Semiconductor device manufacturing: process – Packaging or treatment of packaged semiconductor – Assembly of plural semiconductive substrates each possessing...
Reexamination Certificate
1999-10-15
2001-09-18
Niebling, John F. (Department: 2812)
Semiconductor device manufacturing: process
Packaging or treatment of packaged semiconductor
Assembly of plural semiconductive substrates each possessing...
C438S107000, C438S109000, C438S110000, C438S126000, C438S127000, C438S118000
Reexamination Certificate
active
06291267
ABSTRACT:
TECHNICAL FIELD
The present invention relates generally to semiconductor modules and, more specifically, to a structure and process for improved underfill of chips in semiconductor modules that have the circuit side of a larger chip connected via solder bumps to both the circuit side of a smaller chip and a substrate.
BACKGROUND OF THE INVENTION
Many semiconductor modules use Controlled Collapse Chip Connection (C4) solder balls or bumps to connect the circuitry side of the chip to the substrate. Chips using C4 technology are commonly referred to as either “C4 chips” or “flip chips.” For flip chips designed to be wire bonded to a substrate, the C4 pattern is typically around the perimeter of the chip. For flip chips requiring greater numbers of interconnections, and that have been designed for C4 connections, a fully or partially populated array of C4 connections is used. In modules that are not hermetic, and that use C4 connections, an epoxy underfill is typically used. The underfill serves two purposes: (1) it provides environmental protection of the chip circuitry and the solder connections, and (2) it carries some of the shear loading between the chip and the carrier, thus extending the fatigue resistance of the C4 connections.
For example,
FIG. 1
shows a conventional semiconductor module assembly
8
comprising flip chip
10
. Flip chip
10
is mounted to substrate
14
using C4 solder balls
12
. This mounting is usually accomplished by use of an automated placement tool. A thin layer of flux (not shown) is usually applied to either the top surface of substrate
14
or to solder balls
12
. After placement of flip chip
10
on substrate
14
, the assembly is typically run through a reflow furnace in which oxygen levels are controlled to very low levels. The C4 solder balls
12
are typically a lead-tin alloy, such as 97% lead and 3% tin or 37% lead and 63% tin (by weight). Substrate
14
is typically ceramic, but can be an organic substrate, or may comprise any substrate materials known in the art. Substrate
14
has input-output connections
18
such as Pin Grid Array (PGA), Ball Grid Array (BGA), Column Grid Array (CGA), or Land Grid Array (LGA) connections.
Although the array of C4 solder balls
12
is typically well-aligned to a corresponding array of solder-wettable pads (not shown) on the top surface of substrate
14
by the placement tool, the surface tension of the reflowed solder balls automatically re-aligns the two arrays during the solder reflow step. During cooling, C4 solder balls
12
solidify and rigidly attach flip chip
10
to substrate
14
. Such attachment provides both a mechanical and an electrical connection between the flip chip
10
and the substrate
14
.
Underfill
16
is typically applied to preheated module assembly
8
by an automated dispense tool (not shown). The module assembly
8
is typically preheated to effectively reduce the viscosity of underfill
16
, thereby reducing the time required for and improving the effectiveness of the underfill process. Underfill
16
is typically applied along a single edge of flip chip
10
and then allowed to flow by capillary action to completely fill the spaces between flip chip
10
and substrate
14
, thus surrounding C4 solder balls
12
.
Depending on the size of flip chip
10
and the viscosity of underfill
16
, a single dispense pass may be sufficient. Multiple passes may be required, however, with sufficient delay allowed between each pass to allow the underfill
16
to completely flow under the flip chip
10
. Generally, the presence of a fillet
13
of underfill
16
around the complete perimeter of flip chip
10
indicates that enough underfill
16
has been applied. Then, the underfill
16
is cured by any of several mechanisms known in the art, such as exposure to heat, ultraviolet light, or microwave energy.
Although it is not uncommon for wirebond chips to be mounted in cavities, flip chips have typically been joined to substrates without a need for cavities. In a recent effort to bring the function of separate chips as close together as possible on a package, some modules have been designed with a large chip attached using C4 technology to both a substrate and to a smaller chip. Referring now to
FIG. 2
, there is shown a module assembly
19
comprising a larger chip
20
with a smaller chip
22
attached to the larger chip
20
. The stacked chip structure
23
is joined to a substrate
14
′.
The C4 solder balls
12
around the perimeter of larger chip
20
are attached to substrate
14
′; the C4 solder balls
12
′ near the middle of the larger chip
20
are attached to smaller chip
22
. Solder balls
12
and
12
′ may comprise the same alloy, or solder balls
12
′ between chips
20
and
22
may be of a higher melting temperature alloy than solder balls
12
between larger chip
20
and substrate
14
′ so that solder balls
12
′ do not reflow when larger chip
20
is joined to substrate
14
′. To prevent interference between smaller chip
22
and substrate
14
′; there is a shallow cavity
24
in the substrate
14
′ that can accommodate the smaller chip
22
.
The configuration of stacked chip structure
23
within cavity
24
presents some difficulties, however, in providing underfill (not shown) by conventional underfill techniques. Although underfill applied along the edge of larger chip
20
flows under larger chip
20
and is drawn under and around the perimeter of larger chip
20
by capillary action, capillary action is insufficient to draw underfill into cavity
24
or into the space
25
between chips
20
and
22
. Thus, although the underfill may fill the thin gaps
26
between larger chip
20
and substrate
14
′, it does not provide the desired coverage in cavity
24
or spaces
25
. As a result, air trapped in those regions may compromise the integrity of module assembly
19
if the trapped air expands during subsequent heating steps typically used to cure the underfill.
U.S. Pat. No. 5,760,478 issued to Bozso et al. describes a system comprising two flip chips connected face-to-face by conventional solder connections, with one or both of the chips having a chip-to-substrate connection for power and signal. For instance, a low-power device, such as a Dynamic Random Access Memory (DRAM) chip, may be attached directly to a higher-power logic chip, with the logic chip also connected to a substrate. In the preferred embodiment, both the chip-to-chip connections and the chip-to-substrate connections are solder connections. The low-power chip is smaller than the higher-power chip and fits under the larger chip within a cavity in the substrate, as is shown in
FIG. 2
by Bozso et al. This type of structure presents difficulties in providing underfill by conventional techniques, as described above. Thus,
Thus, there is a need in the art for a process and module structure that enables underfill of such stacked chip structures while avoiding the problems of the prior art.
SUMMARY OF THE INVENTION
To meet this and other needs, and in view of its purposes, the present invention provides a chip-under-chip module comprising a first larger chip, a second smaller chip mounted under the first larger chip, a substrate having a top surface to which the first larger chip is mounted, such as with C4 solder balls, a cavity into which the second smaller chip fits when the first larger chip is mounted on the top surface, and an access channel connecting the cavity to the top surface. The module further comprises underfill, such as a heat-curable or radiation-curable resin, disposed under the first larger chip between the first larger chip and the substrate, between the first and second chips, within the cavity, and within the access channel. The module is preferably produced by a process of dispensing the underfill through the access channel.
The channel may have a constant depth within the substrate or may have a sloping depth from the top surface to the cavity. The channel may extend to an edge of the substrate, or may exten
Dore Kevin A.
Edwards David L.
Blecker, Esq. Ira D.
International Business Machines - Corporation
Niebling John F.
Ratner & Prestia
Zarneke David A
LandOfFree
Process for underfilling chip-under-chip semiconductor modules does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Process for underfilling chip-under-chip semiconductor modules, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Process for underfilling chip-under-chip semiconductor modules will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2483078