Electrodeposition apparatus with virtual anode

Abstract

An electrochemical etching step in a semiconductor device fabrication process increases the radius of curvature of edges of metal lines deposited on the semiconductor device. The metal lines are fabricated by forming a mask, electrodepositing the metal, and removing the mask, and the electro-chemical etching step in performed subsequently. The increased radius of curvature of the metal lines simplifies subsequent planarization and decreases line-to-line capacitance, thereby enhancing device performance. In an apparatus for performing the fabrication process, wires sown into a gasket which secures the semiconductor wafer and prevents electrolyte leakage, allows the gasket to function also as a component of the cathode. A more uniform metal deposition is created by a virtual anode, i.e., a metal plate having an aperture and being located between the anode and the cathode.

Description

BACKGROUND OF THE INVENTIONThe present invention relates to a selective metal electrodeposition process and apparatus.Following is a list of the most important requirements for metallization processes for advanced semiconductor processing:low metal resistivity (comparable to or less than that of aluminum)low contact resistance to the active device area (a requirement of the diffusion barrier and not the current carrying metal)low contact resistance to previous and subsequent metallization stepsexcellent contact and via step coverage capabilityacceptable film morphology, adhesion and stress propertiesdecreased line-to-line capacitanceimproved planarization processingcompatibility with planarization processingcompatibility with other processing (such as salicides and dielectric depositions)competitive economics and throughput considerationsexcellent reliability (electromigration and corrosion concerns)environmentally responsible processCOMPARISON WITH CURRENT AND PROPOSED METALLIZATIO...

AI Technical Summary

Benefits of technology

Briefly, in one preferred embodiment, the present invention is directed toward a process for electrodeposition of metal such as copper, silver or gold into a semiconductor wafer having an active layer such as silicon. The process completes the steps of a diffusion barrier deposition process to provide an electrically conducting layer having contacts and vias on said wafer and to prevent metal diffusion to said silicon, a nucleation layer process to provide a sufficient adhesion surface for the electrodeposited metal, an inert metal mask process to place a first metal mask layer onto said wafer, a selective electrodeposition process to selectively electrodeposit said metal onto said conducting layer of said semiconductor wafer while simultaneously completely filling a contact or via without a standard resist metal etch, a photoresist removal step, and an electrochemical etch process to decrease line-to-line capacitance and simplify planarization processing.

Problems solved by technology

low contact resistance to previous and subsequent metallization steps

These voids can cause an increase in line resistance and ultimately the opening of a line (open circuit).

This directly affects the ability of the metal to carry current into and out of the contacts and vias.

Poor step coverage may lead to the failure (open circuit) of the metal in the hole.

Subsequent processing (planarization and stacking vias on top of contacts) is also complicated by poor step coverage.

They are only partial solutions and introduce other problems such as complicating the metal etch step and increasing the sheet resistance of the metal.

Solving the step coverage problem with standard sputtering techniques has also convincingly failed.

The ability to alloy the aluminum using this technique has not yet been demonstrated nor has its production worthiness.

Hot sputtered aluminum is of questionable production worthiness and electromigration resistance.

Reflowed aluminum requires processing with an expensive laser ($700,000) after standard sputtering.

Blanket LPCVD tungsten is presently economically unattractive as well as exhibiting a much larger resistivity than that of aluminum.

Some economic improvements can be expected, but its large resistance requires thick films which results in large fringing (line to line) capacitances which degrade device performance.

Thick lines and lines with sharp edges and corners also complicate subsequent planarization processing.

Additionally, plasma etching of tungsten has not proven to be as production worthy a process as that of aluminum etching.

Selective LPCVD tungsten plugs with sputtered aluminum addresses the step coverage problem, but has throughput and economic problems, leakage current concerns when used as the first metallization process and electromigration resistance concerns.

Step coverage should be excellent, but electromigration resistance is still an issue.

Images