阅读说明：本技术 用于晶片上芯片处理的分离方法和组件 (Separation method and assembly for chip processing on wafer ) 是由 A·M·贝利斯 B·R·比茨 于 2021-05-28 设计创作，主要内容包括：本公开涉及一种用于晶片上芯片处理的分离方法和组件。本文公开一种用于分离晶片上芯片组合件的半导体管芯堆叠的方法。在一个实例中,分隔壁以图案方式布置在装置晶片的第一表面上,使得所述分隔壁之间的区域限定安装位置。管芯堆叠被安装到所述装置晶片,其中各个管芯堆叠位于所述分隔壁之间的相应安装位置。从与所述装置晶片的所述第一表面相对的第二表面切穿所述装置晶片,并且从所述管芯堆叠之间去除所述分隔壁以在相邻管芯堆叠之间形成空通道。(The present disclosure relates to a separation method and assembly for chip processing on a wafer. A method for separating a semiconductor die stack of a chip-on-wafer assembly is disclosed. In one example, the partition walls are arranged in a pattern on the first surface of the device wafer such that the areas between the partition walls define mounting locations. Die stacks are mounted to the device wafer, with each die stack being located at a respective mounting location between the dividing walls. Cutting through the device wafer from a second surface opposite the first surface of the device wafer, and removing the dividing walls from between the die stacks to form empty channels between adjacent die stacks.)

1. A method for separating a semiconductor die stack of a chip-on-wafer assembly, comprising:

arranging partition walls in a pattern on a first surface of a device wafer such that regions between the partition walls define mounting locations;

mounting die stacks to the device wafer, wherein each die stack is located at a respective mounting location between the partition walls;

cutting through the device wafer from a second surface opposite the first surface of the device wafer; and

removing the partition walls from between the die stacks, thereby forming empty channels between adjacent die stacks.

2. The method of claim 1, wherein the dividing wall is disposed on the first surface of the device wafer prior to mounting the die stack.

3. The method of claim 1, further comprising:

pre-forming the partition walls into the pattern before disposing the partition walls on the first surface of the device wafer; and

adhering the separation wall to the first surface of the device wafer.

4. The method of claim 1, further comprising molding a molding material over a top surface of the die stack and the partition walls, and wherein the cutting through the device wafer is completed after applying the molding material.

5. The method of claim 1, wherein the dividing wall has a height greater than a height of the die stack.

6. The method of claim 1, wherein a width of the dividing walls between the die stacks is less than a distance between the die stacks, the method further comprising molding a molding material over the die stacks and top surfaces of the dividing walls, the molding material extending between the dividing walls and sides of the die stacks.

7. The method of claim 1, wherein the partition walls are water soluble, solvent soluble, or removable using an etching process.

8. The method of claim 1, the method further comprising:

molding a molding material over the top surfaces of the die stack and the partition walls; and

thinning at least the molding material to expose at least a top surface of the partition walls.

9. The method of claim 1, wherein the divider wall forms an integral interface with at least one side of an adjacent die stack.

10. A method for separating a semiconductor die stack of a chip-on-wafer assembly, comprising:

mounting a carrier wafer on the second surface of the device wafer;

mounting a die stack to a first surface of the device wafer, the die stack having channels with cross-lines defining a pattern therebetween;

dispensing divider wall material into the channels between the die stacks to form divider walls; and

cutting through the device wafer from the second surface to the first surface of the device wafer to separate the die stack.

11. The method of claim 10, further comprising removing the partition walls using water, solvent, dry etching, or plasma etching.

12. The method of claim 10, further comprising:

separating the carrier wafer from the device wafer prior to cutting through the device wafer; and

removing the separation wall after cutting through the device wafer.

13. The method of claim 10, further comprising:

applying a carrier film on a top surface of the die stack prior to cutting through the device wafer, the carrier film attached to a dicing frame;

removing the carrier wafer after applying the carrier film to the top surface and before cutting through the device wafer; and

removing the partition walls after removing the carrier wafer.

14. The method of claim 10, further comprising applying a molding material over the top surfaces of the die stack and the partition wall, wherein the molding material does not extend between the die stack and the partition wall.

15. The method of claim 10, further comprising applying a molding material over the top surfaces of the die stack and the partition wall, wherein the molding material forms sides between the die stack and the partition wall.

16. A chip-on-wafer COW assembly, comprising:

a device wafer having a first surface and a second surface opposite the first surface;

die stacks mounted to the first surface of the device wafer in a pattern, wherein individual die stacks are located at respective mounting locations, wherein each of the die stacks includes at least one semiconductor die; and

a partition wall forming an intersection of the pattern between the die stacks on the first surface of the device wafer, the partition wall comprising a partition wall material.

17. The COW assembly of claim 16, wherein the die stacks have sides facing the divider wall, and wherein at least one side of at least one of the die stacks forms an integral interface with the divider wall material.

18. The COW assembly of claim 16, wherein the divider wall material is preformed as the lines of intersection of the pattern prior to adhering to the first surface of the device wafer.

19. The COW assembly of claim 16, wherein the divider wall material is dispensed between the die stacks after the die stacks are mounted to the first surface.

20. The COW assembly of claim 16, wherein the divider wall material comprises a water-soluble divider wall material, a solvent-soluble divider wall material, or an etchable divider wall material.

Technical Field

The present technology relates to semiconductor device packages. More particularly, some embodiments of the present technology relate to techniques for holding a die stack in place to reduce damage to the die during dicing.

Background

Semiconductor dies, including memory chips, microprocessor chips, logic chips, and imager chips, are typically assembled by mounting a plurality of semiconductor dies individually or in a die stack on a substrate in a grid pattern. The mounted die stack is then encased in a polymeric material (e.g., resin) in a wafer-level molding process. As more dies are stacked together to increase capacity, the height of the die stack increases. This can cause the wafer to bend after the molding process, which in turn can cause the die stack to shift position so that they are not properly aligned when diced. Thus, when the stack of die is separated using a rotating blade, chipping or cracking may occur along the edges of the die.

Disclosure of Invention

In one aspect, the present disclosure is directed to a method for separating a semiconductor die stack of a chip-on-wafer assembly, comprising: arranging partition walls in a pattern on a first surface of a device wafer such that regions between the partition walls define mounting locations; mounting die stacks to the device wafer, wherein each die stack is located at a respective mounting location between the partition walls; cutting through the device wafer from a second surface opposite the first surface of the device wafer; and removing the divider walls from between the die stacks, thereby forming empty channels between adjacent die stacks.

In another aspect, the present disclosure is directed to a method for separating a semiconductor die stack of a chip-on-wafer assembly, comprising: mounting a carrier wafer on the second surface of the device wafer; mounting a die stack to a first surface of the device wafer, the die stack having channels with cross-lines defining a pattern therebetween; dispensing divider wall material into the channels between the die stacks to form divider walls; and cutting through the device wafer from the second surface to the first surface of the device wafer to separate the die stack.

In yet another aspect, the present disclosure relates to a Chip On Wafer (COW) assembly, comprising: a device wafer having a first surface and a second surface opposite the first surface; die stacks mounted to the first surface of the device wafer in a pattern, wherein individual die stacks are located at respective mounting locations, wherein each of the die stacks includes at least one semiconductor die; and a partition wall forming an intersection of the pattern between the die stacks on the first surface of the device wafer, the partition wall comprising a partition wall material.

Drawings

Many aspects of the present technology can be better understood with reference to the following drawings. The components in the drawings are not necessarily to scale. Instead, emphasis is placed upon illustrating the principles of the technology.

Fig. 1A is a cross-sectional view of a die stack attached to a device wafer with a separation wall between the die stacks in a chip-on-wafer (COW) assembly in accordance with the present technology.

Fig. 1B is a cross-sectional view of two die stacks of the COW assembly of fig. 1A, in accordance with the present technique.

Fig. 2 is a cross-sectional view of a COW assembly after partition walls have been formed on a surface of a device wafer but before a die stack has been placed on the device wafer, in accordance with the present techniques.

FIG. 3 is a top view of partition walls arranged in a grid pattern in accordance with the present technique.

Fig. 4A illustrates a cross-sectional view of an example of an overmolded COW assembly in accordance with the present techniques.

Fig. 4B illustrates a portion of the overmolded COW assembly of fig. 4A, in accordance with the present technique.

Fig. 5 illustrates a cross-sectional view of an overmolded COW assembly with a portion of the molding material removed, in accordance with the present technique.

Fig. 6 illustrates a cross-sectional view of a COW assembly according to the present technology, in which the partition wall has been removed.

Fig. 7 illustrates a cross-sectional view of the COW assembly inverted from the orientation shown in fig. 6 after removal of the partition walls and the carrier wafer in accordance with the present technique.

Fig. 8 illustrates a cross-sectional view of the COW assembly inverted from the orientation shown in fig. 5, ready for cutting, in accordance with the present technique.

FIG. 9 illustrates the COW assembly of FIG. 8 after the device wafer has been diced from the back side while the dividing walls remain in place, in accordance with the present techniques.

Fig. 10 is a flow diagram of a method for separating a die stack of a COW assembly in accordance with the present technology.

Detailed Description

Specific details of several embodiments for preventing wafer bow and dicing defects of Chip On Wafer (COW) packages are described below. In one example, a method for separating a stack of semiconductor dies in a COW assembly includes arranging partition walls in a pattern (e.g., a grid) on a first surface of a device wafer such that regions between the partition walls define mounting locations. The method also includes mounting each die stack at a respective mounting location such that each die stack is separated from each other by a dividing wall. The die stack may be spaced apart from the divider wall by a gap, and the divider wall may have the same height as the die stack or be taller than the die stack. The die stack and the partition walls are then encased in a polymer material that covers the die stack and fills the gap between the die stack and the partition walls. The dividing walls inhibit bowing of the device wafer and maintain the orientation of the die stack. The die stacks are then separated from each other by etching or dissolving the dividing walls. For example, the partition walls may be made of a material that dissolves in water or another solvent, so that no saw is required to cut the die stack from each other. Thus, COW is expected to have less damage, if any, to the die caused by dicing.

Fig. 1A is a cross-sectional view of a COW assembly 10 having a device wafer 104, partition walls 130 arranged in a grid array on the device wafer 104, and a stack of dies 106 (identified as 106a, 106b, and 106c, respectively) mounted on the device wafer 104 at mounting locations in the region between the partition walls 130. In most applications, hundreds of die stacks 106 are mounted on the device wafer 104 for packaging.

At this stage of packaging the COW assembly 10, the temporary carrier wafer 100 supports and protects the device wafer 104 during processing. More specifically, the second side 136 of the device wafer 104 is attached to the carrier wafer 100 by the adhesive 102. The carrier wafer 100 may be silicon (Si), glass, or other suitable material. The partition walls 130 and die stack 106 are then attached to the device wafer 104. The die stacks 106 may each have several individual semiconductor dies 110. In the embodiment illustrated in fig. 1A, the die stacks 106 may each include four dies 110 (individually identified as dies 110a-d), but it should be understood that the die stacks 106 may have any number of dies (e.g., 2, 3, 4, 5, 6, 7, 8, 10, 12, or more) based on the requirements of the system. Die 110 may be a memory die, including any known type of memory die.

The device wafer 104 may be a silicon wafer having one or more through-silicon vias (TSVs) 107 extending therethrough, and solder balls, posts, and pads (not shown) may be attached to the through-silicon vias (TSVs) 107 at the second surface 136. Thus, the device wafer 104 may serve as an interposer. In other embodiments, device wafer 104 may itself have dies, such as logic dies, processors, or other types of dies for operating memory dies 110.

The die stack 106 may be formed on the device wafer 104 by attaching the die 110a to the device wafer 104, and then stacking the dies 110b-d in series with one another to form the die stack 106 at the appropriate location on the device wafer 104. Alternatively, the dies 110a-d may be stacked together while separated from the device wafer 104, and then the die stack 106 is attached to the device wafer 104.

Fig. 1B shows a portion of the COW assembly 10 including the die stacks 106a and 106B in more detail. The die stack 106 may have a non-conductive film (NCF)120 between the die 110a and the device wafer 104 and between each of the dies 110a-d to adhere the dies 110a-d to the device wafer 104 and to each other. The NCF 120 under die 110a may be the same or different than the NCF 120 between dies 110a-110 d. Other materials, such as underfill materials, may be used in place of NCF 120. The dies 110a-d are also electrically connected to each other and to the device wafer 104 by conductive pillars or bumps (not shown) arranged to correspond to an array of TSVs and/or an array of ball pads between each of the components.

The die stacks 106 are separated from each other by channels 112 (also shown in fig. 1A) that define channel distances. The width of the channels 112 may be uniform to accurately separate the die stacks 106 from one another. In another embodiment, the widths of some of the channels 112 may be different relative to one another.

At this stage of processing the COW assembly 10, the dividing walls 130 may occupy the channels 112 such that one side 142 of a die stack 106a contacts one side of the dividing walls 130, forming an integral interface, while one side 144 of an adjacent die stack 106b contacts the other side of the same dividing walls 130, forming another integral interface. In some embodiments, the separation walls 130 may be formed by dispensing separation wall material into the channels 112 between the die stacks 106 after the die stacks 106 have been mounted to the device wafer 104. For example, the separation walls 130 may be formed in situ on the device wafer 104 using inkjet printing, 3D printing, mask printing, or other suitable processes. In other embodiments, the separation wall 130 may be pre-formed as a complete unit or as a separate wall from the device wafer 104. For example, the sheet of partition wall material may be processed to form a particular pattern of mounting locations having a predetermined size. Alternatively, the divider wall material may be dispensed into a mold and then dried (cured) to have the desired configuration. The pre-formed partition walls may then be adhered to the device wafer 104 prior to mounting the die stack 106 to the device wafer 104. In one embodiment, the separation wall 130 may be glued or adhered to the first surface 138 of the device wafer 104 (fig. 1A) in a single piece or in two or more pieces. In some embodiments, the dividing walls 130 may prevent the NCFs 120 from extending into the channels 112 between adjacent die stacks 106.

The partition wall material may be soluble in water or other solvents, or the partition wall material may be carbon based, silicon (Si), or other material suitable for dry etch removal. An example of a water-soluble partition wall material is HogoMax. Other examples of solvent-soluble partition wall materials are Brewer Science Wafer bound HT-10.10(Brewer Science wave Bond HT-10.10) and Nissan Chemical NAD7009(Nissan Chemical NAD 7009). In further examples, the dry etchable partition wall material may comprise transparent carbon and polyimide. Materials suitable for use as the partition wall material are not limited to these examples.

FIG. 2 is a drawing showingA cross-sectional view of the COW assembly 10 at a processing stage after the partition walls 130 have been formed on the first surface 138 of the device wafer 104, but before the die stack 106 has been placed on the device wafer 104. FIG. 3 is a top view of a grid 140 of partition walls 130 according to the present technique. Referring to fig. 2 and 3 together, the grid 140 of partition walls 130 is defined to have a first dimension W configured to receive the die stack 106_S1(FIG. 3 only) and a second dimension W_S2Mounting area 146. In general, the mounting region 146 is configured to receive one or more die stacks 106. For example, the mounting region 146 may be composed of straight lines (e.g., square or rectangular) to accommodate the footprint of the die stack 106.

The grid pattern 140 may be formed by a plurality of cross-wires configured to reside in the channels 112 between the die stacks 106. In one embodiment, as shown in FIG. 3, substantially parallel lines extending along the first direction 174 may intersect substantially parallel lines extending along the second direction 176 at a 90 degree angle. The partition wall 130 may have a width W_W(see fig. 2), which may be wide enough to contact the sides of adjacent die stacks 106, as previously discussed. In another embodiment, the width W_WMay be smaller than the width of the channels 112 (fig. 1B) between adjacent die stacks 106, leaving a space or gap between the die stacks 106 and the partition walls 130.

The separation walls 130 may have a height H corresponding to the die stack 106 (fig. 1B)_sHeight H of_w(see FIG. 2). In some embodiments, the divider wall height H of the divider wall 130_wMay be equal to die stack height H_sWhile in other embodiments the divider wall height H_wMay be less than or greater than the die stack height H_s。

In some embodiments, as described above, the lattice 140 of partition walls 130 may be a preformed component that is cut, molded, built from layers, or otherwise formed separately to provide the lattice pattern 140. The preformed grid 140 may be adhered to the device wafer 104 prior to attaching the die stack 106 to the device wafer 104, or alternatively, the preformed grid 140 may be attached to the device wafer 104 after some or all of the die stack 106 are in place. The preformed grid 140 may be a single piece or multiple pieces.

Fig. 4A shows a cross-sectional view of another example of an overmolded COW assembly 10 according to the present technique after wafer level molding. In fig. 1A-4B, like reference numerals refer to like parts. In this example, the molding material 150 is molded over the die stack 106 and the partition walls 130. The molding material 150 may cover the top surfaces 158 of the die stack 106 and the top surfaces of the partition walls 130, as well as the sides of the die stack 106. Accordingly, the molding material 150 may have exterior sides 152 and 154 and a top 156, respectively.

As shown in more detail in FIG. 4B, the divider wall width W_WLess than the channel width W of the channels 112 between the die stacks 106_LSo that there is a gap between the sides of the die stack 106 and the dividing wall 130. Accordingly, during the molding process, some of the molding material 150 may flow in the gaps between the sides of the die stack 106 and the partition walls 130 to form side portions 157 (individually labeled 157a and 157b) of the molding material 150.

Fig. 5 illustrates a cross-sectional view of the overmolded COW assembly 10 after a portion of the molding material 150 has been removed to expose the partition walls 130, in accordance with the present technique. The molding material 150 above the uppermost die 110d may be removed while leaving at least a portion of the molding material 150 along the outer sides 152 and 154. The molding material 150 may be completely removed from the top of the die stack 106, or a thin layer of molding material 150 may remain on top of the uppermost die 110 d. In some embodiments, a small amount of the uppermost die 110d may also be removed. This thinning process may be used to achieve a desired thickness or height H of the die stack 106_SAnd the partition wall 130 is exposed. The molding material 150 may be removed using back grinding, chemical mechanical planarization, or other suitable processes. The separation walls 130 between the die stacks 106 may inhibit flexing or bending of the COW assembly 10 during the overmolding process and when the COW assembly 10 cools after the overmolding process.

Fig. 6 shows a cross-sectional view of the COW assembly 10 after the partition walls 130 have been removed to open the channels 112 between the die stacks 106 to form empty channels between adjacent die stacks 106. The dividing walls 130 may be removed before the COW assemblies 10 are attached to the dicing frame and removed from the carrier wafer 100. The divider wall 130 may be removed without sawing or mechanically cutting the divider wall material. For example, water or other solvents may be used to dissolve the partition wall material in a wet process or wet clean. For example, a solvent may be used to remove the photosensitive material or temporary binder. In other embodiments, if the divider wall material is an organo-carbon based material, plasma etching may be used to remove the divider wall material. Other materials may be used for the divider wall material, such as laser ablated material. In each of these cases, the dividing wall 130 is chemically or thermally removed without a rotating blade or other mechanical cutting device. Thus, the die stack can be separated into individual units without chipping or cracking the edges of the die 110.

Fig. 7 shows a cross-sectional view of the COW assembly 10 in an opposite orientation to that in fig. 5 and 6. At this stage of packaging the COW assembly 10, a tape or carrier film 200 supported by a dicing frame 202 has been attached to the top surface 178 of the die stack 106, and the carrier wafer 100 has been removed. The device wafer 104 is then diced from the back side 180 (e.g., the upwardly facing surface in fig. 7) to form spaces 212 that are at least substantially aligned with the channels 112 between the die stacks 106. The spacers 212 extend through the thickness "T" of the device wafer 104 to completely separate the die stacks 106 from one another. The device wafer 104 may be diced using a laser, a rotary saw, or other suitable technique to cut through the thickness T of the device wafer 104 without penetrating the die 110 a. Thus, when the device wafer 104 is cut, the blade does not cut alongside the die stack 106. Because the dividing wall 130 has been removed, in some embodiments, no dicing is required below the thickness T of the device wafer 104. Accordingly, it is desirable for the methods of the present technology to reduce chipping of the dies 110 a-d.

Fig. 8 and 9 show another embodiment of the stage shown in fig. 6 and 7, in which the COW assembly 10 shown in fig. 5 has been attached to the carrier film 200 and the carrier wafer 100 has been removed before the separation walls 130 have been removed. As shown in fig. 8, the COW assembly 10 is in the opposite orientation to that shown in fig. 5, and the separation walls 130 and the die stack 106 contact the carrier film 200. Fig. 9 shows the COW assembly 10 after cutting the device wafer 104 from the back side 220 (e.g., the upward facing surface in fig. 9) to form the spaces 212 at least substantially aligned with the dividing walls 130. The device wafer 104 may be cut as described above, such as by using a laser or a rotating blade, to cut only through the thickness T of the device wafer 104. After the spacers 212 are formed in the device wafer 104, the partition walls 130 may be removed using a wet process or a dry etch to open the channels 112 between the die stacks 106 (e.g., similar to the channels 112 shown in fig. 7). When the divider wall 130 is made of a water soluble material, the wet process will also clean debris from the surface 222 of the device wafer 104 caused by cutting the device wafer 104. Thus, final cleaning of the COW assembly 10 may be accomplished at the dicing tool. This embodiment provides the advantage that no additional cleaning step is added.

One expected advantage of the embodiment shown in fig. 8 and 9 is that holding the dividing walls 130 until the device wafer 104 is cut into COW assemblies 10 provides additional structural support to avoid bending or warping. It is also desirable that the additional structural support provided by the dividing walls 130 prevent or at least inhibit cracking and/or breaking of the device wafer 104 during peeling of the carrier wafer 100. In addition, the dividing walls 130 may also protect the sides of the die stack 106 from cracking or chipping while the device wafer 104 is being diced.

Fig. 10 is a flow diagram of a method 1000 for processing a COW component 10 according to the present technique. The method 1000 includes applying the dividing walls 130 and attaching the die stack 106 to the device wafer 104 (block 1010). The separation walls 130 may be applied to the device wafer 104 before or after attaching the die stack 106 to the device wafer 104. The molding material is then molded over the die stack 106 and the device wafer 104 (block 1020), and an upper portion of the molding material 150 is thinned, such as by back grinding (block 1030). After thinning the molding material, there are two options. Option 1 involves removing the divider wall material (block 1040) and then dicing the device wafer 104 from the back side (block 1042). Option 2 alternatively includes attaching the COW assembly 10 to the carrier strip 200 and removing the carrier wafer 100 (block 1050), dicing the device wafer 104 from the back side before removing the partition walls 130 (block 1052), and then removing the partition walls 130 (block 1054).

It is not intended to be exhaustive or to limit the present technology to the precise form disclosed herein. Although specific embodiments have been disclosed herein for purposes of illustration, various equivalent modifications are possible without departing from the technology, as those skilled in the relevant art will recognize. In some instances, well-known structures and functions have not been shown or described in detail to avoid unnecessarily obscuring the description of the embodiments of the technology. Although the steps of a method may be presented herein in a particular order, alternative embodiments may perform the steps in a different order. Similarly, certain aspects of the technology disclosed in the context of particular embodiments may be combined or removed in other embodiments. Moreover, while advantages associated with certain embodiments of the technology have been disclosed in the context of those embodiments, other embodiments may also exhibit such advantages, and not all embodiments necessarily exhibit such advantages or other advantages disclosed herein, which fall within the scope of the technology. Accordingly, the present disclosure and related techniques may encompass other embodiments not explicitly shown or described herein.

Throughout this specification, the singular terms "a" and "the" include plural referents unless the context clearly dictates otherwise. Similarly, unless the word "or" is expressly limited to mean only a single item exclusive from other items when referring to a list of two or more items, the use of "or" in such a list is to be interpreted as encompassing (a) any single item in the list, (b) all items in the list, or (c) any combination of items in the list. Additionally, the term "comprising" is used throughout to mean including at least the features described, such that any further number of additional types of the same and/or other features is not excluded. Reference herein to "one embodiment," "some embodiments," or similar language means that a particular feature, structure, operation, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present technology. Thus, appearances of such phrases or formulations herein are not necessarily all referring to the same embodiment. Furthermore, the various particular features, structures, operations, or characteristics may be combined in any suitable manner in one or more embodiments.

From the foregoing, it will be appreciated that specific embodiments of the technology have been described herein for purposes of illustration, but that various modifications may be made without deviating from the scope of the invention. The technology is not limited except as by the appended claims.