Method and structure for thru-mask contact electrodeposition

Details Method and structure for thru-mask contact electrodeposition Method and structure for thru-mask contact electrodeposition

2002-10-28

2004-11-09

Whitehead, Jr., Carl (Department: 2813)

Semiconductor device manufacturing: process

Chemical etching

Combined with the removal of material by nonchemical means

C438S669000, C438S672000

Reexamination Certificate

active

06815354

ABSTRACT:

FIELD
The present invention generally relates to semiconductor integrated circuit technology and, more particularly, to a process for electroplating on a workpiece.
BACKGROUND
In IC fabrication, formation of wiring networks on the ICs often presents technological problems. In particular, certain applications require formation of both narrow wires or other structures and wide wires, electrodes, bond pads or other structures, with the narrow and wide structures adjacent to one another and as closely spaced as allowed. Such conductive networks require forming narrow and large gaps through a mask layer, which is typically a photoresist layer, coated on a substrate and filling them with a conductive material such as copper.
As exemplified in
FIG. 1
, in one conventional process, an upper surface
9
of a silicon or device substrate
10
is first coated with a barrier layer
12
and then with a copper seed layer
14
. It should be noted that the substrate
10
typically contains more layers of metal interconnect structures, insulators and devices below the upper surface
9
, which are not shown in detail in FIG.
1
. The barrier layer
12
helps adhesion of copper to the upper surface
9
and also serves as a diffusion barrier for copper. The barrier layer
12
is typically made of a single or multilayer films, such as Ta, TaN, Cr, TiW, NiP, TaN/Ta, WC, WN, WNC, Al/Ni—V, carbon and nitrogen doped tungsten or even, Ti/Cu/Ni layer and their various combinations. In Ti/Cu/Ni, Ti is for adhesion and sealing, Cu for current conduction and Ni serves as the diffusion barrier and wettable layer. In some other applications a dewetting layer or surface or region may also be incorporated. The seed layer
14
may be a thin film copper, or copper alloy, silver or silver alloy layer that provides a base layer on which better nucleation and growth of the subsequently deposited metal layer can take place. A resist layer
16
is formed on the seed layer
14
and defined and etched to form various features by lithographic processes. The resist material may be soft or hard baked organic materials, it could be inorganic materials such as silicon oxide, alumina or even sapphire. The resist material may also be a low or high dielectric constant material. In this invention, a resist material is a material that copper or any metal of interest will not substantially nucleate on without a seed layer coating. In
FIGS. 1 and 2A
two exemplary features etched in the resist, a narrow line or via
18
and a wider structure such as a trench or a via
20
are shown with the seed layer in exposed portions
24
on the bottom of these features. These features require filling with copper.
Filling up such features with copper with the conventional electroplating processes from additive or additive free baths, present problems. As shown in
FIG. 2A
, as copper layer
22
grows on exposed portions
24
of the seed layer to fill the features from bottom to top, wider features or trenches
20
may fill up before narrower features or lines
18
. As seen in
FIG. 2A
, at time t
1
, although the wide trench
20
is already filled, the narrow line
18
is only partially filled. As the electroplating time is extended until time t
2
in order to fill the narrow line
18
, the wide trench
20
is overfilled and a copper dome
26
may form on the wide trench
20
due to the extra electroplating time.
Such filling discrepancies between features of various shapes and sizes may be due to various reasons such as differences in mass transfer of the plating electrolyte into features with different aspect ratios, the location of the features with respect to high current density regions such as contacts, and location of different size features with respect to each other. For example, apart from concerns about nonuniform electric field distribution across the workpiece, mass transfer is not as efficient in small features as in large features. Therefore, current densities tend to be higher in larger features. If a small and large feature are side by side and their size difference is high, the large feature may steal current density from the small one resulting in the phenomenon depicted in FIG.
2
A.
Further, in IC design, if the larger feature density is higher on certain regions of the IC, the severity of this problem becomes higher in such regions. As is apparent, taking into account such factors limit the most efficient circuit layout in terms of electrical circuit performance. This is particularly true since rigid design rules also need to be followed. For instance, all vias may be essentially the same diameter according to the specified design rules, and their minimum spacing from each other may also be specified by the design rules, such as being at least ¼ of their diameter to prevent shorting and the like. Such design rules, while necessary using conventional process technologies, nonetheless result in a loss of flexibility in wiring design and layout, thereby causing losses in wiring densities, but also resulting in performance penalties and the inability to integrate certain functionalities in a package or wafer. For example, it may result in passive and active electrical elements, or the leads thereto, not being capable of being integrated in close proximity to each other. One aspect of this reduced flexibility is that many wide features cannot be fabricated in very close to very narrow lines, as mentioned above. Many very fine lines or features, for example those less than 5 microns in width, cannot be readily attached to very wide lines, vias or features larger than, for example, 20 microns, without incurring severe cost penalties.
If the various features, regardless of dimensions, shape, locations or relative locations on a workpiece could be filled up in a uniform manner at the same rate, many of the problems discussed above would be overcome.
As discussed above, conventional copper deposition methods yield a copper structure that is very non-uniform, i.e., some features are overfilled at the expense of others. The overfilled extra copper needs to be subsequently removed using material removal techniques such as chemical mechanical polishing (CMP) process. These material removal techniques are complicated and costly, and even more costly, especially when a thick layer of the conductive material is to be planarized.
Conventionally, after the metal deposition, the resist layer
16
, and the seed layer
14
and barrier layer portions disposed below the resist layer are removed to isolate the various electrical structures of interest, as shown in FIG.
2
B. The large disparity in the height of large structure and smaller one is obvious in the absence of the masking resist layer. The height H
1
of the copper structure
22
a
is significantly greater than H
2
of the copper structure
22
b
. After seed layer and barrier layer removal, the difference between the small and the large features becomes even more significant. The seed layer thickness is nearly insignificant when compared to the width X
1
and height H
1
of structure
22
a
. Because the removal process can affect all of the conductors, it can affect all structures, not only the seed layer and barrier layer portions being removed. Thus removal of the seed layer portions does not significantly affect the structure
22
a
, in terms of the desired width X
12
and height H
12
in
FIG. 2C
, which shows the structures after the removal of the seed and barrier layer portions. The seed layer portion removal, however, may be catastrophic for finer features such as the structure
22
b
as shown by the significantly reduced size of the structure
22
b
that has been affected by the removal process. Mass transfers and surface area effects cooperate during seed layer removal to produce an undesirable structure, especially with respect to the height H
22
of the structure
22
b
. The large difference ‘&Dgr;H’ between the heights of two structures of differing sizes also represents a severe design rule violation. This is another fundamental reason that limits the choice of p

Affiliated with

Also associated with

Rating

Method and structure for thru-mask contact electrodeposition does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Method and structure for thru-mask contact electrodeposition, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and structure for thru-mask contact electrodeposition will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3361840