffusion. The p-type doped region converts the n-doped first thin film layer 161 near the isolation trench 180 into a p-type doped region.
For further information, refer to the first embodiment.
Although the present invention is disclosed above, the present invention is not limited thereto. Various changes and modifications may be effected therein by one skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.