Semiconductor devices and methods of making and using solder bumps on fo-wlp

Abstract

Semiconductor devices and methods of making and using solder bumps on FO-WLPs. A semiconductor device has a semiconductor die and an encapsulant deposited over the semiconductor die. An interconnect structure is formed over the semiconductor die and encapsulant. A first under bump metallization (UBM) pad is formed over the interconnect structure. The first UBM pad is non-circular. A plate film is disposed over the first UBM pad. The plate film includes a first opening over the first UBM pad, wherein a footprint of the first opening is larger than a footprint of the first UBM pad. A solder paste is deposited in the first opening. The solder paste is reflowed to form solder bumps on the UBM pads.

Description

TECHNICAL FIELD

[0001] The present disclosure relates generally to semiconductor devices, and more particularly, to semiconductor devices and methods of making and using solder bumps having fan-out wafer level packaging (FO-WLP). BACKGROUND

[0002] Semiconductor devices are commonly found in modern electronic products. Semiconductor devices perform a wide range of functions, such as signal processing, high-speed computing, transmitting and receiving electromagnetic signals, controlling electronic

[0003] Fan-out wafer level packaging (FO-WLP) is a common packaging topology for today's semiconductor manufacturing. FO-WLP devices typically have under bump metallization (UBM) pads formed thereover. Solder is printed onto the UBM pads using a stencil. Solder paste is disposed into the stencil openings over the UBM pads and then reflowed to form solder bumps.

[0004] A common problem with stencil printing of solder paste is that the solder paste can be deposited into the stencil openings with voids. Solder paste voids result in inconsistent bump heights due to insufficient volume of solder and irregular shaped bumps that do not completely cover their respective UBM pads. These problems are most prevalent with non-circular UBM pads. Rectangular or oblong UBM pads will typically have the problem that their perimeter or ends are not completely covered in solder after bump reflow. Additionally, within a group of pads on a FO-WLP, the biggest problem is the outermost pads at the edge of the grouping. Therefore, there is a need for improved bump forming devices and methods for FO-WLP devices. BRIEF DESCRIPTION OF DRAWINGS

[0005] Figures la- lc illustrates a semiconductor wafer having multiple semiconductor dies separated by saw lanes;

[0006] Figures 2a- 2j illustrates a process of forming a fan-out wafer level package;

[0007] Figures 3a- 3h illustrates a stencil opening having a larger footprint relative to an underlying under bump metallization pad;

[0008] Figures 4a- 4h illustrates a stencil opening for a rectangular UBM pad;

[0009] Figures 5a- 5ddifferent versions of film opening sizes are illustrated for different cases; and

[0010] Figure 6a and 6b illustrating integration of fan-out wafer level packaging into electronic devices. DETAILED DESCRIPTION

[0011] In the following description of the drawings, the application is described in one or more embodiments where like numbers describe similar elements throughout. The application is described with regard to the best mode contemplated for carrying out the application and to include various alternatives, modifications, and equivalents. The features illustrated in the drawings are not necessarily drawn to scale. Elements having the same reference numbers in different figures have similar functions. The term "semiconductor die" as used herein refers to both the singular and the plural of the term, and thus can refer to both a single semiconductor device and multiple semiconductor devices.

[0012] Semiconductor devices are generally manufactured using two complex manufacturing processes: front-end manufacturing and back-end manufacturing. Front-end manufacturing involves creating multiple dies on the surface of a semiconductor wafer. Each die on the wafer contains active and passive electrical components that are electrically connected to form functional electrical circuits. Active electrical components such as transistors and diodes have the ability to control the flow of current. Passive electrical components such as resistors, capacitors, and inductors have a

[0013] Back-end manufacturing refers to the process of cutting or singulating the finished wafer into individual semiconductor dies and encasing the semiconductor dies into packages. To singulate the wafer, the wafer is scored and broken along non-functional areas of the wafer, called saw streets or scribe lines. The wafer is then singulated using a laser cutting tool or saw blade. After singulation, the individual semiconductor dies are placed in packages that provide structural support, electrical interconnects, and environmental isolation. The contact pads formed on the semiconductor dies are then connected to the package pins or contacts using wirebonds. The electrical connections can be made with conductive layers, bumps, pillars, conductive paste, or solder. A sealant or other encapsulation material is deposited over the semiconductor die to provide environmental isolation and to protect the die from mechanical damage. The finished semiconductor packages are then inserted into electronic systems and the functionality of the semiconductor devices is made available to other system components.

[0014] Figure laA semiconductor wafer 100 is shown having a base substrate material 102 such as silicon, germanium, aluminum phosphide, aluminum arsenide, gallium arsenide, gallium nitride, indium phosphide, silicon carbide, or other bulk material used for structural support. A plurality of semiconductor dies or electrical components 104 are formed on the wafer 100 separated by passive inter-die wafer regions or streets 106. The streets 106 provide cutting areas to singulate the semiconductor wafer 100 into individual semiconductor dies 104. In one embodiment, the semiconductor wafer 100 has a width or diameter of 100-450 millimeters (mm).

[0015] Figure lb A cross-sectional view of a portion of the semiconductor wafer 100 is shown. Each semiconductor die 104 has a back surface or non-active surface 108 and an active surface 110 containing analog or digital circuitry implemented as active devices, passive devices, conductive layers, and dielectric layers formed over or within the die and electrically interconnected according to the electrical design and function of the die. For example, the circuitry can include one or more transistors, diodes, and other circuit elements formed within the active surface 110 to implement analog circuitry or digital circuitry such as a digital signal processor (DSP), an application specific integrated circuit (ASIC), a memory, a power device, or other signal processing circuitry. The semiconductor die 104 can also contain integrated passive devices (IPD) for RF signal processing such as inductors, capacitors, and resistors.

[0016] A conductive layer 112 is formed over the active surface 110 using physical vapor deposition (PVD), chemical vapor deposition (CVD), electrolytic plating, electroless plating processes, or other suitable metal deposition processes. The conductive layer 112 can be one or more layers of aluminum (Al), copper (Cu), tin (Sn), nickel (Ni), gold (Au), silver (Ag), or other suitable conductive material. Any of the conductive layers mentioned below can be formed using any of the materials or methods described for the conductive layer 112.

[0017] The conductive layer 112 operates as a contact pad electrically connected to the circuitry on the active surface 110. Prior to singulation and packaging, the incoming wafer 100 is backgrinded to reduce the thickness of the wafer.

[0018] In Figure lc the semiconductor wafer 100 is singulated into individual semiconductor dies 104 using a saw blade or laser cutting tool 118 through the streets 106. The individual semiconductor dies 104 can be inspected and electrically tested for identification of known good dies (KGD) or known good units (KGU) after singulation.

[0019] Figures 2a- 2j A process of forming fan-out wafer-level semiconductor packages (FO-WLPs) is illustrated. In Figure 2a In the middle, multiple semiconductor dies 104 are picked and placed onto a carrier 120, which contains a sacrificial base material such as silicon, polymer, beryllium oxide, glass, or other suitable low-cost rigid material for structural support. An interface layer or double-sided tape 122 is formed or disposed over the carrier 120 as a temporary bonding film, thermal release layer, or UV release layer. For simplicity of illustration, the formed packages each include a single semiconductor die 104 and do not include other components. In other embodiments, more complex packages can be formed with multiple semiconductor dies, additional discrete active or passive components, or any other desired electrical components. While the carrier 120 is illustrated as large enough to fabricate three devices together, a typical carrier can handle hundreds or thousands of devices at a time. The semiconductor dies 104 are disposed with active surfaces 110 oriented toward the carrier 120.

[0020] In the middle, multiple semiconductor dies 104 are picked and placed onto a carrier 120, which contains a sacrificial base material such as silicon, polymer, beryllium oxide, glass, or other suitable low-cost rigid material for structural support. An interface layer or double-sided tape 122 is formed or disposed over the carrier 120 as a temporary bonding film, thermal release layer, or UV release layer. For simplicity of illustration, the formed packages each include a single semiconductor die 104 and do not include other components. In other embodiments, more complex packages can be formed with multiple semiconductor dies, additional discrete active or passive components, or any other desired electrical components. While the carrier 120 is illustrated as large enough to fabricate three devices together, a typical carrier can handle hundreds or thousands of devices at a time. The semiconductor dies 104 are disposed with active surfaces 110 oriented toward the carrier 120. Figure 2b In the middle, a sealant or molding compound 124 is deposited over and around the semiconductor dies 104 using paste printing, compression molding, transfer molding, liquid sealant molding, vacuum lamination, spin coating, or other suitable applicator. The sealant 124 can be a liquid or particulate polymer composite such as epoxy, epoxy acrylate, or another suitable polymer, with or without suitable fillers. The sealant 124 is electrically non-conductive, provides structural support, and environmentally protects the semiconductor devices from external elements and contaminants. The semiconductor dies 104 are disposed with active surfaces 110 oriented toward the carrier 120 so that the active surfaces are exposed from the sealant 124 for subsequent electrical connection to the contact pads 112 once the carrier is removed. The sealant 124 undergoes a post-mold cure (PMC) process.

[0021] The combination of the semiconductor dies 104 and the sealant 124 forms a reconstituted wafer 126. The term "reconstituted wafer" indicates that the reconstituted wafer 126 can be analogous to a semiconductor wafer 100 from which the semiconductor dies 104 were singulated in Figure lc In a sense, the semiconductor wafer 100 has been reconstituted, where the semiconductor dies 104 have been further dispersed to facilitate the formation of fan-out interconnect structures over the active surfaces 110 after deposition.

[0022] Figure 2cThe carrier 120 is removed or debonded by mechanical peeling, thermal release, UV release, or another suitable means. The reconstituted wafer 126 is flipped so that the active surface 110 is oriented upward. After removal of the carrier 120, an automated optical inspection process is optionally performed on the surface of the reconstituted wafer 126, with the active surface 110 exposed, before a fan-out build interconnect structure 128 is formed over the reconstituted wafer 126.

[0023] The interconnect structure 128 is a fan-out interconnect structure, indicating that the final electrical connection footprint formed at the top of the interconnect structure is wider than the contact pads 112 at the bottom of the interconnect structure. The interconnect structure 128, referred to as a build interconnect structure, refers to the manner in which the interconnect structure is formed by successively building insulating and conductive layers over the reconstituted wafer 126 until the desired signal routing is achieved.

[0024] Formation of the interconnect structure 128 begins with formation of an insulating or passivation layer 130 over the reconstituted wafer 126. The passivation layer 130 includes one or more layers of silicon dioxide (Si02), silicon nitride (Si3N4), silicon oxynitride (SiON), tantalum pentoxide (Ta205), aluminum oxide (AI2O3), solder resist, polyimide (PI), photosensitive polyimide (PSPI), benzocyclobutene (BCB), polybenzoxazole (PBO), and other materials with similar insulating and structural properties. The passivation layer 130 can be formed using PVD, CVD, printing, lamination, spin coating, spray coating, sintering, or thermal oxidation. Any of the insulating, passivation, or dielectric layers mentioned above or below can be formed using any of the materials or methods described for the passivation layer 130.

[0025] Openings are formed through the passivation layer 130 to expose the contact pads 112. The openings can be formed by chemical etching, photolithography, mechanical drilling, laser drilling, or any other suitable means. A conductive layer 132 is formed over the passivation layer 130, including the through openings of the passivation layer physically and electrically coupled to the contact pads 112. The conductive layer 132 is formed from any of the materials and any of the methods described above for the contact pads 112. The conductive layer 132 is etched or selectively formed to include conductive traces that fan out from the contact pads 112. The conductive layer 132 also includes contact pads for subsequent conductive layers to be formed to connect thereto.

[0026] In some embodiments, vias are formed in or through the passivation layer 130 beneath the contact pads of the conductive layer 132, such that the contact pads include portions that extend downward toward the reconstituted wafer 126. The protrusions can be a plurality of vias formed as dots, concentric circles, or any other desired shape. The protrusions beneath the conductive layer 132 reduce warpage during fabrication and improve adhesion of the conductive layer 132 to the passivation layer 130. Any desired number of insulating and conductive layers can be interleaved on top of one another as desired to achieve desired signal routing. More layers allow for more complex signal routing, while also complicating the manufacturing process.

[0027] An insulating layer 134 is formed over the passivation layer 130 and the conductive layer 132. The insulating layer 134 is formed using any of the processes and materials described above for the passivation layer 130. As with the insulating layer 130, openings are formed through the insulating layer 134 to expose the contact pads of the conductive layer 132. In some embodiments, each contact pad forms a plurality of openings, or one or more circular openings, providing a discontinuous opening for each contact pad in some cross-sections. In other embodiments, each contact pad forms one large opening. The top surface of the insulating layer 134 with openings exposing the conductive layer 132 optionally undergoes an automated optical inspection process before proceeding further.

[0028] A contact pad or under bump metallization (UBM) pad 140 is formed over the insulating layer 134 and through the openings to physically and electrically couple to the contact pads of the underlying conductive layer 132. The UBM pad 140 is formed using any of the materials and methods described above for the contact pad 112. In some embodiments, the UBM pad 140 is formed with a wetting layer, a barrier layer, and an adhesion layer. The UBM pad 140 can be any desired size and shape, and is formed in any desired distribution pattern across the top surface of the interconnect structure 128. An optional automated optical inspection process is performed at this stage to inspect the UBM pads 140.

[0029] Figure 2d A start of a bump formation process is shown, which uses a stencil 150 disposed over the interconnect structure 128. The stencil 150 is a sheet of polymer, metal, or other material formed by rolling, molding, slicing, or another suitable method. The stencil 150 includes openings 152 over each UBM pad 140 on which a solder bump is to be formed. The openings 152 are formed by mechanical drilling, laser etching, chemical etching, molding, or another suitable process.

[0030] Figure 2eA close-up view of the UBM pads 140 is shown, which illustrates that the stencil opening 152 is larger than the UBM pad in at least one cross-section. Figure 2e The stencil opening 152 in the cross-section illustrated has a larger dimension than the UBM pad 140, which can be a length, a width, a diagonal, or any other direction. In some embodiments, the opening 152 extends beyond the footprint of the UBM pad 140 in only one direction, i.e., only to the left or only to the right in the illustrated cross-section. The larger dimension of the stencil opening 152 results in a gap between the outside surface of the UBM pad 140 and the inside surface of the stencil 150 in the stencil opening 152.

[0031] The opening 152 allows the solder paste 154 to be deposited onto the UBM pad 140, as shown in Figure 2f In one embodiment, the solder paste 154 is placed over each opening 152 or over the entire stencil 150 in excess, and then a rubber squeegee or other type of blade is slid across the stencil to press the solder paste into the openings and remove the excess solder paste.

[0032] Figure 2g is a close-up view of the UBM pads 140, which illustrates that the solder paste 154 fills the openings 152 and extends beyond the footprint of the UBM pads 140. The solder paste 154 can or can not extend down into the lateral gap between the UBM pads 140 and the stencil 150 or completely fill the lateral gap between the UBM pads 140 and the stencil 150. The openings 152 are formed to be slightly larger than the corresponding UBM pads 140 in at least one direction to increase the amount of solder paste 154 that can be deposited into the openings. The slightly larger openings 152 reduce the likelihood that the solder paste 154 will have voids relative to the corresponding UBM pads 140 and also reduce the negative consequences of any remaining voids by allowing additional solder to fill in any voids.

[0033] In Figure 2h the stencil 150 is removed, and the solder paste 154 is reflowed by heating above its melting temperature to form solder bumps 156. Figure 2i A closer view of the UBM pads 140 is shown. The portion of the solder paste 154 that was deposited beyond the footprint of the UBM pads 140 is pulled in by the surface tension of the molten solder to create solder bumps 156 that are formed only on the top surface of the UBM pads. The UBM pads 140 can have a wetting layer only on the top surface to facilitate the solder paste 154 forming bumps 156 only on the top surface of the UBM pads.

[0034] An optional automated optical inspection process is performed at this stage to inspect the solder bumps 156. Because the stencil 150 has openings 152 that are larger than the corresponding UBM pads 140 in at least one direction, the solder bumps 156 are reliably formed over the UBM pads at a consistent height and completely cover all of the UBM pads. The semiconductor package is completed by optionally taping the reconstituted wafer 126 and back-grinding the encapsulant 124 to reduce the thickness of the package, laser marking the encapsulant, removing the tape, and then mounting the reconstituted wafer for singulation.

[0035] The reconstituted wafer 126 is singulated using a saw blade or laser cutting tool 158 in the Figure 2j to separate the reconstituted wafer 126 into individual FO-WLPs 160. After singulation, the FO-WLPs 160 can be picked up and placed into a tape-and-reel or other container for storage and shipping, or mounted directly onto a PCB or substrate of a larger electronic device.

[0036] Figures 3a- 3h Additional details of the UBM pads 140 and stencil openings 152 in various configurations are illustrated. Figures 3a- 3h A plan view of UBM pads 140 and stencil openings 152 over the UBM pads with various stencil opening sizes is shown. The UBM pads 140 have the same size in each example, while the size of the stencil openings 152 varies depending on the positioning of the particular UBM pad 140 relative to other UBM pads. The UBM pads 140 are long-elliptical, with a width illustrated by line 170 and a length illustrated by line 172. Being long-elliptical means that the UBM pads 140 have two semi-circular ends with two linear sides connecting the two semi-circles, thus extending the length 172 to be greater than the width 170. The UBM pads 140 can also be elliptical, rectangular, or other elongated shapes. In one embodiment, the UBM pads 140 have a width of 290 micrometers (pm) and a length of 396 pm.

[0037] The ends of the UBM pads 140 are rounded, with a radius indicated by arrow 174. Because the radius is half of the width 170, the arrow 174 starts at a point that is centered along the width 170. In other embodiments, the UBM pads 140 can be formed with an angular radius that is less than or greater than half of the width 170. In the example dimensions given in the previous paragraph, the angular radius would be 145 pm, half of the width 170, thus resulting in a semi-circle at each end of the UBM pads 140.

[0038] The stencil openings 152 are formed with a length and width that are larger than the corresponding dimensions of the UBM pads 140. Figure 3aA UBM pad 140 is shown with a stencil opening 152a. The stencil opening 152a includes a width 180a that is greater than the width 170 of the UBM pad 140, and a length 182a that is greater than the length 172 of the UBM pad. Having a greater dimension increases the amount of solder paste 154 deposited, and thereby creates a bump 156 that is more uniform in size and shape.

[0039] The opening 152a is a good default size for the stencil opening 152, which can then be customized for each UBM pad 140 depending on the specific positioning of each UBM pad. In one embodiment, the stencil opening 152a has a width 180a of 300 pm, a length 182a of 400 pm, and a corner radius of 100 pm. The default size of the opening 152a is limited by the minimum gap between immediately adjacent stencil openings 152. In one embodiment, the width of the stencil 150 between adjacent stencil openings 152 needs to be maintained at least 50 pm, and the gap between immediately adjacent UBM pads 140 in the width direction is 70 pm. Thus, the immediately adjacent stencil openings 152 can each grow 10 pm towards each other, and still maintain a 50 pm width of the stencil 150 between the openings.

[0040] Having a smaller corner radius than the UBM pad 140 means that even with the same length and width, the stencil opening 152 will be longer in the diagonal direction than the UBM pad. In some embodiments, simply extending the corners of the stencil opening 152 beyond the footprint of the UBM pad 140, even with the same length and width, is sufficient to increase the amount of solder paste 154 deposited, thereby overcoming the problem of solder paste voids.

[0041] Figure 3b A stencil opening 152b is shown disposed over another UBM pad 140. The stencil opening 152b has the same length 182a as the stencil opening 152a, but has been elongated in the width direction, with a width 180b that is wider than the width 180a. In one embodiment, the width 180b is between 312 pm and 320 pm. In another embodiment, the width 180b is specifically 312 pm or 320 pm. The opening 152b is used for UBM pads 140 that have an immediately adjacent UBM pad on both the top and bottom, but no immediately adjacent UBM pad on the left or right, e.g., pads within a string of pads following a linear path vertically. Thus, the opening 152b can be extended both to the left and to the right relative to the opening 152a, but not upward or downward.

[0042] Figure 3cA photoresist opening 152c that is expandable only to the right and not to the left is shown. The photoresist opening 152c is usable with UBM pads 140 formed along the right edge of the group of pads, which have immediate UBM pads on the top, left, and bottom. Thus, the opening 152c is expandable only to the right. The opening 152c has a width 180c that is greater than the width 180a but less than the width 180b. In one embodiment, the width 180c is between 305 pm and 310 pm. In another embodiment, the width 180c is specifically 305 pm or 310 pm. The opening 152c still has the same length 182a as the openings 152a and 152b.

[0043] Figure 3d A photoresist opening 152d that is disposed over another UBM pad 140 is shown. The photoresist opening 152d has the same width 180a as the photoresist opening 152a, but has been lengthened in the length direction, with a length 182b that is greater than the length 182a. In one embodiment, the length 182b is between 412 pm and 420 pm, inclusive. In another embodiment, the length 182b is specifically 412 pm or 420 pm.

[0044] The opening 152d is used for UBM pads 140 that have immediate UBM pads on the left and right, but not on the top or bottom, such as pads within a string of pads that follow a linear path horizontally. Thus, the opening 152d can be expandable both up and down relative to the opening 152a, but not to the left or right.

[0045] Figure 3e A photoresist opening 152e that is expandable only up and not down is shown. The photoresist opening 152e is usable with UBM pads 140 formed along the top edge of the group of pads, which have immediate UBM pads on the left, bottom, and right. Thus, the opening 152e is expandable only up. The opening 152e has a length 182c that is greater than the length 182a but less than the length 182b. In one embodiment, the length 182c is between 405 pm and 410 pm, inclusive. In another embodiment, the length 182c is specifically 405 pm or 410 pm. The opening 152e still has the same width 180a as the opening 152a.

[0046] Figure 3fA photoresist opening 152f is shown disposed over another UBM pad 140. The photoresist opening 152f has both an elongated length 182b and an elongated width 180b relative to the photoresist opening 152a. The opening 152f can be used with isolated UBM pads 140 that do not have a UBM pad immediately adjacent in either direction, and thus are not limited in the extent of available directions of expansion.

[0047] Figure 3g An opening 152g is shown that expands both upward and to the right, but not downward or to the left. The opening 152g can be used with UBM pads 140 at a corner of a pad group that have an immediately adjacent UBM pad in two non-opposite directions. The opening 152g has a width 180c and a length 182c.

[0048] Figure 3h A photoresist opening 152h is shown formed over a UBM pad 140. The photoresist opening 152h expands in three directions, and can be used at the end of a row of pads where the end UBM pad is immediately adjacent to only another pad. The opening 152h expands in only a single direction in the width dimension, but in both directions in the height dimension, i.e., upward, downward, and to the right, but not to the left. The opening 152h has a length 182b and a width 180c. A similar pad opening can be formed by expanding in the left, right, and upward three directions, and has a length 182c and a width 180b.

[0049] Each of the photoresist openings 152a-152h includes the same 100 pm corner radius 184, while simply elongating the length, width, or both in either or both directions. In other embodiments, the corner radius 184 can also be tailored to fine tune the total area of the opening. In some cases, two photoresist openings can have different lengths or widths, but can be compensated for by changing the corner radius, thereby maintaining a common footprint area. Additional specific cases in using the openings 152a-152h are explained below with reference to Figures 4a- 4h

[0050] Figures 3a- 3h FIGS. 15A-15C illustrate the use of the photoresist openings 152a-152h with UBM pads 140 that are not isolated, i.e., have an immediately adjacent UBM pad in one or more directions. Figure 4a ​A similar concept, but with rectangular UBM pads 190 instead of long-oval UBM pads 140. The UBM pads 190 are described as rectangular because the UBM pads include four linear sides connected at right angles that define a width 200 and a length 202. The corners of the UBM pads 190 are rounded, as with UBM pads 140, but the radius 204 is less than half of the width 200 to impart a rectangular shape to the UBM pads 190. In one embodiment, the UBM pads 190 have a length 202 of 353 μιη, a width 200 of 172 μιη, and a corner radius 204 of 50 μιη. The UBM pads 190 can be used in combination with UBM pads 140 on the same FO-WLP device.

[0051] Figure 4b A photoresist opening 192a above the UBM pads 190 is shown. The photoresist opening 192a has a width 210a that is the same as the width 200, and a length 212a that is the same as the length 202. The photoresist opening 192a is rectangular without rounded corners, i.e., the corner radius is zero. In other embodiments, the photoresist opening 192a has a corner radius that is greater than zero and less than the radius 204, so that the photoresist opening still extends beyond the footprint of the UBM pads 190. The opening 192a does not have a length or width that is greater than the UBM pads 190, because the UBM pads 190 are typically more tightly packed than the pads 140, and do not have room to expand between immediately adjacent pads without violating the 50 μιη separation rule. In other embodiments, especially where the distance between pads is greater than the minimum photoresist width required, the rectangular pads can be allowed to have some expansion by default.

[0052] Figure 4c A photoresist opening 192b with an elongated width 210b is shown, elongated on both the left and right sides, but with the same length 212a as the photoresist opening 192a. The photoresist opening 192b is used with pads 190 that have immediately adjacent pads 190 both above and below, but no adjacent pads to the left or right. In one embodiment, the width 210b is 182 μιη. Other widths are used in other embodiments. The photoresist opening 192b still has a corner radius of zero, but can have a greater corner radius.

[0053] Figure 4d An opening 192c with a width 210c that is expanded only to the right relative to the opening 192a is shown. The opening 192c is used when a pad is on the right edge of a group of pads, and the pad has immediately adjacent pads above, below, and to the left. In one embodiment, the width 210c is 177 μιη.

[0054] Figure 4eA stencil opening 192d is shown with a length 212b that expands both up and down, but with a width 210a that is the same as the stencil opening 192a. The stencil opening 192d is used for a UBM pad 190 when the UBM pad has pads immediately adjacent to the left and right, but not up or down. In one embodiment, the length 212b is 363 μιη. Other lengths are used in other embodiments. The stencil opening 192c still has an angular radius of zero, but can have a larger angular radius.

[0055] Figure 4f An opening 192e is shown with a length 212c that only grows up relative to the opening 192a. The opening 192e is used when a pad is on the top edge of a group of pads, and the pad has pads immediately adjacent to the left, right, and down. In one embodiment, the length 210c is 358 μιη.

[0056] Figure 4g A stencil opening 192f is shown with an elongated width 210b and an elongated length 212b. The stencil opening 192f still has an angular radius of zero, but can have a larger angular radius. The opening 192f is usable with a pad 190 that is isolated with no immediately adjacent UBM pads in any direction.

[0057] Figure 4h A stencil opening 192g is shown with an elongated width 210c and length 212c. Compared to the opening 192a, the stencil opening 192g is elongated up and to the right, but not down or to the left. The opening 192g is usable with a pad 190 on a corner of a group with immediately adjacent UBM pads in exactly two non-perpendicular directions.

[0058] Figures 3e- 3g A stencil opening 192h is shown with an elongated width 210c and length 212b. Compared to the stencil opening 192a, the stencil opening 192h is elongated up, down, and to the right. The stencil opening 192h is usable in situations where a UBM pad 190 has one immediately adjacent pad to the left. A similar stencil opening can be formed with a width 210b and length 212c.

[0059] A general rule for any pad including the UBM pads 140 and 190 is that the stencil opening should expand as far as possible within the group of pads while maintaining a minimum stencil width, for example 50 μιη in the example above. Further, pads that are not in the middle of a group of pads can be further expanded in any direction where there is no other immediately adjacent pad. Making the stencil opening 192 have a larger footprint than the corresponding UBM pad 190 reduces the amount of solder paste voids, and the effect of the remaining solder paste voids.

[0060] The same concepts explained with respect to UBM pads 140 and 190 will also apply to UBM pads of any size and shape. In various embodiments, a reduced corner radius, an increased length, an increased width, any combination of the three, or any other method of increasing the footprint size of the via opening relative to the contact pad or UBM pad can be used. As Figures 5a- 5d illustrated in FIG. 4a, a bump is formed on UBM pad 190.

[0061] Figures 3a- 3h FIGS. 1-5d illustrate various groupings of UBM pads 140 and 190 on FO-WLP 160 to show how one particular embodiment can use various sizes of via openings. Although directional words such as horizontal, vertical, above, below, top, bottom, left, and right are used herein to refer to the relative positioning and orientation of UBM pads 140 and 190, the directional words are used only to describe Figure 5a the relative positioning within the particular angles and views illustrated in FIGS. 4a-4h and 5a-5d, and are not meant to be out of the context of those figures.

[0062] Figure 5a FO-WLP 160 in FIG. 4a includes several groupings of UBM pads 140 and 190. Two rows 220 of UBM pads 140 are formed at the top and bottom of the footprint of FO-WLP 160. Rows 220 each include seven or eight UBM pads 140 that are spread across the entire width of FO-WLP 160. UBM pads 140 in rows 220 are relatively spread out, and the corresponding via openings 152 can extend to any reasonable size in any direction. UBM pads 140 are not really groups of UBM pads, but rather multiple individual UBM pads.

[0063] In FIG. 4b, a group 222 of UBM pads 140 is formed on the left side of FO-WLP 160. Group 222 is a rectangle of UBM pads 140 with two arms 223 extending at the top and bottom of the group toward the right side of FO-WLP 160. In Figure 5a FIG. 4c, another group 224 of UBM pads 140 is formed on the right side of FO-WLP 160. Group 224 is rectangular with an additional shorter column of UBM pads 140 formed along the left-hand edge of the group. A rectangular group 226 of UBM pads 190 is centrally formed in the middle of FO-WLP 160. Group 226 consists of five columns of UBM pads 190, with each column offset vertically from the adjacent column. Figure 5a

[0064] Figure 5a ​The UBM pads 140 and 190 in the design depend on where the UBM pads fall within the group to utilize various film opening sizes. Figures 5b- 5d In the diagram, UBM pads 140 and 190 are too small to actually illustrate the different sizes of the liner openings. Figure 5a have Figures 5b- 5d A magnified view of the section to show how different sizes of plate openings are used. Plate openings 152 and 192 are in... Figure 3a It may not be drawn to scale, but it has enlarged dimensions to better illustrate which film openings are elongated in which directions.

[0065] Generally, the UBM pad group has film openings 152a and 192a, said film openings 152a and 192a being from... Figure 4a and Figures 5b- 5d The default size is within the middle of a group in which UBM pads are surrounded by other UBM pads on all four sides. At the edge of the group, the UBM pads have openings 152c, 152e, 192c, and 192e, which extend outward from the group only in one direction perpendicular to the edge of the group where the UBM pads are located. In the corners of a group in which UBM pads have exactly two non-opposite sides and no other adjacent UBM pads, elongated film openings 152g or 192g are used only on one side of the length and width. Film openings 152b, 152d, 192b, and 192d are used in rows of UBM pads such as arm 223, where two opposite sides of the UBM pads have adjacent UBM pads, and the other two opposite sides do not have adjacent UBM pads. Elongated openings 152f and 192f in both length and width directions are used for UBM pads that do not have adjacent UBM pads in any direction, such as those in row 220. Openings 152h or 192h are used at the ends of line 222 and in other cases where UBM pads 150 or 190 have only one adjacent UBM pad.

[0066] When the UBM pads are in multiple groups, the UBM pads within a group are considered to be immediately adjacent to each other, while the gaps between groups cause UBM pads from different groups to not be immediately adjacent to each other. More generally, two UBM pads are considered to be immediately adjacent to each other when the spacing between them is consistent with the spacing between UBM pads within a group of UBM pads on a given package. In some embodiments, two UBM pads are only considered to be immediately adjacent to each other when the spacing between them is consistent or approximately equal to the minimum spacing between UBM pads of a generally sized group of UBM pads on a package in a given direction (e.g., a horizontal or vertical direction). That is, if the minimum spacing between UBM pads of a generally sized group of UBM pads is different in one direction than in another direction, the required spacing for UBM pads to be considered immediately adjacent can be different in those two directions.

[0067] Figure 5a Figures 1-3 illustrate the application of the sizing rules described above to Figure 5b a detailed view. Figure 5b Figures 1-3 illustrate the application of the sizing rules described above to Figure 5b the lower right corner of group 224. Figure 5a The four UBM pads 140a that form a square in the upper left corner of group 224 each have another UBM pad 140 immediately adjacent in all four cardinal directions. As a result, no additional elongated version of the film opening 152a having any length or width is used over the UBM pads 140a. The same is true for all UBM pads 140 that are positioned internally within a group, i.e., those pads that have an immediately adjacent UBM pad in all four cardinal directions. Figure 5b The four UBM pads 140a that form a square in the upper left corner of group 224 each have another UBM pad 140 immediately adjacent in all four cardinal directions. As a result, no additional elongated version of the film opening 152a having any length or width is used over the UBM pads 140a. The same is true for all UBM pads 140 that are positioned internally within a group, i.e., those pads that have an immediately adjacent UBM pad in all four cardinal directions.

[0068] Figure 5bThe bottom right UBM pad 140g is in the corner of group 224 and has two immediately adjacent UBM pads, one above and one to the left. Thus, UBM pad 140g has two non-opposing sides, the bottom and the right, that do not have another immediately adjacent UBM pad and uses a stencil opening 152g that is elongated in both length and width but only in a direction outward of group 224. All corner UBM pads 140 in groups 222 and 224 follow the same rules, including the top and bottom UBM pads in the shorter columns at the leftmost side of group 224. Because group 224 has shorter columns along the left edge, group 224 actually has six corners that follow the corner rules and use stencil opening 152g, two on the leftmost columns, two on the second column from the left, and two on the rightmost columns. Group 222 has only two such corners, on the leftmost columns of group 222, because arm 223 keeps the upper right and lower right corners of group 222 free of UBM pads 140 that have immediately adjacent UBM pads on only two non-opposing sides. In one embodiment, the eight corners of groups 222 and 224 utilize a stencil opening that has a length of 405 μιη and a width of 305 μιη.

[0069] On the right and left edges of groups 222 and 224 that extend vertically in the figure, a stencil opening 152c is used that is larger than opening 152a only in the horizontal direction (perpendicular to the edge on which the UBM pads sit) and only in a direction outward of the group. The particular UBM pads 140 on the left and right edges are those that have another immediately adjacent UBM pad both above and below and only one immediately adjacent UBM pad to the left or right. Figure 5b The left and right edge UBM pads are shown as UBM pads 140c. In one embodiment, the left and right edge UBM pads 140b use a stencil opening that has a length of 400 μιη and a width of 310 μιη.

[0070] On the top and bottom edges of groups 222 and 224 (the edges that extend horizontally in the figure), a stencil opening 152e is used that is longer than opening 152a only in the vertical direction (perpendicular to the edge on which the UBM pads sit) and only in a direction oriented outward from the group. The particular UBM pads 140 on the top and bottom edges are those that have another immediately adjacent UBM pad both to the left and to the right and only one immediately adjacent UBM pad above or below. Figure 5c The top and bottom edge UBM pads are shown as UBM pads 140e. In one embodiment, the top and bottom edge UBM pads 140e use a stencil opening that has a length of 405 μιη and a width of 300 μιη.

[0071] Figure 5c The arms 223 and pairs of UBM pads 140 from row 220 are shown. The UBM pads 140 in the middle of the arms, those with immediately adjacent UBM pads on only two opposite sides, are elongated only in the direction where there is no immediate adjacent UBM pad. For the middle UBM pads 140 of arm 223, the film mask openings 152d are elongated up and down compared to the openings 152a. The middle UBM pads of arm 223 are labeled as UBM pads 140d in Figure 5c In one embodiment, the middle UBM pads 140 of arm 223 have film mask openings 152d with a length of 410 pm and a width of 300 pm.

[0072] At the ends of arm 223, the UBM pads 140h use film mask openings 152h that are elongated in both length and width, but the width is elongated only to the right and not to the left into the arm. For the end UBM pads 140, those with immediate adjacent UBM pads on only one side, a length of 410 pm and a width of 305 pm are used in one embodiment.

[0073] The UBM pads 140 of row 220, labeled as UBM pads 140f in Figure 5d are completely isolated UBM pads, with no immediate adjacent UBM pads in any direction. The UBM pads 140f can use film mask openings 152f that are elongated in all four directions, up, down, left, and right. In one embodiment, the isolated UBM pads 140f use film mask openings with a length of 405 pm and a width of 310 pm.

[0074] Figure 6a A detailed view of group 226 formed by rectangular UBM pads 190 is illustrated. The UBM pads 190 follow essentially the same general rules as the UBM pads 140, but the specific dimensions used are different. The UBM pads 190 along the top and bottom edges use film mask openings 192e that are vertically elongated instead of horizontally elongated, and are elongated only in the direction outside the group. The UBM pads 190 along the left and right edges use film mask openings 192c that are horizontally elongated instead of vertically elongated, and are elongated only in the direction outside the group. The UBM pads 190 in the corners use film mask openings 192g that are both vertically and horizontally elongated, but again are elongated only in the direction outside the group. The inner UBM pads 190, those with immediate adjacent UBM pads in all four cardinal directions, use the default film mask openings 192a without any elongation.

[0075] Using a larger than bump formed UBM pad opening improves paste coverage by reducing solder voids and reducing the negative effects of persistent solder voids. Consistent bump height and fully covered UBM pads can be more reliably achieved. The resulting package has fewer defects, which reduces costs upstream in the supply chain. The problem of solder voids is typically worse at the edges of a UBM pad set, so additional elongation along the edges and corners of a UBM pad set further mitigates the problem that can be particularly bad in those areas.

[0076] Figure 6a and 6b FIG. illustrates integrating the above semiconductor package (e.g., FO-WLP 160) into a larger electronic device 300. Figure 6b FIG. illustrates a partial cross-section of FO-WLP 160 being mounted as a component of electronic device 300 onto a printed circuit board (PCB) or other substrate 302. Bumps 156 are reflowed onto a conductive layer 304 of PCB 302 to physically attach and electrically connect FO-WLP 160 to PCB. In other embodiments, thermal compression or another suitable attachment and connection method is used. In some embodiments, an adhesive or underfill layer is used between FO-WLP 160 and PCB 302. Semiconductor die 104 is electrically coupled to conductive layer 304 through bumps 156 and interconnect structure 128.

[0077] Figure 6b FIG. illustrates electronic device 300 with a chip carrier substrate or PCB 302, where multiple semiconductor packages are disposed on a surface of PCB 302, including FO-WLP 160. Depending on the application, electronic device 300 can have one type of semiconductor package or multiple types of semiconductor packages.

[0078] Electronic device 300 can be a stand-alone system that uses semiconductor packages to perform one or more electrical functions. Alternatively, electronic device 300 can be a subassembly of a larger system. For example, electronic device 300 can be part of a tablet computer, cellular telephone, digital camera, communication system, or other electronic device. Alternatively, electronic device 300 can be a graphics card, network interface card, or other signal processing card that can be inserted into a computer. The semiconductor packages can include microprocessors, memory, ASICs, logic circuits, analog circuits, RF circuits, discrete devices, or other semiconductor dies or electrical components. Miniaturization and weight reduction are essential for products to be marketable. Distances between semiconductor devices can be reduced to achieve higher densities. PCB 302 can have a more irregular shape to conveniently accommodate more ergonomic and smaller device housings.

[0079] In ​In general, the PCB 302 provides a universal substrate for structural support and electrical interconnection of semiconductor packages disposed on the PCB. Conductive signal traces 304 are formed on the surface of the PCB 302 or within layers of the PCB 302 using evaporation, electroplating, electroless plating, screen printing, or other suitable metal deposition processes. The signal traces 304 provide electrical communication between each of the semiconductor packages, mounted components, and other external system components. The traces 304 also provide power and ground connections to each of the semiconductor packages.

[0080] In some embodiments, the semiconductor device has two levels of packaging. The first level of packaging is a technique for mechanically and electrically attaching a semiconductor die to an intermediate substrate. The second level of packaging involves mechanically and electrically attaching the intermediate substrate to a PCB. In other embodiments, the semiconductor device can have only a first level of packaging, in which the die is mechanically and electrically disposed directly on the PCB.

[0081] For illustrative purposes, several types of first level packages are shown on the PCB 302, including wire bond packages 346 and flip chips 348. Additionally, several types of second level packages are shown disposed on the PCB 302, including ball grid array (BGA) 350, bump chip carrier (BCC) 352, land grid array (LGA) 356, multi-chip module (MCM) or SIP module 358, quad flat no-lead package (QFN) 360, quad flat package 362, and embedded wafer level ball grid array (eWLB) 364. In one embodiment, the eWLB 364 is a fan-out wafer level package (Fo-WLP) or a fan-in wafer level package (Fi-WLP).

[0082] Depending on system requirements, any combination of semiconductor packages configured with any combination of first and second level packaging styles, and other electrical components, can be connected to the PCB 302. In some embodiments, the electronic device 300 includes a single attached semiconductor package, while other embodiments require multiple interconnected packages. By combining one or more semiconductor packages on a single substrate, manufacturers can incorporate pre-fabricated components into electronic devices and systems. Because the semiconductor packages include complex functionality, less expensive components and streamlined manufacturing processes can be used to manufacture the electronic devices. The resulting devices are less likely to fail and are less expensive to manufacture, resulting in lower costs to the consumer.

[0083] While one or more embodiments of the application have been illustrated and described, it will be appreciated that those skilled in the art will be able to devise modifications and adaptations of the embodiments without departing from the scope of the application as set forth in the following claims. While FO-WLP is specifically disclosed, the UBM pad bump formation method disclosed above can also be used with other types of semiconductor packages. For all distance measurements provided herein, the length, width, or other measurement can be any suitable value.

Claims

1. A method of fabricating a semiconductor device, comprising: providing a semiconductor die; depositing an encapsulant over the semiconductor die; forming an interconnect structure over the semiconductor die and the encapsulant; forming a first under bump metallization (UBM) pad over the interconnect structure, wherein the first UBM pad is non-circular; disposing a stencil over the first UBM pad, wherein the stencil includes a first opening over the first UBM pad, wherein a footprint of the first opening is larger than a footprint of the first UBM pad; depositing a solder paste in the first opening; and reflowing the solder paste to form a solder bump on the UBM pad.

2. The method of claim 1, further comprising forming a set of UBM pads over the interconnect structure, wherein: the first UBM pad is formed on an edge of the set of UBM pads; a second UBM pad is formed inside the set of UBM pads; the stencil includes a second opening over the second UBM pad; and the first opening of the stencil is larger than the second opening by being elongated in a direction perpendicular to the edge of the set of UBM pads.

3. The method of claim 1, further comprising forming a set of UBM pads over the interconnect structure, wherein: the first UBM pad is formed at a corner of the set of UBM pads; a second UBM pad is formed inside the set of UBM pads; the stencil includes a second opening over the second UBM pad; and the first opening of the stencil is larger than the second opening by being elongated in both length and width.

4. The method of claim 1, wherein the first UBM pad is oblong or rectangular in shape.

5. The method of claim 1, wherein a length and a width of the first opening of the stencil are equal to a length and a width of the first UBM pad, and a corner radius of the first opening is smaller than a corner radius of the first UBM pad.

6. The method of claim 1, wherein the first opening includes a length or a width that is larger than a parallel dimension of the first UBM pad.

7. A method of fabricating a semiconductor device, comprising: providing a semiconductor die; depositing an encapsulant over the semiconductor die; forming an interconnect structure over the semiconductor die and the encapsulant; forming a first under bump metallization (UBM) pad over the interconnect structure; and disposing a stencil over the first UBM pad, wherein the stencil includes a first opening over the first UBM pad, wherein a footprint of the first opening is larger than a footprint of the first UBM pad.

8. The method of claim 7, further comprising forming a set of UBM pads over the interconnect structure, wherein: the first UBM pad is formed on an edge of the set of UBM pads; a second UBM pad is formed inside the set of UBM pads; the stencil includes a second opening over the second UBM pad; and the first opening of the stencil is larger than the second opening by being elongated in a direction perpendicular to the edge of the set of UBM pads.

9. The method of claim 7, further comprising forming a set of UBM pads over the interconnect structure, wherein: the first UBM pad is formed at a corner of the set of UBM pads; a second UBM pad is formed inside the set of UBM pads; the stencil includes a second opening over the second UBM pad; and the first opening of the stencil is larger than the second opening by being elongated in both length and width. ​ ​ ​ The film includes a second opening over the second UBM pad; and The first opening of the film is larger than the second opening by being elongated in both length and width.

10. The method of claim 7, wherein the first opening includes a length or width that is larger than a parallel dimension of the first UBM pad.

11. A semiconductor device, comprising: a first under bump metallization (UBM) pad; and a film disposed over the first UBM pad, wherein the film includes a first opening over the first UBM pad, wherein a footprint of the first opening is larger than a footprint of the first UBM pad.

12. The semiconductor device of claim 11, further comprising a UBM pad group, wherein: the first UBM pad is formed on an edge of the UBM pad group; the second UBM pad is formed inside the UBM pad group; the film includes a second opening over the second UBM pad; and the first opening of the film is larger than the second opening by being elongated in a direction perpendicular to the edge of the UBM pad group.

13. The semiconductor device of claim 11, further comprising a UBM pad group, wherein: the first UBM pad is formed at a corner of the UBM pad group; the second UBM pad is formed inside the UBM pad group; the film includes a second opening over the second UBM pad; and the first opening of the film is larger than the second opening by being elongated in both length and width.

14. The semiconductor device of claim 11, wherein a length and a width of the first opening of the film are equal to a length and a width of the first UBM pad, and a corner radius of the first opening is smaller than a corner radius of the first UBM pad.

15. The semiconductor device of claim 11, wherein the first opening includes a length or width that is larger than a parallel dimension of the first UBM pad.

Images