JP3763822B2 - Infrared sensor - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an infrared sensor and a method for manufacturing the same, and more particularly to a pixel structure of an uncooled infrared sensor and a method for manufacturing the same, and to provide a highly sensitive uncooled infrared sensor.
[0002]
[Prior art]
Infrared imaging has the advantage of being able to capture images day and night, and has the advantage of being more permeable to smoke and fog than visible light. In addition, it can obtain temperature information on subjects, so it can be monitored in the defense field. Wide application range as camera and fire detection camera.
[0003]
The biggest drawback of conventional mainstream quantum infrared solid-state imaging devices is that they require a cooling mechanism for low-temperature operation, but in recent years they have not required a cooling mechanism. The development of is becoming popular. In an uncooled type or thermal type infrared solid-state imaging device, an incident infrared ray having a wavelength of about 10 μm is converted into heat by an absorption structure, and the temperature change of the heat-sensitive part caused by the weak heat is electrically converted by some thermoelectric conversion means. Infrared image information is obtained by converting into a signal and reading out the electrical signal.
[0004]
In order to improve the sensitivity of such an uncooled infrared sensor, there are roughly three types of methods.
[0005]
One is a method of improving the ratio of the infrared power incident on the infrared detection unit: dP to the temperature change of the subject: dTs, that is, dP / dTs. This method is mainly to improve the sensitivity by the optical system, such as increasing the diameter of the infrared lens, using an antireflection coating, a low-absorption lens material, improving the infrared absorption rate of the infrared detector, and improving the infrared absorption area. Corresponds to this. Along with the recent increase in the number of non-cooled infrared sensors, the unit pixel size is mainly 40 μm × 40 μm. Among the items described above, it is relatively important to improve the infrared absorption area in the infrared detector. It was left as an issue.
[0006]
However, it has been reported that the infrared absorption area has been increased to 90% of the pixel area by forming an infrared absorption layer on top of the pixel (Tomohiro Ishikawa, et al., Proc. SPIE Vol.3698, p.556). , 1999), it is difficult to obtain a greater sensitivity improvement by optical means.
[0007]
The second is a method of improving the ratio of the incident infrared power: dP and the temperature change of the infrared detector: dTd, that is, dTd / dP, whereas the method described above is an optical method. Is a thermal method. In general, in an uncooled infrared sensor mounted in a vacuum package, heat transfer from the infrared detection unit to the support substrate is currently supported by supporting the infrared detection unit on a hollow structure inside the support substrate. It is dominated by the heat conduction of the structure. Therefore, a leg-like support structure made of a material having a low thermal conductivity is laid out to be thinner and longer as long as it can be designed (for example, Tomohiro Ishikawa, et al., Proc.
SPIE Vol.3698, p.556, 1999).
[0008]
However, as the pixel size is being reduced to about 40μm × 40μm, the silicon LSI process level microfabrication has already been performed, so the support structure layout has been improved to further improve sensitivity. Difficult to do. Similarly, it is difficult to further reduce the thermal conductivity, which is a material characteristic of the support structure. Particularly, the wiring for outputting an electrical signal from the infrared detector has a similar electrical conduction mechanism. Therefore, it is difficult to realize a significant improvement in sensitivity in terms of materials.
[0009]
The third is a means for improving the ratio of the change in temperature of the infrared detector: dTd to the change in electrical signal generated by the thermoelectric conversion means: dS, that is, dS / dTd, which is an electrical method. Unlike the other two methods, it is very important for this method to reduce various electrical noises that occur at the same time, while aiming for simple high sensitivity, that is, improvement in dS / dTd. Until now, various thermoelectric conversion means have been studied.
[0010]
The main ones are as follows.
[0011]
(1) Thermopile that converts temperature difference to potential difference by Seebeck effect
(For example, Toshio Kanno, et al., Proc. SPIE Vol.2269, pp.450-459, 1994),
(2) A bolometer that converts a temperature change into a resistance change according to the change temperature of the resistor.
(For example, A.Wood, Proc. IEDM, pp.175-177, 1993),
(3) Pyroelectric element that converts temperature change into electric charge by the pyroelectric effect
(For example, Charles Hanson, et al., Proc. SPIE Vol.2020, pp.330-339, 1993),
(4) Silicon pn junction that converts a temperature change into a voltage change by a constant forward current
(For example, Tomohiro Ishikawa, et al., Proc. SPIE Vol.3698, p.556, 1999)
However, in the comparison of each method, it cannot be said that there is a method that is decisively superior to other methods in terms of its thermoelectric conversion characteristics, noise characteristics, and manufacturing methods. For example, a bolometer is excellent in terms of temperature resolution, but a silicon pn junction that can be manufactured only in a silicon LSI process is excellent in the manufacturing process.
[0012]
[Problems to be solved by the invention]
As described above, as one of the methods for increasing the sensitivity of the uncooled infrared sensor, the ratio of the incident infrared power: dP and the temperature change of the infrared detector: dTd, that is, dTd / dP is improved. There is a thermal method.
[0013]
In general, heat transport from the infrared detector to the support substrate is dominated by the heat conduction of the support structure that supports the infrared detector on the hollow structure inside the support substrate. The leg-shaped support structure is designed to be thinner and longer as long as it can be designed. However, while the pixel size is being reduced to about 40 μm × 40 μm, the silicon LSI process has already been implemented. It is difficult to achieve a further significant improvement in sensitivity by devising the layout of the support structure because of the level of microfabrication.
[0014]
The present invention has been made on the basis of recognition of such a problem, and the purpose thereof is to significantly reduce the cross-sectional area of the support structure that supports the infrared detection unit compared to the conventional case and suppress heat "escape". An object of the present invention is to provide an infrared sensor with significantly improved detection sensitivity.
[0015]
[Means for Solving the Problems]
According to the present invention, on the semiconductor substrate, a plurality of infrared detection units having an absorption unit that absorbs incident infrared rays and converts the infrared rays into heat, and a thermoelectric conversion unit that converts a temperature change caused by the heat generated in the absorption unit into an electrical signal. Are disposed in a passivation film formed on the semiconductor substrate, and a recess formed by etching away a part of the passivation film and a part of the semiconductor substrate. And at least one support for supporting each of the plurality of infrared detection units, and the upper surface of each of the supports is lower than the upper surface of the passivation film formed on the side wall of the recess, And lower than the upper surface of the infrared detection unit disposed in the recess, each lower surface of the support is higher than the lower surface of the infrared detection unit Infrared sensor is provided.
[0018]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to specific examples.
[0019]
(First embodiment)
FIG. 1 is an overall configuration diagram of an infrared sensor according to a first embodiment of the present invention. An infrared detection pixel 1 for converting incident infrared rays into an electrical signal is two-dimensionally arranged on a semiconductor substrate, and a vertical address circuit and a horizontal address circuit for pixel selection are arranged adjacent to the infrared detection pixel array 2 and selected. An output unit is provided for sequentially outputting signals from the pixels.
[0020]
The infrared detection pixel 1 in FIG. 1 is a forward-biased pn junction, and a constant current source for forward-biasing the pn junction of the pixel is also arranged adjacent to the infrared detection pixel array 2. Here, in FIG. 1, an array of 4 pixels of 2 rows × 2 columns is shown as the infrared detection pixel array 2.
[0021]
In the infrared detection pixel row selected by the vertical address circuit, the forward bias current supplied from the constant current source flows through the current path of the vertical signal line 3, the selected pixel, and the horizontal address line 4 and is generated in the vertical signal line 3. The signal voltages thus selected are sequentially selected and output by the horizontal address circuit.
[0022]
In FIG. 1, as a simplest example, a structure in which a signal voltage generated on the vertical signal line 3 is directly output via a column selection transistor 5 sequentially selected by a horizontal address circuit is shown. Since the voltage is weak, a structure for amplifying the signal voltage in units of columns may be provided as necessary.
[0023]
FIG. 2 is an equivalent circuit diagram of the infrared detection pixel 1 of FIG. In order to increase sensitivity, n pn junctions are connected in series, and an additional resistor Ra is present in series with the pn junction.
[0024]
Additional resistance: Ra is pixel internal wiring resistance between the pn junction and the horizontal address line 4 and between the pn junction and the vertical signal line 5: Rl, contact resistance between this wiring and the pn junction: Rc, and p region of the pn junction And n region resistance: Rs.
[0025]
3A is a plan configuration diagram of the infrared detection pixel shown in FIG. 2, and FIG. 3B is a cross-sectional view taken along the line AA ′. The infrared detection pixel includes an infrared absorption layer 18, a pn junction inside the SOI layer 8 formed for thermoelectric conversion on the hollow structure 7 formed inside the single crystal silicon support substrate 6, and the SOI layer 8. A sensor unit 10 composed of a buried silicon oxide film layer 9 supporting the sensor unit 10, a support unit 11 for supporting the sensor unit 10 on the hollow structure 7 and outputting an electrical signal from the sensor unit 10, The sensor unit 10 includes a connection unit (not shown) that connects the vertical signal line 4 and the horizontal address line 5.
[0026]
By providing the sensor unit 10 and the support unit 11 on the hollow structure 7, the temperature of the sensor unit 10 is efficiently modulated by incident infrared rays. FIG. 3 shows a structure in the case of n = 2.
[0027]
And in this invention, it forms so that thickness T1 of the support part 11 may become thinner than thickness T2 of the sensor part 10. FIG. In particular, in the case of this example, the thickness T1 is made smaller than T2 by forming the lower surface 11B of the support portion 11 at a position higher than the lower surface 10B of the sensor portion 10 when viewed from the substrate 6. . In this way, by reducing the thickness T1 of the support portion, it is possible to suppress the “escape” of heat and greatly improve the sensitivity to infrared rays.
[0028]
Next, a manufacturing process of the infrared detection pixel shown in FIG. 3 will be described.
[0029]
4 and 5 are principal part process cross-sectional views showing the manufacturing process of the infrared detection pixel.
[0030]
First, a so-called SOI substrate in which a buried silicon oxide film layer 9 and a single crystal silicon layer 8 are sequentially stacked on a single crystal silicon support substrate 6 as a semiconductor substrate is prepared (FIG. 4A).
[0031]
Next, the same process as STI (Shallow-Trench-Isolation) is performed as element isolation. That is, an element isolation region is defined using a technique such as photolithography, and the single crystal silicon layer 8 in the element isolation area is removed by etching using a technique such as RIE (Reactive-Ion-Etching). The film 14 is embedded by a technique such as CVD (Chemical-Vapor-Deposition), and planarized by a technique such as CMP (Chemical-Mechanical-Polishing) (FIG. 4B). At this time, it goes without saying that the region of the support portion 11 is also defined as an element isolation region and the element isolation silicon oxide film 14 is buried.
[0032]
Next, in the element isolation region, the element isolation silicon oxide film 14 and the buried silicon oxide film layer 9 in the support structure element isolation region for forming the support portion 11 are, for example, RIE (Reactive-Ion-Etching). Then, the sacrificial silicon film 21 is buried by a technique such as CVD (Chemical-Vapor-Deposition) and is flattened by a technique such as CMP (Chemical-Mechanical-Polishing) (FIG. 4B). .
[0033]
Since the sacrificial silicon film 21 is a so-called sacrificial layer to be etched in the etching of the support substrate 6 which will be performed later, it may have any structure of single crystal, polycrystal, and amorphous. .
[0034]
The surface of the single crystal silicon layer 8 in the so-called active region where the single crystal silicon layer 8 is exposed in the planarization step by CMP or the like after the sacrificial silicon film 21 is buried in the element isolation region for the support structure is exposed. In order to protect, it is more preferable to perform a step of protecting the surface of the single crystal silicon layer 8 with a silicon oxide film or the like prior to the silicon oxide film etching step in the support structure element isolation region.
[0035]
Next, the n-type impurity region 15 for the pn junction of the sensor portion is formed by photolithography technology and doping technology such as ion implantation as well as the source / drain regions of the peripheral circuits such as the address circuit, output portion, constant current source, etc. Then, a so-called metallization step of forming contact holes 16 for forming peripheral circuits and pn junction wiring and forming metal wiring 17 is performed.
[0036]
Next, although an infrared absorption layer is formed in the sensor part, it is also possible to divert the interlayer insulating film and the passivation film 18 formed in the metallization process as in this embodiment. (FIG. 4 (d)).
[0037]
Next, the passivation film 18 and the interlayer insulating film are etched by RIE (Reactive-Ion-Etching) or the like in order to form the support portion 11 and the etching hole 19 for forming the hollow structure 7 ( FIG. 5 (a)).
[0038]
Next, the sacrificial silicon film 21 is removed by etching with a chemical such as TMAH (Tetra-Methyl-Ammonium-Hydroxide) through the etching hole 19 (FIG. 5B).
[0039]
Finally, as the anisotropic etchant of the single crystal silicon, the single crystal silicon is anisotropically etched using a chemical such as TMAH (Tetra-Methyl-Ammonium-Hydroxide), for example, so that the inside of the single crystal silicon support substrate 6 The hollow structure 7 is formed in the structure, and the structure of the infrared detection pixel shown in FIG. 3 can be obtained (FIG. 5C).
[0040]
In the description of FIG. 5, the sacrificial silicon film 21 etching removal process and the subsequent anisotropic etching process of the supporting single crystal silicon substrate 6 have been described independently, but the chemicals used for these etchings are basically the same. Since it is the same chemical solution, in the actual process, after it becomes the shape of FIG. 5 (a), by performing etching with a chemical solution such as TMAH, for example, without being aware of the shape of FIG. 5 (b), The structure shown in FIG. 5C, which is the final shape, can be obtained.
[0041]
Of course, it goes without saying that a gate electrode forming step for forming a transistor used in the peripheral circuit is necessary, but since it is not directly related to the manufacturing process of the infrared detection pixel, in the above description, Omitted.
[0042]
As shown in FIG. 3, in the sensor structure of this embodiment, immediately after the element isolation step, a sacrificial silicon film 21 is embedded in the element isolation region for the support portion, and then the sacrificial silicon layer is removed by etching. As a result, the bottom portion of the support portion is formed to have the same height as the surface of the single crystal silicon layer 8. For this reason, the support portion 11 is formed to be significantly thinner. Therefore, the cross-sectional area of the support portion 11 is greatly reduced, and the thermal conductance between the sensor portion 10 and the support substrate 6 is greatly reduced. For comparison with the present embodiment, the reduction in the cross-sectional area is apparent from the cross-sectional structure shown in FIG. 18 as the cross-sectional structure of the conventional example.
[0043]
18A is a plan view of a conventional infrared detection pixel, and FIG. 18B is a cross-sectional view taken along the line AA ′. In this figure, the same elements as those in FIG. 3 are denoted by the same reference numerals, and detailed description thereof is omitted.
[0044]
Comparing the support portion 11 of the present embodiment shown in FIG. 3 with the conventional support portion 11 shown in FIG. 18, the cross-sectional area of the support portion 11 is significantly reduced in this embodiment. I understand. As a result, the “escape” of heat from the sensor unit 10 is greatly reduced, and the ratio of the temperature change of the sensor unit to the incident infrared power, so-called thermal sensitivity, is greatly improved. Therefore, an infrared sensor having high infrared sensitivity can be obtained. Obtainable.
[0045]
Here, after forming the shape of FIG. 5A, an appropriate amount of the passivation film 18 on the surface side of the support portion 11 is etched in a state where the region excluding the support portion 11 is protected by a photoresist or the like, and then sacrificed. It is also possible to obtain the structure shown in FIG. 6 by etching the silicon film 21 and the supporting single crystal silicon substrate 6.
[0046]
That is, in the structure shown in FIG. 6, the bottom surface 11B of the support portion 11 is not only formed at the same height as the top surface of the single crystal silicon layer 8, but the top surface 11T of the support portion 11 is Is formed at a position lower than the upper surface 10T. According to the structure of FIG. 6, the cross-sectional area of the support portion 11 is further reduced as compared with the structure of FIG. 3. Therefore, the “escape” of heat is further suppressed, and the infrared sensor is more sensitive. Can be
[0047]
Thus, it is practically advantageous to etch the upper surface 11T of the support portion so as to be lower than the upper surface 10T of the sensor portion. This is because the silicon oxide film 18 of the sensor unit 10 needs to be formed thick in order to improve the sensitivity of the sensor.
[0048]
FIG. 7 is a conceptual diagram showing an example in which the silicon oxide film 18 is formed thick as described above.
[0049]
That is, in order to increase the sensitivity to infrared rays, it is necessary to form the silicon oxide film 18 of the sensor unit 10 thick so as to sufficiently absorb infrared rays. Actually, the silicon oxide film 18 may be formed to a thickness of 500 nm or more. Further, a silicon nitride film 30 may be deposited as a light absorption film on this to a thickness of about 300 nm. In such a case, if the upper surface of the support portion 11 is etched and the upper surface 11T is formed lower than the upper surface 10T of the sensor portion, the cross-sectional area of the support portion 11 can be greatly reduced to further improve the sensitivity. Is possible.
[0050]
FIG. 8 is a conceptual diagram showing another specific example of the infrared sensor according to the present embodiment. That is, in this specific example, the single crystal silicon layer 8 and the wiring 7 are provided farther apart and are connected by a deep contact hole 16. This configuration may be necessary due to the configuration of peripheral circuits provided around the sensor unit 10, and this will be described in detail later.
[0051]
On the other hand, according to the steps shown in FIGS. 4 and 5, in addition to the effect that the completed infrared sensor having the structure of FIG. 3 is highly sensitive, there is also an effect that the manufacturing process is made easier. In order to explain this point, first, the manufacturing process of the infrared sensor having the conventional structure shown in FIG. 18 will be described for comparison.
[0052]
19 and 20 are process cross-sectional views showing the main part of the manufacturing process of the conventional infrared sensor shown in FIG. These drawings are the same as FIG. 4 and FIG. 5 except that the sacrificial silicon film 21 is not embedded in the support structure element isolation region and there is no step of etching away the sacrificial silicon film 21. Details are omitted here.
[0053]
With reference to these manufacturing steps, the thermal conductivity is calculated from the cross-sectional area of the support portion 11 in the present invention shown in FIG. 3 and the conventional structure shown in FIG. 18 as follows.
[0054]
First, it is assumed that a silicon oxide film is used as the insulating material in the support portion 11 and titanium is used as the material of the wiring 17. The width of the titanium wiring is 0.6 [μm] and the thickness is 0.05 [μm]. The width of the insulating material constituting the support portion 11 is 1.0 [μm] for protecting the titanium wiring.
[0055]
In the conventional structure of FIG. 18, the thickness of the insulating layer is as follows: buried silicon oxide film layer: 0.2 μm, SOI layer: 0.2 μm, silicon oxide film thickness under the wiring: 1 μm + 0.1 μm, silicon oxide film thickness over the wiring When the thickness is 0.5 μm, the total film thickness is 2 μm.
[0056]
In the case of the present invention shown in FIG. 3, the buried oxide film (0.2 μm), the SOI layer (0.2 μm), and the silicon oxide film thickness (1 μm) under the wiring are not required among the above-described conventional insulating films. . Therefore, the total film thickness of the insulating film is 0.6 μm.
[0057]
Here, the thermal conductivity per unit length of the Ti wiring is the same in the present invention and the conventional example, and is obtained by the following equation.
[0058]
2.2 × 10 −5 [W / μm / K] × 0.6 [μm] × 0.05 [μm] = 6.6 × 10 −7 [μm · W / K]
The thermal conductivity per unit length of the silicon oxide film is as follows.
[0059]
In the conventional example:
1.4 × 10 −6 [W / μm / K] × 1.0 [μm] × 2.0 [μm] = 2.8 × 10 −6 [μm · W / K]
In the present invention:
1.4 × 10 −6 [W / μm / K] × 1.0 [μm] × 0.6 [μm] = 8.4 × 10 −7 [μm · W / K]
The total heat conduction of Ti and the silicon oxide film is 3.5 × 10 −6 [W / μm / K] in the conventional example, and 1.5 × 10 −6 [W / μm / K] in the present invention. Become.
[0060]
That is, in the present invention, the heat conduction is reduced to (1.5 / 3.5) times, that is, about ½ or less, and as a result, the sensitivity is improved more than twice.
[0061]
Heretofore, the effect of suppressing the “escape” of heat in the support portion 11 has been quantitatively described.
[0062]
Next, actual merit in production will be described.
[0063]
First, as shown in FIG. 20A, the cross-sectional structure immediately after the silicon oxide film etching that serves as both the formation of the support portion 11 and the formation of the etching hole 19 and the cross-sectional structure of FIG. Compare.
[0064]
The silicon oxide film etching pattern in this step is the minimum dimension that can be processed by photolithography in order to make the sensor portion 8 as large as possible and the support portion 11 as thin and long as possible within the limited pixel region. Designed with. Further, in order to strictly process the layout designed with this minimum dimension, it is common to use RIE which is anisotropic etching for the silicon oxide film etching.
[0065]
In etching by RIE, it is generally known that a so-called aspect ratio, which is defined as a ratio between an opening width and an etching depth, is an index of technical difficulty in performing an RIE process. That is, the conventional process (FIGS. 19 and C) and the conventional structure (FIG. 18), which require a deeper etching depth even with the same opening width, are more difficult RIE processes. The manufacturing process according to the present embodiment is an easier RIE process.
[0066]
Furthermore, according to this embodiment, in addition to the effect of increasing sensitivity by forming the support portion 11 thin and the effect of facilitating the manufacturing process, there are secondary effects resulting from the ease of the manufacturing process. There is also an effect of high sensitivity.
[0067]
That is, RIE, which is anisotropic etching, realizes anisotropic etching perpendicular to the direction of ion incidence, that is, to the substrate, using the ionicity of the etching. However, when the above-mentioned etching depth is deep, the etching opening width is narrow, and the aspect is high, the directionality of incident ions is deteriorated. However, the occurrence of a tapered shape is widely known as a general fact.
[0068]
Here, a silicon oxide film RIE having a high aspect ratio will be considered with reference to FIG. Then, in FIG. 20 (a), the cross section of the etching hole expressed vertically is actually slightly tapered. In consideration of the tapered shape of the etching hole shape, it can be seen that the cross-sectional area of the actually produced support portion 11 is larger than that shown in FIGS. 20A and 18 for the following two reasons. .
[0069]
The first reason is that the taper shape of the cross section of the etching hole 19 results in a reverse taper shape of the cross section of the support portion 11, so that the cross section of the support portion 11 has a trapezoidal shape with the lower bottom longer than the upper base. This is because it becomes larger than the cross-sectional area designed by the upper base and the thickness.
[0070]
Another reason is that when the etching hole 19 has a tapered shape, the opening area at the bottom of the etching hole 19 to which the etching chemical is supplied is reduced in the etching process of the supporting single crystal silicon substrate 6. That is, in consideration of the tapered shape of the cross section of the etching hole 19, in order to secure the opening area at the bottom of the etching hole, it is necessary to perform a slightly enlarged layout on the bottom of the etching hole 19. Therefore, the cross-sectional area of the support portion 11 is further increased, and the sensitivity of the infrared sensor is further deteriorated.
[0071]
(Second Embodiment)
Next, a second embodiment of the present invention will be described. The overall configuration of the infrared sensor according to the present embodiment and the equivalent circuit of the infrared detection pixel are the same as those shown in FIGS. 1 and 2 as in the first embodiment, and the description thereof is omitted.
[0072]
FIG. 9 is a schematic structural diagram for explaining a cross-sectional structure and a planar structure of an infrared detection pixel according to the present embodiment. The structure shown in FIG. 9 is substantially the same as FIG. 3 shown in the first embodiment, but the bottom surface 11B of the support portion 11 is higher than the height of the single crystal silicon layer 8 when viewed from the substrate 6, and the wiring The difference is that it is formed to be almost the same height as 17.
[0073]
Also in the present embodiment, the thickness T1 of the support portion 11 is formed to be thinner than the thickness T2 of the sensor portion 10, and is configured to be able to suppress heat “escape”.
[0074]
Further, the depth of the contact hole 16 in the sensor unit 10 is deeper than that in FIG. This is for the purpose of facilitating understanding of the difference from the first embodiment in describing this embodiment.
[0075]
The difference between the present embodiment and the first embodiment described above is easier to understand when the present embodiment shown in FIG. 9 is compared with the first embodiment shown in FIG. The difference between FIG. 3 and FIG. 6 describing the first embodiment is only that the depth of the contact hole 16 in the sensor unit 10 is deep. 8 can be said to represent a structure closer to the actual cross-sectional structure than FIG. This reflects the situation that the contact hole 16 inside the sensor unit 8 becomes deep due to the presence of transistors and capacitors that constitute peripheral circuits such as an address circuit and an output circuit, which are not shown in FIG. is there. The situation will be described with reference to FIG.
[0076]
FIG. 10 is a conceptual diagram in which sensor parts, transistors, and capacitors are arranged from the left and their cross-sectional structures are compared. Since the structure of the sensor unit 10 is as described above, description thereof is omitted. The n-channel type MOS transistor portion Tr is composed of an N-type impurity region constituting source / drain and a gate electrode 22 formed on the gate oxide film, and the capacitor C is formed on the element isolation silicon oxide film 14. The capacitor lower electrode 23 and the capacitor upper electrode 24 stacked on the lower electrode via a dielectric film.
[0077]
The wiring 17 formed in each of these regions is a common wiring, and the surface of the insulating film before the wiring 17 is formed is planarized by a technique such as CMP. Therefore, as shown in FIG. 10, it is necessary to form the wiring 17 at a position higher than the gate electrode 22 of the transistor and the capacitor upper electrode 24 through the insulating film. Therefore, the actual cross-sectional structure of the sensor portion is a structure having a deep contact hole 16 as shown in FIG. 10 or FIG. In consideration of the actual capacitor structure, the depth is about 1 μm.
[0078]
Similarly, FIG. D shows a pixel structure of an infrared sensor having a conventional structure in consideration of a deep contact hole in a sensor portion formed due to the presence of a peripheral circuit.
[0079]
Returning to the description of the present embodiment shown in FIG.
[0080]
The structure shown in FIG. 9 is almost the same as FIG. 3 shown as the first embodiment, but the bottom of the support structure 11 is higher than the height of the single crystal silicon layer 8 and is almost the same as the wiring 17. The part formed in height is different. Therefore, the cross-sectional area of the support part 11 can be reduced further than in the first embodiment, and an infrared sensor with higher sensitivity can be obtained. Here, as can be understood from the description given above with reference to FIG. 10, the thickness of the support portion 11 in the present embodiment is formed to be about 1 μm thinner than the thickness of the support portion 11 in the first embodiment. It will be.
[0081]
Next, a process for manufacturing the pixel of the infrared sensor according to the present embodiment will be described with reference to process cross-sectional views in FIGS.
[0082]
First, a so-called SOI substrate is prepared in which a buried silicon oxide film layer 9 and a single crystal silicon layer 8 are sequentially stacked on a single crystal silicon support substrate 6 as a semiconductor substrate (FIG. 11A).
[0083]
Next, the same process as STI (Shallow-Trench-Isolation) is performed as element isolation. That is, an element isolation region is defined using a technique such as photolithography, and the single crystal silicon layer 8 in the element isolation area is removed by etching using a technique such as RIE (Reactive-Ion-Etching). The film 14 is embedded by a technique such as CVD (Chemical-Vapor-Deposition), and is planarized by a technique such as CMP (Chemical-Mechanical-Polishing) (FIG. 11B). At this time, it goes without saying that the region of the support portion 11 is also defined as an element isolation region and the element isolation silicon oxide film 14 is buried.
[0084]
Next, the n-type impurity region 15 for the pn junction of the sensor portion is formed by photolithography technology and doping technology such as ion implantation as well as the source / drain regions of the peripheral circuits such as the address circuit, output portion, constant current source, etc. After the formation, an insulating layer 25 is formed (FIG. 11C).
[0085]
Next, in the element isolation region, the insulating layer 25, the element isolation silicon oxide film 14, and the buried silicon oxide film layer 9 in the element isolation region for supporting structure for forming the support portion 11 are, for example, RIE ( Etching is removed by a technique such as Reactive-Ion-Etching (FIG. 11D).
[0086]
Next, the sacrificial silicon film 21 is embedded by a technique such as CVD (Chemical-Vapor-Deposition), and is planarized by a technique such as CMP (Chemical-Mechanical-Polishing) (FIG. 12A). Since the sacrificial silicon film 21 is a so-called sacrificial layer to be etched in the etching of the support substrate 6 which will be performed later, it may have any structure of single crystal, polycrystal, and amorphous. .
[0087]
Next, an insulating film 26 is formed in the entire region, and an insulating layer between the sacrificial silicon layer 21 and the wiring 17 is formed. In this support part 11, the insulating layer 26 formed under the wiring 17 finally becomes the bottom surface of the support part 11 (FIG. 12B).
[0088]
Next, a so-called metallization process of forming contact holes 16 and wirings 17 for peripheral circuit and pn junction wiring is performed.
[0089]
Next, although an infrared absorption layer is formed in the sensor part, it is also possible to divert the interlayer insulating film and the passivation film 18 formed in the metallization process as in this embodiment. (FIG. 12 (c)).
[0090]
Next, the passivation film 18 and the insulating film 26 are etched by RIE (Reactive-Ion-Etching) or the like in order to form the support portion 11 and the etching hole 19 for forming the hollow structure 7 (FIG. 13 (a)).
[0091]
Next, the sacrificial silicon film 21 is removed by etching with a chemical such as TMAH (Tetra-Methyl-Ammonium-Hydroxide) through the etching hole 19 (FIG. 13B).
[0092]
Finally, as the anisotropic etchant of single crystal silicon, for example, the single crystal silicon is anisotropically etched using a chemical solution such as TMAH (Tetra-Methyl-Ammonium-Hydroxide), so that the inside of the single crystal silicon support substrate 6 is inside. The hollow structure 7 is formed in the structure, and the structure of the infrared detection pixel shown in FIG. 9 can be obtained (FIG. 13C).
[0093]
In the description of FIGS. 13B to 13C, the sacrificial silicon film 21 is removed independently of the etching process and the subsequent anisotropic etching process of the supporting single crystal silicon substrate 6. Since the chemical solution to be used is basically the same chemical solution, in the actual process, after forming the shape of FIG. 13A, for example, by etching with a chemical solution such as TMAH, the chemical solution of FIG. The structure shown in FIG. 13C, which is the final shape, can be obtained without being conscious of the shape.
[0094]
Of course, it goes without saying that a gate electrode forming step for forming a transistor used in the peripheral circuit is necessary, but since it is not directly related to the manufacturing process of the infrared detection pixel, in the above description, Omitted.
[0095]
As shown in FIG. 9, in the sensor structure of the present embodiment, a sacrificial silicon film 21 is embedded in the support element isolation region in the planarization process of the insulating film 25 performed before the metallization process. After that, the sacrificial silicon layer is removed by etching, so that the bottom portion 11B of the support portion is formed at the same height as the surface of the wiring 17. Therefore, the support portion 11 is formed to be much thinner. Therefore, the cross-sectional area of the support portion 11 is further greatly reduced, and the thermal conductance between the sensor portion 10 and the support substrate 6 is further greatly reduced. The For comparison with the present embodiment, the reduction in the cross-sectional area of the support portion 11 is apparent from a comparison with the cross-sectional structure diagram shown in FIG.
[0096]
As a result, the ratio of the temperature change of the sensor unit with respect to the incident infrared power, so-called thermal sensitivity, is greatly improved, so that an infrared sensor having high infrared sensitivity can be obtained.
[0097]
In addition, after forming the shape of FIG. 13A, an appropriate amount of the passivation film 18 on the surface side of the support portion 11 is etched in a state where the region excluding the support portion 11 is protected by a photoresist or the like, and then the sacrificial silicon is etched. It is also possible to obtain the structure shown in FIG. 15 by removing the film 21 (FIG. 14B) and further etching the supporting single crystal silicon substrate 6 (FIG. 14C).
[0098]
That is, in the structure shown in FIG. 15, the bottom surface 11B of the support portion 11 is not only formed at the same height as the surface of the wiring 17, but the top surface 11T of the support portion 11 is the surface 10T of the sensor portion 10. It is formed at a lower height. According to this structure, the cross-sectional area of the support portion 11 is further reduced as compared with the structure of FIG. 9, and thus the sensitivity of the infrared sensor can be further increased, which is more preferable.
[0099]
The embodiments of the present invention have been described above with reference to specific examples. However, the present invention is not limited to these specific examples.
[0100]
For example, further supplementing the wiring 17 provided in the infrared sensor of the present invention, the wiring in the sensor unit 10 and the wiring in the support unit 11 are not limited to the same material. For example, a metal wiring material such as aluminum (Al) used in a normal LSI process is used as a material of the wiring 17 in the sensor unit 10, and titanium (Ti) having a low thermal conductivity is used as a material of the wiring 17 of the support unit 11. ) Is more preferable because it is possible to further reduce the “escape” of heat.
[0101]
In this case, the wiring 17 is formed twice by the first wiring forming process for forming the aluminum wiring in the sensor unit 10 and the second wiring process for forming the titanium wiring in the support unit 11. Become. At this time, a method of forming an interlayer insulating film between the aluminum wiring and the titanium wiring and forming a contact hole for connecting both wirings is possible, but the following method is also possible and more preferable.
[0102]
This is a method in which the second wiring step is performed immediately after the first wiring forming step after designing the second wiring pattern to be surely formed on the first wiring pattern. Of course, only the second wiring pattern is formed on the support portion 11. According to this method, not only the above-mentioned insulating layer formation and contact formation steps are unnecessary, but also the occurrence of contact failure due to the disconnection of the titanium wiring in the contact hole for forming the titanium wiring with a very thin film thickness on the upper layer is prevented. Is possible. Needless to say, the purpose of reducing the thickness of the titanium wiring is to reduce the heat conduction of the support portion 11 and increase the sensitivity.
[0103]
On the other hand, each of the first and second embodiments described above is an infrared sensor configured by two-dimensionally arranging infrared detection pixels. Of course, the infrared detection pixels are arranged one-dimensionally. It goes without saying that the same effect can be obtained even when applied to a dimension sensor or a single infrared sensor not arranged in an array.
[0104]
Further, in addition to the specific examples described above, an infrared absorption layer may be stacked.
[0105]
FIG. 16 is a conceptual diagram illustrating a configuration of an infrared sensor in which an infrared absorption layer is stacked. That is, in any of the configurations described above as the embodiment of the present invention, the infrared absorption layer 100 is laminated on the sensor unit 10. The infrared absorption layer 100 has a configuration in which a reflective layer 100 </ b> A, an insulating layer 100 </ b> B, and an absorption layer 100 </ b> C are stacked, absorbs infrared rays, and supplies heat to the sensor unit 10. Since such an infrared absorption layer 100 has a detection area far larger than that of the sensor unit 10, it is possible to obtain an infrared sensor with a higher sensitivity in combination with an optical sensitivity improvement effect by expanding the infrared absorption area. It becomes.
[0106]
Further, the pn junction used as the thermoelectric conversion means in the present invention is not limited to the planar pn junction, and a lateral pn junction can be similarly employed as the thermoelectric conversion means.
[0107]
FIG. 17 is a perspective perspective view conceptually showing the main part of such a lateral structure. That is, in the example illustrated in the figure, a plurality of lateral pn junction diodes 220 made of SOI films are provided on a buried oxide film 210 formed on a substrate 200 and are connected in series by a metal strap 230. In the present invention, a pn junction having such a lateral structure can be similarly employed as the thermoelectric conversion means.
[0108]
In addition, various modifications can be made without departing from the scope of the present invention.
[0109]
【The invention's effect】
As described above in detail, according to the present invention, the support for supporting the sensor unit on the hollow structure is significantly thinner than the conventional structure, the cross-sectional area is greatly reduced, and the thermal conductance is reduced. As a result, an extremely sensitive infrared sensor can be obtained.
[0110]
Further, according to the present invention, since the insulating layer in the support region is etched and the sacrificial silicon film is embedded, the aspect ratio of the insulating layer RIE for forming the support legs is greatly reduced, and the process is facilitated. As a secondary effect, it is possible to further reduce the cross-sectional area of the support leg and increase the sensitivity.
[0111]
That is, according to the present invention, it is possible to easily and surely obtain an uncooled infrared sensor having higher sensitivity than before, and it is possible to provide a high-performance sensor at various costs in various application fields. The above benefits are enormous.
[Brief description of the drawings]
FIG. 1 is an overall configuration diagram of an infrared sensor according to a first embodiment of the present invention.
FIG. 2 is an equivalent circuit diagram of the infrared detection pixel 1 of FIG.
3A is a plan configuration diagram of the infrared detection pixel shown in FIG. 2, and FIG. 3B is a cross-sectional view taken along line AA ′ of FIG.
FIG. 4 is a cross-sectional view of a main part showing a manufacturing process of an infrared detection pixel.
FIG. 5 is a cross-sectional view of a main part showing a manufacturing process of an infrared detection pixel.
6A is a plan view of an infrared detection pixel, and FIG. 6B is a cross-sectional view taken along line AA ′ of FIG.
FIG. 7 is a conceptual diagram showing an example in which a silicon oxide film 18 is formed thick.
FIG. 8 is a conceptual diagram showing another specific example of the infrared sensor according to the first embodiment of the present invention.
FIG. 9 is a schematic structural diagram for explaining a planar configuration and a cross-sectional structure of an infrared detection pixel according to a second embodiment of the present invention.
FIG. 10 is a conceptual diagram in which sensor parts, transistors, and capacitors are arranged from the left and their cross-sectional structures are compared.
FIG. 11 is a process cross-sectional view illustrating a process for manufacturing a pixel of an infrared sensor according to a second embodiment of the invention.
FIG. 12 is a process cross-sectional view illustrating a process of manufacturing a pixel of an infrared sensor according to a second embodiment of the invention.
FIG. 13 is a process cross-sectional view illustrating a process for manufacturing a pixel of an infrared sensor according to a second embodiment of the invention.
FIG. 14 is a process cross-sectional view illustrating a process for manufacturing a pixel of an infrared sensor according to a second embodiment of the invention.
15 is a conceptual diagram of an infrared sensor in which an upper surface 11T of a support part 11 is formed at a height lower than a surface 10T of a sensor part 10. FIG.
FIG. 16 is a conceptual diagram illustrating a configuration of an infrared sensor in which an infrared absorption layer is stacked.
FIG. 17 is a perspective perspective view conceptually showing a main part of a lateral structure.
18A is a plan configuration diagram of a conventional infrared detection pixel, and FIG. 18B is a cross-sectional view taken along the line AA ′ of FIG.
FIG. 19 is a process cross-sectional view illustrating a main part of a manufacturing process of a conventional infrared sensor.
FIG. 20 is a process cross-sectional view illustrating a main part of a manufacturing process of a conventional infrared sensor.
FIG. 21 (a) is a plan configuration diagram of a conventional infrared detection pixel, and FIG. 21 (b) is a cross-sectional view taken along line AA ′.
[Explanation of symbols]
1 Infrared detection pixel
2 Infrared detection pixel array
3 Vertical signal lines
4 Horizontal address lines
5 column selection transistor
6 Single crystal silicon support substrate
7 Hollow structure (hollow part)
8 Single crystal silicon layer
9 Embedded silicon oxide layer
10 Sensor part
11 Support part (support)
14 Device isolation silicon oxide film
15 N-type impurity region
16 Contact hole
17 Metal wiring
18 Passivation film
19 Etching hole
21 Sacrificial silicon film
22 MOS transistor gate electrode
23 Capacitor lower electrode
24 Capacitor upper electrode
25, 26 Insulating layer

Claims (8)

An infrared sensor provided with a plurality of infrared detectors on a semiconductor substrate having an absorption part that absorbs incident infrared rays and converts the infrared rays into heat, and a thermoelectric conversion part that converts a temperature change caused by heat generated in the absorption part into an electrical signal Because
A passivation film formed on the semiconductor substrate;
And at least one support that is disposed in a recess formed by etching and removing a part of the passivation film and a part of the semiconductor substrate, and supports each of the plurality of infrared detection units. ,
Wherein each of the upper surface of the support is lower than the upper surface of the passivation film formed on the side wall of the recess, and rather low from the upper surface of the infrared detection unit which is disposed in said recess,
The infrared sensor according to claim 1, wherein a lower surface of each of the supports is higher than a lower surface of the infrared detection unit .   2. The thickness of a surface of the support body that is spaced and opposed to the infrared detection unit is thinner than a thickness of a surface of the infrared detection unit that is disposed to face the surface of the support. Infrared sensor. The semiconductor substrate is an SOI substrate having a single crystal silicon substrate, a buried oxide film formed on the single crystal silicon substrate, and a single crystal silicon layer formed on the buried oxide film,
The infrared sensor according to claim 1, wherein the recess is formed to a depth that reaches the single crystal silicon substrate of the SOI substrate.   The infrared sensor according to claim 1, wherein the thermoelectric conversion part includes a pn junction formed by forming a second conductivity type region in a part of the first conductivity type single crystal semiconductor layer. . The lower surface of the support is provided at the same height as the upper surface of the single crystal semiconductor layer as viewed from the semiconductor substrate or at a position higher than the upper surface of the single crystal semiconductor layer. The described infrared sensor. The infrared detection unit includes a first wiring extending in a direction parallel to the semiconductor substrate surface in order to output an electrical signal from the thermoelectric conversion unit,
The support has a second wiring connected to the first wiring;
The infrared sensor according to claim 1, wherein the second wiring is made of a material having a lower thermal conductivity than that of the first wiring. The infrared sensor according to claim 1, wherein the support body is formed around the infrared detection unit. The infrared sensor according to claim 1, wherein the support extends in parallel to the surface of the semiconductor substrate.

Images