CROSS REFERENCE TO RELATED APPLICATIONS This application is a Divisional of, claims priority to, and incorporates by reference U.S. patent application Ser. No. 13/947,235 filed Jul. 22, 2013.
BACKGROUND Microelectromechanical systems (MEMS) is the technology of very small devices, typically within a size range of about 1 μm to 1 mm. MEMS devices are produced with semiconductor fabrication technology. There are currently many types of MEMS devices used in many different applications, for example, microphones, pressure sensors, and mechanical resonators.
One common type of MEMS device is a BAW (Bulk Acoustic Wave) filter. BAW filters are electromechanical devices that can be implemented as ladder or lattice filters. BAW filters typically operate at frequencies from around 2 to around 16 GHz, and may be smaller or thinner than equivalent SAW filters. The two main variants of BAW filters are Thin Film Bulk Acoustic Resonator or FBAR and Solid Mounted Bulk Acoustic Resonator or SMR. BAW filters are frequently used in wireless applications. Various methods of designing and making BAW filters are known in the art. BAW filters are typically implemented at the “wafer level.” Identical BAW structures are formed on a large semiconductor substrate (wafer) that is subsequently cut (singulated) into identical separate dies (dice).
BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a top plan view of a first semiconductor wafer.
FIG. 2 is a top plan view of a second semiconductor wafer.
FIG. 3 is a detail cross sectional elevation view of a portion of the first semiconductor wafer showing the location of a BAW structure and bond pad.
FIG. 4 is a detail cross sectional elevation view of a portion of the second semiconductor wafer showing a cavity and void formed therein.
FIG. 5 is a detail cross sectional elevation view showing portions of the second semiconductor wafer mounted on portions of the first semiconductor wafer.
FIG. 6 is a cross sectional elevation view similar to FIG. 5, but with a thinned first semiconductor wafer and showing saw streets.
FIG. 7 is a cross sectional elevation view like FIG. 6, except showing a hole formed in a portion of the first semiconductor wafer below the BAW structure.
FIG. 8 is a cross sectional elevation view of an integrated circuit package formed using a singulated portion of the two attached semiconductor wafer portions shown in FIG. 7.
FIG. 9 is a flow chart of a method of forming a wafer level package.
DETAILED DESCRIPTION In this specification directional terms such as up, down, vertical, lateral and the like are used in a relative sense to describe the relationship between objects, parts, surfaces, etc., usually with reference to a drawing figure. Such terms, unless expressly indicated otherwise, do not indicate any particular orientation of the object being described with respect to the earth.
This specification, in general, discloses a wafer level assembly that includes, as best illustrated by FIG. 6, a first semiconductor wafer substrate 10 having a top surface 14 and a bottom surface 16 (17 after thinning) with a BAW structure 18 formed on the top surface 14. The assembly includes a second semiconductor wafer substrate 30 having a top surface 34 and a bottom surface 36 and having a cavity 38 with an opening 40 in the bottom surface 36. The bottom surface 36 of the second substrate 30 is attached to the top surface 14 of the first substrate 10 and the BAW structure 18 is positioned inside a cavity 38.
Also, this specification, in general, discloses, as illustrated by FIG. 9 a method of forming a wafer level package. The method includes, as shown at block 122, fabricating a BAW structure on a first semiconductor wafer substrate. The method also includes, as shown at block 124, forming a cavity in a second semiconductor wafer substrate. As shown at block 126, the method further includes mounting the second substrate on the first substrate such that the BAW structure is positioned inside the cavity in the second substrate.
Having thus generally described a wafer level assembly embodiment and a method of forming a wafer level package, details thereof and various embodiments of such assemblies and methods will now be described in detail.
FIG. 1 illustrates a first semiconductor wafer 10. The wafer has a plurality of identical, integrally connected wafer portions 12 that each contain identical structure and circuitry. These identical portions 12 are later separated in a die forming process. FIG. 2 is a top plan view of a second wafer 30 which also has a plurality of identical, integrally interconnected wafer portions 32.
FIG. 3 is a detail cross sectional view of a portion of wafer 10. Wafer 10 has a generally flat top surface 14 and a generally flat bottom surface 16. A MEMS structure, such as bulk acoustic wave structure (“BAW”) 18, has been formed at the top of the wafer 10, as by conventional processes well known in the art. BAW structure 18 has a top surface 20 and a bottom surface 22. The bottom surface 22 interfaces with and is integrally connected to the wafer top surface 14. A bond pad 19 is formed on top surface 14 adjacent to the BAW structure 18. The first wafer 10 has an initial thickness “a”. Thickness “a” is typically in a range of between 400 μm and 1000 μm.
FIG. 4 is a detail cross sectional view of a portion of the second wafer 30 including wafer portions 32, which are later separated in a die forming process. The original configuration of one of the wafer portions 32, prior to etching operations, is illustrated in dash lines in FIG. 4. The second wafer 30 has a generally flat top surface 34 and a generally flat bottom surface 36 A cavity 38 having a generally rectangular cross section is formed in the second wafer 30 as by etching bottom surface 36. The cavity 38 has a cavity opening 40. Cavity side walls 44, 46 extend generally perpendicular to a generally flat cavity top wall surface 42. A lateral void 50 may also be formed on the second wafer 30. The void 50 may be formed within each wafer portion 32 such that it completely eliminates an entire lateral section of each wafer portion 32. The second wafer 30 may have a height “b” which is typically in a range of between about 200 μm and 1000 μm.
In FIG. 5, the second wafer 30 has been mounted on top the first wafer 10 with integrally connected second wafer portions 32 positioned directly above integrally connected first wafer portions 12. When thus situated, the BAW structure 18 on the first wafer 10 is positioned inside the cavity 38 of the second wafer 30. This attachment process may be done in a controlled environment such as a vacuum or a controlled gas environment such as nitrogen, in order to carefully control the contents of the cavity 38. Also, the die pad 19 on the first wafer 10 is positioned within the lateral void 50 of the second wafer 30 and is thus exposed. The generally flat bottom surface 36 of the second wafer 30 interfaces with the top surface 14 of the first wafer 10. The two surfaces are fixedly attached as by a wafer-to-wafer bonding process, which joins the dielectric portions of the wafers. This attachment process, in one embodiment, does not include any electrical interconnect. There are many available methods known in the art to perform such bonding. See, for example, U.S. Pat. No. 6,603,847 of Yu Chen, et al. issued Jan. 7, 2003, which is hereby incorporated by reference for all that it discloses. In one method embodiment, the assembly shown in FIG. 5 is formed prior to a thinning operation that is performed on the bottom surface 16 of the first wafer 10. The thickness and configuration of the attached second wafer 30 helps to stabilize the first wafer 10 while it is being thinned.
FIG. 6 illustrates the assembly shown in FIG. 5 after a thinning operation. The position of the original bottom surface 16 of the first wafer 10 is illustrated in dash lines in FIG. 6. The first wafer 10 has been thinned as by either or both of mechanically grinding and chemically etching surface 16 to provide a new bottom surface 17. The height “a” of the thinned first wafer 10 may be from about 50 μm to 400 μm. The new surface 17 may also be a generally planar surface. This thinning process may take place either before or after the first and second wafers 10, 12 are attached. However, as mentioned above, when the second wafer 30 is attached to the first wafer 10 before thinning, it stabilizes the first wafer 10 during thinning. After thinning the height a′ of the first wafer 10 may be less than the height “b” of the second wafer 30, for example, it may be between about 100 μm to 400 μm thick. Thinning a wafer to such a low thickness presents challenges, which may include warping and/or cracking. Therefore the second wafer 30 may be designed in a way, as illustrated by FIGS. 5 and 6, which provides mechanical stability during the process of thinning the first wafer 10 to the desired thickness. For example, the ratio of the attached area of the second wafer 30 to the total top surface area of the first wafer 10 may be between about 25% and 75%; the ratio of the thickness of the second wafer 30 to the thickness of the first wafer 10 may be between about 1:2 and 3:1 before grinding and between about 3:1 and 10:1 after grinding; and the total thickness of the combined wafers after grinding may be between about 200 μm and 600 μm. Design considerations include which parts of each second wafer portion 32 are to be bonded to first wafer portions 12. Another consideration is the positioning of these connected parts of wafer portions 32 with respect to singulation cuts 70. In the illustrated embodiments, the second wafer 30 may have no circuitry and may be used for forming an enclosure around the BAW structure 18 and for increasing the combined wafer 10/30 thickness for providing mechanical stability during thinning of the first wafer 10.
As illustrated in dash lines in FIG. 6, singulation cuts 70 are made to separate the various portions 12/32 of the attached wafers into rectangular dice 62. The singulation cuts 70 generally follow the rectangular grid saw streets shown in FIGS. 1 and 2. The singulation cuts 70 may be made by saw cutting, punch cutting, stealth laser fracturing or other means. Each of the attached bottom and top wafer portions 12/32 form a separate die 62.
FIG. 7 shows a second structural embodiment of the combined wafers 10/30, which is identical to that shown in FIG. 4 except that a cavity 56 has been provided below each BAW structure 18 exposing a portion of the BAW bottom surface 22. The cavity 56 may be formed by etching or other means. The cavity 56 has an opening 58 at the lower surface 17 of the first wafer 10. Thus a plurality of dies 64 having a slightly different structure than the dies 62 of FIG. 6 may be provided. The dies 64 may be singulated along saw streets 70.
As illustrated by FIG. 8, an IC package 80 may be provided by connecting a die such as die 62 or die 64 to a lead frame 82 and encapsulating the die 62 or 64 and the lead frame 82 in a layer of encapsulant 96. In FIG. 8, the lead frame 82 comprises a die attachment pad portion 84 and a plurality of surrounding lead portions 86. A die attachment film 88, such as an epoxy resin filled with a filler material such as silver, may be used to attach the die 64 to the die attachment pad portion 84 of a lead frame 82. This epoxy may be applied using any die attach techniques known in the art, such as dispensing by a needle to the leadframe 82, by film attachment to the back surface 12 of first wafer 10, or by screen printing to leadframe 82 or to first wafer 10. This epoxy may be any suitable material for die attachment to a leadframe, and may have a range of possible chemical formulations, as is known in the art.
A bond wire 90 has a first end 92 that may be welded to the contact pad 19 on the die 62/64 as by a ball bond. A second end 94 of the bond wire 90 may be welded to one of the lead portions 86 by a second weld such as a stitch bond. In one embodiment, the bond wire 90 has a low wire loop 91, which significantly reduces inductance interference in the BAW structure 18. As used herein “low wire loop” means a loop profile that has a height less than the thickness (height) of the associated die 64 so that the overall thickness of the package mold 96 is not limited by the wire loop profile. Loop profiles of 25 to 100 microns can be achieved with traditional ball bonding or with techniques such as “reverse bonding” where a stitch bond is placed on a small ball bond which is first placed on the die. These techniques are well known in the art.
Methods for forming wafer level packages and different embodiments of integrated circuit packages and wafer level assemblies therefor have been described in detail herein. Various alternative embodiments will occur to those skilled in the art after reading this disclosure. For example, multiple bond pads could be provided on each second wafer die 32, multiple bond wires could be connected between each bond pad 19 and lead 86 and various combinations of die pads 19, leads 86 and bond wires 90 could be used. Similarly multiple MEMS structures, such as BAW structures 18 or others, could be provided on each first wafer portion 12 and such multiple MEMS structures could be contained in one or multiple cavities in second wafer portions 32. The appended claims are intended to be broadly construed to cover all such alternative embodiments, except to the extent limited by the prior art.
