FACILITATING CHIP DICING FOR METAL-METAL BONDING AND HYBRID WAFER BONDING

A method of forming a stacked assembly of semiconductor chips can include juxtaposing and metallurgically joining kerf metal elements exposed in kerf regions of a first wafer with corresponding kerf metal elements exposed in kerf regions of a second wafer, and affixing undiced semiconductor chips of the first wafer with corresponding undiced semiconductor chips of the second wafer. The assembled wafers are then cut along the dicing lanes thereof into a plurality of individual assemblies of stacked semiconductor chips, each assembly including an undiced semiconductor chip of the first wafer and an undiced semiconductor chip of the second wafer affixed therewith.

Skip to: Description  ·  Claims  · Patent History  ·  Patent History
Description
DOMESTIC PRIORITY

This application is a divisional of U.S. patent application Ser. No. 14/096,325, filed Dec. 4, 2013, the disclosure of which is incorporated by reference herein in its entirety.

BACKGROUND

The present invention relates to a method for performing direct wafer-to-wafer bonding, including wafer multi-stacking, through metal features such as metal pads, lines, or patterns, on surfaces of the wafer to one another, prior to severing the joined wafers into individual units each containing a chip of a wafer and a chip of another wafer bonded to the wafer.

Assemblies of vertically stacked semiconductor chips having direct vertical electrical interconnections using through silicon vias (TSV) offer improvements in integration density and speed of information access.

Wafer-to-wafer bonding techniques can be used to join wafers together in vertically stacked wafer assemblies, which can then be diced into individual stacked semiconductor chip assemblies containing stacks of two or more semiconductor chips each. Each individual stacked semiconductor chip assembly may have through silicon vias extending in a vertical direction of the assembly for electrically connecting the chips therein. Wafer-to-wafer bonding techniques include metal-to-metal bonding in which flat metal elements, typically a plurality of discrete metal pads, at a surface of one wafer are joined with corresponding flat metal elements at a surface of a second wafer.

Metal-to-metal bonding techniques are subject to variations in bond strength between wafers due to misalignment between the metal elements of one wafer relative to the metal elements of another wafer to which they are to be bonded. Such misalignment can result in metal to oxide contact in a non-controlled fashion, or metal elements of one wafer not bonding with those of the other wafer.

Hybrid bonding, in which metal elements of respective wafers bond together as well as oxide elements of the respective wafers, leave large portions of the respective wafers unbonded, which include among others, kerf regions disposed between adjacent undiced semiconductor chips of a wafer. The unbonded areas can provide a potential source of cracking or chipping defects when dicing a stacked wafer assembly into a plurality of individual stacked semiconductor chip assemblies, especially in the case of wafer multi-stacking.

Despite these existing ways of joining wafers to one another, further improvements can be made.

SUMMARY

According to an aspect of the invention, a method of forming a stacked assembly of semiconductor chips includes metallurgically joining kerf metal elements exposed in a kerf region of a first wafer with corresponding kerf metal elements exposed in kerf regions of a second wafer, and affixing semiconductor chips of the first wafer with corresponding semiconductor chips of the second wafer which they face. The assembled wafers are then cut along the dicing lanes thereof into a plurality of individual assemblies of stacked semiconductor chips, each assembly including a semiconductor chip of the first wafer and a semiconductor chip of the second wafer affixed therewith. The simultaneous joining of such kerf metal elements, and optionally crackstop elements, which can be done simultaneously or at a different time than the joining of electrical connection elements for inter-chip connections between respective chips of the assembly, provides mechanical support to the assembly during the subsequent cutting into individual stacked chip assemblies.

In accordance with another aspect of the invention, a wafer subassembly is provided for use in forming a plurality of individual assemblies of stacked semiconductor chips. The wafer subassembly may include a semiconductor wafer and a plurality of kerf metal elements disposed in kerf regions at the surface of the wafer, the kerf regions disposed in dicing lanes between adjacent undiced semiconductor chips of the wafer, the kerf metal elements being configured for direct metal to metal bonding with corresponding kerf metal elements of a second wafer subassembly to form a stacked wafer assembly.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a sectional view illustrating first and second wafer assemblies each having metal elements at corresponding surfaces thereof.

FIG. 2 is a sectional view illustrating a joined stacked wafer assembly formed by joining the first and second wafer assemblies shown in FIG. 1.

FIG. 3 is a plan view illustrating a wafer corresponding to the stacked wafer assembly shown in FIG. 2.

FIG. 4 is a detailed partial plan view further illustrating a portion of the wafer shown in FIG. 3.

FIG. 5 is a plan view illustrating a structure of a wafer in accordance with a variation of the embodiment shown in FIG. 3.

FIG. 6 is a detailed partial plan view further illustrating a portion of the wafer shown in FIG. 5 in accordance with one variation thereof.

FIG. 7 is a detailed partial plan view further illustrating a portion of the wafer shown in FIG. 5 in accordance with one variation thereof.

FIG. 8 is a detailed partial plan view further illustrating a portion of the wafer shown in FIG. 5 in accordance with one variation thereof.

FIG. 9 is a plan view illustrating a structure of a wafer in accordance with a variation of the embodiment shown in FIG. 3.

FIG. 10 is a detailed partial plan view further illustrating a portion of the wafer shown in FIG. 9.

DETAILED DESCRIPTION

As will be described further below, first and second wafers can be assembled with one another by metallurgically joining metal elements in the kerf and in the chips. The elements in the chip serve as primarily electrical connections, while the elements in the kerf serve primarily as mechanical support and additional crack-stop and delamination protection structures. The kerf elements are disposed in dicing lanes of each wafer with corresponding kerf metal elements of the other wafer and affixing undiced semiconductor chips of the first semiconductor wafer with corresponding undiced semiconductor chips of the second wafer. Then the assembled wafers are cut along the dicing lanes into a plurality of assemblies of stacked semiconductor chips. Cutting the assembled wafers with a saw along the dicing lanes typically removes the kerf metal elements such that they are absent from the individual assemblies of stacked semiconductor chips. However, some of these additional kerf structures may also remain on the chip.

The method can be used in conjunction with assembly techniques in which chip metal elements disposed in undiced semiconductor chips regions of a first wafer are joined with corresponding chip metal elements of a second wafer which they face. In another example, the method can be used in conjunction with assembly techniques in which oxide elements in undiced semiconductor chips regions of a first wafer face and are joined with corresponding oxide elements of a second wafer. The wafers may be thinned prior to joining For example, with the wafers shown in FIG. 1, the thickness of the first wafer 100 in a direction orthogonal to its surface 101 can be less than 100 microns between opposite surfaces of the first wafer. The corresponding thickness of the second wafer can be less than 100 microns. Referring to FIG. 1, first and second wafers 100, 200 each having metal elements at corresponding surfaces thereof, are shown in a state prior to the metal elements of the first wafer being metallurgically joined with corresponding metal elements of the second wafer. First wafer 100 includes undiced semiconductor chips 110 and 112 which are integral portions of a semiconductor wafer, as well as kerf region 120 being another integral portion of the first semiconductor wafer disposed between the undiced semiconductor chips 110, 120. “Edges”, i.e., boundaries of each of the undiced semiconductor chips 110, 112 with the adjacent kerf region 120 of the first wafer, are shown at 111 and 113, respectively. These boundaries 111, 113 may be somewhat fluid as the width of the kerf region, and hence the exact location of the boundaries can depend upon the width of the cut produced by a saw used to sever the semiconductor chips from one another. In the wafer 100 shown in FIG. 1, a bulk semiconductor layer of the first wafer 100 can extend continuously within each of the regions occupied by the undiced first and second semiconductor chips and the kerf region 120. The second wafer 200 has a similar structure of undiced semiconductor chips 210 and 212 which are integral portions of a semiconductor wafer, as well as kerf region 220 being another integral portion of the first semiconductor wafer disposed between the undiced semiconductor chips 210, 220. Metal crack stop barriers 130, 230 of each chip are further shown in FIG. 1, these being within the areas of the undiced semiconductor chips 110, 112 of the second wafer adjacent to the boundaries 111, 113 thereof. The chips in FIGS. 1 and 2 are shown as aligned and/or bonded in a slightly misaligned state. The slight misalignment that results is not intentional, but rather a tolerable result of processing in accordance with the embodiments of the invention.

The first wafer 100 further includes a plurality of kerf metal elements 122 in the kerf region 120 at a surface 101 of the first wafer which face corresponding kerf metal elements 222 in the kerf region 220 at a surface 201 of the second wafer. In addition, a plurality of chip metal elements 132 at surfaces of the undiced semiconductor chips 110, 112 of the first wafer can be facing corresponding chip metal elements 232 at surfaces of the undiced semiconductor chips 210, 212 of the second wafer. Typically, the kerf metal elements and the chip metal elements are metal pads which may be formed of copper and may project above a surface of a dielectric layer, e.g., an oxide layer, at the surface of the wafer. In particular examples, the kerf metal elements and the chip metal elements can be passivated with a bonding agent or defluxing agent such as BTA, or may be coated with a metal barrier such as, for example, a barrier or manganese and manganese oxide (e.g., Mn/MnO).

FIG. 2 illustrates a further stage of processing in which the kerf metal elements 122, 222 of the first and second wafers, respectively, are metallurgically joined with one another. Typically, the kerf metal elements of each wafer are joined with those of the other wafer by direct metal-to-metal bonding. In some cases, joining can be performed using one or more of heat, pressure, and typically without a solder being disposed between the surfaces of the corresponding joined kerf metal elements of each wafer.

Undiced semiconductor chips of the first wafer 100 are also affixed with corresponding undiced semiconductor chips of the second wafer 200. In one example, corresponding chip metal elements 132, 232 of the first and second wafers are joined to one another. Typically, and the chip metal elements of each wafer are joined with one another by direct metal-to-metal bonding. In another example, undiced semiconductor chips of the respective wafers can be joined with one another using oxide elements at the facing surfaces of the respective wafers. In one example, the affixing of the semiconductor chips and the metallurgically joining are performed simultaneously.

FIG. 3 further illustrates the corresponding wafer assembly in which kerf regions 120 which extend in a horizontal layout direction of the wafer 100 are seen therein being generally coextensive with horizontal dicing lanes, and kerf regions 120′ which extend in a vertical layout direction of the wafer 100 and are generally coextensive with vertical dicing lanes. Kerf metal elements in FIG. 3 extend full widths of chords which traverse the generally circular area of the wafer 100. Further detail of the structure shown in FIG. 3 can be seen in the detailed partial plan view provided in FIG. 4.

Referring to FIG. 3, in one example, the kerf metal elements of each wafer may include at least one continuous kerf metal element 322 which has a first portion extending in a first direction of a length of a first dicing lane 120′ parallel to an edge of a first undiced semiconductor chip 312 and parallel to an edge of a second undiced semiconductor chip 312′, and a second portion extending in the first direction, the second portion crossing a second dicing lane 120 that extends in a second direction between the first undiced semiconductor chip 312 and the second undiced semiconductor chip 322.

After joining the kerf metal elements and affixing the undiced semiconductor chips of the respective wafers with one another, the resulting stacked wafer assemblies are cut along the dicing lanes represented by kerf regions 120, 120′ to form a plurality of individual stacked semiconductor chip assemblies. The cutting process, e.g., sawing process, can result in the removal the kerf metal elements 122 such that the kerf metal elements may be absent from the stacked semiconductor chip assemblies.

The joined kerf metal elements can provide additional support to the chips of each wafer when dicing the stacked wafer assembly into the individual stacked assemblies of chips and also serve as additional crack-stops and delamination protection. This support can be especially advantageous when, as is common, the wafers have been thinned prior to being joined such that each wafer has a thickness 100 microns or less between its opposite major surfaces. FIG. 5 illustrates a variation of the embodiment shown in FIG. 3, in which the kerf metal elements 522 are provided in the respective kerf regions 120 as a plurality of discrete metal elements which are disposed adjacent to each of the undiced semiconductor chips 100 of each wafer. FIGS. 6, 7, and 8 provide a more detailed view of the kerf metal elements of FIG. 5 for the portion of FIG. 5 in the dashed outline therein. FIG. 6 depicts a particular example in which the length and width dimensions of each kerf metal element 622 in the horizontal and vertical layout directions of the wafer are about the same. FIG. 7 depicts another example in which the length dimension of each kerf metal element 722 is much greater than the width dimension of each kerf metal element. FIG. 8 depicts yet another example in which the width dimension of each kerf metal element 822 can be greater than the width dimension of each kerf metal element.

FIGS. 9 and 10 illustrate another variation in which the kerf metal elements include individual rectangular frame elements 922. In this case, each individual rectangular frame element 922 surrounds an individual undiced semiconductor chip 100 of one of the wafers. In addition to the rectangular frame elements, the kerf metal elements of each wafer can also include discrete metal elements 924 provided as individually separated pads in the kerf region between the rectangular frame elements 922.

While the invention has been described in accordance with certain preferred embodiments thereof, those skilled in the art will understand the many modifications and enhancements which can be made thereto without departing from the true scope and spirit of the invention, which is limited only by the claims appended below.

Claims

1. A wafer subassembly for use in forming a plurality of individual assemblies of stacked semiconductor chips, comprising:

a semiconductor wafer and a plurality of kerf metal elements disposed in kerf regions at the surface of the wafer, the kerf regions disposed in dicing lanes between adjacent undiced semiconductor chips of the wafer, the kerf metal elements being configured for direct metal-to-metal joining with corresponding kerf metal elements of at least one other wafer subassembly to form a stacked wafer assembly.

2. The wafer subassembly of claim 1, wherein the wafer subassembly has a thickness of less than 100 microns between opposite exposed surfaces thereof.

3. The wafer subassembly of claim 1, wherein the kerf metal elements include individual rectangular frame elements, each individual rectangular frame element surrounding an individual undiced semiconductor chip of one of the wafers.

4. The wafer subassembly of claim 1, wherein the kerf metal elements of each wafer include at least one continuous kerf metal element which has a first portion extending in a first direction of a length of a first dicing lane parallel to an edge of a first undiced semiconductor chip and parallel to an edge of a second undiced semiconductor chip, and a second portion extending in the first direction, the second portion crossing a second dicing lane that extends in a second direction between the first undiced semiconductor chip and the second undiced semiconductor chip.

Patent History
Publication number: 20150255417
Type: Application
Filed: May 20, 2015
Publication Date: Sep 10, 2015
Inventors: Mukta G. Farooq (Hopewell Junction, NY), Erdem Kaltalioglu (Newburgh, NY), Wei Lin (Albany, NY), Spyridon Skordas (Wappingers Falls, NY), Kevin R. Winstel (East Greenbush, NY)
Application Number: 14/716,959
Classifications
International Classification: H01L 23/00 (20060101); H01L 23/544 (20060101);