The invention relates generally to a process for fabricating an integrated circuit structure, and more specifically to a fabrication process for directly forming nickel monosilicide (NiSi).
In the semiconductor processing art, low resistivity metal silicide regions are commonly formed on silicon-containing features to enable efficient electrical interconnection of components in an electronic device. Silicides are compound materials formed from a chemical reaction between various forms of silicon (e.g., single-crystal or polycrystalline) with a metal. Self-aligned silicides (referred to as salicides) are formed on silicon-containing features, such as transistor gates and source/drain regions, to provide a layer of low resistivity material on the feature.
For example, nickel monosilicide (NiSi) is often used as a contact material in silicon-based fabrication. NiSi has a resistivity of 14-20 μohm-cm and is thus comparable to titanium silicide (TiSi2) and cobalt silicide (CoSi2). Moreover, NiSi has the lowest formation temperature of the three silicides—roughly 350° C. to 750° C. Further, NiSi consumes less silicon (about 1.82 nm of Si is consumed per nm of metal) than the other two compounds. Nickel silicide has three main phases depending on formation temperature (Ni2Si, NiSi, and NiSi2). Nickel monosilicide (NiSi) is the desired phase partially due to its having the lowest resistivity of the three phases.
In a self-aligned silicide processing method, a blanket metal is deposited on exposed portions of silicon-containing features. The metal is then reacted with portions of the features to form silicide regions. Portions of the features that are not exposed, for example, portions covered by a spacer, do not form a silicide region. In this manner, self-aligned silicides are selectively formed on the features without patterning or etching silicide to define low resistivity regions. As discussed above, self-aligned silicides can be formed from metals that include nickel, titanium, cobalt, as well as other metals that react with silicon to form silicides.
With reference to
In
In another conventional prior art process (not shown but similar to
As semiconductor technology advances, smaller feature sizes (i.e., smaller design rules) have become increasingly important. Smaller feature sizes allow an increased density of electronic devices and concomitant increases in execution speeds. However, neither the one-step nor the two-step RTA processes are adequate for silicidation steps at extremely small design rules. For example, the one-step RTA process is particularly troublesome for certain silicide-forming metals, such as nickel. At rapid thermal anneal temperatures ranging from 350° C. to 700° C., the reaction rate between the nickel and silicon is difficult to control, resulting in an excessive formation of nickel silicide. Control of the reaction rate can be especially problematic with metals such as cobalt and titanium. As indicated by
Further, particular metals present certain challenges. For example, the use of titanium in the two-step RTA process to form titanium silicide (TiSi2) in a self-aligned manner is ineffective with smaller semiconductor structures. Neither titanium metals nor titanium alloys fully react with small areas of silicon. Referring to
Cobalt is also used to react with silicon (not shown) to form self-aligned cobalt silicide (CoSi2) regions utilizing a two-step RTA process. However, temperatures at which the first and second anneals are performed are relatively high. For example, the first anneal for cobalt is typically at temperatures ranging from 450° C. to 510° C. The second anneal is at temperatures ranging from 760° C. to 840° C. These high temperatures induce stress on the semiconductor structure and can destroy functionality of the semiconductor device. Additionally, these relatively high temperatures may not be compatible or desirable with either pre-existing components of the device or subsequent fabrication steps. More particularly, these high temperatures may deleteriously diffuse materials of the existing semiconductor device.
Formation of CoSi2 has two additional problems. First, formation of CoSi2 as a silicide has a large silicon consumption rate. The large consumption rate is especially problematic with varying silicon feature sizes (discussed above with reference to
Accordingly, what is needed is a method to control formation rates of silicides to reduce silicide formation in and around the features, reduce interfacial roughness due to the silicide growth, and produce thermally stable and low resistivity silicides.
In an exemplary embodiment, the invention is a method of forming a nickel monosilicide layer on silicon-containing features of an electronic device. The method includes depositing a nickel film over the silicon-containing features where the nickel film is co-deposited with a selected material. The selected material is chosen to have an atomic percentage in a range of about 10% to 25%. The nickel film is then reacted with the underlying silicon-containing features in a single anneal step to directly form the nickel monosilicide layer.
In another exemplary embodiment, the invention is a method of forming a nickel monosilicide layer on silicon-containing features of an electronic device where the method includes depositing a nickel film over the silicon-containing features. The nickel film is co-deposited with a selected material chosen from a group including platinum, palladium, zirconium, and germanium. The selected material has an atomic percentage in a range of about 10% to 15%. A single anneal step of less than about 500° C. is applied to the nickel film to directly form the nickel monosilicide layer.
In another exemplary embodiment, the invention is a method of forming a nickel monosilicide layer on silicon-containing features of an electronic device where the method includes depositing a nickel film over the silicon-containing features. The nickel film is co-deposited with platinum. The platinum is chosen to have an atomic percentage in a range of about 10% to 25%. A single anneal step in a range of 250° C. to 350° C. is applied to the nickel film to directly form the nickel monosilicide layer without first forming any other nickel silicide phase.
In another exemplary embodiment, the invention is a method of forming a nickel monosilicide layer on silicon-containing features of an electronic device where the method includes depositing a nickel film over the silicon-containing features. The nickel film is co-deposited with a selected material chosen from a group including platinum, palladium, zirconium, germanium, tungsten, tantalum, and titanium. The selected material has an atomic percentage in a range of about 10% to 15%. A single anneal step of less than about 500° C. is applied to the nickel film to directly form the nickel monosilicide layer.
As described above, self-aligned silicidation (salicidation) is widely used in integrated circuit fabrication to reduce single-crystal and polycrystalline silicon interconnect and contact resistance values. The nickel monosilicide formation process of the present invention has a sheet resistance value which remains constant for linewidths as small as 30 nm and has a low silicon consumption rate. Unlike the prior art, which typically forms NiSi from a metal-rich Ni2Si phase, NiSi is produced directly. Further, various embodiments include an alloy and a composition for a salicide process based on NiSi. In one embodiment, the alloy is comprised of nickel with a platinum (Pt) concentration of between about 10 atomic percent and 15 atomic percent. In other embodiments, other elements such as palladium, zirconium, germanium, tungsten, tantalum, or titanium are used with Ni in atomic percentages of between about 10% and 25%. (Note that an important distinction is made between atomic percentage and percentage by weight. For example, 15% Pt by weight in 85% Ni by weight corresponds to 5 atomic % Pt in 95 atomic % Ni. Therefore atomic percentages will be used exclusively and designated as “at. %” herein.)
In an exemplary embodiment, one or more NiPt layers are formed over silicon-containing areas of a semiconductor device. The one or more layers may be co-deposited (e.g., co-sputtered) from separate Ni and Pt targets and are formed with 10 at. % to 15 at. % Pt. The separate targets are typically pure Ni and pure Pt. Alternatively, the layers may be co-deposited from a single target comprised of Ni1-xPtx such that a proportion of Pt is produced from 10 at. % to 15 at. %.
Referring to
The substrate 501 may be comprised of various materials known in the semiconductor art. Such materials include silicon (or other group IV semiconducting materials), compound semiconductors (e.g., compounds of elements, especially elements from periodic table Groups III-V and II-VI), quartz photomasks (e.g., with a deposited and annealed polysilicon layer or a deposited/sputtered metal layer over one surface), or other suitable materials. Frequently, the substrate 501 will be selected based upon an intended use of a finalized semiconducting product. For example, a memory cell used as a component in an integrated circuit for a computer may be formed on a silicon wafer. A memory cell used for lightweight applications or flexible circuit applications, such as a cellular telephone or personal data assistant (PDA), may form the memory cell on a polyethyleneterephthalate (PET) substrate deposited with silicon dioxide and polysilicon followed by an excimer laser annealing (ELA) anneal step. For purposes of exemplary embodiments described herein, only the doped silicon-containing regions 503A, and the silicon-containing feature 505A need be comprised at least partially of silicon. In a specific exemplary embodiment, the substrate 501 may be selected to be a silicon wafer. A preferential chemical etch or, alternatively, an in-situ sputter etch may be applied to the substrate 501 prior to any metal deposition steps.
Spacers 507 are formed along sidewalls of the silicon-containing feature 505A. Fabrication of the spacers 507 is known in the art. The spacers 507 are frequently formed from a dielectric material such as a chemical vapor deposition (CVD) deposited silicon dioxide. A blanket metal layer 509 is formed over the semiconductor device 500. The blanket metal layer 509, as described above, may be co-deposited from separate Ni and Pt targets and is formed with 10 at. % to 15 at. % Pt or may be co-deposited from a single target comprised of Ni1-xPtx. In a specific exemplary embodiment, a power density applied to the one or more targets is between two and ten watts/cm2 with an ambient argon partial pressure of between 0.5 to 5 millitorr. The blanket metal layer 509 is formed to a thickness of between 1 nm and 100 nm but may vary depending upon device type, design rules, and other factors which may be readily determined by a skilled artisan.
In
In a specific exemplary embodiment, the RTA step is performed at between 250° C. to 350° C. The RTA step produces partially-consumed doped silicon-containing regions 503B and a silicon-containing feature 505B. However, in other specific exemplary embodiments, temperatures as high as 500° C. may be employed. Temperatures greater than 600° C. (including back-end-of-line processes) are generally not employed primarily for three reasons: (1) the NiSi layer can agglomerate at temperatures around 500° C. to 600° C. causing a discontinuous NiSi layer with Ni islands formed as described above; (2) an enhanced grain growth of silicon due to the higher temperature may lead to an inversion phenomenon resulting in large grains of silicide across polycrystalline silicon; and (3) a high resistivity Ni2Si phase of silicide is formed above about 750° C. These phenomena increase the contact resistance of the film, increase interfacial roughness levels, and decrease the thermal stability of the NiSi film and are therefore unacceptable for advanced semiconductor processing. These high temperature results will occur with any NiSi film. Advantageously, the present invention limits or eliminates such concerns.
Referring now to
In the foregoing specification, the present invention has been described with reference to specific embodiments thereof. It will, however, be evident to a skilled artisan that various modifications and changes can be made thereto without departing from the broader spirit and scope of the invention as set forth in the appended claims. For example, skilled artisans will appreciate that various types of annealing treatments other than RTA may be employed. Additionally, sputtering power densities, partial pressures, film thicknesses, and other fabrication details are merely exemplary and may be changed for a particular device type or fabrication environment as needed and known by one of skill in the art. The specification and drawings are, accordingly, to be regarded in an illustrative rather than a restrictive sense.