Semiconductor device and method for production of semiconductor device

Abstract

A semiconductor device with a connection pad in a substrate, the connection pad having an exposed surface made of a metallic material that diffuses less readily into a dielectric layer than does a metal of a wiring layer connected thereto.

Claims

1. A semiconductor device, comprising: a first semiconductor substrate; and a first wiring layer formed on the first semiconductor substrate, wherein: the first wiring layer includes a first pad, a first insulating film, and a diffusion preventing layer, the first pad and the diffusion preventing layer comprise at least a section of a first surface of the first wiring layer, and a surface of the first pad contacts the diffusion preventing layer and the first insulating film.

2. The semiconductor device of claim 1, wherein a first portion of the diffusion preventing layer contacts a second insulating film, and wherein the second insulating film is included in a second wiring layer of a second semiconductor layer.

3. The semiconductor device of claim 2, wherein a second portion of the diffusion preventing layer forms a surface opposite to the first surface of the first wiring layer and contacts the first insulating film.

4. The semiconductor device of claim 3, wherein the surface of the first pad extends in a direction perpendicular to the first surface of the first wiring layer.

5. The semiconductor device of claim 1, wherein the diffusion preventing layer includes at least one of Si, N, O, and C.

6. The semiconductor device of claim 1, wherein the diffusion preventing layer includes one of SiN and SiOC.

7. The semiconductor device of claim 1, further comprising a barrier metal.

8. The semiconductor device of claim 7, wherein the barrier metal extends in a direction perpendicular to the first surface of the wiring layer.

9. The semiconductor device of claim 7, wherein the barrier metal extends in a direction parallel to the first surface of the wiring layer.

10. The semiconductor device of claim 7, wherein the barrier metal includes at least one of Ti, N, and Ta.

11. The semiconductor device of claim 10, wherein the barrier metal includes one of TiN and TaN.

12. The semiconductor device of claim 7, wherein a portion of the barrier metal contacts the first pad.

13. The semiconductor device of claim 7, wherein the first pad is partially covered by the barrier metal.

14. The semiconductor device of claim 1, wherein a surface of a second pad of a second semiconductor substrate is exposed through the diffusion preventing layer.

15. The semiconductor device of claim 1, wherein the diffusion preventing layer prevents metal of the first pad from diffusing into a dielectric film of a second wiring layer of a second semiconductor substrate.

16. The semiconductor device of claim 1, wherein the first pad includes at least one of Cu, Au, Ag, Al, Ta, Ti, W, Sn, Mo, Ni, In, Co and an alloy of any of Cu, Au, Ag, Al, Ta, Ti, W, Sn, Mo, Ni, In, and Co.

17. The semiconductor device of claim 1, wherein a width of a second pad of a second semiconductor in a direction parallel to the first surface of the first wiring layer is greater than a width of the first pad in the direction parallel to the first surface of the wiring layer.

18. The semiconductor device of claim 1, wherein a width of a second pad of a second semiconductor in a direction parallel to the first surface of the first wiring layer is equal to a width of the first pad in the direction parallel to the first surface of the wiring layer.

19. The semiconductor device of claim 1, wherein a first portion of the diffusion preventing layer is disposed between the first insulating film and a second insulating film.

20. A semiconductor device, comprising: an imaging device comprising a first semiconductor substrate; and a first wiring layer formed on the first semiconductor substrate, wherein: the first wiring layer includes a first pad, a first insulating film, and a diffusion preventing layer, the first pad and the diffusion preventing layer comprise at least a section of a first surface of the first wiring layer, and a surface of the first pad contacts the diffusion preventing layer and the first insulating film.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIGS. 1A and 1B are a schematic perspective view and a schematic sectional view, respectively, showing the laminated wafer pertaining to one embodiment of the present invention;

(2) FIG. 2A is a sectional view of the region IIa in FIG. 1B, and FIGS. 2B and 2C are plan views of the region IIa in FIG. 1B;

(3) FIG. 3 is a schematic diagram illustrating the method for production of chips to be fabricated from laminated wafers shown in FIG. 1;

(4) FIG. 4 is a sectional view showing a first modification of the embodiment shown in FIG. 2A;

(5) FIG. 5 is a sectional view showing a second modification of the embodiment shown in FIG. 2A;

(6) FIG. 6 is a sectional view showing a third modification of the embodiment shown in FIG. 2A;

(7) FIGS. 7A and 7B are plan views showing a fourth modification of the embodiments shown in FIGS. 2B and 2C, respectively; and

(8) FIGS. 8A and 8B are schematic perspective views showing an example of chips bonded together.

DETAILED DESCRIPTION OF THE PRESENTLY PREFERRED EMBODIMENTS

(9) FIG. 1A is a schematic perspective view showing the laminated wafer 1 pertaining to one embodiment of the present invention.

(10) The laminated wafer 1 is composed of the first wafer 3A and the second wafer 3B which are bonded together. (These wafers will be simply referred to a “wafer 3” without discrimination in some cases hereinafter.) The laminated wafer 1 is divided into a plurality of chips (semiconductor device 5) by dicing.

(11) FIG. 1B is a schematic sectional view taken along the line Ib to Ib in FIG. 1A.

(12) The first wafer 3A is composed of the first semiconductor substrate 7A and the multiple layers laminated thereon which are the wiring layers 9 and the interlayer dielectric films 11. Incidentally, FIG. 1B does not show the boundary line between the multiple layers, which may constitute an interlayer dielectric film 11. Like the first wafer 3A, the second wafer 3B is also composed of the second semiconductor substrate 7B and the multiple layers laminated thereon which are the wiring layers 9 and the interlayer dielectric films 11. In the first and second wafers 3A and 3B, the multiple wiring layers 9 are connected to one another through the vias 13 penetrating the interlayer dielectric films 11.

(13) Incidentally, the term “semiconductor substrate 7” will be occasionally used hereinafter to denote both the first semiconductor substrate 7A and the second semiconductor substrate 7B without discrimination between them.

(14) The semiconductor substrate 7 is an unfabricated wafer (or wafer in a narrow sense), which is formed from silicon, for instance. The wiring layers 9 and the vias 13 are formed from copper, for instance. The interlayer dielectric films 11 are formed from any material containing at least one of silicon, nitrogen, oxygen, and carbon. Their examples include silicon oxide film.

(15) The wiring layers 9 and the vias 13 in the wafers 3A and 3B have the barrier metal 21 which prevents their diffusion into the interlayer dielectric film 11. See FIG. 2A. The barrier metal 21 may be formed from materials, such as TiN or TaN, for instance.

(16) Incidentally, the semiconductor substrate 7 and the interlayer dielectric film 11 will be referred by the same name or symbol for brevity hereinafter both before dicing and after dicing.

(17) Each of the wafers 3A and 3B will be made into a monofunctional LSI by fabrication of semiconductor elements (not shown), the wiring layers 9, and the vias 13 on the semiconductor substrate 7. For example, the monofunctional LSI on the wafer 3A and the wafer 3B may be the memory 31 and the logic device 33, respectively. The memory 31 may be DRAM, SRAM, and flash memory, for instance, and the logic device 33 may be MPU and peripheral circuit, for instance. If the wafers 3A and 3B each having the monofunctional LSI are bonded together, it will be possible to produce an LSI chip with multiple functions and a high degree of integration.

(18) FIG. 2A is an enlarged view of the region IIa shown in FIG. 1B. FIG. 2B is a plan view (seen from the second wafer 3B) of the first wafer 3A in the region IIa shown in FIG. 2A. FIG. 2C is a plan view (seen from the first wafer 3A) of the second wafer 3B in the region IIa shown in FIG. 2A.

(19) The two wafers 3A and 3B form an integral body, with the interlayer dielectric films 11 (as the uppermost layers) bonded to each other. Incidentally, terms “the first dielectric film 15A” and “the second dielectric film 15B” will be used hereinafter to denote respectively the uppermost interlayer dielectric film 11 of the first wafer 3A and the uppermost interlayer dielectric film 11 of the second wafer 3B. Also, a term “the dielectric film 15” will be used occasionally hereinafter to simply denote both “the first dielectric film 15A” and “the second dielectric film 15B” without discrimination between them.

(20) Also, the two wafers 3A and 3B are electrically connected to each other through the first pad 17A (in the first wafer 3A) and the second pad 17B (in the second wafer 3B) which are in contact with each other. Incidentally, the first pad 17A and the second pad 17B will be simply referred to as the “pad 17” hereinafter without discrimination between them.

(21) The first pad 17A is formed from metal filled into the recess (groove) formed in the first dielectric film 15A. The first pad 17A has an exposed face which is flush with the boundary between the first dielectric film 15A and the second dielectric film 15B. The first pad 17A has a rectangular shape in plan view, for instance.

(22) The first pad 17A is connected to the via 13 formed right above it. In this way the first pad 17A is connected to the first wiring conductor 19A shown in FIG. 1B, which is composed of the wiring layer 9 and the via 13 and is covered with the first dielectric layer 15A (and other interlayer dielectric film 11).

(23) The second pad 17B is also formed from metal filled into the recess (groove) formed in the second dielectric film 15B. The second pad 17B has an exposed face which is flush with the boundary between the first dielectric film 15A and the second dielectric film 15B. The second pad 17B has a rectangular shape in plan view, for instance.

(24) The second pad 17B is formed from a metal which is less diffusible into the interlayer dielectric film 11 than copper. Such a metal includes, for example, Au, Ag, Al, Ta, Ti, W, Sn, Mo, Ni, In and Co and an alloy containing at least one of them.

(25) The second pad 17B is connected to the second wiring conductor 19B shown in FIG. 1B, which is composed of the wiring layer 9 and the via 13 and is covered with the second dielectric layer 15B.

(26) The first and second wiring conductors 19A and 19B may be simply referred to as “the wiring conductor 19” hereinafter without discrimination between them.

(27) The second pad 17B is formed wider than the first pad 17A. In other words, the two pads 17 are formed in such a way that the first pad 17A remains covered by the second pad 17B even though they get out of the position when the two wafers 3 are bonded together.

(28) Misalignment at the time of bonding is about ±3 μm, for instance. Therefore, the second pad 17B should be larger than the first pad 17A by 6 μm or more.

(29) The result of the two pads 17 being formed as mentioned above is that only the second pad 17B comes into contact with the dielectric film 15 in case of misalignment at the time of bonding. The second pad 17B is less diffusible into the dielectric film 15 than the first pad 17A.

(30) Incidentally, the first pad 17A may be connected to the first wiring conductor 19A through the connecting part in the same way as the second pad 17B. In this case, the connecting part 18 for the second pad 17B should be larger than the connecting part for the first pad 17A by more than the amount of misalignment.

(31) The second pad 17B may have the via 13 right below it in the same way the first pad 17A, so that it is connected to the second wiring conductor 19B.

(32) FIG. 3 is a conceptual diagram illustrating the process for producing the chip 5.

(33) The wiring step forms by repeated photolithography the interlayer dielectric film 11 (in multiple layers), the wiring layer 9 (in multiple layers), the wiring conductor 19, and the pad 17. The wiring step employs the film deposition system 51, the exposure system 53, the etching system 55, and the planarization system 57.

(34) In the wiring step, the pad 17 is formed by damascene process (either single damascene process or dual damascene process). FIG. 2A shows the first pad 17A which was formed by dual damascene process.

(35) Each of the wafers 3 undergoes the bonding pretreatment step, which involves activation of the surface of the wafer 3 and removal of oxide film from the pad 17. This step is accomplished by reduction or annealing. Reduction employs hydrogen plasma, NH.sub.3 plasma, or formic acid plasma. Annealing employs hydrogen or forming gas (N.sub.2 or H.sub.2).

(36) Incidentally, FIG. 3 schematically shows the volumetric plasma reduction treatment system 59, which is run under the following conditions. Gas: H.sub.2/Ar=100/170 sccm Microwave: 2.8 kW (2.45 GHz) Pressure: 0.4 Pa Substrate temperature: 400° C. Duration: 1 min

(37) The pretreatment step is followed by the bonding step in which the pretreated wafers 3 are bonded together by the bonding system 61, such as any bonding system that is effective to bond an interposer to the semiconductor substrate.

(38) The bonding system 61 has the function to position the two wafers 3 and the function to bond them together with heating under pressure.

(39) The positioning is accomplished by causing the wafer's notch or orientation flat to engage with a relevant engaging member, or by causing the wafer 3 to fit into a relevant positioning member (such as a V-shaped frame), or by detecting the notch, orientation flat, and/or entire edge and properly moving the wafer 3 according to the results of detection.

(40) FIG. 3 shows an example of the equipment which has the detector 63 to detect the notch of the second wafer 3B, the second table 65B supporting the second wafer 3B, and the drive unit 67 to achieve positioning by translational or rotational movement of the second wafer 3B according to the results of detection.

(41) The positioning unit may be combined with or separated from the heating and pressurizing unit. In the case of the equipment shown in FIG. 3, each of the first and second tables 65A and 65B, which are intended for positioning, is provided with the heater 69 and the first and second tables 65A and 65B apply heat and pressure.

(42) The bonding equipment (or the positioning unit in the bonding equipment) varies in positioning accuracy depending on the principle of positioning, the tolerance of the members, and the accuracy of the constituents of the equipment. In the wiring step mentioned above, the second pad 17B is formed in such a way that it is larger than the first pad 17A by more than the accuracy of positioning by the bonding equipment 61.

(43) Incidentally, the accuracy of positioning may be based on experimental values available from the producer of the bonding equipment or obtained by the user of the bonding equipment.

(44) Thus the two wafers 3 are bonded together, and the resulting laminated wafer 1 undergoes dicing by the dicing blade 71, so that it is divided into a plurality of chips 5.

(45) According to the foregoing embodiment, the resulting chip 5 is composed of the first semiconductor substrate 7A and the second semiconductor substrate 7B, both facing each other. In addition, the chip 5 has the first wiring conductor 19A formed in the first semiconductor substrate 7A and the second wiring conductor 19B formed in the second semiconductor substrate 7B. The chip 5 also has the first dielectric film 15A, which covers the first wiring conductor 19A, and the second dielectric film 15B, which covers the second wiring conductor 19B and faces the first dielectric film 15A bonded thereto. The chip 5 also has the first pad 17A, which is connected to the first wiring conductor 19A and faces toward the second pad 17B. The chip 5 also has the second pad 17B, which is connected to the second wiring conductor 19B and faces toward the first pad 17A bonded thereto. The second pad 17B is formed from a metal which is less diffusible into the first dielectric film 15A than the second wiring conductor 19B.

(46) The method for producing the chip 5 includes a step of forming the first pad 17A in the first semiconductor substrate 7A which has the first wiring conductor 19A formed therein and the first dielectric film 15A formed therein which covers the first wiring conductor 19A. The first pad 17A is connected to the first wiring conductor 19A and exposes itself from the first dielectric film 15A. The method for producing the chip 5 also includes a step of forming the second pad 17B in the second semiconductor substrate 7B which has the second wiring conductor 19B formed therein and the second dielectric film 15B formed therein which covers the second wiring conductor 19B. The second pad 17B is connected to the second wiring conductor 19B and exposes itself from the second dielectric film 15B. Moreover, the method for producing the chip 5 includes a step of bonding together the first semiconductor substrate 7A and the second semiconductor substrate 7B, with the first pad 17A and the second pad 17B kept in contact with each other. Then, the second pad 17B is formed from a metal which is less diffusible into the first dielectric film 15A than the second wiring conductor 19B.

(47) Therefore, even though misalignment occurs at the time of positioning for bonding and the second pad 17B comes into contact with the first dielectric film 15A, metal diffusion into the first dielectric film 15A is less significant than in the case where the second pad 17B is formed from the metal constituting the second wiring conductor 19B. Although any metal less diffusible into the dielectric film is usually expensive, such a metal is not used for the entire wiring but is used only for the second pad 17B according to an embodiment of the present invention. This contributes to cost reduction and protection of the second pad 17B from oxidation.

(48) Since the second pad 17B is wider than the first pad 17A, the first pad 17A does not come into contact with the second dielectric film 15B even in the case of positioning misalignment. As the result, if either of the first pad 17A and the second pad 17B is formed from a metal less diffusible into the dielectric film 15, it is possible to prevent metal from diffusing into the dielectric film 15 due to positioning misalignment. Consequently, the first pad 17A can be formed from a metal (such as the same one as used for the first wiring conductor 19A) which is as diffusible into the second dielectric film 15B as the first wiring conductor 19A into the second dielectric film 15B. The first pad 17A can also be formed from a metal which is more diffusible into the second dielectric film 15B than the second pad 17B is diffusible into the first dielectric film 15A.

(49) The low-diffusible metal includes Au, Ag, Al, Ta, Ti, W, Sn, Mo, Ni, In and Co and an alloy containing at least any one of them. These metals are not only less diffusible than copper constituting the wiring conductor 19 but also better in bonding performance than other low-diffusible metals (such as Al and W).

(50) Bonding of the wafers 3A and 3B or bonding of the first and second semiconductor substrates 7A and 7B is accomplished by using the bonding equipment 61 capable of accurate positioning. Moreover, the second pad 17B is wider than the first pad 17A by more than the positioning accuracy. These measures effectively prevent metal diffusion due to positioning errors.

First Modified Example

(51) FIG. 4 is a sectional view showing a first modification of the embodiment shown in FIG. 2A.

(52) According to the embodiment of FIG. 2A, the second pad 17B is formed entirely from a low-diffusible metal. By contrast, according to a first modified embodiment, the second pad 117B has a surface layer which is formed from a low-diffusible metal.

(53) In other words, the second pad 117B has the base part 123, which accounts for a large portion (by volume) thereof, and the base part 123 is formed from the same material (such as copper) as the second wiring conductor 19B. The second pad 117B also has the covering layer 125 which is formed on that side of the base part 123 which faces the first pad 17A. The covering layer 125 is formed from a low-diffusible metal (such as Au).

(54) The structure mentioned above saves expensive low-diffusible metal and contributes to cost reduction.

Second Modified Example

(55) FIG. 5 is a sectional view showing a second modification of the embodiment shown in FIG. 2A.

(56) According to the embodiment of FIG. 2A, either of the first pad 17A or the second pad 17B is formed from a low-diffusible metal. By contrast, according to a second modified embodiment, both of the first pad 217A and the second pad 217B are formed from a low-diffusible metal.

(57) This structure prevents metal from diffusing into the dielectric film 15 even when either of the two pads 217 comes into contact with the dielectric film 15. In other words, this structure prevents metal diffusion from either of the pads 217 even though one of the two pads 217 is not made wider than the other despite the possibility of positioning errors. This permits the design with a high degree of freedom for dimensions. FIG. 5 shows an instance in which the first pad 217A and the second pad 217B have the same area.

Third Modified Example

(58) FIG. 6 is a sectional view showing a third modification of the embodiment shown in FIG. 2A.

(59) According to this modified embodiment, the second dielectric film 15B has the diffusion preventing layer 325 formed thereon which prevents the metal of the first pad 17A from diffusing into the second dielectric film 15B. The diffusion preventing layer 325 may be formed from SiN or SiOC, for instance. Incidentally, the second pad 317B exposes itself toward the first dielectric film 15A through the diffusion preventing layer 325.

(60) In this case, metal diffusion from the first pad 17A is inhibited by the diffusion preventing layer 325. Therefore, as in the case of the second modified embodiment, it is possible to prevent metal from diffusing from the first pad 17A without expanding the second pad 317B. This permits the design with a high degree of freedom for dimensions. FIG. 6 shows an instance in which the first pad 17A and the second pad 317B have the same area.

Fourth Modified Example

(61) FIGS. 7A and 7B are plan views showing a fourth modification of the embodiment shown in FIGS. 2B and 2C.

(62) According to this modified embodiment, the first pad 417A has a round shape. In this case, the first pad 417A overreaches the second pad 17B a little even though its position deviates by rotation about the axis perpendicular to it. The result is positive inhibition of metal diffusion.

(63) The scope of the present invention is not limited to the embodiments mentioned above. It covers other various embodiments.

(64) The foregoing embodiments demonstrate the bonding of wafers to each other. The present invention may be applied to another embodiment, as shown in FIG. 8A, in which the chip 503A is bonded to the wafer 3B or the chips 503A and 503B are bonded together.

(65) Incidentally, the chip-to-wafer bonding will involve a larger positioning misalignment than the wafer-to-wafer bonding. The accuracy of alignment will be larger than ±10 μm in the former case, whereas it will be about ±3 μm. in the latter case. This difference should be taken into account when the area of the pad is designed.

(66) The scope of the present invention is not limited to bonding two wafers (or chips) together. For example, the scope may cover the bonding of three or more wafers together. Each wafer (or chip) may include any type of circuit and is not limited to memory or logic devices. For example, the circuit may be for an imaging device.

(67) The embodiments or modifications thereof disclosed herein may be properly combined with one another. For example, the technology shown in FIG. 4 (in which a low-diffusible metal is used only for the surface of the pad) may be combined with the technology shown in FIG. 5 (in which a low-diffusible metal is used only for the surfaces of the two pads).

(68) It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and alterations may occur depending on design requirements and other factors insofar as they are within the scope of the appended claims or the equivalents thereof.